England and Wales High Court (Queen's Bench Division) Decisions

IN THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION

		Royal Courts of Justice Strand, London, WC2A 2LL
		19/02/2010

The Hon. Mrs Justice Swift DBE :

The claim

1. This is a claim for damages on behalf of the widow and estate of Leslie Sabin (the deceased), who died on 10 May 2006. It is alleged that, as a result of exposure to asbestos during his employment with the British Railways Board (BRB), the deceased contracted asbestosis from which he died. Proceedings were commenced on 2 October 2007. A Defence was filed, in which the defendant (the successor in title of BRB) admitted the deceased's employment and exposure to asbestos. It was also admitted that such exposure was in breach of the statutory and common law duties of care owed to the deceased by BRB. Damages, subject to liability, have been agreed in the sum of £100,000.
2. The issue in the case relates to causation. It is agreed that, prior to his death, the deceased was suffering from diffuse interstitial pulmonary fibrosis and that it was that condition which caused his death. The claimant's case is that the deceased's fibrosis was caused by exposure to asbestos and was, therefore, the condition known as asbestosis. The defendant's case is that the deceased's fibrosis was not related to his asbestos exposure, but was a condition known as usual interstitial pneumonia (UIP), the cause of which is unknown.

The background

3. All types of asbestos give rise to the risk of asbestosis if a sufficient dose is inhaled. The amphiboles (i.e. crocidolite (blue asbestos), amosite (brown asbestos) and anthophyllite) appear to be more fibrogenic (i.e. liable to cause fibrosis) than chrysotile (white asbestos). However, persons exposed only to commercial chrysotile can develop asbestosis.
4. Persons exposed to asbestos inhale asbestos fibres which are then deposited in the tissue of the lungs. Those asbestos fibres tend to become coated in iron. The coated fibres are known as "asbestos bodies". Some individuals have a poor coating mechanism. As a result, the number of asbestos bodies found in their lung tissue will be fewer than would be expected when compared with the number of asbestos fibres present. The formation of asbestos bodies also varies according to the type and characteristics of the relevant fibres. Crocidolite fibres may form asbestos bodies that are small and difficult to detect in routine histological sections. Chrysotile fibres do not tend to become coated as readily as crocidolite or amosite fibres.
5. A large proportion of members of the public will have had some background, non-occupational, exposure to asbestos and may therefore have some asbestos fibres and/or bodies present in their lung tissue. A small proportion of the population will have had higher levels of exposure resulting from their work. However, only a small proportion of persons who experience occupational exposure to asbestos go on to develop asbestosis. The reason why some people develop the condition while others do not is not fully understood. It is believed to depend on individual susceptibility.
6. The severity of asbestosis present in an individual is classified according to four Grades. Grade 1 or Grade 2 asbestosis is not accompanied by clinical or radiological signs of disease. Diagnosis of Grade 1 or Grade 2 disease can be made only by histological examination of tissue obtained by biopsy in life or at post-mortem. Grade 3 and Grade 4 cases have radiological changes and symptomatic clinical signs of the disease. There is a broad correlation between the severity of disease and the level of asbestos exposure. However, within that spectrum, there are wide variations, reflecting differing individual susceptibilities.
7. In the past, the terms UIP, idiopathic pulmonary fibrosis (IPF) and cryptogenic fibrosing alveolitis (CFA) were used virtually interchangeably to describe fibrosis which was of unknown origin. UIP is now recognised by pathologists as the most common subtype of IPF/CFA.

The history of events

The deceased's exposure to asbestos dust

8. The deceased was employed by BRB as a lorry driver. In the course of his work, he drove loads of hessian sacks full of raw asbestos from the BRB depot at Wigan to the premises of Turner and Newall at nearby Hindley Green. On some days, he would make four or five such deliveries, each consisting of about 100-120 sacks of raw asbestos. His job entailed helping to load the sacks of asbestos onto the wagon and securing them. When he arrived at Turner and Newall's premises, he would drive into a large warehouse and reverse the wagon up to a conveyor. He then had to unload the asbestos onto the conveyor. That process took about an hour. He worked a six-day week. The sacks were very dusty. Sometimes, they broke as they were being loaded onto the conveyor, creating even more dust. The deceased estimated that he spent about 50% of his time delivering asbestos to Turner and Newall. During the remainder of his employment with BRB, he did not have exposure to asbestos. He had no asbestos exposure in any other employment during his working life.

The development of the deceased's lung condition

9. The deceased first attended his general practitioner (GP) in connection with his lung condition in late January 2004. He had a productive cough and, on examination of his chest, fine crepitations were noted. He gave a history of asbestos exposure. A chest x-ray undertaken on 30 January 2004 showed extensive bilateral pulmonary fibrosis. He was referred to a chest physician. Various investigations were then undertaken, including lung function tests which revealed that he had substantial impairment of lung function.
10. In May 2004, the deceased was reviewed by Dr Wolstenholme, consultant physician, who suspected that his condition was asbestos-related. In August 2004, a CT scan showed appearances of diffuse interstitial pulmonary fibrosis with bilateral calcified pleural plaques typical of asbestos exposure. In October 2004, the deceased was told that the diagnosis was thought to be asbestosis. He was given information about claiming compensation. Over the subsequent 18 months, his condition worsened until his death on 10 May 2006.
11. A post-mortem examination was carried out by Dr Mills, consultant pathologist. That examination revealed the presence of fibrosis and bilateral pleural plaques. Examination of lung tissue showed fibrosis with a degree of "honeycombing". There were areas of chronic inflammation. No asbestos bodies or fibres were identified at that time. Dr Mills concluded that the cause of death was respiratory failure caused by lung fibrosis due to asbestos. In the absence of any finding of asbestos bodies or fibres, his conclusion was based solely on the clinical information available to him.
12. Subsequently, lung tissue was sent to the University Hospital, Cardiff (known as Llandough Hospital) for analysis of the asbestos fibre content by electron microscopy. Dr Allen Gibbs, consultant histopathologist, carried out mineral analysis of the lung tissue using transmission electron microscopy with energy dispersive x-ray spectrometry. The purpose of this highly sophisticated technique is to identify the presence of asbestos fibres which would not be visible by light microscopy and to identify the type of fibres present. The number of fibres seen (in millions of fibres per gram of dry lung) was:

Amosite: 1.46 Chrysotile 1.32

Crocidolite: 3.96 Anthophyllite: 0.29

13. Having considered the post-mortem findings and the Llandough Hospital fibre count, Dr Wolstenholme confirmed his opinion that the pulmonary fibrosis from which the claimant had suffered had been caused by exposure to asbestos.

Subsequent histological investigations

14. Subsequently, further histological investigations were carried out by two pathologists of international repute, both specialists in the field of lung disease. Professor Roggli, who was instructed on behalf of the claimant, practises at the Department of Pathology at the Duke University Medical Center, North Carolina, in the United States. He and his team have published widely in connection with their research into the diagnosis of asbestos-related disease. Dr Attanoos, who was instructed by the defendant, is a consultant histopathologist at Llandough Hospital. Clinicians at Llandough Hospital have been working in the field of asbestos-related disease for many years. It is the only centre in the UK which is equipped to carry out analysis of asbestos content in lung tissue by means of electron microscopy.
15. Dr Attanoos received the four blocks of lung tissue which had previously been used by Dr Gibbs for his analysis. Dr Attanoos had four sections (with a total lung section area of about 10 sq cm) cut to a thickness of five microns. He viewed those four sections of tissue on slides by light microscopy (a procedure routinely used by pathologists) in order to examine the morphological features and to search for asbestos bodies. He saw a total of 15 scattered asbestos bodies in the four sections, representing 1.5 asbestos bodies per sq cm of lung tissue.
16. Professor Roggli received the same four blocks of tissue as those examined by Dr Attanoos. He used part of one of them for his investigations. He did not count the asbestos bodies visible on tissue slides initially. It is the practice at the Duke University laboratory to use the lung digestion method to count asbestos bodies. That process involves dissolving the lung tissue, leaving a residue comprising asbestos bodies and mineral dust. By using light microscopy to examine half the residue, Professor Roggli identified 3,970 asbestos bodies per gram of wet lung tissue.
17. Professor Roggli then examined the other half of the residue by scanning electron microscopy (SEM) and energy dispersive x-ray analysis (EDXA). By that method he detected 60,100 coated and 278,000 uncoated asbestos fibres per gram of wet lung tissue, a total of 338,000 fibres per gram. Twenty consecutive uncoated fibres were examined by EDXA. Seventeen fibres had a chemical composition indicative of crocidolite and the composition of the other three fibres was indicative of amosite. Ten consecutive coated fibres were also examined by EDXA, of which six were crocidolite and four were amosite.
18. In the course of his preparations for the hearing, Professor Roggli performed a count of asbestos bodies by light microscopy on one randomly selected slide out of the four which had been sent to him by Dr Attanoos. In that one slide, he found six asbestos bodies, the equivalent of 2.8 asbestos bodies per sq cm.

The medical evidence

19. In addition to the pathologists, both parties instructed consultant chest physicians. Dr Robin Rudd was instructed on behalf of the claimant and Dr Charles Hind by the defendant. Neither Dr Rudd nor Dr Hind had examined the deceased during his life. All the medical experts provided reports and supplementary material and gave oral evidence. Each pair of experts discussed the case and provided Joint Statements setting out those respects in which they agreed and disagreed.
20. In forming their views about the causation of the deceased's fibrosis, the medical experts considered in particular:

i) The extent of the deceased's exposure to asbestos and his likely asbestos dose;

ii) The radiological findings;

iii) The morphological features of the deceased's lung tissue evident on histological examination;

iv) The clinical course of the deceased's disease; and

v) The evidence of the presence of asbestos bodies and asbestos fibres in the deceased's lung tissue.

Some of the medical experts also gave evidence about the effect of clearance of asbestos fibres and the balance of risk. I shall address each of these topics separately.

The extent of the deceased's exposure to asbestos and his likely asbestos dose

Duration of exposure

21. The Particulars of Claim stated that the deceased had been employed by BRB between 1958 and 1962. The Defence admitted employment between those dates. It is clear from the Inland Revenue records that the deceased's employment with BRB ceased during the tax year 1961/62. There is no documentary evidence as to the start date of his employment.
22. In March 2004, when the deceased first attended hospital in connection with his respiratory condition, he gave a history of having been exposed to asbestos for seven years, some 40 years previously. However, when he made an application for Industrial Injuries Disablement Benefit in January 2006, he recorded on the application form the dates of his employment with BRB as 1958 to 1962. He stated that, between 1951 and 1958, he had been a long distance lorry driver and had not been exposed to asbestos. The form completed in March 2006 by the doctor who carried out a home assessment of the deceased for the purpose of his benefit application recorded that he had been exposed to asbestos for a period of four years. The deceased's witness statement, prepared for the purpose of these proceedings and signed on 19 January 2006, gave the same history as the benefit application form.
23. The deceased's widow, Mrs Della Sabin (the claimant), made a witness statement for these proceedings in which she stated that she and the deceased married in September 1959, having become engaged at Christmas 1958. She said that they had been going out for two years before that. For the whole of that time, the deceased had been employed by BRB. The claimant gave oral evidence. She was adamant that at no time during her relationship with him had the deceased worked as a long distance lorry driver. She was unable to say when he had started to work for BRB. All she could say was that it must have been some time before she started to go out with him at or about the end of 1956.
24. Although the claimant was adamant about that point, her evidence was in other respects very vague. At one point, she appeared to believe that the deceased had gone to work for Readymix Concrete at some time in the early 1960s, whereas it was clear from the Inland Revenue documentation that his employment with that company had not started until 1971 or 1972. She also suggested that he had been driving a van at the time when she first met him, although she later agreed that he did not drive a van when employed by BRB. He did so during his employment with Readymix Concrete. She could not say when he finished working at BRB.
25. For the claimant, Mr Allan QC suggested that the claimant was unlikely to be mistaken about the work that the deceased was doing when she first met him since she would have remembered if he had been working away from home. He pointed out that her evidence was consistent with the estimate of seven years initially given by the deceased to the doctor treating him in March 2004. For the defendant Mr Kent QC submitted that I should prefer the evidence of the deceased as set out in his witness statement and in his benefit application form.
26. In his witness statement and in the benefit application form completed in the same month, the deceased gave a detailed account of his employment history. He set out the dates of his employments before he worked for BRB and the dates of his subsequent periods of employment. I am satisfied that, at that time, he would have been aware of the importance of giving an accurate account of his working history and would have tried to do so as far as he was able. It is likely that, by January 2006, he had had time to reflect on the accuracy of the first estimate of the length of his asbestos exposure which he had given almost two years previously. I infer that he had come to the conclusion that his initial recollection was wrong and that he set out in the January 2006 documents what he believed to be the correct position. Whilst, in usual circumstances, I would accept Mr Allan's contention that a wife is likely to remember the job which her husband was doing when she first met him (especially if it involved working away from home), the claimant's evidence was so vague and confused that I am unable to place any reliance on it.
27. I find that the deceased's period of employment with BRB was as stated in his witness statement, i.e. from 1958 to 1962, a period of approximately four years.

The deceased's likely asbestos dose

28. Both parties instructed experienced consulting engineers (Mr Peter Deary for the claimant and Mr Robert Hanson for the defendant) to estimate the extent of the deceased's exposure to asbestos (or "asbestos dose") during his employment with BRB. In reaching their estimates, they took account of the quantities of asbestos dust likely have been produced by the work carried out by the deceased, together with the frequency and the duration of his exposure to asbestos dust.
29. They also took into account documents from Turner and Newall dating from the mid-1960s which recorded the average dust concentration levels produced by the unloading of sacks of asbestos from lorries and the stacking by hand of asbestos at their Rochdale and Hindley Green premises. It should be noted that those tasks were recorded as producing the highest asbestos levels (from six to 50 fibres per cc) of any work done at the two premises.
30. In their Joint Statement, the engineers set out their respective positions. They expressed their calculations in terms of the average exposure level over a working year (fibres/ml years). Mr Deary estimated the deceased's mean exposure to asbestos dust at 42 fibre/ml years, while Mr Hanson estimated it at 47.2 fibre/ml years. Both these estimates were based on a period of four years' employment.
31. Professor Roggli, together with Dr Rudd and Dr Hind, was content to accept the engineers' estimate of the deceased's asbestos dose. However, Dr Attanoos expressed doubt about the general accuracy of dose estimates undertaken by engineers. He said that they make the false assumption that all airborne asbestos is capable of being inhaled, whereas in fact only some of the asbestos fibres measured by their equipment are respirable. Furthermore, the equipment used to measure asbestos fibre counts in the workplace does not record the presence of smaller fibres which can be inhaled and are detectible by transmission electron microscopy. Nor does it permit a distinction to be made between asbestos and other fibres. He described the engineers' evidence as necessarily subjective and suggested that it should be given far less weight than the objective analysis of the presence of asbestos fibres in lung tissue.

Evidence relating to the threshold asbestos dose level giving rise to a risk of asbestosis

32. There is evidence as to the level of asbestos dose which gives rise to a risk of asbestosis. Thus, having reached a conclusion about the deceased's likely asbestos dose, it is possible to compare that dose with that level.
33. The Royal Commission (1984) on Matters of Health and Safety Arising from the Use of Asbestos in Ontario examined the available evidence and concluded that there was a threshold level below which an individual exposed to asbestos was not at risk of developing asbestosis. They stated that their best judgement as to the lifetime occupational exposure to asbestos at which the fibrotic process could not advance to the point of clinical manifestation of asbestosis was in the range of 25 fibre/ml years and below. Professor Doll and Professor Peto in their authoritative paper, Effects on health of exposure to asbestos, published in 1985, saw no reason to disagree with the Commission's conclusion. The CAP-PPS Report (as to which see paragraph 42 below) states that clinical asbestosis can be induced by cumulative asbestos exposure amounting to an estimated 25 fibre/ml years, whilst research in China, the US and Germany suggests that sub-clinical (i.e. Grade 1 or Grade 2) asbestosis can result from an asbestos dose as low as 10 fibre/ml years.
34. There was some dissent between the experts as to the type of fibres to which the threshold level of 25 fibre/ml years related. Professor Roggli and Dr Attanoos said that the level was appropriate to exposure to commercial amphiboles. They suggested that the threshold level giving rise to a risk of asbestosis as a result of exposure to chrysotile would be higher than 25 fibre/ml years. The evidence of Dr Rudd and Dr Hind was that the Commission's threshold level was based primarily on data relating to chrysotile exposure. Dr Rudd said that a recent unpublished study had suggested that, for amphiboles, a lower threshold might be appropriate. For practical purposes, however, where there was uncertainty as to the type(s) of asbestos fibre(s) to which the individual had been exposed, he regarded the threshold of 25 fibre/ml years as appropriate. Dr Attanoos plainly had reservations about the threshold, which he said had been based on only a limited number of epidemiological studies. Asked whether he considered that the threshold should be changed, he responded that he believed that there was good evidence that the level should vary in accordance with the type of asbestos to which there had been exposure.

The accepted criteria for diagnosing asbestosis

35. Before going on to deal with the other matters considered by the medical experts, it is necessary to refer to the criteria which have been developed for the diagnosis of asbestosis.
36. A paper published in 1982, The Pathology of Asbestos-Associated Disease of the Lungs and Pleural Cavities: Diagnostic Criteria and Proposed Grading Scheme, by Dr John Craighead and others, set out pathological criteria for the diagnosis of asbestosis, together with criteria for grading the severity of the disease. The paper defined the minimal features which would permit a diagnosis of asbestosis as:

"… the demonstration of discrete foci of fibrosis in the walls of respiratory bronchioles associated with accumulations of asbestos bodies".

Those criteria for diagnosis became the accepted standard and remained so until 1997. I shall refer to them as the 1982 Criteria.

37. In 1997, an international experts' meeting, whose 19 participants included Professor Roggli, was convened in Helsinki to discuss conditions of the lungs and pleura associated with asbestos exposure and to agree on "state-of–the-art" criteria for their diagnosis. I shall refer to the Report of this meeting as the 1997 Consensus Report and to the criteria as the 1997 Helsinki Criteria.
38. The 1997 Consensus Report observed that work histories, coupled where possible with dust measurements, were valuable tools for assessing occupational asbestos exposure. It added that a calculation of cumulative fibre dose was "an important parameter of asbestos exposure". The Report discussed the respective roles of radiology, histology and analysis of lung tissue for asbestos fibres and asbestos bodies.
39. The 1997 Consensus Report identified the criteria for the histological diagnosis of asbestos as follows (page 312):

"A histological diagnosis of asbestos requires the identification of diffuse interstitial fibrosis in well inflated lung tissue remote from a lung cancer or other mass lesion, plus the presence of either two or more asbestos bodies in tissue with a section area of 1 cm² or a count of uncoated asbestos fibres that falls into the range recorded for asbestosis by the same laboratory."

40. The criterion of a finding of two asbestos bodies per sq cm referred to histological examination of slides of lung tissue by light microscopy. The criterion originated from work done by Professor Roggli and his team and reported in a paper, Number of Asbestos Bodies on Iron-Stained Tissue Sections in Relation to Asbestos Body Counts in Lung Tissue Digests, published in 1983. The authors used the lung digestion method to count the numbers of asbestos bodies present in specimen lung tissue. Using logarithmic methods, they attempted to correlate those numbers with the numbers of asbestos bodies visible on examination of slides of lung tissue by light microscopy. They found that one asbestos body found on a sq cm iron-stained tissue section corresponded to about 800 asbestos bodies per gram of wet fixed lung tissue using the lung digestion method. Professor Roggli reported to the Helsinki meeting that, in the Duke University laboratory, 95% of asbestosis cases were found to have the equivalent of five asbestos bodies per sq cm on examination of a tissue section, whilst 100% of the cases had the equivalent of two asbestos bodies per sq cm. The criterion of two asbestos bodies per sq cm was adopted as part of the 1997 Helsinki Criteria. The criterion was expressed by reference to histological examination of tissue slides, rather than to the lung digestion technique, because the former is a comparatively simple technique used by laboratories all over the world, whereas lung digestion is a more sophisticated process, which is less widely available.
41. So far as asbestos fibre analysis was concerned, the 1997 Consensus Report recommended (at page 312) that:

"Each laboratory should establish its own reference values. The median values for occupationally exposed populations should be substantially above the reference values. Efforts to standardize analytical methods for fiber burden analysis by different laboratories are recommended."

42. Recently, an international committee organised under the auspices of the Asbestos Committee of the College of American Pathologists and Pulmonary Pathology Society has produced a Report (which I shall refer to as the CAP-PPS Report) updating the 1997 Helsinki Criteria and, inter alia, defining the morphological features of asbestosis at its various stages. Both Professor Roggli and Dr Attanoos were members of the international committee and co-authors of the Report. Although as yet unpublished, the CAP-PPS Report is now in its final form and has been accepted for publication. It is plainly an authoritative work in its field. The CAP-PPS Report is directed in particular at asbestosis and disorders with which asbestosis might be confused.
43. The CAP-PPS Report makes clear (at page18) that, in most cases, asbestosis can be diagnosed on the basis of the clinical and radiological evidence without recourse to histological examination. A history of moderate to heavy asbestos exposure, together with the presence of diffuse opacities in the lower zones of the lung fields, often accompanied by end-inspiratory crackles, restrictive impairment of lung function and/or the presence of pleural plaques, is sufficient to make a diagnosis, The need to undertake histological examination arises only when the history of asbestos exposure is equivocal or when the clinical or radiological features are atypical or non-diagnostic.
44. The CAP-PPS Report states (at page 27) that, in a case where histological examination is necessary, a microscopic diagnosis of asbestosis should only be made when:

"(a) there is an acceptable pattern of alveolar septal fibrosis"

Until now, the criteria have permitted a diagnosis of asbestosis in cases where fibrosis has been confined to the bronchiolar wall although there has been a difference of opinion between experts as to whether such a pattern of fibrosis is sufficient for a diagnosis of asbestosis. However, the CAP-PPS Report states the view of its authors that fibrosis which is confined to the bronchiolar wall should not be referred to as asbestosis.

"(b) an average of at least two asbestos bodies per sq cm of lung tissue can be seen on histological examination of lung tissue."

I shall refer to these criteria as the "CAP-PPS Criteria".

45. The CAP-PPS Report goes on to say that, in rare cases, fewer than two bodies per sq cm of a five micron-thick lung section, are seen and the heavy fibre burden necessary for a diagnosis of asbestosis is only demonstrated by more sophisticated analytical techniques. It points out that asbestos bodies are often distributed unevenly within the lungs so that more than one section of lung tissue may need to be examined. The Report does not specify a minimum number of sections that should be examined. It states (at page 2) that:

"Fewer asbestos bodies (i.e. less than two or more per cm²) do not necessarily exclude a diagnosis of asbestosis but evidence of excess asbestos would then require quantitative studies performed on lung digests."

46. The role of asbestos fibre analysis is discussed at page 31 et seq:

"Methods for detecting the presence and quantities of asbestos fibers in lung tissue samples were reviewed in an earlier section. Suffice it here to say that fiber analysis should be considered an adjunctive technique in the assessment of asbestosis as outlined above. Most studies have shown that parties with asbestosis have in excess of a million fibers per gram of dry lung tissue. Fiber analysis may also be useful for excluding a diagnosis of asbestosis in individuals with diffuse pulmonary fibrosis and a history of asbestos exposure, but lacking the necessary histopathological criteria. As noted in a previous section, some individuals are poor coaters of asbestos fibers and thus do not readily form asbestos bodies. In such cases, light microscopy has a limited role in the assessment of the overall lung fiber burden.

It is the consensus of the Committee that cases of asbestosis (i.e. asbestos-induced fibrosis) not meeting the histological criteria outlined in this document are rare. In such cases, analysis of lung tissue samples by an experienced laboratory using electron microscopic techniques may be useful. Cases with diffuse interstitial fibrosis and an asbestos fiber burden within the range of values observed for bona fide cases of asbestosis as determined for a given experienced laboratory are likely examples of asbestos-induced pulmonary fibrosis (i.e. asbestosis). The asbestos range for a laboratory refers to the retained amphibole fiber counts in cases of asbestosis (meeting the aforementioned morphological criteria). The chrysotile count is not included due to the low biopersistence of the fiber. Conventional biological range values are defined as including 95% of observed values for that group. The critical value to determine as the lower range value is the 5th percentile, i.e. the value below which the lowest 5% of cases fall and 95% of cases are above."

The purpose of using the fifth percentile value (rather than the minimum value of the range) is to allow for errors of diagnosis or other inaccuracies which might have occurred at the lower end of the range.

The radiological findings

47. The CT scan on the deceased performed in August 2004 showed appearances of diffuse interstitial pulmonary fibrosis with bilateral calcified pleural plaques typical of asbestos exposure. Pleural plaques are evidence of asbestos exposure. However, they can develop after asbestos doses well below that necessary to cause asbestosis. Dr Rudd's view was that their presence does not imply that the cause of an individual's fibrosis is asbestosis. In their first Joint Statement, the pathologists agreed that the presence of pleural plaques did not "allow one to conclude that an associated diffuse pulmonary fibrosis was asbestosis". However, Professor Roggli expressed the view, under "Points of Disagreement" that, although pleural plaques were not a definitive finding, they "added to the weight of the evidence" in favour of a diagnosis of asbestosis. In their second Joint Statement, the pathologists agreed that pleural plaques, although not determinative of causation, were more common in asbestosis than in UIP. In evidence, Professor Roggli said that 12% of individuals diagnosed with UIP had pleural plaques in contrast to 80% of individuals diagnosed with asbestosis. His view was that the presence of pleural plaques "leaned in favour" of a diagnosis of asbestosis.

The morphological features

48. Both pathologists examined the morphological features of the deceased's lung tissue. They agreed that there was evidence of established diffuse interstitial fibrosis. They both noted the presence of features which are typically associated with UIP. UIP is a disease in which inflammation tends to advance aggressively, but at different rates in different areas of lung tissue. Typically, the tissue will have a "honeycomb" appearance, with areas of pathologically advanced fibrosis adjacent to areas where lung tissue is normal or in the early stages of fibrosis. Fibroblastic foci (i.e. buds seen at the advancing edge of the fibrosing process) will be present, as will evidence of inflammatory cell infiltrates. Asbestosis, on the other hand, usually has a slower rate of progress and typically produces a more homogeneous, collagenous pattern in the lungs with less evidence of inflammation. Fibroblastic foci are not usually seen.
49. The pathologists agreed that there was a profusion of fibroblastic foci in the deceased's lung tissue, together with honeycombing and inflammatory cell infiltrates. They further agreed that these features, taken on their own, would in general favour a diagnosis of UIP over one of asbestosis. The difference between the pathologists was as to the weight to be attached to the morphological evidence.
50. The 1997 Consensus Report (at page 312) stated:

"Asbestosis is defined as diffuse interstitial fibrosis of the lung as a consequence of exposure to asbestos dust. Neither the clinical features nor the architectural tissue abnormalities sufficiently differ from those of other causes of interstitial fibrosis to allow confident diagnosis without a history of significant exposure to asbestos dust in the past or the detection of asbestos fibers or bodies in the lung tissue greatly in excess of that commonly seen in the general population."

51. The CAP-PPS Report discusses the difficulties associated with distinguishing between asbestosis and other forms of fibrosis on the basis of morphological features alone. It observes (at page 26):

"Some cases [of asbestosis] resemble UIP whereas others are more like fibrotic non-specific interstitial pneumonia (NSIP) and still others do not match any other form of interstitial fibrosis. Asbestosis is characterized as having a lower lobe and peripheral distribution similar to UIP, but with the temporal and spatial homogeneity of the fibrotic variant of NSIP. Fibroblast foci are uncommon, only occasionally being seen …. If these foci of immature fibrosis are at all conspicuous, another diagnosis (such as UIP) should be considered. Honeycombing may be seen in advanced cases but it is seldom as severe as in UIP."

52. At page 30, the CAP-PPS Report states:

"A more difficult area is the distinction between idiopathic pulmonary fibrosis and asbestosis …. The most common pattern of idiopathic pulmonary fibrosis is usual interstitial pneumonia (UIP). This is characterized by temporal heterogeneity, represented by densely hyalinized areas of fibrosis alternating with areas showing fibroblastic foci and yet others that consist of nearly normal lung. Honeycombing changes are frequently found in UIP. As noted, some cases resemble UIP and others fibrotic NSIP but in general the presence of readily identified asbestos bodies permits the distinction of asbestosis from these other interstitial lung disorders. As noted above, the presence of frequent fibroblast foci is against a diagnosis of asbestosis. Pleural plaques … provide evidence of asbestos exposure but they develop at relatively low levels of exposure and may therefore be present in patients with other fibrotic lung disorders. In difficult cases fiber analysis may be necessary to determine the etiology of the fibrotic process."

53. Microscopic diagnosis is discussed (at page 25):

"The microscopic diagnosis of asbestosis requires an appropriate pattern of interstitial fibrosis plus the finding of asbestos bodies. Both components must be present. It may be added that the fibrosis in asbestosis is always paucicellular, lacking any significant degree of inflammation and being collagenous rather than fibroblastic."

54. Professor Roggli acknowledged that the morphological features in the deceased's case were more typical of UIP than of asbestosis. He said that he had considered that fact when reaching his diagnosis. However, he said that honeycombing of the lung tissue was not uncommon in cases of more advanced (i.e. Grade 4) asbestosis. In the Duke University laboratory, honeycombing had been observed in 11% of cases of Grade 3.5/4 asbestosis. Professor Roggli conceded that the presence of fibroblastic foci was suggestive of UIP. However, although uncommon in asbestosis, they could occur. His view was that the overall pattern was predominately collagenous rather than inflammatory. He observed that both the 1997 Consensus Report and the CAP-PPS Report made clear that there were some "problem cases" where it was not possible to differentiate between asbestosis and other types of fibrosis (in particular UIP) on the basis of morphological features alone. The deceased's was, he said, such a case.
55. Professor Roggli said that, in his view, the correct approach to take in such cases was to see whether there were asbestos bodies present in the lung tissue. If abundant asbestos bodies were present, a diagnosis of asbestosis was appropriate even if the morphological features were more typical of UIP. If no asbestos bodies were seen, the diagnosis should be UIP. If only a few asbestos bodies were seen, it was necessary to carry out mineral analysis in order to find out how many asbestos fibres were present in the lung tissue. If the asbestos fibre count fell within the relevant reference range for the laboratory, the correct diagnosis would be asbestosis.
56. Dr Attanoos considered that the morphological features in the deceased's case militated strongly against a diagnosis of asbestosis. His view was that the presence of fibroblastic foci with asbestosis was "exceptional" rather than (as the final version of the CAP-PPS Report states) "uncommon" or "infrequent". He acknowledged that, in the Llandough Hospital series of asbestosis cases, there were a small number of cases which had morphological features of the UIP type. They had been included because, despite those features, the cases fulfilled the 1997 Helsinki Criteria. The Criteria did not provide for the exclusion of cases with atypical morphological features. However, Dr Attanoos expressed doubt as to whether the atypical features were in fact the product of asbestosis. He pointed out that it was possible that the individuals had been exposed to asbestos (as a result of which the criterion requiring the presence of two asbestos bodies in a sq cm of lung tissue was fulfilled) and then, quite independently, they had developed UIP. He said that, given the relatively high incidence of UIP among the general population, that was a perfectly plausible explanation.
57. Dr Attanoos accepted that, if he had found an average of two or more asbestos bodies in one sq cm of lung tissue, or if the asbestos fibre count had exceeded the fifth percentile of the Llandough Hospital laboratory reference range, he would have diagnosed asbestosis in the deceased's case despite the atypical morphological features. That would, he said, have accorded with the approach set out in the 1997 Consensus Report and the CAP-PPS Report.

The clinical course

58. The deceased first consulted a doctor about his respiratory symptoms in January 2004. Asbestosis was diagnosed by Dr Wolstenholme in October 2004. In August 2005, the deceased saw Dr Wolstenholme for the purposes of a medical report. Dr Wolstenholme noted that he described increasing breathlessness over the past two years which had progressed more over the previous couple of months. He noted that there appeared to be an increasingly rapid progression of the disease. He estimated the deceased's future life expectancy at around three years. In the event, the deceased died nine months later.
59. Dr Attanoos said that the rapid clinical progression of the disease was typical of UIP. The rate of progression was linked to the morphological features and was caused by the aggressive nature of the disease. By contrast, asbestosis tended to progress slowly over a long period, often many decades. He said that he did not find it "biologically plausible" that asbestos fibres should elicit a rapidly aggressive process in some individuals but not in others. His view was that, in cases where that appeared to have occurred, UIP was in fact present in an individual who happened to have been previously exposed to asbestos.
60. Dr Hind said that the clinical course of the deceased's disease was in favour of a diagnosis of UIP. The progression had been rapid and he had died of his disease. Few asbestosis sufferers went on to die of their disease. He said that both these factors suggested that the deceased had been suffering from UIP, not asbestosis.
61. Professor Roggli acknowledged that the progression of the deceased's disease as more rapid than in the typical case of asbestosis. However, he suggested that its progression might not have been as rapid as it appeared. No radiological evidence was available before 2004, so there was no way of knowing whether he had signs of the disease before that time. He said that, in any event, it was recognised that there was some cases of asbestosis in which the disease was more aggressive and progressed rapidly. He said that it was clear that the nature and extent of a person's response to asbestos exposure depended on individual susceptibility, the mechanisms of which were not well understood.
62. Dr Rudd accepted that, on average, UIP progresses more rapidly than asbestosis. However, he said that there was a wide range of rates of progression in both diseases. He pointed out that the x-rays taken in January 2004 showed clear evidence of interstitial fibrosis of the lungs which would have been developing for some considerable time before then. In April 2004, the deceased had been found to have only half his predicted lung function. Thus, there must, he said, have been considerable progression of the disease before 2004. He accepted, however, that the clinical course in the deceased's case was not typical of asbestosis but said that it fell within the range of rates of progress which were known to occur in asbestosis.
63. The CAP-PPS Report refers (at page 20) to the clinical courses of UIP and asbestosis, observing:

"One point of distinction is that UIP tends to be more rapidly progressive than asbestosis, which is usually either static or only slowly progressive, although cases of rapidly progressive asbestosis do occur."

This tends to support Dr Rudd's view.

The evidence of the presence of asbestos bodies and asbestos fibres in the deceased's lung tissue.

64. As I have said, Dr Attanoos carried out an examination of four slides of lung tissue by light microscopy in order to ascertain the number of asbestos bodies therein. Professor Roggli carried out a similar exercise with one slide. Both pathologists considered the results of mineral analysis carried out at their respective laboratories by electron microscopy in order to measure the quantity of asbestos fibres in the deceased's lung tissue.
65. The laboratories at both Llandough and Duke University have established reference ranges for the levels of asbestos fibres found on histological examination in the cases of control subjects who have had exposure to asbestos and have been diagnosed as suffering from asbestosis ("asbestosis controls"), as well as for control subjects with no occupational exposure to asbestos. Professor Roggli and Dr Attanoos compared the levels of asbestos fibres found in the deceased's lung tissue with the ranges recorded in their respective laboratories.
66. Professor Roggli compared his finding of 338,000 fibres per gram of wet lung tissue with the Duke University laboratory reference range for asbestosis controls. The median value for commercial amphibole exposure is 206,000 fibres per gram of wet lung tissue for fibres five microns or greater in length. The fifth percentile value is 20,800 fibres per gram. The deceased's fibre count was well above the median value for the range. It was above the fifth percentile value for Grade 4 cases and a little below the median value of 398,000 fibres per gram.
67. Dr Attanoos used Dr Gibbs' analysis of the deceased's lung tissue. He had found a total amphibole (i.e. crocidolite, amosite and anthophyllite) asbestos count of 5.71 million fibres per gram of dry lung tissue of which commercial amphiboles (crocidolite and amosite) comprised 5.42 million fibres per gram. Dr Attanoos compared those results with the reference ranges held at the Llandough Hospital laboratory for non-occupationally exposed controls. The range for total amphiboles extended up to one million fibres per gram of dry lung tissue. Thus, the deceased's total amphibole asbestos count was more than five times the upper range of the non-occupationally exposed control population.
68. Dr Attanoos went on to compare the results of Dr Gibbs' analysis with the Llandough Hospital reference range for asbestosis controls. The fifth percentile of that range is 21.4 million fibres per gram of dry lung tissue. The deceased's commercial amphibole fibre count (5.42 million fibres per gram) fell well below that value. Applying 95% confidence limits round the fifth percentile value produced a range of 14-31 million fibres per gram. The deceased's commercial amphibole fibre count fell well below even the lower end of that range.
69. Each of the experts argued that the exercise carried out in the other's laboratory was flawed in several important respects. The alleged flaws relate to two factors in particular, namely the validity of the reference ranges of asbestosis controls held by each laboratory and the methods used in each laboratory to examine and analyse the deceased's lung tissue.

The validity of the Duke University reference range of controls

70. The reference range of asbestosis controls at the Duke University laboratory has been established over three decades. It is based on 206 cases which cover a wide range of employments and levels of asbestos exposure. Sixty eight per cent of the cases came to the laboratory as medico-legal cases and 20% came when specimens of lung tissue were sent for resection in cases of suspected lung cancer or mesothelioma.
71. There has been no peer review of the cases that make up the reference range. The cases have, however, been the subject of several publications based on research done using the asbestosis control cases. A paper published by Professor Roggli and Dr Robin Vollmer in 2008, Twenty-five years of fiber analysis : what have we learned?, reported on a review carried out in order to ascertain whether asbestos fibre counts had changed over the previous 25 years. The cases were divided equally according to their chronological sequence. The first (earlier) half had significantly higher median asbestos body and fibre counts than the second half. A study of the severity of asbestosis in those cases which had been classified by grade revealed more cases of Grade 4 in the first half (13%) than in the second half (9%). These findings were consistent with the levels of asbestos exposure and the lower severity of asbestosis following the US ban on the use of asbestos in insulation products from 1971.
72. For the purposes of these proceedings, Professor Roggli produced a spreadsheet of the mineral analysis of 192 of the 206 reference range control cases which, according to him, demonstrated histologically confirmed asbestosis with evidence of exposure to commercial amphiboles (i.e. crocidolite and/or amosite). The median value of the findings on electron microscopy was 206,000 fibres per gram of wet lung tissue for fibres five microns or greater in length. The fifth percentile was 20,800 fibres per gram.
73. The defendant made a number of criticisms of the Duke University reference range. The main criticisms were:

(i) The Duke University reference range contains cases which do not meet the 1997 Helsinki Criteria in that they do not have two asbestos bodies per sq cm of dry lung tissue.

(ii) The Duke University reference range controls have not been reconsidered in the light of developments in pathological opinion. Current mainstream pathological opinion (as evidenced by the CAP-PPS Report) is that, for asbestosis to be diagnosed, there must be at least first tier involvement of alveolar septal fibrosis. Dr Attanoos pointed out that Professor Roggli's previous publications made clear that 18 out of 100 of his control cases did not have any evidence of alveolar septal fibrosis. Thus, it may well be that as many as 18% of the total control group (perhaps more) lacked this evidence.

(iii) A large proportion of the Duke University laboratory asbestosis control cases were classified as Grade 1. Dr Attanoos suggested that there may be Grade 1 cases which had been wrongly classified as being cases of asbestosis. He pointed out that Grade 1 fibrotic lesions are commonly seen in smokers who have had no asbestos exposure. Such cases might have been wrongly classified as cases of early asbestosis.

(iv) There had been no systematic system of referral of cases to the Duke University laboratory. Many cases were medicolegal referrals. A significant proportion had other, often malignant, diseases which may have resulted in misdiagnosis of asbestosis. There had been no peer review of the histology in the Duke University laboratory asbestosis controls cases.

74. Dr Attanoos argued that the effect of (i), (ii) and (iii) was that cases had or might have been classified as cases of asbestosis which did not in fact meet the criteria. These were likely to be early (i.e. Grade 1) cases where the levels of asbestos fibres in the lungs were likely to be lower. The misclassification of such cases as "true" cases of early asbestosis would have the effect of reducing the minimum asbestos fibre level in the Duke University reference range. He said that the only means of validating Professor Roggli's asbestosis control cohort would be by a formal review of his control cases.
75. Professor Roggli accepted that, when his laboratory started to collect its asbestosis control cases, the 1982 Criteria, which required only an accumulation of asbestos bodies in lung tissue, were in force. The control cases had not been "weeded out" to take account of the 1997 Helsinki Criteria. Thus, there would be a number of the Duke University asbestosis controls that would not have fulfilled the 1997 Helsinki Criteria in that examination of tissue slides by light microscopy would not have revealed two asbestos bodies per sq cm. He estimated that this would be no more than about 5% of the control cases. In any event, those cases would have had high asbestos fibre counts, coupled with the presence of diffuse interstitial fibrosis, so that a diagnosis of asbestosis was justified according to the 1997 Consensus Report. He accepted also that at least some (possibly as many as 18%) of the Duke University asbestosis controls had bronchiolar wall fibrosis only and therefore would not meet the new criteria contained in the CAP-PPS Report.
76. Professor Roggli did not accept that the Duke University laboratory asbestosis controls were unrepresentative or that the presence of malignant disease or other lung conditions, where it occurred, was liable to cause confusion. Analysis was carried out only on sections of lung tissue well away from any tumour that was present. He did not consider that there had been any confusion with fibrosis caused by cigarette smoking. The appearance of such fibrosis could, he said, readily be distinguished from the appearance of asbestosis.
77. Professor Roggli maintained that, even if the number of the Duke University asbestosis controls was reduced to remove those cases that did not meet the 1997 Helsinki Criteria and/or the CAP-PPS Criteria, this would have little effect on the fifth percentile value of the reference range. The cases to be excluded would inevitably be the less severe cases, i.e. cases classified at Grade 1. The Duke University laboratory asbestosis controls included a substantial number of Grade 1 cases (84 out of the 167 cases in which the grading was known). This was because the finding of asbestosis was often incidental to a malignant disease such as mesothelioma or lung cancer. He referred to a diagram in a 2009 paper (as yet unpublished, but accepted for publication), Asbestos Fiber Content of Lungs with Diffuse Interstitial Fibrosis, written by himself and colleagues from Duke University. He said that the diagram (at page 154, Figure 2) demonstrated that the Duke University laboratory asbestosis controls with Grade 1 disease showed a wide spread of asbestos fibre counts and therefore, by implication, a wide range of levels of asbestos exposure. Quite a number of the Grade 1 cases had fibre counts suggestive of a high level of asbestos exposure. He said that, in those circumstances, the exclusion of a number of Grade 1 cases would not necessarily have much impact on the fifth percentile value of the Duke University reference range.
78. Professor Roggli said that, in any event, the issue of whether or not some of the Duke University laboratory Grade 1 asbestosis control group should be excluded was of marginal, if any, relevance to this case. The deceased's asbestosis (if it were asbestosis) would be classified as Grade 4. The relevant group of asbestosis controls to examine was therefore the group of 18 Grade 4 cases contained within the Duke University laboratory asbestosis controls. I shall refer to Professor Roggli's evidence about that group later in this judgment.
79. Professor Roggli accepted that there had been no peer review of the diagnoses in the Duke University laboratory asbestosis control cases. He considered that there were dangers in undertaking a review of a control group "by committee" (as undertaken at the Llandough Hospital laboratory) since that might result in the exclusion of some "true" asbestosis cases about which one expert had doubts. He pointed out that there had been peer reviews of a number of publications based on his cases. The analysis of the 206 cases carried out in 2008 (see paragraph 70 above) had produced the results which might have been expected; that provided some support for the fact that the control group was robust.
80. Professor Roggli said that the Duke University laboratory had done a lot of work in attempting to differentiate between cases of asbestosis and cases of IPF/UIP. A paper, Scanning Electron Microscopic Analysis of Mineral Fibre Content of Lung Tissue in the Evaluation of Diffuse Pulmonary Fibrosis, written by him and published in 1990, described work done to compare the asbestos fibre content of lung tissue from 24 cases of diffuse interstitial fibrosis of unknown cause (IPF) in which the individuals concerned had had some exposure to asbestos with the asbestos fibre content of lung tissue from 36 cases of histologically confirmed asbestosis. The fibre count was below the 95% confidence limit for asbestosis in each of the 24 IPF cases. Professor Roggli said that the findings suggested that the Duke University reference range was robust in that non-asbestosis cases did not come within the asbestosis range, even when there had been some asbestos exposure.
81. The work described in the unpublished 2009 paper (see paragraph 77 above) followed up the 1990 study. The asbestos fibre content of lung tissue from (i) 163 cases of histologically confirmed asbestosis and (ii) 86 cases of diffuse interstitial fibrosis with a history of asbestos exposure (but where tissue obtained on biopsy did not meet established criteria for asbestosis) were compared. EDXA was used to identify the types of asbestos fibres observed. In this study, three cases of Grade 3/4 diffuse interstitial fibrosis cases fell within the 95% confidence limit for a count of total commercial amphibole asbestos fibres in the asbestosis control group. In the paper, Professor Roggli and his colleagues suggested that the three cases were probably "true examples of occult asbestosis" and had been wrongly diagnosed because the individuals were poor coaters of asbestos fibres and therefore did not meet the 1997 Helsinki Criteria or because of some sampling error related to the variability of asbestos bodies in histological section. Professor Roggli said that the fact that the majority of diffuse interstitial fibrosis cases fell outside the 95% confidence limit again tended to confirm the robustness of the Duke University asbestosis reference range.
82. In evidence, Dr Rudd acknowledged that the fact that some of Professor Roggli's asbestosis control cases would not meet current criteria for the diagnosis of asbestosis constituted a "limitation".

The validity of the Llandough Hospital reference range of controls

83. The Medical Research Council Pneumoconiosis Unit was set up at the Llandough Hospital shortly after the war and was active until 1985. It was a leading centre for research into industrial disease. Much of that research was conducted by Dr Chris Wagner, a pathologist of great repute. He looked at a large group of cases arising from employment at the Devonport dockyard. They were referred to the Unit by regional pathologists; they were all post-mortem cases. Dr Wagner diagnosed 170 cases of asbestosis arising from that group. Since the closure of the Unit, clinicians based at the Llandough Hospital (among them Dr Gibbs and Dr Attanoos) have continued to work in the field of asbestos-related disease. As I have previously said, Llandough Hospital is the only centre in the UK which is equipped to carry out analysis of asbestos content in lung tissue by means of electron microscopy.
84. Up to 2006-2007, the Llandough Hospital reference range was based on work done by Dr Gibbs, who concluded that, in a case of Grade 1 asbestosis, the asbestos fibre count was usually in excess of 50 million fibres per gram of dry lung tissue. That conclusion was based upon the 170 Devonport dockyard cases. The figure was subject to criticism from some quarters as being significantly too high. This led to an attempt to establish a more formalised reference range.
85. In 2006-2007, Dr Attanoos, together with Dr Gibbs and another pathologist, Dr Corrin, reviewed the histological evidence in the 170 cases previously diagnosed by Dr Wagner so as to ensure that they complied with the 1997 Helsinki Criteria. They excluded all those cases that did not match the requirement for diffuse interstitial fibrosis set out in the 1997 Helsinki Criteria and where two asbestos bodies per sq cm were not found on histological examination of lung tissue sections. That exercise resulted in a reduction of the 170 cases to just 47. For a short time, those 47 cases formed the Llandough Hospital asbestosis control group. The fifth percentile value for that group was 19.9 million fibres per gram of dry lung tissue. Dr Attanoos said that that value was a "provisional value" pending the addition of further cases to the group.
86. Subsequently, Dr Attanoos was given the task of identifying further cases for inclusion in the asbestosis control group. A number of cases involving individuals who had worked at the Cape Uxbridge, Cape Acre Mill or Nottingham Gas Mask factories were identified and subjected to careful screening to ensure that they met the 1997 Helsinki Criteria. This exercise resulted in the addition of a further 33 cases to the asbestosis control group, making a total of 80 in all. The fifth percentile value for the expanded group was 21.4 million fibres per gram of dry lung tissue.
87. The claimant makes a number of criticisms of the Llandough Hospital reference range. Those criticisms can be summarised thus:

a) The number of asbestosis controls used to produce the range is very small, only 80;

b) Whilst it may be that there was a range of employments and levels of asbestos exposure among the Devonport dockyard cases, those who were employed at the Cape Uxbridge, Cape Acre Mill and Nottingham Gas Mask factories would have had very heavy exposure indeed; all the cases involved persons who had died more than 20 years ago and whose asbestos exposure would have occurred much earlier, when the levels of asbestos exposure were likely to have been very high;

c) The exercise carried out by the three pathologists may well have ensured that every case in the control group was a "true" asbestosis case but may also have excluded some cases of less severe asbestosis. It was surprising that as many as 123 of Dr Wagner's 170 cases had been excluded, especially since some had been excluded because it was considered that they had been misdiagnosed initially.

88. It was suggested that the small number of asbestosis controls raised a question about the reliability of the reference range. Dr Rudd said that the minimum number of cases required for a reliable indication of the fifth percentile value was 200; 400 cases would produce an even greater degree of reliability. A range based on only 80 cases was statistically very unreliable. If such a small number of controls was to be used as a reference range, it would be appropriate to use the minimum value, rather than the fifth percentile value, as a comparator. However, that minimum value was likely to be greater than the minimum value would have been if more asbestosis control cases had been used. Dr Rudd observed that, given the small number of cases in the control group, the statistical confidence in both the fifth percentile and minimum values must be low.
89. Furthermore, it was said that the effect of b)-c) above was that the Llandough Hospital asbestosis controls were an unrepresentative group of individuals with very high levels of exposure to asbestos and therefore very high asbestos fibre counts. Dr Rudd suggested that the group did not contain a sufficient range of exposure levels. The result was that the median and fifth percentile values of the Llandough Hospital reference range were set too high. In the course of his clinical career, Dr Rudd has taken thousands of histories of asbestos exposure from men suspected of having asbestos-related disease. He said that exposure at the Cape Uxbridge, Cape Acre Mill and Nottingham Gas Mask factories was amongst the heaviest ever recorded. He pointed out that, once the 33 very heavily exposed asbestosis controls who had worked at those factories were added to the 47 Devonport asbestosis controls, the Llandough Hospital laboratory reference range fifth percentile value increased from 19.9 million to 21.4 million fibres per gram of dry lung tissue. That, he said, demonstrated the potential effect of altering the distribution of exposure among the control group.
90. Dr Rudd acknowledged that employees at the Devonport dockyard would have had varying levels of exposure. However, he said that most of the 47 dockyard workers included in the Llandough Hospital laboratory asbestosis control group had died in the late 1960s or 1970s, making it likely that they had been exposed to the high levels of asbestos recorded in the data collected by Surgeon Commander Harries in the 1960s. Dr Rudd said that, in his clinical practice, he had encountered individuals suffering from asbestosis who had been employed in trades such as plumbing or construction. Their asbestos exposure, although over the recognised threshold level giving rise to the risk of asbestosis, had occurred at much lower levels than those likely to have been encountered by the Llandough Hospital asbestosis controls and had often been intermittent rather than continuous. He said that asbestosis cases resulting from lower levels of exposure were not represented in the Llandough Hospital asbestosis control group. The result was that the Llandough Hospital reference range was inadequate for the purpose of identifying the minimum level of asbestos fibre count to be found in individuals who had been diagnosed with asbestosis. It was impossible to say whether, if the Llandough Hospital control group had been larger with a wider distribution of levels of exposure, the deceased's fibre count would have fallen within the fifth percentile value for the range.
91. Dr Attanoos said that he considered that the Devonport dockyard cases represented a very heterogeneous group of workers with varying levels of asbestos exposure. As to the 33 additional cases, they had been chosen from cases which had been referred to the Llandough Hospital in the past. He accepted that all three groups of workers would have had very high levels of exposure to asbestos and could therefore be expected to have very high asbestos fibre counts. The fact that 31 out of the 80 asbestosis control cases had been classified as Grade 4 disease reflected the fact that there had probably been a high level of asbestos exposure in those cases. It also reflected the fact that the Llandough Hospital asbestosis controls were all post-mortem cases and would therefore include a number of individuals who had died of their asbestosis.
92. Dr Attanoos accepted that, in order to reflect the full range of levels of asbestos fibre counts associated with asbestosis, it would probably have been preferable to examine cases involving individuals who had been exposed to lower levels of asbestos. He explained that constraints of time and resources had meant that the control cases had to be identified from within the material held at the Llandough Hospital laboratory. He observed that many cases with lesser exposure would not fulfil the 1997 Helsinki Criteria and would not therefore be eligible for inclusion in the Llandough Hospital asbestosis control group.
93. Dr Attanoos was asked about the age of the cases in the Llandough Hospital asbestosis control group. He agreed that the latest date of death among the 47 Devonport dockyard asbestos controls was 1980. The dates of death in the eight Cape Acre Mill and Nottingham Gas Mask factory cases were not apparent from the data; dates in 1990 and 1991 did appear but these could have been the dates when research involving those cases had previously been carried out. The most recent death among the 25 Cape Uxbridge group of cases was 1987. He agreed that (with the possible exception of the Acre Mill and Nottingham Gas Mask factory cases) all the cases within the asbestosis control group involved asbestos exposure which had occurred several decades previously.
94. Dr Attanoos said that, provided that the asbestosis control group contained some cases of lower levels of asbestos exposure, the fact that many of the cases in the group involved very high levels of exposure would not affect the usefulness of the reference range. The median of the range would of course be higher than if more cases involving lower levels of asbestos exposure had been included. However, the important value is the fifth percentile value. He said that that value would be unaffected by the inclusion in the control group of a large number of cases involving high levels of asbestos exposure.
95. So far as the review of Dr Wagner's asbestosis cases was concerned, Dr Attanoos said that the reviewing pathologists had initially been perplexed to find that, in a large number of cases, they could not confirm Dr Wagner's diagnosis of asbestosis. However, Dr Wagner had been applying the standards of his time, when there was a general acceptance that the presence of even one asbestos body in lung tissue denoted asbestosis. In addition, he had been a research – rather than a diagnostic – pathologist and was not as well versed in routine pathology. In the event, in 23 cases, the reviewing pathologists considered that the presence of other pathologies (such as silicosis) precluded the diagnosis of asbestosis. In the majority of cases rejected, examination of the available histological material did not confirm the presence of diffuse interstitial fibrosis. In those cases, the fibrosis was very focal, suggesting that it had been caused by smoking. In some cases the available histological material was insufficient to confirm the diagnosis. In many of the cases rejected, the reason for rejection was that less than two asbestos bodies per sq cm were seen on examination of the tissue sections. Such cases were rejected by the reviewing pathologists regardless of the asbestos fibre count. Dr Attanoos said that there were a number of cases where there was a high asbestos fibre count but less than two asbestos bodies per sq cm. He said that the number of such cases would be small since cases with a high amphibole count and an asbestos body count of less than two asbestos bodies per sq cm are unusual. Dr Attanoos said that he and his colleagues had examined 19 cases in which Dr Wagner had not diagnosed asbestosis; they had concluded that some of those cases (he could not say how many) met the 1997 Helsinki Criteria and had included them in the reference range.

The results of examination under light microscopy

96. In his initial search for asbestos bodies, Dr Attanoos used hematoxylin-eosin (HE) stained tissue slides, whereas Professor Roggli used the lung digestion method. There was some debate at the hearing as to the relative accuracy of the two techniques.
97. Professor Roggli said that it was important to remember that the criterion of "two asbestos bodies per sq cm" had been adopted as a surrogate for the number of asbestos bodies that would be found by means of the lung digestion method. The latter, he said, was a far more accurate technique for the reasons that he and his colleague had set out in their 1983 Paper (see paragraph 40). They had emphasised (page 359) that the major source of error in the quantification of asbestos bodies by examination of slides under light microscopy was the marked variation of asbestos bodies present in different blocks of tissue or at different levels within the same block. Such variation was, they observed, less of a problem with the lung digestion method since a much larger amount of tissue was analysed. Even so, lung digestion often showed a substantial variation in the range of asbestos body counts from different sites in the lung. They had concluded that the examination of sections of lung tissue was "an appropriate qualitative screening technique for the detection of asbestos bodies". However, they said that they were unable to recommend it "as a substitution for digestion techniques with respect to quantitative determination of tissue asbestos body counts". The method had been adopted as part of the 1997 Helsinki Criteria in preference to the lung digestion technique only because it was a widely available method.
98. Professor Roggli said that the 1983 Paper had been concerned primarily with amosite asbestos bodies. (In general, crocidolite was less widely used in the US than in the UK and there was more use of amosite.) He suggested that the correlation contained in the 1983 Paper may not be as relevant to cases involving significant crocidolite exposure. He referred to a suggestion made by a leading pathologist, Professor Henderson, and reported in the "parent document" to the 1997 Consensus Report, that crocidolite bodies may be very small and less easy than amosite bodies to identify on histological examination of tissue slides, leading to under-reporting of the number of asbestos bodies present when that method is used. Professor Roggli said that, with the lung digestion technique, it was much easier to see the shape of an asbestos body. When examining a tissue slide, asbestos bodies can be obscured by the tissue itself and there may be other particulate material present which makes identification difficult. In addition, some individuals are "poor coaters" of asbestos fibres, making those asbestos bodies that do form less easy to see on examination of a tissue slide.
99. Dr Rudd observed that the lung digestion method enabled the pathologist to sample a much greater volume of lung tissue. He said that an asbestos body count carried out by that method was therefore much less prone to statistical error than the counting of asbestos bodies in a limited number of routine lung tissue sections. He said that the digestion method provided a much more reliable indicator of the concentration of asbestos bodies in lung tissue.
100. Dr Attanoos agreed that the lung digestion method was the more accurate technique for the reasons given by Professor Roggli and Dr Rudd. However he stressed that this was only the case if a large number of samples of tissue were taken from all parts of the lung. If this was not done, the results would be unrepresentative. He said that, in the deceased's case, he had taken four sections of tissue from different parts of the lung for the purposes of counting the asbestos bodies therein. He found 15 asbestos bodies in 10 sq cm of dry lung tissue which, when averaged out, produced a figure of 1.5 asbestos bodies per sq cm. In using this method, he had followed the criteria set out in both the 1997 Helsinki Consensus Report and the CAP-PPS Report. He agreed that, if he had found 20 asbestos bodies in the same amount of lung tissue, he would have diagnosed asbestosis in the deceased's case.
101. Dr Attanoos used HE-stained sections when examining the deceased's lung tissue for asbestos bodies. This was, he said, because iron staining would have made the tissue unsuitable for the purpose of examining its morphological features. He agreed that iron staining makes asbestos bodies – in particular crocidolite bodies – easier to see than in HE-stained sections. However, he said that he was able readily to identify some asbestos bodies in the sections of lung tissue so that it was not necessary to use iron staining. He said that, if he had not seen any asbestos bodies or if they had been very infrequent, he would probably have used an iron stain to check if any were present. His evidence was that, during the review of the Llandough Hospital asbestosis control group, in cases where asbestos bodies were not evident or were infrequent on HE-stained slides and where more tissue was available, he and his colleagues had examined iron-stained tissue slides to see whether asbestos bodies could be seen by that method. He estimated that this had been done in about 10% of cases. However, since some asbestos bodies were visible in the deceased's lung tissue, he had not done this. He dismissed the suggestion that, if he had examined iron-stained sections, it was very likely that he would have seen more than 20 asbestos bodies (i.e. two asbestos bodies per sq cm) as "purely speculative".
102. In their 1983 Paper (see paragraph 40 above) Professor Roggli and his colleagues suggested that, when HE staining was used, a finding of one asbestos body on histological examination of tissue slides was probably equivalent to about 1,000 asbestos bodies identified by the digestion method. Using iron-staining, the approximate equivalence was one asbestos body on examination of tissue slides to about 800 asbestos bodies by the digestion method. This, Professor Roggli said, demonstrated that iron-staining was the more accurate method. The CAP-PPS Report refers (at page 27) to the two methods:

"The second feature necessary for a histological diagnosis of asbestosis is the finding of asbestos bodies. Asbestos bodies are golden-brown, beaded or dumbbell shaped structures with a thin translucent core (Figure 13 A-D). They form from the deposition of an iron-protein-mucopolysaccharide coating on the surface of an inhaled asbestos fiber by alveolar macrophages. In asbestosis, these bodies are typically found embedded within fibrous tissue, but they may also be observed within alveolar spaces or within the cytoplasm of macrophages or multinucleate giant cells (Figure 14). They are most numerous around the bronchioles but here their presence is often masked by deposits of carbon, their distinction from which is facilitated by the use of iron-stains. Although asbestos bodies are typically formed on amphibole cores (Figure 15 A, B), chrysotile asbestos bodies are also observed in cases when chrysotile-induced asbestos (Figures 16 A-C). Asbestos bodies may also be observed within hilar lymph nodes, but this does not constitute asbestosis.

In the majority of cases, asbestos bodies are readily identified in hematoxylin and eosin-stained sections, and several can commonly be found in a 2x2 cm area of an iron-stained section."

103. This passage tends to support the suggestion that iron staining is the more sensitive method. Professor Roggli accepted that the correlation between the two methods was subject to considerable variation in individual cases. However, he emphasised that, as between the two methods, the digestion method was much more accurate.
104. As I have explained, Professor Roggli examined one of the tissue slides which had been sent to him by Dr Attanoos by light microscopy. The slide was already HE-stained so that he could not use iron staining. In that two sq cm slide, he found six asbestos bodies, the equivalent of 2.8 asbestos bodies per sq cm. He accepted that Dr Attanoos had adopted the correct method in averaging out the results of his asbestos body count over the four tissue slides in order to reach his result of 1.5 asbestos bodies per sq cm of dry lung tissue. However, the fact that he himself had identified six asbestos bodies (i.e. 40% of the total of 15 asbestos bodies found by Dr Attanoos) in just one of the tissue slides (representing 12% of the total material examined by Dr Attanoos) led him to believe that Dr Attanoos had failed to recognise some of the asbestos bodies that were present. This led him in turn to question – not only Dr Attanoos' failure to diagnose asbestosis in the deceased's case – but also whether some of Dr Wagner's 170 cases had been erroneously excluded from the Llandough Hospital asbestosis control group on the ground that it had not been possible to identify two asbestos bodies per sq cm of dry lung tissue.
105. Professor Roggli said that the process of counting the asbestos bodies found on examination of slide tissues had produced uncertainty in the deceased's case. In that situation, more weight should be placed on the results of the examination by the digestion process. The identification of 3,970 asbestos bodies per gram of wet lung tissue using that process was approximately equivalent (on the basis of the 1983 paper) to a finding of four asbestos bodies per sq cm of dry lung tissue by examination of HE-stained tissue slides. This supported his view that Dr Attanoos' examination of the tissue slides had resulted in an underestimate of the number of asbestos bodies present. His examination by EDXA and that of Dr Gibbs had revealed that the majority of the fibres in the deceased's lung tissue were crocidolite fibres. This led him to believe that the underestimate might have been caused by a failure to detect crocidolite fibres on the HE-stained tissue slides.
106. By electron microscopy, Professor Roggli found 60,100 asbestos bodies per gram of wet tissue. His evidence was that this was an unexpected finding, given the finding of 3,970 asbestos bodies per gram by light microscopy. He would usually expect to find about twice as many asbestos bodies by electron microscopy than by light microscopy. In this case, there were about 15 times as many. He said that this could be explained by the presence of crocidolite bodies which were evident on examination by the more sensitive electron microscopy, but not by light microscopy.

The results of examination by electron microscopy

107. The pathologists agreed that the most sensitive investigation – and the means by which the diagnosis can be determined in a "borderline" case – is by electron microscopy. Different methods were used in the two laboratories. Dr Gibbs used transmission electron microscopy, whereas Professor Roggli used SEM. It was common ground that transmission electron microscopy enables the operator to see the very smallest asbestos fibres whereas, with SEM, only asbestos fibres measuring five microns or more can be identified. There is a widely held view – not shared by Dr Attanoos - that asbestos fibres measuring less than 5 microns play no part in the causation of asbestosis.
108. Dr Gibbs found a total amphibole (i.e. crocidolite, amosite and anthophyllite) asbestos count of 5.71 million fibres per gram of dry lung tissue, of which commercial amphiboles (crocidolite and amosite) comprised 5.42 million fibres per gram of dry lung tissue. Professor Roggli found 338,000 asbestos fibres per gram of wet lung tissue. A direct comparison of the findings is not possible because of the different techniques used and the fact that the Llandough Hospital result is expressed in terms of dry lung tissue and the Duke University result in terms of wet lung tissue. Even if two laboratories use the same electron microscopy technique, however, differences in the preparation of the material to be examined and other variations of practice mean that comparisons of their results are not usually helpful. That is why the 1997 Consensus Report advocated the development of individual reference ranges for each laboratory.
109. Professor Roggli's evidence was that the median value of the Duke University laboratory reference range is 206,000 fibres per gram of wet lung tissue. The fifth percentile value is 20,800 fibres per gram. The deceased's asbestos fibre count of 338,000 fibres per gram was more than 16 times the fifth percentile value and significantly higher than the median value. Thus, he said, even after adjustment of the range to take account of those cases which did not comply with the 1997 Helsinki Criteria and/or the CAP-PPS Report, the overwhelming likelihood was that the deceased's asbestos fibre count would be above the fifth percentile value for the range.
110. As I have said, Professor Roggli's evidence was that the relevant group to be used for comparison in the deceased's case was the group of Duke University controls who had been diagnosed as suffering from Grade 4 asbestosis. He had prepared a Table (which I shall call Professor Roggli's Table or the Table) of those cases and had inserted details of the deceased's case on it for comparison purposes. In the four cases where tissue slides were available, he examined them to ensure that the diagnoses had been correct. All four had asbestos body counts well over two asbestos bodies per sq cm of HE-stained dry lung tissue. He was confident that, in the majority of the other cases listed in the Table, the two asbestos bodies per sq cm criterion would have been met had the relevant slides been available. There were only two possible exceptions but their fibre counts had been well within the reference range. There had been another case which he had been a little uncertain about but he was satisfied that it would have met the 1997 Helsinki Criteria.
111. Professor Roggli was questioned closely about his Table and, in particular, about the fact that the number of asbestos bodies per gram of wet lung tissue found on light microscopy in the deceased's case was significantly lower than in all but five of the other 18 Grade 4 cases. He pointed out that, by contrast, only five of the Grade 4 group had higher asbestos fibre counts than the deceased. He referred to his unpublished 2009 paper (see paragraph 77 above) which demonstrated that the deceased's asbestos fibre count fell well within the Duke University laboratory fifth percentile for Grade 4 asbestosis and only slightly below the median value of 398,000 amphibole fibres per gram of wet lung tissue.
112. The fifth percentile value of the Llandough Hospital laboratory range of 80 asbestosis controls is 21.4 million fibres per gram of dry lung tissue. The deceased's commercial amphibole fibre count of 5.42 million fibres per gram, as measured by Dr Gibbs, fell well below that value. Applying 95% confidence limits round the fifth percentile value produced a range of 14-31 million fibres per gram. The deceased's amphibole fibre count fell below even the lower end of that range.
113. Dr Attanoos expressed confidence that the Llandough Hospital reference range provided a reliable indicator as to whether an individual's fibre count put him at risk of developing asbestosis. It was suggested to him that, given the fact that the engineers had estimated the deceased's asbestos dose at a level considerably in excess of the recognised threshold level giving rise to a risk of asbestosis, it was surprising that his fibre count was only approximately a quarter of the Llandough Hospital asbestosis control group's fifth percentile value. This was all the more so since the threshold level related to the risk of clinically evident asbestosis, whereas the fifth percentile value related to all grades of asbestosis. Having expressed his reservations about the validity of the threshold level and of engineers' estimates of asbestos dose, Dr Attanoos said that there was no correlation between estimates of asbestos dose as estimated by an engineer and the results of analysis of asbestos fibre counts. The latter was an objective measurement and the results of fibre counts were to be preferred to subjective estimates made by engineers. Thus he did not find the apparent discrepancy surprising.

The effect of clearance

114. It is generally accepted that, over time, natural mechanisms cause asbestos fibres spontaneously to clear from the lungs. It is widely recognised that the rate of clearance varies as between the different type of asbestos fibres.
115. In their second Joint Statement, the pathologists said at paragraphs 12 and 13:

"12 Fiber clearance of amphibole fibers is prolonged. Neither Professor Roggli nor Dr Attanoos take into account clearance factors when interpreting individual mineral or fiber counts. This is because both Pathologists considered that the effects of fiber clearance would be negligible with respect of the determination of the disease of asbestosis.

13 Fiber clearance factors are not considered to diminish the validity of the asbestosis range in either Professor Roggli's or Dr Attanoos' laboratory."

116. Dr Rudd referred to studies which suggest that half time for fibre clearance (i.e. the time by which the fibre count will fall to half its initial level) is 20 years for amosite, around six-seven years for crocidolite and only a few months for chrysotile. Dr Attanoos did not dispute these periods although he said that both fibrosis and disease – whether asbestosis or UIP – could have the effect of slowing the clearance process.
117. Dr Rudd said that the Llandough Hospital asbestosis controls had had high levels of asbestos exposure many years ago. It was reasonable to assume that the time between the cessation of their asbestos exposures and their deaths was significantly shorter than the period of about 44 years that had elapsed between the deceased's last exposure to asbestos (1962) and his death (2006). Thus, the effects of clearance would have been significantly less in the case of the Llandough Hospital asbestosis controls than in that of the deceased, giving rise to higher asbestos fibre counts among the asbestosis controls. If the deceased had died, say, 20 years earlier than he actually did, and his lung tissue had been analysed, then it is likely that his asbestos fibre count would have been significantly higher than it was in 2006. His crocidolite level would have been about eight times as high and his amosite level would have been about twice as high. His total fibre count would have been approximately 35,000,000 fibres per gram, comfortably within the Llandough Hospital fifth percentile value. Thus, he would have been diagnosed as having asbestosis.
118. Dr Attanoos said that it was the practice for correlative pathological and mineralogical studies to be presented without allowance for fibre clearance. The 1997 Consensus Report did not suggest that each individual case of asbestosis required a back extrapolation of the fibre count to a period of the last asbestos exposure for the individual. The CAP-PPS Report, while referring to clearance, states that the asbestosis reference range for a laboratory refers to the "retained" amphibole fibre counts. He said that back extrapolation of values was not considered to be a scientifically robust approach and any difference in fibre count levels was, in any event, unlikely to be significant.
119. In cross-examination of Dr Rudd, Mr Kent questioned him about a Joint Statement which had been agreed by him and Dr Attanoos in connection with other legal proceedings. The Points of Agreement in that Joint Statement included the following:

"On present evidence the result in [Mr W's] case would appear to be a borderline case, around the lower end of the redefined asbestosis range in the Dock Yard series [i.e. the 47 Devonport dockyard cases]. There is clearance of fibres from the lungs by natural mechanisms over time. This factor is usually ignored when defining the "asbestosis range" because information about when exposure ceased is not available for the subjects from whom the range is defined. In the case of [Mr W] in which, unusually, exposure ceased more than 50 years before death it is reasonable to take this factor into consideration, reinforcing the conclusion that his count should be regarded as approximating to the lower end of the asbestosis range."

120. It is apparent, therefore, that, in that case, Dr Attanoos was prepared to take into account the time which had passed since the individual's last asbestos exposure and the clearance of fibres which would have taken place during that time.
121. Dr Rudd accepted that published ranges do not make any allowance for clearance. He said that it would not be practicable for them to do so since the date of death would not be known in all cases. However, he considered that it was appropriate to take the clearance factor into account when considering the appropriate diagnosis in an individual case.

The balance of risk

122. The chest physicians addressed the relative risk of an individual developing UIP and asbestosis. They agreed that, among the general population, UIP is more common than asbestosis. A recent textbook quoted US mortality rates as 10.7/100,000 per annum for men and 7.4/100,000 for women. The risk of UIP increases with age.
123. The risk of asbestosis arises in a relatively small group of the population, namely those who have been exposed to asbestos at or above the threshold level. Not all those who are exposed to asbestos at and above that level will go on to develop asbestosis. Only a small minority will do so. This is because of the wide range of individual susceptibility that exists. Dr Rudd and Dr Hind were in broad agreement that asbestos exposure at or about the threshold level of 25 fibre/ml years gave rise to a 1% risk of developing clinically evident asbestosis. They agreed also that about 10% of persons exposed to asbestos at levels of 100 fibre/ml years go on to develop clinically evident asbestosis. They agreed that, having been exposed to a level estimated at about 42-47.2 fibre/ml years, the deceased's risk of developing asbestosis would have lain somewhere between 1% and 10%, probably (on the basis of four years' exposure to asbestos) at about 3-4%. It follows therefore that his risk of developing asbestosis was prima facie much greater than his risk of developing UIP.
124. Dr Hind argued that, even taking into account his estimated asbestos dose, the deceased's risk of developing asbestosis that (a) was of the atypical UIP-like type and (b) would lead to death was significantly less than 3-4% and was also less than his risk, as a member of the general population, of developing UIP. He said that UIP is more frequently fatal than is asbestosis. Thus, he said, looking at the balance of risk, it was more likely that the deceased's disease was UIP than asbestosis. Dr Rudd disagreed. He said that, even if one assumed that only 5% of asbestos cases had morphological and clinical features typical of UIP, the balance of risk would still favour asbestosis. Even taking into account the risk of death, the balance of risk faced by the deceased lay in favour of asbestosis.

The opinions of the medical experts

The pathologists

125. I have already set out the views of the pathologists in some detail. Professor Roggli concluded that, despite the admittedly atypical morphological and clinical features of the deceased's case, the history and extent of asbestos exposure, together with his findings of asbestos bodies and fibres, led him to the conclusion that the correct diagnosis was one of asbestosis. Dr Attanoos did not agree. He placed considerable emphasis on the morphological and clinical features although he acknowledged that, if he had observed two asbestos bodies per sq cm on the examination of tissue slides or had found an asbestos fibre count within the Llandough Hospital laboratory reference range, he would have diagnosed asbestosis. In the circumstances, however, he considered that the appropriate diagnosis was UIP.

The chest physicians

126. Dr Rudd expressed the view that the balance of probabilities favoured a diagnosis of asbestosis, rather than UIP, in the deceased's case. He said that, in reaching that conclusion, he had taken into account, first, the deceased's occupational history and the fact that two engineers' estimates, independently conducted, had concluded that he had had sufficient exposure to give rise to the risk of asbestosis. Professor Roggli's findings of asbestos bodies and fibres within the reference range for the Duke University laboratory provided further support. Furthermore (and despite his reservations about the Llandough Hospital reference range), if clearance of asbestos fibres over time was taken into account, the deceased's notional asbestos fibre count would be within that range. He said that he had not overlooked the fact that the morphological features and clinical course were more typical of UIP. However, there were cases of asbestosis where those features were present.
127. Dr Hind's opinion was that the deceased had suffered from UIP. He said that the clinical picture was more in keeping with UIP than with asbestosis. This was supported by the pathologists' evidence relating to the morphological features. Since the pathologists did not agree about the results of the histological investigations, Dr Hind said that he had to fall back on the clinical picture. The rapid clinical course was in favour of UIP. He accepted that, on the basis of the engineers' evidence, the deceased's asbestos exposure had put him at risk of contracting asbestosis. However, he said that few patients exposed to that level of asbestos would go on to develop asbestosis ending in death.

The parties' cases

The claimant's case

128. Mr Allan contrasted the approach of the various medical experts. He submitted that Professor Roggli and Dr Rudd had looked at all the relevant evidence and, having taken it into account, had reached their diagnoses. He suggested that, by contrast, the defendant's experts had taken a somewhat rigid stance. Dr Attanoos' diagnosis appeared to hinge largely on the outcome of the counting of asbestos bodies on tissue slides. His evidence had been that, if he had found an average of two asbestos bodies per sq cm of dry lung tissue, he would have diagnosed asbestosis in the deceased's case. Similarly, if he had not found two asbestos bodies per sq cm on examination of tissue sections but the deceased's fibre count had fallen within the Llandough Hospital laboratory reference range, he would have diagnosed asbestosis. In neither of those circumstances would the atypical morphological features or clinical course have prevented him from making the diagnosis. Mr Allan suggested that this was a very narrow approach. He argued that Dr Hind had taken a similarly narrow approach. He had made clear that, since the pathologists could not agree, he was basing his opinion solely on the clinical course of the deceased's disease, coupled with what he perceived to be the greater risk that the deceased would have developed UIP. Mr Allan suggested that the correct approach – and that advocated in the 1997 Consensus Report and the CAP-PPS Report – was to have regard to all the relevant evidence.
129. Mr Allan submitted that the evidence clearly demonstrated a general acceptance – albeit not, apparently, shared by Dr Attanoos – of the existence of a risk threshold level of 25 fibre/ml years, arguably less for amphibole exposure. He referred to the passage in the 1997 Consensus Report (at page 311) which emphasised that considerable weight should be attached to reasonably reliable evidence about an individual's exposure. He said that it was clear from the engineers' evidence that the deceased's exposure had been over the threshold level and had put him at risk of developing asbestosis.
130. Mr Allan contrasted the results of Dr Attanoos' counting of asbestos bodies visible on tissue slides with Professor Roggli's count using the lung digestion method. He argued that the latter was a far more reliable technique, involving as it did the examination of considerably more tissue than the sections examined by Dr Attanoos which were only five microns thick. He questioned why Dr Attanoos had not gone on to examine iron-stained tissue slides, despite the fact that he had accepted that iron staining made asbestos bodies easier to identify. He referred to Professor Roggli's finding of over 60,000 asbestos bodies by electron microscopy which, he said, provided further support for the suggestion that Dr Attanoos' asbestos body count had been an underestimate.
131. Mr Allan suggested that little weight could be attached to the Llandough Hospital laboratory reference range for the reasons to which I have already referred. He stressed that it was essential that a reference range should contain a sufficient number of cases and as broad a range as possible of the levels of exposure capable of causing asbestosis. The available evidence did not suggest that the Llandough Hospital laboratory reference range fulfilled those requirements. It was likely that, as a result, the minimum and fifth percentile values were set far too high. In any event, the small number of asbestosis controls meant that no reliance could be placed on the fifth percentile or minimum values. By contrast, the Duke University laboratory reference range contained larger numbers of controls and a greater range of occupations and levels of asbestos exposure. This meant that its fifth percentile value was likely to be more reliable. The reference range included some more recent deaths. It was true that some of the controls would not satisfy the new requirement for alveolar septal fibrosis or the "two asbestos bodies per sq cm" requirement. However, it was significant that the deceased's fibre count and asbestos body count both fell within the Duke University laboratory fifth percentile value for Grade 4 cases and exceeded by a very wide margin indeed the fifth percentile value for all cases.
132. Mr Allan said that it was clear from the evidence that the deceased had been exposed mainly to crocidolite, together with some amosite. That was, he said, highly relevant when considering the issue of clearance, since clearance of crocidolite takes place faster than clearance of amosite. Dr Rudd's evidence had made clear that, two decades ago, the deceased's fibre count would have been very substantially greater than it was at the time of his death – high enough to have exceeded comfortably the fifth percentile value of the Llandough Hospital laboratory reference range. The length of time that had elapsed between the deceased's asbestos exposure and his death was unusually long, allowing clearance to take place over more than four decades. He said that this was an important factor to take into account.
133. Mr Allan submitted that I was entitled to take account also of the relative risks of the deceased developing asbestosis and UIP. He said that, if one looks at the relative risks in the population who have been heavily exposed to asbestos, the risk of disabling asbestosis must be greater than the risk of UIP. He said that the literature dealing with diagnosis stated that UIP (or IPF/CFA) should be diagnosed only when all other possible causes for the diffuse interstitial fibrosis have been excluded. That, he said, was because, if there had been significant exposure to asbestos, the fibrosis was more likely to result from asbestosis.
134. Mr Allan said that it was clear from the literature that the typical morphological features and rapid clinical course of UIP could sometimes be seen in cases of asbestosis. Professor Roggli and Dr Rudd had taken that into account but had, nevertheless, concluded that the appropriate diagnosis was asbestosis. That would have been Dr Attanoos' diagnosis also, had the deceased's case met the other requirements of the 1997 Helsinki Criteria.

The defendant's case

135. Mr Kent accepted that, despite the uncertainties and imprecision, I should proceed on the basis of the asbestos dose estimated by the engineers. He accepted also that the generally accepted threshold level for the risk of asbestosis is 25 fibre/ml years. He said that the deceased's asbestos dose had not been very much in excess of that threshold level. He urged me to bear in mind that most people exposed to that asbestos dose do not go on to develop asbestosis. He emphasised that the deceased had been exposed only for a short period. He suggested that he would have been "at the lower end of the bracket of risk".
136. Mr Kent argued that the approach adopted by the Llandough Hospital laboratory of including within their reference range only those cases that complied with the 1997 Helsinki Criteria " (in particular, the "two asbestos bodies per sq cm" requirement) must be correct. Strict adherence to the 1997 Helsinki Criteria meant that all the control cases were "true" asbestosis cases. He submitted that, while some of the Llandough Hospital laboratory asbestosis controls may have had very high levels of asbestos exposure which would affect the median value for the reference range, the fifth percentile value would not be affected since there were some cases of lighter exposure within the control group. He emphasised that the deceased had failed to meet the "two asbestos bodies per sq cm" criterion and that his fibre count had fallen well below the fifth percentile for the group, even with a 95% confidence interval.
137. Mr Kent went on to refer to the criticisms of the Duke University laboratory reference range to which I have already referred. He submitted that Professor Roggli had not initially carried out an examination of the tissue slides in the deceased's case as required by the 1997 Helsinki Criteria. When he did, he looked at only one slide. He had sought to quantify the asbestos bodies in the deceased's lung tissue using the digestion method which was not explicitly allowed for by the 1997 Helsinki Criteria. He submitted that the 1983 Paper which had formed the basis of the correlation between the number of asbestos bodies found by examining tissue slides and the number found by the lung digestion method was flawed because it was based on only six cases and the correlation could be seen only on logarithmic analysis. He suggested that Professor Roggli had used an insufficient amount of tissue for the lung digestion exercise. The techniques used at the Duke University laboratory were, he said, generally less rigorous than those used at the Llandough Hospital laboratory. He said that the risk of an unreliable fifth percentile or median value was therefore much greater for the Duke University laboratory reference range.
138. Mr Kent said that it was agreed that the morphological features and the course of the deceased's disease were against a diagnosis of asbestosis and in favour of a diagnosis of UIP. There was disagreement as to whether his asbestos body and fibre counts were within the range which would put him at risk of developing asbestosis. He submitted that, in those circumstances, the inevitable conclusion must be that the appropriate diagnosis was UIP. He suggested that the only way I could conclude on a balance of probabilities that he had suffered from asbestosis would be if I were to find that he had been at a higher risk of developing atypical asbestosis leading to death than of developing fatal UIP. He said that, while Dr Rudd may be correct in saying that the risk of developing fatal asbestosis had been higher than the risk of contracting fatal UIP, the matter was so uncertain that it could not properly provide the basis for a finding on the balance of probabilities.

Discussion and conclusions

The approach

139. There are two possible explanations for the diffuse interstitial fibrosis which resulted in the deceased's death. Either it resulted from his asbestos exposure or it had some other unknown cause. My task is to decide whether the claimant has succeeded in establishing, on the balance of probabilities, that the deceased was suffering from asbestosis. I shall endeavour to do that by considering all the available evidence, including the extent of his asbestos exposure, the risk to which he was exposed, the clinical and morphological features and the results of the histological investigations carried out after his death.

Exposure and asbestos dose

140. I have found that the deceased was exposed to asbestos for a period of four years, between 1958 and 1962. The consulting engineers estimated his asbestos dose at, respectively, 42 and 47.2 fibre/ml years.
141. Dr Attanoos' comments about the shortcomings of engineers' estimates of asbestos dose clearly have some validity. They are estimates, rather than exact measurements. However, the calculation of asbestos dose in a case such as this is a well recognised exercise. The engineers had access to some measurements recording the levels of asbestos dust produced by tasks which were undertaken by the deceased. Both engineers drew attention in their reports to the difficulties associated with calculating asbestos dose and warned that caution should be exercised both in making and interpreting their calculations. In their Joint Statement, however, they noted that their final estimates, which they had revised as a result of the information about dust measurements, were very similar. Given their considerable joint experience in the field of asbestos exposure, this gives me confidence that, whilst their estimates do not represent a completely accurate measurement of the level of asbestos inhaled by the deceased, they represent the best available indication of the extent of his exposure.
142. It is now over 25 years since the threshold of 25 fibre/ml years giving rise to a risk of asbestosis was first promulgated. It is clear from the CAP-PPS Report that it is still widely accepted as the level over which asbestos exposure gives rise to the risk of the development of clinically evident (i.e. Grade 3 or Grade 4) asbestosis. It seems that there is research in progress which may lead to the setting of different threshold levels for different types of asbestos in the future. For present purposes, however, the Commission's threshold level remains the recognised standard for asbestos exposure in general and I consider that it is entirely appropriate to adopt it. The deceased's exposure was significantly in excess of the threshold level. I accept of course that only a small minority of individuals exposed to such a level of asbestos will go on to develop asbestosis.

The radiological findings

143. The radiological findings, including the presence of pleural plaques, appear to me to be neutral. I do not consider that they assist me in reaching a conclusion as to the causation of the deceased's disease.

The morphological features and the clinical course of the disease

144. The pathologists agreed that the morphological features of the deceased's lung tissue favoured a diagnosis of UIP. The chest physicians agreed that the rapid clinical course of his disease was also more characteristic of UIP than of asbestosis. These factors are linked, in that it is the typically aggressive nature of UIP that produces both its rapid course and the inflammation that results in its characteristic morphological appearance. Professor Roggli's evidence was that, in some individuals, asbestosis can progress aggressively, producing both a rapid clinical course and morphological features typical of UIP. Dr Rudd, who has very considerable clinical experience of both asbestosis and IPF/UIP, said that there were cases in which asbestosis progressed as quickly as had the deceased's disease.
145. Dr Attanoos expressed considerable scepticism as to whether asbestosis could ever cause the morphological features or rate of progression typical of UIP. He suggested that cases which in the past had been classified (even in his own laboratory) as being examples of asbestosis with UIP-like morphological features and clinical course had in fact been wrongly categorised. He considered that they were more likely in reality to be UIP cases where the individual had coincidentally been exposed to occupational levels of asbestos.
146. I recognise the possibility that Dr Attanoos' thesis may, in the future and after further research, prove to be correct. However, the literature to which I referred earlier in this judgment – in particular, the authoritative CAP-PPS Report – indicates that there are cases where, for reasons that are not fully understood, asbestosis is accompanied by morphological features and a clinical rate of progression which are typical of UIP. I am satisfied that that represents the view currently held by the majority of those practising in this field. In such cases, the two conditions can only be distinguished by examining lung tissue for the presence of levels of asbestos fibres and/or bodies sufficiently high to justify a diagnosis of asbestosis.

The presence of asbestos bodies

147. The 1997 Helsinki Criteria for the diagnosis of asbestosis were (1) the identification of diffuse interstitial fibrosis and (2) the presence of two or more asbestos bodies per sq cm of dry lung tissue or a count of asbestos fibres falling into the range recorded for asbestosis by the relevant laboratory. The CAP-PPS Criteria are (1) an acceptable pattern of alveolar septal fibrosis and (2) an average of at least two asbestos bodies per sq cm of dry lung tissue. The CAP-PPS Report makes clear that the presence of fewer than two asbestos bodies per sq cm does not necessarily exclude a diagnosis of asbestosis. In such a case, quantitative studies performed by lung digestion will be necessary, together possibly with asbestos fibre analysis.
148. There is no dispute that the deceased fulfilled the first requirement of both sets of criteria. The second requirement (two asbestos bodies per sq cm) has assumed considerable significance in this case. Dr Attanoos made clear that, if he had detected an average of two asbestos bodies per sq cm in the tissue slides he examined, he would have made a diagnosis of asbestosis in the deceased's case. He would have made that diagnosis irrespective of his views on the morphological features and the rapid clinical course of the deceased's disease. He would have done so in accordance with the 1997 Helsinki Criteria and CAP-PPS Criteria.
149. I can well understand why the "two asbestos bodies per sq cm" criterion was introduced. Examination of tissue slides by light microscopy is a relatively simple and inexpensive process which can be performed in most laboratories world wide. In many cases of prolonged and heavy asbestos exposure, the required number of asbestos bodies will be readily visible by this method, permitting immediate diagnosis of asbestosis. It is clear, however – in particular from the CAP-PPS Report – that the method is not well suited to cases where the asbestos exposure was not so heavy or prolonged or where the individual was a "poor coater" of asbestos fibres. Given the wide variations in the distribution of asbestos bodies within the lungs, examination of only a small amount of tissue from a few sites on tissue slides can lead to an underestimate (or, indeed, an overestimate) of the concentration of asbestos bodies. In addition, it can be difficult to detect asbestos bodies – particularly crocidolite - by this method. These difficulties are increased if HE staining – rather than iron staining –is used.
150. Dr Attanoos' examination of the tissue slides revealed the presence of 15 asbestos bodies in 10 sq cm of lung tissue. The examination was carried out using a HE stain. Having observed less than two asbestos bodies per sq cm, Dr Attanoos did not proceed to examine any slides using an iron stain. I found his reasons for not doing so difficult to understand, especially since he said that, during his review of the Llandough Hospital asbestosis controls, he had, in "borderline" cases, proceeded to examination using iron staining whenever the necessary additional lung tissue was available. There is, therefore, no evidence of the average number of asbestos bodies per sq cm that would have been visible on iron-stained slides.
151. I am satisfied that by far the more accurate technique for identifying and quantifying asbestos bodies in lung tissue is by means of the lung digestion method used at the Duke University laboratory. I have referred to the work carried out by Professor Roggli and his team on the correlation between the number of asbestos bodies found by the lung digestion method and by examination of tissue slides. It was that work that gave rise to the adoption in the 1997 Consensus Report and the subsequent CAP-PPS Report of the "two asbestos bodies per sq cm" criterion. It is, in my view, highly significant that, in the CAP-PPS Report, it is generally recognised that, in cases where fewer than two asbestos bodies per sq cm are found on examination of tissue slides, it will be necessary to proceed to quantification of asbestos bodies by the digestion method. That confirms that the digestion method is generally recognised to be the more sensitive and accurate technique.
152. Using the lung digestion method, Professor Roggli found 3,970 asbestos bodies per gram of wet lung tissue. On the basis that 1,000 asbestos bodies per gram found by lung digestion are equivalent to one asbestos body per sq cm of HE-stained dry lung tissue, his finding represented about four asbestos bodies per sq cm seen on examination of tissue slides. I take into account the criticisms made by the defendant of the work which forms the basis for the correlation between the concentrations of asbestos bodies found by the two methods, in particular the fact that the relevant study was a very small one. However, it seems that the correlation is widely accepted by practitioners in the field. It formed the basis for the second criterion of both the 1997 Helsinki Criteria and the CAP-PPS Criteria. It is referred to with apparent approval in the CAP-PPS Report, of which Dr Attanoos was a co-author. Furthermore, it is important to remember that it was the "two asbestos bodies per sq cm" that was derived from the findings on examination by the lung digestion method and not vice versa. For these reasons, I am satisfied that Professor Roggli's finding of 3,970 asbestos bodies per gram of wet lung tissue (if accurate) strongly suggests that Dr Attanoos' count of asbestos bodies on tissue slides was an underestimate. That would not be surprising since we know from the EDXA examinations carried out by Professor Roggli and Dr Gibbs that a high proportion of the asbestos bodies present in the deceased's lungs were crocidolite bodies, which are difficult to identify on tissue slides.
153. It is suggested on behalf of the defendant that Professor Roggli's count may well not be accurate. He is criticised for having used tissue from only one of the blocks sent to him by Dr Attanoos for his examination by the digestion method. It is said that this may have given an unrepresentative result. For the claimant it is said that the amount of tissue examined by Professor Roggli was significantly greater than that contained in the four slides, each five microns thick, examined by Dr Attanoos. In this context, it is relevant to consider the Duke University laboratory finding of 60,100 asbestos bodies per gram of wet lung tissue on electron microscopy. Professor Roggli's evidence was that this was far in excess of what would be expected, given the concentration of asbestos bodies found by light microscopy. It suggested that the light microscopy count had seriously underestimated the concentration of asbestos bodies in the deceased's lung tissue.
154. Looking at the picture as a whole, I am satisfied that Professor Roggli's findings demonstrate that the deceased had in his lungs a concentration of asbestos bodies which was significantly in excess of the "two asbestos bodies per sq cm" criterion or its equivalent by the digestion method. The suggestion that, because of the amount of lung tissue Professor Roggli used, his light microscopy count may have been an overestimate is completely inconsistent with the results of his examination by electron microscopy, which suggests that it was quite the opposite. It is probable in my view that the underestimates (as I am satisfied they were) produced by Dr Attanoos's examination of tissue slides and Professor Roggli's examination by the lung digestion method resulted from the difficulties in identifying crocidolite bodies and, possibly, from an inability by the deceased's body efficiently to coat asbestos fibres. It is impossible to say whether the fact that Professor Roggli found as many as six asbestos bodies on only one of the four HE stained tissue slides examined by Dr Attanoos means that Dr Attanoos failed to see asbestos bodies which should have been visible even without iron staining. It may just be that Professor Roggli happened to select a slide where the concentration was unusually high. However, it is in my view probable that, if Dr Attanoos had used the more sensitive technique of iron staining when examining the lung tissue slides, he would have detected a concentration of asbestos bodies which would have fulfilled the 1997 Helsinki Criteria and the CAP-PPS Criteria. In that event, he would have diagnosed asbestosis.

The presence of asbestos fibres

155. I come now to the fibre analyses carried out at the Llandough Hospital and Duke University Laboratories. The reliability of a laboratory's reference range inevitably depends on the extent to which the control group is truly representative of the spread of severity of disease and of the different extents of asbestos exposure experienced by a sufficient number of individuals who have been accurately diagnosed as suffering from asbestosis. I have already referred to the criticisms made by the respective parties of the asbestosis control groups that provide the references ranges for the Duke University and Llandough Hospital laboratories.
156. It is clear from the evidence that the two laboratories have adopted different approaches towards the acceptance of cases into their asbestosis control group. The Llandough Hospital laboratory uses the 1997 Helsinki Criteria and precludes the entry of any case which does not fulfil the "two asbestos bodies per sq cm" requirement, irrespective of the asbestos fibre count found on analysis. Dr Attanoos explained that such cases were excluded from entry so as to ensure that all cases within the asbestosis control group were "true" cases of asbestosis.
157. The practice of examining tissue slides by light microscopy is not generally used at the Duke University laboratory. The "two asbestos bodies per sq cm" criterion is not, therefore, applied. Instead, asbestos bodies are identified and quantified by means of the lung digestion method. Inclusion in the asbestosis control group is determined by looking at all the available evidence (including the results of investigations by light and electron microscopy) and reaching a conclusion as to whether a diagnosis of asbestosis can be justified. If it can, the case is included within the control group. No specific criteria are applied. Professor Roggli explained that the emphasis at the Duke University laboratory was to achieve maximum sensitivity (i.e. to pick up all asbestosis cases right across the spectrum of extent of exposure and severity of disease). He contrasted this with the approach at the Llandough Hospital laboratory where, he suggested, the emphasis was on specificity (i.e. on ensuring that every case within the control group was a "true" case of asbestosis). He observed that the tension between specificity and sensitivity was a common dilemma in medical research and epidemiology.
158. The Llandough Hospital asbestosis reference range is of relatively recent origin. Until 2006/2007, its minimum value, in a case of Grade 1 asbestosis, was a fibre count of 50 million fibres per gram of dry lung tissue, i.e. well over twice its current fifth percentile value of 21.4 million fibres per gram. The current fifth percentile value was adjusted upwards from 19.9 million fibres per gram of dry lung tissue in about 2008 after the addition of the 33 further cases. The asbestosis control group as a whole has not been the subject of any published research. However, the controls themselves have been subject to peer review with a view to ensuring that they are "true" asbestosis cases. I share the surprise expressed by others that the original 170 Devonport dockyard cases diagnosed by Dr Wagner as asbestosis have been whittled down by the review process to only 47 cases, although I have no basis for finding that cases were wrongly excluded from the new asbestosis control group. The effect has been, however, that, even with the addition of the 33 cases identified by Dr Attanoos, the size of the control group is not, for the reasons advanced by Dr Rudd, sufficient to provide a statistically reliable fifth percentile or minimum value.
159. It is accepted that, of the 80 control cases, the 33 added most recently had had very heavy exposure to asbestos. This would not be true of all the original 170 Devonport dockyard cases, some of whom are known to have had less heavy exposure than others. However, the fact that most of the dockyard control subjects died more than 20 years ago (some considerably earlier), and are likely therefore to have been exposed to asbestos in the 1940s, 1950s and 1960s, would tend to suggest that their levels of exposure would have been higher than those encountered in later years. Moreover, there was no evidence before me as to how many of those individuals (if any) who had had less heavy exposure to asbestos were included within the final 47 control cases. The fact that 31 out of the 80 Llandough Hospital asbestosis controls had been classified as having Grade 4 disease (compared with only 18 out of the 192 Duke University asbestos controls) tended to confirm that the group included many individuals who had experienced high levels of asbestos exposure.
160. The application of the criterion of "two asbestos bodies per sq cm" on examination of tissue slides has probably led to some "true" cases of asbestosis (particularly asbestosis caused or contributed to by crocidolite exposure) being excluded from the Llandough Hospital asbestosis control group. I have already referred to the difficulties of identifying asbestos bodies (particularly crocidolite bodies) by that method and to the difference in the readiness with which individuals are able to "coat" asbestos fibres so as to form asbestos bodies. There was no system at the Llandough Hospital laboratory for "checking" the asbestos body concentration in "borderline" cases by the lung digestion method. Such "checking" is, as I have said, advocated in the CAP-PPS Report, which makes clear that the presence of fewer than two asbestos bodies per sq cm does not necessarily exclude a diagnosis of asbestosis. The cases excluded from the Llandough Hospital asbestosis control group on the basis that they did not comply with the "two asbestos bodies per sq cm" criterion may well have been "borderline" cases in which the asbestos fibre count would have been towards the bottom end of the reference range and would therefore have had an effect on both the fifth percentile and/or minimum value for the range.
161. In short, I cannot be confident, on the basis of the available evidence, that the Llandough Hospital laboratory reference range contains a sufficiently representative range of controls (in terms of types of employment and/or levels of asbestos exposure) to provide a reliable fifth percentile or minimum value. Indeed, it is highly likely in my view that those values are set too high, probably significantly too high. That is leaving aside the issue of statistical unreliability to which I have already referred.
162. I have some reservations also about the reference range which has been established at Duke University. It is accepted by Professor Roggli that a significant number (unknown, but possibly as many as 30 or more) of his asbestosis control cases would not meet the CAP-PPS Criteria for asbestosis by reason of the absence of alveolar septal fibrosis. Although it may well be, as Professor Roggli suggested, that the removal of those cases would have little effect on the fifth percentile value of the reference range, it is impossible to be entirely confident about that until the necessary adjustments have been made.
163. The absence of specific criteria for inclusion in the control group may also be a weakness. I can well understand why Professor Roggli and his team consider the "two asbestos bodies per sq cm" criterion too blunt and imprecise an instrument to be used as a requirement for entry into the asbestosis control group. I can see also the good sense in looking at all the available evidence before deciding whether a case justifies a diagnosis of asbestosis and therefore inclusion in the asbestosis control group. By that means, one would hope to include within the group the widest spectrum of asbestosis cases. However, there is, as Dr Attanoos pointed out, a certain circularity in using the size of an individual's asbestos fibre count (among other considerations) as a reason for including him in a group of "asbestosis" controls when the very purpose of that control group is to establish the range of asbestos fibre counts in cases of asbestosis. These competing considerations present difficult questions to which there is no easy answer.
164. The system used by the Duke University laboratory depends very much on the accuracy of the diagnosis in the asbestosis control cases. This fact is linked to the further criticism made by the defendant, namely that there has been no peer review of the diagnosis of asbestosis in the Duke University control cases. I can understand Professor Roggli's concerns about the potential adverse effects of diagnosis "by committee". His contention is that the fifth percentile is used in order to allow for any misdiagnoses at the bottom end of the range. I accept that point. Nevertheless, it would in my view be possible to have greater confidence in the reliability of the Duke University laboratory reference range – particularly the very bottom end of the range – if the diagnosis of asbestosis in the control group had been subject to independent verification by one of Professor Roggli's peers.
165. I consider that there are a number of important respects in which the Duke University laboratory reference range is to be preferred to that of the Llandough Hospital laboratory. Its asbestosis control group is significantly larger than that at the Llandough Hospital laboratory. The 192 cases currently comprised in the group are only just below the number identified by Dr Rudd as required for a reliable indication of the fifth percentile value. With the removal of 30 or so cases, however, its statistical reliability will inevitably be reduced to some extent. It is clear from the published literature that the Duke University asbestosis controls cover a wider range of different employments (and, therefore, in all probability, levels of asbestos exposure) than do the Llandough Hospital controls. Provided that the Duke University laboratory controls have been correctly diagnosed as asbestosis cases, they should provide a wider spectrum of asbestos fibre counts found in sufferers from all grades of asbestosis and thus constitute a far more representative mix of cases than is found in the Llandough Hospital reference range. The Duke University asbestosis control cases have been collected over a number of years. They have been the subject of work which has been described in peer-reviewed publications and which (certainly in the case of the 2008 study: see paragraph 71 above) has shown precisely the trends which would have been expected from the decrease in asbestos use over time. The fact that the control group includes individuals who died relatively recently (some in the 2000s) is also an advantage.
166. In making these observations about the reference ranges of the two laboratories, I intend no criticism at all of Professor Roggli, Dr Attanoos or their respective teams. I recognise that the task of compiling an appropriate reference range is complex and that resources are limited. It is however, necessary, for the purposes of this case, for me to give anxious consideration to the extent, if any, I can rely on those reference ranges – or either of them – when reaching my conclusions in this case.
167. As will be clear from what has gone before, I have considerable doubts about the reliability of the fifth percentile and minimum values of the Llandough Hospital reference range. It is highly probable, in my view, that they are set too high, possibly significantly too high. I do not regard them as a reliable indicator of the lowest concentrations of asbestos fibres that give rise to a risk of asbestosis.
168. It is possible in my view that the Duke University values are set too low. I take the view that they should be approached with a degree of caution.
169. As Professor Roggli emphasised, however, it is important to keep in mind that the deceased's asbestosis was classified as Grade 4. Whilst that does not necessarily mean that, in an individual case, the level of asbestos exposure (and the asbestos fibre count) will be high, there is a broad correlation between severity of disease and the level of exposure and/or fibre count. This is borne out by the fact that, while the Duke University median value for all grades of asbestosis is 206,000 fibres per gram of wet lung tissue, the median value for Grade 4 cases is 398,000 fibres per gram. At 338,000 fibres per gram, the deceased's fibre count was very close to the Duke University median value for Grade 4 cases, and significantly above the fifth percentile value for such cases. Any misdiagnoses which have occurred amongst the controls are likely to have occurred amongst the cases of Grade 1, possibly Grade 2, asbestosis, rather than the Grade 4 cases. They would be unlikely materially to affect the fifth percentile for Grade 4 cases.
170. It is true that the Duke University reference range contained only 18 Grade 4 cases which is a small number for statistical purposes; the figures must therefore be treated with some caution. Nevertheless, the fact that the deceased's fibre count (a) exceeded the fifth percentile for the whole reference range by a factor of over 16 and (b) exceeded the fifth percentile value for the Grade 4 cases (it was the sixth highest of the 18 cases) and was relatively near the median value for the Grade 4 cases does provide support for Professor Roggli's contention that, whatever adjustments were made to his asbestosis control group, the deceased's fibre count would – at the very least – fall within the fifth percentile of the reference range.

Clearance

171. It is necessary also to consider the effects of clearance. There is no dispute between the experts that clearance of asbestos fibres from the lungs occurs and at a rate which differs as between the various types of asbestos. Dr Rudd's evidence about the rates of clearance for crocidolite and amosite was unchallenged. It highlighted the fact that an individual's asbestos fibre count can reduce very significantly with the passage of time. Thus, one can have a situation whereby an individual who dies 40 years after his last exposure has an asbestos fibre count which falls below a laboratory's reference range whereas, had he died 20 years earlier, the fibre count would have been well within that range. This is particularly so if a high proportion of his exposure has been to crocidolite which clears significantly faster than amosite. I note that, in the case of Mr W (who died 50 years after his last asbestos exposure), Dr Attanoos agreed that it was reasonable to take the clearance factor into consideration. Mr W's case was a borderline case, falling below the (then) fifth percentile for the Llandough Hospital reference range. Dr Attanoos agreed that the effect of clearance enabled his asbestos fibre count to be treated as falling at the lower end of the reference range.
172. The reference ranges of the Llandough Hospital and Duke University laboratories make no allowance for clearance and none is advocated by the 1997 Consensus Report or the CAP-PPS Report. Plainly, it would be impossible to make such an allowance in the absence of accurate evidence as to the period which had elapsed since each control subject's last asbestos exposure and the type(s) of asbestos to which he had been exposed. Dr Attanoos made the point that it would be impermissible to compare an extrapolated fibre count in the case of an individual with the reference range for an asbestosis control group for which no such extrapolation has been made. It would not be comparing like with like. I can see the force of that contention. However, in the case of most of the Llandough Hospital asbestosis controls, the period between their last exposure to asbestos and their death is likely to have been significantly shorter than in the deceased's case. That is because of the heaviness of their exposure (which is likely to have resulted in an earlier death) and its duration, which was often significantly longer than that of the deceased. The opportunity for clearance to occur in the case of those controls would have been significantly less than for the deceased. In other words, their asbestos fibre counts would be significantly higher than if, like the deceased, they had survived for 44 years after their last asbestos exposure.
173. The deceased's long period of survival after his last asbestos exposure makes his a somewhat unusual case. I am satisfied that, in the intervening years, the number of commercial amphibole fibres - both crocidolite and amosite but especially crocidolite - in the deceased's lungs will have reduced very considerably. I accept Dr Rudd's evidence that, had he died 20 years before he did, his asbestos fibre count would have been within – probably well within – the fifth percentile value for the current Llandough Hospital laboratory reference range, a value which, as I have said, is in any event probably set too high.

The balance of risk

174. I accept the quantification of the risk of developing asbestosis at approximately 1% for an individual who has had a dose of 25 fibre/ml years and about 10% in the case of a person who has had a dose of 100 fibres/ml years. The deceased's risk of developing asbestosis was, on the basis of the engineer's estimates, between 3% and 4%. It seems to me likely that Dr Rudd is correct in saying that, even taking into account the fact that the deceased would have been at a lower risk of developing (a) asbestosis with atypical features and (b) asbestosis of such severity as to cause his death, his risk of developing asbestosis would still have been greater than his risk of developing UIP. However, the evidence about the balance of risk was not detailed and I am not sufficiently confident about the issue to rely upon it when reaching my final conclusion. As it happens, I do not need to do so.

Final conclusion

175. The extent of the deceased's asbestos exposure (which, although relatively short, was sufficient to take him well over the recognised threshold level giving rise to the risk of developing asbestosis), the presence (as I find) of very significant numbers of asbestos bodies in his lung tissue, and his asbestos fibre counts (which are well within the reference range for the Duke University Grade 4 asbestosis control cases and which, when the effects of clearance are taken into account, would be within the (probably too high) Llandough Hospital reference range) make me entirely satisfied, on a balance of probabilities, that his fibrosis resulted from his asbestos exposure. In reaching that conclusion, I have taken into account the fact that the morphological and clinical features were more typical of UIP. It is clear from the literature that, for reasons which are not understood, these features are sometimes present in cases of asbestosis. I am satisfied that the deceased's was one of those cases.
176. There will therefore be judgment for the claimant in the sum of £100,000.