England and Wales Court of Appeal (Criminal Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CRIMINAL DIVISION)
ON APPEAL FROM BIRMINGHAM CROWN COURT
(Mr Justice Owen)

		Royal Courts of Justice Strand, London, WC2A 2LL
		20th December 2002

Lord Justice Dyson: giving the judgment of the court

Outline of the case

1. The appellant appeals against her conviction on 19 July 2000 at Birmingham Crown Court for the murder of her husband, Dominic. The prosecution case was that she caused his death by injecting insulin on the night of 30 January 1997. He died a few days later on 9 February. By the morning of 31 January, he was in a coma. There is no dispute that this was caused by hypoglycaemia (low blood glucose or sugar). He never regained consciousness. His death was caused by adult respiratory distress syndrome (ARDS). It was common ground that the hypoglycaemia contributed to the ARDS and, therefore, to his death. The cause of the hypoglycaemia was a central issue at the trial. It was the prosecution case that the hypoglycaemia was caused by injected insulin (referred to as "exogenous" insulin).
2. The deceased was 35 years of age at the time of his death. He and the appellant had been married for 12 years. He was paralysed from the waist down as a result of a traffic accident in 1984. He was a large man, weighing 20 stones. He was a qualified nurse. He had obtained university qualifications and had become manager at the Kingfisher Centre in Peterborough, which served disabled people. He did not suffer from diabetes. The appellant was a ward sister at Kettering Hospital. In 1993, she had given birth to their son, Tadgh. He was a much-loved child of the family, which, so far as was disclosed by the evidence that was adduced at the trial, was stable and happy. The deceased played a full part in caring for the child. The family lived at 55 Main Street, Stonely. There was no evidence that the appellant had any motive for the murder, or to suggest that the deceased might have committed suicide, or that he was killed by the act of anyone else.
3. The jury heard a great deal of medical evidence, some of which was extremely complex. The prosecution case was based on the fact that the appellant was experienced with and had access to insulin, and above all, on the scientific evidence for the administration of exogenous insulin. We have heard yet further scientific evidence to which we shall come in due course. It is sufficient at this stage to say that central to the prosecution case were two immuno-assay tests carried out on serum produced from a sample of the deceased's blood taken at 11.40 am on 31 January. The tests were performed by Dr Teale and Mr Hiscutt. Dr Teale's results showed an inappropriately high concentration of insulin of 887 picamoles/litre, and a very low concentration of an enzyme known as C peptide, which was less than the level of measurability which was set at 94 picamoles/litre. A somewhat different test was carried out by Mr Hiscutt: he found a high insulin level of 616, and an "undetectable" level of C peptide. On the basis of these results, the expert witnesses called on behalf of the prosecution said that they were sure that the insulin was exogenous. The body, by its pancreas, produces insulin and C-peptides in equal amounts, whereas exogenous or bottle insulin does not contain C-peptides at all. The appellant's expert, Professor Marks, disagreed with the prosecution experts. His opinion was that, although the results of the two tests were consistent with the prosecution theory of exogenous insulin, they did not by any means prove it. Since there was no direct evidence implicating the appellant in the death of her husband, the importance of these tests was obvious. Owen J, the trial judge, pointed out that without the results of these tests, there would have been no prosecution case. We shall return to the tests when we discuss the scientific evidence in greater detail.
4. Before we come to the grounds of appeal, we need to describe the events of late January and early February 1997.

Events

5. The deceased had been suffering from influenza in late January. Some witnesses who saw him on 30 January said that he seemed well, although one of them (Anthea Taylor) thought that he was probably still suffering from influenza. Patrick McCarthy, one of his brothers, said that the deceased had told him that day that he was tired and was going to bed. His other brother Albert said that the deceased telephoned him at 8pm and said that he was well but "knackered". They agreed to meet the following day. The deceased also telephoned his mother at 9.06 pm. She said that he seemed to be his usual self, although he said that he was tired. It is likely that he had his evening meal between 7.30 and 9pm.
6. The appellant had agreed with Mary Newall, a colleague, that they would go to a medical party at Kettering hospital that evening, and that she would stay overnight with the Newalls. The appellant told him to go to bed. She left for the party at about 8.45 pm. She called at the Newalls' house (25 miles away) at about 9.15. At 9.20, the deceased telephoned his brother Patrick and said that he was going to bed because he felt extremely tired. At 9.30, Mr Newall drove his wife and the appellant to the party at the hospital. In Mrs Newall's presence, she telephoned the deceased, who, she told Mrs Newall, had said that he was feeling slightly better. Various witnesses confirmed that she was at the party. Mrs Newall was with her most of the time.
7. At 0.39 hrs in the morning of 31 January, the appellant and Mrs Newall took a taxi and returned to the Newalls' home. They went to bed at about 3 am. By the time Mr Newall had got up at 6.50 am, the appellant had left the house. She was required to be on duty at 7 am. Between 8.53 and 9.36, the appellant made 10 telephone calls in quest of her husband. She called the nursery. As a result, Ms Jane Lancaster-Adlam went from the nursery to the family home. Tadgh passed the keys to her so she could let herself in. She found the deceased, apparently unconscious, on the bed upstairs. She telephoned the appellant, who told her to contact their GP, Dr Roberts. Dr Roberts went to the house straightaway. He thought that the deceased had aspirated vomit. He noticed that the deceased's catheter bag was full of urine. He found a 25g syringe on the bedroom floor: this was later examined and found to contain no traces of insulin. No syringe or other needles were ever found. Dr Roberts called an ambulance which arrived at 10.26 am. The crew noticed vomit on the sheets, and that when the deceased was moved he vomited again. On his way to Hinchingbroke Hospital, he vomited yet again. He arrived at the hospital at 11.36. His blood glucose level was found to be very low. A blood sample was taken at 11.40. He was then given 50ml of glucose, but his blood glucose concentration did rise. Further glucose was given during the day, but, although his glucose was restored to a normal level, he remained unconscious. A sample of his urine was taken some time later on 31 January: it was found to contain a high concentration of C-peptide (about 7500 pcms/l).
8. On 2 and 4 February, there occurred what have been referred to as the first and second IMED incidents. We shall return to these when we consider one of the grounds of appeal.
9. The deceased never regained consciousness and, as we have already said, he died on 9 February. Post mortems were conducted by Dr Harris on 11 February and by Dr Carey on 13 February.

The principal issues raised by the defence

10. These were:

(a) Did the deceased suffer an insulin-induced hypoglycaemic coma, or was there an alternative cause of the hypoglycaemia?

(b) Did the insulin immuno-assay tests sufficiently prove the existence of exogenous insulin in the serum?

(c) Did the evidence prove that it was the appellant who injected the deceased?

11. As we have said, it is now common ground that the deceased suffered from an episode of hypoglycaemia on the night of 30/31 January, and that this contributed to his death.

The Grounds of Appeal

12. On behalf of the appellant, Mr Steven Kay QC submits that the conviction is unsafe on the grounds that:

(1) The judge was wrong not to accede to the defence submission, made before the trial began, that there should be a stay on the grounds that to permit the trial to proceed would be an abuse of process.

(2) The judge was wrong not to accede to the defence submission that the evidence of the two IMED incidents should be excluded.

(3) The judge was wrong not to withdraw the case from the jury at the close of the prosecution case following a submission of no case to answer.

(4) The prosecution failed to produce sufficient evidence of the presence of insulin in the serum: the immuno-assay tests were not capable of providing conclusive proof of the existence of insulin in the serum.

(4A) More fundamentally, Mr Kay seeks leave to adduce fresh expert evidence which, he submits, casts doubt on the prosecution case that exogenous insulin caused the hypoglycaemia.

(5) The prosecution introduced misleading and inappropriate prejudice into the case, by suggesting as a possible motive that the appellant was having an affair and that this had aroused jealousy in the deceased.

(6) The judge failed to deal adequately with the defence in the summing up.

(7) The summing up was not a sufficiently cogent exposition of the salient features of the case.

We shall first deal with grounds (1)-(3) and (5)-(7). We shall then turn to grounds (4) and (4A).

Ground 1: abuse of process

13. The application to stay the proceedings was based on the fact that the prosecution had failed to examine or preserve the deceased's brain. The appellant was therefore denied the opportunity of examining it, and was denied important evidence which could have advanced her case. In fact, five samples of the brain were retained. In support of the application, Mr Kay placed before the judge witness statements by Professor Marks and Dr Geddes. Dr Geddes, a neuropathologist, said that the brain had not been adequately examined at post mortem, and that the few samples that were taken showed pathology attributable to poor oxygen supply to the brain; there were no changes in the samples that were typical of the effects of severe hypoglycaemia. Had the brain been preserved and properly examined, it was possible that hypoglycaemia could have been confirmed or excluded as a cause of death. Dr Anderson, a neuropathologist who made a report on behalf of the prosecution, said that it was impossible to confirm or refute that the deceased suffered an initial hypoglycaemic insult, not only because the examination of the brain was incomplete, but also because hypoxic-ischaemic changes will have compounded and masked any initial pure hypoglycaemic injury. Dr Anderson's report was before the judge.
14. The judge gave his ruling on 24 May. He observed that there had been no allegation of bad faith on the part of the prosecution. He said (page 5D of the transcript of the ruling) that the question that he had to decide was whether it was possible for there to be a fair trial. He considered whether the examination of the brain would have been of benefit to the appellant and noted that the experts differed as to this. These were arguments that would be available at the trial. He could see no unfairness in the circumstances of this case. He referred to what Mantell LJ said in R v Damian Paul Medway [2000] Crim LR 415 when giving the judgment of this court in relation to an application to stay for abuse of process on the grounds of loss of evidence. Mantell LJ said:

"We recognise that in cases where evidence has been tampered with, lost or destroyed it may well be that a defendant will be disadvantaged. It does not necessarily follow that in such a case the defendant cannot have a fair trial or that it would be unfair for him to be tried."

15. The judge then continued at page 8C:

"We would think that there would need to be something wholly exceptional about the circumstances of the case to justify a stay on the ground that evidence had been lost or destroyed. One such circumstance might be if the interference with the evidence was malicious."

16. Mr Kay concedes that the judge purported to apply the correct test, but submits that he applied it incorrectly. In particular, Mr Kay criticises the judge for holding that there would be no unfairness in allowing the trial to continue.
17. In our judgment, the judge's decision was plainly correct and is unassailable. We start by observing that the brain was destroyed on 12 February shortly after the completion of the post mortem. This was long before the possibility of criminal proceedings against the appellant was contemplated. Accordingly, this is not one of those cases where evidence is lost by some act of the police once a prosecution is underway. But more fundamentally, it has been emphasised by this court time and again that a stay on the ground that a fair trial is not possible by reason of the loss of evidence will only be granted in the most exceptional circumstances. The circumstances relied on in the present case are far from exceptional. If the judge was required to grant a stay in the present case, then applications for a stay on such grounds would succeed routinely in most, if not all, cases where evidence has been lost which might have been of assistance to a defendant. That is not the law. We reject the first ground of appeal.

Ground 2: The two IMED pump incidents

18. The deceased was connected in hospital to an Imed infusion pump which supplied him with a sedative drug called propofol. On 2 February, Nurse Davies found that the pump had been turned off just after the appellant had visited him in the intensive care unit. There was conflicting evidence as to whether it was possible to switch the machine off by accident. The appellant denied that she had switched it off. The second incident occurred on 4 February when Nurse Amphlett found air in the line from the pump of a similar machine. The appellant denied that she was responsible for this. The pump was not defective. There was evidence that, although the two incidents were not in fact life-threatening, nurses understood that air embolisms did create a risk for patients. According to one of the interviews which she had with the police, the appellant said that she thought that withdrawal of the drug could be harmful. It was uncertain how much air was found in the line, but it was insufficient to cause the alarm to sound. The cause of the air being in the line was unknown, although Mr Highmoor, an expert witness, said that it could not have happened by accident.
19. On 7 June, the judge was asked to rule on the admissibility of the evidence about the two IMED pump incidents. On behalf of the prosecution, it was submitted that the evidence was relevant and admissible. If the jury were satisfied that the appellant had been responsible for the two incidents, then they could find that this indicated that she was seeking either to hasten her husband's death, or to ensure that he could not recover consciousness, and explain what the appellant had done to him. On behalf of the appellant, it was submitted to the judge that, although the evidence had some probative value, this was outweighed by its prejudicial effect.
20. In allowing the evidence to be admitted, the judge said:

"But certain it seems to me that they could say, if they found that it must have been the accused who turned off the machine in the one case and made the air go into the machine in the other case, if they found that those were done deliberately by her, then it does seem to me that this would be a situation in which a jury would say, "We must now consider whether those were done deliberately by this woman to this man. If they were, what conclusion can we draw as to what happened earlier on 30^th or 31^st January?"

21. Mr Kay submits that the evidence of the appellant's responsibility for the two incidents was tenuous. This was borne out by the fact that in his summing up, the judge said (page 371) the evidence was such that it made one wonder whether there was anything in the point at all. Mr Kay also makes the point that, if the appellant had been charged with an offence arising from her alleged involvement in the two incidents, the case would have been with drawn from the jury at the ensuing trial. By admitting the evidence, the judge merely added to the baseless suspicions engendered by the prosecution.
22. The judge was called on to weigh the probative value of the evidence against its prejudicial effect. The decision whether to exclude the evidence under section 78 of PACE 1984 involved an exercise of discretion on his part. This court will only interfere with a judge's exercise of discretion when he or she has made a decision that is plainly wrong. In our view, far from being plainly wrong, the judge's conclusion was plainly right. It was correctly conceded by Mr Kay that the evidence was probative. It was only prejudicial if the jury considered that the appellant had been responsible for one or both of the incidents. But if that was their view, then the evidence had very considerable probative value. The question facing the judge was not whether the jury would be sure of her responsibility for the IMED incidents to the criminal standard of proof. He had to assess the probative value of the evidence and weigh it against its prejudicial effect. The less the probative value and the greater the prejudicial effect, the weaker the case for admitting the evidence. In our judgment, he was fully entitled to conclude that the scales came down in favour of admitting the evidence.

Ground 3: no case to answer

23. We should introduce this ground of appeal by saying something about insulin and immuno-assay tests. The results of the tests conducted by Dr Teale and Mr Hiscutt formed the central plank of the prosecution case. Proinsulin is a protein produced in the human pancreas. When it is released into the blood system following stimulation by the ingestion of food, it splits into equal parts of insulin and C-peptide. Insulin is a compound which controls the blood sugar levels, and keeps them within the normal range. C-peptide, which serves no known useful purpose, is removed from the blood system into the urine. The purpose of the tests is to discover the amount of each of these two components. The technique employed in immuno-assay tests is to raise in an animal an antibody to the compound or antigen that it is sought to find and measure in the human sample. The antibody will bind to the antigen. The amount of the antigen present can be measured by adding molecules of the same compound with an identifying radio-isotope attached to them.
24. There can be interference with the results of immuno-assay tests. As became clear during the evidence that we heard, there are three different mechanisms of interference. First, an endogenous antibody can mimic the antigen. Thus in the context of insulin, the molecule can behave like insulin, and be taken for insulin. Secondly, the antibody can bind the insulin receptors. Thirdly, the antibody, although not an insulin antibody, can bind the insulin. This type of antibody is sometimes referred to as a heterophil.
25. With this introduction in mind, we turn to summarise the prosecution evidence in relation to the tests. Dr Teale is the clinical scientist who received the serum sample from the deceased's blood at Guildford Laboratory. This is the reference laboratory for hypoglycaemia for the UK. Dr Teale said in relation to the results of his test (887 pcms/l (insulin), and less than 94 pcms/l (C-peptide)):

"But the fact that the C peptide is completely unmeasurable suggests to me that taken with the measurable insulin that the patient had been given insulin."

26. He acknowledged the possibility of interference by antibodies recognising "phantom" insulin antigens and thereby giving false readings. But he said that this sort of interference does not usually occur with a specific antibody in the system, and in any event that if this had been a problem in the present case, it would have been revealed by the dilutions:

"Q. Because you have applied the double dilution you cannot say that it specifically identified insulin? A. I cannot say that it specifically; nothing specifically identifies insulin 100 per cent, but the fact that the dilutions were so similar told me that it is likely to be insulin."

27. Mr Hiscutt gave evidence about the somewhat different test that he conducted which yielded a result of 616 pcms/l (insulin) and "undetectable" (C-peptide).
28. Professor Forrest is professor of forensic toxology at the University of Sheffield. He said that the double dilution technique enabled one to go from being (99%) sure to being absolutely (99.8%) sure that exogenous insulin was present. The presence of interfering antibodies would be revealed by the double dilution technique.
29. The next prosecution expert was Dr Ferner, a consultant physician with a particular interest in focal pharmacology. Apart from exceptional circumstances, the level of C-peptide in blood is likely to be at least as high as the level of insulin. Apart from the case of exogenous insulin, he did not know of any circumstances in which this would not occur. He acknowledged the existence of the insulin auto-immune syndrome ("IAS") where there could be higher levels of insulin than C-peptide as a result of natural disease. But the C-peptide would always be measurable.
30. At the close of the prosecution case, it was submitted to the judge on behalf of the appellant that since (a) there was no direct or indirect evidence to show that she had injected the deceased with insulin, and (b) the prosecution scientific evidence recognised the possibility of interference in assay tests, no jury could reasonably convict her on the prosecution evidence. The judge refused to withdraw the case from the jury. He said that the expert evidence was such that the jury could certainly accept that somebody had injected the deceased, and that it was open to them to infer that it must have been the appellant. He went through the possibilities of injection by an intruder, by accident, by suicide and by the deceased injecting himself for other purposes (eg body building), and concluded that it was open to the jury to reject each of these possibilities, and to decide that the injection had been administered by the appellant.
31. In our judgment, the judge reached the right conclusion. The jury were entitled to accept the prosecution expert evidence. If they did so, then they would find that the assay tests were consistent only with the insulin being exogenous. Once it was established to their satisfaction that the insulin was exogenous, it was open to them to decide that the appellant had injected the deceased with insulin. The possibility of injection by an intruder could be dismissed as fanciful. The judge was right to decide that the jury might properly reject accident, suicide or self-administration for body-building purposes, and might properly conclude that the only realistic explanation was that the appellant had injected the insulin. In our view, there plainly was a case for the appellant to answer.

Ground 5: misleading and inappropriate prejudice introduced by the prosecution.

32. In opening the case, Mr Coward QC told the jury that the appellant had no known motive to murder the deceased. As the case started, so it ended, because at the end of his summing up (page 370H), the judge reminded the jury of the prosecution case, and told them that it was accepted on behalf of the crown that it had been unable to show a motive. But Mr Kay complains of two incidents during the trial. The first occurred during the cross-examination of the appellant. She was asked what she thought was going on between the deceased and Nadine Jay. The appellant replied: "Well it may have been the beginning of an affair—but I knew nothing about it". Mr Kay then interjected that the evidence showed that it was Claire Fixter, not Nadine Jay who had invited the deceased to her house on the evening of 30 January. To which counsel for the prosecution said: "so it was Claire then?" Mr Kay complains that this was an unwarranted innuendo that the deceased was suspected of having an affair with another woman, and that this gave the appellant a motive to murder him. The judge later said to the jury that he doubted that they would rely on that evidence as the basis for finding that jealousy was the motive.
33. We agree with Mr Kay that Mr Coward should not have made these remarks. There was no evidential foundation for them, and they were inconsistent with the prosecution case which was, from start to end, that they could not point to a motive, but they did not have to do so. Regrettable though these remarks were, they do not affect the safety of the conviction.
34. The second complaint made by Mr Kay is that when the appellant gave evidence, she described how the family doctor, Dr Roberts, had attended the house. She referred to him as "David". Thereupon, Mr Coward said "Ah, David is it? I see. I see". The appellant explained that all the doctors in the Practice were known to her family by their first names. Mr Kay submits that this exchange illustrates the way in which the prosecution seized on any opportunity to insinuate unfounded motives. We have found this complaint mystifying. We fail to see how the exchange implied anything about the appellant's possible motive for murdering her husband. We think that the remark by counsel was inappropriate, but it is no more than that.

Ground 6: failure to deal with the defence adequately in the summing up.

35. The first criticism is that the judge failed to remind the jury of the appellant's evidence in a "clear, fair and appropriate" manner. She gave evidence for some 3 or 4 days. The judge did not summarise her evidence as a block in a discrete part of his summing up. Instead, he summarised her evidence as it touched the various events and issues in the case, when he dealt with those events and issues in his summing up. Mr Kay accepts that the judge covered all the salient parts of her evidence in this way. His complaint is that it was unfair to omit some of the detail. We confess that we are totally unable to understand this complaint. If the judge had dealt with the appellant's evidence in a single block, no doubt he would have been criticised for that, and probably with justification. Since it is accepted that the judge reminded the jury of all the salient parts of the appellant's evidence, it seems to us that there can be no substance in this complaint.
36. The second criticism concerns the way in which the judge dealt with the evidence of Professor Marks. The summary of his evidence occupies 79 pages of a transcript that occupies about 400 pages. 44 of these pages include substantial passages, quoted more or less verbatim, from what Mr Kay describes as a hostile cross-examination. The evidence of the other expert witnesses is not the subject of so much verbatim quotation in the summing up. For example, the judge referred to only one question that was put in cross-examination to Dr Teale. In the result, Mr Kay submits, the summing up was unbalanced and unfair. It gave undue prominence to the cross-examination of Professor Marks, when compared with the cross-examination of the prosecution experts, thereby encouraging the jury to form a more favourable view of them than of Professor Marks. We reject this complaint. It is not suggested that the judge distorted the evidence of Professor Marks, or that he omitted material parts of it from his summing up. It is true that the technique that he adopted to remind the jury of the evidence of Professor Marks did differ to a certain extent from that which he adopted in relation to the evidence of the other experts. But we do not consider that this was unfair, or that there is any reason to suppose that the jury drew any inferences adverse to the appellant from these differences.
37. The third criticism is that the judge failed to put the defence case that the deceased might have administered insulin to himself either in order to commit suicide or as a body-building exercise. These possibilities had not been withdrawn by defence counsel. But when we examined the passages in the summing up of which complaint is made, we found that there was no substance in this point. The first passage is at page 38E-39H. But when asked about this, Mr Kay accepted that no criticism could properly be made. The second passage is at page 89H-90F. Here too, Mr Kay disavows any criticism. It is clear that in these two passages the judge was expressly inviting the jury to consider the possibilities of self-administration whether by suicide or otherwise. During the course of argument, Mr Kay reduced his criticism of the judge to the passage at page 90G-91C of the summing-up, where the judge returned to the possibility that the deceased injected himself with insulin as a body-building aid. The judge said at the beginning of this passage: "Next Mr Hooper asks you to consider the taking of insulin as a bodybuilding aid". He continued:

"You know that it can be used in that way. You know that Dominic had weights and he used the weights often. He belonged to gym clubs, two of them, which he would attend during the week. So it could be said that here is somebody into bodybuilding. But no syringe was found. Of course, it could be said there again he would not have left a syringe lying around if he had injected himself with insulin. On the other hand, if he had injected himself with insulin for the purposes of bodybuilding, he would not have expected anything to go wrong. So that is obviously a factor. If he had wanted to get rid of the syringe it is not difficult to place a syringe (or anything else for that matter) in a place where a three year old child cannot obtain it. So you must consider, if you come to this stage, says Mr Hooper, that that is not likely."

38. In our judgment, it is clear that the judge did not withdraw this issue from the jury. He was entitled to remind the jury that the appellant's own counsel had said that this was an unlikely theory, as indeed it clearly was.
39. The final criticism concerns two remarks made by the judge which, it is said, were inappropriate to a fair trial. At page 50E, he said:

"It is in some ways a simple case, made more complicated inevitably by the need felt by the defence to test evidence here, evidence there and so on. But that is what it is in general terms."

And at page 373G he said:

"We have had an enormous amount of evidence. I repeat what I said before: Nobody can criticize the defence for querying everything. It makes it very difficult for the jury. It makes it very difficult for the judge but they are not appearing for us; they are appearing for the accused and they are entitled to do that and, what is more, if they did not do it, they would be failing in their duty so we have to put up with it."

40. We would agree that the judge should not have spoken of the "need felt" by the defence to test the evidence, or having to "put up" with the difficulty caused by the defence challenging everything. That was inappropriate language, said no doubt in an unguarded moment. But the effect of these remarks must be considered in the context of a 6 week long trial, and a summing up that lasted 5 days. We are in no doubt that these remarks did not render the trial unfair or the conviction unsafe.

Ground 7: summing up not a sufficiently cogent exposition of the salient features of the case.

41. Mr Kay does not submit that the summing up was incoherent or confusing: he accepts that the jury would have been able to follow it. His point is that there was an imbalance in the sense that, although he carefully analysed the evidence of the prosecution experts by reference to the technical issues in the case, when he came to deal with the evidence of Professor Marks, he merely read great tracts of the evidence verbatim. But, as Mr Kay acknowledged, this is not in truth a discrete ground of appeal; it is merely a repeat of the second criticism made under ground 6. We do not, therefore, need to say any more about it.

Ground 4: no evidence of insulin.

42. We have already summarised the relevant prosecution evidence (paragraphs 25-29 above). In our judgment, it was plainly open to the jury to conclude that they were sure that the test results were reliable. The possibility of interference had been explored in considerable detail during the trial. Moreover, for the reasons that we give later, we have ourselves reached the conclusion that the hypoglycaemia was caused by an excess of insulin. Accordingly, we reject ground 4.
43. A linked criticism made by Mr Kay concerns the way in which the judge in his summing up dealt with the evidence of Dr Teale as to the possibility of interference with the tests by heterophils, notwithstanding the fact that linear results obtained following double dilution. We have already quoted the relevant part of the evidence of Dr Teale (paragraph 26 above). The judge reminded the jury of what Dr Teale had said in these terms at page 160H:

"But in order to try and circumvent this possibility, I come back to the dilution system that we used, which, if this were a problem, would show up, in that the dilutions would not produce the same result. You would get a sudden change in values."

44. Mr Kay submits that the judge did not remind the jury that Dr Teale had accepted that the results following double dilution did not prove conclusively that insulin was present; merely that it was "likely" to be present. We would agree that the judge did not reflect the evidence of Dr Teale with complete accuracy. But we make the following points: (a) this was, at most, a modest inaccuracy, and one which (unsurprisingly) did not strike Mr Hooper as sufficiently significant to warrant his seeking a correction, although he had ample opportunity to do so; (b) it is fanciful to suppose that, if the jury had been reminded of the precise words used by Mr Hooper, they would have acquitted the appellant; and (c) in any event, as will become clear shortly, we have concluded, on the basis of a great deal of further expert evidence adduced before us, that the jury reached the correct decision on the medical and technical issues in the case.

Ground 4A: the fresh evidence

Should the fresh evidence be received?

45. We heard evidence from four expert witnesses on behalf of the appellant, and four witnesses called in rebuttal on behalf of the prosecution. The four prosecution experts had all given evidence at the trial, whereas all the defence witnesses were new to the case. We decided to hear the evidence before reaching a conclusion as to whether we should formally receive it pursuant to section 23 of the Criminal Appeal Act 1968. Mr Curtis submits that we should not receive the evidence since (a) most of it was available at the time of the trial, and (b) it is wrong in principle to allow a defendant to have a second bite at the cherry, in particular where this involves calling evidence to make good perceived shortcomings in the expert evidence that was adduced at trial.
46. Mr Curtis makes a powerful case for not receiving the fresh evidence. All the evidence given by the four experts who were called before us on behalf of the appellant was, in substance, available at the trial. It is true that a few of the epidemiological papers to which we have been referred postdate the trial, but most of them do not. It is true to say that there may now be a somewhat greater appreciation than there was at the time of the trial of the extent of the possibility of interference with immuno-assay tests, and the limitations of double-dilution as a means of reducing the risk of invalidation by interference. But the change since the trial is marginal.
47. Moreover, as we shall explain, the technical case advanced on behalf of the appellant in reliance on her four new witnesses is in at least two crucial respects, wholly at odds with the technical case advanced at trial on the basis of the evidence of Professor Marks. Mr Curtis has drawn our attention to the case of Steven Jones [1997] Cr App R 86 where Lord Bingham CJ, giving the judgment of this court, said at page 93D:

"Expert witnesses, although inevitably varying in standing and experience, are interchangeable in a way in which factual witnesses are not. It would clearly subvert the trial process if a defendant, convicted at trial, were to be generally free to mount on appeal an expert case which, if sound, could and should have been advanced before the jury".

We are mindful of this salutary passage. But we note that Lord Bingham was careful not "to circumscribe the operation of a statutory rule enacted to protect defendants against the risk of wrongful conviction" (page 93B).

48. This is a most unusual case. The evidence against the appellant was entirely circumstantial. The absence of any apparent motive and the other factual features strongly relied on by Mr Kay (to which we shall later refer) meant that the technical evidence assumed critical importance. Further, this evidence was not only unusually complex, but some of it touched areas of medicine which are not well understood. In these circumstances, we are in no doubt that it is necessary and expedient in the interests of justice to receive all the fresh material in evidence.

The evidence of Professor Marks

49. Professor Marks was not called to give evidence before us, and important parts of his evidence are no longer relied on by the appellant. Nevertheless, we think that it is important to have in mind what he said. We should say at the outset that Professor Marks is acknowledged to be a world-renowned expert in hypoglycaemia and immuno-assay tests. The following summary does not do justice to his evidence at trial, but we think that it contains its essential elements.
50. He said that his initial reaction on learning about this case and seeing the results of the tests was that it was a case of exogenous insulin. But when he saw the clinical notes, he changed his mind. The picture that they disclosed did not fit a diagnosis of insulin induced hypoglycaemia. He mentioned a number of features which indicated to him that the hypoglycaemia had not been caused by insulin. These included: (a) a person suffering from insulin-induced hypoglycaemia sweats or at least has a clammy skin, and (b) vomiting is inconsistent with insulin-induced hypoglycaemia
51. Professor Marks therefore looked for a different explanation. A possibility that he espoused at one stage of his evidence was that the hypoglycaemia was caused by an infection such as septicaemia, although he conceded that he was not expert in the field of septicaemia, and that that did not fit very well either. Nevertheless, he thought that septicaemia was the most likely cause (summing up page 346E). But in the end, he acknowledged that he did not know what had caused the hypoglycaemia although the deceased had "almost certainly died from natural causes" (page 320A). He thought that the hypoglycaemia was not induced by insulin, although he could not rule it out as a possibility (page 341A).
52. He gave a great deal of evidence about immuno-assay tests. He said that double dilutions could "rule things out. They don't rule them in" (pages 310F and 320F). There were simple tests that could have been carried out which would have shown whether there were interfering antibodies present, but none of these was carried out. He accepted that Dr Teale's dilution test would have revealed whether IAS was present. Since it did not reveal any interference by IAS, the possibility of IAS could be ruled out (pages 361-2). He said that he had never said that IAS was "even a starter". He went on to say (page 368B) that there had been a big upsurge of interest in the heterophil antibodies "and that has revealed how common this is as a cause of serious interference". He was then asked about the issue of interference in the context of the present case: "is it just a wild card to be considered or is it a real cause for grave concern in the context of this case?" His answer as recorded at page 368D of the summing up was: "I think it is a very grave concern. I believe that this is what has got everybody involved in a wild goose chase".

Fresh evidence called before this court

53. The following is no more than the barest outline of what the witnesses said.

The appellant's experts

54. The first witness called on behalf of the appellant was Dr Ismael. He is a consultant in clinical biochemistry and chemical endocrinology at Pinderfields and Leeds NHS Trusts. He has very considerable expertise and 35 years' experience in the field of immuno-assay tests. With reference to the facts of the present case, he said that he would have suspected analytical interference in the tests taking all the facts into account, and would have given an appropriate warning to his clinical colleagues. He agreed with Professor Marks that double dilution testing could rule interference in, but could not rule it out. The possibility of interference was now widely appreciated, and had to be taken into account by those interpreting the results of tests. It can be eliminated by conducting tests for the presence of anti-bodies, but no such test was carried out in the present case. The fact that similar results were obtained from the two tests carried out (by Dr Teale and Mr Hiscutt) does not rule out the possibility that the results were falsified by interference. Dr Ismael agreed with Professor Marks that the possibility of interference with the C-peptide results yielded by the two tests could be ignored. He disagreed with Professor Marks, however, as to the cause of the hypoglycaemia. He thought that it had been induced by insulin, and as a result of IAS rather than exogenous insulin. He gave three reasons for preferring IAS to exogenous insulin. These were: (a) no injection site was found; (b) the blood sugar concentrations were restored to normal by the administering of limited amounts of glucose to the deceased while he was in hospital on 31 January; and (c) the fact that a high concentration of C-peptide was found in the urine on that day.
55. Dr Wood is consultant clinical biochemist at the Regional Endocrine Unit, Southampton General Hospital. He too has great experience in the field of immuno-assay tests. He substantially agreed with Dr Ismael on the subject of the tests and the possibility of interference. In his report dated 3 October 2001, he discussed the possible cause of the raised insulin shown by the test results. He wrote: "although the most likely cause of raised insulin in the blood sample from the deceased is a raised "unbound" insulin (as argued by the prosecution) it is not the only possible cause". He went on to discuss IAS, and on the next page of the report, he wrote: "the concentrations reported in this case are compatible with and perhaps more consistent with the presence of insulin auto-antibodies than they are with insulin administration". The reason he gave in his oral evidence was that more typically the insulin levels in the case of exogenous insulin are much higher than 887 pcms/l.
56. Dr Clark is consultant clinical scientist and SAS director at University Hospital Birmingham NHS Trust. She too has considerable expertise and experience in the field of immuno-assay tests. Her evidence was confined to the reliability of the tests, and was substantially to the same effect on this topic as that of Dr Ismael and Dr Wood.
57. The last of the four witnesses was Professor Bagshawe. He is Emeritus Professor of Medical Oncology at London University. He supported the other defence witnesses in saying that immuno-assay tests were not conclusive, since, although they were generally reliable, they were not infallible. In particular, interference could not be ruled out. He said that the evidence in the present case was equally consistent with the cause of the hypoglycaemia being endogenous as exogenous insulin. He thought that there were possible causes of endogenous insulin which could not be ruled out. In particular, he mentioned the possibility of a small insulinoma in the pancreas, or other cancers which had escaped the notice of the two doctors who conducted the post mortems; and hyperplasia of the beta cells in the islets in the pancreas, something that would have been difficult to see at autopsy. But he acknowledged that, if any of these conditions had been responsible for the raised insulin levels, there would have been a corresponding rise in the C-peptide concentrations too. He sought to explain the low C-peptide results produced on the assay tests by saying that these results might have been the subject of interference. Other possible causes were canvassed by him. In particular, he referred to the fact that the autopsy report stated that the deceased had fatty change in the liver, and said that this indicated that there might have been a failure to degrade insulin leading to an accumulation of insulin in the blood, not matched by an increase in C-peptide. But he conceded that, if the hypoglycaemia in this case had been as a result of fatty change in the liver, he would have expected that condition to have become manifest at an earlier stage.
58. In his report dated 23 August 2001, Professor Bagshawe wrote that he "strongly supported" the view of Professor Marks that the features described in the clinical records of the deceased "differed substantially from the usual clinical features associated with insulin induced hypoglycaemia". As he made clear during his evidence in chief, he was referring in particular to the absence of sweating and clammy skin, and the fact that the deceased had vomited. In his later report, however, and as made clear by him in cross-examination, Professor Bagshawe did not feel able to maintain this opinion: he accepted that the clinical features were consistent with insulin induced hypoglycaemia.

The Prosecution experts

59. Dr Carey is a consultant forensic pathologist. He said that both he and Dr Harris (who had conducted the first post mortem) were concerned to look for a source of endogenous insulin. They therefore examined the pancreas very carefully. They could find no tumour on examination both with the naked eye, and with the microscope. Taking into account all the evidence, that is to say the laboratory tests, and all the clinical material, it was his opinion that the insulin had been administered exogenously. He relied particularly on the fact that the deceased had had no history of hypoglycaemia, and that once his glucose levels had been restored to normal on 31 January, they remained normal for the rest of his life. This showed that, as from 31 January, his pancreas was working normally. IAS or some insulin-producing tumour could explain the test results if the tests were looked at in isolation, but they had to be ruled out once the clinical features of the case were taken into account. He accepted that auto-immune disorders can occur as a bolt out of the blue, but the deceased did not have any of the conditions that cause or are associated with IAS, such as rheumatoid arthritis, Grave's disease, Lupus and others. All the findings, both the laboratory test results and the clinical evidence, were entirely explained by an injection of insulin some hours before the deceased was found unconscious.
60. Professor Forrest is a registered medical practitioner and a chartered chemist at the University Department of Forensic Pathology, Medico-Legal Centre, Sheffield. He said that he had no doubt that the prime cause of the deceased's death was exogenous insulin. He ruled out all other possible causes of hypoglycaemia, taking them one by one. It is unnecessary to refer to each of them. It is sufficient to mention insulinoma and IAS. He said that insulinoma tumours are relatively slow growing. It could be ruled out, not only because it would be accompanied by high C-peptide, but also because it is inconceivable that the first manifestation of a hypogycaemic episode attributable to an insulinoma would have been a major attack such as occurred in the present case. In any event, it is ruled out by the fact that the deceased achieved a normal euglycaemic state so easily on 31 January. The fact that his blood sugar levels returned to normal during the afternoon of 31 January indicates that he had been given an insulin overdose some hours before, and that the insulin had run its course.
61. Professor Forrest also ruled out IAS because it did not fit the clinical picture (ie no previous history, and restoration to normal pancreatic function in hospital), and none of the medical conditions or drug therapies associated in the literature with IAS was present in this case. All the case studies relied on by the appellant's experts were ones where there was reason to believe that the patient's immune system may have been disturbed, so as to generate auto-immune anti-bodies. The nearest to a case where a patient suffered spontaneous hypoglycaemia is Dr Ismael's case 5, an Italian case, which is the subject of the paper by Dozio and others (Vol 83 of Journal of Clinical Endocrinolgy and Metabolism, page 643). Extremely high levels of insulin were found with positive and elevated insulin anti-bodies. Professor Forrest observed that the patient was in traction for "arthrosis", which may be rheumatic arthritis. It is clear that the paper is by no means easy to interpret, but Professor Forrest concluded that the treatment described by the paper as having been given "implies some sort of immunological condition going on, even if they have not been able to characterise it exactly".
62. The finding of high levels of C-peptide in the urine sample was entirely consistent with exogenous insulin. As Professor Forrest explained, once the blood sugar concentration returned to normal, the pancreas would have started producing insulin and C-peptide again. Thus it was to be expected that the C-peptide would be secreted into the blood and via the kidney into the urine. If the urine sample had been taken some time after the blood sugar concentration was restored to normal, then that would explain the high C-peptide concentration in the urine.
63. The next prosecution witness was Dr Ferner. He is a consultant physician and clinical pharmacologist at the City Hospital, Birmingham. He substantially agreed with Professor Forrest on all issues. He agreed that the hypoglycaemia must have been induced by insulin, and that the insulin must have been exogenous. As regards IAS specifically, this could be ruled out because (a) it is very rare especially in white men; (b) none of the conditions or disorders that have been found to be associated with IAS was present in the deceased (although he accepted that the category of factors or conditions which are associated with IAS may not be closed); (c) no case has been reported where IAS has caused death; (d) C-peptide concentrations are almost always measurable, and sometimes high, in cases of IAS (although he accepted that in the Dutch case of Schlemper, the C-peptide concentrations were low, and might have been classified as "unmeasurable" by Dr Teale and Mr Hiscutt); and (e) the fact that the hypoglycaemia came "out of the blue" and the blood sugar concentrations were restored so quickly would have been "extraordinary" if this had been a case of IAS: where a hypoglycaemic episode occurs caused by IAS, there "is usually a programme of attacks for several weeks, months or years".
64. The final prosecution expert witness was Dr Teale. He described the tests carried out by himself and Mr Hiscutt. He explained that neutralising reagents had been added in both cases to reduce the possibility of heterophilic interference. Although no test was absolutely infallible, confidence in the reliability of the tests was further justified by the fact that (a) linear results were obtained following the application of the double dilution technique, and (b) the results of the two tests were broadly in line with each other. There was nothing to suggest that there had been interference with either test in the present case. He said that he had a particular interest in cancerous tumours which cause hypoglycaemia. His experience was that, where this occurs, the insulin and C-peptide secretion from the pancreas is switched off, so that the concentrations of insulin and C-peptide are very low. It followed that cancerous tumours could not have caused the hypoglycaemia in the present case. Dr Teale was sure that the cause of the hypoglycaemia was exogenous insulin.

Our findings on the medical evidence

Was the hypoglycaemia caused by insulin?

65. It is common ground that the deceased suffered from a major hypoglycaemic attack during the night of 30/31 January, and that this contributed to his death. We are entirely satisfied that this attack was induced by insulin. As we have said, the evidence of Professor Marks at the trial was that he did not know what caused the hypoglycaemia, but the cause was not insulin. He listed a number of features described in the clinical records which, to quote from Professor Bagshawe's report, "differed substantially from the usual clinical features associated with insulin induced hypoglycaemia". But as we have said, when he was cross-examined, Professor Bagshawe did not maintain his reliance on these features. The possibility that the hypoglycaemia was not induced by insulin was explored in evidence before us. A list of the known other causes was considered by Professor Forrest and Professor Bagshawe. It is not necessary for us to rehearse the list. They were all excluded by Professor Forrest on the grounds that there was no evidence to support any of them. Professor Bagshawe eventually agreed that the cause of the hypoglycaemia was "the inappropriate production of insulin" (Tr 4/128). On the evidence before us, it is indisputable that the hypoglycaemia was induced by insulin. Once this is recognised, the issue of interference in the assays recedes into the background so far as concerns the insulin figures found on the two forms of test.

If the insulin was endogenous, was it caused by IAS?

66. It is convenient next to consider whether, if the insulin was endogenous, it was the result of IAS or some other disorder. Professor Bagshawe considered that the following possible causes for the insulin (apart from IAS) existed: a small insulinoma or other cancer in the pancreas; hypoplasia of the beta cells in the islets; or fatty change in the liver. In our view, the first two can safely be excluded. First, Doctor Harris and Doctor Carey both examined the pancreas very carefully, and found no evidence of either. Secondly, we accept the evidence of Professor Forrest that it is inconceivable that the first manifestation of an attack of hypoglycaemia attributable to an insulinoma would have been in an attack of the severity that occurred in the present case. Thirdly, Professor Bagshawe accepted that, if either of these disorders had caused the raised insulin in the present case, there would also have been an accompanying equivalent rise in the concentration of C-peptide. He had to contend, therefore, that the C-peptide results produced by the assay tests must have been distorted by interference. The trouble with this theory is that it was not supported by Dr Ismael or indeed by Professor Marks at the trial. They were both of the opinion that, although the insulin results were questionable, the possibility of interference with the C-peptide results could be ignored.
67. As regards Professor Bagshawe's reference to the fatty change in the liver, we have already pointed out that he conceded that, if this had been the cause, he would have expected to see some prior manifestation of it, and there was none.
68. We conclude, therefore, that, if the insulin was endogenous, then (IAS apart), it cannot have been caused by any of the disorders suggested by Professor Bagshawe.

Was it IAS or exogenous insulin?

69. We now turn to the fundamental question of whether the insulin was exogenous or endogenous (ie IAS). As we have seen, much of the evidence (both before us and at trial) was directed to the question of whether the assay tests were reliable, and, in particular, whether the inappropriately high recorded levels of insulin concentration might have been inaccurate on the grounds that there had been interference. Since we have concluded that the cause of the hypoglycaemia was elevated concentrations of insulin, the accuracy of the tests is irrelevant. We have already dealt with Professor Bagshawe's suggestion that the tests may have yielded inaccurate C-peptide results.
70. All four prosecution experts were firmly of the opinion that the insulin which caused the hypoglycaemia was exogenous. Professor Bagshawe conceded that the evidence was consistent with exogenous insulin. Dr Ismael and (possibly) Dr Wood said that the more plausible explanation was that the insulin was endogenous, although they did not rule out the possibility that it was exogenous. We shall start with Dr Ismael. As we have seen (paragraph 54 above), he gave three principal reasons for favouring endogenous insulin. The first was the failure to discover an injection site. This does not seem to have been regarded as a factor of significance by any of the other experts in the case. One of the problems is that, if the insulin was injected, it is uncertain by what means this was done. The various possibilities were discussed in some detail by Dr Ferner. He explained that injections can be given intravenously, or subcutaneously. If done subcutaneously, they can be done by a number of small injections, possibly in the buttocks. Subcutaneous injections usually leave a small lump under the skin, but this might not be noticeable if the injection is deep into the buttock, and in a large person might not be noticeable at all.
71. The second reason given by Dr Ismael was the fact that the blood sugar concentrations were relatively easily restored to normal on 31 January. We need to set out the facts in a little more detail than we have already done. The deceased arrived at the hospital at 11.40am. Very soon thereafter, he was given two 50 ml shots of 50% glucose. At 12.45 pm, his blood sugar concentration was normal. But by 1.45 pm, it had fallen to less than 1 (severely hypoglycaemic). He was then given a further 50 ml shot of 50% glucose. When the test for blood sugar concentration was repeated at 3.05 pm, the result was within the normal range. And so it remained during all subsequent tests. It is also relevant to note that on the second or third day after the deceased was admitted to the ITU in the hospital, nasogastric feeding began and was continued thereafter. Professor Forrest said that, until towards the end of his time in the ITU, he was receiving reasonably substantial amounts of nasogastric feed". He said (Tr 4/51):

"100 millilitres per hour is not an insubstantial amount of calories, glucose, carbohydrate etc. So he was being fed. He was not being starved in ITU. And he would have had to have a normal response to carbohydrate loads in order to have been able to tolerate that".

72. Dr Ismael's reasoning was as follows. In cases of IAS, the bulk of the circulating serum insulin is bound to insulin antibodies, and is therefore physiologically inactive: only a relatively small amount (ie the unbound or free insulin) is physiologically active. If the 887 pcms/l of insulin revealed by Dr Teale's test were exogenous, it would all have been free and physiologically active, and would have been too much to permit the restoration of a euglycaemic state so readily early in the afternoon of 31 January. We do not find this reasoning at all convincing. It does not account for the fact that once the blood sugar concentrations returned to normal, they remained normal even though the deceased was receiving substantial nutrition. In other words, the pancreas started to work perfectly normally again, producing appropriate amounts of insulin in response to changes in blood sugar levels and nutrition. This is entirely consistent with the previous excess of insulin being attributable to an injection, and is difficult to explain if the previous excess was attributable to endogenous insulin. The literature shows a recurrent pattern of hypoglycaemia occurring some 3 – 7 hours after a meal, when caused by IAS.
73. The third reason given by Dr Ismael as to why the evidence was more consistent with IAS than with exogenous insulin relates to the high C-peptide concentrations found in the urine sample. He said that IAS patients have a very high blood concentration of C-peptide after a meal, and he would expect such patients to show higher concentrations of C-peptide in the urine than non-IAS patients. It was for this reason that he thought that the high C-peptide result was more consistent with IAS than with exogenous insulin. On the other hand, the prosecution experts explained the high concentration of C-peptide in the urine as attributable to the resumption of normal working of the pancreas following the administration of glucose to the deceased in hospital. In our opinion, it is impossible to base any conclusions on the level of C-peptide in the urine, because the evidence does not disclose when the urine sample was taken.
74. Accordingly, we reject Dr Ismael's reasons for saying that IAS fits the facts better than exogenous insulin. We did not find the evidence of Dr Wood very satisfactory on this issue. It seems to us that the two passages in his report to which we have already referred (paragraph 55 above) are to some extent inconsistent with each other. In his oral evidence, Dr Wood sought to resolve the inconsistency by distancing himself from what he had written. He said that the presence of unbound insulin was "a likely cause" or "a cause" of the raised insulin, and not as he had written "the most likely cause" of it. We do feel able to place any weight on this part of Dr Wood's evidence.
75. It does not, however, follow from our rejection of Dr Ismael's reasons for saying that the facts are more consistent with endogenous insulin that the insulin must have been exogenous. That would be impermissible reasoning. We must now, therefore, consider whether the insulin must indeed have been exogenous. We are entirely persuaded by the evidence of the prosecution experts that the insulin must have been exogenous and was not the result of IAS. We summarise our reasons as follows. First, the deceased had no previous attack of hypoglycaemia. Secondly, he had had no prior symptoms of IAS: this was a "bolt out of the blue". There was no evidence of any of the auto-immune disorders that are almost invariably associated with IAS. Examples of these are rheumatoid arthritis, Lupus disease, and Graves' disease. Nor was there any evidence of any serious viral or bacterial infection which is often associated with IAS. Every reported case of IAS, with the possible exception of the Dozio case to which we have already referred (paragraph 61) above, shows a history of some manifestation of the disorder by one means or another prior to an episode of hypoglycaemia. There is no reported case of an unheralded attack of IAS, still less of one so severe as to lead to death. It is true that the deceased had had influenza about a fortnight earlier, and that he complained of feeling tired on the night of 30 January, but we are satisfied that these were of no significance.
76. Thirdly, there is no reported case of an attack of hypoglycaemia attributable to IAS having been fatal. It usually responds to treatment and/or a carefully monitored diet. It would have been quite extraordinary if such a severe attack could have been caused by a previously unheralded IAS.
77. Fourthly, the fact that the blood sugar concentrations were restored to normal on 31 January, and were maintained at normal levels thereafter despite significant infusions of nutrition points very strongly against this having been a case of IAS. It is clear that the pancreas was working normally after the blood sugar levels were raised on 31 January. That is very difficult to explain if the attack of hypoglycaemia was caused by IAS. Why should a previously unsuspected IAS have caused a massive attack of hypoglycaemia on the night of 30/31 January, and then within hours have ceased to have any potency, and thereafter permit the pancreas to work as normally as, apparently, it had been working throughout the deceased's life? We think that there is no answer to this question.
78. Fifthly, IAS is an extremely rare condition in any event. In about 30 years, there have been 240 registered cases in Japan, and about 30 published cases of Caucasians, of whom the majority have been women. There has been no published case in the United Kingdom. There is a greater genetic disposition to IAS among Japanese than among Caucasians.
79. It seems to us that the combined effect of these five points irresistibly proves that the hypoglycaemic attack from which the deceased suffered on the night of 30/31 January was not caused by IAS, but was caused by exogenous insulin.
80. We should add that we derive further support for our conclusion from the fact that Professor Marks, a world expert on hypoglycaemia rejected IAS as not even being a starter in this case. It is true that his evidence on other points, in particular in relation to the cause of the hypoglycaemia, has not been endorsed by other experts and has not been accepted by us. Nevertheless, we believe that we are entitled to rely on his evidence in relation to IAS as supportive of the conclusion that we have reached.

Is the conviction safe?

81. We must now consider whether our conclusions on the medical evidence leave any room for doubting the safety of the conviction. As Mr Kay points out, the prosecution case against the appellant was based entirely on circumstantial evidence. He submits that there are a number of features of the case which make it impossible to be sure of the appellant's guilt.
82. First, there was no known motive for the appellant to murder her husband. We have already summarised the evidential position as to motive (such as it was) when dealing with Ground 5 above. The reality is that, as the judge reminded the jury, the prosecution could not identify any motive.
83. Secondly, although as a nurse, the appellant had access to insulin and had the skill to inject it effectively, she had very little opportunity to administer the injection on the night of 30/31 January. Although the prosecution did not formally concede that the injection could not have been before the appellant left the house for the party at about 8.45 pm, she could not realistically have injected the deceased before she left for the party. Professor Bagshawe said that the insulin was very likely to have been the soluble, rather than the slow acting variety, and that a large dose would not have had an effect for longer than 12 hours at most. Since the blood serum sample was taken at 11.40am the following morning, the injection (if that is what it was) must have been made between about midnight, and the time when the deceased was found unconscious. It is true that Dr Ferner said that it was not possible to be sure what variety of insulin was injected. But we consider it to be so unlikely that the injection was given before 8.45 pm that this possibility can safely be ruled out. Not only is that hypothesis inconsistent with the evidence of Professor Bagshawe, but it is also inherently improbable in the light of certain other facts. In particular, we have in mind the fact that the deceased made a number of telephone calls in the evening of 30 January, including one to his mother at 9.06pm, and another to his brother Patrick at 9.20 pm. He did not complain of being unwell, although he told his brother that he was feeling extremely tired. If he had felt unwell later in the evening, he could have telephoned for help, but he did not.
84. The alternative hypothesis advanced by the prosecution was that the injection had been given by the appellant some time between about 3.30 and 6.30am on 31 January. Mr Newall said that she and the appellant retired to bed at about 3am after they had returned from the party at the hospital. On this alternative hypothesis, the appellant must have left the Newalls' house some time after 3am, driven to her home in Stonely (a journey of about 25 minutes), given the injection, and driven back to the hospital in time for work at 7am. Mr Kay submits that there are a number of aspects of the case which compel the conclusion that it is most unlikely that this could have happened. As he points out, it was a very risky strategy for the appellant to adopt to effect the murder of her husband. She could easily have been heard creeping out of the house and/or driving off in her car which was parked just outside the house (the Newalls' bedroom looked out on to the street). She might have disturbed the Newalls' dogs. She might have been seen or heard arriving at her home in the dead of night, but nobody was called to say that they had seen her. Further, as was explained by Professor Bagshawe, she could not have injected the deceased intravenously in the dark: she would have had to put the light on, and obtain his co-operation. (Subcutaneous injection could, however, have been given in the dark, without his co-operation and while he was asleep to a part of his body where he, as a paraplegic, had no sensation). If the deceased had been awake and aware of what was going on, he would have asked her about the drugs that she was giving: in other words, his knowledge and understanding about drugs would have been a further impediment to the appellant.
85. All of these are cogent points, although it is right to point out that, on her story, the appellant was able to leave the Newalls' house at about 6.30am and drive to the hospital without disturbing the dogs and without being heard by Mr or Mrs Newall. But more fundamentally, it is important to bear in mind that none of these points is new. They were all canvassed in great detail at the trial. The lack of known motive and the inherent unlikelihood of the appellant returning to the matrimonial home at dead of night to murder her husband were amongst the many points that were strongly deployed on behalf of the appellant before the jury. Moreover, the jury had the benefit (denied to us) of seeing the appellant give evidence for about 3 days. It is clear that they did not believe her account. It has not been argued by Mr Kay that this was a perverse verdict. The jury must have been sure that the hypoglycaemia was caused by exogenous insulin. Having considered all the points made on behalf of the appellant about lack of known motive and lack of ready opportunity, and having heard her give evidence, the jury were sure that she had injected the deceased with insulin. For the reasons that we have given, we are sure that the hypoglycaemia was caused by exogenous insulin. Having reached this conclusion we remind ourselves of what Lord Bingham of Cornhill said in R v Pendleton [2002] 1 WLR.72,83E:

"I am not persuaded that the House laid down any incorrect principle in Stafford, so long as the Court of Appeal bears very clearly in mind that the question for its consideration is whether the conviction is safe and not whether the accused is guilty. But the test advocated by counsel in Stafford and by Mr Mansfield in this appeal does have a dual virtue to which the speeches I have quoted perhaps gave somewhat inadequate recognition. First, it reminds the Court of Appeal that it is not and should never become the primary decision-maker. Secondly, it reminds the Court of Appeal that it has an imperfect and incomplete understanding of the full processes which led the jury to convict. The Court of Appeal can make its assessment of the fresh evidence it has heard, but save in a clear case it is at a disadvantage in seeking to relate that evidence to the rest of the evidence which the jury heard. For these reasons it will usually be wise for the Court of Appeal, in a case of any difficulty, to test their own provisional view by asking whether the evidence, if given at the trial, might reasonably have affected the decision of the trial jury to convict. If it might, the conviction must be thought to be unsafe."

86. It seems to us that, applying this approach, we are driven to the conclusion that this conviction is safe. It follows from our view that the cause of the hypoglycaemia was exogenous insulin that, if the fresh evidence had been heard by the jury (whether in addition to, or substitution for, the evidence that they did hear), they would still have convicted her. Having concluded that the insulin was exogenous, the jury had to consider who administered it. Accident or deliberate injection by an intruder are not pursued by Mr Kay as realistic possibilities. There remained two other candidates. Either the deceased injected himself in order to commit suicide or for some other reason, or the injection was by the appellant. The possibilities of suicide or self-administration for body-building purposes were considered by the jury. It is not surprising that they were both rejected. The appellant herself said that she did not think that he could have committed suicide. He was apparently happy, loved the child, and generally had a great deal to live for. Mr Hooper (who represented the appellant at trial) accepted before the jury that self-administration for body-building was not likely. One of the problems with this theory was that no insulin syringe was ever found. It follows that it is not at all surprising that the jury concluded that they could be sure that the exogenous insulin was administered by the appellant. Once the jury had decided that they could be sure that the insulin was exogenous, they were entitled on the evidence to conclude that it had been administered by the appellant. It follows that, since we too have decided that we are sure that the insulin was exogenous, there is no basis for concluding that the verdict is unsafe. Accordingly, this appeal must be dismissed.