Gay & Anor, R v

On 13 January 2005 at Worcester Crown Court, following a trial before Pitchers J and a jury, Angela and Ian Gay were convicted of the manslaughter of Christian Blewitt, a young boy who had been placed with them with a view to adoption (count 2 on the indictment). They were each sentenced to 5 years’ imprisonment. They were acquitted of the murder of Christian (count 1). No verdict was entered in respect of an alternative count of cruelty to a person under 16 (count 3).

They subsequently sought leave to appeal against conviction on three grounds. Ground 1 relates to the issue of joint enterprise; ground 2 is based on fresh evidence from a medical expert as to the cause of Christian’s death; and ground 3 relates to the judge’s directions to the jury concerning the burden of proof. The single judge refused leave on grounds 1 and 3, which come before us by way of renewed applications. The single judge referred the application for leave on ground 2 to the full court, which is how that ground now comes before us.

For reasons explained below, we have decided to refuse the renewed applications for leave on grounds 1 and 3 but to grant leave on ground 2 and to deal with the substantive appeals on that ground. We will therefore refer to Mr and Mrs Gay as appellants.

The position in relation to ground 2 was complicated by the existence of further applications on behalf of the appellants to be allowed to instruct three additional medical experts and for the adjournment of the leave application pending receipt of reports from those additional experts. We decided not to rule on those applications as a threshold issue but to hear the existing expert evidence de bene esse, together with full argument from counsel, before reaching any concluded view on the matter. In the event, because of our decision to grant leave and the conclusion we have reached on the substantive appeals, as set out below, the applications in respect of additional experts fall away.

Christian Blewitt was born on 8 June 1999. During the first ten months of his life he was admitted to hospital on four occasions. Such was the concern about the quality of the parenting he was receiving that he was removed from his mother and placed with foster parents in June 2000. His two younger siblings were removed from hospital at birth and also placed with the foster parents. An attempt to rehabilitate Christian with his mother proved a failure and he was returned to his foster parents. In summer 2002 the foster parents were informed that the children were to be put up for adoption.

The appellants had married in 1990. Mr Gay was an electrical engineer and Mrs Gay became an actuary. They were unable to have children and eventually considered adopting a child. Mrs Gay contacted social services in January 2002. In June 2002 an adoption social worker, Mrs Gillian Jones, assessed them with a view to placing Christian and his two siblings. She thought the appellants suitable but arranged for them to spend time with an experienced foster carer, Mrs Capper, so as to gain some experience of caring for young children. There was some dispute at trial about what happened at the sessions with Mrs Capper. It suffices to note that only Mr Gay attended the sessions, though both applicants had been expected to attend, and that Mrs Capper left a message for the social worker expressing concerns about the appellants’ ability to cope. Mrs Jones spoke to Mr Gay about Mrs Capper’s concerns but came to the conclusion that there was no real cause for concern. She obtained permission for the appellants to visit the nursery attended by her own child in order for them to gain experience of small children. Mr Gay attended on five or six occasions and was praised by nursery staff.

As matters progressed, the appellants met Ms Melrose Gray, the social worker responsible for Christian. They also looked after another infant for three and a half weeks, and moved to a new address where they adapted the house to make it suitable for any children they might adopt. Mrs Jones raised the question of Mrs Gay taking time off work following adoption of Christian because of his past history of hospitalisation and neglect. Mr Gay assured her that he and his wife would be together. They indicated in documentation that Mrs Gay planned to take adoption leave for a number of months.

The appellants were approved by the adoption panel on 9 September and were approved specifically to adopt Christian and his siblings on 20 September. They had an initial meeting with the children on 22 October at the foster carers’ home. On 29 October the first unaccompanied visit took place at the appellants’ house.

The children moved in with the appellants on 1 November. The appellants looked after the children together and arrangements were made for social services to visit on a regular basis. On 2 November Mr Gay telephoned Mrs Jones to the effect that the children had settled in well. During a visit by Mrs Jones and Ms Gray on 6 November, Christian became upset but Mrs Gay was able to reassure them and it was noted that Christian was calling her “mummy”.

Two days later, on 8 November, Mr Gay telephoned Mrs Jones to say that they were very worried. Mr Gay said in evidence that Christian’s behaviour on 6 and 8 November had been odd and he would not communicate despite encouragement. He also that Christian’s sleeping behaviour during this period was odd. Mrs Gay said that Christian had panicked during the visit of 6 November but calmed down after being comforted, but that on 7 and 8 November his manner had become very withdrawn. When Mrs Jones went to see the appellants following the telephone call, they told her that they were not sure that Christian was the child they had in mind and that he presented difficulties: they had expected him to be able to talk better and to recognise colours and letters.

Mrs Jones spoke firmly to the appellants and made clear that there was no prospect of the children being split up. She left the appellants with a video. Mr Gay telephoned the following day, 9 November, to say they had viewed the video and felt they had not given Christian the chance he deserved. Mr Gay also said in evidence that on 9 November Christian behaved like a different person and communicated perfectly normally.

Mrs Jones made a further visit on 11 November. Mrs Gay had in fact just returned to work, but Mrs Jones was not informed of it. She learned of it only on a further visit on 27 November, when Mr Gay told her and, according to her, said that he was cross about it. His own evidence was that he might have given the impression that he was furious about his wife’s return to work, but it was not true: he had been trying to appease Mrs Jones by appearing to agree with her. He said that they had wanted to give the impression to social services that they would both remain at home as long as was required. Mrs Gay denied deliberately misleading social services about her return to work. She said that she worked short days and there were no problems in the first week: the children accepted that she was not there and the atmosphere was relaxed.

Mrs Jones reported back to social services that Mrs Gay had returned to work. They formed the view that this was not acceptable, as it went against everything that had been told to the adoption panel. When Mrs Jones next visited the appellants, on 5 December, she told Mr Gay that she would have to discuss his wife’s return to work because it was contrary to the agreement. The events leading to Christian’s death supervened before that discussion could take place.

In the course of his police interviews, Mr Gay said that they had told Mrs Jones at the meeting on 5 December that they found Christian’s behaviour to be odd. He believed that Christian had a problem with attention seeking when Mrs Gay returned home in the evening. Christian refused to do anything for her and became introverted. This behaviour continued until bed-time and was very upsetting for her.

Mr Gay said that there had been a number of incidents on Saturday 7 December involving Christian and Mrs Gay. Christian struck her on the head with his chin, causing her to leave the room in distress. She did not feel that he had done it on purpose but was upset because she had wanted to cuddle him and felt rejected. Mr Gay did not see the second incident, when Christian bit his wife’s hand on the web of the thumb, but was told about it by his wife. He asked Christian to apologise, but Christian would not do so and kept turning his head away. Apart from this there was nothing wrong with Christian’s behaviour, so he felt able to go to his boat.

Mrs Gay’s account was that Christian had been withdrawn on 7 December. She asked him for a hug. He came slowly and when she picked him up, he jerked his body so that his head came into contact with her mouth. On another occasion she tried to hug him, but he objected and his knee or foot came into contact with her ribs. There had been an earlier occasion when the boys had got over-excited and Christian had nipped her hand. She thought she had said something about it to her husband and may have told him not to tell Mrs Jones in case Mrs Jones wanted to bring the placement to an end. Mrs Gay agreed that she was a bit upset about the bite. Apart from that there was nothing wrong about Christian’s behaviour; otherwise her husband would not have gone to the boat. While her husband was out, Christian stayed with her in the kitchen. He seemed a bit grizzly, so she put him upstairs with his brother. When Mr Gay returned, they spent the afternoon with the children. Christian played with his food, which was unusual, and he only ate a fraction of it.

Mr and Mrs Gay gave virtually identical accounts of events on Sunday 8 December, though the account given by Mrs Gay was in some respects more detailed. She said that she got up at 8.30 or 9.00 am, showered and spent two or three minutes with the boys. Christian seemed a bit quiet. He did not drink his tea and kept spilling it, but insisted that he did not require help with it. He did not finish his drink. Nor did he eat much of his cereal for breakfast, which was unusual. At about 11.30 am he started to ask for drinks. He came to her begging for water, which was unusual, and followed her to the sink. He guzzled water from a tumbler she filled for him, to the extent that he spilled some of it. He asked for another drink and she filled the tumbler a second time. She could not understand why he wanted water and she gave him some Ribena from a jug in the fridge. After this he asked again for water and she gave him some more. She crouched down to support him while he drank. He had four drinks in total. He asked for a fifth but she said no because it would make him poorly. He went to the family room. He then wet himself, so she took him upstairs to the bathroom to clean him, while Mr Gay cleaned up downstairs and brought some clean clothing for Christian. Christian asked for a shower so that he could wash his hair. Normally he had a phobia about the shower.

They had lunch between 1.00 and 1.30 pm. Christian was suddenly sick: it was mostly liquid. He was put to bed. Later Mr Gay brought him down to the family room and they reheated his food. He did not eat his meal, but put his hands in the gravy and smeared his face. They left him alone in the kitchen to see if he would eat the meal if he were on his own. After five or ten minutes they looked in and it appeared that he had thrown his meal on the floor. Nothing like this had happened before. Mr Gay described Christian as looking up and smiling and laughing at him, which he attributed to Christian being cheeky. Mrs Gay said when she saw Christian, he was in his chair with his head slightly down. Mr Gay stood Christian on the floor and wiped him, then suggested he be removed from the situation and picked him up and took him upstairs. They put him in his baby sister’s cot. Mrs Gay asked Christian why he had done it. He tried to speak but words would not come out. He was in a sitting position when Mr and Mrs Gay returned downstairs.

At about 3.15 pm Mrs Gay went upstairs to check on Christian. She found him lying on his back diagonally, with his head almost in the corner of the cot. His eyes were half open as if rolled up. She picked him up. He was totally floppy. She shouted to her husband, and when he came she asked him to get some milk. Nothing happened when she put the milk to Christian’s lips. Mr Gay shone a torch into Christian’s eyes but got little reaction. He told his wife they had to get Christian to hospital. He drove Christian to Russells Hall Hospital, arriving within 20 minutes.

From the time of his admission to hospital Christian did not recover consciousness. The nature of his condition, the tests carried out on him and the efforts to treat him are considered below. He was transferred to Birmingham Children’s Hospital during the evening of 8 December. Over the next three days his condition deteriorated, and at 3.28 pm on 12 December he was declared dead.

The appellants had been arrested on 11 December, the day before Christian died. They were interviewed at length and were subsequently charged with his murder.

The focus of the evidence at trial was on medical issues. Multiple areas of medical expertise were covered and a large number of experts gave evidence. There were, however, three areas of particular importance:

There was evidence of trauma to Christian’s head. The post mortem revealed 11 sub-scalp bruises which appeared to be recent. There were areas of subdural haemorrhaging and bruising, and the brain was grossly swollen. There were also retinal haemorrhages.

Christian had severe hypernatraemia, i.e. an exceptionally high concentration of sodium in his blood. That is the area central to the present appeal.

There was an area of infarction (in effect, death of the muscle) in the left ventricle of the heart, which had probably occurred while Christian was in hospital. It was the kind of injury that might be caused by a heart attack in an adult, but was most unusual in a young child. There was also some older infarction of a capillary muscle elsewhere in the heart.

The issues revolved around the first and second of those areas, the causation of them, and their relationship to the third.

The charge of murder (count 1) was based on the trauma to the head, the essential questions being whether it had been a substantial cause of death, whether it had been caused by an unlawful act or unlawful acts of the appellants (shaking or banging) and, if so, whether such act or acts had been carried out with the intention requisite for murder.

The charge of manslaughter (count 2) was put on alternative bases. The first was that the death was caused by trauma to the head resulting from the unlawful act or acts of the appellants, but carried out without the intention requisite for murder. The second was that the death was caused by the high sodium concentration in Christian’s blood, which in turn had been caused by the unlawful administration of salt by the appellants: the evidence was that Christian would have had to ingest 30-40 grams of salt, or at least 4½ teaspoonfuls, to account for such a high level of sodium. Although the charge was left to the jury on that dual basis, the reality by the end of the trial was that the manslaughter charge depended on the salt issue. As the judge expressed it in summing up: “it is the head injury that founds murder and salt manslaughter” (104B).

In relation to both counts (count 1 and count 2) the prosecution case was put on the basis of joint enterprise. The prosecution could not say which of the appellants was the principal and which was the secondary party in relation to any of the unlawful acts alleged. At the close of the prosecution case, a submission of no case to answer was made on the ground, inter alia, that there was insufficient evidence to enable a jury to conclude that the appellants were parties to a joint enterprise. The judge rejected that submission. As part of the first ground of appeal, it is submitted that he was wrong to do so.

In his summing up the judge reminded the jury at some length of the medical evidence, but concluded with a helpful summary, whilst stressing that it was “inevitably an over-simplification”. What he said in relation to salt is merely the start of what we will need to consider. What he said in relation to the head trauma, however, provides a sufficient overview of the position at trial. The judge said this (145-147):

As regards the salt issue, the question left open by the judge in that passage was “how did he get it into his system?” The appellants denied that they had given salt to Christian. The prosecution invited the jury to infer from all the circumstances that Christian could not have taken such a large amount of salt accidentally and that it must have been deliberately administered to him. The suggestion was that it was done as a punishment. Commonsense and experience, together with the experience of the experts and the absence of any such case in the literature, were said to support the view that a child would not voluntarily take such a large amount of salt. The prosecution pointed to the evidence that the appellants were not coping well that week, and particularly that weekend. They also relied on the appellants’ own statements that Christian himself did not have access to such a large amount of salt within the house.

It is evident from the jury’s verdicts that they did not accept the prosecution’s case on count 2 in respect of trauma to the head but did accept the case on count 2 in respect of hypernatraemia caused by unlawful administration of salt (or “salt poisoning”, as it has been described in some of the evidence and submissions before us).

With that introduction we turn to consider the grounds of appeal.

Mr Mansfield QC, for the appellants, indicated that grounds 1 and 3 were not freestanding grounds in the sense of providing a basis for an appeal in the absence of the fresh evidence under ground 2. It seems to us, however, that they do fall to be considered separately and that the most convenient course is to examine them before moving on to ground 2. As we have already mentioned, the single judge felt able to dispose of them while at the same time referring ground 2 to the full court.

Ground 1 addresses the issue of joint enterprise. It is submitted that, even if salt was administered unlawfully to Christian, there was no evidence to show that the applicants were parties to a joint enterprise to administer it. Although neither of them left the house during the relevant period, there was no evidence that they remained together with Christian throughout that period; on the contrary, there was clear evidence that there were times when they were apart. This should have led to the judge accepting the submission of no case to answer at the close of the prosecution case. Further, the argument is said to be even stronger in the light of the evidence called by the defence and the jury’s verdicts on count 1. It is submitted that the argument that there was no evidence capable of establishing joint enterprise is stronger in this case than in any previous reported case involving two parents jointly accused of the murder or manslaughter of a child.

In his ruling rejecting the submission of no case to answer, the judge referred to relevant authorities from which he extracted a number of principles. Since it is not contended that he misdirected himself, we do not set out those principles. The judge said, correctly, that the whole of the picture must be examined. He then identified a number of features of the evidence, on the basis of which he concluded that the evidence as a whole was sufficient to enable a jury properly to conclude that there was a joint enterprise:

In our judgment those matters provided an entirely adequate basis for the judge’s rejection of the submission of no case to answer on this issue. Nor do we accept that anything material was added on this issue by the defence evidence or by the jury’s verdicts on count 1, so as to affect the safety of the convictions on count 2 so far as this aspect of the case is concerned.

The judge did observe in his ruling that a proper direction would be needed on the issue of joint enterprise. No complaint is made of the direction he gave in his summing up, save as regards one specific point. That point concerns the prosecution’s reliance on the lie about placing Christian in the cot rather than on his bed (see the sixth of the points mentioned in the judge’s reasons quoted above). It was asserted by the prosecution that the reason for the lie was what had occurred in the cot, namely, as was alleged, an assault with intent to cause grievous bodily harm. The appellants submit that that related only to count 1 and cannot logically support an assertion of joint enterprise in relation to count 2. It is said that the judge erred in failing to give the jury a direction to that effect. It is also said that the judge should have directed the jury in terms of R v Strudwick and Merry (1999) 99 Cr App R 326 that lies could not make good any deficiencies in the evidence of joint enterprise.

As to the specific point concerning the lie about the placing of Christian in the cot, we are satisfied that the matter was adequately dealt with by the judge when he told the jury (158C):

In so far as the appellants seek to make any more general point, we are satisfied from our consideration of the summing up as a whole that there was no risk of the jury relying on lies otherwise than as additional support for the prosecution case on count 2, including the case as to joint enterprise.

Those are the reasons for our decision to refuse the renewed application for leave to appeal on ground 1.

It is not in dispute that the judge gave the jury clear and correct general directions on the burden of proof. The appellants contend, however, that there are two passages in his summing up in which he inadvertently misstated the position or gave a misleading impression.

In spelling out how the jury should approach disputes between the expert witnesses, the judge said this:

Much later, in the summary of the medical evidence which we have already quoted at paragraph 27 above, the judge said this in relation to the evidence of the defence expert, Dr Chambers, on the salt issue:

Mr Mansfield submitted that the italicised passages were erroneous. It was not enough for the jury to be sure that Dr Chambers was wrong. They had to be sure that the prosecution experts were right, which did not follow automatically from a finding that Dr Chambers was wrong. The difference was of real importance in this case, where there was an issue as to whether the prosecution had excluded all possible causes of the excessive sodium concentration other than the ingestion of salt – an issue the importance of which is underlined by the fresh evidence under ground 2.

In our judgment there is no substance to those submissions. When looked at as a whole, the judge’s directions on the burden of proof can have left the jury in no doubt whatsoever that they had to be sure of the prosecution case before they could convict. There was no disguised shifting of the burden to the defence.

As a narrower point, the submissions would fail even if attention were confined to what the judge said about the jury having to be sure that Dr Chambers was wrong. It was Dr Chambers’s evidence that, although he could not identify an alternative mechanism, “he could not be sure that Christian was given an excess of salt from outside his body” (see the judge’s summing up at 128B, which fairly reflects the transcript of the evidence). It follows that in order to be sure that Dr Chambers was wrong the jury had to be sure that the prosecution were right on the issue of ingestion of salt.

Those are the reasons for our decision to refuse the renewed application for leave to appeal on ground 3.

That clears the way for the true focus of this appeal: ground 2 and the linked application for the court to receive fresh evidence. The issues raised are of considerable scientific complexity, but we think it unnecessary and unhelpful to go into the level of detail canvassed in the expert reports placed before the court. We propose to concentrate on the broader features of the debate.

We have already referred in general terms to the fact that, on his admission to hospital on 8 December 2002, Christian was suffering from hypernatraemia, an abnormally high concentration of sodium in his blood. His blood (or plasma) sodium on admission to the Russells Hall Hospital was at the dangerously high level of 184 millimoles per litre (“mmol/l”). A normal level is in the region of 140 mmol/l. He was given an intravenous saline solution (which, on the face of it, may seem odd, but is explained by the need to bring down the level of sodium gradually in order to reduce the risk of damage to the cells). By the time of his transfer to the Birmingham Children’s Hospital in the evening of 8 December the level was down to 173 mmol/l. Although the regime adopted at that hospital was expected to continue to bring the level down, it remained at or above 170 mmol/l for the next 24 hours, i.e. throughout 9 December. This was the “plateauing effect” to which reference is made in the expert evidence considered below. On 10-11 December the level did fall further but did not get below 157 mmol/l, still much higher than normal.

The issue at trial was what had caused this high sodium concentration and in particular whether the prosecution had proved its case that it was caused by the ingestion of salt. When introducing the medical evidence in his summing up, the judge explained the position as follows (104-105):

Among the prosecution experts, of particular significance for present purposes was Professor Haycock, a professor of paediatrics and honorary consultant paediatrician and paediatric nephrologist. It was his opinion that Christian’s hypernatraemia was due to ingestion of salt, and it was he who calculated (or recalculated, making a substantial upward revision of his initial calculations) that it would have needed 30-40 grams of salt taken orally to raise Christian’s plasma sodium concentration to the level it was on admission to hospital.

Professor Haycock also said that there was every reason to suppose that the kidneys were normal. In those circumstances, and having regard to the regime applied in hospital, Christian’s body would have been expected to excrete the excess sodium, and the sodium level would have been expected to fall back to a normal level. Yet the sodium readings showed that the level did not fall as expected (in particular, there occurred the plateauing effect mentioned above), and a further set of readings for the “fractional excretion of sodium” showed likewise that the excess sodium was not being excreted. Professor Haycock thought that the most likely explanation for this was severe cardiac dysfunction. It was in this way that he sought to explain the known details of Christian’s condition.

The importance of Professor Haycock’s evidence is indicated by the fact that he is the one prosecution expert singled out in the judge’s overview of the medical evidence on the issue of salt which we have quoted in paragraph 27 above.

The defence expert on this issue was Dr Chambers, a consultant paediatrician and renal physician, to whose evidence the judge directed repeated attention in the various passages that we have quoted. Dr Chambers was unhappy about Professor Haycock’s explanation that the failure of the body to continue to excrete the excess sodium was due to cardiac dysfunction. He accepted that “it could conceivably have been an explanation” and said that cardiac physiology was not his field, yet he did not think that the heart dysfunction could have brought in such compensatory mechanisms as to override the impulse of the body to excrete a sodium load, and he referred in this connection to the fact that Christian’s circulation was not badly compromised. He could not explain the plateauing effect and why the excess sodium did not continue to be excreted. Whilst unable to think of other causes of the hypernatraemia, he was not sure it was due to the ingestion of salt. He said that medicine cannot always provide all the answers.

What the appellants seek to do by the fresh evidence now tendered is, in effect, to provide a possible alternative explanation for the hypernatraemia. The evidence comes from Dr Walters, a retired specialist consultant in chemical pathology with particular experience of fluid and electrolyte disturbances. There are two linked aspects to his evidence. First, he rejects the view that the hypernatraemia was caused by ingestion of salt. Secondly, he puts forward an alternative medical hypothesis which was not considered at the trial and which he says is capable of explaining the pattern of events without any unlawful conduct on the part of the appellants.

Mr Mansfield’s case, put simply, is that if the jury had been faced with the competing expert views which are now before this court on the issue of hypernatraemia, it is possible that they would have reached different verdicts on count 2, just as the competing expert views on the issue of trauma to the head may have led to the acquittals on count 1.

In order to decide whether to receive the evidence of Dr Walters under section 23 of the Criminal Appeal Act 1968, we heard that evidence de bene esse. In rebuttal, the Crown called Professor Haycock, whose evidence we likewise heard de bene esse. Both witnesses had submitted detailed written reports (in exchanges that continued up to the time of the hearing before us), but they also dealt with the main features of their evidence in oral testimony. We turn to consider their evidence.

Dr Walters said he was sure that Christian’s hypernatraemia was not the result of salt poisoning. There were a number of reasons for that view.

First, there was the failure to excrete the load of salt allegedly taken. Provided the regulatory systems are working normally, the body rejects excess sodium. Christian would have had to ingest an enormous amount of salt for his plasma sodium concentration to be so high. That should have led to a high rate of excretion of sodium. There was a high rate initially, but it tailed off and was back at a more normal rate after about 12 hours. It should have continued at the higher rate for at least 48 hours after admission.

At trial, Professor Haycock had suggested heart failure as a possible explanation for the body’s retention of excess sodium. Dr Walters rejected that view. He accepted that there was damage to Christian’s heart and that this was an extremely unusual finding in a young child, but he did not consider that it could account for the retention of the excess sodium. He explained that a feature of severe heart failure (but not, he said, of mild heart failure) is the release of a hormone called aldosterone which promotes the reabsorption of sodium in the kidneys and thus the retention of sodium within the body. But he did not consider that this mechanism could have occurred in Christian’s case:

An echocardiogram carried out by Dr Dhillon at about 3.00 pm on 9 December showed that the left ventricular function was mildly impaired, but in Dr Walters’s view the changes overall were not such as would give rise to severe heart failure.

This was borne out by Dr Dhillon’s findings on clinical examination. In heart failure there is a falling cardiac output which stimulates the sympathetic nervous system and, in turn, causes widespread constriction of small blood vessels, with the result that the skin becomes pale and cold; and when the skin is pressed and the blood is squeezed out of the skin capillaries, the colour returns very slowly. But Dr Dhillon described Christian as being pink and warm, with good capillary return. Most important of all, Dr Dhillon recorded the arterial pulses as “full”, which is an indication of a good cardiac output. Dr Dhillon brought all his clinical observations together with the description of Christian’s circulation as “hyperdynamic”, which means there was increased blood flow, not decreased blood flow.

Thus, Christian had severe heart disease, but not heart failure, and his condition was not such as could stimulate the enormous output of aldosterone that would be necessary to override the body’s desire to get rid of excess sodium.

Further, the measurement of aldosterone itself was close to the middle of the normal range, rather than at the high level one would expect in a case of heart failure.

Dr Walters’s reliance on the aldosterone measurement went further than the point that it would have been expected to be high if there had been heart failure. His view was that if there had been a massive load of ingested salt still in the body, then the level of aldosterone would have been expected to be low: he referred to cases where aldosterone had been shown to be suppressed to very low levels indeed in the case of patients undergoing experiments in the course of which they were given massive salt loads. The fact that the level was normal (at a time when, in Dr Walters’s view, the impact of the infusion of saline solution would have effectively worn off) was one of the matters that he took into account in rejecting the explanation of salt poisoning and arriving at an alternative explanation.

Another principal reason why Dr Walters rejected salt poisoning was the failure to bring the sodium level down to the normal range. It is usually easy to bring the sodium down to normal in such a case by giving the patient enough water. Two mechanisms then operate. First, the excretion of sodium in the urine is enhanced: there is a limit to the extent to which the kidneys can concentrate the sodium in the urine, and to get the maximum rate of excretion one has to ensure that enough urine is formed to carry the sodium out in its maximum concentration. Secondly, water is also retained through the action of what is called the anti-diuretic hormone (“ADH”): the effect is to dilute the concentration of sodium in the extracellular fluid, with the result that water moves back into the cells themselves. The fluid balance charts show that towards the end Christian was being given only very dilute solutions of salt, and he passed large volumes of urine, but without a correspondingly high concentration of sodium. At some stage during his admission he retained a lot of water, with some sodium, but still the sodium concentration did not fall below about 160 mmol/l.

This leads to the alternative hypothesis put forward by Dr Walters. The fact that the sodium concentration did not come down below about 160 mmol/l suggested to him a fault in the body’s mechanism for maintaining the concentration of sodium in the body. The sodium concentration is regulated by a part of the brain called the hypothalamus, which achieves this by changing not the amount of sodium but the amount of water in the body. Specialist cells in the hypothalamus, known as osmoreceptors or collectively as the osmostat, trigger the release of ADH, which in turn, as mentioned above, causes water to be retained in the body through its action on the kidney tubules. Thus, if the plasma sodium concentration rises, ADH is released so that a small volume of concentrated urine is formed, the amount of water retained in the body is increased and the sodium concentration is thereby brought down. If the sodium concentration falls, the release of ADH is inhibited so that a large volume of dilute urine is passed, the amount of water retained in the body is reduced and the sodium concentration is thereby brought up.

Another group of osmoreceptors in the hypothalamus regulates the thirst mechanism. There is some dispute as to whether this is a separate osmostat or part of the same osmostat, but nothing turns on that. We will refer for convenience to a single osmostat. A high sodium concentration results in the sensation of thirst, which leads to drinking. In concert with the increased action of ADH on the kidney tubules, this enables lost water to be replaced.

The normal sodium concentration, as explained above, is in the region of 140 mmol/l, and the secretion of ADH to regulate the body water content normally takes place by reference to that level. This is sometimes referred to as the osmostat setting.

There are cases, however, in which the osmostat setting is known to be disturbed. Most of the reported cases involve physical damage to the hypothalamus, e.g. as a result of surgery. There is no evidence that Christian fell into that category. There are other, less common, cases in which no identifiable cause has been found: the description “idiopathic” is given to such cases. Most of those other cases involve the downward resetting of the osmostat, so that the patient has a persistently low sodium concentration. The upward resetting of the osmostat, resulting in a high sodium concentration, is much rarer. There are, however, two documented cases of it: a 1985 paper by Gill, Baylis and Burn, A case of ‘essential’ hypernatraemia due to resetting of the osmostat, and a 1987 paper by Thompson, Freeman, Record and Baylis, Hypernatraemia due to a reset omsostat for vasopressin release and thirst, complicated by nephrogenic diabetes insipidus. Both concerned adults, but in Dr Walters’s opinion there is no reason in principle why the same event could not happen in a child, and he thought it only a matter of time before resetting was described in a child. He stressed that the relevant mechanisms are still not known.

Dr Walters referred to another disturbance of the hypothalamus which is more common than resetting of the osmostat, namely reduced sensitivity of the osmostat, such that the release of ADH or increase in thirst occurs much more slowly than with a normal osmostat. He also said that in theory there could be an upward resetting of the osmostat together with a sluggish response to concentrations of sodium above that reset level. Whilst his hypothesis in relation to Christian was presented principally in terms of resetting of the osmostat, he indicated in the course of his evidence that it might not be a resetting but simply a reduced sensitivity: it was something happening to the hypothalamus that resulted in gross disturbance of the normal mechanisms that regulate secretion of ADH and thirst.

Thirst is important for the other aspect of Dr Walters’s hypothesis. The hypothesis involves not just an upwards resetting of the osmostat relating to the plasma sodium concentration (perhaps to a level of 160-165 mmol/l instead of the normal 140 mmol/l, though he did not postulate any specific reset value), but superimposed water depletion due to defective thirst. Thirst is usually set slightly above the ADH setting, but in Dr Walters’s view there is no basis for assuming that the two will always be reset in parallel. In any event, if the osmostat relating to thirst is reset upwards, thirst will not be stimulated until the sodium concentration reaches a higher level than would normally stimulate it. In the absence of thirst, water will not be drunk and the body will suffer from water depletion. Such water depletion, which does not involve the loss of sodium, is to be distinguished from dehydration, in which both water and sodium are lost. Dr Walters accepted that there were no clinical signs of dehydration in Christian.

Dr Walters said that the appellants’ description of Christian’s attitude to drinking made him think that Christian had an abnormality in his thirst mechanism. It may be that Christian did not know what thirst was until the late stages when his plasma sodium concentration went abnormally high and he displayed frantic thirst on the morning of 8 December.

Dr Walters made clear that he could not prove his hypothesis about the resetting of the osmostat or similar dysfunction of the hypothalamus, because of the limited data available in respect of Christian’s condition (though he was not criticising the hospital staff for not taking the relevant measurements at the time). Equally, however, he maintained that the hypothesis provided an alternative explanation which could not be excluded.

The cross-examination of Dr Walters by Mr Davis QC for the Crown brought out various weaknesses or difficulties in Dr Walters’s approach. They included the following:

The two reported cases of upward resetting of the osmostat not only concerned adults but were factually very different. Neither involved a sudden onset of coma and death resulting from the elevated level of sodium in the blood. In response to this, Dr Waters observed that there may have been other cases where people have died but the true diagnosis may have been missed because it was not thought of at the time.

Routine tests carried out on Christian in 1999 and 2000 showed normal sodium levels (135 to 140 mmol/l). Dr Walters’s hypothesis therefore required something to have happened thereafter to the hypothalamus to cause the osmostat to reset to a higher level or some other disturbance to occur to the normal mechanisms that regulate thirst and the release of ADH. Nonetheless Dr Walters considered that a reset could occur at any time (the trigger being unknown) and that a change of this kind would not have affected Christian’s ability to live a normal life.

It is still difficult, however, to explain the sudden onset of acute illness, with the sodium concentration at the dangerously high level of 184 mmol/l on Christian’s admission to hospital. Again, however, Dr Walters considered that this was possible if Christian had been running with a plasma sodium level of, say, 160 or 165 mmol/l and then for some reason had become water depleted. If that happened there would be a surge in the plasma sodium and it could reach 184 mmol/l without clinical signs of dehydration.

A similar point applies to the evidence of Christian’s extreme thirst on the morning of 8 December. It suggests a surge in the plasma sodium level (a “quantum leap”, as it was put in cross-examination). But Dr Walters observed that we do not actually know what Christian’s plasma sodium was doing at that time; and also that, although Christian was clamouring for drinks in the morning, his thirst appeared to have been alleviated in the afternoon, before he became unconscious.

It was put to Dr Walters that Christian’s pattern of thirst prior to 8 December did not support the suggestion of a reset osmostat or water depletion, in that he would regularly take a drink. As to that, it is true that there is evidence of Christian taking drinks, but it is also relevant to note that on 27 November Mr Gay had told Mrs Jones that he was worried that neither of the boys asked for drinks and that they would have to be reminded to drink, but then Christian’s brother (not, it would seem, Christian) would drink excessively. There is also some evidence of Christian spurning a drink of milk on 5 December; and at breakfast on 8 December he did not finish his tea.

If there was severe water depletion, as Dr Walters postulated, there would have been a contraction in the volume of Christian’s extracellular fluid. Professor Haycock’s view was that there was no reason to suppose that such a contraction occurred and, on the contrary, there were some signs tending to indicate an expansion in the extracellular fluid. Professor Haycock referred in this connection to readings for blood urea and creatinine, haemoglobin and plasma albumin. These various points were put to Dr Walters, who did not agree with Professor Haycock with regard to the interpretation or relevance of the figures.

Various points were also raised with Dr Walters concerning his view that heart failure could not account for the body’s failure to excrete the excess sodium. They included the existence of evidence that Christian’s blood pressure was consistently low. They also included the evidence of Dr McLellan at trial that it was difficult to say what was preventing the secretion of sodium because so many things were happening: he was receiving powerful drugs, he was being ventilated, his heart was failing to pump progressively with time, his brain function was deteriorating, and he was dying. Dr Walters did not accept that any of these matters explained the failure to excrete the sodium on 9 December.

We can deal more briefly with the evidence given by Professor Haycock, who said that overall he still held the same view as he did at trial, that the high sodium level found on 8 December was due to the ingestion of salt.

As to the issue raised by Dr Walters concerning osmoreceptor dysfunction, Professor Haycock said he was familiar with the condition and had seen and treated forms of it. The two examples of upward resetting of the osmostat which Dr Walters had cited from the literature were the only two examples of which he was aware, but it was not possible to refute Dr Walters’s statement that there might be other cases that had not been investigated properly.

He dealt with various of the matters that had been put to Dr Walters in the course of cross-examination. He explained why in his view the data tended to indicate an expansion of Christian’s extracellular fluid, rather than the contraction that would be expected if there had been water depletion. He could not see an event or illness in Christian’s history that might have led to a change in the osmostat setting. In all the cases of osmoreceptor dysfunction that he had seen himself or had studied in the literature (i.e. not limited to the two reported cases of upward resetting of the osmostat), the patients had shown little or no thirst; and in none had there been anything like the frantic thirst displayed by Christian on 8 December. Nor did any of them involve the kind of catastrophic collapse seen in Christian’s case. Christian’s thirst on 8 December suggested a rapid rise in sodium concentration over a very short space of time. Ingestion of salt would do it. So would lack of access to water, but it would take a few hours at least to get to the point of severe thirst.

As to Professor Haycock’s view that heart failure might account for the fact that the excess sodium was not excreted, he made clear that he did not agree with Dr Walters’s interpretation of the aldosterone measurements. In his view, there were too many factors at work to enable any proper interpretation of the results.

In cross-examination he accepted that if heart failure did cause the ordinary mechanisms for the excretion of excess sodium to be overridden, it would be unique in his experience and he was unaware of any case in the literature of a similar situation. He had put it forward at trial as a hypothesis. If the hypothesis was wrong, he accepted that it would leave a number of anomalies to be explained.

He also accepted that hypernatraemia can occur in the absence of dehydration (in the sense of loss of water and sodium) and in the absence of salt poisoning, i.e. it may arise where there is simple water depletion. It may be caused by or associated with a disturbance of the hypothalamus, which may take different forms and the mechanism of which is in many cases not clearly understood.

There then occurred the following important exchange:

Professor Haycock was asked a number of questions about why he had not mentioned the other possible cause of hypernatraemia, namely water depletion, in his witness statement for the trial. His answer, in essence, was that he did not think it was applicable in this case, because Christian had experienced thirst, and for a child not to take in sufficient water he would have to have had a thirst deficiency.

Although a number of other matters were canvassed with Professor Haycock in cross-examination, including criticisms of certain of his analyses and calculations, we think it unnecessary to set them out.

Section 23 of the Criminal Appeal Act 1968 provides:

For the Crown, Mr Davis took the point that the first strand in Dr Walters’s evidence, namely his rejection of salt poisoning as a cause of the hypernatraemia, should not be admitted as fresh evidence since it is simply a re-run of an issue ventilated fully at trial. The matters referred to by Dr Walters were raised by Dr Chambers or in the cross-examination of Professor Haycock. The fact that Dr Chambers’s view did not prevail at trial cannot provide a proper basis for allowing evidence to the same effect from another expert on appeal.

Mr Davis drew attention to what was said by Judge LJ giving the judgment of the court in R v Kai-Whitewind [2005] EWCA Crim 1092, at para 97 (this is in a passage omitted from the version of the judgment reported at [2005] 2 Cr App R 457):

Those are important observations and we have borne them very much in mind, but we do not think that they bite on the particular circumstances of this case. It seems to us that Dr Walters’s rejection of salt poisoning as the cause of the hypernatraemia cannot sensibly be separated from the second strand in his evidence, namely the alternative hypothesis he puts forward as a possible explanation of the events that occurred. The one leads into the other and it would be wholly artificial to look at the second strand in isolation. Moreover, even in relation to the first strand Dr Walters goes beyond the defence position at trial: whereas Dr Chambers was not sure that the cause was salt poisoning, Dr Walters is sure that it was not. Accordingly, although the first strand does involve a substantial repetition of matters raised at trial, that cannot be viewed as an insuperable objection to it in the very special situation with which we are faced.

We therefore take Dr Walters’s evidence as a whole when considering the various criteria in section 23(2) of the 1968 Act.

As to section 23(2)(a), we are satisfied that the evidence of Dr Walters is capable of belief. He is a reputable expert and his hypothesis derives some support from the scientific literature, even if the condition on which he relies is extremely rare and has not hitherto been observed in a child. Professor Haycock, as we have said, accepted that what Dr Walters had done was a legitimate exercise in postulating an alternative hypothesis and that he was not in a position to say that Dr Walters was wrong. It is plain that Dr Walters’s hypothesis faces substantial difficulties, as the cross-examination brought out, but it cannot in our view be dismissed as untenable or simply incredible.

That takes us to the second and crucially important criterion, in section 23(2)(b): whether it appears to the court that the evidence may afford a ground for allowing the appeal.

Medically this was on any view, as Mr Mansfield submitted, an exceptional case. Not only was there a remarkable combination of features, including trauma to the head as well as hypernatraemia, but the damage to Christian’s heart was in itself a highly unusual feature in a child of this age. Moreover the prosecution case on the cause of the hypernatraemia relied not simply on the existence of that damage to the heart, but on the hypothesis that there was heart failure sufficient to override the body’s normal mechanisms for the excretion of excess sodium. If that hypothesis is correct, then Professor Haycock has accepted before us that the situation would be unique in his experience and that he is not aware of any similar case in the literature. If the hypothesis is not correct, it leaves a number of anomalies that the prosecution’s case does not explain.

In so far as Dr Walters challenges Professor Haycock’s hypothesis that the failure to excrete the excess sodium was accounted for by heart failure, his evidence does add materially to that given at trial but, if taken alone, it would not in our view be sufficient to provide a ground of appeal.

As we have said, however, that strand in Dr Walters’s evidence cannot be taken alone. Dr Walters not only rejects salt poisoning as the explanation for the hypernatraemia, but has moved on from that to a different hypothesis which provides a possible alternative explanation for the hypernatraemia. Although that alternative hypothesis faces substantial difficulties, the same is true of Professor Haycock’s hypothesis as to heart failure; and the fact is that Dr Walters’s hypothesis adds an important additional dimension to the case as considered by the jury at the trial.

At the trial Professor Haycock put forward heart failure as the most plausible explanation he could think of for the failure to excrete the excess sodium, and he effectively threw down the gauntlet to the defence, saying it was for someone else to come up with another explanation. Dr Chambers, although not sure of salt poisoning, was unable to say that Professor Haycock was wrong and was unable to think of an alternative explanation. All other known causes appeared to have been ruled out. In those circumstances it is perhaps unsurprising that the jury accepted the explanation given by Professor Haycock. There is a marked contrast in this respect between the position on count 2 and the position on count 1, where the jury were faced with a major conflict of expert opinion which may have contributed to their decision to acquit. Had the jury been faced on count 2 with Dr Walters’s evidence positively disagreeing with Professor Haycock and putting forward an alternative explanation for the hypernatraemia, then in our judgment there is a real possibility that their conclusion on that count would have been different.

We have not lost sight of the fact that the medical evidence, although important, formed only part of the evidence in the case and must be assessed in the context of the case as a whole. The appellants, previously childless and having recently taken on the immense burden of looking after three children with a view to adoption, were in a position of considerable stress. The particular difficulties they had experienced with Christian had added to that stress. There was evidence that they had sought to minimise the stress they were under. The evidence of sub-scalp bruising was on one view consistent with gripping of Christian’s head in a manner that might be associated with unlawful administration of salt (though this does not seem to have been a significant feature of the prosecution case on count 2 at the trial). A small cut on Christian’s mouth was also consistent with such action (though this, too, does not seem to have been a significant feature of the case at trial). The prosecution were able to gain some support for their case from the appellants’ lies or concealment of the truth, in particular about Christian biting Mrs Gay’s hand. All of those matters serve to underline why we put it no higher than a possibility that the jury might have reached different verdicts on count 2 if the evidence of Dr Walters had been placed before them. Such matters do not, however, serve to remove that possibility.

It follows that the criterion in section 23(2)(b) weighs heavily in favour of receipt of the fresh evidence.

As to section 23(2)(c), there is no issue concerning the admissibility of Dr Walters’s evidence.

That leaves us with section 23(2)(d), the question of reasonable explanation for the failure to adduce the evidence at trial. For the Crown, Mr Davis very fairly and realistically acknowledged that the lack of a reasonable explanation should not stand in the appellants’ way if, having heard the evidence de bene esse, the court were of the view that it threw doubt on the safety of the conviction. Nonetheless he did make the point that the hypothesis put forward by Dr Walters is based not on new science but on something that has been known about for many years and he submitted that there was no reasonable excuse for the failure to adduce it at trial.

There was some oblique reference in the evidence at trial to the condition on which Dr Walters’s hypothesis is based. Mr Punt, a consultant neurosurgeon, in explaining that it was improbable that the hypernatraemia was the consequence of head injury, referred to a very small number of cases within his experience in which hypernatraemia had followed severe surgical damage in the region of the hypothalamus. Dr Chambers picked this up in his evidence, stating (23G-H):

Dr Chambers went on to say, however, that this was simply an educated guess, speculation, and he did not develop the point. It was evidently not something about which he had any detailed knowledge. Professor Haycock, for his part, accepts that he knew that hypernatraemia could be caused by disturbance of the hypothalamus, but he did not mention it because he did not think it was applicable on the facts of this case. In the result, neither he nor anyone else said anything that ought reasonably to have alerted the attention of the defence to the point.

The fact is, therefore, that the science was there, but its potential significance was not recognised by those involved in the trial. This was not through any want of diligence by the appellants or their legal team. It was only subsequently that Dr Walters was approached, by a different legal team, and expressed the opinions that he has now given in evidence to us. In all the circumstances we take the view that a reasonable explanation does exist in this case for the failure to adduce the evidence at trial.

Having regard to the conclusions we have reached on the application of the various criteria under section 23(2), we are in no doubt that it is “necessary or expedient in the interests of justice” to receive the evidence of Dr Walters under section 23(1).

Our decision that the evidence should be received leads inevitably to the grant of leave to appeal on ground 2. We therefore go on to deal with the substantive appeals against conviction, treating the evidence that we heard de bene esse as evidence duly received under section 23. The submissions addressed to us by counsel covered this contingency. In reality we need add very little to what we have already said.

The question is whether, in all the circumstances, the evidence of Dr Walters causes us to doubt the safety of the appellants’ convictions on count 2. In considering that question we have had regard to the guidance in R v Pendleton [2002] 1 WLR 72. We have to make our own assessment, but can properly test it by reference to whether the evidence, if given at trial, might reasonably have affected the decision of the jury. We have already explained the reasons why we take the view that the jury might have reached a different conclusion if Dr Walters’s evidence had been placed before them at the trial. Having regard to the various factors that we have discussed in that context, and after anxious consideration of the case as a whole, we have reached the conclusion that the convictions on count 2 are unsafe.

We should, however, stress that if Dr Walters’s evidence had been available at trial, that would not in our view have made it inappropriate for the prosecution to proceed on count 2 or have meant that the judge had to withdraw the case on that count from the jury. There is a passage in the appellants’ written grounds of appeal which suggests, by reference to R v Cannings [2004] 2 Cr App R 63, that the existence of a genuine conflict of opinion between distinguished and reputable experts should have that effect. Very wisely, Mr Mansfield did not pursue that suggestion in his oral submissions. Such an argument has been knocked firmly on the head by the judgment of the court in Kai-Whitewind (above) at paras 73-91, where the true scope of what was said in Cannings has been explained at some length. In the present case, as it seems to us, the conflict in expert evidence on count 2 would have been a proper matter for the jury to evaluate in the context of the evidence in the case as a whole; and we cannot say that Dr Walters’s evidence would necessarily have carried the day.

For the reasons we have given, the appeals are allowed on the ground relating to fresh evidence and the appellants’ convictions of manslaughter on count 2 are quashed. We will hear submissions from counsel on whether a retrial should be ordered.