- Email: [email protected]
Use of psychotropic drugs and pregnancy outcome
J Clin Epidemiol Vol. 49, No. 11, pp. 1309-1313, 1996 Copyright 0 1996 Elsevier Science Inc.
0895-4356/96/$15.00 PII SO895-4356(96)00203-X ELSEVIER
PHARMACOEPIDEMIOLOGY
Use of Psychotropic
Drugs and Pregnancy
Pekka Larivaara, ’ Anna-Lisa ‘DEPARTMENT
OF PUBLIC
HEALTH
‘DEPARTMENT
SCIENCE
OF OBSTETRICS
REPORTS
AND
Hartikainen, ‘z2and Paula Rantakallio’ GENERAL
AND
Outcome
PRACTICE,
GYNAECOLOGY,
UNIVERSITY UNIVERISITY
OF OULU, OF OULW,
OULU,
FINLAND,
AND
FINLAND
ABSTRACT. This study reports the use of psychotropic drugs and pregnancy outcome in a prospective survey carried out in northern Finland in 1985-1986, consisting of 7933 pregnant women and their 8030 births. Of the mothers, 120 (1.50/ o ) used psychotropic drugs during pregnancy and of them 26 (0.3% of the total) used these drugs before and during the pregnancy. According to the logistic regression analysis, the biological and social background variables associating significantly with the use of these drugs were maternal advanced age and multiparity (>35 years and >4 parous), overweight (body mass index >the 90th percentile), smoking, alcohol use, belonging to social class IV, and failure to ensure contraception. The users needed hospital observation significantly more often during pregnancy (regular users, 80.8%; occasional users, 38.3; nonusers, 27.4%) and the adjustment of the background variables did not change this result. Of the pregnancy complications bleeding was significantly more common among users than nonusers (23.3 vs. 13.2%) and this difference was not explained by diff erence in background variables. The mean birth weight of infants of the regular users was significantly lower (255 g) when background variables were standardized by linear regression analysis. No association of increased number of birth defects was found with usage of psychotropic drugs. J CLIN EPIDEMIOL 49;11:1309-1313, 1996. KEY
WORDS.
Birth weight,
malformation,
mental
One of the problems for any physician who becomes involved in the care of pregnant women is drug therapy and drug safety. It has been estimated that 3% of all newborns have congenital birth defects [l] and only 3% of them may be caused by drugs and environmental chemicals [2]. Most of the developmental defects are of unknown origin. Undoubtedly psychotropic drugs may have a number of other effects on fetuses and newborns, but they have been less discussed and are difficult to diagnose. In Finland 7% of women use prescribed psychotropic drugs. The usage increases with advancing age. Of the youngest women (1534 years) 1.5% use them. The corresponding figure in the 35- to 44-year-old group is 6.3% and women 45 years of age or more have had an 11 times higher level of use than the youngest women [3]. This prospective survey was carried out to investigate the usage of psychotropic drugs before and during pregnancy, to characterize the uses and course of pregnancy, and to examine the effect of these drugs on fetuses and newborns. MATERIALS
AND
METHODS
The dara belong to a l-year birth cohort in the two most northern provinces in Finland, collected prospectively during pregnancy. The membership of the cohort was defined by the expected date of delivery falling between July 1, 1985 and June 30, 1986. The cohort consisted of 9362 pregnant women and covered 99% of all deliveries in that region during the study period [4,5]. The pregnant women completed two questionnaires in maternity health centers, the first up to week 24 of gestation and the second ar the end of the pregnancy, before or after delivery. The first qurstionnaire contained 42 items about demographic, biological, and
Address reprint reqrrests to: Pekka Larivaara, Department of Public Health Science and General Pracrice, University of Zulu, Aapisne 1, FIN-90220 Zulu, Fmland. Accepted for publication on 30 April 1996.
illness, pregnancy
outcome,
psychotropic
drugs
socioeconomic characteristics of the women and their families including the use of drugs, tobacco, and alcohol; these questions were also asked in the second questionnaire. Both prescription and nonprescription drugs were included. The second questionnaire contained 23 items about the mothers’ health during pregnancy. The nurses helped the mothers to fill in the questionnaire and ensure that all questions were answered. The third questionnaire, which contained 52 items about the pregnancy, the delivery, and the newborn baby, was completed in the maternity hospital by the midwives. Psychiatric hospital diagnoses in 1972-1986 were collected from a national hospital discharge register, which covers the data for 95100% of all hospital admissions in the country [6]. The diagnoses were established both in general and mental hospitals but diagnoses given at outpatient visits are not included in the register. Of the study population, 85% answered both first and second questionnaires. These 7933 mothers and their 8030 infants are analyzed in this article. Data on various demographic variables were missing in 0.1 to 11.5% of the cases. Socioeconomic status was classified by the woman’s occupation (I and II, professional; III, skilled worker; lV, unskilled worker; and V, farmer [7], examples of which are given in an earlier paper [S]. In the statistical analyses of the data the chi-square and Student’s t tests were used. Logistic regression analysis was used to obtain risk estimates for the use of psychotropic drugs during pregnancy. The effect of psychotropic drugs on the birth weight was determined by linear regression analysis.
RESULTS One hundred and twenty mothers (12 1 children) who used psychotropic drugs during pregnancy were identified, representing 1.5% of all pregnant women in our study population, 7813 women (7909 children) serving as a control group (nonusers). Of them, 26 (27 children), 0.3% of the total, used these drugs regularly before and during the pregnancy (regular users). Another 94 women (94 children) among the psychotropic drug users took these drugs for short
1310 TABLE
P. Larivaara 1. Biological
characteristics
and social
background
variables
Regular Maternal
variable
Age (years) <2Q 20-35 >35 Parity 0 l-3 <3 Previous induced abortions Yes NO Contraception in use before pregnancy Yes No Body mass index < 10th percentile Yes NO >90th percentile Yes NO Social class I + II III IV Farmers School education Primary school More than primary school Unemployed or no disability pension Yes NO Marital status Married Cohahiting Unmarried, widowed, or divorced Smoking after second month of pregnancy Yes No Drinking during pregnancy” Yes No
users
of psychotropic
drug users and nonusers
Occasional
%
N
0 17 9
0.0 65.4 34.5
1 67 26
6 15 5
23.1 57.7 19.2
7 19
et al.
users
%
Nonusers
N
%
1.1 71.3 27.7
326 6701 786
4.2 85.8
23 54 17
24.5 57.5 18.1
2654 4540 607
34.0 58.2 7.8
26.9 73.1
14 80
14.9 85.1
867 6934
11.1 88.9
6 18
25.0 75.0
40 52
43.5 56.6
3732 3989
48.3 51.7
0 25
0.0 100.0
6 85
6.6 93.4
725 6944
9.6 90.4
8 17
32.0 68.0
17 74
18.7 81.3
784 6895
10.2 89.8
2 8 13 0
8.7 34.8 56.5 0.0
27 34 26 6
29.0 36.6 28.0 6.5
1950 3082 2184 414
25.6 40.4 28.6 5.4
9 10
47.4 52.6
31 58
34.8 65.2
2042 5689
26.4 73.6
6 16
27.3 72.7
7 82
7.9 92.1
615 6964
8.1 91.9
16 5 5
61.5 19.2 19.2
80 11 3
85.1 11.7 3.2
6324 1110 379
80.9 14.2 4.9
17 9
65.4 34.6
18 76
19.2 80.9
1715 6093
22.0 78.0
10 13
43.5 56.5
24 70
25.5 74.5
1099 6612
14.3 85.8
N
-
-
-
“The itcln containcJ the quesrion: “Have you used alcohol during the pregnancy?” “Y es or no.” Exact quantities haven’t heen taken into consideration.
perioda or only occasionally (occasional users). Of the 120 women, 70 (58%) used psychotropic drugs during the first 16 weeks of pregnancy. Most of the psychotropic drugs used were tranquilizers and hypnotics (n = 79, 74.2%). Neuroleptics were prescribed for 21 mothers (17.5%), antidepressants for 17 (14.2%), and lithium for only 1 mother. Psychostimulants and sleeping pills had been used in six cases (4.2%). The biological characteristics and social background variables of the psychotropic drug users and nonusers are presented in Table 1. Use of psychotropic drugs increased significantly along with age. The mean age of the regular users was 33.6 (SD 4.9) years, of occasional users was 31.6 (SD 5.7) years, and of nonusers was 27.7 (SD 5.5) years. The users of psychotropic drugs were more often multipa-
rious, obese, had had more induced ahortions, and were less frequent users of contraception than the nonusers. The regular users were clearly more often unmarried or cohabiting. Similarly they were more often smokers and alcohol users than the others. The trend of increased alcohol usage was also seen among occasional users. The psychiatric diagnoses of the mothers are presented in Table 2. Of the regular users, 15 women (57.7%) had heen admitted to a hospital because of psychiatric illness. Hospital admission with psychiatric diagnosis was less frequent in the group of the occasional users and only a few had had such an illness among nonusers. From the background variables described in Table 1, logistic regression analysis differentiated six variables that had a significant association with the regular use of psychotropic drugs (Table 3). Smoking and drinking during pregnancy, age over 35 years com-
Use of Psychotropic
Drugs and Pregnancy
1311
Outcome
TABLE 2. Psychiatric diagnoses given at hospital to the users and nonusers of psychotropic drugs Regular users (N= 26)
admissions
Nonusers (N= 7813)
USCXS
(N=
94)
N
%
N
%
N
%
No diagnoses Diagnoses (total)’ Psychosis Neurosis Mental retardation
11 15 9 12 I
42.3 57.7 34.6 46.2 3.8
85 9 4 6 -
90.4 9.6 4.3 6.4
7723 90
98.8 1.2 0.3 0.9 0.1
;: 6
‘The sane patient may have several diagnoses.
bined with multiparity, body mass index over the 90th percentile, and social class IV associated with increased use of these drugs regularly, and the usage of contraceptives before the pregnancy associated with an opposite effect. Age over 35 years (OR 2.86, 95% CI 1.71-4.78) and alcohol drinking (OR 2.45, 95% CI 1.52-3.96), associated with occasional usage. With respect to different pregnancy complications the users of psychotropic drugs had more frequent bleeding than the nonusers (N = 28 vs. N = 1031; 23.3 vs. 13.2%; p = 0.001). Among the users, 22 (18.3%) had taken the drugs before bleeding and in 12 cases the medicine had been used during the first trimester. Of the other users, five women had the hleeding first and the drugs had been prescribed later. One had placenta previa and the bleeding occurred during the last weeks of pregnancy. From the total nine variables the logistic regression analysis differentiated three that had a significant association with the bleeding (Table 4). Both groups of users of psychotropic drugs had significant and independent association with the bleeding. Of the regular users of psychotropic drugs, 23 (88.5%) had visited outpatient clinics of delivery hospitals during pregnancy. The corresponding figures for occasional users and nonusers were 76 (80.9%) and 5694 (72.9%), respectively. These differences were not significant. The regular users needed hospital observation in 21 cases (80.8%) as compared with 2 139 cases (27.4%) among the nonusers (p < 0.001). Also, the occasional users needed hospital care more than nonusers did (N = 36; 38.3%; p = 0.025). Logistic regression analysis differentiated five variables (among the total of nine) that had a significant association with hospital admissions (Table 5). Regular use of psychotropic drugs was the most powerful variable that was associated with admissions.
TABLE 3. Odds ratios of background variables in logistic ree gression analysis on the regular use of psychotropic drugs dure ing pregnancy variable
Smoking Age and parity (235 years and 24 parous) Social class IV Drinking Body mass index (~90th percentile) Usage of contraceptives hefore the pregnancy
4. Odds ratios of background analysis on bleeding during
Occasional
Psychiatric diagnoses
Explanatory
TABLE gression
Odds ratio
95% Confidence interval
6.13
2.52-14.90
5.72 2.93 2.80
2.43-13.46 1.31-6.56 1.19-6.57
2.90
1.05-8.04
0.35
0.14-0.89
Explanatory
variable
Parity 24 Regular use of psychotropic drugs Occasional use of psychotropic drugs Mother’s age 235 years Social class IV Mother has a psychiatric diagnosis Smoking Drinking BMI >90th percentile
variables pregnancy Odds ratio
in logistic
re-
95% Confidence interval
0.5 4.2
0.4-0.7 1.7-10.4
1.9 1.2 0.9 0.6 0.9 1.1 0.9
1.1-3.1 1.0-1.5 0.8-1.0 0.3-1.2 0.8-1.1 0.9-l.) 0.7-1.2
The mean birth weight was 3055 g (SD 758) among the regular users, 3527 g (SD 528) among the occasional users, and 3544 g (SD 571) among the nonusers. The difference between the regular users and nonusers was significant (P = 0.012). When the background variables presented in Table 1, bleeding during pregnancy, and the data on the frequency of marriages and the gestational week at birth were standardized, the estimated difference in hirth weight of singletons between the regular users and nonusers diminished to 255 g, hut remained significant (p < 0.001). The percentage of low birth weight infants (<2500 g) was significantly higher among the regular users and that of preterm births almost significantly higher (Table 6). The infants of regular users were also admitted more frequently to the newborn nursery during the neonatal period than the others, the figures being 8 infants (29.6%) for regular users, 13 infants (13.8%) for occasional users, and 1124 infants (14.2%) for the nonusers. When the admission rate was studied by logistic regression analysis and the nine background variables mentioned ahove were adjusted, the maternal use of psychotropic drugs had, however, no significant association with the admission rate. Congenital anomalies were found in three newborns of the users of psychotropic drugs. A twin pair among the regular users had inguinal hernias and one of them also had a clubfoot. One newhorn of the occasional users had Down’s syndrome with a congenital heart anomaly. The number of congenital anomalies among the newborns of the users was 3
TABLE 5. Odds ratios of background variables in logistic regression analysis on the mother’s hospital admissions during pregnancy Explanatory
variable
Bleeding during pregnancy drugs Regular use of psychotropic Mother’s age 235 years BMI >90th percentile Parity 24 Mother has a psychiatric diagnosis Occasional use of psychotropic drugs Drinking Smoking Social class IV
Odds ratio
95% Confidence interval
2.9 7.8 1.4 1.4 0.8 1.5
2.5-3.3 2.8-21.6 1.2-1.6 1.1-1.7 0.7-1.0 1 .o-2.2
1.4 0.9 0.9 1.0
0.9-2.2 0.8-1.1 0.8-1.1 0.9-1.1
1312
P. Larivaara
TABLE 6. Rate of preterm and nonusers of psychotropic
births, low birth drugs Regular
Pregnancy
outcome
Preterm birth (~37 weeks) Birth weight (~2500 g) SGA (< 10th percentile)
weight,
and smallness
of infants
users (N = 27)
relative
Occasional
to gestational
users (N = 94)
age (SGA)
among
et al. users
Nonusers (N= 7909)
N
%
p Value a
N
%
p Valuea
N
%
4 5 4
14.8 18.5 14.8
0.078 0.001 0.592
4 5 8
4.3 5.3 8.5
0.824 0.705 0.789
420 317 781
5.3 4.0 9.9
“e Values testing rhe difference for nonusers.
(2.5%, 95% CI 0.5-7.1) and among those of the nonusers was 246 (3.1%, 95% CI 2.7-3.5), the figures thus showing no significant difference between the groups. Of the 71 newborns who were exposed to psychotropic drugs during the first 16 gestational weeks, 2 had a birth defect (the twins mentioned above), the percentage being 2.8 (95% CI 0.3-9.8). There were no perinatal deaths among the regular users and only one stillbirth among the occasional users, perinatal mortality being 0.8 in the study group and 0.9% in the group of the nonusers. The only mother who had used lithium had taken it during the first trimester. The course of her pregnancy was normal and the newborn had no perinatal problems. DISCUSSION About 10% of pregnant women used psychotropic drugs during pregnancy in Finland in 1965 [9]. S’mce then the publicity aimed at unnecessary medication during pregnancy had affected both the mothers and the health personnel. As a result of that the usage of tranquilizers and analgesics has had a continuous decreasing trend in Finland [9]. In the United States as late as the 1970s very high rates of use were reported, e.g., in Nashville, Tennessee 13% of pregnant women received psychotropic drugs, most frequently diazepam
[lOI. Our results on the usage of psychotropic drugs during pregnancy (1.5%) corresponded to those earlier findings and to the usage of prescribed psychotropic drugs among the youngest women (15-34 years of age) in Finland in 1976 [3]. At the end of 1983, 2.68 per 1000 of the Finnish population were admitted to mental hospitals because of schizophrenia or other psychoses; 0.63 per 1000 were admitted because of other psychiatric diseases. Among women in those hospitals 83.8% had been diagnosed as having a psychosis and 16.2% of them had other diagnoses [l 11. In our study group more nonpsychotic diagnoses were found, which presumably means that most sick women with psychotic disease become pregnant less frequently than those suffering from less severe mental disorders. The most frequently prescribed drugs in our study were tranquilizers and hypnotics. Brocklebank and co-workers observed some years earlier that during pregnancy in most cases the prescribed minor tranqutlizer was dlazepam, accounting for 86% of the women who received anxiolytic drugs [lo]. The results concerning the biological characteristics and social hackground of the users of psychorropic drugs show that these mothers are mostly older and multiparous compared to nonusers. They are socially multiprohlematic; despite helonging frequently to the fourth social class there is a trend to lower school education, unemployment or to being on a disability pension. These women have also more often vague conjugal relationships with their partners and
they neglect contraception. Smoking and drinking are common among them, too. In Sandahl’s study in Sweden it has been found similarly that older pregnant women use psychotropic drugs more frequently than do younger women. Their results showed no correlation hetween the use of these drugs and smoking habits [12]. Of the pregnancy complications bleeding was notable, hut although the users of psychotropic drugs more frequently had bleeding during pregnancy than did nonusers evidence about a marked association between the drugs and bleeding remained obscure. However, the users also had many other problems during the course of pregnancy, because they had been admitted to maternity hospital in several cases. The mean birth weight of the newhorns of the regular users was smaller than that of the nonusers when background variables influencing birth weight such as smoking, drinking, parity, and gestational week were taken into account. The study gives no answer to the question as to whether the drugs have a direct growthretarding effect or whether there are some further effective variables that have not appeared in this survey. The infants of the regular users also had a tendency to be born prematurely. Hypotonia or floppy infant syndrome, seen earlier in the infants of mothers consuming more than 30 mg of henzodiazepines per day [13], was not detected among the infants of our survey. The usage of large amount of benzodiazepines per day [13], was not detected among the infants of our survey. The usage of large amount of henzodiazepines may also lead to the consequences of withdrawal symptoms in a newborn with tremor, irritahllity, hyperactivity, hypotonicity, tachypnea, and voracious feeding [14]. Sandahl’s study in Sweden included 2436 women, 74 of whom were exposed to psychotropic drugs [12]. One stillhorn and five (6.8%) malformed infants were horn among these women. Among the users in our study group only one still birth was verified, and this outcome gives us no right to conclude that mortality arises from the usage of psychotropic drugs during pregnancy. Our malformation rates among the users were lower than those in Sandahl’s study and did not differ from nonusers. Early reports as to the increased incidence of cleft lip and palate in infants of first trimester abuse of benzodiazepines have not been substantiated by more recent reports [15,16]. Our results confirm these findings. On the basis of our outcomes we conclude that the use of psychotropic drugs during pregnancy may harm both the mother and the mfant and therefore a continuous decreasing trend in the usage is favorable. On the other hand, in case of using no psychotropic drugs, it is assumed that problems may appear among mothers who suffet from psychiatric problems during pregnancy.
References 1. Kalter H, Warkany J. G)ngemtnl malformations. 1. Etiologic factors and their role in prevention. N Engl J Med 1983; 108: 424-431.
Use of Psychotropic
Drugs and Pregnancy
Outcome
2. Golhur M. Teratology for the obstetrician. Curr Status Gynecol 1980; 55: 269-277. 3. Riska E, Klaukka K. Use of psychotropic drugs in Finland. Sot Sci Med 1984; 19: 983-989. 4. Rantakallio P, Oja H. I’ermaral risk for infants of unmarried mothers over a period of 20 years. Early Hum Dev 1990; 22: 157-169. 5. Hartikainen-Surri A-L, Rantakalho P, Sip& P. Changes III progmws of twin hirths over 20 years. Ann Med 1990; 22: 131&135. 6. An) S, Koskinen R, Keskungkl I. Sairaalastapoistorekisterln dlagnoosi-, toimenpldeja tapaturmatietojen luotettavuus. Duodecim 1990; 21: 1443-1450. 7. Sosiaaliryhmitys. Helsingin kaupungin tilastotolmlsto. Helsinki 1954. 8. Rantakallio P. Social hackground of mothers who smoke during pregnancy and Influence of these factors on the offspring. Sot Sci Med 1979; 13A: 423-429. 9. Harjulehto T, An) T, Sax& L. Long-term changes in medication during pregnancy. Teratology 1988; 37: 145-148.
1313 10. Brocklehank JC, Ray WA, Feclerspiel CF, Schaffner W. Drug prescrihmg durmg pregnancy. A controlled study of Tennessee Medicaid recipeents. Am J Obstet Gynecol 1978; 132: 235. II. Hakkarainen A. Changes in the Psychiatric Patients Population: A Theoretlcal Model and Its Apphcation in Finland. Helsinki: National Board of Health in Finland; 1989. 12. Sandahl B. A prospectwe study of drug use, smoking and contraceptives during early pregnancy. Acta Obstet Gynecol Stand 1985; 64: 381386. 13. Lee RV. Drug abuse. In: Burrow GN, Ferris TF, eds. Medical Complications during Pregnancy. Philadelphia: W.B. Saunders; 1982: 538-549. 14. Guillozet N. Drug risks in pregnancy revisited. J Fam Pratt 1977; 4: 1043-1052. 15. Cordero JF, Oakley GP. Drug exposure dung pregnancy: Some epdemiologic considerations. Clin Obstet Gynecol 1982; 26: 418-428. 16. Rosenberg L, Mitchell H, Parseus J. Lack of relation of oral clefts to dlazepam use during pregnancy. N Engl J Med 1983; 309: 1282-1285.
0895-4356/96/$15.00 PII SO895-4356(96)00203-X ELSEVIER
PHARMACOEPIDEMIOLOGY
Use of Psychotropic
Drugs and Pregnancy
Pekka Larivaara, ’ Anna-Lisa ‘DEPARTMENT
OF PUBLIC
HEALTH
‘DEPARTMENT
SCIENCE
OF OBSTETRICS
REPORTS
AND
Hartikainen, ‘z2and Paula Rantakallio’ GENERAL
AND
Outcome
PRACTICE,
GYNAECOLOGY,
UNIVERSITY UNIVERISITY
OF OULU, OF OULW,
OULU,
FINLAND,
AND
FINLAND
ABSTRACT. This study reports the use of psychotropic drugs and pregnancy outcome in a prospective survey carried out in northern Finland in 1985-1986, consisting of 7933 pregnant women and their 8030 births. Of the mothers, 120 (1.50/ o ) used psychotropic drugs during pregnancy and of them 26 (0.3% of the total) used these drugs before and during the pregnancy. According to the logistic regression analysis, the biological and social background variables associating significantly with the use of these drugs were maternal advanced age and multiparity (>35 years and >4 parous), overweight (body mass index >the 90th percentile), smoking, alcohol use, belonging to social class IV, and failure to ensure contraception. The users needed hospital observation significantly more often during pregnancy (regular users, 80.8%; occasional users, 38.3; nonusers, 27.4%) and the adjustment of the background variables did not change this result. Of the pregnancy complications bleeding was significantly more common among users than nonusers (23.3 vs. 13.2%) and this difference was not explained by diff erence in background variables. The mean birth weight of infants of the regular users was significantly lower (255 g) when background variables were standardized by linear regression analysis. No association of increased number of birth defects was found with usage of psychotropic drugs. J CLIN EPIDEMIOL 49;11:1309-1313, 1996. KEY
WORDS.
Birth weight,
malformation,
mental
One of the problems for any physician who becomes involved in the care of pregnant women is drug therapy and drug safety. It has been estimated that 3% of all newborns have congenital birth defects [l] and only 3% of them may be caused by drugs and environmental chemicals [2]. Most of the developmental defects are of unknown origin. Undoubtedly psychotropic drugs may have a number of other effects on fetuses and newborns, but they have been less discussed and are difficult to diagnose. In Finland 7% of women use prescribed psychotropic drugs. The usage increases with advancing age. Of the youngest women (1534 years) 1.5% use them. The corresponding figure in the 35- to 44-year-old group is 6.3% and women 45 years of age or more have had an 11 times higher level of use than the youngest women [3]. This prospective survey was carried out to investigate the usage of psychotropic drugs before and during pregnancy, to characterize the uses and course of pregnancy, and to examine the effect of these drugs on fetuses and newborns. MATERIALS
AND
METHODS
The dara belong to a l-year birth cohort in the two most northern provinces in Finland, collected prospectively during pregnancy. The membership of the cohort was defined by the expected date of delivery falling between July 1, 1985 and June 30, 1986. The cohort consisted of 9362 pregnant women and covered 99% of all deliveries in that region during the study period [4,5]. The pregnant women completed two questionnaires in maternity health centers, the first up to week 24 of gestation and the second ar the end of the pregnancy, before or after delivery. The first qurstionnaire contained 42 items about demographic, biological, and
Address reprint reqrrests to: Pekka Larivaara, Department of Public Health Science and General Pracrice, University of Zulu, Aapisne 1, FIN-90220 Zulu, Fmland. Accepted for publication on 30 April 1996.
illness, pregnancy
outcome,
psychotropic
drugs
socioeconomic characteristics of the women and their families including the use of drugs, tobacco, and alcohol; these questions were also asked in the second questionnaire. Both prescription and nonprescription drugs were included. The second questionnaire contained 23 items about the mothers’ health during pregnancy. The nurses helped the mothers to fill in the questionnaire and ensure that all questions were answered. The third questionnaire, which contained 52 items about the pregnancy, the delivery, and the newborn baby, was completed in the maternity hospital by the midwives. Psychiatric hospital diagnoses in 1972-1986 were collected from a national hospital discharge register, which covers the data for 95100% of all hospital admissions in the country [6]. The diagnoses were established both in general and mental hospitals but diagnoses given at outpatient visits are not included in the register. Of the study population, 85% answered both first and second questionnaires. These 7933 mothers and their 8030 infants are analyzed in this article. Data on various demographic variables were missing in 0.1 to 11.5% of the cases. Socioeconomic status was classified by the woman’s occupation (I and II, professional; III, skilled worker; lV, unskilled worker; and V, farmer [7], examples of which are given in an earlier paper [S]. In the statistical analyses of the data the chi-square and Student’s t tests were used. Logistic regression analysis was used to obtain risk estimates for the use of psychotropic drugs during pregnancy. The effect of psychotropic drugs on the birth weight was determined by linear regression analysis.
RESULTS One hundred and twenty mothers (12 1 children) who used psychotropic drugs during pregnancy were identified, representing 1.5% of all pregnant women in our study population, 7813 women (7909 children) serving as a control group (nonusers). Of them, 26 (27 children), 0.3% of the total, used these drugs regularly before and during the pregnancy (regular users). Another 94 women (94 children) among the psychotropic drug users took these drugs for short
1310 TABLE
P. Larivaara 1. Biological
characteristics
and social
background
variables
Regular Maternal
variable
Age (years) <2Q 20-35 >35 Parity 0 l-3 <3 Previous induced abortions Yes NO Contraception in use before pregnancy Yes No Body mass index < 10th percentile Yes NO >90th percentile Yes NO Social class I + II III IV Farmers School education Primary school More than primary school Unemployed or no disability pension Yes NO Marital status Married Cohahiting Unmarried, widowed, or divorced Smoking after second month of pregnancy Yes No Drinking during pregnancy” Yes No
users
of psychotropic
drug users and nonusers
Occasional
%
N
0 17 9
0.0 65.4 34.5
1 67 26
6 15 5
23.1 57.7 19.2
7 19
et al.
users
%
Nonusers
N
%
1.1 71.3 27.7
326 6701 786
4.2 85.8
23 54 17
24.5 57.5 18.1
2654 4540 607
34.0 58.2 7.8
26.9 73.1
14 80
14.9 85.1
867 6934
11.1 88.9
6 18
25.0 75.0
40 52
43.5 56.6
3732 3989
48.3 51.7
0 25
0.0 100.0
6 85
6.6 93.4
725 6944
9.6 90.4
8 17
32.0 68.0
17 74
18.7 81.3
784 6895
10.2 89.8
2 8 13 0
8.7 34.8 56.5 0.0
27 34 26 6
29.0 36.6 28.0 6.5
1950 3082 2184 414
25.6 40.4 28.6 5.4
9 10
47.4 52.6
31 58
34.8 65.2
2042 5689
26.4 73.6
6 16
27.3 72.7
7 82
7.9 92.1
615 6964
8.1 91.9
16 5 5
61.5 19.2 19.2
80 11 3
85.1 11.7 3.2
6324 1110 379
80.9 14.2 4.9
17 9
65.4 34.6
18 76
19.2 80.9
1715 6093
22.0 78.0
10 13
43.5 56.5
24 70
25.5 74.5
1099 6612
14.3 85.8
N
-
-
-
“The itcln containcJ the quesrion: “Have you used alcohol during the pregnancy?” “Y es or no.” Exact quantities haven’t heen taken into consideration.
perioda or only occasionally (occasional users). Of the 120 women, 70 (58%) used psychotropic drugs during the first 16 weeks of pregnancy. Most of the psychotropic drugs used were tranquilizers and hypnotics (n = 79, 74.2%). Neuroleptics were prescribed for 21 mothers (17.5%), antidepressants for 17 (14.2%), and lithium for only 1 mother. Psychostimulants and sleeping pills had been used in six cases (4.2%). The biological characteristics and social background variables of the psychotropic drug users and nonusers are presented in Table 1. Use of psychotropic drugs increased significantly along with age. The mean age of the regular users was 33.6 (SD 4.9) years, of occasional users was 31.6 (SD 5.7) years, and of nonusers was 27.7 (SD 5.5) years. The users of psychotropic drugs were more often multipa-
rious, obese, had had more induced ahortions, and were less frequent users of contraception than the nonusers. The regular users were clearly more often unmarried or cohabiting. Similarly they were more often smokers and alcohol users than the others. The trend of increased alcohol usage was also seen among occasional users. The psychiatric diagnoses of the mothers are presented in Table 2. Of the regular users, 15 women (57.7%) had heen admitted to a hospital because of psychiatric illness. Hospital admission with psychiatric diagnosis was less frequent in the group of the occasional users and only a few had had such an illness among nonusers. From the background variables described in Table 1, logistic regression analysis differentiated six variables that had a significant association with the regular use of psychotropic drugs (Table 3). Smoking and drinking during pregnancy, age over 35 years com-
Use of Psychotropic
Drugs and Pregnancy
1311
Outcome
TABLE 2. Psychiatric diagnoses given at hospital to the users and nonusers of psychotropic drugs Regular users (N= 26)
admissions
Nonusers (N= 7813)
USCXS
(N=
94)
N
%
N
%
N
%
No diagnoses Diagnoses (total)’ Psychosis Neurosis Mental retardation
11 15 9 12 I
42.3 57.7 34.6 46.2 3.8
85 9 4 6 -
90.4 9.6 4.3 6.4
7723 90
98.8 1.2 0.3 0.9 0.1
;: 6
‘The sane patient may have several diagnoses.
bined with multiparity, body mass index over the 90th percentile, and social class IV associated with increased use of these drugs regularly, and the usage of contraceptives before the pregnancy associated with an opposite effect. Age over 35 years (OR 2.86, 95% CI 1.71-4.78) and alcohol drinking (OR 2.45, 95% CI 1.52-3.96), associated with occasional usage. With respect to different pregnancy complications the users of psychotropic drugs had more frequent bleeding than the nonusers (N = 28 vs. N = 1031; 23.3 vs. 13.2%; p = 0.001). Among the users, 22 (18.3%) had taken the drugs before bleeding and in 12 cases the medicine had been used during the first trimester. Of the other users, five women had the hleeding first and the drugs had been prescribed later. One had placenta previa and the bleeding occurred during the last weeks of pregnancy. From the total nine variables the logistic regression analysis differentiated three that had a significant association with the bleeding (Table 4). Both groups of users of psychotropic drugs had significant and independent association with the bleeding. Of the regular users of psychotropic drugs, 23 (88.5%) had visited outpatient clinics of delivery hospitals during pregnancy. The corresponding figures for occasional users and nonusers were 76 (80.9%) and 5694 (72.9%), respectively. These differences were not significant. The regular users needed hospital observation in 21 cases (80.8%) as compared with 2 139 cases (27.4%) among the nonusers (p < 0.001). Also, the occasional users needed hospital care more than nonusers did (N = 36; 38.3%; p = 0.025). Logistic regression analysis differentiated five variables (among the total of nine) that had a significant association with hospital admissions (Table 5). Regular use of psychotropic drugs was the most powerful variable that was associated with admissions.
TABLE 3. Odds ratios of background variables in logistic ree gression analysis on the regular use of psychotropic drugs dure ing pregnancy variable
Smoking Age and parity (235 years and 24 parous) Social class IV Drinking Body mass index (~90th percentile) Usage of contraceptives hefore the pregnancy
4. Odds ratios of background analysis on bleeding during
Occasional
Psychiatric diagnoses
Explanatory
TABLE gression
Odds ratio
95% Confidence interval
6.13
2.52-14.90
5.72 2.93 2.80
2.43-13.46 1.31-6.56 1.19-6.57
2.90
1.05-8.04
0.35
0.14-0.89
Explanatory
variable
Parity 24 Regular use of psychotropic drugs Occasional use of psychotropic drugs Mother’s age 235 years Social class IV Mother has a psychiatric diagnosis Smoking Drinking BMI >90th percentile
variables pregnancy Odds ratio
in logistic
re-
95% Confidence interval
0.5 4.2
0.4-0.7 1.7-10.4
1.9 1.2 0.9 0.6 0.9 1.1 0.9
1.1-3.1 1.0-1.5 0.8-1.0 0.3-1.2 0.8-1.1 0.9-l.) 0.7-1.2
The mean birth weight was 3055 g (SD 758) among the regular users, 3527 g (SD 528) among the occasional users, and 3544 g (SD 571) among the nonusers. The difference between the regular users and nonusers was significant (P = 0.012). When the background variables presented in Table 1, bleeding during pregnancy, and the data on the frequency of marriages and the gestational week at birth were standardized, the estimated difference in hirth weight of singletons between the regular users and nonusers diminished to 255 g, hut remained significant (p < 0.001). The percentage of low birth weight infants (<2500 g) was significantly higher among the regular users and that of preterm births almost significantly higher (Table 6). The infants of regular users were also admitted more frequently to the newborn nursery during the neonatal period than the others, the figures being 8 infants (29.6%) for regular users, 13 infants (13.8%) for occasional users, and 1124 infants (14.2%) for the nonusers. When the admission rate was studied by logistic regression analysis and the nine background variables mentioned ahove were adjusted, the maternal use of psychotropic drugs had, however, no significant association with the admission rate. Congenital anomalies were found in three newborns of the users of psychotropic drugs. A twin pair among the regular users had inguinal hernias and one of them also had a clubfoot. One newhorn of the occasional users had Down’s syndrome with a congenital heart anomaly. The number of congenital anomalies among the newborns of the users was 3
TABLE 5. Odds ratios of background variables in logistic regression analysis on the mother’s hospital admissions during pregnancy Explanatory
variable
Bleeding during pregnancy drugs Regular use of psychotropic Mother’s age 235 years BMI >90th percentile Parity 24 Mother has a psychiatric diagnosis Occasional use of psychotropic drugs Drinking Smoking Social class IV
Odds ratio
95% Confidence interval
2.9 7.8 1.4 1.4 0.8 1.5
2.5-3.3 2.8-21.6 1.2-1.6 1.1-1.7 0.7-1.0 1 .o-2.2
1.4 0.9 0.9 1.0
0.9-2.2 0.8-1.1 0.8-1.1 0.9-1.1
1312
P. Larivaara
TABLE 6. Rate of preterm and nonusers of psychotropic
births, low birth drugs Regular
Pregnancy
outcome
Preterm birth (~37 weeks) Birth weight (~2500 g) SGA (< 10th percentile)
weight,
and smallness
of infants
users (N = 27)
relative
Occasional
to gestational
users (N = 94)
age (SGA)
among
et al. users
Nonusers (N= 7909)
N
%
p Value a
N
%
p Valuea
N
%
4 5 4
14.8 18.5 14.8
0.078 0.001 0.592
4 5 8
4.3 5.3 8.5
0.824 0.705 0.789
420 317 781
5.3 4.0 9.9
“e Values testing rhe difference for nonusers.
(2.5%, 95% CI 0.5-7.1) and among those of the nonusers was 246 (3.1%, 95% CI 2.7-3.5), the figures thus showing no significant difference between the groups. Of the 71 newborns who were exposed to psychotropic drugs during the first 16 gestational weeks, 2 had a birth defect (the twins mentioned above), the percentage being 2.8 (95% CI 0.3-9.8). There were no perinatal deaths among the regular users and only one stillbirth among the occasional users, perinatal mortality being 0.8 in the study group and 0.9% in the group of the nonusers. The only mother who had used lithium had taken it during the first trimester. The course of her pregnancy was normal and the newborn had no perinatal problems. DISCUSSION About 10% of pregnant women used psychotropic drugs during pregnancy in Finland in 1965 [9]. S’mce then the publicity aimed at unnecessary medication during pregnancy had affected both the mothers and the health personnel. As a result of that the usage of tranquilizers and analgesics has had a continuous decreasing trend in Finland [9]. In the United States as late as the 1970s very high rates of use were reported, e.g., in Nashville, Tennessee 13% of pregnant women received psychotropic drugs, most frequently diazepam
[lOI. Our results on the usage of psychotropic drugs during pregnancy (1.5%) corresponded to those earlier findings and to the usage of prescribed psychotropic drugs among the youngest women (15-34 years of age) in Finland in 1976 [3]. At the end of 1983, 2.68 per 1000 of the Finnish population were admitted to mental hospitals because of schizophrenia or other psychoses; 0.63 per 1000 were admitted because of other psychiatric diseases. Among women in those hospitals 83.8% had been diagnosed as having a psychosis and 16.2% of them had other diagnoses [l 11. In our study group more nonpsychotic diagnoses were found, which presumably means that most sick women with psychotic disease become pregnant less frequently than those suffering from less severe mental disorders. The most frequently prescribed drugs in our study were tranquilizers and hypnotics. Brocklebank and co-workers observed some years earlier that during pregnancy in most cases the prescribed minor tranqutlizer was dlazepam, accounting for 86% of the women who received anxiolytic drugs [lo]. The results concerning the biological characteristics and social hackground of the users of psychorropic drugs show that these mothers are mostly older and multiparous compared to nonusers. They are socially multiprohlematic; despite helonging frequently to the fourth social class there is a trend to lower school education, unemployment or to being on a disability pension. These women have also more often vague conjugal relationships with their partners and
they neglect contraception. Smoking and drinking are common among them, too. In Sandahl’s study in Sweden it has been found similarly that older pregnant women use psychotropic drugs more frequently than do younger women. Their results showed no correlation hetween the use of these drugs and smoking habits [12]. Of the pregnancy complications bleeding was notable, hut although the users of psychotropic drugs more frequently had bleeding during pregnancy than did nonusers evidence about a marked association between the drugs and bleeding remained obscure. However, the users also had many other problems during the course of pregnancy, because they had been admitted to maternity hospital in several cases. The mean birth weight of the newhorns of the regular users was smaller than that of the nonusers when background variables influencing birth weight such as smoking, drinking, parity, and gestational week were taken into account. The study gives no answer to the question as to whether the drugs have a direct growthretarding effect or whether there are some further effective variables that have not appeared in this survey. The infants of the regular users also had a tendency to be born prematurely. Hypotonia or floppy infant syndrome, seen earlier in the infants of mothers consuming more than 30 mg of henzodiazepines per day [13], was not detected among the infants of our survey. The usage of large amount of benzodiazepines per day [13], was not detected among the infants of our survey. The usage of large amount of henzodiazepines may also lead to the consequences of withdrawal symptoms in a newborn with tremor, irritahllity, hyperactivity, hypotonicity, tachypnea, and voracious feeding [14]. Sandahl’s study in Sweden included 2436 women, 74 of whom were exposed to psychotropic drugs [12]. One stillhorn and five (6.8%) malformed infants were horn among these women. Among the users in our study group only one still birth was verified, and this outcome gives us no right to conclude that mortality arises from the usage of psychotropic drugs during pregnancy. Our malformation rates among the users were lower than those in Sandahl’s study and did not differ from nonusers. Early reports as to the increased incidence of cleft lip and palate in infants of first trimester abuse of benzodiazepines have not been substantiated by more recent reports [15,16]. Our results confirm these findings. On the basis of our outcomes we conclude that the use of psychotropic drugs during pregnancy may harm both the mother and the mfant and therefore a continuous decreasing trend in the usage is favorable. On the other hand, in case of using no psychotropic drugs, it is assumed that problems may appear among mothers who suffet from psychiatric problems during pregnancy.
References 1. Kalter H, Warkany J. G)ngemtnl malformations. 1. Etiologic factors and their role in prevention. N Engl J Med 1983; 108: 424-431.
Use of Psychotropic
Drugs and Pregnancy
Outcome
2. Golhur M. Teratology for the obstetrician. Curr Status Gynecol 1980; 55: 269-277. 3. Riska E, Klaukka K. Use of psychotropic drugs in Finland. Sot Sci Med 1984; 19: 983-989. 4. Rantakallio P, Oja H. I’ermaral risk for infants of unmarried mothers over a period of 20 years. Early Hum Dev 1990; 22: 157-169. 5. Hartikainen-Surri A-L, Rantakalho P, Sip& P. Changes III progmws of twin hirths over 20 years. Ann Med 1990; 22: 131&135. 6. An) S, Koskinen R, Keskungkl I. Sairaalastapoistorekisterln dlagnoosi-, toimenpldeja tapaturmatietojen luotettavuus. Duodecim 1990; 21: 1443-1450. 7. Sosiaaliryhmitys. Helsingin kaupungin tilastotolmlsto. Helsinki 1954. 8. Rantakallio P. Social hackground of mothers who smoke during pregnancy and Influence of these factors on the offspring. Sot Sci Med 1979; 13A: 423-429. 9. Harjulehto T, An) T, Sax& L. Long-term changes in medication during pregnancy. Teratology 1988; 37: 145-148.
1313 10. Brocklehank JC, Ray WA, Feclerspiel CF, Schaffner W. Drug prescrihmg durmg pregnancy. A controlled study of Tennessee Medicaid recipeents. Am J Obstet Gynecol 1978; 132: 235. II. Hakkarainen A. Changes in the Psychiatric Patients Population: A Theoretlcal Model and Its Apphcation in Finland. Helsinki: National Board of Health in Finland; 1989. 12. Sandahl B. A prospectwe study of drug use, smoking and contraceptives during early pregnancy. Acta Obstet Gynecol Stand 1985; 64: 381386. 13. Lee RV. Drug abuse. In: Burrow GN, Ferris TF, eds. Medical Complications during Pregnancy. Philadelphia: W.B. Saunders; 1982: 538-549. 14. Guillozet N. Drug risks in pregnancy revisited. J Fam Pratt 1977; 4: 1043-1052. 15. Cordero JF, Oakley GP. Drug exposure dung pregnancy: Some epdemiologic considerations. Clin Obstet Gynecol 1982; 26: 418-428. 16. Rosenberg L, Mitchell H, Parseus J. Lack of relation of oral clefts to dlazepam use during pregnancy. N Engl J Med 1983; 309: 1282-1285.