American Journal of Epidemiology Copyright C 1991 by The Johns Hopkins University School of Hygiene and Pubic Health AOJignts reserved
Vol 133, No 8 Printed in USA
Cocaine Use during Pregnancy: Perinatal Outcomes
Arden Handler,1 Naomi Kistin,1 Faith Davis,2 and Cynthia Ferre2
The relation between maternal cocaine use and perinatal outcomes was investigated among 17,466 non-Asian singleton deliveries in 1988 from the University of Illinois Perinatal Network data base in the metropolitan Chicago area. Elevated adjusted relative risks (RR) of low birth weight (RR = 2.8, 95% confidence interval (Cl) 2.2-3.7), prematurity (RR = 2.4, 95% Cl 1.9-3.1), abruptfo placentae (RR = 4.5, 95% Cl 2.48.5), and perinatal death (RR = 2.1, 95% Cl 1.1-4.0) were observed for "any" cocaine users (n = 408) compared with women who did not use cocaine or any other drugs or alcohol (n = 17,058). There was an increased (although unstable) risk of intrapartum placenta previa not previously reported (RR = 2.3, 95% Cl 1.0-5.1). The relative risk of small-for-gestational-age births for cocaine users who did not smoke (RR = 3.4, 95% Cl 1.8-6.5) was greater than that for cocaine users who did (RR = 2.1, 95% Cl 1.14.1). Irrespective of smoking status, cocaine use during pregnancy increased the risk of small-for-gestational-age births. Am J Epidemiol 1991 ;133:818-25. cocaine; infant, small for gestational age; perinatal mortality; placenta praevia; pregnancy outcome; smoking; substance abuse; tobacco use disorder
Cocaine use in the United States has risen to epidemic proportions over the last decade (1). Although the extent of cocaine addiction during the prenatal period is not known, the National Institute on Drug Abuse estimates that 1.96 million women aged 18 years and over used cocaine in 1985, representing a 59 percent increase from 1982 when 1.23 million women were estimated to be users (2).
Received for publication November 27, 1989, and in final form October 22, 1990. Abbreviations O, confidence interval; RR, relative risk. 1 Program in Community Health Sciences, University of Illinois School of Public Health, Chicago, IL. 2 Program in Epidemiology and Biostatisbcs, University of Illinois School of Public Health, Chicago, IL Reprint requests to Dr. Arden Handler, M/C 922, School of Public Health, University of Illinois, Box 6998, Chicago, IL 60680. This research was partially funded by an ASPH-CDC Cooperative Agreement Ms. Ferre was the recipient of a University of Illinois Graduate Fellowship during the completion of this manuscript. The authors thank the University of Illinois Perinatal Management Group for access to the data base. Presented at the Second Annual Meeting of the Society for Pediatric Eptdemiologic Research, Birmingham, Alabama, June 1989.
818
While cocaine use has risen, the use of other illicit substances has remained "stable" or has declined since 1982 (1). Most cocaine users abuse more than one drug, primarily marijuana and/or alcohol (3, 4). Prospective and retrospective studies of prenatal cocaine exposure have identified a variety of adverse maternal, fetal, and neonatal outcomes. These toxic effects include an increased incidence of spontaneous abortion (5-7), fetal death (7-10), prematurity (6, 9-12), abruptio placentae (5, 8-13), and congenital malformations (8, 12). Decreased fetal growth, as evidenced by decreased birth weight (6-8, 10-15), birth length (7, 8, 10, 12, 13, 15), and head circumference (7, 8, 10-13, 15), and an increase in small-forgestational-age births (6, 8, 9, 12), has been documented. Adverse sequelae associated with the neurologic development of infants exposed to cocaine in utero have also been identified (5, 12). Many studies of the impact of maternal cocaine exposure have been hampered by problems of small sample size and difficulties in exposure ascertainment. Insufficient
Cocaine and Perinatal Outcomes
sample size has frequently prevented researchers from properly controlling for confounders through multivariable analysis. This has impeded estimation of the independent effect of cocaine in the presence of other risk factors for adverse perinatal outcomes such as no prenatal care, cigarette smoking, malnutrition, and poverty. Likewise, problems related to both exposure ascertainment and sample size have frustrated efforts to assess the effect of cocaine exposure independent of other drugs during pregnancy. Because of difficulties related to the timing and methods used to collect information about maternal substance abuse and the fact that women who use cocaine rarely use only cocaine, a cocaine exposure group which has been commonly studied is the "any" cocaine group. This group includes women who use only cocaine as well as those who use cocaine and other drugs (or alcohol) (5-7, 13). The present study uses information from a large perinatal registry to assess the effect of exposure to any cocaine on perinatal morbidity and mortality independent of other risk factors: race, age, gravidity, smoking, and prenatal care. The use of a large perinatal registry does not eliminate difficulties related to exposure ascertainment; however, a large sample size does permit an evaluation of the effect of any cocaine use on perinatal outcomes while adjusting for other risk factors.
MATERIALS AND METHODS
A retrospective cohort study was conducted by using data from a large urban perinatal registry. Medical records of all obstetric events at 12 hospitals in the University of Illinois Hospital Perinatal Network were abstracted daily in 1988 by trained personnel employed by the University of Illinois School of Public Health. The perinatal abstract form includes over 450 maternal, fetal, and neonatal variables. Substance abuse variables documented on the perinatal abstract and included in this analysis are antepartum or intrapartum ma-
819
ternal use of cocaine, heroin, marijuana, amphetamines, alcohol, and tobacco cigarettes. Cocaine exposure is recorded on the abstract form if it is noted anywhere in the prenatal or labor and delivery records; it is based on patient self-report or the results of urine toxicology screenings of mothers or neonates. No data are available in the Registry on the dose, route, type of cocaine, or time of administration during pregnancy. Exposures to heroin, marijuana, and amphetamines are ascertained in the same way as exposure to cocaine. Cigarette smoking and alcohol exposures are based on selfreport. Because of the large proportion of cigarette smokers among cocaine users in this study (70 percent), smoking is considered as a confounder in this analysis. Given the possibility of differential ascertainment of cocaine and other drug exposures at the 12 participating hospitals, risk estimates are also adjusted for hospital of delivery. Other risk factors available on the perinatal abstract form and considered as potential confounders are maternal race (black, Hispanic, or white), maternal age (less than 20, 20-29, or 30 years or more), gravidity (primigravid or multigravid), and prenatal care experience (none or some). Selected perinatal outcomes evaluated include low birth weight (2,500 g or less); prematurity (less than 37 weeks gestation); small for gestational age (birth weight less than the 10th percentile of the birth weight distribution of the singleton deliveries included in this analysis at each gestational age); decreased birth length (birth length less than the 10th percentile at each gestational age); decreased head circumference (head circumference less than the 10th percentile at each gestational age); abruptio placentae (yes or no); intrapartum placenta previa (yes or no); the occurrence of any congenital anomaly (yes or no); and perinatal death (antepartum or intrapartum fetal death and neonatal death prior to hospital discharge and without infant readmission). Perinatal deaths are reported as rates per 100 livebirths plus fetal deaths. The percentage of missing values for all variables is relatively small,
820
Handler et al.
TABLE 1. Characteristics of maternal cocaine users and nonusers among 17,466 non-Asian singleton deliveries, Chicago, Illinois, metropolitan area, 1988* Cocaine users (n = 408)
Race Black Hispanic White Age (years) National Institute on Drug Abuse, 1985. (National Institute on Drug Abuse Research Monograph 57). (DHHS no. (ADM) 85-1402). 24. Chasnoff LI, Landress HJ, Barrett ME. The prevalence of illicit-drug or alcohol use during pregnancy and discrepancies in mandatory reporting in Pinellas County, Florida. N Engl J Med 1990;322; 1202-6.
Vol 133, No 8 Printed in USA
Cocaine Use during Pregnancy: Perinatal Outcomes
Arden Handler,1 Naomi Kistin,1 Faith Davis,2 and Cynthia Ferre2
The relation between maternal cocaine use and perinatal outcomes was investigated among 17,466 non-Asian singleton deliveries in 1988 from the University of Illinois Perinatal Network data base in the metropolitan Chicago area. Elevated adjusted relative risks (RR) of low birth weight (RR = 2.8, 95% confidence interval (Cl) 2.2-3.7), prematurity (RR = 2.4, 95% Cl 1.9-3.1), abruptfo placentae (RR = 4.5, 95% Cl 2.48.5), and perinatal death (RR = 2.1, 95% Cl 1.1-4.0) were observed for "any" cocaine users (n = 408) compared with women who did not use cocaine or any other drugs or alcohol (n = 17,058). There was an increased (although unstable) risk of intrapartum placenta previa not previously reported (RR = 2.3, 95% Cl 1.0-5.1). The relative risk of small-for-gestational-age births for cocaine users who did not smoke (RR = 3.4, 95% Cl 1.8-6.5) was greater than that for cocaine users who did (RR = 2.1, 95% Cl 1.14.1). Irrespective of smoking status, cocaine use during pregnancy increased the risk of small-for-gestational-age births. Am J Epidemiol 1991 ;133:818-25. cocaine; infant, small for gestational age; perinatal mortality; placenta praevia; pregnancy outcome; smoking; substance abuse; tobacco use disorder
Cocaine use in the United States has risen to epidemic proportions over the last decade (1). Although the extent of cocaine addiction during the prenatal period is not known, the National Institute on Drug Abuse estimates that 1.96 million women aged 18 years and over used cocaine in 1985, representing a 59 percent increase from 1982 when 1.23 million women were estimated to be users (2).
Received for publication November 27, 1989, and in final form October 22, 1990. Abbreviations O, confidence interval; RR, relative risk. 1 Program in Community Health Sciences, University of Illinois School of Public Health, Chicago, IL. 2 Program in Epidemiology and Biostatisbcs, University of Illinois School of Public Health, Chicago, IL Reprint requests to Dr. Arden Handler, M/C 922, School of Public Health, University of Illinois, Box 6998, Chicago, IL 60680. This research was partially funded by an ASPH-CDC Cooperative Agreement Ms. Ferre was the recipient of a University of Illinois Graduate Fellowship during the completion of this manuscript. The authors thank the University of Illinois Perinatal Management Group for access to the data base. Presented at the Second Annual Meeting of the Society for Pediatric Eptdemiologic Research, Birmingham, Alabama, June 1989.
818
While cocaine use has risen, the use of other illicit substances has remained "stable" or has declined since 1982 (1). Most cocaine users abuse more than one drug, primarily marijuana and/or alcohol (3, 4). Prospective and retrospective studies of prenatal cocaine exposure have identified a variety of adverse maternal, fetal, and neonatal outcomes. These toxic effects include an increased incidence of spontaneous abortion (5-7), fetal death (7-10), prematurity (6, 9-12), abruptio placentae (5, 8-13), and congenital malformations (8, 12). Decreased fetal growth, as evidenced by decreased birth weight (6-8, 10-15), birth length (7, 8, 10, 12, 13, 15), and head circumference (7, 8, 10-13, 15), and an increase in small-forgestational-age births (6, 8, 9, 12), has been documented. Adverse sequelae associated with the neurologic development of infants exposed to cocaine in utero have also been identified (5, 12). Many studies of the impact of maternal cocaine exposure have been hampered by problems of small sample size and difficulties in exposure ascertainment. Insufficient
Cocaine and Perinatal Outcomes
sample size has frequently prevented researchers from properly controlling for confounders through multivariable analysis. This has impeded estimation of the independent effect of cocaine in the presence of other risk factors for adverse perinatal outcomes such as no prenatal care, cigarette smoking, malnutrition, and poverty. Likewise, problems related to both exposure ascertainment and sample size have frustrated efforts to assess the effect of cocaine exposure independent of other drugs during pregnancy. Because of difficulties related to the timing and methods used to collect information about maternal substance abuse and the fact that women who use cocaine rarely use only cocaine, a cocaine exposure group which has been commonly studied is the "any" cocaine group. This group includes women who use only cocaine as well as those who use cocaine and other drugs (or alcohol) (5-7, 13). The present study uses information from a large perinatal registry to assess the effect of exposure to any cocaine on perinatal morbidity and mortality independent of other risk factors: race, age, gravidity, smoking, and prenatal care. The use of a large perinatal registry does not eliminate difficulties related to exposure ascertainment; however, a large sample size does permit an evaluation of the effect of any cocaine use on perinatal outcomes while adjusting for other risk factors.
MATERIALS AND METHODS
A retrospective cohort study was conducted by using data from a large urban perinatal registry. Medical records of all obstetric events at 12 hospitals in the University of Illinois Hospital Perinatal Network were abstracted daily in 1988 by trained personnel employed by the University of Illinois School of Public Health. The perinatal abstract form includes over 450 maternal, fetal, and neonatal variables. Substance abuse variables documented on the perinatal abstract and included in this analysis are antepartum or intrapartum ma-
819
ternal use of cocaine, heroin, marijuana, amphetamines, alcohol, and tobacco cigarettes. Cocaine exposure is recorded on the abstract form if it is noted anywhere in the prenatal or labor and delivery records; it is based on patient self-report or the results of urine toxicology screenings of mothers or neonates. No data are available in the Registry on the dose, route, type of cocaine, or time of administration during pregnancy. Exposures to heroin, marijuana, and amphetamines are ascertained in the same way as exposure to cocaine. Cigarette smoking and alcohol exposures are based on selfreport. Because of the large proportion of cigarette smokers among cocaine users in this study (70 percent), smoking is considered as a confounder in this analysis. Given the possibility of differential ascertainment of cocaine and other drug exposures at the 12 participating hospitals, risk estimates are also adjusted for hospital of delivery. Other risk factors available on the perinatal abstract form and considered as potential confounders are maternal race (black, Hispanic, or white), maternal age (less than 20, 20-29, or 30 years or more), gravidity (primigravid or multigravid), and prenatal care experience (none or some). Selected perinatal outcomes evaluated include low birth weight (2,500 g or less); prematurity (less than 37 weeks gestation); small for gestational age (birth weight less than the 10th percentile of the birth weight distribution of the singleton deliveries included in this analysis at each gestational age); decreased birth length (birth length less than the 10th percentile at each gestational age); decreased head circumference (head circumference less than the 10th percentile at each gestational age); abruptio placentae (yes or no); intrapartum placenta previa (yes or no); the occurrence of any congenital anomaly (yes or no); and perinatal death (antepartum or intrapartum fetal death and neonatal death prior to hospital discharge and without infant readmission). Perinatal deaths are reported as rates per 100 livebirths plus fetal deaths. The percentage of missing values for all variables is relatively small,
820
Handler et al.
TABLE 1. Characteristics of maternal cocaine users and nonusers among 17,466 non-Asian singleton deliveries, Chicago, Illinois, metropolitan area, 1988* Cocaine users (n = 408)
Race Black Hispanic White Age (years) National Institute on Drug Abuse, 1985. (National Institute on Drug Abuse Research Monograph 57). (DHHS no. (ADM) 85-1402). 24. Chasnoff LI, Landress HJ, Barrett ME. The prevalence of illicit-drug or alcohol use during pregnancy and discrepancies in mandatory reporting in Pinellas County, Florida. N Engl J Med 1990;322; 1202-6.
