- Email: [email protected]
Effects of birth weight and ethnicity on incidence of sudden infant death syndrome
Effects of birth weight and ethnicity incidence of sudden infant death syndrome
on
Sudden infant death syndrome occurs with increased frequency in low birth weight infants and in black infants. The degree to which the higher LBW rate among blacks might explain this higher SIDS rate is unknown. To address this question, we analyzed the 4233 SIDS deaths that occurred among 252,376 neonatal survivors in C o o k County from 1975 to 1980, using computer-coded matched infant birth and death records. Birth weight and ethnic group w e r e identified. The overall SIDS rates in blacks, Hispanics, and whites were 5.1, i.2, and 1.3/1000 neonatal survivors, respectively. Within e a c h ethnic group, the SIDS rates increased progressively with decreasing birth weight. Within the _< 1500 gm birth weight groups, the SIDS rates w e r e 46.4, 3.9, and 5.5/1000 neonatal survivors in blacks, Hispanics, and whites. Using direct standardization, we found that 27% of the SIDS rate disparity between blacks and whites could be explained by the higher LBW rate in blacks (14% vs 6% in whites). The g o o d outcomes in both LBW and SIDS rates for the Hispanic population were unexpected because, like blacks, Hispanics are socioeconomically disadvantaged. Findings for this group suggest that the remaining 73% of the increased SIDS rate in blacks cannot be attributed in a straightforward manner to differences in income or educational attainment. (J PEDIATR1986;108:209-214) Lehman Black, M.D., Richard J. David, M.D,, Robert T. Brouillette, M.D., a n d Carl E. Hunt, M.D. From the Department of Pediatrics, Children's Memorial Hospital, and the Center for Health Services and Policy Research, Northwestern University, Chicago
Sudden infant death syndrome remains the single leading cause of infant death beyond the neonatal period, accounting for nearly one half of all deaths from 28 days to 1 year of age. Since the convening of the Second International Congress on Sudden Infant Death S3}ndrome in 1969,1 several epidemiologic risk factors have been identified, but little reduction in the SIDS rate has occurred. Several studies have shown an increased risk for SIDS in low birth
Supported in part by the Children's Research Guild and the Lillian S. Wells Foundation. Presented in part at the Midwest Society for Pediatric Research, November 1983. Submitted for publication March 4, 1985; accepted July 24, 1985. Reprint requests: Richard J. David, M.D., Children's Memorial Hospital, 2300 Children's Plaza, Chicago, IL 60614.
weight babies, especially in the group weighing <1500 gm28; one preliminary report showed the SIDS risk to increase exponentially as birth weight decreased. 8 There is also an increased risk for SIDS in black infants.3,4,7,9-11 Because the LBW rate is significantly higher in blacks than in whites) 2,13 we examined the contribution of LBW LBW SIDS VLBW
Low birth weight Sudden infant death syndrome Verylow birth weight
to the racial disparity in SIDS rate between blacks and whites. We also examined the birth weight distribution and SIDS rates in Hispanics, the second largest ethnic minority in the United States and one having many socioeconomic risk factors. To assess the interrelationships between birth weight, race, and SIDS risk, we have examined a large
209
2 10
B l a c k et al.
The Journal o f Pediatrics February 1986
Table. SIDS rates in Cook County for black, white, and Hispanic infants, according to birth weight Birth weight group
_<1500 gm 1501-2500 gm >2500 gm All weights
Ethnic group
Neonatal survivors
SIDS deaths
SIDS rate (per 1000)
P
Black White Hispanic Black White Hispanic Black White Hispanic Black White Hispanic
2,477 1,096 260 19,286 10,944 2,470 150,177 214,123 51,573 171,910 226,163 54,303
40 6 1 209 33 13 629 250 52 878 289 66
16.3 5.5 3.8 10.8 3.0 5.3 4.2 1.2 1.0 5.1 1.3 1.2
0.0083* 0.6818~" 0.1164~ <0.001" 0.0836"~ 0.0093~ <0.0001 * 0.3371t <0.0001~: <0.0001 * 0.7039~ <0.0001:~
*Compared with white. tC0mPared with Hispanic. $c0mpared with black. postneonatal population in an ethnically diverse metropolitan area. i
METHODS , Population at risk. The population of Cook County, Illinois, in 1980 was ~ 5.2 million, with an ethnic composition that included 3.0 million white non-Hispanics, 1.3 million blacks, and 500,000 Hispanics.~4 Between 1975 and !980, the number of live births per year ranged between 84,000 and 86,000. Included in our study population were all infants born during the years 1975 through 1980 to mothers residing in Cook County. Deaths were analyzed for members of this cohort who died in the county between the ages of 28 and 364 days. Data sources. We obtained tapes from the Illinois Department of Public Health containing live birth files and matched infant birth-death files for the years 1975 through 1980. These annual birth-death files contain a record for each infant of the birth c o h o r t year w h o died before reaching 1 year of age. These consolidated records contain data from the infant's birth certificate and death certificate, even though these may have been filed in different years. Names and other identifying data are not included. We used the cause of death code 795.0 ("sudden death, cause unknown") from the International Classification of Diseases, eighth revision, for deaths occurring during the years 1975 through 1978 and 798.0 ("sudden infant death syndrome, cot death, crib death") from the ninth revision for the years 1979 and 1980. The autopsy rate for cases reported as S1DS was >95% for each ethnic group. Infants were divided into three different ethnic groups: black, white, and Hispanic, on the basis of the variables "infant race" and "maternal country of origin" derived from the birth record. Hispanic ethnicity as such was not
coded in the Illinois vital records before 1978 and had to be created by using South America, Central America, Cuba, Mexico, and Puerto Rico as the countries of maternal origin. Hispanic was further divided into black-Hispanic and white-Hispanic, but; because of the Small number of births in the former group (0.09%), we included only white-Hispanic in this study. Computations. w e generated a cross-t~tbulation of neonatal deaths, postneonatal SIDS deaths, and postneonatal deaths from all remainingcauses by 500 gm birth weight interVals from 500 to 5500 gm. This was done for each ethnic grbup in' each of the 6 years studied. SiDS rates were cal6ulated from the 6-year totals as follows: SIDS rate = (No. of SIDS infants/Neonatal survivors) • 1000. Neonatal survivors were infants >28 days and <1 year of age. The number Of neonatal survivors was calculated by subtracting neonatal deaths from live births. Live births by weight and ethni c group were obtained from a crosstabulation of all 1977 Cook County resident live births. Because o f the large number of cases in each cell of this b i r t h matrix per year, only 1 year's birth data were examined; taken from the midpoint of the study period. We then multiplied the number of 1977 live births by 6 to obtain an estimate of the 6-year total Of live births. From that estimated birth total and the yearly neonatal death totals, we calculated the number of neonatal survivors at risk for SIDS. Our calculations of the overall SIDS rate for Cook County included only blacks, whites, and Hispanics. All other ethnic groups were excluded because they were from a variety of" ethnic backgrounds and comprised only 3% of all Cook County SIDS deaths in these 6 years. We used the technique of direct standardization to examine the effect of birth weight distribution on the SIDS rate in the black population. 15 To do this we multipled each
Volume 108 Number 2
Birth weight, ethnicity, and S I D S
211
48
35
g 38
D-
25
,.z, ~e
g L
,8 m 5
B
g
g
g
g
g
g
g
g
g
u3
BIRTH
WEIGHT
GROUP
Fig. I. Frequency distributions for 1977 live births in Cook County by 500 gm birth weight categories for black ( - - ) , white (---), and Hispanic (---) ethnic groups.
12
CE
o I 8 w 0_ w
G
nt ~o
4
2
B
i
i
m ~
I
i
u1 a/
N
BIRTH
m
WEIGHT
I
i
i
m
N
to
m
~-
r
i
to
GROUP
Fig. 2. SIDS rate as a function of birth weight for all black, white, and Hispanic infants combined (Cook County infants aged 28 to 364 days, 1975 through 1980).
black birth weight-specific SIDS rate times the proportion of births in that weight category, assuming the same birth weight distribution as actually observed in the white population. Summing across birth weight categories then yielded an overall "expected" black SIDS rate. The difference between this expected rate and the actual black SIDS rate rePresents the frequency by which the black SIDS rate would be reduced if blacks had the same birth weight distribution as whites. Dividing this number by the difference between the actual black and actual white SIDS rates yields the percent of SIDS rate differences in the black and white populations explained by differences in birth weight distribution between the two populations.
RESULTS There were 1233 SIDS deaths among the 252,376 neonatal survivors in Cook County between 1975 and 1980. The overall SIDS rate for the 6 Years of this study was 2.7/1000 neonatal survivors. The overall SIDS rates for the black, white, and Hispanic infants were 5.1, 1.3, and 1.2, respectively (Table). The black SIDS rate was 4 times the white SIDS rate; the Hispanic rate did not differ significantly from that in white infants. Ethnic group differences in SIDS rates were also observed for moderatelY low birth weight (1501 to 2500 gm) and very low birth weight (__<1500 gm) groupS. In the moderately LBW group the black SIDS rate was 3.6 times higher than the white SIDS rate; among VLBW births the black rate was nearly
2 12
B l a c k et al.
The Journal o f Pediatrics February 1986
18
Q Z E U] D O -r tE bJ n Ld Frr n,
15
12
9
6
O3
i (g
{}
BIRTH
WEIGHT
GROUP
Fig. 3. SIDS rates as a function of birth weight for black ( - - ) , white (---), and Hispanic (---) ethnic groups (Cook County infants aged 28 to 364 days, t975 through 1980).
3 times the white rate. Again, the differences between white and Hispanic infants were not statistically significant. Birth weight distributions. Fig. 1 shows the three ethnic group birth weight distribution curves for 1977 live births in Cook County. The nearly identical shape of the white and Hispanic curves was reflected in similar rates of LBW for the two populations: 5.9% and 5.6%, respectively. The black LBW rate was 13.8%, more than twice the rate in either white or Hispanic infants. A similar relationship held for the VLBW group: the white an d Hispanic rates were identical at 0.4%, whereas the black rate was more than 6 times as high at 2.5%. Relation of SIDS to birth weight. Fig. 2 presents the overall b i r t h weight-specific SIDS mortality rates for Cook County. The SIDS rate was increased among survivors of lower birth weight, a trend that became most evident in the < 1500 gm weight category. The relationship between SIDS rate and birth weight is best described by an exponential model: Rate = 31.5e -~ (r =0.97, P <0.0001). Birth weight-specific SIDS mortality curves for the black, white, and Hispanic populations are depicted in Fig. 3. White and HisPanic curves were similar, at least above 2500 gm, but the black SIDS rates were higher throughout virtually the entire range of birth weights reported. The black and white rate curves are described better b y an exponential than by a linear model, but the coefficients (reflecting magnitude of birth weight effect) vary considerably. The Hispanic curve is fitted slightly better bY a linear model. Standardizing for birth weight differences between eth-
nic groups, we calculated that the black SIDS rate would be reduced from its actual level of 5.11/1000 to 4.08/1000 if the black infants had the same birth weight distribution as the white population. That is, 27% of the difference between the black and white SIDS rates can be explained by the difference in birth weight distributions in the black and white infants. DISCUSSION We have shown that the risk of SIDS increases exponentially with decreasing birth weight, with qualitatively similar birth weight-risk relationships for the three ethnic groups studied. The exponential relationship has been reported in one previous study, but those authors examined a relatively small, homogeneous white population? In contrast, our study population contains the largest number of SIDS cases yet analyzed and has enough ethnic diversity to permit confirmation of the exponential relationship not only for the population as a whole but also for individual ethnic groups. Comparisons between ethnic groups revealed a large black-white discrepancy in SIDS rates for nearly all birth weight groups. Hispanics, a group in whom SIDS has not been well studied, had rates similar to those in whites. The overall black/white SIDS ratio in this Study was ~ 4:1, the largest such ratio reported. To our knowledge, earlier investigators have not attempted to quantitate the contribution of birth weight to the observed racial disparities in SIDS incidence. We found that some 27% of the blackwhite SIDS rate difference could be attributed to the different birth weight distributions in these two populations. Earlier studies have demonstrated similarities in
Volume 108 Number 2
white and Hispanic neonatal and postneonatal mortality rates,~6. ~7 but SIDS rate similarities have not been previously reported. The strong relationship between birth weight and SIDS risk, which is confirmed by our data, may provide important clues in the search for an underlying mechanism for SIDS. Theoretically, the association between birth weight and SIDS could be explained by any one of three basic mechanisms: (1) LBW (or, more precisely, some component of incomplete maturation associated with LBW) causes SIDS, (2) LBW causes some other condition and this condition causes SIDS, or (3) some other condition causes both LBW and SIDS. Arguments favoring the first theory, a primary role of LBW in causing SIDS, are based on increased incidence of disordered respiratory control among preterm infants ("apnea of prematurity"), ~8as well as evidence of impaired central respiratory control in SIDS, 19 leading to a general theory that abnormal maturation of central respiratory control causes SIDS. Our data indicate that, although the incidence of SIDS among LBW babies is much higher than among babies of normal birth weight, most SIDS cases (76%) occur among babies whose birth weights are normal. Similarly, 73% of the difference in SIDS incidence between black and white infants is not mediated through differences in birth weight. Thus, although a primary causal relationship between prematurity and SIDS may exist in some cases, our epidemiologic data indicate that this is not the main mechanism operating in the population at large. The second type of mechanism, an indirect causal relationship between LBW and SIDS, has also been proposed. According to this hypothesis, delivery at LBW may cause subtle hypoxic-ischemic injury to the brainstem areas controlling respiration, and such damage causes SIDS. 2~ The appeal of this theory over the direct causal theory is that, although risk of perinatal brain injury goes up sharply at lower birth weights, most cases of hypoxicischemic injury occur in term infants2~; the much larger number of babies born at normal birth weights with relatively low risk outweighs the increased specific risk experienced by a much smaller number of LBW babies. Our results are numerically more compatible with this hypothesis than with one of direct causality, but they provide no specific support for perinatal brain injury as the intervening variable between LBW and SIDS. The other causal mechanism that could explain the association of LBW and SIDS is that both are caused by a third condition. Socioeconomic status has been pointed to as a likely common underlying cause, because it is known to be strongly associated both with SIDS 6,9-11,22,23and with LBW. 12,13,24.25 Although Hispanics in Cook County are
Birth weight, ethnicity, and SIDS
2 13
socioeconomically similar to blacks, the birth weight distribution and SIDS rate among their infants were more comparable to those seen among white infants. The median income for Hispanic households in the Chicago area in 1979 was 73% that of white households; income of blacks was 56% that of whites. Hispanic educational attainment was less than among either whites or blacks/4 According to our third hypothesis of causality, we might have expected Hispanic infants to experience outcomes intermediate between those of whites and blacks. In fact, their incidence of LBW and of SIDS was very similar to rates among white births. For the smaller weight categories, this lack of difference in rates could have resulted from the lack of statistical power in the comparisons that involved small numbers of neonatal survivors, but for the babies weighing more than 2500 gm or for all weights combined, the likelihood of missing an important difference is quite small. Does the similarity of outcomes for whites and Hispanics indicate then that socioeconomic status is not an underlying cause of SIDS and of LBW? We think not. More likely, it suggests that median income and education, traditional indicators of socioeconomic status, have different significance in black and Hispanic populations. Future studies that examine black-Hispanic differences in factors that might mediate the effects of social class on health outcomes could prove quite revealing in this regard. Such factors include access to medical care, environmental stress, social support systems, diet, cigarette smoking, and many others. What are the implications of this study for management of SIDS risk? With the increasing use of home monitors, diagnostic tests to delineate abnormal cardiorespiratory control, and respiratory stimulant drugs in infants thought to be at high risk of SIDS, 26 there is considerable interest in prospectively identifying groups at risk for SIDS who might benefit from further evaluation or treatment. Our findings indicate that more attention should be devoted to screening of small black infants, because being both LBW and black increases SIDS risk. Similarly, clinical trials to determine the efficacy of various interventions must be controlled for birth weight and race. Although populationwide preventive programs specific for SIDS are not yet possible on the basis of current knowledge, our data indicate that programs that reduce the incidence of LBW, such as the Special Supplemental Program for Women, Infants, and Children,27,28 might reduce SIDS incidence in some high-risk groups by a substantial amount, translating into avoidance of hundreds of SIDS deaths annually. Such efforts are clearly of major importance at present, even as we continue to pursue various leads toward a basic understanding and prevention of SIDS.
2 14
B l a c k et al.
We thank Drs. Richard Cooper and Leo Estrada for their comments, and the Division of Data Processing and Vital Records, Illinois Department of Public Health, for data tapes. REFERENCES 1. Bergman AB, Beckwith JB, Ray GC. Sudden infant death syndrome: proceedings of the second international conference on causes of sudden death in infants. Seattle: University of Washington Press, 1970. 2. Bergman AB, Ray CG, Pomeroy MA, et al: Studies of the sudden infant death syndrome in King County, Washington. Ill. Epidemiology. Pediatrics 1972;49:860-870. 3. Kraus JF, Frante CE, Borhani NO. Discriminatory risk factors in post-neonatal sudden unexplained death. Am J Epidemiol 1972;96:328-333. 4. Arsenault PS. Maternal and antenatal factors in the risk of sudden infant death syndrome. Am J Epidemiol 1980; 111:279-284. 5. Spiers PS, Wang L. Short pregnancy interval, low birth weight, and the sudden infant death syndrome. Am J Epidemiol 1976;104:15-21. 6. Standfast S J, Jereb S, Zanerich DT. The epidemiology of sudden infant death in Upstate New York. lI. Birth characteristics. Am J Public Health 1980;70:1061-t064. 7. Kraus JF, Borhani NO. Post-neonatal sudden unexplained death in California: a cohort study. Am J Epidemiol 1972; 95:497-510. 8. Yount JE, Flanagan W J, Dingley EF, et al. Evidence for an exponentially increasing incidence of sudden infant death syndrome with decreasing birth weight [abstr]. Pediatr Res 1979;13:510. 9. Blok JH. The incidence of sudden infant death syndrome in North Carolina's cities and counties: 1972-1974. Am J Public Health 1978;68:367-372. 10. Valdes-Dapena M, Bule LJ, McGovern JA, et al. Sudden unexpected death in infancy: a statistical analysis of certain socioeconomic factors. J PEDIATR 1968;73:387-394. 11. Naeye RL, Ladis B, Drage JS. Sudden infant death syndrome--a prospective study. Am J Dis Child 1976;130:12071210. 12. Kessner DM, Singer J, Kalk CE, et al. Infant death: an analysis by maternal risk and health care. Washington D.C.: Institute of Medicine, National Academy of Sciences, 1973. 13. Placek P. Maternal and infant health factors associated with low infant birth weight: findings from the 1972 national natality survey. In: Reed DM, Stanley FJ, eds. The epidemiology of prematurity. Baltimore: Urban & Sehwarzenberg, 1977.
The Journal o f Pediatrics February 1986
14. U.S. Department of Commerce, Bureau of the Census. Census of population, 1980. Vol. I. Characteristics of the population, part D-15: Illinois. Government Printing Office, 1983. 15. Fleiss JL. Statistical methods for rates and proportions, 2nd ed. New York: John Wiley & Sons, 1981:244. 16. Powell-Griner E, Streck D. A closer examination of neonatal mortality rates among Texas Spanish surname population. Am J Public Health 1982;72:993-999. 17. Selby ML, Lee ES, Tuttle DM, et al. Validity of the Spanish surname infant mortality rate as a health status indicator for the Mexican American population. Am J Public Health 1984;74:998-1002. 18. Henderson-Smart D J, Pettigrew AG, Campbell DJ. Clinical apnea and brainstem neural dysfunction in preterm infants. N Engl J Med 1983;308:353-357. 19. Shannon DC, Kelly DH. SIDS and near-SIDS. N Engl J Med 1982;306:959-1028. 20. Kelly DH, Shannon DC. Epidemiology clues to the etiology of SIDS. Am J Public Health 1980;70:1047-1048. 21. Stanley FJ. An epidemiological study of cerebral palsy in Western Australia, 1956-1975. I. Changes in total incidence of cerebral palsy and associated factors. Dev Med Child Neurol 1979;21:701-713. 22. Froggatt P, Lynas MA, Mackenzie G. Epidemiology of sudden unexpected death in infants in Northern Ireland. Br J Prev Soc Med 1971;25:119-134. 23. Sorensen-Biering F, Jorgensen T, Hilden J. Sudden infant death in Copenhagen 1956q971. 11. Social factors and morbidity. Aeta Paediatr Stand 1979;68:1-9. 24. Brooks CH. Social, economic, and biologic correlates of infant mortality in city neighborhoods. J Health Soc Behav 1980;21:2-11. 25. Kehrer BH, Wolin CM. Impact of income maintenance on low birth weight: evidence from the Gary Experiment. J Hum Resour 1979;14:434-462. 26. Hunt CE, Brouillette RT, Hanson D. Theophylline improves pneumogram abnormalities in infants at risk for SIDS. J PEDIATR 1983:103:967-974. 27. Kotelehuck M, Schwarz JB, Anderka MT, et al. WIC participation and pregnancy outcomes: Massachusetts statewide evaluation project. Am J Public Health 1984;74:10861092. 28. Berkenfield J, Schwartz JB. Nutrition intervention in the community--the,"WIC" program. N Engl J Med 1980; 302:579-581.
on
Sudden infant death syndrome occurs with increased frequency in low birth weight infants and in black infants. The degree to which the higher LBW rate among blacks might explain this higher SIDS rate is unknown. To address this question, we analyzed the 4233 SIDS deaths that occurred among 252,376 neonatal survivors in C o o k County from 1975 to 1980, using computer-coded matched infant birth and death records. Birth weight and ethnic group w e r e identified. The overall SIDS rates in blacks, Hispanics, and whites were 5.1, i.2, and 1.3/1000 neonatal survivors, respectively. Within e a c h ethnic group, the SIDS rates increased progressively with decreasing birth weight. Within the _< 1500 gm birth weight groups, the SIDS rates w e r e 46.4, 3.9, and 5.5/1000 neonatal survivors in blacks, Hispanics, and whites. Using direct standardization, we found that 27% of the SIDS rate disparity between blacks and whites could be explained by the higher LBW rate in blacks (14% vs 6% in whites). The g o o d outcomes in both LBW and SIDS rates for the Hispanic population were unexpected because, like blacks, Hispanics are socioeconomically disadvantaged. Findings for this group suggest that the remaining 73% of the increased SIDS rate in blacks cannot be attributed in a straightforward manner to differences in income or educational attainment. (J PEDIATR1986;108:209-214) Lehman Black, M.D., Richard J. David, M.D,, Robert T. Brouillette, M.D., a n d Carl E. Hunt, M.D. From the Department of Pediatrics, Children's Memorial Hospital, and the Center for Health Services and Policy Research, Northwestern University, Chicago
Sudden infant death syndrome remains the single leading cause of infant death beyond the neonatal period, accounting for nearly one half of all deaths from 28 days to 1 year of age. Since the convening of the Second International Congress on Sudden Infant Death S3}ndrome in 1969,1 several epidemiologic risk factors have been identified, but little reduction in the SIDS rate has occurred. Several studies have shown an increased risk for SIDS in low birth
Supported in part by the Children's Research Guild and the Lillian S. Wells Foundation. Presented in part at the Midwest Society for Pediatric Research, November 1983. Submitted for publication March 4, 1985; accepted July 24, 1985. Reprint requests: Richard J. David, M.D., Children's Memorial Hospital, 2300 Children's Plaza, Chicago, IL 60614.
weight babies, especially in the group weighing <1500 gm28; one preliminary report showed the SIDS risk to increase exponentially as birth weight decreased. 8 There is also an increased risk for SIDS in black infants.3,4,7,9-11 Because the LBW rate is significantly higher in blacks than in whites) 2,13 we examined the contribution of LBW LBW SIDS VLBW
Low birth weight Sudden infant death syndrome Verylow birth weight
to the racial disparity in SIDS rate between blacks and whites. We also examined the birth weight distribution and SIDS rates in Hispanics, the second largest ethnic minority in the United States and one having many socioeconomic risk factors. To assess the interrelationships between birth weight, race, and SIDS risk, we have examined a large
209
2 10
B l a c k et al.
The Journal o f Pediatrics February 1986
Table. SIDS rates in Cook County for black, white, and Hispanic infants, according to birth weight Birth weight group
_<1500 gm 1501-2500 gm >2500 gm All weights
Ethnic group
Neonatal survivors
SIDS deaths
SIDS rate (per 1000)
P
Black White Hispanic Black White Hispanic Black White Hispanic Black White Hispanic
2,477 1,096 260 19,286 10,944 2,470 150,177 214,123 51,573 171,910 226,163 54,303
40 6 1 209 33 13 629 250 52 878 289 66
16.3 5.5 3.8 10.8 3.0 5.3 4.2 1.2 1.0 5.1 1.3 1.2
0.0083* 0.6818~" 0.1164~ <0.001" 0.0836"~ 0.0093~ <0.0001 * 0.3371t <0.0001~: <0.0001 * 0.7039~ <0.0001:~
*Compared with white. tC0mPared with Hispanic. $c0mpared with black. postneonatal population in an ethnically diverse metropolitan area. i
METHODS , Population at risk. The population of Cook County, Illinois, in 1980 was ~ 5.2 million, with an ethnic composition that included 3.0 million white non-Hispanics, 1.3 million blacks, and 500,000 Hispanics.~4 Between 1975 and !980, the number of live births per year ranged between 84,000 and 86,000. Included in our study population were all infants born during the years 1975 through 1980 to mothers residing in Cook County. Deaths were analyzed for members of this cohort who died in the county between the ages of 28 and 364 days. Data sources. We obtained tapes from the Illinois Department of Public Health containing live birth files and matched infant birth-death files for the years 1975 through 1980. These annual birth-death files contain a record for each infant of the birth c o h o r t year w h o died before reaching 1 year of age. These consolidated records contain data from the infant's birth certificate and death certificate, even though these may have been filed in different years. Names and other identifying data are not included. We used the cause of death code 795.0 ("sudden death, cause unknown") from the International Classification of Diseases, eighth revision, for deaths occurring during the years 1975 through 1978 and 798.0 ("sudden infant death syndrome, cot death, crib death") from the ninth revision for the years 1979 and 1980. The autopsy rate for cases reported as S1DS was >95% for each ethnic group. Infants were divided into three different ethnic groups: black, white, and Hispanic, on the basis of the variables "infant race" and "maternal country of origin" derived from the birth record. Hispanic ethnicity as such was not
coded in the Illinois vital records before 1978 and had to be created by using South America, Central America, Cuba, Mexico, and Puerto Rico as the countries of maternal origin. Hispanic was further divided into black-Hispanic and white-Hispanic, but; because of the Small number of births in the former group (0.09%), we included only white-Hispanic in this study. Computations. w e generated a cross-t~tbulation of neonatal deaths, postneonatal SIDS deaths, and postneonatal deaths from all remainingcauses by 500 gm birth weight interVals from 500 to 5500 gm. This was done for each ethnic grbup in' each of the 6 years studied. SiDS rates were cal6ulated from the 6-year totals as follows: SIDS rate = (No. of SIDS infants/Neonatal survivors) • 1000. Neonatal survivors were infants >28 days and <1 year of age. The number Of neonatal survivors was calculated by subtracting neonatal deaths from live births. Live births by weight and ethni c group were obtained from a crosstabulation of all 1977 Cook County resident live births. Because o f the large number of cases in each cell of this b i r t h matrix per year, only 1 year's birth data were examined; taken from the midpoint of the study period. We then multiplied the number of 1977 live births by 6 to obtain an estimate of the 6-year total Of live births. From that estimated birth total and the yearly neonatal death totals, we calculated the number of neonatal survivors at risk for SIDS. Our calculations of the overall SIDS rate for Cook County included only blacks, whites, and Hispanics. All other ethnic groups were excluded because they were from a variety of" ethnic backgrounds and comprised only 3% of all Cook County SIDS deaths in these 6 years. We used the technique of direct standardization to examine the effect of birth weight distribution on the SIDS rate in the black population. 15 To do this we multipled each
Volume 108 Number 2
Birth weight, ethnicity, and S I D S
211
48
35
g 38
D-
25
,.z, ~e
g L
,8 m 5
B
g
g
g
g
g
g
g
g
g
u3
BIRTH
WEIGHT
GROUP
Fig. I. Frequency distributions for 1977 live births in Cook County by 500 gm birth weight categories for black ( - - ) , white (---), and Hispanic (---) ethnic groups.
12
CE
o I 8 w 0_ w
G
nt ~o
4
2
B
i
i
m ~
I
i
u1 a/
N
BIRTH
m
WEIGHT
I
i
i
m
N
to
m
~-
r
i
to
GROUP
Fig. 2. SIDS rate as a function of birth weight for all black, white, and Hispanic infants combined (Cook County infants aged 28 to 364 days, 1975 through 1980).
black birth weight-specific SIDS rate times the proportion of births in that weight category, assuming the same birth weight distribution as actually observed in the white population. Summing across birth weight categories then yielded an overall "expected" black SIDS rate. The difference between this expected rate and the actual black SIDS rate rePresents the frequency by which the black SIDS rate would be reduced if blacks had the same birth weight distribution as whites. Dividing this number by the difference between the actual black and actual white SIDS rates yields the percent of SIDS rate differences in the black and white populations explained by differences in birth weight distribution between the two populations.
RESULTS There were 1233 SIDS deaths among the 252,376 neonatal survivors in Cook County between 1975 and 1980. The overall SIDS rate for the 6 Years of this study was 2.7/1000 neonatal survivors. The overall SIDS rates for the black, white, and Hispanic infants were 5.1, 1.3, and 1.2, respectively (Table). The black SIDS rate was 4 times the white SIDS rate; the Hispanic rate did not differ significantly from that in white infants. Ethnic group differences in SIDS rates were also observed for moderatelY low birth weight (1501 to 2500 gm) and very low birth weight (__<1500 gm) groupS. In the moderately LBW group the black SIDS rate was 3.6 times higher than the white SIDS rate; among VLBW births the black rate was nearly
2 12
B l a c k et al.
The Journal o f Pediatrics February 1986
18
Q Z E U] D O -r tE bJ n Ld Frr n,
15
12
9
6
O3
i (g
{}
BIRTH
WEIGHT
GROUP
Fig. 3. SIDS rates as a function of birth weight for black ( - - ) , white (---), and Hispanic (---) ethnic groups (Cook County infants aged 28 to 364 days, t975 through 1980).
3 times the white rate. Again, the differences between white and Hispanic infants were not statistically significant. Birth weight distributions. Fig. 1 shows the three ethnic group birth weight distribution curves for 1977 live births in Cook County. The nearly identical shape of the white and Hispanic curves was reflected in similar rates of LBW for the two populations: 5.9% and 5.6%, respectively. The black LBW rate was 13.8%, more than twice the rate in either white or Hispanic infants. A similar relationship held for the VLBW group: the white an d Hispanic rates were identical at 0.4%, whereas the black rate was more than 6 times as high at 2.5%. Relation of SIDS to birth weight. Fig. 2 presents the overall b i r t h weight-specific SIDS mortality rates for Cook County. The SIDS rate was increased among survivors of lower birth weight, a trend that became most evident in the < 1500 gm weight category. The relationship between SIDS rate and birth weight is best described by an exponential model: Rate = 31.5e -~ (r =0.97, P <0.0001). Birth weight-specific SIDS mortality curves for the black, white, and Hispanic populations are depicted in Fig. 3. White and HisPanic curves were similar, at least above 2500 gm, but the black SIDS rates were higher throughout virtually the entire range of birth weights reported. The black and white rate curves are described better b y an exponential than by a linear model, but the coefficients (reflecting magnitude of birth weight effect) vary considerably. The Hispanic curve is fitted slightly better bY a linear model. Standardizing for birth weight differences between eth-
nic groups, we calculated that the black SIDS rate would be reduced from its actual level of 5.11/1000 to 4.08/1000 if the black infants had the same birth weight distribution as the white population. That is, 27% of the difference between the black and white SIDS rates can be explained by the difference in birth weight distributions in the black and white infants. DISCUSSION We have shown that the risk of SIDS increases exponentially with decreasing birth weight, with qualitatively similar birth weight-risk relationships for the three ethnic groups studied. The exponential relationship has been reported in one previous study, but those authors examined a relatively small, homogeneous white population? In contrast, our study population contains the largest number of SIDS cases yet analyzed and has enough ethnic diversity to permit confirmation of the exponential relationship not only for the population as a whole but also for individual ethnic groups. Comparisons between ethnic groups revealed a large black-white discrepancy in SIDS rates for nearly all birth weight groups. Hispanics, a group in whom SIDS has not been well studied, had rates similar to those in whites. The overall black/white SIDS ratio in this Study was ~ 4:1, the largest such ratio reported. To our knowledge, earlier investigators have not attempted to quantitate the contribution of birth weight to the observed racial disparities in SIDS incidence. We found that some 27% of the blackwhite SIDS rate difference could be attributed to the different birth weight distributions in these two populations. Earlier studies have demonstrated similarities in
Volume 108 Number 2
white and Hispanic neonatal and postneonatal mortality rates,~6. ~7 but SIDS rate similarities have not been previously reported. The strong relationship between birth weight and SIDS risk, which is confirmed by our data, may provide important clues in the search for an underlying mechanism for SIDS. Theoretically, the association between birth weight and SIDS could be explained by any one of three basic mechanisms: (1) LBW (or, more precisely, some component of incomplete maturation associated with LBW) causes SIDS, (2) LBW causes some other condition and this condition causes SIDS, or (3) some other condition causes both LBW and SIDS. Arguments favoring the first theory, a primary role of LBW in causing SIDS, are based on increased incidence of disordered respiratory control among preterm infants ("apnea of prematurity"), ~8as well as evidence of impaired central respiratory control in SIDS, 19 leading to a general theory that abnormal maturation of central respiratory control causes SIDS. Our data indicate that, although the incidence of SIDS among LBW babies is much higher than among babies of normal birth weight, most SIDS cases (76%) occur among babies whose birth weights are normal. Similarly, 73% of the difference in SIDS incidence between black and white infants is not mediated through differences in birth weight. Thus, although a primary causal relationship between prematurity and SIDS may exist in some cases, our epidemiologic data indicate that this is not the main mechanism operating in the population at large. The second type of mechanism, an indirect causal relationship between LBW and SIDS, has also been proposed. According to this hypothesis, delivery at LBW may cause subtle hypoxic-ischemic injury to the brainstem areas controlling respiration, and such damage causes SIDS. 2~ The appeal of this theory over the direct causal theory is that, although risk of perinatal brain injury goes up sharply at lower birth weights, most cases of hypoxicischemic injury occur in term infants2~; the much larger number of babies born at normal birth weights with relatively low risk outweighs the increased specific risk experienced by a much smaller number of LBW babies. Our results are numerically more compatible with this hypothesis than with one of direct causality, but they provide no specific support for perinatal brain injury as the intervening variable between LBW and SIDS. The other causal mechanism that could explain the association of LBW and SIDS is that both are caused by a third condition. Socioeconomic status has been pointed to as a likely common underlying cause, because it is known to be strongly associated both with SIDS 6,9-11,22,23and with LBW. 12,13,24.25 Although Hispanics in Cook County are
Birth weight, ethnicity, and SIDS
2 13
socioeconomically similar to blacks, the birth weight distribution and SIDS rate among their infants were more comparable to those seen among white infants. The median income for Hispanic households in the Chicago area in 1979 was 73% that of white households; income of blacks was 56% that of whites. Hispanic educational attainment was less than among either whites or blacks/4 According to our third hypothesis of causality, we might have expected Hispanic infants to experience outcomes intermediate between those of whites and blacks. In fact, their incidence of LBW and of SIDS was very similar to rates among white births. For the smaller weight categories, this lack of difference in rates could have resulted from the lack of statistical power in the comparisons that involved small numbers of neonatal survivors, but for the babies weighing more than 2500 gm or for all weights combined, the likelihood of missing an important difference is quite small. Does the similarity of outcomes for whites and Hispanics indicate then that socioeconomic status is not an underlying cause of SIDS and of LBW? We think not. More likely, it suggests that median income and education, traditional indicators of socioeconomic status, have different significance in black and Hispanic populations. Future studies that examine black-Hispanic differences in factors that might mediate the effects of social class on health outcomes could prove quite revealing in this regard. Such factors include access to medical care, environmental stress, social support systems, diet, cigarette smoking, and many others. What are the implications of this study for management of SIDS risk? With the increasing use of home monitors, diagnostic tests to delineate abnormal cardiorespiratory control, and respiratory stimulant drugs in infants thought to be at high risk of SIDS, 26 there is considerable interest in prospectively identifying groups at risk for SIDS who might benefit from further evaluation or treatment. Our findings indicate that more attention should be devoted to screening of small black infants, because being both LBW and black increases SIDS risk. Similarly, clinical trials to determine the efficacy of various interventions must be controlled for birth weight and race. Although populationwide preventive programs specific for SIDS are not yet possible on the basis of current knowledge, our data indicate that programs that reduce the incidence of LBW, such as the Special Supplemental Program for Women, Infants, and Children,27,28 might reduce SIDS incidence in some high-risk groups by a substantial amount, translating into avoidance of hundreds of SIDS deaths annually. Such efforts are clearly of major importance at present, even as we continue to pursue various leads toward a basic understanding and prevention of SIDS.
2 14
B l a c k et al.
We thank Drs. Richard Cooper and Leo Estrada for their comments, and the Division of Data Processing and Vital Records, Illinois Department of Public Health, for data tapes. REFERENCES 1. Bergman AB, Beckwith JB, Ray GC. Sudden infant death syndrome: proceedings of the second international conference on causes of sudden death in infants. Seattle: University of Washington Press, 1970. 2. Bergman AB, Ray CG, Pomeroy MA, et al: Studies of the sudden infant death syndrome in King County, Washington. Ill. Epidemiology. Pediatrics 1972;49:860-870. 3. Kraus JF, Frante CE, Borhani NO. Discriminatory risk factors in post-neonatal sudden unexplained death. Am J Epidemiol 1972;96:328-333. 4. Arsenault PS. Maternal and antenatal factors in the risk of sudden infant death syndrome. Am J Epidemiol 1980; 111:279-284. 5. Spiers PS, Wang L. Short pregnancy interval, low birth weight, and the sudden infant death syndrome. Am J Epidemiol 1976;104:15-21. 6. Standfast S J, Jereb S, Zanerich DT. The epidemiology of sudden infant death in Upstate New York. lI. Birth characteristics. Am J Public Health 1980;70:1061-t064. 7. Kraus JF, Borhani NO. Post-neonatal sudden unexplained death in California: a cohort study. Am J Epidemiol 1972; 95:497-510. 8. Yount JE, Flanagan W J, Dingley EF, et al. Evidence for an exponentially increasing incidence of sudden infant death syndrome with decreasing birth weight [abstr]. Pediatr Res 1979;13:510. 9. Blok JH. The incidence of sudden infant death syndrome in North Carolina's cities and counties: 1972-1974. Am J Public Health 1978;68:367-372. 10. Valdes-Dapena M, Bule LJ, McGovern JA, et al. Sudden unexpected death in infancy: a statistical analysis of certain socioeconomic factors. J PEDIATR 1968;73:387-394. 11. Naeye RL, Ladis B, Drage JS. Sudden infant death syndrome--a prospective study. Am J Dis Child 1976;130:12071210. 12. Kessner DM, Singer J, Kalk CE, et al. Infant death: an analysis by maternal risk and health care. Washington D.C.: Institute of Medicine, National Academy of Sciences, 1973. 13. Placek P. Maternal and infant health factors associated with low infant birth weight: findings from the 1972 national natality survey. In: Reed DM, Stanley FJ, eds. The epidemiology of prematurity. Baltimore: Urban & Sehwarzenberg, 1977.
The Journal o f Pediatrics February 1986
14. U.S. Department of Commerce, Bureau of the Census. Census of population, 1980. Vol. I. Characteristics of the population, part D-15: Illinois. Government Printing Office, 1983. 15. Fleiss JL. Statistical methods for rates and proportions, 2nd ed. New York: John Wiley & Sons, 1981:244. 16. Powell-Griner E, Streck D. A closer examination of neonatal mortality rates among Texas Spanish surname population. Am J Public Health 1982;72:993-999. 17. Selby ML, Lee ES, Tuttle DM, et al. Validity of the Spanish surname infant mortality rate as a health status indicator for the Mexican American population. Am J Public Health 1984;74:998-1002. 18. Henderson-Smart D J, Pettigrew AG, Campbell DJ. Clinical apnea and brainstem neural dysfunction in preterm infants. N Engl J Med 1983;308:353-357. 19. Shannon DC, Kelly DH. SIDS and near-SIDS. N Engl J Med 1982;306:959-1028. 20. Kelly DH, Shannon DC. Epidemiology clues to the etiology of SIDS. Am J Public Health 1980;70:1047-1048. 21. Stanley FJ. An epidemiological study of cerebral palsy in Western Australia, 1956-1975. I. Changes in total incidence of cerebral palsy and associated factors. Dev Med Child Neurol 1979;21:701-713. 22. Froggatt P, Lynas MA, Mackenzie G. Epidemiology of sudden unexpected death in infants in Northern Ireland. Br J Prev Soc Med 1971;25:119-134. 23. Sorensen-Biering F, Jorgensen T, Hilden J. Sudden infant death in Copenhagen 1956q971. 11. Social factors and morbidity. Aeta Paediatr Stand 1979;68:1-9. 24. Brooks CH. Social, economic, and biologic correlates of infant mortality in city neighborhoods. J Health Soc Behav 1980;21:2-11. 25. Kehrer BH, Wolin CM. Impact of income maintenance on low birth weight: evidence from the Gary Experiment. J Hum Resour 1979;14:434-462. 26. Hunt CE, Brouillette RT, Hanson D. Theophylline improves pneumogram abnormalities in infants at risk for SIDS. J PEDIATR 1983:103:967-974. 27. Kotelehuck M, Schwarz JB, Anderka MT, et al. WIC participation and pregnancy outcomes: Massachusetts statewide evaluation project. Am J Public Health 1984;74:10861092. 28. Berkenfield J, Schwartz JB. Nutrition intervention in the community--the,"WIC" program. N Engl J Med 1980; 302:579-581.