Dear Editor,

We are glad to respond to the comments of Dr. George Morley [1] on our paper entitled “Multiorgan Dysfunction in Infants with Post- asphyxial Hypoxic Ischaemic Encephalopathy”.[2] Dr. Morley raised the possibility that the cause of the HIE in our patients was hypovolaemia due to deprivation of the placental transfusion (with or without tight nuchal cord) rather than intrapartum asphyxia. We will attempt to respond to most of the items in Dr. Morley’s two letters.

First, we wish to define “asphyxia”. Asphyxia is defined in the literature on either a pathophysiological or syndromal basis. The former consists of hypoxia, hypercapnia and metabolic acidaemia. The latter is defined in various consensus statements of national obstetrical and other professional societies, and in the so-called International Consensus Statement. [3] It is our strong impression that noone today accepts depression at birth or, for example a low 5-minute Apgar score as a definition of asphyxia. We wish to point out that Dr. Morley’s statement regarding the 5-minute Apgar score is not relevant to our study, as a score of <_5 was="was" not="not" an="an" essential="essential" inclusion="inclusion" criterion="criterion" it="it" one="one" of="of" three="three" sub-criteria.="sub-criteria." in="in" our="our" response="response" to="to" dr.="dr." morley="morley" the="the" unqualified="unqualified" term="term" asphyxia="asphyxia" indicates="indicates" either="either" pathophysiological="pathophysiological" or="or" syndromal="syndromal" severe="severe" means="means" both="both" and="and" mild="mild" moderate="moderate" implies="implies" only.="only." p="p"/> In his letter, Dr. Morley questions the plausibility of the diving reflex as a response to asphyxia. Knowledge of the fetal and neonatal diving reflex is based mainly on experiments in animals, and on clinical observations in the human newborn infant (we compare the two in more detail below). For animal data on the diving reflex in severe asphyxia, the interested reader is referred to the paper of Behrman et al3, and a detailed recent review by Jensen et al on fetal circulatory responses to hypoxia and asphyxia induced by various means.4 A considerable body of evidence suggests that in the human with hypoxia and asphyxia, the circulation is centralized to brain, heart and adrenal gland (the so-called “vital organs”).[2,4] Dr. Morley reiterated some of our criteria for organ involvement and pointed to some differences; owing to limitations of time we will not deal with our relatively minor differences of opinion. Dr. Morley points out that the diving reflex “does not last very long”. At some point in the progression of severe asphyxia in animals, the reflex breaks down. [4] When the diving reflex breaks down, the non-vital organs are likely to be more acidotic than the vital organs, so the latter have a relative advantage. It is reasonable to think that during the period that the reflex is active and for some period after, this compensatory mechanism provides relative protection to the vital organs.

Dr. Morley relates to repeated cord compression, the diving reflex and organ damage/ dysfunction. Before addressing this issue, we wish to point out that in contrast to “dysfunction”, organs other than brain and heart rarely sustain irreversible damage in severe asphyxia. We therefore prefer to use the term “dysfunction” to “damage”.

In animal models, repeated hypoxic episodes with intervening episodes of recovery or partial recovery (the prolonged partial asphyxial model of Myers[5]) was more strongly associated with multi-organ dysfunction than continued severe asphyxia with prolonged bradycardia (the “total” asphyxia animal model of Windle[6]).[4] Analogies have been drawn to human conditions,[7,8] but to our knowledge, no empirical evidence has been published that compares the multi-organ dysfunction after acute near- total vs. prolonged partial insults. To clarify this issue, we reviewed and analyzed our data. Since all of our patients with post-asphyxial hypoxic ischemic encephalopathy (HIE) had multiorgan dysfunction (MOD), the subgroups defined by the above time-courses of asphyxial insult were identical regarding the presence of MOD. We also determined from our data the proportion of infants with two or more dysfunctional organs in the two groups. Of infants in the acute near-total group, 70 percent had two or more dysfunctional organs, whereas 80 percent of infants in the prolonged partial asphyxia group had two or more dysfunctional organs. This difference was not statistically significant, nor do we consider it clinically significant.

Dr. Morley points to hypovolaemia as a pathophysiological mechanism that may supplement the adverse effects of co-existing asphyxia. He refers to two causes of hypovolaemia: cord compression as a cause of acute fetal blood loss into the placenta, and early clamping of the umbilical cord by depriving the newborn infant of the placental transfusion. We are glad that Dr. Morley draws attention to these phenomena; it is our impression that both are under-recognized as potential contributing causes of brain injury, especially when co-existing with asphyxia.

Nuchal cord, especially if tight, results in blood loss due to compression of the umbilical vein but not the arteries, as well as asphyxia due to failure of placental gas exchange, either or both varying from none to severe.[4,5] Haemorrhagic shock representing a threat to the brain occurs only in extreme cases. [6] In our study, we excluded infants with haemorrhagic shock. Infants thought to have “pure” haemorrhagic shock, which presents with relatively mild depression at birth and a progressive postnatal depression until blood volume is replenished, were excluded. Babies with tight nuchal cord and haemoglobin concentrations within normal limits were considered to have suffered “pure” asphyxia and were included in the study. The mean haemoglobin and haematocrits of our patients (including those with tight nuchal cord) were 161 g/l and 0.47, higher than the mean values for infants with tight nuchal cord studied by Cashore and Usher [4] (haematocrit 0.39), and similar to the values observed by Shepherd et al5 (mean haematocrit 0.48 and haemoglobin 164 g/L), and Linderkamp et al7 (haematocrit 0.46). It is notable that the infants studied by Cashore and Usher [4] (with an average blood loss of 20 percent of blood volume) showed only pallor and mild hypotension in the hours after birth; post-asphyxial HIE was not reported. Of the 27 cases of Shepherd et al5 with tight nuchal cord eligible for their study, [5] had haemoglobin values less than 132 g/L (which we consider to be well below the lower limit of normal) and [3] received emergency blood transfusion for pallor, tachycardia and hypotension. These authors also did not report HIE in their subjects. In our experience term infants can tolerate acute blood losses of up to 40 percent of blood volume and even more without developing obvious haemorrhagic shock, provided that they are free of preceding or concurrent asphyxia.

Early clamping of the umbilical cord with deprivation of the placental transfusion is also a well established cause of peri-partum hypovolaemia. Many published studies have compared early with delayed timing of cord clamping of apparently healthy infants born after uncomplicated pregnancy and parturition or Caesarean section. A number of studies of preterm infants have also been performed. In many studies, early clamping was accomplished within seconds of birth. Delayed clamping was defined in most studies as three minutes after birth or by cessation of cord pulsation. The effects of “delayed” and “early” cord clamping have been compared for a great variety of end-points including haematological, respiratory, cardiovascular, rheological and renal function measurements. Early clamping undoubtedly leaves some infants “with an uncomfortably small blood volume”.8 On the other hand, late clamping is associated with more jaundice [9] and hyperviscosity. [10] To our knowledge, a comprehensive health outcome has not been studied in relation to placental transfusion.

Studies of the effects of timing of cord clamping in severely asphyxiated infants and experimental animals have provided mixed results. According to Smith [11] “perinatal asphyxia may speed and increase the process of placental transfusion, even unto a prenatal transfusion.”[11] Hey [8] stated that “fatally asphyxiated babies usually have a relatively high haematocrit and circulating blood volume at birth.” The studies of Ackerman and colleagues [12,13] showed that infants with low Apgar scores and low umbilical arterial pH at birth tended to have small residual placental blood volumes, suggesting that a shift of blood volume from placenta to foetus had preceded cord clamping. These observations accord with those of Behrman et al in the experimental fetal primate. [14]

By contrast, acute intrapartum asphyxia defined as a 1- minute Apgar score of 5 or less was found to be associated with a reduced haematocrit of 0.49 (cord clamped at 15 seconds) as compared with 0.54 in infants with scores of 6 or greater (both were delivered vaginally). [7] This was considered to be caused by a shift of blood from fetus to placenta. In view of the early cord clamping at 15 seconds, another interpretation of this finding is possible. The infants with low 1-minute Apgar scores may represent the infants to which Dr. Morley refers. The depressed 1-minute Apgar scores and lower haematocrits may be attributable to early clamping of the cord and deprivation of a relatively large portion of the placental transfusion in some infants. As with the instances cited above from the literature, no mention is made of HIE in the 17 term infants studied. In addition, the definition of “acute intrapartum asphyxia” would not be acceptable on either pathophysiological or syndromal grounds.

The appropriate time to clamp the cord of babies with severe asphyxia has, to our knowledge, not been the subject of clinical trial. In a review of placental transfusion, Yao and Lind [15] stated: “In the presence of foetal distress or birth asphyxia, even during a vaginal delivery, it is also advisable not to delay clamping of the cord to avoid hypovolaemia except when there is obvious evidence of fetoplacental bleeding before and during birth and the infant appears pale and in shock. In such circumstances, cord clamping should be delayed if resuscitation can be given simultaneously. Otherwise, early separation of the infant from the cord will facilitate the resuscitation.”

Non-response to initial resuscitation measures suggests the possibility of hypovolaemia and the need for blood volume expansion. During placement of the umbilical venous catheter, central venous pressure can be measured; a low pressure is an additional indication for emergency blood volume expansion. Blood volume expansion is initially provided by colloid or crystalloid, soon followed by blood if indicated. [16] Group O rhesus negative whole blood should be given with great urgency if there is historical or physical evidence of a large acute blood loss (e.g., velamentous insertion of the cord with a tear in a large blood vessel; incision of the placenta prior to the delivery of the infant during the Caesarean section) and if the infant has signs of haemorrhagic shock (extreme pallor, weak peripheral pulses, hypotension, etc.).

The appropriate time to clamp the cord in an intermediate group of infants born after complicated deliveries with relatively mild depression has also not been the subject of clinical trial. We have seen babies born with mild to moderate asphyxia associated with tight nuchal cord, whose condition deteriorated after birth until the hypovolaemia was corrected, and who proved to be moderately severely anaemic. The severity of the anaemia caused by deprivation of blood volume in some cases of tight nuchal cord is far greater than the physiological placental transfusion would have been. A similar effect can be caused by gravity (holding the newly born infant well above the level of the placenta before clamping the cord).[17] Another potential cause of neonatal blood loss is the removal of cord blood for banking for possible future autologous cord blood transplantation (blood volumes as large as 110 ml, or about one- third of a 4 kg birth weight infant, have been banked). Hypovolaemia associated with mild to moderate intrapartum asphyxia, if not speedily recognised and treated by blood volume expansion, may cause neonatal morbidity.

We did not find the statement attributed by Dr. Morley to Peltonen18 that “In the compromised neonate, immediate cord clamping may result in fatality” was not found on review of the cited article. Peltonen [18] did allude to the “highly unphysiological” effects of early clamping on the neonatal heart during the first three or four cardiac cycles after cord clamping and before the first breath. He also referred to the animal study of Born et al [19] in which clamping of the cord before the onset of pulmonary respiration caused “profound asphyxia”. Despite these observations, Peltonen [18] concluded his review of early cord clamping with the citation that “It appears that umbilical cord clamping should be done at the discretion of the obstetrician.”

To our knowledge no studies have been conducted that compared measures of neonatal well being including morbidity and mortality in early-clamped with late-clamped term infants born after uncomplicated deliveries. This is not surprising in light of the large sample size that would be required for such a trial, and ethical hesitations about depriving some infants of large blood volumes, and causing polycythemia in others. In light of the observed adverse associations with very early and with delayed clamping, avoiding these extremes would appear to be sensible. Likewise, delay in resuscitation of a severely asphyxiated infant to await a placental transfusion that is unlikely to occur owing to constricted umbilical arteries is likely to cause more harm than benefit.

One circumstance that may justify an attempt to resuscitate the term infant with an intact foetal circulation may be after delivery complicated by tight nuchal cord in which it proved possible to deliver the infant without ligating and cutting the cord. Even in this situation, one can speculate that delay in clamping the cord may be harmful. For example, re-establishing the umbilical-foetal placental circulation after a period of stasis may cause placental thromboembolism and consequent harm such as neonatal stroke.

Randomized controlled trials comparing different timings of placental transfusion may be justified for some populations. One example referred to above is mild to moderate asphyxia where it is likely that the infant is in primary rather than secondary apnoea. Another is the newly born small for gestational age infants (presumably the “intra-uterine asphyxia” or placental insufficiency group of Linderkamp [7]). On the other hand, in our opinion it would be far more difficult to persuade a research ethics committee today to approve a comparative trial of delayed clamping (with attendant suboptimal conditions for resuscitation) versus immediate clamping (to optimize the resuscitation conditions) in severely asphyxiated infants who are most likely in secondary apnoea. Likewise, it would be difficult to justify a comparison of measures of neonatal well being, morbidity and mortality in very early-clamped with very late- clamped apparently healthy term infants, especially as so much is known about the effects of deprivation or maximization of placental transfusion in these infants. Whatever happens to the depressed infant at birth with regard to cord-clamping, the team progressing through the steps of resuscitation should not overlook the possibility that blood volume expansion may urgently be indicated. With regard to sampling of blood gases Dr. Morley refers to the “later withdrawn” ACOG Practice Bulletin 138 which suggested that obstetricians clamp the cord quickly in order to send samples for blood gases. In the process they “amputate a functioning placenta”. He advises “the prudent obstetrician” to “approach the pulsating cord with due caution”. We agree with this advice with emphasis on “a pulsating cord”. Parenthetically, repeated research has indicated that delayed cord clamping is advantageous for the adaptation of the preterm infant to extra-uterine life.[20,21] The reason often given in the past for early clamping of preterm infants was the perceived threat of the increased levels of jaundice associated with late clamping in the pre-phototherapy era. This is of much less concern today than it was when phototherapy was unavailable and before it was found to be efficacious and safe. Late clamping of preterm infants is probably safer than early clamping.

We agree that for the infants who are more likely than not to benefit from delayed cord clamping, the technical difficulty that may be caused by the delay in sampling from the cord artery may be relatively unimportant, and the potential enhancement of the circulating blood volume is the greater priority. To complete the diagnosis (pathophysiological or syndromal) of asphyxia, and if necessary its treatment, a sample can be taken from the infant rather than the cord. This has the advantage of avoiding errors such as sampling from the umbilical vein instead of an artery. Moreover, even umbilical arterial samples after severe cord compression are occasionally unrepresentative of the infant’s condition at birth, and capillary sampling is likely to reflect the infant’s current condition more accurately.

Dr. Morley also addressed the issue of hypoxic versus ischemic aetiopathogenesis of brain lesions. Considerable ambiguity exists between the terms “hypoxia” and “ischaemia”. From the aetiological standpoint (aetiology meaning “root cause”) we interpret Dr. Morley’s statements to differentiate between the hypoxia of asphyxia, and the ischaemia of hypovolaemia (in the context of his letter, hypovolaemia is caused by acute blood loss due to deprivation of the placental transfusion, associated or not with early cord clamping or tight nuchal cord). From the aetiological standpoint, we prefer the term “haemorrhagic shock” to describe the latter condition. Haemorrhagic shock causes tissue hypoxia, the end-results of which on brain and other organs are very similar to those of severe asphyxia (incidentally, one intermediate result that asphyxia and hypovolaemic shock have in common is the centralization of the blood circulation).

On the other hand, severe asphyxia is often associated with hypotension, the pathogenesis (mechanism as opposed to aetiology) of which is vasodilatation. Whatever the pathogenesis of ischaemia in severely asphyxiated infants (eg, cerebral vasoconstriction due to break-down in the protective diving reflex, cardiogenic shock, or hypovolaemic shock of the various causes considered above), it is considered to be the more important pathogenetic mechanism of brain injury than the hypoxia that is an essential part of the definition of asphyxia (with mild to moderate asphyxia, the fetus is unlikely to develop ischaemia and to sustain brain injury). The complexity of these overlapping conditions is nicely illustrated by the example of extreme cases of tight nuchal cord in which severe asphyxia and haemorrhagic shock coexist. To our knowledge brain MRI cannot differentiate between the end-results of severe asphyxia with secondary brain ischaemia, and haemorrhagic shock. As indicated above, we wish to compare the primate fetus with severe asphyxia studied mainly between the 1960s and the 1970s as a model for brain injuries associated with severe fetal asphyxia in the human. Dr. Morley’s statement that “the degree of asphyxia in the human HIE neonate is minor compared to that required to damage a monkey’s brain” is surprising. The relative immaturity of the term human fetus and the larger stores of glycogen compared with the monkey fetus would be expected to provide the human with greater tolerance to severe asphyxia. This appears to be supported by empirical data. Perusal of recently published case series that report durations of sustained bradycardia and outcomes in a variety of near-total asphyxia scenarios shows that the human fetus with sustained bradycardia seems to have greater staying power than the experimental fetal term monkey with total asphyxia and sustained bradycardia. For example, in a report of the short-term outcomes of the babies of 97 women with uterine rupture (82 of whom had a trial of labour); it was observed that “significant neonatal morbidity” occurred only in those cases in which “18 or more minutes elapsed between the onset of the prolonged deceleration and delivery”. This was the case even in cases that had prior severe late or variable decelerations. Of the 13 babies with sustained decelerations for 18 or more minutes without prior severe late or variable decelerations, only one (with 32 minutes of deceleration) had a severe neonatal outcome, viz, HIE (the long-term outcome was not reported). [22] Review of published case series of severe total or near-total asphyxia provides a similar perspective. For example, Cases 3 and 4 of Pasternak and Gorey [23] had uterine rupture and known durations of bradycardia (18 and 15 minutes); their outcomes were relatively moderate (respectively, mild spasticity and dystonia with normal cognition and head circumference at 3.5 years of age, and mild transient hypotonia and normal cognition and head circumference at 4 years of age). Of particular interest is Case 9 with umbilical cord rupture (an acute total asphyxia; blood loss was not mentioned); sustained bradycardia was present for 31 minutes; this infant had mild transient hypotonia with normal cognition and head circumference at 18 months. [23] These durations of bradycardia are beyond the duration of bradycardia seen in experimental total asphyxia that proved lethal in term rhesus monkeys.

We summarize by stating where we agree and disagree with Dr. Morley. We disagree that “Most, if not all neonates in this study had the placenta amputated at the moment of birth together with a large volume of blood.” In our opinion, it is likely that the umbilical cord had lost its function in the great majority of our cases by the time the cord was clamped; these were cases of severe asphyxia. In our opinion, maintaining the umbilical cord connection was unlikely to provide a placental transfusion. “Amputation” of a non-functioning placenta with poor potential for recovery of function is appropriate. Based on the eligibility and exclusion criteria of our study, and on the haemoglobin and haematocrit values reported here, we are confident that immediate cord clamping did not cause significant blood losses in the great majority of our cases.

On the other hand we emphasize that we agree with Dr. Morley that early clamping of the pulsating cord of an infant may cause harmful blood losses and postnatal haemorrhagic shock, particularly in infants who were mildly to moderately asphyxiated at birth with a shift of blood volume from fetus to placenta. Similarly, placement of an infant above the level of the placenta without clamping the still-functioning cord, or “transplanting” cord blood in large volume may cause harmful blood losses. Dr. Morley has drawn attention to aspects of neonatal care that are sometimes overlooked in contemporary practice. We hope that this correspondence will stimulate clinicians to review the older literature on placental transfusion, cord clamping, and haemorrhagic shock. Where questions still remain to be answered, there is a place for new clinical and even animal studies to elucidate the role of sequestration of blood in the placenta in the aetiopathogenesis of maladaptation at birth, and of neonatal brain illness and injury.

References

(1). Morley GM. Hypovolemia:The Cause of Multiorgan Dysfunction. 2004. BMJ Publishing Group. http://fn.bmjjournals.com/cgi/eletters/89/2/F152#407

(2). Shah P, Riphagen S, Beyene J, Perlman M. Multiorgan dysfunction in infants with post-asphyxial hypoxic-ischaemic encephalopathy. Arch Dis Child Fetal Neonatal Ed 2004;89:F152-F155.

(3). MacLennan A. A template for defining a causal relation between acute intrapartum events and cerebral palsy: international consensus statement. BMJ 1999;319:1054-9.

(4). Cashore WJ, Usher RH. Hypovolemia resulting from tight nuchal cord at birth. Pediatr Res 1973;7:339.

(5). Shepherd AJ, Richardson J, Brown JP. Nuchal cord as a cause of neonatal anemia. Am J Dis Child 1985;139:71-3.

(6). Vanhaesebrouck P, Vanneste K, De Praeter C, Van Trappen Y, Thery M. Tight nuchal cord and hypovolaemic shock. Arch Dis Child 1987;62:62-3.

(7). Linderkamp O, Versmold HT, Messow-Zahn K, Muller-Holve W, Riegel KP, Betke K. The effect of intra-partum and intra-uterine asphyxia on placental transfusion in premature and full-term infants. Eur J Pediatr 1978;127:91-9.

(8). Hey E. Resuscitation at birth. Br J Anaesth 1977;49:25- 33.

(9). Saigal S, O'Neill A, Surainder Y, Chua LB, Usher R. Placental transfusion and hyperbilirubinemia in the premature. Pediatrics 1972;49:406-19.

(10). Linderkamp O, Nelle M, Kraus M, Zilow EP. The effect of early and late cord-clamping on blood viscosity and other hemorheological parameters in full-term neonates. Acta Paediatr 1992;81:745-50.

(11). Smith CA, Nelson NM. The physiology of newborn infant. In: Thomas CC, 4th ed. 1976:157-77.

(12). Chou PJ, Ackerman BD. Perinatal acidosis and placental transfusion. Acta Paediatr Scand 1973;62:417-21.

(13). Flod NE, Ackerman BD. Perinatal asphyxia and residual placental blood volume. Acta Paediatr Scand 1971;60:433-6.

Dear Editor,

We have read with interest Forsyth et al’s article in your journal [1]. They have concluded that “angiotensin converting enyzme (ACE) is increasingly identified as a risk factor for cardiovascular disease, and serum activity in infancy may contribute to the link between low birth weight and later cardiovascular events.”

We think, in this manner that it is inappropriate to come to a conclusion by means of Forsyth et al’s study methods since none of the cardiologic parameters of any healthy term infants were assessed in the study. Though emphasizing the ACE being important in regulation of peripheral blood pressure, arterial pressure measurements on infants were done neither at birth, nor 1st and 3rd months. Nevertheless, some influent questions mentioned below remained unanswered:

1.How many infants of high ACE level had any cardiologic problems?
2.How many infants were assessed echocardiographically?
3.Are all of the infants of lower ACE levels healthy cardiologically?

The authors stressed on their conclusion that especially “later” cardiovascular events had a relation with ACE level. However “later” cardiovascular events could not be foreseen by following up the infants for 3 months.

Overall, mentioning the relation would be impossible. There is a need for controlled studies having longitudinal extended case numbers and well- monitorized cardiologic parameters.

Reference

(1). Forsyth JS, Reilly J, Fraser CG, Struthers AD. Angiotensin converting enyzme activity in infancy is related to birth weight. Arch Dis Child Fetal Neonatal Ed 2004;89:F442-F444.

Dear Editor,

The primary immunisation schedule in the UK is about to change for the seventh time since the accelerated schedule was introduced in 1990. The latest change, to a combined DTPaHibIPV vaccine (Pediacel), provoked a flurry of lay concern.

Given the rapid advancement of vaccine technology that has prompted such frequent changes to our primary immunisation policy it is inevitable that we would have lessons to learn along the way. Amongst the important ones are that not all populations respond to vaccines alike, (preterm populations are different from those born at term [1], and ethnicity matters [2]), interactions occur if vaccines are combined [3], and ‘minor’ changes may have ‘major’ consequences (DTPaHib combinations result in far poorer Hib protection than DTPwHib combinations [4]).

It is sometimes impossible to determine precisely which immunisations have been given to a study population, and how they were administered – for example whether as separate or combined immunisations. Whilst we read with great interest further data from Clarke and Robinson [5] on the effect of steroid treatment on preterm infants’ immunisation responses several important points are illustrated. The twelve infants reported share identical demographic characteristics with (and are presumably the same infants as) 12 infants previously reported in this journal after a fourth dose of conjugate Haemophilus influenzae type b (Hib) vaccine (PRP- T), where the implication was that a booster dose of PRP-T in such infants would not increase Hib protection [6]. In the original paper no mention is made of the fact that the extra PRP-T dose was actually administered with DTPw, a fact only deducible with care 18 months later, yet one which may alter the interpretation of the original paper. Likewise, the historical data from both Ramsay [7] and Booy [8] predate the introduction of Hib and Men C into the schedule. Comparison of the post primary immunisation GMTs presented by Clarke [5] (immunised with the Trivax AD DTP vaccine) with those of a more contemporaneous group of preterm steroid treated infants (immunised with the Pasteur Meriuex DTPHib vaccine [9]) shows that only the 95% confidence intervals for FIM and FHA overlap:

Vaccine Study

As the schedule changes yet again may we suggest that future publications ensure clear and explicit identification by name, manufacturer and method of administration of all vaccines administered to study subjects, whether or not the antibody response to a particular component is the major focus of the paper. Such information would clearly aid interpretation of the data.

References

(1). Slack MH, Schapira D, Thwaites RJ, Burrage M, Southern J, Andrews N, et al. Immune response of premature infants to meningococcal serogroup C and combined diphtheria-tetanus toxoids-acellular pertussis-Haemophilus influenzae type b conjugate vaccines. Journal of Infectious Diseases 2001;184(12):1617-20.

(2). Ward J, Brenneman G, Letson G, Heyward W, and the Alaskan Haemophilus influenzae group. Limited efficacy of a Haemophilus influenzae type b conjugate vaccine in Alaskan native infants. New England Journal of Medicine 1990;323:1393-1401.

(3). Bell F, Martin A, Blondeau C, Thornton C, Chaplais J and Finn A. (1996) Combined diphtheria, tetanus, pertussis, and Haemophilus influenzae type b vaccines for primary immunisation. Archives of Disease in Childhood;75(4):298-303.

(4). Eskola J, Olander RM, Hovi T, Litmanen L, Peltola S, Kayhty H. Randomised trial of the effect of co-administration with acellular pertussis DTP vaccine on immunogenicity of Haemophilus influenzae type b conjugate vaccine. Lancet 1996;348(9043):1688-92.

(5). Clarke P, Robinson M. DTP immunisation of steroid treated preterm infants. Archives of Disease in Childhood Fetal & Neonatal Edition 2004;89:F468-470.

(6). Clarke P, Powell P, Goldblatt D, Robinson M. Effect of a fourth Haemophilus influenzae type b immunisation in preterm infants who received dexamethasone for chronic lung disease. Archives of Disease in Childhood 2003;88(S1):58-61.

(7). Ramsay ME, Miller E, Ashworth LA, Coleman TJ, Rush M, Waight PA. Adverse events and antibody response to accelerated immunisation in term and preterm infants. Archives of Disease in Childhood 1995;72(3):230-2.

(8). Booy R, Aitken SJ, Taylor S, Tudor-Williams G, Macfarlane JA, Moxon ER, et al. Immunogenicity of combined diphtheria, tetanus, and pertussis vaccine given at 2, 3, and 4 months versus 3, 5, and 9 months of age. Lancet 1992;339(8792):507-10.

(9). Robinson M, Heal C, Gardener E, Powell P, Sims D. Antibody response to diphtheria-tetanus-pertussis immunisation in preterm infants who receive dexamethasone for chronic lung disease. Pediatrics 2004;113(4):733-737.

Dear Editor

We are grateful to colleagues for their comments on our annotation.[1] We would stress that we merely abstracted the views of others so any criticisms (apart from our brevity) will be of the lawyers, doctors, nurses, physiotherapists and parents who contributed to the Royal Commission Report. We found it to be systematic, rational and objective.

We strongly refute any suggestion that any of the New Zealand professionals should be criticised let alone made scapegoats (witness our final paragraph). We are puzzled that Drs Rosenbloom and Ryan discount the quoted witness statements of the parents and involved clinicians. The lawyers and doctors are clear that the physiotherapy and nursing practices did occur and that the levels of head shaking were not monitored.

We are concerned with infant brain injuries not lung disease and consider this to be topical. We share colleagues’ concern at the need to base opinions on speculative presumption extrapolated from animal or accident research and are aware of the limited evidence that identifies the minimal forces needed to cause shaken brain damage in neonates or older infants. We found the reported experiences to be a helpful insight.

We are delighted that Dr Rushton (eLetter) has taken this opportunity to state he thought vigorous chest physiotherapy without supporting the head was responsible for the porencephalic lesions and to inform of his pivotal involvement in advising New Zealand colleagues. We understand there were earlier concerns that publishing the speculation about physiotherapy would open liability to litigation. Lawyers might consider the inference that fear of litigation led to suppression of information that might have prevented the New Zealand deaths and the dilemma facing clinicians who reported the cerebral implications of vigorous physiotherapy. Dr Knight reports (eLetter) their unit has been ‘subject to a long official public inquiry, law suits and had twenty medical, nursing and physiotherapy staff investigated by registration authorities, lasting 8 years.’

We do not accept criticisms of inaccurate references. The Cochrane review we both cited was last updated in 1997. There has been an updated review this year (dealing with lung not brain disease), which was unavailable to the editors or us at the time of submission. Dr Knight (eLetter) states there was no change in the vigour of chest physiotherapy from 1985 until the end of 1994 but he co-authored the paper [2] we cited that states that there was no policy to support the head during chest physiotherapy and no data on the extent the head moved during physiotherapy whether given by nurses or physiotherapists. The Royal Commission Report found no record of the vigour of chest percussion and understood there was considerable variation with no standardisation of training.

We recommend interested colleagues to read this Report and the publications of Knight and Harding et al before dismissing the possibility that vigorous chest physiotherapy without supporting the head may cause brain injuries in certain circumstances.

References

(1) Williams AN, Sunderland R, Neonatal shaken baby syndrome: an aetiological view from Down Under. Arch Dis Child 2002;87:F29-30

(2) Knight DB, Bevan CJ, Harding JE, et al. Chest physiotherapy and porencephalic lesions in very preterm infants. J Paediatr Child Health 2001;37:554–8.