Getahun, Darios MD, MPH*; Amre, Devendra MD, PhD†; Rhoads, George G. MD, MPH‡; Demissie, Kitaw MD, PhD‡§
Sudden infant death syndrome (SIDS) is a leading cause of infant death during the postneonatal period. It is defined as an infant death for which a careful history and a postmortem investigation (an autopsy and an examination of the scene of death) fail to explain the cause of death.1 The incidence of SIDS in the United States is reported to be 0.62 per 1,000 live births in 2000 (n = 2,523).2 Although a national campaign aimed at reducing prone sleeping has led to reductions in the incidence of SIDS,3–5 major potential etiological factors remain unidentified. Some studies have suggested that perinatal events could play an important role. Information regarding these, with the exception of smoking, has been inconsistent. Interpretation of findings from previous studies has been hampered by limited sample sizes, potential misclassification of risk factors, lack of control for confounding variables, and an inability to account for potential interaction between risk factors. To further clarify the role of perinatal events as risk factors for SIDS, we conducted a population-based study of singleton births and linked infant deaths for the United States from 1995 to 1998.
MATERIALS AND METHODS
Data for this analysis were obtained from the National Center for Health Statistics Linked Birth and Infant Death file for the years 1995–1998. Using data provided by individual states under the Vital Statistics Cooperative Program, National Center for Health Statistics routinely links infant deaths (less than 365 days of age) to birth certificates. The linked birth and infant death files contain the following information on the mother and infant: sociodemographic, obstetric medical history, complications of the index pregnancy, and labor and neonatal outcomes. Information on cause of infant's death and time of death is also included in the data set. The variables studied included maternal age, maternal race, maternal education, prenatal care, gestational age, infant sex and weight, maternal smoking during pregnancy, diabetes, gestational anemia, chronic hypertension, pregnancy-induced hypertension, placenta previa, abruptio placentae, premature rupture of membranes, and maternal fever.
A case-control analysis was conducted to identify potential maternal and obstetric risk factors for SIDS. Sudden infant death syndrome cases were identified from the data set by using the classification recommended by the International Collaborative Effort.6 The International Collaborative Effort used the International Classification of Diseases, Ninth Revision (ICD-9-CM)7 diagnosis codes to identify cases of SIDS. In addition to the recoded ICD-9-CM code of 798.0 for SIDS, the International Collaborative Effort classification includes ICD-9-CM code of E913 (accidental mechanical suffocation) as SIDS deaths. This classification is believed to identify nearly all SIDS cases. We excluded infants with gestational age less than 20 weeks and birth weights of less than 500 g. The justification for excluding this group was to avoid errors in gestational age estimation and to minimize interstate differences in reporting live births that were at borderline of viability. Substantial interstate variability has been noted in reporting these infants as fetal deaths or live births. Furthermore, we excluded infants born to twin or higher-order births from the analyses. In each calendar year, infants with an ICD-9-CM diagnosis code of 798.0 or E9130 (SIDS cases) were identified (n = 3,329 for 1995; n = 3,161 for 1996; n = 3,018 for 1997; and n = 2,896 for 1998). Within each year, we selected a random control sample that was 4 times larger than the SIDS sample. This was done by assigning a random number to each non-SIDS births, sorting into order by this number, and selecting the required sample from the beginning of the list. Last, each year selected controls were combined to form the combined control group. Cases (n = 12,404) and controls (n = 49,616) over the 4-year period were combined for subsequent analysis.
The outcome variable of interest was being a case or a control (SIDS or non-SIDS infant, respectively). Independent variables included in this analysis were medical and obstetric complications, including diabetes (juvenile onset, adult onset, and gestational diabetes), gestational anemia (hemoglobin level of less than 10 g/dL during pregnancy or a hematocrit of less than 30% during pregnancy), chronic hypertension (blood pressure persistently greater than 140/90 mm Hg, diagnosed before the 20th week of gestation), pregnancy-induced hypertension (blood pressure increase of at least 30 mm Hg systolic or 15 mm Hg diastolic on 2 measurements taken 6 hours apart after the 20th week of gestation), placenta previa (implantation of the placenta over or near the internal opening of the cervix), abruptio placentae (premature separation of a normally implanted placenta), premature rupture of membranes (rupture of the membranes more than 12 hours before the onset of labor), maternal fever (greater than 100°F or 38°C occurring during labor and/or delivery). For this analysis, small for gestational age was defined as birth weight that is less than the 3rd and 10th percentile, respectively, based on the 1995–1998 race- and sex-specific normogram.
We considered maternal age (less than 20 years, 20–24 years, 25–29 years, 30–34 years, and 35 years or greater), maternal race (white, black, and others), maternal education (less than 12 years, equal to 12 years, 13–15 years, equal to 16 years, and 17 years or greater), and maternal smoking during pregnancy (yes or no) as potential risk and confounding variables.
Potential confounding variables on the association between placental factors and SIDS were chosen a priori. We then performed a univariable analysis to compare the distribution of the potential confounders among the cases and the controls (Table 1). Because the distribution of the variables was different among cases and controls, variables were not removed from the unconditional multiple logistic regression model even if they did not achieve statistical significance. This approach is likely to increase the precision of the estimates.8 An estimate was determined to be statistically significant if the 95% confidence interval (CI) did not contain the null value of 1. To assess temporal trend in SIDs, we conducted Mantle–Haenszel χ2 statistics for trend. An analysis was conducted by using SAS statistics software (SAS institute, Cary, NC). The study was approved by the University of Medicine and Dentistry of New Jersey Institutional Review Board.
There was a total of 15,188,888 singleton live births in the United States for the period 1995–1998. Exclusion of infants with gestational age less than 20 weeks and birth weights of less than 500 g resulted in 15,186,869 infants. Of these infants, 12,404 died of SIDS.
The overall incidence of SIDS was 81.7 per 100,000 live births during the study period. There was a continuous year-to-year decrease in the incidence of SIDS (87.6 per 100,000 for 1995, 83.5 per 100,000 for 1996, 80.0 per 100,000 for 1997, and 75.7 per 100,000 for 1998), and the decreasing trend in the incidence of SIDS was statistically significant (P < .001).
Table 1 outlines the sociodemographic characteristics of the cases and controls. Compared with the controls, the cases were more likely to be African Americans and their mothers were significantly younger and less likely to have completed high school or university education. In comparison with infants of nonsmoking mothers, infants of smoking mothers during the index pregnancy were more likely to die of SIDS, and SIDS mothers were more likely to have gestational anemia, placenta previa, abruptio placentae, and premature rupture of membranes. Male babies were also more frequently seen in cases. Surprisingly, the results showed that the frequency of intrapartum fever was higher in the control group compared with the case group.
Table 2 outlines the results obtained after adjusting for each individual risk factor and for potential confounding factors listed in Table 1 by using unconditional multiple logistic regression models. Risks for SIDS were significantly associated with younger maternal age at pregnancy, no prenatal care (odds ratio [OR] 1.70; 95% CI 1.44, 2.00), maternal smoking during pregnancy (OR 3.19; 95% CI 3.03, 3.37), small for gestational age (OR of SIDS among infants less than the 3rd percentile for their gestational age as compared with infants appropriate for their gestational age: 1.33; 95% CI 1.19, 1.49 and OR of SIDS among infants less than the 10th percentile for their gestational age as compared with infants appropriate for their gestational age: 1.40; 95% CI 1.30, 1.50), events complicating pregnancy and delivery, such as placenta previa (OR 1.70; 95% CI 1.24, 2.33), abruptio placentae (OR 1.57; 95% CI 1.24, 1.98), premature rupture of membranes (OR 1.48; 95% CI 1.33, 1.66), and male sex (OR 1.47; 95% CI 1.40, 1.53). To further understand the negative association seen between intrapartum fever and SIDS in our analysis, we reiterated the analysis after stratifying infants into preterm and term categories. Intrapartum fever remained protective in term infants (OR 0.68; 95% CI 0.54, 0.85); however, it was nonsignificantly associated with SIDS in preterm infants (OR 1.22; 95% CI 0.83, 1.79). Risks for SIDS were not elevated in pregnant mothers who had diabetes, gestational anemia, pregnancy-induced hypertension, and/or chronic hypertension.
Numerous perinatal complications are associated with tobacco use. Considering that the elevated risk of SIDS observed for lower gestational age at delivery (OR of SIDS among gestational age less than 37 weeks compared with 37 or greater equals 1.83, 95% CI 1.73, 1.94) and placental abnormalities (placenta previa, abruptio placentae, and premature rupture of membranes) could be the result of their association with maternal smoking during pregnancy,9 we evaluated the models with and without maternal smoking during pregnancy. The estimates for the former did not change substantially (placenta previa OR 1.73; 95% CI 1.27, 2.36; abruptio placentae OR 1.79; 95% CI 1.43, 2.25; and premature rupture of membranes OR 1.57; 95% CI 1.41, 1.75), indicating that their associations with SIDS were independent of maternal smoking during pregnancy.
The overall incidence of SIDS using ICD-9-CM code of 798.0 alone was 74.2 per 100,000 live births during the study period. The same analysis using this code alone did not change the magnitude of the result (placenta previa OR 1.74; 95% CI 1.25, 2.43; abruptio placentae OR 1.42; 95% CI 1.11, 1.80; and premature rupture of membranes OR 1.49; 95% CI 1.33, 1.68).
Our study showed a decreasing trend in the incidence of SIDS during the study period, and this decreasing trend is likely to be real and unlikely to be an artifact of the data acquisition methods used by the National Center for Health Statistics because these are quite standardized and were not known to vary during the years represented by the study birth cohorts (1995–1998). Our finding of decreasing trend in the incidence of SIDS in the United States is consistent with results from earlier studies using other datasets.10–12
In this nested case-control study, we found that young maternal age, low education, and black race were substantial risk factors for SIDS. In addition, prenatal events, such as lack of prenatal care, smoking during pregnancy, placenta previa, abruptio placentae, and premature rupture of membranes increased the risks for SIDS. Similarly, male sex was associated with a higher risk for SIDS.
Strengths of the present study include 1) a large sample size; 2) a population-based case-control design; 3) a more complete definition of SIDS, including cases previous inaccurately classified as non-SIDS; and 4) control for potential confounding variables.
Conversely, a limitation of this study is its dependence on the Linked Birth and Infant Death file that is likely to contain some coding errors, which could introduce random or systematic bias. Because using the International Collaborative Effort classification codes for SIDS is more inclusive, there is a possibility of overestimation of SIDS. However, use of ICD-9-CM code 798.0 alone could fail to identify some SIDS cases and underestimate its true impact. We reanalyzed the data using ICD-9-CM code of 798.0 alone to evaluate any discrepancies between these 2 approaches. The magnitudes of the ORs were similar using both approaches.
We were able to confirm the consistently reported association between maternal smoking during pregnancy and SIDS. Our findings also support the previously reported associations between lack of or delayed prenatal care and small for gestational age and SIDS. Although some studies indicate that the latter 2 could be associated with SIDS through their relationship with maternal smoking during pregnancy, our analyses indicates that they most likely play an independent role.
A limited number of studies have examined the role of placental abnormalities as risk factors for SIDS. Klonoff-Cohen et al13 recently reported an 8-fold increase in the risk for SIDS associated with placental abruption in a study conducted in California. These large estimates, however, were based on a very small number of subjects (8 cases and 1 control), and the use of a single control precluded the adjustment for potential confounding variables. In an earlier study conducted among a similar population from California, Li and Wi14 reported a 2-fold elevation in risk for SIDS in association with placental abruption and placenta previa. Their study was based on a large sample of SIDS cases (2,028) and 21,037 controls, and they adjusted for potential confounding variables. Earlier, Buck et al15 reported a 21-fold increase in risk from placenta previa (OR 21.8; 95% CI 2.80, 169.78) and a more than 3-fold increase in risks associated with abruptio placentae (OR 3.65; 95% CI 1.09, 12.22) in their case-control study. The wide confidence intervals indicate substantial uncertainty in these results. Our estimates, based on a very large sample, support the findings reported by Li and Wi.14 These observations suggest that pregnancies complicated by long-standing placental abnormalities might predispose to SIDS. Perhaps, fetal hypoxia resulting from long-standing abnormalities of the placenta could result in subtle neurological damage that contributes to later infant demise.
In contrast to our findings, premature rupture of membranes was not found to be a factor associated with SIDS in the study done by Buck et al.16 However, their small sample size could have precluded the determination of significantly elevated risks.
There have been reports suggesting that maternal anemia during pregnancy could increase the risk for SIDS by 2- to 3-fold. Our study indicated that maternal anemia does not increase the risk for SIDS. Only 2 previous studies have evaluated risks for SIDS associated with preeclampsia/eclampsia. Li and Wi17 reported a 50% increase in risk (OR 1.5; 95% CI 1.1, 2.0), based on a case-control study. However, on controlling for preterm delivery and low birth weight, the magnitude of the latter was reduced and was no longer significant (OR 1.3, 95% CI 0.9, 1.8). Previously, Buck et al15 reported elevated risks associated with eclampsia, but these finding were unstable because of small sample sizes and inadequate control for confounding variables. We did not find any association between pregnancy-induced hypertension and SIDS. We expected an association between intrapartum fever during labor and SIDS occurrence, but our finding did not support the anticipated association.
Because intrapartum fever is a surrogate measure of chorioamnionitis and chorioamnionitis is associated with neurodevelopmental sequelae,18,19 an association between intrapartum fever and SIDS would be expected. To further understand the negative association observed between intrapartum fever and SIDS in our analysis, we reiterated the analysis after stratifying infants into preterm and term categories. Intrapartum fever remained to be protective in term infants; however, it was associated with SIDS in preterm infants. Similarly, our group20 previously has shown that intrapartum fever is a risk factor for SIDS in preterm infants (OR 1.43; 95% CI 0.90, 2.27), but was protective from SIDS among term infants (OR 0.75; 95% CI 0.51, 1.12). One potential explanation for the negative association seen in term infants in our study could have been the stress associated with intrapartum fever on the unborn child of term pregnancy. Fetal stress associated with intrapartum fever may trigger increased activity of the adrenocortical system and as a result increase the level of glucocorticotropin hormone in the fetal circulation. The increased steroid hormone may accelerate fetal organ maturation (eg, lung maturation). This and other unexplained physiological changes may work in favor of the newborn infant and enable him or her to overcome postnatal adverse events. However, the positive association seen in preterm infants might be the impact of chorioamnionitis on fetal brain.19,21 It is logical that this effect would be highest in very preterm infants. Naef et al22 and Nicaise et al23 found higher neonatal morbidity and mortality associated with preterm premature rupture of membranes.
Numerous hypotheses concerning the pathophysiology of SIDS have been put forth. Recently, Filiano et al24 proposed a triple-risk hypothesis stating that SIDS results from the intersection of 3 overlapping factors: 1) a vulnerable infant; 2) a critical development period in homeostatic control; and 3) an exogenous stressor. Death from SIDS occurs only in the presence of these 3 factors. Fitted within the “vulnerable infant” concept are epidemiologic findings of increased risks for SIDS from certain prenatal and postnatal events that are likely to make some children more susceptible. These susceptible children potentially undergo critical changes during 2–6 months of age when the incidence of SIDS is most common and finally succumb to SIDS from exposure to exogenous stressors, such as infections or sleeping in prone positions. This 3-tier hypothesis has recently been challenged by Guntheroth et al,25 who stated that SIDS was more likely to be of multifactorial origin with interactions between various risk factors (a hypothesis previously put forward by Bergman26 in 1970 and Raring27 in 1975) and unlikely to be a 3-tier process.
Our epidemiological observations based on a large database do not provide a single all-encompassing hypothesis for the development of SIDS. Nevertheless, some recent findings indicating increased risk associated with placental abnormalities along with risks associated with maternal smoking and preterm birth suggest an important role for factors that lead to “hypoxic conditions” either in the fetus or in the newborn that could influence the adequate development of the brain–cardiorespiratory axis and that eventually contributes to the development of SIDS. From a public health perspective, elevated risks for SIDS related to young maternal age, lack of education, and absence of or late prenatal care are important. Increasing community awareness and patient acceptance of the benefits of prenatal services might contribute to reducing risks from modifiable factors, such as maternal smoking and to the earlier detection of less-modifiable risks resulting from placental abnormalities and other prenatal and postnatal factors. Whatever the mechanisms underlying the pathophysiology of SIDS, in addition to the “back to sleep” initiative, efforts should be made toward the control and prevention of potentially hazardous prenatal and postnatal events to further reduce the incidence of SIDS.
1. Willinger M, James LS, Catz C. Defining the sudden infant death syndrome (SIDS): deliberations of an expert panel convened by the National Institute of Child Health and Human Development [review]. Pediatr Pathol 1991;11:677–84.
2. Minino AM, Arias E, Kochanek KD, Murphy SL, Smith BL. Deaths: final data for 2000. Natl Vital Stat Rep 2002;50:1–119.
3. Changing concepts of sudden infant death syndrome: implications for infant sleeping environment and sleep position. American Academy of Pediatrics. Task Force on Infant Sleep Position and Sudden Infant Death Syndrome [review]. Pediatrics 2000;105:650–6.
4. Irgens LM, Markestad T, Baste V, Schreuder P, Skjaerven R, Oyen N. Sleeping position and sudden infant death syndrome in Norway 1967–91. Arch Dis Child 1995;72:478–82.
5. Oyen N, Markestad T, Skaerven R, Irgens LM, Helweg-Larsen K, Alm B, et al. Combined effects of sleeping position and prenatal risk factors in sudden infant death syndrome: the Nordic Epidemiological SIDS Study. Pediatrics 1997;100:613–21.
6. Cole S, Hartford RB, Bergsjo P, McCarthy B. International collaborative effort (ICE) on birth weight, plurality, perinatal, and infant mortality. III: A method of grouping underlying causes of infant death to aid international comparisons. Acta Obstet Gynecol Scand 1989;68:113–7.
7. Centers for Disease Control, United States. Health Care Financing Administration, National Center for Health Statistics. ICD-9-CM International Classification of Diseases, 9th revision, Clinical Modification, 6th ed. Washington, DC: U.S. Department of Health and Human Services Centers for Disease Control and Prevention Health Care Financing Administration; 1997.
8. Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other multivariable methods. The Duxbury series in statistics and decision sciences. 2nd ed. Boston (MA): PWS-Kent Pub Co; 1988. p. 369.
9. Naeye RL. Abruptio placentae and placenta previa: frequency, perinatal mortality, and cigarette smoking. Obstet Gynecol 1980;55:701–4.
10. Paris CA, Remler R, Daling JR. Risk factors for sudden infant death syndrome: changes associated with sleep position recommendations. J Pediatr 2001;139:771–7.
11. Pollack HA, Frohna JG. A competing risk model of sudden infant death syndrome incidence in two US birth cohorts. J Pediatr 2001;138:661–7.
12. Sudden infant death syndrome: United States, 1983–1994. MMWR Morb Mortal Wkly Rep 1996;45:859–63.
13. Klonoff-Cohen HS, Srinivasan IP, Edelstein SL. Prenatal and intrapartum events and sudden infant death syndrome. Paediatr Perinat Epidemiol 2002;16:82–9.
14. Li DK, Wi S. Maternal placental abnormality and the risk of sudden infant death syndrome. Am J Epidemiol 1999;149:608–11.
15. Buck GM, Cookfair DL, Michalek AM, Nasca PC, Standfast SJ, Sever LE. Assessment of in utero hypoxia and risk of sudden infant death syndrome. Paediatr Perinat Epidemiol 1989;3:157–73.
16. Buck GM, Michalek AM, Kramer AA, Batt RE. Labor and delivery events and risk of sudden infant death syndrome (SIDS). Am J Epidemiol 1991;133:900–6.
17. Li DK, Wi S. Maternal pre-eclampsia/eclampsia and the risk of sudden infant death syndrome in offspring. Paediatr Perinat Epidemiol 2000;14:141–4.
18. Patrick LA, Smith GN. Proinflammatory cytokines: a link between chorioamnionitis and fetal brain injury [review]. J Obstet Gynaecol Can 2002;24:705–9.
19. Dammann O, Leviton A. Role of the fetus in perinatal infection and neonatal brain damage [review]. Curr Opin Pediatr 2000;12:99–104.
20. Petrova A, Demissie K, Rhoads GG, Smulian JC, Marcella S, Ananth CV. Association of maternal fever during labor with neonatal and infant morbidity and mortality. Obstet Gynecol 2001;98:20–7.
21. Leviton A, Paneth N, Reuss ML, Susser M, Allred EN, Dammann O, et al. Maternal infection, fetal inflammatory response, and brain damage in very low birth weight infants: Developmental Epidemiology Network Investigators. Pediatr Res 1999;46:566–75.
22. Naef RW3rd, Allbert JR, Ross EL, Weber BM, Martin RW, Morrison JC. Premature rupture of membranes at 34 to 37 weeks’ gestation: aggressive versus conservative management. Am J Obstet Gynecol 1998;178:126–30.
23. Nicaise C, Gire C, Fagianelli P, Debriere R, Thomachot L, d'Ercole C, et al. [Neonatal consequences of preterm premature rupture of membrane (PPROM) at 24–34 WG: 118 singleton pregnancies]. J Gynecol Obstet Biol Reprod (Paris) 2002;31:747–54.
24. Filiano JJ, Kinney HC. A perspective on neuropathologic findings in victims of the sudden infant death syndrome: the triple-risk model. Biol Neonate 1994;65:194–7.
25. Guntheroth WG, Spiers PS. The triple risk hypotheses in sudden infant death syndrome [review]. Pediatrics 2002;110:e64.
26. Bergman AB. Synthesis. In: Bergman AB, Beckwith JB, Ray CG, editors. Sudden infant death syndrome. Seattle (WA): University of Washington Press; 1970. p. 210–1.
27. Raring RH. Crib death: scourge of infants-shame of society. Hicksville (NY): Exposition Press; 1975. p. 93–7.