65
1981have found single-dose daily oral cyclophosphamide starting 2 mg/kg to be more effectiveand have no worse side-effects, than my previous use of 3 mg/kg daily azathioprine (each given with high single-dose, alternate dayl-3 prednisone, which raises circulating neutrophil counts and helps to depress lymphocyte counts). The cyclophosphamide dose is adjusted down or up according to the 2000/700 guidelines. Important safety adjuncts with oral cyclophosphamide include a weekly differentiated bloodcount and the patient drinking at least 36 litres of liquid daily. Because the opportunity for substantial benefit of oral immunosuppression in multiple sclerosis patients might have been missed, I suggest a trial with chronic (1-3 years) oral at
cyclophosphamide (plus high single-dose alternate-day prednisone) according to these indices. (The Canadian cooperative 22-week trial including 1-5-2 mg/kg oral cyclophosphamide" and another oral trials used the outmoded leucocyte count, and recent intravenous cyclophosphamide studies6,7have used only "white blood cell count".) University of Southern California, Neuromuscular Center, Los Angeles, California 90017, USA
W. KING ENGEL
Engel WK, Dalakas MC. Treatment of neuromuscular diseases. In. Wiederholt WC, ed. Therapy for neurologic disorders. New York: Wiley, 1982: 51-101. 2. Engel WK. Treating "untreatable" neuromuscular disease patients. Muscle Nerve 1
1986; 9: 95. DC, Engel WK. Remarkable
3. De Vivo
4.
recovery of
a
steroid
responsive
recurrent
polyneuropathy J Neurol Neurosurg Psychiatry 1970; 33: 62-69. The Canadian Cooperative Multiple Sclerosis Study Group. The Canadian cooperative tnal of cyclophosphamide and plasma exchange in progressive multiple
sclerosis Lancet 1991; 337: 441-46. 5. Khatn BO, McQuillen MD, Hoffmann RG, Harrington GJ, Schmoll D. Plasma exchange in chronic progressive multiple sclerosis: a long-term study. Neurology
1991, 41: 409-14. 6. Mackin GA, Weiner HL, Orav JA, et al. Cyclophosphamide/ACTH plus maintenance cyclophosphamide boosters in progressive MS: final report of the Northeast Cooperative MS Treatment Group I. Neurology 1991; 41: 147. 7. Likosky WH. Experience with cyclophosphamide in multiple sclerosis the cons. Neurology 1988; 38: 14-18
Diagnosis of urinary tract infection in children SIR,-I would point out some drawbacks to Dr Vickers and colleagues’ report (Sept 28, p 767). With their definition of urinary tract infection (UTI)-the presence of both a positive culture (> 105 colony-forming units/ml of one organism) and positive microscopy (8 organisms per high-power field or 100 organisms per grid)—they had 24 UTIs in 342 children. 13 other urine specimens had a pure growth of one organism that subsequently proved negative on repeat culture that was done because the initial microscopy was negative. There were therefore 13 false-positive cultures on the initial urine specimens in conjunction with true negative microscopy. What is not controlled for in this study is the possibility of false-positive microscopy in conjunction with a false-positive urine culture. If 13 of 37 specimens were known to be initial false positives on culture, then how do we know whether 13 of the remaining 24 were not also false-positive cultures but had enough bacteria to result in positive microscopy? Since urine specimens with this possibility were not repeated, the question cannot be conclusively answered by this study. I think Vickers and colleagues’ study warrants this question being looked at again, perhaps with a more reliable collection method such as suprapubic aspiration or urethral catheterisation in the young child, to reduce the possibility of false-positive culture and microscopy. Pediatrics, UAH Medical Clinics,
BOYCE A. HORNBERGER
Huntsville, Alabama 35801, USA
SlR,—We agree with Dr Sheppard and Dr Kelly (Nov 2, p 1144) that close collaboration between clinicians and microbiologists is essential for an accurate diagnosis of urinary tract infection (UTI). The first requirement is correct sampling and transportation, followed by careful microscopy and culture. Accurate diagnosis is
especially important
in
infancy.’
It is
important
babies who should be investigated for urinary
to
identify those anomalies,
tract
especially vesicoureteric reflux, without subjecting those who have false-positive tests to unnecessary procedures. Suprapubic aspiration or urine is regarded as the gold standard of diagnostic tests for UTI. Aronson and colleagues2 compared such sampling with simultaneously obtained clean-bag specimens in 120 patients, and noted 31 false-positive and 8 false-negative tests in samples from the bag. The recommendationto use ultrasonography to show urine in the bladder of newborn babies before suprapubic aspiration improved the success rate of first attempt from 36% to 100%. We have used the same technique in infants admitted for febrile illness. One of us (N. A.) has taught all paediatric residents (senior house officers) to use a ward-based ultrasound scanner to help with the suprapubic aspiration technique. Previously untrained staff find this easy to do, which increases the success rate and avoids upset to infants, parents, or staff by having to repeat the test and uncertainties of less accurate collection techniques. During four weeks in October and November, 1991, we obtained urine by suprapubic aspiration in 35 male and 26 female babies; 25 were newborn, and 36 were aged 1-16 months. 11 specimens (from 9 boys and 2 girls) yielded a positive culture. All isolates were coliforms; bacterial count was greater than 105/ml in 4 and less than 105/ml in 7. These findings confirm that low bacterial counts in urine specimens from children may be clinically significant and should not always be dismissed as contaminants. Dr Vickers and colleagues (Sept 14, p 674) misunderstood a previous paper on low bacterial counts in boys.’ Of the 73 boys investigated, 12 had counts of 104/ml on presentation to their general practitioners. 4 of them subsequently proved to have radiological abnormalities-the same percentage as boys who presented with counts of 105/ml. The association of low bacterial counts with radiological abnormalities has also been reported by Cohen.5 Department of Paediatrics and Child Health, and Public Health Laboratory, St Mary’s Hospital, Portsmouth P03 6AD, UK
NICOLA AUSTIN ROSALIND MASKELL RICHARD J. HALLETT
Working Group of the Royal College of Physicians. Guidelines of the management of acute urinary tract infection in childhood. J R Coll Phys Lond 1991; 25: 36 2. Aronson AS, Gustafen B, Svenningsen NW. Combined suprapubic aspiration and clean-voided urine examination in infants and children Acta Paediatr Scand 1973; 1.
62: 396-400. 3.
O’Callaghan C, McDougall PN. Successful suprapubic aspiration of urine. Arch Dis
Child 1987; 62: 1072-73. RJ, Paed L, Maskell R. Urinary infection in boys: a three year prospective study Lancet 1976; ii: 1107-10. 5. Cohen M. The first urinary tract infection in male children Am J Dis Child 1976, 130:
4. Hallett
810-13.
Neonatal bilirubin SiR,—Your Nov 16 editorial notes the lack of conclusive data on association between moderately increased neonatal bilirubin concentrations and inferior intelligence; our results support that view.1 However, your conclusion that current treatment policies "only achieve inefficient removal of a potentially helpful antioxidant" may be misleading. Current evidence on a beneficial role of bilirubin as an antioxidant is indirect and retrospectiveor has been obtained in vitro.3 Phototherapy is most effective in infants with substantial hyperbilirubinaemia but it works gradually, and once bilirubin concentrations exceed 375 jjmol/1 prolonged phototherapy treatment may be necessary. Advice suggesting that early phototherapy should be avoided may therefore place the infant at greater risk of extended exposure to potentially harmful levels of bilirubin and of prolonged hospital stay. An alternative strategy may soon be possible by using methods that identify, within a few hours of birth, infants at high risk of hyperbilirubinaemia. In the USA, the NICHHD Neonatal Research Network has launched a multicentre study to find out if hyperbilirubinaemia can be predicted by non-invasive measurement of respiratory carbon monoxide excretion.4 If we could estimate the risk of clinically important neonatal jaundice developing we could not only allow early discharge from hospital but also be at the first stage of a strategy of chemoprevention.57 In the meantime, we can resort to fibreoptic technology for delivery of high-intensity phototherapy. Fibreoptic an
66
devices not only provide more efficient phototherapy but may also allow maximum exposure of body surface areas without interfering with maternal-fetal interactions. Thus, concern about the negative effect of phototherapy on breast feeding can be negated and a convenient form of home phototherapy can be made practicable. Department of Pediatrics, Stanford University School of Medicine, Stanford, California 94305, USA
DANIEL S. SEIDMAN DAVID K. STEVENSON
DS, Paz I, Stevenson DK, Laor A, Danon YL, Gale R. Neonatal hyperbilirubinemia and physical cognitive performance at 17 years of age.
1. Seidman
Pediatrics 1991; 88: 828-33. 2. Benaron DA, Bowen FW Variation of initial serum bilirubin in newborn infants with type of illness. Lancet 1991; 338: 78-81 3 Stocker R, Yamamoto Y, McDonagh AF. Bilirubin is an antioxidant of possible physiologic importance. Science 1987; 235: 1043-46 4. Smith DW, Martin D, Inguillo D, Vreman HJ, Cohen RS, Stevenson DK. Use of noninvasive tests to predict significant jaundice in full-term infants: preliminary studies. Pediatrics 1985; 75: 278-80. 5 Kappas A, Drummond GS, Manola T, Petmezaki S, Valaes T. Sn-protoporphyrin use in the management of hyperbilirubinemia in term newborns with direct Coombs-positive ABO-incomparibility. Pediatrics 1988; 81: 485-97. 6. Vreman HJ, Rodgers PA, Stevenson DK. Zinc protoporphyrin administration for suppression of increased bilirubin production by latrogenic hemolysis in rhesus neonates J Pediatr 1990, 117: 292-97. 7 Johnson L, Dworanczyk R, Abbasi M, et al. Bilirubin oxidase (BOX) feeding significantly decrease serum bilirubin (B) levels in jaundiced infant Gunn rats. Pediatr Res 1988; 23: 412A.
Limb-reduction defects and chorionic villus sampling SiR,—Firth et all have described 5 infants with severe limbreduction defects among 289 pregnancies in which chorionic villus sampling (CVS) at between 56 and 66 days of gestation had been done. 4 infants also had micrognathia and microglossia. The occurrence of this combination of defects after CVS is of serious concern to
everybody doing prenatal diagnosis.
The population incidence figure cited by Firth et al as the context in which to judge these findings is based on our study.2 We found that 1 in 175 000 live births had the combination of limb-reduction defects with hypoglossia/micrognathia or hypoglossia-hypodactyly syndrome. However, this is a minimum estimate for this combination of defects because in registry data minor anomalies such as hypoglossia or micrognathia are underascertained. Registration of cases is in part from sources unskilled in picking up dysmorphic features or minor anomalies. However, it provides the only incidence figure for this anomaly; to the best of our knowledge there are no other consecutive live-birth population studies on such defects. In considering what categories of limb defect to compare with, other information needs to be taken into account. Instead of using only the very low incidence figure of a specific defect combination known to be underestimated it appears more appropriate to use also the figure for particular defects of the hand and fingers. This is because these almost always come to the attention of the registry and thus present an upper limit of frequency of anomalies to compare with the observation in cases after CVS. The incidence figures for defects of various levels of the hand and fingers are summarised in the table In interpreting and comparing incidence figures, the classification system of limb-reduction defects used to derive the INCIDENCE FIGURES OF LIMB-REDUCTION DEFECTS OFTHE HAND AND FINGERS
*Excludes familial cases or recognised syndromes, also excludes recognised amniotic band sequences tDoes not exclude familial cases or specific syndromes A case may have more than one reduction defect registered these incidence figures reflect incidence of defect, not individuals
numbers must be kept in mind. We used the classification system based on the suggestions of the American Society of Surgeons of the Hand and of the International Society of Prosthetics and Orthotics3 and extended it to a hierarchy of categorising limb defects with respect to prenatal timing of development (unpublished). If incidences for limb reduction defects were calculated simply on the basis of the ICD-9 coding for bone defects, these figures would differ (table). The complete data and details of the classification system used will be published elsewhere. Since CVS has been raised as the possible cause for very specific birth defects, we felt it useful to bring to attention these additional incidence figures from our study. Klinik fur Frauenheilkunde und Geburtshilfe, Medizinische Universitat zu Lubeck,
Lubeck, Germany
URSULA G. FROSTER
Department of Medical Genetics, University of British Columbia, Vancouver, British Columbia, Canada
P. A. BAIRD
HV, Boyd PA, Chamberlain P, MacKenzie IZ, Lindenbaum RH, Huson SM Severe limb abnormalities after chorion villus sampling at 56-66 days’ gestation. Lancet 1991; 337: 762-63. 2. Froster-Iskenius UG, Baird PA. Limb reduction defects in over one million consecutive livebirths. Teratology 1989; 39: 127-35. 3 Swanson AR. A classification for congenital limb malformations. J Hand Surg 1976, 1: 8-22. 1. Firth
among HIV-infected mothers SiR,—Three retrospective studies have revealed an unexpectedly high frequency of twins among HIV-infected children born in the USA.1-3 However, among six published prospective cohort studies only one, by our group in Brazzaville, Congo, has revealed a significantly greater number of twins among infants bom to HIV-infected mothers compared with controls born to HIV seronegative mothers.4 Because all HIV-infected women do not transmit the virus to their children during pregnancy, one focus of research has been to identify the determinants of perinatal
Twinning
transmission. Transmission may be related to the mother’s clinical immunological status5 but discordance in HIV-infectious status in twins6 indicates that other factors, related to the fetus or to the circumstances of delivery, may also have a role. Wiznia et al, in the Bronx (New York City), found a high frequency of twins among children born to seropositive mothers registered in foster-care programmes (7-9%, 6/80).1 Thomas et al made a similar observation in paediatric AIDS cases, also in New York (4-3%, 10/235)."In Los Angeles, Frederick and Mascola, using a community-based active surveillance for paediatric HIV infection, also reported a high rate of twins. In all these studies, the comparison was with the frequency of twins reported in the general population of the study’s catchment area. The main hypothesis proposedl-3 was that HIV-infected ..’omen may be more likely to have twin pregnancies than HIV-negative women because of a confounding factor such as drug use. Indeed, hormonal perturbations such as recent discontinuation of oral contraception have been hypothesised to induce twinning, and drug use might also be related to such hormonal changes. Nevertheless, many biases could account for the above findings. For example, twin pregnancies might receive more care than singleton pregnancies and thus have a higher probability of being screened for HIV. Similarly, twins are often born premature or small for age; as a result they stay longer in the hospital and HIV infection is more likely to be tested for or diagnosed than it is in non-twins. Furthermore, if the probability of being reported relies on the probability of being infected (with symptoms or not), and on the assumption of an independent probability of perinatal infection for each twin, the probability of HIV infection within a twin pair (at least one twin infected) is greater than the probability of infection for a non-twin. With a rate of transmission of 30%, the probability would be 51% (1 [Hl3]2) for twins instead of 30% among non-twins. Finally, the risk of transmission of HIV from mother to infant might be increased in twin pregnancies since twins are more likely to be born premature and prematurity is related to HIV infection. or
1981have found single-dose daily oral cyclophosphamide starting 2 mg/kg to be more effectiveand have no worse side-effects, than my previous use of 3 mg/kg daily azathioprine (each given with high single-dose, alternate dayl-3 prednisone, which raises circulating neutrophil counts and helps to depress lymphocyte counts). The cyclophosphamide dose is adjusted down or up according to the 2000/700 guidelines. Important safety adjuncts with oral cyclophosphamide include a weekly differentiated bloodcount and the patient drinking at least 36 litres of liquid daily. Because the opportunity for substantial benefit of oral immunosuppression in multiple sclerosis patients might have been missed, I suggest a trial with chronic (1-3 years) oral at
cyclophosphamide (plus high single-dose alternate-day prednisone) according to these indices. (The Canadian cooperative 22-week trial including 1-5-2 mg/kg oral cyclophosphamide" and another oral trials used the outmoded leucocyte count, and recent intravenous cyclophosphamide studies6,7have used only "white blood cell count".) University of Southern California, Neuromuscular Center, Los Angeles, California 90017, USA
W. KING ENGEL
Engel WK, Dalakas MC. Treatment of neuromuscular diseases. In. Wiederholt WC, ed. Therapy for neurologic disorders. New York: Wiley, 1982: 51-101. 2. Engel WK. Treating "untreatable" neuromuscular disease patients. Muscle Nerve 1
1986; 9: 95. DC, Engel WK. Remarkable
3. De Vivo
4.
recovery of
a
steroid
responsive
recurrent
polyneuropathy J Neurol Neurosurg Psychiatry 1970; 33: 62-69. The Canadian Cooperative Multiple Sclerosis Study Group. The Canadian cooperative tnal of cyclophosphamide and plasma exchange in progressive multiple
sclerosis Lancet 1991; 337: 441-46. 5. Khatn BO, McQuillen MD, Hoffmann RG, Harrington GJ, Schmoll D. Plasma exchange in chronic progressive multiple sclerosis: a long-term study. Neurology
1991, 41: 409-14. 6. Mackin GA, Weiner HL, Orav JA, et al. Cyclophosphamide/ACTH plus maintenance cyclophosphamide boosters in progressive MS: final report of the Northeast Cooperative MS Treatment Group I. Neurology 1991; 41: 147. 7. Likosky WH. Experience with cyclophosphamide in multiple sclerosis the cons. Neurology 1988; 38: 14-18
Diagnosis of urinary tract infection in children SIR,-I would point out some drawbacks to Dr Vickers and colleagues’ report (Sept 28, p 767). With their definition of urinary tract infection (UTI)-the presence of both a positive culture (> 105 colony-forming units/ml of one organism) and positive microscopy (8 organisms per high-power field or 100 organisms per grid)—they had 24 UTIs in 342 children. 13 other urine specimens had a pure growth of one organism that subsequently proved negative on repeat culture that was done because the initial microscopy was negative. There were therefore 13 false-positive cultures on the initial urine specimens in conjunction with true negative microscopy. What is not controlled for in this study is the possibility of false-positive microscopy in conjunction with a false-positive urine culture. If 13 of 37 specimens were known to be initial false positives on culture, then how do we know whether 13 of the remaining 24 were not also false-positive cultures but had enough bacteria to result in positive microscopy? Since urine specimens with this possibility were not repeated, the question cannot be conclusively answered by this study. I think Vickers and colleagues’ study warrants this question being looked at again, perhaps with a more reliable collection method such as suprapubic aspiration or urethral catheterisation in the young child, to reduce the possibility of false-positive culture and microscopy. Pediatrics, UAH Medical Clinics,
BOYCE A. HORNBERGER
Huntsville, Alabama 35801, USA
SlR,—We agree with Dr Sheppard and Dr Kelly (Nov 2, p 1144) that close collaboration between clinicians and microbiologists is essential for an accurate diagnosis of urinary tract infection (UTI). The first requirement is correct sampling and transportation, followed by careful microscopy and culture. Accurate diagnosis is
especially important
in
infancy.’
It is
important
babies who should be investigated for urinary
to
identify those anomalies,
tract
especially vesicoureteric reflux, without subjecting those who have false-positive tests to unnecessary procedures. Suprapubic aspiration or urine is regarded as the gold standard of diagnostic tests for UTI. Aronson and colleagues2 compared such sampling with simultaneously obtained clean-bag specimens in 120 patients, and noted 31 false-positive and 8 false-negative tests in samples from the bag. The recommendationto use ultrasonography to show urine in the bladder of newborn babies before suprapubic aspiration improved the success rate of first attempt from 36% to 100%. We have used the same technique in infants admitted for febrile illness. One of us (N. A.) has taught all paediatric residents (senior house officers) to use a ward-based ultrasound scanner to help with the suprapubic aspiration technique. Previously untrained staff find this easy to do, which increases the success rate and avoids upset to infants, parents, or staff by having to repeat the test and uncertainties of less accurate collection techniques. During four weeks in October and November, 1991, we obtained urine by suprapubic aspiration in 35 male and 26 female babies; 25 were newborn, and 36 were aged 1-16 months. 11 specimens (from 9 boys and 2 girls) yielded a positive culture. All isolates were coliforms; bacterial count was greater than 105/ml in 4 and less than 105/ml in 7. These findings confirm that low bacterial counts in urine specimens from children may be clinically significant and should not always be dismissed as contaminants. Dr Vickers and colleagues (Sept 14, p 674) misunderstood a previous paper on low bacterial counts in boys.’ Of the 73 boys investigated, 12 had counts of 104/ml on presentation to their general practitioners. 4 of them subsequently proved to have radiological abnormalities-the same percentage as boys who presented with counts of 105/ml. The association of low bacterial counts with radiological abnormalities has also been reported by Cohen.5 Department of Paediatrics and Child Health, and Public Health Laboratory, St Mary’s Hospital, Portsmouth P03 6AD, UK
NICOLA AUSTIN ROSALIND MASKELL RICHARD J. HALLETT
Working Group of the Royal College of Physicians. Guidelines of the management of acute urinary tract infection in childhood. J R Coll Phys Lond 1991; 25: 36 2. Aronson AS, Gustafen B, Svenningsen NW. Combined suprapubic aspiration and clean-voided urine examination in infants and children Acta Paediatr Scand 1973; 1.
62: 396-400. 3.
O’Callaghan C, McDougall PN. Successful suprapubic aspiration of urine. Arch Dis
Child 1987; 62: 1072-73. RJ, Paed L, Maskell R. Urinary infection in boys: a three year prospective study Lancet 1976; ii: 1107-10. 5. Cohen M. The first urinary tract infection in male children Am J Dis Child 1976, 130:
4. Hallett
810-13.
Neonatal bilirubin SiR,—Your Nov 16 editorial notes the lack of conclusive data on association between moderately increased neonatal bilirubin concentrations and inferior intelligence; our results support that view.1 However, your conclusion that current treatment policies "only achieve inefficient removal of a potentially helpful antioxidant" may be misleading. Current evidence on a beneficial role of bilirubin as an antioxidant is indirect and retrospectiveor has been obtained in vitro.3 Phototherapy is most effective in infants with substantial hyperbilirubinaemia but it works gradually, and once bilirubin concentrations exceed 375 jjmol/1 prolonged phototherapy treatment may be necessary. Advice suggesting that early phototherapy should be avoided may therefore place the infant at greater risk of extended exposure to potentially harmful levels of bilirubin and of prolonged hospital stay. An alternative strategy may soon be possible by using methods that identify, within a few hours of birth, infants at high risk of hyperbilirubinaemia. In the USA, the NICHHD Neonatal Research Network has launched a multicentre study to find out if hyperbilirubinaemia can be predicted by non-invasive measurement of respiratory carbon monoxide excretion.4 If we could estimate the risk of clinically important neonatal jaundice developing we could not only allow early discharge from hospital but also be at the first stage of a strategy of chemoprevention.57 In the meantime, we can resort to fibreoptic technology for delivery of high-intensity phototherapy. Fibreoptic an
66
devices not only provide more efficient phototherapy but may also allow maximum exposure of body surface areas without interfering with maternal-fetal interactions. Thus, concern about the negative effect of phototherapy on breast feeding can be negated and a convenient form of home phototherapy can be made practicable. Department of Pediatrics, Stanford University School of Medicine, Stanford, California 94305, USA
DANIEL S. SEIDMAN DAVID K. STEVENSON
DS, Paz I, Stevenson DK, Laor A, Danon YL, Gale R. Neonatal hyperbilirubinemia and physical cognitive performance at 17 years of age.
1. Seidman
Pediatrics 1991; 88: 828-33. 2. Benaron DA, Bowen FW Variation of initial serum bilirubin in newborn infants with type of illness. Lancet 1991; 338: 78-81 3 Stocker R, Yamamoto Y, McDonagh AF. Bilirubin is an antioxidant of possible physiologic importance. Science 1987; 235: 1043-46 4. Smith DW, Martin D, Inguillo D, Vreman HJ, Cohen RS, Stevenson DK. Use of noninvasive tests to predict significant jaundice in full-term infants: preliminary studies. Pediatrics 1985; 75: 278-80. 5 Kappas A, Drummond GS, Manola T, Petmezaki S, Valaes T. Sn-protoporphyrin use in the management of hyperbilirubinemia in term newborns with direct Coombs-positive ABO-incomparibility. Pediatrics 1988; 81: 485-97. 6. Vreman HJ, Rodgers PA, Stevenson DK. Zinc protoporphyrin administration for suppression of increased bilirubin production by latrogenic hemolysis in rhesus neonates J Pediatr 1990, 117: 292-97. 7 Johnson L, Dworanczyk R, Abbasi M, et al. Bilirubin oxidase (BOX) feeding significantly decrease serum bilirubin (B) levels in jaundiced infant Gunn rats. Pediatr Res 1988; 23: 412A.
Limb-reduction defects and chorionic villus sampling SiR,—Firth et all have described 5 infants with severe limbreduction defects among 289 pregnancies in which chorionic villus sampling (CVS) at between 56 and 66 days of gestation had been done. 4 infants also had micrognathia and microglossia. The occurrence of this combination of defects after CVS is of serious concern to
everybody doing prenatal diagnosis.
The population incidence figure cited by Firth et al as the context in which to judge these findings is based on our study.2 We found that 1 in 175 000 live births had the combination of limb-reduction defects with hypoglossia/micrognathia or hypoglossia-hypodactyly syndrome. However, this is a minimum estimate for this combination of defects because in registry data minor anomalies such as hypoglossia or micrognathia are underascertained. Registration of cases is in part from sources unskilled in picking up dysmorphic features or minor anomalies. However, it provides the only incidence figure for this anomaly; to the best of our knowledge there are no other consecutive live-birth population studies on such defects. In considering what categories of limb defect to compare with, other information needs to be taken into account. Instead of using only the very low incidence figure of a specific defect combination known to be underestimated it appears more appropriate to use also the figure for particular defects of the hand and fingers. This is because these almost always come to the attention of the registry and thus present an upper limit of frequency of anomalies to compare with the observation in cases after CVS. The incidence figures for defects of various levels of the hand and fingers are summarised in the table In interpreting and comparing incidence figures, the classification system of limb-reduction defects used to derive the INCIDENCE FIGURES OF LIMB-REDUCTION DEFECTS OFTHE HAND AND FINGERS
*Excludes familial cases or recognised syndromes, also excludes recognised amniotic band sequences tDoes not exclude familial cases or specific syndromes A case may have more than one reduction defect registered these incidence figures reflect incidence of defect, not individuals
numbers must be kept in mind. We used the classification system based on the suggestions of the American Society of Surgeons of the Hand and of the International Society of Prosthetics and Orthotics3 and extended it to a hierarchy of categorising limb defects with respect to prenatal timing of development (unpublished). If incidences for limb reduction defects were calculated simply on the basis of the ICD-9 coding for bone defects, these figures would differ (table). The complete data and details of the classification system used will be published elsewhere. Since CVS has been raised as the possible cause for very specific birth defects, we felt it useful to bring to attention these additional incidence figures from our study. Klinik fur Frauenheilkunde und Geburtshilfe, Medizinische Universitat zu Lubeck,
Lubeck, Germany
URSULA G. FROSTER
Department of Medical Genetics, University of British Columbia, Vancouver, British Columbia, Canada
P. A. BAIRD
HV, Boyd PA, Chamberlain P, MacKenzie IZ, Lindenbaum RH, Huson SM Severe limb abnormalities after chorion villus sampling at 56-66 days’ gestation. Lancet 1991; 337: 762-63. 2. Froster-Iskenius UG, Baird PA. Limb reduction defects in over one million consecutive livebirths. Teratology 1989; 39: 127-35. 3 Swanson AR. A classification for congenital limb malformations. J Hand Surg 1976, 1: 8-22. 1. Firth
among HIV-infected mothers SiR,—Three retrospective studies have revealed an unexpectedly high frequency of twins among HIV-infected children born in the USA.1-3 However, among six published prospective cohort studies only one, by our group in Brazzaville, Congo, has revealed a significantly greater number of twins among infants bom to HIV-infected mothers compared with controls born to HIV seronegative mothers.4 Because all HIV-infected women do not transmit the virus to their children during pregnancy, one focus of research has been to identify the determinants of perinatal
Twinning
transmission. Transmission may be related to the mother’s clinical immunological status5 but discordance in HIV-infectious status in twins6 indicates that other factors, related to the fetus or to the circumstances of delivery, may also have a role. Wiznia et al, in the Bronx (New York City), found a high frequency of twins among children born to seropositive mothers registered in foster-care programmes (7-9%, 6/80).1 Thomas et al made a similar observation in paediatric AIDS cases, also in New York (4-3%, 10/235)."In Los Angeles, Frederick and Mascola, using a community-based active surveillance for paediatric HIV infection, also reported a high rate of twins. In all these studies, the comparison was with the frequency of twins reported in the general population of the study’s catchment area. The main hypothesis proposedl-3 was that HIV-infected ..’omen may be more likely to have twin pregnancies than HIV-negative women because of a confounding factor such as drug use. Indeed, hormonal perturbations such as recent discontinuation of oral contraception have been hypothesised to induce twinning, and drug use might also be related to such hormonal changes. Nevertheless, many biases could account for the above findings. For example, twin pregnancies might receive more care than singleton pregnancies and thus have a higher probability of being screened for HIV. Similarly, twins are often born premature or small for age; as a result they stay longer in the hospital and HIV infection is more likely to be tested for or diagnosed than it is in non-twins. Furthermore, if the probability of being reported relies on the probability of being infected (with symptoms or not), and on the assumption of an independent probability of perinatal infection for each twin, the probability of HIV infection within a twin pair (at least one twin infected) is greater than the probability of infection for a non-twin. With a rate of transmission of 30%, the probability would be 51% (1 [Hl3]2) for twins instead of 30% among non-twins. Finally, the risk of transmission of HIV from mother to infant might be increased in twin pregnancies since twins are more likely to be born premature and prematurity is related to HIV infection. or
