- Email: [email protected]
Congenital Heart Disease Infant Death Rates Decrease as Gestational Age Advances from 34 to 40 Weeks
Congenital Heart Disease Infant Death Rates Decrease as Gestational Age Advances from 34 to 40 Weeks James F. Cnota, MD, Resmi Gupta, MS, Erik C. Michelfelder, MD, and Richard F. Ittenbach, PhD Objectives To describe congenital heart disease death rates in infants born between 34 and 40 weeks, estimate the relationship between gestational age and congenital heart disease infant death rates, and compare congenital heart disease death rates across 1- and 2-week intervals in gestational age. Study design The 2000 to 2003 national linked birth/infant death cohort datasets were obtained. Congenital heart disease deaths were identified by using International Statistical Classification of Diseases, 10th Revision codes. Proportional death rates were calculated by using congenital heart disease deaths and all live births. The relationship between congenital heart disease death rates and gestational age was determined. Death rates were compared across intervals. Results A total of 14.9 million records were analyzed. Congenital heart disease deaths occurred in 4736 infants (0.04%) born between 34 and 40 weeks. There was a significant, negative linear relationship between congenital heart disease death rate and gestational age (R2 = 0.97). Comparisons across 1-week intervals varied (P = .02-.23). All 2-week intervals were statistically significant (P < .01). Conclusions Congenital heart disease death rates decrease as gestational age approaches 40 weeks. These results should be considered before elective delivery for the sole indication of prenatally diagnosed congenital heart disease. (J Pediatr 2011;159:761-5).
G
estational age, when expressed as either premature or not premature, has been found to predict outcome after surgery for congenital heart disease.1,2 In this dichotomous form, gestational age has been included in both the Risk Adjustment for Congenital Heart Surgery (RACHS-1)3 and Aristotle4 risk assessment models. This ‘‘patient specific’’ factor assigns increased risk to premature infants who are born before 37 weeks gestation. After 37 weeks, all increments of gestational age are considered equivalent. In contrast, population-based data for all live births suggest that morbidity and mortality rates continue to decline until 40 weeks gestation.5 Even a 1-week interval appears to improve outcome. To date, few studies report the significance of gestational age in late preterm (34-36 weeks) and term ($37 weeks) infants with congenital heart disease.6 Such data could significantly impact risk assessment in the newborn with congenital heart disease. A revised perspective of gestational age in the late preterm and term infant may affect current perinatal management strategies after the prenatal diagnosis of critical fetal heart disease. Prenatal diagnosis allows the family and medical team to make more informed decisions during pregnancy, at the time of delivery, and in the immediate newborn period.7,8 For families referred to a tertiary care center, a transfer of obstetric care and elective delivery is often recommended when neonatal intervention is anticipated.9 Such a plan ensures experienced personnel in the delivery room, prompt initiation of treatment, and availability of interventional catheterization,10 surgical support, or both as needed. However, prenatal diagnosis has also been associated with increased cesarean delivery rates,11 lower birth weights,12 and lower gestational age at delivery.13 Whether these factors, specifically the modest decrease in gestational age, affect the anticipated benefit of prenatal diagnosis remains unclear. Because of the lack of population based data on the relationship between gestational age and congenital heart disease, the purpose of this study was to: (1) describe congenital heart disease death rates in infants born between 34 and 40 weeks; (2) estimate the relationship between gestational age and congenital heart disease infant death rates; and (3) compare congenital heart disease death rates across 1- and 2-week intervals in gestational age. We hypothesized that gestational age and congenital heart disease infant death rates would have a negative, statistically significant relationship between 34 and 40 weeks gestation.
Methods The 2000 to 2003 national linked birth/infant death cohort datasets were obtained as public use data files from the National Center for Health Statistics (NCHS). The datasets link death certificates for infants <1 year of age with their respective birth certificates. In 2003, 99% of all infant death records were successfully matched to their corresponding birth certificate.14 Additionally, the files included data from all birth certificates within the defined period for the specific purpose of calculating From the The Heart Institute (J.C., R.G., E.M., R.I.) and Division of Biostatistics/Epidemiology (R.G., R.I.), rates. These datasets were chosen for analysis because they broadly represent conCincinnati Children’s Hospital Medical Center, Cincinnati, OH
ICD-10 NCHS
International Statistical Classification of Diseases, 10th Revision National Center for Health Statistics
The authors declare no conflicts of interest. 0022-3476/$ - see front matter. Copyright ª 2011 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2011.04.020
761
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 159, No. 5
Table I. Death rate by gestational age in weeks with untransformed and log-transformed scales for the period 2000 to 2003 Untransformed scale Gestational age (weeks) 34 35 36 37 38 39 40
Transformed scale
Mean (SE)
95% CI
0.1244 (0.0116) 0.0926 (0.0096) 0.0716 (0.0020) 0.0571 (0.0059) 0.0399 (0.0028) 0.0289 (0.0013) 0.0235 (0.0007)
0.08-0.16 0.06-0.12 0.06-0.07 0.03-0.07 0.03-0.04 0.02-0.03 0.02-0.02
Mean (SE) 2.0938 (0.0959) 2.3932 (0.1149) 2.6292 (0.0276) 2.8712 (0.1028) 3.2253 (0.0702) 3.5439 (0.0462) 3.7484 (0.0304)
95% CI 2.392.752.713.193.443.693.84-
1.78 2.02 2.54 2.54 3.00 3.39 3.65
Untransformed death rates represent the number of deaths per 100 live births for a given gestational age. Untransformed CI values were rounded to two decimal places for consistency in reporting.
genital heart disease in the United States by including both surgical and non-surgical congenital heart disease. Additionally, they have the added benefit of capturing mortality outcomes from inpatient, emergency department, and out-ofhospital locations. Infant deaths were attributed to congenital malformations of the cardiovascular system based on International Statistical Classification of Diseases, 10th Revision (ICD-10) codes used for underlying cause-of-death on the death certificate.15 In accordance with World Health Organization regulations for classification of death, ICD-10 codes have been used by NCHS for cause-of-death statistics since 1999, allowing for the consistent application of diagnostic codes across the study period. Proportional death rates were calculated by using congenital heart disease deaths as the numerator and all infant births as the denominator. Gestational age subgroups were defined by using the standard NCHS algorithm for gestational age.16 The primary determinant of gestational age was the interval between the first day of the mother’s last menstrual period and the date of birth. When that value is clearly inconsistent with the infant’s plurality and birth weight, the result is edited by NCHS and replaced by a clinical estimate of gestational age, which is determined by examination of the newborn. In 2003, gestational age was determined by clinical estimate in 4.6% of births. The date of last menstrual period was not reported in 97% of those cases. Only 3% of editing was required for the inconsistencies aforementioned. The level of editing required in 2003 was similar to earlier years.17 Statistical Analyses Data analysis proceeded in two distinct phases: a descriptive phase and an inferential phase. Descriptive statistics for death rates from all congenital heart disease causes from 2000 to 2003 were computed to describe the basic features of the data (Table I). A quantilequantile plot was used to determine whether infant death rates were normally distributed and amenable to analysis with traditional parametric techniques. On the basis of the observed plot, log transformations were then used as a means of normalizing the death rate distribution for subsequent modeling. 762
A trend analysis with orthogonal contrasts was used to estimate the linear, quadratic, or higher level polynomial effects. A regression-based analysis of variance model was then used as a means of assessing the relationship between infant death rates from congenital heart disease and gestational age at birth from 34 to 40 weeks, after adjusting for birth year. Pre-specified subgroup comparisons to the statistically significant omnibus F-test were computed with the TukeyKramer test, a derivation of the Tukey honestly significant difference technique, selected specifically for its ability to maintain a consistent type I error-rate across all post-hoc comparisons. The experiment wise error rate was set at the nominal (a = 0.05) level. All analyses were conducted with SAS software version 9.1.3 (SAS Institute Inc, Cary, North Carolina).
Results Of the 14.9 million birth records analyzed for 2000 to 2003, 12.5 million (84%) occurred between 34 and 40 weeks gestational age. Congenital heart disease deaths occurred in 4736 infants (0.04%) born between 34 and 40 weeks. These deaths
Figure 1. Proportion of congenital heart disease deaths by gestational age in weeks. CHD, congenital heart disease. Cnota et al
ORIGINAL ARTICLES
November 2011 represent 75% of all infant congenital heart disease deaths across all gestational ages. Within the study population of infants born 34 to 40 weeks gestational age, most congenital heart disease deaths occurred in infants born at full term, gestational age 37 to 40 weeks (Figure 1). The study consisted of 51% male and 49% female infants born at 34 to 40 weeks gestational age. Approximately 79% of the infants were white, 15% were black, and the remaining 6% consisted of other races. Proportional infant death rates attributed to congenital heart disease are presented in Table I. The untransformed death rates represent the number of deaths per 100 live births for a given gestational age. Untransformed death rates range from two in 10 000 at 40 weeks gestational age to 12 in 10 000 at 34 weeks gestational age. Log-transformed death rates are also presented because those rates were used specifically to normalize data for the analysis. When assessing the trend across gestational age groups, only the linear component of orthogonal polynomial contrast coefficients was statistically significant (P < .01), suggesting that there was a statistically significant linear trend in the means of congenital heart disease death rate, but no higher order trend. Results of the analysis of variance test (Figure 2) indicated that for infants born at 34 to 40 weeks gestational age, estimates of congenital heart disease death rates decreased with each 1-week increase in gestational age (R2 = 0.97). Follow-up analysis to the omnibus test was performed by using the Tukey-Kramer post hoc technique. Whereas the difference in congenital heart disease death rates across 1-week intervals (ie, 34 weeks versus 35 weeks) ranged in significance (P = .02-.23, Table II), all 2-week intervals (ie, 38 versus 40 weeks) were markedly significant (P < .01). To assess whether the overall observations were also seen in high-risk subgroups, we performed two additional analyses. First, only deaths that occurred during the first 28 days of life were considered. This neonatal population demonstrated the same negative linear relationship with gestational age that
Figure 2. Linear relationship between congenital heart disease death rates and gestational age. Error bars represent the variability of annual death rate across the years 2000 to 2003.
Table II. Post hoc analysis of congenital heart disease death rates in a 1-week interval in gestational age (log scale) Comparison (weeks) 34 versus 35 35 versus 36 36 versus 37 37 versus 38 38 versus 39 39 versus 40
Estimate 0.29 0.23 0.24 0.35 0.31 0.20
P value .02 .11 .10 <.01 .01 .23
All P values were modified with the Tukey-Kramer adjustment technique. All 2-week interval differences in death rate were statistically significant (P < .01).
was observed in the overall congenital heart disease population (R2 = 0.95, P < .01). Second, only deaths with a primary diagnosis of hypoplastic left heart syndrome were considered. This diagnosis was the single most common cause of infant death from congenital heart disease and was the only individual diagnosis with a sufficient number of deaths to allow NCHS reporting of diagnosis-specific rates. Death rates attributed to hypoplastic left heart syndrome demonstrated the same negative linear relationship with gestational age (R2 = 0.75, P < .01) as the overall congenital heart disease population.
Discussion This study demonstrated a significant, negative linear relationship between congenital heart disease death rates and gestational age between 34 and 40 weeks. Furthermore, the data suggest that even an additional week of gestation is associated with lower congenital heart disease death rates at some gestational age intervals. An additional 2 weeks of gestation is highly statistically significant across all intervals examined. An explanation for why some gestational age intervals are statistically significant and others are not remains unclear. Perhaps developmental milestones unique to a given week of gestation can have relatively greater or lesser effect on mortality rates. These findings appear to parallel observations in the general population, in which the public health significance of such findings has already been appreciated. Although the highest risk of morbidity and mortality has been attributed to infants born at gestational age <34 weeks, recent findings have demonstrated increased morbidity and mortality rates in the late preterm infant born 34 to 36 weeks.18 Rates of pulmonary complications19 and infant mortality continue to decline to a nadir at 39 to 40 weeks. Such findings are particularly concerning because there has been a 25% increase in late preterm deliveries since 199020 and an overall decrease in the mean gestational age at delivery. Complementing the population-based methods of this study, a recent large single center study6 suggested that infants with major congenital heart disease may be particularly vulnerable to the medical complications associated with gestational age <39 weeks. Respiratory insufficiency superimposed on the complex cardiopulmonary interaction of neonatal congenital heart disease results in poor outcome.21,22 Additionally, recent findings suggest that the brain maturation in infants with
Congenital Heart Disease Infant Death Rates Decrease as Gestational Age Advances from 34 to 40 Weeks
763
THE JOURNAL OF PEDIATRICS
www.jpeds.com
complex congenital heart disease may be delayed by as long as 4 weeks.22,23 In effect, many of our ‘‘term’’ infants with congenital heart disease may have neurological findings and risks similar to premature infants. As we continue to learn about the neurodevelopmental consequences of congenital heart disease, even small differences in gestational age may turn out to have clinical importance. In the current era, a large portion of patients with a complicated single ventricle are diagnosed with fetal echocardiogram—as many as 75% of patients with hypoplastic left heart syndrome.13 Those babies appear to be delivered earlier than their undiagnosed counterparts. And, although the overall difference in gestational age between infants with prenatal and postnatal diagnosis may appear relatively modest, those born after labor induction may deliver as much as 1.5 weeks earlier.11 The results of this study suggest that such differences can have a significant effect on postnatal outcomes. Therefore, although labor and delivery management must be considered in the context of each maternal-fetal dyad, these data support the recommendations from the American College of Obstetricians and Gynecologists Practice Bulletin on Induction of Labor. Even in the presence of prenatally diagnosed congenital heart disease, scheduled delivery before 39 weeks gestation should be avoided.24 Because of the known delayed brain maturation and potential for continued fetal growth, further research should address the question of whether elective delivery for the sole indication of fetal congenital heart disease is appropriate at any gestational age. The use of federal datasets, including those derived from birth certificates and death certificates, introduces limitations. Estimates of gestational age are derived from imperfect criteria.25 NCHS does apply a diagnostic algorithm that is well described16 and reduces the likelihood of significantly erroneous assignment of gestational age. Primary cause of death relies on individual practicing physicians to complete death certificates. Data quality is improved by using standardized logic algorithms designed to arrive at the most likely cause of death.26 Congenital heart disease incidence is not available because the time course and clinical presentation of congenital heart disease does not readily allow for documentation of defects on the birth certificate. Such incidence data are necessary to calculate true death rates (deaths of infants with congenital heart disease/births of infants with congenital heart disease); therefore these data are only able to describe proportional death rates. Although earlier studies have indicated that infants with congenital heart disease are not born at significantly different gestational age overall,27 diagnosis-specific differences in incidence, gestational age, and death rates could contribute to the findings. Finally, birth certificate data do not allow for reliable assessment of prenatal diagnosis; therefore we are unable to determine how this important variable influences outcome. These findings provide a broad perspective on the relationship between gestational age and congenital heart disease. They highlight the idea that perinatal factors impacting the general population can be important, or are even more important, in a smaller population of high-risk infants with 764
Vol. 159, No. 5 congenital heart disease. Although the relatively simple model has limited ability to address confounding factors that could influence the observed relationship, the association is very strong and the decision to analyze gestational age as a discrete variable allows for direct translation of weekly increments to the clinical setting. Further work is necessary to understand how characteristics of the mother, pregnancy and fetus influence gestational age in this population. Although these results are similar to observations from the general population, supporting the validity of the findings, they also extend the relationship to a unique, high-risk population. An increase of even 1-week in gestational age may result in reduced deaths in these late preterm and term infants. These findings challenge the common approach of scheduling elective delivery before 40 weeks after prenatal diagnosis of complex congenital heart disease. n Submitted for publication Oct 5, 2010; last revision received Mar 28, 2011; accepted Apr 18, 2011. Reprint requests: James Cnota, MD, The Heart Institute, Division of Cardiology, Cincinnati Children’s Hospital Medical Center, MLC 2003, 3333 Burnet Ave, Cincinnati, OH 45229-3039. E-mail: [email protected]
References 1. Andrews RE, Simpson JM, Sharland GK, Sullivan ID, Yates RW. Outcome after preterm delivery of infants antenatally diagnosed with congenital heart disease. J Pediatr 2006;148:213-6. 2. Dees E, Lin H, Cotton RB, Graham TP, Dodd DA. Outcome of preterm infants with congenital heart disease. J Pediatr 2000;137:653-9. 3. Jenkins KJ, Gauvreau K, Newburger JW, Spray TL, Moller JH, Iezzoni LI. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg 2002;123:110-8. 4. Lacour-Gayet F, Clarke D, Jacobs J, Comas J, Daebritz S, Daenen W, et al. The Aristotle score: a complexity-adjusted method to evaluate surgical results. Eur J Cardiothorac Surg 2004;25:911-24. 5. Mathews TJ, MacDorman MF. Infant mortality statistics from the 2005 period linked birth/infant death data set. Natl Vital Stat Rep 2008;57:1-32. 6. Costello JM, Polito A, Brown DW, McElrath TF, Graham DA, Thiagarajan RR, et al. Birth before 39 weeks’ gestation is associated with worse outcomes in neonates with heart disease. Pediatrics 2010;126:277-84. 7. Allan L, Dangel J, Fesslova V, Marek J, Mellander M, Oberhansli I, et al. Recommendations for the practice of fetal cardiology in Europe. Cardiol Young 2004;14:109-14. 8. Rychik J, Ayres N, Cuneo B, Gotteiner N, Hornberger L, Spevak PJ, et al. American Society of Echocardiography guidelines and standards for performance of the fetal echocardiogram. J Am Soc Echocardiogr 2004;17:803-10. 9. Cohen MS, Frommelt MA. Does fetal diagnosis make a difference? Clin Perinatol 2005;32:877-90, viii. 10. Bonnet D, Coltri A, Butera G, Fermont L, Le Bidois J, Kachaner J, et al. Detection of transposition of the great arteries in fetuses reduces neonatal morbidity and mortality. Circulation 1999;99:916-8. 11. Raboisson MJ, Samson C, Ducreux C, Rudigoz RC, Gaucherand P, Bouvagnet P, et al. Impact of prenatal diagnosis of transposition of the great arteries on obstetric and early postnatal management. Eur J Obstet Gynecol Reprod Biol 2009;142:18-22. 12. Bartlett JM, Wypij D, Bellinger DC, Rappaport LA, Heffner LJ, Jonas RA, et al. Effect of prenatal diagnosis on outcomes in D-transposition of the great arteries. Pediatrics 2004;113:e335-40. 13. Atz AM, Travison TG, Williams IA, Pearson GD, Mahle WT, Sklansky MS, et al. Impact of prenatal diagnosis and risk factors for preoperative death in neonates with single right ventricle: screening results
Cnota et al
November 2011
14.
15.
16. 17. 18.
19.
20.
from the Pediatric Heart Network Single Ventricle Reconstruction Trial. J Am Coll Cardiol 2009;53:A355-66. Mathews TJ, MacDorman MF. Infant mortality statistics from the 2003 period linked birth/infant death data set. Natl Vital Stat Rep 2006;54: 1-29. National Center for Health Statistics. United States Linked Birth/Infant Death Period Public Use Files, Years 2000, 2001, 2002, and 2003. In: Service USPH, editor. Hyattsville, Maryland: National Center for Health Statistics; 2002, 2003, 2004, and 2005. Martin JA. United States vital statistics and the measurement of gestational age. Paediatr Perinat Epidemiol 2007;21(Suppl. 2):13-21. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2003. Natl Vital Stat Rep 2005;54:1-116. Davidoff MJ, Dias T, Damus K, Russell R, Bettegowda VR, Dolan S, et al. Changes in the gestational age distribution among US singleton births: impact on rates of late preterm birth, 1992 to 2002. Semin Perinatol 2006;30:8-15. Morrison JJ, Rennie JM, Milton PJ. Neonatal respiratory morbidity and mode of delivery at term: influence of timing of elective caesarean section. Br J Obstet Gynaecol 1995;102:101-6. MacDorman MF, Menacker F, Declercq E. Cesarean birth in the United States: epidemiology, trends, and outcomes. Clin Perinatol 2008;35:293-307, v.
ORIGINAL ARTICLES 21. McMahon CJ, Penny DJ, Nelson DP, Ades AM, Al Maskary S, Speer M, et al. Preterm infants with congenital heart disease and bronchopulmonary dysplasia: postoperative course and outcome after cardiac surgery. Pediatrics 2005;116:423-30. 22. Miller SP, McQuillen PS, Hamrick S, Xu D, Glidden DV, Charlton N, et al. Abnormal brain development in newborns with congenital heart disease. N Engl J Med 2007;357:1928-38. 23. Licht DJ, Shera DM, Clancy RR, Wernovsky G, Montenegro LM, Nicolson SC, et al. Brain maturation is delayed in infants with complex congenital heart defects. J Thorac Cardiovasc Surg 2009;137:529-36, discussion 36-37. 24. ACOG Practice Bulletin No. 107: induction of labor. Obstet Gynecol 2009;114:386-97. 10.1097/AOG.0b013e3181b48ef5. 25. Wier ML, Pearl M, Kharrazi M. Gestational age estimation on United States livebirth certificates: a historical overview. Paediatr Perinat Epidemiol 2007;21:4-12. 26. National Center for Health Statistics. Vital statistics, instructions for classifying the underlying cause of death. NCHS instruction manual; part 2a. Hyattsville, MD: Public Health Service. Published annually. 27. Reller MD, Strickland MJ, Riehle-Colarusso T, Mahle WT, Correa A. Prevalence of congenital heart defects in metropolitan Atlanta, 19982005. J Pediatr 2008;153:807-13.
Congenital Heart Disease Infant Death Rates Decrease as Gestational Age Advances from 34 to 40 Weeks
765
G
estational age, when expressed as either premature or not premature, has been found to predict outcome after surgery for congenital heart disease.1,2 In this dichotomous form, gestational age has been included in both the Risk Adjustment for Congenital Heart Surgery (RACHS-1)3 and Aristotle4 risk assessment models. This ‘‘patient specific’’ factor assigns increased risk to premature infants who are born before 37 weeks gestation. After 37 weeks, all increments of gestational age are considered equivalent. In contrast, population-based data for all live births suggest that morbidity and mortality rates continue to decline until 40 weeks gestation.5 Even a 1-week interval appears to improve outcome. To date, few studies report the significance of gestational age in late preterm (34-36 weeks) and term ($37 weeks) infants with congenital heart disease.6 Such data could significantly impact risk assessment in the newborn with congenital heart disease. A revised perspective of gestational age in the late preterm and term infant may affect current perinatal management strategies after the prenatal diagnosis of critical fetal heart disease. Prenatal diagnosis allows the family and medical team to make more informed decisions during pregnancy, at the time of delivery, and in the immediate newborn period.7,8 For families referred to a tertiary care center, a transfer of obstetric care and elective delivery is often recommended when neonatal intervention is anticipated.9 Such a plan ensures experienced personnel in the delivery room, prompt initiation of treatment, and availability of interventional catheterization,10 surgical support, or both as needed. However, prenatal diagnosis has also been associated with increased cesarean delivery rates,11 lower birth weights,12 and lower gestational age at delivery.13 Whether these factors, specifically the modest decrease in gestational age, affect the anticipated benefit of prenatal diagnosis remains unclear. Because of the lack of population based data on the relationship between gestational age and congenital heart disease, the purpose of this study was to: (1) describe congenital heart disease death rates in infants born between 34 and 40 weeks; (2) estimate the relationship between gestational age and congenital heart disease infant death rates; and (3) compare congenital heart disease death rates across 1- and 2-week intervals in gestational age. We hypothesized that gestational age and congenital heart disease infant death rates would have a negative, statistically significant relationship between 34 and 40 weeks gestation.
Methods The 2000 to 2003 national linked birth/infant death cohort datasets were obtained as public use data files from the National Center for Health Statistics (NCHS). The datasets link death certificates for infants <1 year of age with their respective birth certificates. In 2003, 99% of all infant death records were successfully matched to their corresponding birth certificate.14 Additionally, the files included data from all birth certificates within the defined period for the specific purpose of calculating From the The Heart Institute (J.C., R.G., E.M., R.I.) and Division of Biostatistics/Epidemiology (R.G., R.I.), rates. These datasets were chosen for analysis because they broadly represent conCincinnati Children’s Hospital Medical Center, Cincinnati, OH
ICD-10 NCHS
International Statistical Classification of Diseases, 10th Revision National Center for Health Statistics
The authors declare no conflicts of interest. 0022-3476/$ - see front matter. Copyright ª 2011 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2011.04.020
761
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 159, No. 5
Table I. Death rate by gestational age in weeks with untransformed and log-transformed scales for the period 2000 to 2003 Untransformed scale Gestational age (weeks) 34 35 36 37 38 39 40
Transformed scale
Mean (SE)
95% CI
0.1244 (0.0116) 0.0926 (0.0096) 0.0716 (0.0020) 0.0571 (0.0059) 0.0399 (0.0028) 0.0289 (0.0013) 0.0235 (0.0007)
0.08-0.16 0.06-0.12 0.06-0.07 0.03-0.07 0.03-0.04 0.02-0.03 0.02-0.02
Mean (SE) 2.0938 (0.0959) 2.3932 (0.1149) 2.6292 (0.0276) 2.8712 (0.1028) 3.2253 (0.0702) 3.5439 (0.0462) 3.7484 (0.0304)
95% CI 2.392.752.713.193.443.693.84-
1.78 2.02 2.54 2.54 3.00 3.39 3.65
Untransformed death rates represent the number of deaths per 100 live births for a given gestational age. Untransformed CI values were rounded to two decimal places for consistency in reporting.
genital heart disease in the United States by including both surgical and non-surgical congenital heart disease. Additionally, they have the added benefit of capturing mortality outcomes from inpatient, emergency department, and out-ofhospital locations. Infant deaths were attributed to congenital malformations of the cardiovascular system based on International Statistical Classification of Diseases, 10th Revision (ICD-10) codes used for underlying cause-of-death on the death certificate.15 In accordance with World Health Organization regulations for classification of death, ICD-10 codes have been used by NCHS for cause-of-death statistics since 1999, allowing for the consistent application of diagnostic codes across the study period. Proportional death rates were calculated by using congenital heart disease deaths as the numerator and all infant births as the denominator. Gestational age subgroups were defined by using the standard NCHS algorithm for gestational age.16 The primary determinant of gestational age was the interval between the first day of the mother’s last menstrual period and the date of birth. When that value is clearly inconsistent with the infant’s plurality and birth weight, the result is edited by NCHS and replaced by a clinical estimate of gestational age, which is determined by examination of the newborn. In 2003, gestational age was determined by clinical estimate in 4.6% of births. The date of last menstrual period was not reported in 97% of those cases. Only 3% of editing was required for the inconsistencies aforementioned. The level of editing required in 2003 was similar to earlier years.17 Statistical Analyses Data analysis proceeded in two distinct phases: a descriptive phase and an inferential phase. Descriptive statistics for death rates from all congenital heart disease causes from 2000 to 2003 were computed to describe the basic features of the data (Table I). A quantilequantile plot was used to determine whether infant death rates were normally distributed and amenable to analysis with traditional parametric techniques. On the basis of the observed plot, log transformations were then used as a means of normalizing the death rate distribution for subsequent modeling. 762
A trend analysis with orthogonal contrasts was used to estimate the linear, quadratic, or higher level polynomial effects. A regression-based analysis of variance model was then used as a means of assessing the relationship between infant death rates from congenital heart disease and gestational age at birth from 34 to 40 weeks, after adjusting for birth year. Pre-specified subgroup comparisons to the statistically significant omnibus F-test were computed with the TukeyKramer test, a derivation of the Tukey honestly significant difference technique, selected specifically for its ability to maintain a consistent type I error-rate across all post-hoc comparisons. The experiment wise error rate was set at the nominal (a = 0.05) level. All analyses were conducted with SAS software version 9.1.3 (SAS Institute Inc, Cary, North Carolina).
Results Of the 14.9 million birth records analyzed for 2000 to 2003, 12.5 million (84%) occurred between 34 and 40 weeks gestational age. Congenital heart disease deaths occurred in 4736 infants (0.04%) born between 34 and 40 weeks. These deaths
Figure 1. Proportion of congenital heart disease deaths by gestational age in weeks. CHD, congenital heart disease. Cnota et al
ORIGINAL ARTICLES
November 2011 represent 75% of all infant congenital heart disease deaths across all gestational ages. Within the study population of infants born 34 to 40 weeks gestational age, most congenital heart disease deaths occurred in infants born at full term, gestational age 37 to 40 weeks (Figure 1). The study consisted of 51% male and 49% female infants born at 34 to 40 weeks gestational age. Approximately 79% of the infants were white, 15% were black, and the remaining 6% consisted of other races. Proportional infant death rates attributed to congenital heart disease are presented in Table I. The untransformed death rates represent the number of deaths per 100 live births for a given gestational age. Untransformed death rates range from two in 10 000 at 40 weeks gestational age to 12 in 10 000 at 34 weeks gestational age. Log-transformed death rates are also presented because those rates were used specifically to normalize data for the analysis. When assessing the trend across gestational age groups, only the linear component of orthogonal polynomial contrast coefficients was statistically significant (P < .01), suggesting that there was a statistically significant linear trend in the means of congenital heart disease death rate, but no higher order trend. Results of the analysis of variance test (Figure 2) indicated that for infants born at 34 to 40 weeks gestational age, estimates of congenital heart disease death rates decreased with each 1-week increase in gestational age (R2 = 0.97). Follow-up analysis to the omnibus test was performed by using the Tukey-Kramer post hoc technique. Whereas the difference in congenital heart disease death rates across 1-week intervals (ie, 34 weeks versus 35 weeks) ranged in significance (P = .02-.23, Table II), all 2-week intervals (ie, 38 versus 40 weeks) were markedly significant (P < .01). To assess whether the overall observations were also seen in high-risk subgroups, we performed two additional analyses. First, only deaths that occurred during the first 28 days of life were considered. This neonatal population demonstrated the same negative linear relationship with gestational age that
Figure 2. Linear relationship between congenital heart disease death rates and gestational age. Error bars represent the variability of annual death rate across the years 2000 to 2003.
Table II. Post hoc analysis of congenital heart disease death rates in a 1-week interval in gestational age (log scale) Comparison (weeks) 34 versus 35 35 versus 36 36 versus 37 37 versus 38 38 versus 39 39 versus 40
Estimate 0.29 0.23 0.24 0.35 0.31 0.20
P value .02 .11 .10 <.01 .01 .23
All P values were modified with the Tukey-Kramer adjustment technique. All 2-week interval differences in death rate were statistically significant (P < .01).
was observed in the overall congenital heart disease population (R2 = 0.95, P < .01). Second, only deaths with a primary diagnosis of hypoplastic left heart syndrome were considered. This diagnosis was the single most common cause of infant death from congenital heart disease and was the only individual diagnosis with a sufficient number of deaths to allow NCHS reporting of diagnosis-specific rates. Death rates attributed to hypoplastic left heart syndrome demonstrated the same negative linear relationship with gestational age (R2 = 0.75, P < .01) as the overall congenital heart disease population.
Discussion This study demonstrated a significant, negative linear relationship between congenital heart disease death rates and gestational age between 34 and 40 weeks. Furthermore, the data suggest that even an additional week of gestation is associated with lower congenital heart disease death rates at some gestational age intervals. An additional 2 weeks of gestation is highly statistically significant across all intervals examined. An explanation for why some gestational age intervals are statistically significant and others are not remains unclear. Perhaps developmental milestones unique to a given week of gestation can have relatively greater or lesser effect on mortality rates. These findings appear to parallel observations in the general population, in which the public health significance of such findings has already been appreciated. Although the highest risk of morbidity and mortality has been attributed to infants born at gestational age <34 weeks, recent findings have demonstrated increased morbidity and mortality rates in the late preterm infant born 34 to 36 weeks.18 Rates of pulmonary complications19 and infant mortality continue to decline to a nadir at 39 to 40 weeks. Such findings are particularly concerning because there has been a 25% increase in late preterm deliveries since 199020 and an overall decrease in the mean gestational age at delivery. Complementing the population-based methods of this study, a recent large single center study6 suggested that infants with major congenital heart disease may be particularly vulnerable to the medical complications associated with gestational age <39 weeks. Respiratory insufficiency superimposed on the complex cardiopulmonary interaction of neonatal congenital heart disease results in poor outcome.21,22 Additionally, recent findings suggest that the brain maturation in infants with
Congenital Heart Disease Infant Death Rates Decrease as Gestational Age Advances from 34 to 40 Weeks
763
THE JOURNAL OF PEDIATRICS
www.jpeds.com
complex congenital heart disease may be delayed by as long as 4 weeks.22,23 In effect, many of our ‘‘term’’ infants with congenital heart disease may have neurological findings and risks similar to premature infants. As we continue to learn about the neurodevelopmental consequences of congenital heart disease, even small differences in gestational age may turn out to have clinical importance. In the current era, a large portion of patients with a complicated single ventricle are diagnosed with fetal echocardiogram—as many as 75% of patients with hypoplastic left heart syndrome.13 Those babies appear to be delivered earlier than their undiagnosed counterparts. And, although the overall difference in gestational age between infants with prenatal and postnatal diagnosis may appear relatively modest, those born after labor induction may deliver as much as 1.5 weeks earlier.11 The results of this study suggest that such differences can have a significant effect on postnatal outcomes. Therefore, although labor and delivery management must be considered in the context of each maternal-fetal dyad, these data support the recommendations from the American College of Obstetricians and Gynecologists Practice Bulletin on Induction of Labor. Even in the presence of prenatally diagnosed congenital heart disease, scheduled delivery before 39 weeks gestation should be avoided.24 Because of the known delayed brain maturation and potential for continued fetal growth, further research should address the question of whether elective delivery for the sole indication of fetal congenital heart disease is appropriate at any gestational age. The use of federal datasets, including those derived from birth certificates and death certificates, introduces limitations. Estimates of gestational age are derived from imperfect criteria.25 NCHS does apply a diagnostic algorithm that is well described16 and reduces the likelihood of significantly erroneous assignment of gestational age. Primary cause of death relies on individual practicing physicians to complete death certificates. Data quality is improved by using standardized logic algorithms designed to arrive at the most likely cause of death.26 Congenital heart disease incidence is not available because the time course and clinical presentation of congenital heart disease does not readily allow for documentation of defects on the birth certificate. Such incidence data are necessary to calculate true death rates (deaths of infants with congenital heart disease/births of infants with congenital heart disease); therefore these data are only able to describe proportional death rates. Although earlier studies have indicated that infants with congenital heart disease are not born at significantly different gestational age overall,27 diagnosis-specific differences in incidence, gestational age, and death rates could contribute to the findings. Finally, birth certificate data do not allow for reliable assessment of prenatal diagnosis; therefore we are unable to determine how this important variable influences outcome. These findings provide a broad perspective on the relationship between gestational age and congenital heart disease. They highlight the idea that perinatal factors impacting the general population can be important, or are even more important, in a smaller population of high-risk infants with 764
Vol. 159, No. 5 congenital heart disease. Although the relatively simple model has limited ability to address confounding factors that could influence the observed relationship, the association is very strong and the decision to analyze gestational age as a discrete variable allows for direct translation of weekly increments to the clinical setting. Further work is necessary to understand how characteristics of the mother, pregnancy and fetus influence gestational age in this population. Although these results are similar to observations from the general population, supporting the validity of the findings, they also extend the relationship to a unique, high-risk population. An increase of even 1-week in gestational age may result in reduced deaths in these late preterm and term infants. These findings challenge the common approach of scheduling elective delivery before 40 weeks after prenatal diagnosis of complex congenital heart disease. n Submitted for publication Oct 5, 2010; last revision received Mar 28, 2011; accepted Apr 18, 2011. Reprint requests: James Cnota, MD, The Heart Institute, Division of Cardiology, Cincinnati Children’s Hospital Medical Center, MLC 2003, 3333 Burnet Ave, Cincinnati, OH 45229-3039. E-mail: [email protected]
References 1. Andrews RE, Simpson JM, Sharland GK, Sullivan ID, Yates RW. Outcome after preterm delivery of infants antenatally diagnosed with congenital heart disease. J Pediatr 2006;148:213-6. 2. Dees E, Lin H, Cotton RB, Graham TP, Dodd DA. Outcome of preterm infants with congenital heart disease. J Pediatr 2000;137:653-9. 3. Jenkins KJ, Gauvreau K, Newburger JW, Spray TL, Moller JH, Iezzoni LI. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg 2002;123:110-8. 4. Lacour-Gayet F, Clarke D, Jacobs J, Comas J, Daebritz S, Daenen W, et al. The Aristotle score: a complexity-adjusted method to evaluate surgical results. Eur J Cardiothorac Surg 2004;25:911-24. 5. Mathews TJ, MacDorman MF. Infant mortality statistics from the 2005 period linked birth/infant death data set. Natl Vital Stat Rep 2008;57:1-32. 6. Costello JM, Polito A, Brown DW, McElrath TF, Graham DA, Thiagarajan RR, et al. Birth before 39 weeks’ gestation is associated with worse outcomes in neonates with heart disease. Pediatrics 2010;126:277-84. 7. Allan L, Dangel J, Fesslova V, Marek J, Mellander M, Oberhansli I, et al. Recommendations for the practice of fetal cardiology in Europe. Cardiol Young 2004;14:109-14. 8. Rychik J, Ayres N, Cuneo B, Gotteiner N, Hornberger L, Spevak PJ, et al. American Society of Echocardiography guidelines and standards for performance of the fetal echocardiogram. J Am Soc Echocardiogr 2004;17:803-10. 9. Cohen MS, Frommelt MA. Does fetal diagnosis make a difference? Clin Perinatol 2005;32:877-90, viii. 10. Bonnet D, Coltri A, Butera G, Fermont L, Le Bidois J, Kachaner J, et al. Detection of transposition of the great arteries in fetuses reduces neonatal morbidity and mortality. Circulation 1999;99:916-8. 11. Raboisson MJ, Samson C, Ducreux C, Rudigoz RC, Gaucherand P, Bouvagnet P, et al. Impact of prenatal diagnosis of transposition of the great arteries on obstetric and early postnatal management. Eur J Obstet Gynecol Reprod Biol 2009;142:18-22. 12. Bartlett JM, Wypij D, Bellinger DC, Rappaport LA, Heffner LJ, Jonas RA, et al. Effect of prenatal diagnosis on outcomes in D-transposition of the great arteries. Pediatrics 2004;113:e335-40. 13. Atz AM, Travison TG, Williams IA, Pearson GD, Mahle WT, Sklansky MS, et al. Impact of prenatal diagnosis and risk factors for preoperative death in neonates with single right ventricle: screening results
Cnota et al
November 2011
14.
15.
16. 17. 18.
19.
20.
from the Pediatric Heart Network Single Ventricle Reconstruction Trial. J Am Coll Cardiol 2009;53:A355-66. Mathews TJ, MacDorman MF. Infant mortality statistics from the 2003 period linked birth/infant death data set. Natl Vital Stat Rep 2006;54: 1-29. National Center for Health Statistics. United States Linked Birth/Infant Death Period Public Use Files, Years 2000, 2001, 2002, and 2003. In: Service USPH, editor. Hyattsville, Maryland: National Center for Health Statistics; 2002, 2003, 2004, and 2005. Martin JA. United States vital statistics and the measurement of gestational age. Paediatr Perinat Epidemiol 2007;21(Suppl. 2):13-21. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2003. Natl Vital Stat Rep 2005;54:1-116. Davidoff MJ, Dias T, Damus K, Russell R, Bettegowda VR, Dolan S, et al. Changes in the gestational age distribution among US singleton births: impact on rates of late preterm birth, 1992 to 2002. Semin Perinatol 2006;30:8-15. Morrison JJ, Rennie JM, Milton PJ. Neonatal respiratory morbidity and mode of delivery at term: influence of timing of elective caesarean section. Br J Obstet Gynaecol 1995;102:101-6. MacDorman MF, Menacker F, Declercq E. Cesarean birth in the United States: epidemiology, trends, and outcomes. Clin Perinatol 2008;35:293-307, v.
ORIGINAL ARTICLES 21. McMahon CJ, Penny DJ, Nelson DP, Ades AM, Al Maskary S, Speer M, et al. Preterm infants with congenital heart disease and bronchopulmonary dysplasia: postoperative course and outcome after cardiac surgery. Pediatrics 2005;116:423-30. 22. Miller SP, McQuillen PS, Hamrick S, Xu D, Glidden DV, Charlton N, et al. Abnormal brain development in newborns with congenital heart disease. N Engl J Med 2007;357:1928-38. 23. Licht DJ, Shera DM, Clancy RR, Wernovsky G, Montenegro LM, Nicolson SC, et al. Brain maturation is delayed in infants with complex congenital heart defects. J Thorac Cardiovasc Surg 2009;137:529-36, discussion 36-37. 24. ACOG Practice Bulletin No. 107: induction of labor. Obstet Gynecol 2009;114:386-97. 10.1097/AOG.0b013e3181b48ef5. 25. Wier ML, Pearl M, Kharrazi M. Gestational age estimation on United States livebirth certificates: a historical overview. Paediatr Perinat Epidemiol 2007;21:4-12. 26. National Center for Health Statistics. Vital statistics, instructions for classifying the underlying cause of death. NCHS instruction manual; part 2a. Hyattsville, MD: Public Health Service. Published annually. 27. Reller MD, Strickland MJ, Riehle-Colarusso T, Mahle WT, Correa A. Prevalence of congenital heart defects in metropolitan Atlanta, 19982005. J Pediatr 2008;153:807-13.
Congenital Heart Disease Infant Death Rates Decrease as Gestational Age Advances from 34 to 40 Weeks
765