IMPORTANT NOTICE
This judgment may be published provided that family members are not identified. All persons, including representatives of the media, must ensure that this condition is strictly complied with. Failure to do so will be a contempt of court.
Before :
THE RT. HON. LORD JUSTICE PETER JACKSON
Sitting as a Judge of the Family Court
at the Royal Courts of Justice
Between : M -and- ST HELENS BOROUGH COUNCIL -and- F -and- C (a child) | Applicant Respondents |
Karl Rowley QC and Tammi Bannon (instructed by Stephenson’s Solicitors) for the Mother
Andrew Loveridge (instructed by St Helens Borough Council) for the Local Authority
Frank Feehan QC and Janet Reaney (instructed by Berkson Family Law Solicitors) for the Father
Clive Baker (instructed by Morecrofts LLP) for the Child
Hearing dates: 6-10, 13-14 November 2017 Judgment date: 22 January 2018
JUDGMENT
JUDGMENT: St Helens Council v M and F
(Baby with Multiple Fractures: Rehearing)
Lord Justice Peter Jackson:
Introduction
This is the fourth judgment concerning C, a little boy now 3 years old. It arises from an application made on 27 July 2016, by which C’s mother applies to discharge a care order made on 23 September 2015. The application has become a vehicle for the reconsideration of the medical evidence that led to the making of the care order.
The background is set out in my first fact-finding judgment, given on 21 April 2015 and now published at [2015] EWFC 33:
On 27 June 2014, the local authority issued care proceedings in relation to C, a baby boy born in February 2014. His parents had brought him to hospital on 20 June with a swollen right arm. X-rays identified several old and new fractures of the arm and shoulder. A full skeletal survey was carried out. This showed that C had suffered no fewer than 26 fractures, affecting all four limbs, the ribs and the shoulder blades. The fractures had occurred on at least two separate occasions.
C had been born prematurely at 30 weeks gestation with a birthweight of 1660g (3lb 10oz). At the date of his admission to hospital, his chronological age was 4 months and his corrected age was 6 weeks.
The purpose of this hearing is to establish whether the local authority has proved that C’s injuries were inflicted by his parents. They both strongly deny this and rely on expert medical opinion that the injuries may have been the manifestation of a condition as yet unknown to medical science that caused transient fragility in C’s bones. Against this, other expert medical opinion considers it much more probable that the fractures were the result of C being assaulted.
All the doctors agree that extensive investigations have not shown C to have any known medical condition that might explain the fractures. At the same time they all agree that the radiological appearances seen in the skeletal survey are highly unusual. It is also agreed by everyone that C’s mother and father have been consistently observed to be caring parents, both before and after the hospital admission. The basis for intervention in this case rests on the origin of the fractures, and nothing else.
Having heard the medical and non-medical evidence over seven days in March 2015, I found in my first judgment (i) that C did not have a known or unknown medical condition predisposing him to fractures, (ii) that the fractures and other appearances were caused by the application of excessive force to his limbs and torso, and (iii) that one of the parents was responsible but that it was not possible to say which.
The parents’ application for permission to appeal from these findings was refused by King LJ on 26 June 2015.
There was then a welfare hearing over three days in September 2015. I heard evidence from an independent social worker who favoured C’s rehabilitation to his parents, and from C’s social worker and his Children’s Guardian, who did not. For the reasons given in my second judgment, given on 23 September 2015 and now published as [2015] EWFC 78, I preferred the evidence of the social worker and Guardian and made a care order with a care plan for C to remain living with an aunt and have regular visiting time with his parents.
At the time of those family proceedings, the parents were facing criminal charges, in the course of which they obtained two expert opinions that had not been before this court. These were from Dr David Ayoub, an American radiologist, and Professor Stephen Nussey, a consultant endocrinologist. Dr Ayoub disagreed markedly with the previous interpretation of the x-rays, while Professor Nussey raised the possibility that a rare condition (hypervitaminosis A, or HVA) or the interplay of polypharmacy (and in particular the effects of anti-reflux medicine) might be relevant to C’s presentation.
The criminal trial took place in December 2015. After the prosecution medical witnesses (Dr L, Dr Halliday, Dr Ng and Professor Bishop) had been cross-examined with reference to the opinions of Dr Ayoub and Professor Nussey, the parents were acquitted on the direction of the judge. Accordingly, neither Dr Ayoub nor Professor Nussey gave evidence at the trial and they did not undergo cross-examination.
After the criminal proceedings, I published a judgment containing extracts from the fact-finding and welfare judgments that related to practice issues: Re BR (Proof of Facts) [2015] EWFC 41. The publication of the two judgments themselves, which had been postponed to avoid prejudice to the criminal trial, was held back again because the parents gave notice that they would seek a rehearing of the family proceedings in the light of the new medical evidence.
So it was that the mother’s application (in reality a joint one with the father) was issued in July 2016. Careful consideration had been given by the parents’ representatives to the question of whether that application was preferable to a renewed application to appeal out of time, the conclusion, rightly in my view, being that it was.
The reopening of settled findings is a serious matter, but it is a step that will be taken in those rare cases where justice dictates. When considering an application to reopen findings, the court must first be satisfied that the applicant has shown a solid case to justify the step (Re ZZ [2014] EWFC 9). It must then ensure that the scope of the rehearing is appropriate. The staged approach taken in this case mirrors that taken by Cobb J in Re AD and AM (Fact-finding: Rehearing) [2016] EWHC 2912 (Fam).
The application came before me four times:
On 12 September 2016, I directed that Professor Nussey should provide a report that identified his view more closely.
On 30 January 2017, I considered Professor Nussey’s further report and was satisfied (contrary to the submission of the local authority) that it constituted the necessary solid grounds for the original findings to be reviewed. As I said at the time, I would not have been satisfied of this on the basis of Dr Ayoub’s evidence alone. I gave directions for an experts’ meeting to be convened so that the medical issues could be understood and the scope of the rehearing settled.
On 7 April, after a not altogether satisfactory experts’ meeting on 28 March, I determined that the review of the findings should proceed by way of a rehearing of the medical evidence. I did not accept the submissions of the parents that there should be a rehearing of all the evidence, or that the rehearing should take place before a different judge. However, I left open the possibility that the parents might give further evidence if the medical evidence gave rise to a new issue on which they might be able to add to their earlier evidence. In the event, that did not arise.
On 27 July, pre-trial directions were given. In fact, the court could have accommodated the rehearing locally in July, but the parents understandably wanted continuity of counsel which could not be achieved then.
Approach to fact-finding
I again apply the principles set out at [5-14] of my first judgment, emphasising these paragraphs:
Each piece of evidence must be considered in the context of the whole. The medical evidence is important, and the court must assess it carefully, but it is not the only evidence. The evidence of the parents is of the utmost importance and the court must form a clear view of their reliability and credibility.
When assessing alternative possible explanations for a medical finding, the court will consider each possibility on its merits. There is no hierarchy of possibilities to be taken in sequence as part of a process of elimination. If there are three possibilities, possibility C is not proved merely because possibilities A and B are unlikely, nor because C is less unlikely than A and/or B. Possibility C is only proved if, on consideration of all the evidence, it is more likely than not to be the true explanation for the medical findings. So, in a case of this kind, the court will not conclude that an injury has been inflicted merely because known or unknown medical conditions are improbable: that conclusion will only be reached if the entire evidence shows that inflicted injury is more likely than not to be the explanation for the medical findings.
Where, as here, there is a genuine dispute about the origin of a medical finding, the court should not assume that it is always possible to know the answer. It should give due consideration to the possibility that the cause is unknown or that the doctors have missed something, or that the medical finding is the result of a condition that has not yet been discovered. These possibilities must be held in mind to whatever extent is appropriate in the individual case.
To this I add these observations from Re U; Re B (Serious Injury: Standard of Proof) [2004] EWCA 567:
Particular caution is necessary where the medical experts disagree, one opinion declining to exclude a reasonable possibility of natural cause.
The court must always be on guard against the over-dogmatic expert, or where the expert’s reputation or amour propre is at stake, or where an expert has developed a scientific prejudice.
The Judge in care proceedings must never forget that today’s medical certainty may be discarded by the next generation of experts and that scientific research may throw light into corners that are at present dark.
I also recall Webster v Norfolk County Council [2009] 1 FLR 1378, a case of late discovery of scurvy, where Wilson LJ said:
“The moral I draw from this case and will never forget is that a hypothesis in relation to the causation of a child’s injuries must not be dismissed only because such causation would be highly unusual and that, where his history contains a demonstrably rare feature, that the possible nexus between that feature and his injuries must be the subject of specialist appraisal at an early stage.”
Thus, regarding the standard of proof (discussed in my first judgment), I accept that caution is required when considering the significance of evidence from very experienced specialists that a condition is very rare, not seen in practice, or indeed never seen nowadays. Self-evidently, rare and unknown conditions do exist. In each case, the court must give proper consideration to the possibility that it is dealing with a case where something has been missed because the search that might have found it has not been carried out, or because we do not know what we are looking for. That discipline is particularly needful here, where medical witnesses on both sides of the argument agree that C’s case is unusual to the point that several of them described it as unique, whether it is a case of inflicted injury or not. This characteristic undoubtedly heightens the need for the most careful scrutiny of the evidence, with particular attention being paid to the possibility that the presentation results from an unknown cause.
In that regard, I bear in mind the following passage from R (D) v Life Sentence Review Commissioners [2008] 1 WLR 1499 [2008] UKHL 33 at [28]:
“28. … in some contexts a court or tribunal has to look at the facts more critically or more anxiously than in others before it can be satisfied to the requisite standard. The standard itself is, however, finite and unvarying. Situations which make such heightened examination necessary may be the inherent unlikelihood of the occurrence taking place (Lord Hoffmann's example of the animal seen in Regent's Park), the seriousness of the allegation to be proved or, in some cases, the consequences which could follow from acceptance of proof of the relevant fact. The seriousness of the allegation requires no elaboration: a tribunal of fact will look closely into the facts grounding an allegation of fraud before accepting that it has been established. The seriousness of consequences is another facet of the same proposition: if it is alleged that a bank manager has committed a minor peculation, that could entail very serious consequences for his career, so making it the less likely that he would risk doing such a thing. These are all matters of ordinary experience, requiring the application of good sense on the part of those who have to decide such issues. They do not require a different standard of proof or a specially cogent standard of evidence, merely appropriately careful consideration by the tribunal before it is satisfied of the matter which has to be established.”
This court’s fact-finding enquiry, now amounting to 14 days of evidence extending across two hearings, has surely been the ‘heightened examination’ appropriate to the seriousness of these allegations and of their consequences.
As this is a rehearing:
My previous findings are the starting point. This judgment and its predecessor should be read together.
The parents must present evidence to challenge those findings (the evidential burden).
However, the legal burden of proving its case on the balance of probabilities remains on the local authority throughout.
All the evidence, medical and non-medical, must be fully assessed.
There is no priority as between evidence arising at the earlier and later hearings, nor between evidence that was given orally at the rehearing and evidence that was not.
This hearing
The written evidence, which is extensive, encompasses the papers from the earlier family proceedings, the criminal proceedings and the current proceedings. There are transcripts of previous evidence and copies of many articles from the medical literature.
The following witnesses gave evidence (*witness who gave evidence in the 2015 family proceedings; †witness who gave evidence in the criminal trial):
Dr David Ayoub, radiologist, Springfield, Illinois
*Dr Jeremy Allgrove, consultant paediatrician and paediatric endocrinologist, Great Ormond Street Hospital
*†Dr Sze May Ng, consultant paediatrician and paediatric endocrinologist, Southport and Ormskirk NHS Trust
†Professor Nicholas Bishop, Professor of Paediatric Bone Disease, Sheffield Children’s Hospital
Prof Stephen Nussey, Emeritus Professor of Endocrinology, St Georges Hospital Medical School
*†Dr Katharine Halliday, consultant paediatric radiologist, University College Nottingham
Approach to the medical evidence
Inevitably in terms of length, this judgment is predominantly concerned with the medical evidence. Before embarking on that, I reemphasise that this is only part of the overall evidence – as Mr Baker for the Guardian puts it, the expert evidence can only take the court so far.
I also reiterate what I said during the hearing: the outcome does not narrowly turn upon whether or not a new theory (for example HVA) is accepted or rejected. It is not for the parents or their witnesses to prove anything. Once the rehearing has been embarked upon, all the medical and other evidence must be fully reassessed, and the same ultimate question must be answered: Has the local authority proved that inflicted injury is more likely than not to be the cause of C’s condition?
In that regard, I agree with the submission made on behalf of the parents that, since the answer to the ultimate question is contentious, all theories are hypotheses to be tested against the information available from all sources. Where, as here, there is no direct evidence of assault, a diagnosis of inflicted injury is as much a hypothesis as any other diagnosis, as is a hypothesis of ‘unknown cause’.
Having said that, it is rightly no part of the parents’ case that the lack of medical unanimity relieves the court of the duty to sift the evidence and reach conclusions. It is that task that must now be performed.
Lastly, hearings of this kind may throw up memorable phrases that find their way from the evidence into the submissions. Examples in this case are:
Dr A, treating paediatrician, describing C in the clinical note of an examination six days before admission: “a picture of health”
Dr Ayoub, in written reply to questions in April 2017: “That C suffered from metabolic bone disease is beyond certain.” [E378]
Dr Allgrove, in reply to Mr Feehan: for a paediatrician to describe a child with so many fractures to normal bones as ‘a picture of health’ six days before admission is “unthinkable”.
While vivid remarks of this kind have their place amidst the other evidence, they should not attract more weight than they deserve simply because they are memorable.
The expertise and approach of the medical witnesses
Evidence has been gathered from the full range of relevant specialisms: paediatrics, paediatric surgery, paediatric radiology, paediatric endocrinology, paediatric bone disease, genetics, haematology, endocrinology and radiology.
I discuss the position of Dr Ayoub separately.
Drs Allgrove, Ng and Halliday, and Professors Bishop and Nussey all gave evidence appropriately from their professional standpoint. All are considerable specialists in their own disciplines and respected the frontiers of their expertise when giving evidence. Each of them was willing to acknowledge the claims of other disciplines and each possessed a working knowledge of the science beyond their own specialisms that enabled them to express a coherent overall opinion. So, the radiologists were able to situate their interpretations of the x-rays within the general context of the biochemistry, while the other experts demonstrated a good understanding of the general significance of the x-rays. During the course of the hearing, the disciplinary borders were appropriately policed by counsel, but in the end I am satisfied that these five specialists went no further outside their own territory than was necessary to address the questions on which the court required their help.
It is convenient at this point to address concerns expressed on behalf of the parents about the approach taken by the three medical witnesses upon whom the local authority principally relies: Dr Ng, Dr Halliday and Professor Bishop.
In the case of all three experts, the suggestion is that they arrived at the conclusion that this C suffered inflicted injuries by a process of exclusion, and that consequently they did not give sufficient weight to the possibility of an unknown cause. In Dr Ng’s case, the submission rests upon her reference to normal child protection procedures whereby tests are carried out to identify metabolic conditions that might rule out inflicted injury. In Dr Halliday’s case, it rests upon her evidence that with findings of this kind she started from a hypothesis of inflicted injury and then sought to rule it out. With Professor Bishop, his approach was that in cases referred to him, he was looking for an explanation of bone disease.
I considered the submission of this kind in relation to Dr Ng at paragraph 250 of my first judgment. My view then in relation to her, and my view now in relation to all three witnesses, is that this concern is more apparent than real. Anyone considering a case of this kind must start somewhere, and doctors and lawyers may not necessarily start from the same places. What is important when considering the weight to attach to an expert witness’s evidence is the substance of their investigation and the way in which they have formed their views. What matters is the journey and not just the starting point. Has the witness given open-minded consideration to all the realistic possibilities, or has their opinion been limited by an impermissible thought-process? Having readdressed this issue, I reach the clear conclusion that whatever proper reliance can be placed on the views of these three witnesses should not be discounted on this score. Each of them justified their opinions on the basis of what they saw as positive supportive evidence. Each was willing to consider all the available information and none of them arrived at their view by a process of exclusion.
In the case of Professor Bishop, additional criticism was directed at him for not looking at the x-rays again after receipt of Dr Ayoub’s report, or subsequently. He is also taken to task for points of inaccuracy in his police statement in March 2015 (the details appearing in paragraph 28(a)(ii) of the parents’ closing submissions). Finally, he is criticised for a comment he made at the experts meeting about the parents’ presentation when they attended his clinic with C.
As to these matters, my view is this. I do not consider that Professor Bishop’s evidence was materially hampered by the fact that he had not looked at the radiographs recently. He was not alone in this, and had it been considered important for him and the other experts to do so, it should have formed part of the extensive series of preparatory directions given for the rehearing. Further, had a point been reached during the hearing where any witness was unable to deal with an issue because of the absence of the original radiographs in court, arrangements could have been made to remedy this, but that did not arise. Secondly, insofar as there are limited inaccuracies in Professor Bishop’s summary of the x-rays, I bear in mind that he is not a radiologist and that he was commenting on the overall position alongside more detailed radiology evidence. Lastly, the Professor accepted and regretted the comment made at the experts meeting. It is not suggested that this compromised his independence; rather, it formed part of an overall submission that he had not approached this case with the necessary care. Having heard Professor Bishop giving evidence for over three hours, mostly under cross-examination, I do not accept that submission. I consider that he, like the other witnesses, has fully engaged in applying his expertise to the unusual facts of the case.
I therefore set aside these issues and shall assess the opinions of the five domestic experts on their merits.
Dr David Ayoub
Dr Ayoub is a general radiologist practising in Springfield, Illinois. He was certified by the American Board of Radiology in 1990 and is licensed to practice medicine in the states of Illinois, Iowa and Missouri. Most of his time is spent in the clinical practice of general radiology, with the remainder devoted to legal work (currently by his estimate 5-10%, though it has been as much as 25%).
Dr Ayoub accepted the proposition that his clinical practice concerns adults and his legal work concerns babies. He has never trained in paediatric radiology and his clinical work does not include infants (children under one year old). He has seen just one case of severe rickets in his clinical practice. By contrast, his legal work, stretching back over 10 years, is entirely focussed on infants. This involves him in writing reports and giving evidence, always for the defence, in cases where adults have been accused of causing fractures in infants. 70% of these cases are civil, the remainder criminal. This is unpaid work that he performs out of an avowed conviction that many cases of alleged child abuse are misdiagnosed and are in fact cases of metabolic bone disorder, particularly rickets, which he described as “vastly underdiagnosed”. He considers that he has an ethical duty to correct this situation. When I asked him where he considered his views would be considered controversial, he said that they would to those who are unaware or ignorant of the published support for them.
Dr Ayoub stated that he is sent 3-5 legal cases every week to consider, has been consulted in about 500 cases, has written approximately 200 court reports and has appeared approximately 80 times as a witness in proceedings in the United States. He has been engaged in two other cases in this country and one in Sweden. In every case in which he has written a report, he has expressed the opinion that the child in question suffered from a metabolic bone disorder. In a television interview given in about 2010, he said this: “I’ve not seen any high-risk family. I don’t believe any case of fractures I’ve seen has been as a result of real physical child abuse, that it’s metabolic.”
Dr Ayoub has researched the literature extensively and, according to his CV, has published these letters and articles:
Ayoub D, Plunkett J, Keller KA, Barnes PD. Letter to editor: Are Paterson's critics too biased to recognize rickets? Acta Paediatr. 2010;99(9):1282-3.
Hyman CJ, Ayoub D, Miller ME. Letter to editor: Response to Taylor et al, Comments on Making the Diagnosis of Rickets in Asymptomatic Young Children. Clin Pediatr. 2010;50(5):474.
Hyman CJ, Ayoub D, Miller M. Response to Vinchon. Letter to editor: Hyman CJ, Ayoub D, Miller M. Childs Nerv Syst. 2011;27(2):201.
Ayoub D. Letter to editor: Fractures: Abuse or rickets? Radiology 2012. 264(2):614-5.
Ayoub D. Limitations of radiology in rickets. Letter to editor: Pediatr Dev Pathol. 2013 Sep-Oct;16(5):397
Ayoub D, Hyman CJ, Cohen M, and Miller M. A critical review of the classic metaphyseal lesion (CML): Traumatic or metabolic? AJR Am J Roentgenol. 2014 Jan;202(1):185-96.
Ayoub D, Hyman C, Miller M. Reply: To PMID 24370143. AJR Am J Roentgenol. 2014 Jun;202(6):W604.
Ayoub D, Miller M, Hyman C. Reply: SPR Child Abuse Committee Response Regarding Classic Metaphyseal Lesion. AJR Am J Roentgenol. 2014 Aug;203(2):W233.
Miller M, Ward T, Stolfi A, Ayoub D. Overrepresentation of multiple birth pregnancies in young infants with four metabolic bone disorders: further evidence that fetal bone loading is a critical determinant of fetal and young infant bone strength. Osteoporos Int. 2014 Jul;25(7):1861-73.
Paterson CR, Ayoub D. Congenital rickets due to vitamin D deficiency in the mothers. Clin Nutr. 2014 Dec 17. [Epub ahead of print] Review
Ayoub D. Letter to editor: Misinterpretation of birth-related growth plate injuries as classic metaphyseal lesions. Journal of Forensic Radiology and Imaging, 2015 in press.
Those working in the field will recognise the references to Dr Colin Paterson, proponent of the theory of temporary brittle bone disease, who was criticised by Singer J in Re X (Non-accidental Injury: Expert Evidence) [2001] 2 FLR 90 and who was struck off the medical register in 2004, and to Dr Charles Hyman, the California paediatrician, whose approach as an expert witness was the subject of criticism by Cobb J in Re AD and AM (above). Dr Ayoub said that he knows and sympathises with Dr Paterson and also with Dr Andrew Wakefield, whose views on the effects of childhood vaccination led to him being struck off the medical register in 2010.
Dr Ayoub’s forensic work has not surprisingly attracted public attention in the United States, and he regularly speaks publicly on how in his view metabolic bone conditions, such as rickets, mimic child abuse.
The role of expert witness is a privileged one, allowing the witness to give an opinion on questions outside the court’s own expertise. Such opinions are likely to be influential on the court’s thinking. Where they are soundly based, they serve the interests of justice. Where they are not, the opposite is the case. Judges and juries can be misled, parties may come to believe that their cases are stronger or weaker than they truly are, and subsequent assessments may be confounded.
It is accordingly essential that expert opinions arise from a trustworthy process. Rules made in our civil, family and criminal jurisdictions aim to provide quality control. In family proceedings, these appear at Rule 25 of the Family Procedure Rules and the associated Practice Direction 25B. In criminal proceedings, Rule 19 and PD 19 to the Criminal Procedure Rules 2015 apply. These codes are not identical, but they have common features in that
the court retains control of whether evidence will be admitted (at the outset in family proceedings, ultimately in criminal proceedings),
the expert must give an opinion that is objective and unbiased, and
where the expert propounds a controversial hypothesis, s/he must draw attention to the fact.
Dr Ayoub’s involvement arose from an approach by the parents themselves ahead of their criminal trial. He knows that there are rules governing the terms on which experts give evidence in our courts, but he did not profess to know what they contain.
It is not seriously disputed between the parties that if the Family Court had been asked to approve the prior instruction of Dr Ayoub as an expert witness, it would have been unable to do so. There are two fundamental reasons. Firstly, he does not have the necessary expertise to offer an opinion to a court on the origin of radiological appearances in infants, particularly pre-term infants, as they are a patient cohort of which he has no clinical experience. Secondly, his approach is shot through with the dogma that child abuse is over-diagnosed. It does not matter for this purpose whether he is right or wrong. The expert with a scientific prejudice may perform a service to science by asking questions that challenge orthodoxy, but be unsuited to be an expert witness, a role that requires objectivity when giving answers.
Nothing in Dr Ayoub’s evidence in the present case led me to a different view. He made himself available at an early hour at personal inconvenience and gave his evidence in a serious manner. However, his evidence was characteristic of his general approach. Having taken up a position, he advanced it with the tenacity of an advocate and was dismissive of alternative possibilities. He entertained no doubts about the correctness of his opinion, a dangerous mindset for any expert witness.
I therefore conclude that the family or criminal courts in England and Wales are unlikely to find that Dr Ayoub meets the requirement that an expert witness must be objective and unbiased. At all events, if it is proposed that he should give evidence in any future case concerning fractures in infants or young children in this jurisdiction, the relevant court should be made aware of the matters contained in this judgment.
Here however, his evidence has been heard and I will refer to it as appropriate.
My previous findings in relation to the medical evidence
As to the radiological abnormalities, based on Dr Halliday’s opinion (similar to that of Dr L), I found the following:
The initial x-ray on 16 June covered the right arm and shows the following fractures, with * indicating fractures that show no sign of a healing reaction:
1* Right acromion (upper part of the shoulder blade)
Metaphyseal fracture of proximal right humerus
Metaphyseal fracture of distal right humerus
4* Comminuted fracture of the shaft of the right humerus occurring through the PR
Fracture of distal right radius and ulna
The skeletal survey of 16 June shows the following further fractures:
Right middle metacarpal
7* Left acromion
Metaphyseal fracture of proximal left humerus
Metaphyseal fracture of distal left humerus
Distal left radius
Metaphyseal fracture distal right femur
Distal right tibial shaft
Distal right fibula shaft
Metaphyseal fracture distal left femur
Metaphyseal fracture proximal left tibia
Metaphyseal fracture distal left tibia
Distal left tibia shaft
Distal left fibula shaft
Left 6th, 7th, 8th, 9th and 10th ribs laterally
Right 6th and 7th ribs laterally
The skeletal survey of 16 June also shows quite florid PR at the following further sites where no fractures are seen: the clavicles, the ulnae, the first metatarsals and the scapulae. There are no wormian bones in the skull.
…
There are no significant disagreements between the experts about the radiological appearances themselves. …
Dr Halliday considered the reports of the original radiologists that the bones on the June x-rays appeared osteopenic (white). She agreed that the right arm looked a bit osteopenic but says that this was a very subjective and inaccurate way of assessing bone strength and did not indicate a predisposition to fracture. It might be the result of disuse of the limb following fracture. Osteopenia of prematurity is common in premature babies but there is no generalised osteopenia in the skeletal survey.
…
On C’s June x-rays, the PR is abundant at the fracture sites and also on bones that are not fractured, for example the clavicles and the 1st metatarsals in the arches of the feet. These appearances are extremely striking. The PR around each fracture is greater than usually seen and can be described as 'florid' or 'exuberant' in terms of the width of the reaction and the extent of the bones affected. Dr Halliday had never seen a case with so much PR on so many bones. She had seen PR this wide in isolated bones but not in so many bones. Its width and extent is unique in her experience.
The bones without PR are most of the unfractured ribs, most of the hand and feet bones, the spine, the pelvis and the bones at the base of the skull.
As to the sites of injury, again following Dr Halliday’s opinion, my findings were these:
As to the fracture sites:
Eight of the limb fractures are metaphyseal. These are associated with inflicted injury (pulling or twisting).
Metaphyseal fractures are not associated with bone disorders, where midshaft fractures are common. C had only one midshaft fracture (the upper right arm).
Fractures to the acromia are also associated with inflicted injury and can be caused by the arms being yanked.
The rib fractures are of a kind that would be caused by squeezing or squashing.
Fractures to the 1st metatarsals are also associated with inflicted injury resulting from the foot being twisted. Dr Halliday speculated that the PR on these bones might be associated with fractures that cannot be seen.
Fractures to the clavicles are also associated with inflicted injury. PR on the clavicles is an unusual appearance without an identified fracture, which Dr Halliday could not fully explain, unless it was the result of adult thumb pressure without fracture, or of rapid growth, or a combination of the two.
…
As to the extent of the PR, Dr Halliday thought it likely that this was the result of repeated handling affecting fractures that had not been immobilised, leading to much more PR in the area. This may explain why the PR is exuberant in the limbs, which are subject to twisting, but not on the ribs, where there are only blobs of callus at the fracture sites. Less likely as a cause of extensive PR is rapid growth, though this may contribute.
As to known differential diagnoses, I found that C did not have:
Scurvy
Caffey’s Disease
Neuroblastoma
Leukaemia
Congenital syphilis
Healing rickets
I-cell disease
Job Disease
Osteogenesis imperfecta
Osteomyelitis
Temporary brittle bone disease (sic)
Metabolic bone disease of prematurity
Osteopetrosis
Nor was there evidence of significant deficiency in vitamin C, vitamin D or copper.
My overall conclusions from the medical evidence were these:
…
C had normal bones.
The cause of the fractures was the application of excessive force to his limbs and torso.
The mechanisms described by Dr Halliday and Dr Ng (gripping or squeezing of the chest, twisting or bending of the limbs) would cause fractures of this nature.
The absence of external injury (bruising or other marks) is noteworthy but such injuries would not necessarily be expected unless there were blows to the body.
The eight metaphyseal limb fractures and the fractures to the shoulder blades are of a kind associated with inflicted injury.
C did not suffer from any of the numerous rare medical conditions considered by the specialists. There are no grounds for believing that the doctors have missed something.
The highly distinctive radiological picture provides a good reason to consider the possibility that C may have suffered from an as yet undiscovered transient metabolic bone disorder, but having considered the possibility in depth, I find that there is no good reason to believe that he did in fact suffer from such a condition.
The only reason for particular attention having been paid to Caffey’s Disease is that it is a transient metabolic bone condition. In all other respects, C’s condition is quite different from Caffey’s, which affects whole bones in different parts of the anatomy and is not associated with fractures.
A diagnosis of "unknown cause" is a respectable one in any case when the cause of a medical condition is unknown, and will particularly arise in the context of controversy at the frontiers of medical science: the authorities referred to at paragraph 17 above arise from cases of baby deaths involving "the triad" of head injuries and possible Sudden Infant Death Syndrome.
In this case, I find that the cause of the fractures is not unknown, even in relation to the clavicles and the metatarsals. Given that the injuries to the other bones were the result of assaults, it is entirely possible that the PR seen on the clavicles was the result of force applied by adult thumbs that disrupted the periosteum without causing any identifiable fracture. Likewise, the PR seen on the metatarsals can be explained as being the result of an application of force, perhaps by a twisting of the feet, with or without fracture.
Such symmetry as exists in the distribution of C's fractures is not inconsistent with injury being inflicted during momentary losses of control. I do not accept that there would have to be a systematic approach. Each of the injuries could readily have been caused by sudden, momentary force applied by an adult who was holding the child’s arms, legs or chest with both hands.
The wide extent and florid nature of the PR in this case is explained by the profusion of the fractures and the fact that they were not immobilised and will have been disturbed by repeated handling. It is not possible to know whether rapid growth played a part, but it may have done.
Even if C may have had a mild degree of osteopenia of prematurity, it would not have significantly predisposed him to fracture.
C suffered from gastro-oesophageal reflux which proved intractable. Medical personnel believed that this was his only complaint, and treated him accordingly. It is now clear that he was suffering from far greater difficulties, beginning at some point in the six weeks before his admission. This may coincide with the falloff in his weight gain shown by the growth charts, but that would be speculation. Nonetheless, the presence of so many untreated fractures undoubtedly affected his response to being handled, so that he has been described by his parents and grandfather as being "antisocial" or as wanting to be left alone. It is also very likely that the fractures made him even more difficult to feed.
The routine examinations by Dr O and Dr A did not cause fractures. I accept that iatrogenic fractures have been known to occur, particularly with extremely preterm infants, but, having heard from these doctors, I am satisfied that that is not the case here.
The investigations marshalled by Dr A were organised and thorough. There was no question of any bias towards a diagnosis of inflicted injury; on the contrary, he began from the working assumption that the cause might be an organic one. He took appropriate specialist advice from tertiary centres in reaching his clinical conclusions.
Overview of the medical evidence at this hearing
The evidence of Dr Ayoub and Professor Nussey added two new strands to the debate:
Dr Ayoub fundamentally challenged the radiology assessment, arguing that C suffered from healing rickets, amongst other things. This view had little or no support from the other witnesses. Nevertheless, two particular aspects of the radiology received closer attention on this occasion: (a) the appearance of PR on the shoulder blades (scapulae), and (b) the appearance of the third metacarpal in the right hand. The latter led to a debate about whether C may have had secondary hyperparathyroidism (SHPT), the effects of which are often seen in the long bones of the hands, and if secondary to low calcium, can be in response to or evidence of vitamin D deficiency.
Professor Nussey’s theory that HVA may have caused or contributed to C’s presentation was extensively considered and gained some qualified support from Dr Allgrove.
Even taking the above into account, the evidence of the three original witnesses (Drs Halliday, Allgrove and Ng) was very similar to their evidence in 2015. Having reconsidered the matter, they have each arrived at the same conclusions. Likewise, the oral evidence of Professor Bishop confirmed the written opinion he had previously given.
An overview of the medical opinions is given in this chart, prepared by Mr Baker following the experts meeting and subsequent exchanges:
On the balance of probabilities: | Ayoub | Nussey | Bishop | Ng | Halliday | Allgrove |
The number of fractures has been overestimated | Agree | N/A | Disagree | N/A | Disagree | N/A |
It is likely that C’s bones were adversely affected by vitamin D deficiency in utero | Agree | Agree | Disagree | Disagree | Disagree | Disagree |
C had rickets | Agree | Disagree | Disagree | Disagree | Disagree | Disagree |
A reasonably possible alternative and/or additional diagnosis is that C had vitamin C deficiency/scurvy however it is not possible to establish this retrospectively | Agree | Disagree | Disagree | Disagree | Disagree | Disagree |
A reasonably possible alternative and/or additional diagnosis is that C had copper deficiency | Agree | Disagree | Disagree | Disagree | Disagree | Disagree |
C suffered from hypervitaminosis A leading to the radiological appearances of hyperostosis and contributing to fractures | N/A | Agree | Disagree | Disagree | Disagree | Equivocal |
C has or had an undiagnosed metabolic bone disease | Agree | Equivocal | Disagree | Disagree | Disagree | Agree |
The cause of C’s fractures is inflicted injury | Disagree | Disagree | Agree | Agree | Agree | Disagree |
I shall shortly summarise the evidence of each witness. In the case of those who gave evidence in 2015, the accounts should be read together with their earlier evidence as described in the first judgment.
Evidence of Dr Ayoub
Dr Ayoub opined as a matter of certainty that C had an extreme form of rickets that was by June in the healing stage. He said that rickets is one of the least understood diseases in Western medicine and is highly prevalent on autopsy: see Marta Cohen and others Vitamin D Deficiency and Sudden Unexpected Death in Infancy and Childhood : A Cohort Study; Pediatric and Developmental Pathology 16, 292–300, 2013.
Significant features of his interpretation of the x-rays included:
A smaller number of fractures. He accepted that there were 11 fractures – the right radius, left radius, right fibula, five left ribs and three right ribs
Two appearances – the right humerus and left tibia – are interpreted as being fractures of PR, not bone.
Trabecular bone is seen resulting from severe osteopenia.
There is striated cortical bone on most of the long bones.
The findings at the bone ends are staged ratchitic growth plate recovery with pericondrial rings and slipped physes/epiphyses. They are not what has been described as the classic metaphyseal lesion. Metaphysis abnormalities are often not fractures but instead features of rickets. Dr Ayoub does not accept the assertion (based on the work of Kleinman) that metaphyseal fractures are commonly abuse injuries and he has published a review of the medical literature contending that such fracture are regularly misinterpreted: see A Critical Review of the Classic Metaphyseal Lesion: Traumatic or Metabolic? AJR:202, January 2014: David M. Ayoub, Charles Hyman, Marta Cohen and Marvin Miller.
In the February x-rays, there are signs of diminished mineralisation.
There is bowing of the scapulae – not a normal variant, and not commented upon by other radiologists.
There is a Looser zone deficit in the acromion. (Looser zones are wide, transverse lucencies traversing part way through a bone, usually at right angles to the involved cortex and are associated most frequently with osteomalacia and rickets. They are pseudo-fractures and considered a type of insufficiency fracture. Typically, they have sclerotic irregular margins and are often symmetrical.)
There are signs of scurvy in the forearm.
There is dramatic under-mineralisation in the hands.
There is striking, symmetrical rickets in the femoral shafts – pathognomonic of rickets.
There is a uniformity of PR that would not be produced by fractures.
Dr Ayoub said that these were in his opinion all signs of healing rickets, and that such signs did not have to be evenly spread. As to the number of fractures, it did not surprise him that Dr Halliday sees many more fractures: he would himself have said the same thing 10 years ago. He strongly disagreed with Professor Bishop, who excluded bone disease of prematurity, describing this as a reckless conclusion.
Dr Ayoub raised the possibility (unprovable but acknowledged by Professor Nussey) of HPT resulting from suggested insufficiency of vitamin D and calcium absorption leading to elevated parathyroid hormone levels in the blood and consequent reclaiming of calcium from the bones, seen for example in the right third metacarpal.
Dr Ayoub expressed the view that the exaggerated PR should have led to consideration of collagen disease including scurvy and copper deficiency. He described copper deficiency as a reasonable proposition for C. He says that though the copper and vitamin C levels were normal when taken, the timing of the obtaining of those results is important - i.e. that they were taken when C had been an inpatient in hospital for over 2 weeks, and was on supplementation with Dalivit to the maximum dosage.
Overall, based on there being so many fractures and so much PR, he favoured a combination of deficiency diseases, including (principally) rickets, scurvy, vitamin D and copper deficiency, causing a transient disease of early infancy, with the fractures slowing down as time passed with a natural history of resolution by the age of eight months. He attached great significance to the absence of external indicators of abuse. He did not attach any significance to the coincidence of fractures and parental care, or to the absence of accounts of C being in pain, saying that is well-known that undisplaced fractures hurt less.
Evidence of Professor Nussey
Professor Nussey is Professor of Endocrinology at St Georges Hospital Medical School. At pages [E4-6] of his initial report in the criminal proceedings in November 2015, he gives useful background information about endocrinology, calcium, parathyroid hormone and vitamin D, vitamin A, and hypocalcaemia.
The thesis that C’s condition may have been caused or contributed to by vitamin A excess (HVA) was first proposed in that report:
“A factor that has not been considered by Professor Bishop is the supply of calcium. It has long been recognised that the inhibition of gastric acid production reduces calcium absorption. Fibre within the diet may also reduce calcium absorption. Used as a feed thickener specifically for reflux, locust bean gum safety in neonates and young infants has been reviewed (Meunier et al. 2014). However, this is not been done in the context of other medications such as Gaviscon or Ranitidine used in the treatment of gastro-oesophageal reflux. In the case of C it is to be noted that its use is not recommended in pre-term infants (ibid). Studies in vitro have shown that locust bean gum reduces the availability of materials such as calcium and iron and this is likely to be exacerbated by a reduction in gastric acid production (Bosscher et al. 2001).
Professor Bishop excluded the diagnosis of Caffey’s disease because of the lack of associated clinical features. However, it is noteworthy that neither he nor indeed Dr Halliday included hypervitaminosis A in the differential diagnosis. Whilst I am not a radiologist, this is usually included in the textbook differential diagnosis of the characteristic exuberant periosteal hyperostosis. Furthermore, there are reports of hypervitaminosis A affecting the metaphyses of long bones.
Vitamin A is rapidly absorbed and slowly cleared. It is stored in the liver. In excess, it is the only toxic substance known in rats to lead to spontaneous fractures. Although vitamin A toxicity has been known about for centuries and characterised since the 1940s, there remain large lacunae in knowledge. It is recognised that toxicity is seen at lower doses in the young and elderly, and that there may be other factors exacerbating toxicity. Some of these may be genetic (Hathcock et al., 1990). In vulnerable groups such as children toxicity has been reported with doses of the order of 1500IU/kg/d (ibid). The presentation of toxicity varies with the dose and duration of exposure, as well as the age of the individual exposed. C was noted to have irritability, failure to thrive, persisting, anaemia, and by inference, bone pain and tenderness.
The Committee on Toxicity of the Food Standards Agency (CoT 2013) has recommended for infants, a Tolerable Upper Limit (TUL) for preformed vitamin A of 200µg RE/kg/day. Human breast milk contains circa 80µg RE/100ml; Cow & Gate Stage 1 55µg RE/100ml; human breast milk with fortifier contains 315µg RE/100ml (Cow & Gate; Reference). I am not a Neonatologist, and it is not possible accurately to calculate vitamin A intake in the SCBU from the available material, but it is to be noted that C received human breast milk with fortifier during this admission and after discharge until about the middle of April, when he was eight weeks old. At this time he was receiving 2500IU vitamin A in Dalivit and this was continued subsequently. Depending on the number of sachets of fortifier added to each 100ml of breast milk. It may be roughly calculated that he was receiving about fourfold the TUL and an amount of vitamin A that has been associated with toxicity in reported cases.
It is noted that the Dalivit dose was increased in June 2014 and yet the radiological appearances of C’s bones were reported to have improved 5 months later. This may be accounted for by a number of factors. His growth meant that the dose of vitamin A per kg body weight was reduced. There may also have been maturation of changes in the liver. In addition, it is known that there is a potential antagonism between vitamins A and D. He was likely to have been born with a marked vitamin D deficiency and thus to have been more sensitive to vitamin A excess. As the vitamin D stores improved. He may have become more resistant to vitamin A. In addition, he was no longer taking any medication for reflux after discharge. As there is evidence that histamine has physiological effects on bone cells (both osteoclasts and osteoblasts) it is possible that histamine (H2) receptor antagonism may have played a role (perhaps idiosyncratically in C) in the effects on bone. This effect was removed when the ranitidine was stopped in June 2014.”
As can be seen, the hypothesis as described above has a number of component steps:
That C was born with bones adversely affected by vitamin D deficiency in utero, given the autumn/winter pregnancy and the lack of prescription of vitamin D supplements to his mother.
That vitamin D deficiency and subsequent insufficiency may have created increased sensitivity in C to vitamin A excess.
That C’s calcium absorption may have been adversely affected by a ‘polypharmacy’ created by the acid neutralisers Gaviscon and Domperidone and the acid inhibitors Ranitidine and Omeprazole, breast-milk with supplements, formula feed, fortifier/feed thickener containing locust bean gum, and Dalivit.
That the absence of pathology five months later may be explained by improvements in C’s vitamin D levels alongside a maturational process.
It would be fair to say that Professor Nussey’s view in favour of HVA as the best explanation for C’s presentation had firmed up from the time of his first report to the time when his evidence at this hearing ended.
These are some of the principal aspects of his further evidence:
This is a very unusual case, with no precedent in the child abuse literature and difficulty in finding a similar case of hyperostosis (PR) arising from HVA.
If HVA is associated with such severe hyperostosis, C would be the first premature infant to be thus affected. He would also be the first with so many fractures. Whatever its cause, the case should be reported in the medical literature.
Inflicted injury is a possible cause, either with or without weakened bones.
None of the biochemical readings on admission (raised ALP and low phosphate levels, and the lack of any parathyroid hormone result) is incompatible with the theory of HVA.
The dosages administered in the papers referred to (Caffey, Hathcock) were exponentially higher and more prolonged than those in this case.
The Caffey study shows a long latent period (six months) before vitamin A toxicity took effect.
None of the clinical signs of HVA, and in particular no bulging fontanelle, was seen.
As his evidence proceeded, it became clearer that the Professor accepted the necessity for his theory that C had some transient, unique and possibly genetic hypersensitivity to vitamin A. He observed that the possibility of something uniquely idiosyncratic in C might accord with the unique radiology.
At the end of the his evidence, an abstract from a recent unpublished US study was produced (Malchodi L, et al. ‘Early Antacid Exposure Increases Fracture Risk in Young Children’ May 2017). Professor Nussey was understandably reluctant to comment on an abstract, but said that some contributory effect to demineralisation from antacid use is to be expected.
Evidence of Dr Allgrove
Dr Allgrove, consultant paediatrician and paediatric endocrinologist at Great Ormond Street Hospital, is a specialist in disorders of calcium. He remains of the view that C was the victim of an undiscovered metabolic bone disease, based upon the radiological appearances. He considers it extraordinarily unlikely that the appearances could be accounted for solely on the basis of inflicted injury.
Based on his considerable experience of children with rickets, he does not think that C had rickets. The appearance of the skeleton at birth were normal, and though his vitamin D was relatively deficient on admission at 38 nmol, it was not striking. The later x-rays do not have the classic appearances of rickets.
Taking full account of the timings of tests, copper levels were normal and the vitamin C level was normal. Scurvy is vanishingly rare. There is no evidence that C was not receiving normal calcium intake through breastmilk and formula feeds. He is not aware that use of the anti-reflux medications give rise to calcium deficiency, though they may be a factor in reducing calcium absorption.
Dr Allgrove conceives of the possibility that what Dr Ayoub describes as Looser zones are possibly Looser zones, and reiterated his view that the acromial fractures are in fact accessory ossification.
Dr Allgrove said he had never seen a child with HVA and had not considered it at an earlier stage in C’s case. He did not himself espouse Professor Nussey’s theory, but said that it was an interesting and not unreasonable one. He said that one would have to bear as in mind as a possibility. His only knowledge comes from literature relating to increased bone fragility in animals, with PR in association with vitamin A excess. He still felt that there was ‘something going on’ and the theory of a transient problem that got better would ‘fit quite nicely’. Maybe something else in C made him susceptible to vitamin A. However, there is no evidence positively suggesting excess vitamin A in C’s case.
He described the question of whether the PR resulted from the fractures or was associated with their cause as being a chicken and egg situation. He himself did not believe there could be so many episodes of injury without external injuries. One or two episodes might be different. Repeatedly during his evidence, Dr Allgrove returned to the extent of the PR and the symmetry of the fractures as being influential on his thinking.
Dr Allgrove commented on the biochemistry, which did not indicate malabsorption, and on aspects of the radiology, which did not for him show signs of poor mineralisation. However, there were some signs of hyperparathyroidism.
In relation to pain, Dr Allgrove confirmed that children with weak bones get pain on fracture, and that is likely that C was in pain when his bones were broken. The description by the paediatrician that he was ‘a picture of health’, suggests that the pain was non-specific and not necessarily apparent to a parent.
More generally, Dr Allgrove is unable to conceive of a description of C as ‘a picture of health’ days before his admission, if in fact he had been the subject of a number of deliberate severe assaults; if deliberately inflicted, this degree of periosteal reaction would not have happened without external signs.
In conclusion, Dr Allgrove said that the balance of probabilities in his mind was very much in favour of an underlying bone difficulty that for one reason or another resolved itself.
Evidence of Dr Ng
Dr Ng described her day-to-day clinical experience of premature babies and acute paediatrics and explained that her PhD had been in extreme paediatric endocrinopathies. She said that at 30 weeks gestation, C would have a very good prognosis. His vitamin D reading in June 2014 was insufficient at 38 nmol. He had one high ALP reading and low phosphate. Taken alone, these could be a marker of a mineralisation problem, but you have to look at the results as a whole, which are not abnormal. She cites recent advice given by the British Paediatric and Adolescent Bone Group that unless there is radiological and biochemical evidence of rickets, vitamin D is not relevant to the causation of fractures.
As to HVA, Dr Ng said that this theory was based on old case studies where bone mineralisation problems arose from massive doses given over prolonged periods. The supplements given to premature infants nowadays could not in her view have these side-effects. Many drugs that are in routine use by paediatricians are not licensed for infants, but are prescribed with the approval of the Royal College and are not associated with low levels of calcium and extensive fractures. She has looked after thousands of neonates and never come across supplements causing toxicity. And even if HVA was relevant to the PR, it is not associated with fractures, even where very high doses (up to 5000IU daily) are given to premature babies to counteract the risk of vitamin A deficiency.
She also challenged Professor Nussey’s reliance on the Meunier and Bosscher in vitro studies as being relevant to C’s case. Gaviscon and Ranitidine do not cause low calcium levels and extensive fractures in children. Dr Ng authored the Cochrane review entitled “Drug therapies for reducing gastric acid in cystic fibrosis.” None of the 39 trials reported low calcium in this population, who are prone to malabsorption.
If an inability to maintain vitamin D levels had led to hyperparathyroidism, one would see severe clinical signs of rickets in many bones.
Scurvy can be ruled out on the basis of the normal vitamin C reading of 65 nmol on admission, together with the absence of any gum disease or bleeding.
Copper deficiency can be ruled out on the basis of the normal result in July.
From her experience as a paediatrician involved in child protection cases, Dr Ng spoke of a number that involved multiple fractures with no external injury.
Evidence of Professor Bishop
Professor Bishop is the Professor of Paediatric Bone Disease at Sheffield and he leads the largest paediatric metabolic bone unit in Europe. He examined C in December 2014.
Professor Bishop explained that there have been continuous advances in our understanding of the genetics of metabolic bone disorder over the last 15 years. 15 new genes have been discovered, identifying pathways of control for bone cell activity, but these are persistent and not transient processes. In C’s case, he healed normally and his bones were straight after that.
C’s case is exceptional for the number of fractures and the extent of the PR, together with a virtually normal skeleton six months later. Where there is bone fragility, the condition mostly persists. He asked himself what biological process could explain these findings. If the number of fractures in this case was due to fragility, fragility does not simply go away. That would be a condition completely outside our knowledge.
If this was a metabolic condition, the puzzle is why it appears in some places and not others. There is periosteal reaction only where the ribs are fractured. It is remarkable that there are no changes to the spine.
He differs from Dr Ng in relation to the presence of mild osteopenia of prematurity.
He sees three or four cases of rickets a year. Here, there is no radiological or biochemical evidence for rickets. Rachitic change is primarily driven by a lack of phosphate (not by vitamin deficiency), but here C was maintained by multi-nutrients and his phosphate measurement was within the normal range (145 nmol).
Moreover, the PR in rickets is subtle and is not seen over the ends of the long bones. Here, the appearances of the periosteal membranes are thick. The number of fractures in the June x-rays was striking. You would have to have really severe rickets to cause fractures of this sort. Nor does one see metaphyseal fractures with rickets.
The medications C was given are not associated with fractures.
There are no known cases of copper deficiency in breastfed infants.
Scurvy is highly unlikely in a child receiving normal feed and supplements.
He fundamentally disagreed with Dr Ayoub on the interpretation of the x-rays.
He did not understand why there was PR on the scapulae and clavicles and could not identify a metabolic cause other than speculating that they had undergone trauma insufficient to cause fractures. He said that you have to look at all the bones and to go on looking for known causes.
The November x-rays were pretty normal. He described some irregularity of the distal metaphyses of the left forearm bones. He interpreted this to represent the healing of what he (wrongly, it is said on behalf of the parents) asserts to be the original metaphyseal fractures.
In relation to HVA, like the other witnesses, he had never come across it, except in the old literature. Here, C showed no signs, no hypocalcaemia and no clear evidence of vitamin D deficiency and its consequences. Vitamin A toxicity could relate to the PR but is not associated with fractures. Professor Nussey’s theory could be possible if we were simply debating excessive PR and metaphyseal changes, but ‘multiple fractures is not right’ and does not fit with the known facts.
The lack of bruising makes no difference either way.
Children with OI suffer pain, swelling and immobility when they experience fractures. It may be difficult for first-time parents to distinguish irritability arising from pain from the disturbance of a fracture site from other causes of pain and irritability in babies. (The surgeon, Mr J, gave similar evidence at the earlier hearing).
Professor Bishop concluded that C does not have any known disorder. After extensive testing, C’s biochemistry reflects his fractures. Instances of metabolic bone disease, presenting at this age with multiple fractures are evident both clinically and radiologically. There is no biochemistry or radiography to support metabolic bone disease of prematurity or inherited propensity to bone fragility. The metaphyseal and acromial fractures are typically reported in cases of inflicted injury. One can never be entirely certain, but there is no credible explanation for the appearances other than inflicted injury.
Evidence of Dr Halliday
Dr Halliday is a Consultant Paediatric Radiologist at Nottingham Children’s Hospital and is the Chair of the British Society of Paediatric Radiology. She described her own practice as involving the constant review of neonatal x-rays and the x-rays of children under 16. She said that children’s bones look very different to adult bones on x-ray. The February x-rays of C’s bones are completely normal for a baby of 30 weeks.
Dr Halliday maintained her identification of the fractures. She accepted certain variations between her findings and those of the reporting radiologists, Dr U and Dr L (see first judgment at [81, 87]), but did not consider these to be of any overall significance. She described the November x-rays as normal, disagreeing with Professor Bishop only to that extent.
Responding to Dr Ayoub, she said that his opinion was not in line with the mainstream. He does not look at paediatric x-rays day in day out, and he does not look at a normal population. She absolutely disagreed with him that there are signs of rickets in C’s case. There is no bowing, widening, cupping or flaying. Bone is being laid down where it would not be if rickets was present. Looser zones occur in adult rickets and have been found in the scapula, hips acromion. The appearances here do not look like Looser zones. The appearance of the scapula, described by Dr Ayoub as significantly bowed, is entirely normal.
She understands what Dr Ayoub means by slipped epiphyses, which can occur as insufficiency fractures in rickets but, if that is what is seen here, there are no other radiological signs of healing rickets.
Dr Halliday reaffirmed that the PR is very striking and that it is unusual to find it at the clavicles and scapulae, and on their interior surfaces. She has not seen this before. It is possible that twisting, rather than a fracture, could cause this, but if there is not a fracture (and she has never seen a fracture of the acromion) she cannot imagine the nature of the trauma. It is, as she put it, a dot we can’t join up. She agreed that inflicted injury was not a total explanation.
She did not think that C had hyperparathyroidism as this is typically seen in tunnelling in the phalanges, which is not present.
Copper deficiency might arise from very, very marked osteopenia, but would be generalised and associated with other clinical features.
She has never seen a case of scurvy and, based on the radiology, this is not one.
She has never seen a case of HVA. If C had it, she would not understand why. It is not associated with fractures. She has never seen anything like it in metabolic bone disease.
The parties’ submissions
Counsel have made extensive written submissions, supplemented by brief oral submissions. I am grateful to them for their very considerable assistance.
For the local authority, Mr Loveridge relies on the evidence of Dr Halliday, Dr Ng and Professor Bishop. He submits that the court can find that the fractures sustained by C were inflicted injuries caused by the application of excessive force to his limbs and torso on at least two occasions, and that the other radiological findings are consequent upon the fractures.
Mr Loveridge points to the fact that
C did not suffer perinatal fractures.
The two February x-rays do not show any abnormality in his bones.
All the fractures were sustained in the seven-week period between 2 May and 16 June 2014, a time spent entirely in the parents’ care.
If C had a metabolic bone disorder, it manifested itself only then.
Specifically, if he had a hypersensitivity to vitamin A, it did not display itself at any other time or in any other way.
There is no cogent evidential basis to support the hypothesis of vitamin A toxicity. A hypothesis that might raise doubt in the minds of a jury must be carefully analysed in these proceedings.
Mr Feehan QC and Ms Reaney for the father, and Mr Rowley QC and Ms Bannon for the mother, have helpfully made a joint submission. They argue that:
Any hypothesis must account for both the fractures and the PR, and their symmetry.
The only hypothesis that explains (i) the symmetrical raised PR; (ii) the raised PR in non-fractured bone; (iii) the noted osteopenia; (iv) the unusual bone texture; and (v) the lack of observable external symptoms; is that of Professor Nussey. The failure to account for all of the above is a real problem for those who assert abuse.
The existence of PR in apparently uninjured sites, notably the scapulae, shows that something must be wrong with the bone.
Transient conditions do exist – osteopenia of prematurity, rickets and Caffey’s.
Sensitivity to vitamin A may be temporary, due to the effects of maturation.
The court should be cautious before placing too much weight on the February x-rays. They did not include the lower limbs, which may have been fractured.
Further, the November x-rays, if abnormal, as suggested by Prof Bishop, show that there may have been further fractures at some time after admission.
The Guardian does not put a positive case on the causation of the injuries but Mr Baker has again presented a detailed analysis of the evidence.
Conclusions on the medical evidence
This remains a difficult and unusual case. As was said in Re U (above), particular caution is necessary where the medical experts disagree, one opinion declining to exclude a reasonable possibility of natural cause. I have therefore taken time to reflect on the medical evidence and the parties’ submissions. Long though it is, this judgment cannot reflect the full detail of the scientific information that has been provided, but I have sought to outline its main contours.
Having done so, I step back and consider the landscape as a whole. In doing so, I am struck by the essential similarity between the medical debate now and in 2015. The challenge for the court remains one of discriminating between two bodies of medical opinion in an area of some scientific refinement.
As ever, an interpretation of the medical evidence must work with the available information. Here, the biochemical and radiology investigations, mainly focus on three periods: February, June/July and November 2014. It would of course be valuable, though hardly realistic, to have had sequential data for the whole of C’s early months. I have therefore borne in mind the possibility that the available test results, broadly normal or only somewhat abnormal, might not be the whole story, and that had tests been carried out at other times, they might have disclosed grossly abnormal readings.
I now give my findings on the medical issues in the order they are described above.
The number of fractures
I accept the opinion of Dr Halliday. I prefer her interpretation of the radiology to that of Dr Ayoub because she has the necessary clinical experience, because her view is supported by that of the clinical specialist Dr L, because the radiological interpretations of Dr Ayoub have little or no support from other witnesses, and because I find him to be a witness with strong preconceived notions.
I also find no reason to depart from my previous finding in relation to the presence of metaphyseal fractures.
Again accepting Dr Halliday’s evidence, I do not accept Dr Ayoub’s identification of a Looser zone in the acromion, nor his characterisation of C’s scapula as being abnormally bowed. Nor do I accept that the February x-rays show diminished mineralisation of any significance or that the June x-rays show dramatic under-mineralisation in the hands, where there were in any event no fractures.
Vitamin D deficiency
It is common case that most of a new baby’s supply of vitamin D is acquired during the third trimester of pregnancy, and that a premature birth will interfere with this process. It is also known that vitamin D levels in the population of the northern hemisphere are reduced in winter. There is no doubt that throughout the relevant period, C’s vitamin D levels were lower than normal. The measurement of these levels is problematic and the debate during the evidence has referred to various definitions of ‘normal’, ‘insufficient’ and ‘deficient’. However, even those who favoured ‘deficiency’ cannot point to C’s vitamin D level being at a critically low level for any sustained period, such as to have a substantial effect on his bone mineralisation.
Some attention was paid at this hearing to the appearance of the third metacarpal in the right hand, and the question of whether this was evidence of hyperparathyroidism and vitamin D deficiency. There was a range of views on this appearance, and whether it had any significance. Taking the evidence at its highest, I find that this isolated finding offers only very slight support for the possibility of an overall metabolic pathology.
Rickets
The possibility that C suffered from healing rickets was fully considered in 2014/2015. Now, Dr Ayoub alone argues that the diagnosis should be made, along perhaps with copper and vitamin C deficiency. I do not accept this view. It is contradicted by all the other witnesses and has no sound radiological or biochemical foundation.
Scurvy
I remain satisfied that C did not suffer from scurvy. There is no biochemical or radiological support for this theory, and such assay results as there are speak against it. Given the vitamin C that was being administered to C, it is extremely improbable that he was deficient.
Copper deficiency
Based on the normal copper level result and the lack of clinical signs of this rare haematological condition, there is no support for Dr Ayoub’s opinion that this may be a feature of C’s condition.
Hypervitaminosis A (HVA)
Professor Nussey’s thesis that a cascade of events leading to vitamin A toxicity caused the condition of C’s bones, alone or in combination with other pathology, deserves and has received careful consideration. Mr Baker offers this formulation of the factors for and against:
For
The appearance of the PR is suggestive of a systemic difficulty.
HVA does come within the rare differential diagnosis of generalised PR.
C’s vitamin A consumption was above ‘normal’ at periods after his birth due to the supplementation.
If Dr Ayoub’s evidence is accepted, Professor Nussey’s view of the biochemistry and Dr Ayoub’s view of the radiology may be considered mutually supportive.
Rarity does not exclude possibility.
Against
None of the research papers referred to comes close in terms of dose of vitamin A, time of exposure, or age of the child.
None of the research papers referred to shows an association with fractures in humans.
HVA is not associated with fractures.
No other presenting symptoms are asserted as being consistent with excess vitamin A reaction.
It appears universally accepted that such a condition is very rare indeed.
My findings on this issue are these:
Such testing as was carried out was sufficient to allow a fair appreciation to be gained of C’s biochemical status at the time of his birth and his later admission. The available data shows a relatively mundane picture, indistinguishable from any number of newborn babies. This does not exclude the presence of HVA, but is nonetheless of significance.
The radiology in February was entirely in keeping with that of a normal premature baby. The possibility that C, who had no upper limb fractures, may at the same time have been carrying lower limb fractures outside the x-rays is so improbable that it can be discounted.
Similarly, the biochemistry and radiology following C’s admission to hospital in June, showed an essentially normal metabolism at work.
Insofar as Professor Bishop identifies a minor abnormality in the November x-rays (irregularity of the distal metaphyses of the left forearm bones), I prefer the evidence of Dr Halliday to the effect that these were normal. Professor Bishop is a specialist in metabolic bone disorder, not radiology.
It is an indispensable element of Professor Nussey’s theory that C should have had a unique, transient, and possibly genetic hypersensitivity to vitamin A, and that he should then have recovered from it, perhaps as a result of maturation. Though nothing is impossible in medicine, there is no dependable theoretical foundation for this conjecture and no primary evidence of a pathology which causes vitamin A toxicity with such devastating results, and then abates.
It is not an indispensable element of the theory that there should have been an interplay of polypharmacy arising from the anti-reflux medication, but I nonetheless find that this relatively common form of medication cannot have contributed significantly to the catastrophic condition of C’s bones. Dr Ng has particular experience and expertise in relation to premature babies, and I accept her evidence on this issue.
There were no clinical signs of HVA, such as a bulging fontanelle.
There is no indication that C’s vitamin D status was so compromised that it would have had a knock-on effect on his vitamin A levels.
HVA is only associated with extensive PR if vitamin A is administered in very high, prolonged doses. That was not remotely the case here.
HVA is not associated with fractures.
After careful consideration, my conclusion is that, with all respect to Professor Nussey, I not only have reservations about the theoretical basis for HVA as an explanation for C’s condition, but, more to the point, find that the theory does not have solid support in the actual evidence. On this issue, I prefer the opinions of Professor Bishop and Dr Ng to that of Professor Nussey and, to the limited extent that he contributes a view on this issue, to that of Dr Allgrove.
I should make clear that this conclusion is not based upon consideration of how rare it would be if C had had a condition of this kind. None of the experts had seen the condition before, but they were rightly prepared to consider it as a possible explanation. My rejection of the theory is not based upon it being unprecedented – I would have been glad to find that it had been established – but on the lack of primary evidence to support it, and on the evidence that speaks against it.
I therefore conclude that C’s presentation is not explained by him having suffered from hypervitaminosis A.
Undiagnosed metabolic bone disease
The possibility of an undiagnosed metabolic bone disease has been propounded by Dr Allgrove from the outset. It includes a known condition that has been missed or a condition that has as yet not been discovered. On the unusual facts of this case, these possibilities again deserve careful attention.
Inadequate calcium absorption is a known condition. Professor Nussey postulates that the use of gastric acid inhibitors may be relevant in C’s case. In this regard, it is in my view important to note the periods when these medications were prescribed:
Ranitidine (H2 receptor blocker) 14 April-6 May (3 weeks)
Domperidone (H2 receptor blocker) 25 April-16 June (8 weeks)
Omeprazole (Protein Pump Inhibitor) 10-16 June (1 week)
It was Prof Nussey’s evidence that PPI medication was more likely to affect calcium absorption than H2 receptor blockers, and that in this case Omeprazole (PPI) had been prescribed for too short a time to have any effect. I would add that its prescription also postdates many of the fractures.
None of the experts relies heavily on the Malchodi extract, which deals with older children and speaks of relatively marginal effects on fracture probability with dosages of this kind. Professor Nussey undertook some rough calculations, which indicated an effect of a few percentage points on incidence of fractures in older children. Dr Ayoub writes that “if the evidence of bone disease already presented does not influence state experts, this abstract will not likely do much to open their eyes.”
When considering undiscovered conditions, the court is in the realm of the unknown. Consideration must be given to the possibility of a new unknown condition, and also to a syndrome that results from a combination of known conditions whose interplay is not understood. As already stated, Dr Allgrove is struck by the extent and symmetrical distribution of the PR and the fractures, which leads him to conclude that there must have been some sort of underlying metabolic bone disease. It may be multifactorial, may involve vitamin A toxicity, and certainly involved vitamin D insufficiency; while one alone may not account for the presentation, they possibly provide an explanation in combination. As before, I understand the reason for Dr Allgrove’s perplexity, and I set his opinion beside the others.
Inflicted injury
The witnesses again considered the extremely extensive nature of the fractures and PR, their bilateral/symmetrical distribution, the significance of metaphyseal fractures, the absence of bruising and the question of pain reaction. Leaving aside Dr Ayoub, whose evidence I have rejected, the evidence on these issues on this occasion was largely indistinguishable from that given at the previous hearing.
The question of the apparently unfractured bones on which PR is seen again received attention. Particular consideration was given to the fact that this observation related not only to the clavicles but also to the scapulae. Although the scapulae were referred to in my first judgment at [112], discussion then focused more on the clavicles: see, for example, [122, 127(6) and 253(10)].
In their evidence on this occasion, Dr Halliday and Professor Bishop remained puzzled by these findings. In relation to the scapulae, Dr Halliday was not able to explain how the PR could have arisen.
To summarise
Having considered the medical evidence at this rehearing, these are my conclusions:
I am unable to accept that C’s condition resulted from an identified metabolic cause. Specifically, for the reasons given above, I find that hypervitaminosis A does not offer an explanation.
The other conditions that were ruled out in 2015 remain ruled out now. They are listed at [48] above. In particular, I confirm the finding that C did not suffer from rickets or from vitamin D deficiency of such significance that it could have set in train a demineralisation in his bones that might explain the medical picture.
The investigation has been exhaustive and I am satisfied that no other known condition has been overlooked. Having considered the possibility that testing at other times would have revealed radically different biochemical information, I reject it as overwhelmingly improbable
There are consequently only three realistic explanations for C’s condition: that it is the result of a pathology that is beyond our present knowledge or the result of force inflicted by an adult, or both.
In considering which of these alternatives is to be preferred, I remind myself of my previous findings, set out at [46] above.
In the light of the evidence at this hearing, the conclusion at [253(10)] needs to be expanded to include reference to the PR on the scapulae. This appreciation is not new, but it continues to represent a discrepancy in the hypothesis that C’s injuries were caused by inflicted force. It is a piece that does not neatly fit into that jigsaw or, from Dr Halliday’s perspective, a dot we cannot join up.
However, coming to my ultimate conclusion on the medical evidence, I consider that there are far greater difficulties in the way of a conclusion that the medical findings are the result, in whole or in part, of an unknown medical condition. None of the witnesses absolutely ruled out this possibility, but I am persuaded by the opinions of Dr Halliday, Dr Ng and Professor Bishop that C did not suffer from such a condition. There are five main reasons for this:
The condition would have to have had devastating effects on C’s metabolism without leaving any concomitant biochemical markers.
At the same time, it would leave him (on any view) with multiple fractures of a kind that were bound to be accompanied by significant PR: it is the extent of the PR, not its presence, that is so unusual.
It is particularly problematic to envisage a condition that could cause so many fractures with normal handling, but which has nonetheless escaped detection by medical science.
Such a condition would need to be transient in order to fit with the reliable medical information. In this regard, I found the evidence of Professor Bishop about the nature of transient conditions to be particularly helpful. He explained that we understand the reasons for transience in the few transient conditions that are known to exist. As he put it, fragility does not simply go away. The likelihood of there being an unknown transient condition of such severity that it caused appearances of this kind and then disappeared is extraordinarily unlikely, and cannot in my view be plausibly explained by normal maturation.
I accept that the lack of the PR on the scapulae and clavicles is a finding that tends to point the other way. However, one has to look at the evidence as a whole, and in a complex case of this kind it is not unusual for there to be some anomalies. In this case, the absence of a ready explanation for the anomaly does not in my view invalidate the hypothesis.
Lastly, there is no dispute among the domestic experts that the medical findings could be the result of inflicted force: what divided them sharply was the assessment of probability.
I therefore cannot accept the invitation of Mr Rowley to revise my findings in relation to subparagraphs [253(1) and (12)].
For completeness, I should say that the possibility of C’s condition being the result of both metabolic abnormality and inflicted force does not attract any real support in the evidence and I consider it to be no more than a theoretical possibility.
Having considered all the medical evidence, old and new, I find it more likely than not that C had normal bones and that he suffered inflicted injuries.
The other evidence
I have reviewed the non-medical findings made in 2015. There is no basis for revising them. Relevant paragraphs read as follows:
As to pain, at [261]:
… I find that an adult was undoubtedly in the closest proximity to him whenever the injuries occurred and that he will have cried in pain when his bones were fractured. The inability of either parent to describe a single occasion when he reacted in this way is highly suspicious. The only incidents that they have offered for consideration occurred in doctors’ surgeries.
As to C’s environment, at [267]:
… it is not difficult to envisage how one of the parents might have lost control for the first time, quite possibly in exasperation at his slowness or inability to feed, and how this led to a vicious circle in which C’s injuries made it harder for him to feed, and so on.
As to the parents’ credibility as witnesses, at [268]:
There are cases where a witness’s evidence is of such quality that it outweighs all other evidence and the court is able to accept it unconditionally. That is not the case here. The parents are no doubt credible narrators of normal life events. However, with regard to the specific issue of C’s injuries, I did not consider that I could rely on the evidence of either of them.
As to opportunity, at [273]:
I accordingly conclude that there was plainly the opportunity for C to suffer injury from one or both of his parents, despite the level of surveillance and the protective factors.
As to perpetrator, at [275]:
If it was possible to identify one parent as the perpetrator, that would be desirable for C’s sake, but unfortunately neither has been fully forthcoming. As matters stand, it is possible that either of them caused his injuries and at this stage I cannot say that one is more likely than the other to be responsible.
Outcome
Having considered all the evidence, I return to the essential question: Has the local authority proved that inflicted injury is more likely than not to be the cause of C’s condition? My answer is that it has done so. The findings on which the care order is based must stand and the mother’s application will accordingly be dismissed.