Azathioprine, mercaptopurine and birth outcome: a population-based cohort study
Dr B. Nørgård, Department of Clinical Epidemiology, Vennelyst Boulevard 6, Aarhus University Hospital, DK-8000 Aarhus C, Denmark. E-mail: firstname.lastname@example.org
Background : Data on the safety of azathioprine and mercaptopurine during pregnancy are very sparse.
Aim : To examine the risk of adverse birth outcomes in women who took up prescriptions for azathioprine or mercaptopurine during pregnancy.
Methods : This is a Danish cohort study based on data from a population-based prescription registry, the Danish Birth Registry and the Hospital Discharge Registry. To examine the risk of congenital malformations, we included nine pregnancies exposed 30 days before conception or during the first trimester. To examine perinatal mortality, pre-term birth and low birth weight, we included 10 pregnancies exposed during the entire pregnancy. Eleven different exposed women were included in the study. Outcomes were compared with those of 19 418 pregnancies in which no drugs were prescribed to the mothers.
Results : Fifty-five per cent of the exposed women had inflammatory bowel disease and 45% other diseases. Adjusted odds ratios for congenital malformations, perinatal mortality, pre-term birth and low birth weight were 6.7 (95% confidence interval, 1.4–32.4), 20.0 (2.5–161.4), 6.6 (1.7–25.9) and 3.8 (0.4–33.3), respectively.
Conclusions : Our results suggest that there is an increased risk of congenital malformations, perinatal mortality and pre-term birth in children born to women treated with azathioprine or mercaptopurine during pregnancy. More data are needed to determine whether the associations are causal or occur through confounding.
Azathioprine is one of the most commonly used immunomodulatory drugs in the treatment of severe cases of inflammatory bowel disease. Furthermore, azathioprine is widely used in the prevention of organ transplant rejection, and in the treatment of various autoimmune disorders. Such diseases often affect women in their fertile years, and the safety of azathioprine use during pregnancy is therefore important. Data regarding the safety of azathioprine during pregnancy are very sparse; there is a surprising absence of human studies with control groups.
After absorption, azathioprine, an inhibitor of purine metabolism, is rapidly converted in the liver into a number of metabolites, including the active metabolite mercaptopurine.1–3 There are no data to suggest that azathioprine and mercaptopurine have different efficacy profiles, and both drugs are usually discussed together with regard to their efficacy and safety.4, 5 Both drugs cross the placenta, and amounts of mercaptopurine can be measured in foetal blood.2, 5, 6 In theory, the foetus is protected from the adverse effects of azathioprine in early pregnancy as its liver lacks the enzyme, inosinate pyrophosphorylase, that converts azathioprine to its active metabolites, and only trace amounts of its metabolites are detected in foetal serum.6–8
Azathioprine teratogenicity has been demonstrated in animals (malformations such as cleft palate, hydrocephalus, skeletal defects and ocular anomalies).6, 9–11 Furthermore, supported by animal and human data (on renal transplant recipients only), the use of azathioprine in pregnancy may be related to intrauterine growth retardation.6, 9, 12, 13 Nearly all human data on the safety of azathioprine and mercaptopurine come from pregnancies in organ transplant recipients, but the studies involve case reports and case series,9, 11, 14–26 apart from one cohort study by Sgro et al.27 That study associated the use of immunosuppressive drugs during pregnancy with an increased risk of stillbirth, pre-term birth and low birth weight, together with a 2.8% prevalence of malformations in the exposed.27 Data on the safety of azathioprine and mercaptopurine in patients with inflammatory bowel disease are even more sparse: they include two case series of two and 14 women exposed during pregnancy,28, 29 and results from a case–control study which have been presented in abstract form only.30
We therefore examined the risk of malformations, perinatal mortality, pre-term birth and low birth weight in women who took up prescriptions for azathioprine or mercaptopurine during pregnancy in a population-based design.
We conducted the study in North Jutland County, Denmark (population of approximately 490 000). We included data on all women who, between 1 January 1991 and 31 December 2000, had a live birth after the 28th week of gestation, together with information on perinatal mortality (58 328 pregnancies). We identified all singleton pregnancies in the county during the study period. Data on drug use and outcome were obtained from the population-based registries in North Jutland County.
Use of azathioprine and mercaptopurine
The population-based Pharmaco-Epidemiological Prescription Database of North Jutland was used to identify all pregnancies in which women had taken up prescriptions for azathioprine or mercaptopurine in the county from 1 January 1991 to 31 December 2000. The county is served by pharmacies equipped with computerized accounting systems, from which data are sent to the Danish National Health Service. The National Health Service provides tax-supported health care for all inhabitants of the country. Apart from guaranteeing free access to general practitioners and hospitals, the insurance programme refunds part of the costs associated with the purchase of most prescribed drugs. Data are transferred to the Prescription Database from the accounting system that is maintained by the pharmacies, and include the patient's civil registry number (which incorporates the date of birth), the type of drug prescribed, according to the anatomical therapeutic chemical classification system, and the date of the prescription. The present study identified all prescriptions with the code L04A X01 (azathioprine) and L01B B02 (mercaptopurine). Both azathioprine and mercaptopurine are only available on prescription.
The Danish Medical Birth Registry. The Birth Registry in the county contains information on all births, including the main variables of maternal age, birth weight, parity, gestational age, perinatal mortality and smoking status of the mother.31 Information about smoking is collected at the first antenatal care visit. Cases of perinatal mortality, pre-term birth and low birth weight in exposed women were reviewed in the hospital records.
The County Hospital Discharge Registry. Data on the type of malformation and on the mother's underlying disease were extracted from the Discharge Registry, including the dates of hospitalization and the discharge diagnoses according to the Danish version of the eighth revision of the International Classification of Diseases (ICD-8) until the end of 1993 and ICD-10 after this date.32, 33 In exposed women, cases of children with malformations were validated by review of the hospital records. With regard to the type of underlying maternal disease, the hospital records were reviewed if there was any doubt about the classification.
The codes for malformations were 740.00–759.99 in ICD-8 and Q0.00–Q99.9 in ICD-10. Diagnoses of dislocation of the hip and undescended testis were excluded because of their low validity.
Linkage of exposure and outcome data
The personal 10-digit civil registry number, which is assigned to all citizens in Denmark shortly after birth, was used to link prescription records with both outcome registries.
The study was approved by the Danish Data Protection Agency (record no. 1995-1200-362) and the Regional Ethics Committee (record no. 1995/104).
The women were classified according to the stage of gestation at which they took up prescriptions for azathioprine or mercaptopurine: (i) the ‘early pregnancy’ group, which took up prescriptions from 30 days before conception to the end of the first trimester, was used to estimate the risk of malformations, because this is the period during which the organs are especially vulnerable to teratogenic exposure; (ii) the ‘entire pregnancy’ group, which took up prescriptions during the first to the third trimesters, was used to estimate the risk of other birth outcomes.34 The main control group comprised all pregnant women who had not been prescribed any kind of reimbursed medicine from three months before conception to the end of pregnancy. A second control group comprised all pregnant women, apart from those treated with azathioprine or mercaptopurine from three months before conception to the end of pregnancy, thereby allowing use of other drugs in the control group. A third control group comprised pregnant women treated with azathioprine or mercaptopurine before pregnancy, but not during pregnancy (that is three months before pregnancy and during pregnancy). This third control group aimed to represent controls with a similar pattern of diseases to the drug-exposed cohort.
Logistic regression analyses were performed to estimate the risk of malformations, perinatal mortality (number of stillbirths and deaths during the first week), pre-term birth (fewer than 37 completed weeks of gestation) and low birth weight (< 2500 g) associated with the use of azathioprine or mercaptopurine, adjusted for potential confounders. Every pregnancy was included in the analyses as an independent event.
In nine pregnancies in the ‘early pregnancy’ group and 10 pregnancies in the ‘entire pregnancy’ group, women had taken up prescriptions for azathioprine or mercaptopurine, and the first control group comprised 19 418 pregnancies (Table 1). Exposed women smoked more, and adverse birth outcomes were more prevalent in users than non-users of azathioprine or mercaptopurine (Table 1). Eleven different azathioprine- or mercaptopurine-exposed women were included in both exposure periods, six of whom had ulcerative colitis or Crohn's disease (55%); the underlying diseases in the other women were vasculitis, myasthenia, glomerulonephritis, nephritis and autoimmune hepatitis (Table 2). Ninety-one per cent of the exposed women had taken up prescriptions for azathioprine and 9% for mercaptopurine (Table 2).
Table 1. Characteristics of the study cohort
|Mother's age (years)|
| Mean (s.d.)||26.7 (4.8)||27.7 (5.3)||28.7 (4.7)|
|Smokers, n (%)||4 (44.4)||4 (40.0)||5102 (26.3)|
|Parity > 1, n (%)||1 (11.1)||3 (30.0)||9191 (47.3)|
|Pre-term birth (gestational age < 37 weeks), n (%)||2 (22.2)||3 (30.0)||1062 (5.5)|
|Low birth weight (< 2500 g), n (%)||2 (22.2)||3 (30.0)||844 (4.4)|
|Malformations, n (%)||2 (22.2)||2 (20.0)||711 (3.7)|
|Perinatal mortality, n (%)||1 (11.1)||1 (10.0)||109 (0.6)|
Table 2. Characteristics of the different azathioprine (AZA)- or mercaptopurine (MP)-exposed women and their birth outcomes
| 1||Ulcerative colitis||Period 1, AZA||No||No||No||No|
| 2||Crohn's disease||Periods 1 and 2, AZA||No||No||No||No|
| 3||Crohn's disease||Periods 1 and 2, AZA||No||No||No||No|
| 4||Crohn's disease||Periods 1 and 2, AZA||No||No||No||No|
| 5||Crohn's disease||Periods 1 and 2, AZA||Yes, aphakia||No||No||No|
| 6||Crohn's disease||Periods 1 and 2, AZA||No||No||No||No|
| 7||Myasthenia||Periods 1 and 2, AZA, 1st birth||No||No||No||No|
|Myasthenia||Period 2, AZA, 2nd birth||No||No||No||No|
| 8||Vasculitis||Period 1, AZA||No||No||No||No|
| 9||IgA nephritis||Period 2, AZA||No||No||Yes||Yes|
|10||Autoimmune hepatitis||Periods 1 and 2, MP||Yes, multiple malformations||Yes||Yes||Yes|
|11||Glomerulonephritis and renal transplant recipient||Period 2, AZA||No||No||Yes||Yes|
The three cases of low birth weight in the ‘entire pregnancy’ group were all born pre-term (none was medically induced). The underlying diseases for the women with these adverse birth outcomes included immunoglobulin A nephritis (severe disease with nephrotic syndrome during pregnancy), autoimmune hepatitis (diagnosed two years before pregnancy; severe disease during pregnancy, treated with 125 mg/day mercaptopurine during the first six months of pregnancy, and with 20 mg/day prednisolone, 200 mg/day ciclosporin and 500 mg/day ursodeoxycholic acid during the entire pregnancy) and renal transplant (transplantation due to glomerulonephritis 10 years before pregnancy, treated with 25 mg/day azathioprine, 5 mg/day prednisone and antihypertensive drugs) (Table 2).
Two children with malformations were recorded in the ‘early pregnancy’ group. A child with aphakia was born to a mother with Crohn's disease, who had been treated with azathioprine for six weeks during the first trimester and with 5-aminosalicylic acid during the entire pregnancy. The other child, with several severe malformations (hypoplasia and dysplasia of the lungs, and malformations of the urinary bladder and urethra), was born at a gestational age of 31 weeks and died 6 h after birth; the mother had chronic aggressive autoimmune hepatitis.
Table 3 shows the adjusted odds ratios (OR) for the different birth outcomes in azathioprine- or mercaptopurine-exposed women using the main control group. When we used the second control group, the risk of malformations, perinatal mortality, pre-term birth and low birth weight was still increased: adjusted OR = 6.3 [95% confidence interval (CI), 1.3–30.1], OR = 18.3 (95% CI, 2.3–146.0), OR = 6.7 (95% CI, 1.7–26.0) and OR = 4.0 (95% CI, 0.4–35.4), respectively. Using the third control group, comprising women with similar diseases to the exposed (women who used azathioprine or mercaptopurine before pregnancy, but not three months before pregnancy or during pregnancy), the adjusted OR remained increased (Table 4).
Table 3. Adjusted odds ratios (OR) for birth outcome in women treated with azathioprine (AZA) or mercaptopurine (MP) during pregnancy, using the main control group of women who had not been prescribed any kind of reimbursed medicine from 3 months before conception to the end of pregnancy
|Low birth weight†||3/10 (30.0)||844/19 418 (4.4)||3.8 (0.4–33.3)|
|Pre-term birth‡||3/10 (30.0)||1062/19 418 (5.5)||6.6 (1.7–25.9)|
|Perinatal mortality||1/10 (10.0)||109/19 418 (0.6)||20.0 (2.5–161.4)|
|Congenital malformations§||2/9 (22.2)||711/19 418 (3.7)||6.7 (1.4–32.4)|
Table 4. Adjusted odds ratios (OR) for birth outcome in women treated with azathioprine (AZA) or mercaptopurine (MP) during pregnancy, using a control group of women treated with AZA or MP before pregnancy but not during pregnancy
|Low birth weight‡||3/10 (30.0)||6/30 (20.0)||2.3 (0.4–13.6)|
|Pre-term birth§||3/10 (30.0)||6/30 (20.0)||2.8 (0.4–19.4)|
|Perinatal mortality||1/10 (10.0)||1/30 (3.3)||3.2 (0.2–56.9)|
|Congenital malformations¶||2/9 (22.2)||1/30 (3.3)||7.7 (0.6–102.1)|
In a sub-analysis, the most severely diseased patient (aggressive autoimmune hepatitis), who had received other potentially harmful drugs, was omitted from the exposed cohort. After excluding this patient from the analyses, there were no perinatal deaths, and the adjusted risks of low birth weight, pre-term birth and malformations were OR = 3.8 (95% CI, 0.4–33.5), OR = 4.4 (95% CI, 0.9–21.4) and OR = 3.4 (95% CI, 0.4–27.3), respectively.
Our data indicate an increased risk of malformations, perinatal mortality and pre-term birth in women using azathioprine or mercaptopurine during pregnancy. However, the associations could be confounded, at least partly, by disease activity or the concurrent use of other drugs (co-medication). By including a group of controls with similar diseases to the exposed, we have tried to take into account the possible influence of the type of underlying disease. For ethical reasons, no randomized trials can be designed to evaluate the safety of azathioprine or mercaptopurine during pregnancy. We must therefore base our clinical decisions on observational studies that are vulnerable to bias and to problems with statistical precision due to the low prevalence of adverse birth outcomes. Thus, it is important to report all available data on possible adverse birth outcomes after maternal exposure to these drugs.
Studies of teratogenesis require special attention because malformations cannot be regarded as a single homogeneous outcome; teratogens do not uniformly increase the rates of all malformations, but rather increase the rates of selected malformations. Therefore, under ideal circumstances, we should consider specific rather than overall rates of malformations.35 Cohort studies can thus only detect considerable increases in the risk of specific defects, and are limited in their ability to provide an assurance of safety.
Our study was based on a complete prescription database, which prevents selection bias. One strength is that exposure measurement was based on prescriptions and not on recall. Drug exposure based on self-reported use may lead to recall bias or under-ascertainment, which is a serious threat in case–control studies.36 Another strength is that the outcome data were obtained independently of exposure measurement, and most outcome data obtained from the Danish Medical Birth Registry are valid.37 On the other hand, we have no information on compliance. However, patient non-compliance is not likely to have a major impact, because the drugs are used for long-term treatment in these usually co-operative patients with severe chronic diseases, and a potential misclassification of exposure would tend to under-estimate our risk estimates. One weakness is that most of our estimates are influenced by low statistical precision.
The finding of an increased risk of perinatal mortality and pre-term birth after azathioprine or mercaptopurine exposure supports the data from the study by Sgro et al.27 In that cohort study (44 exposed, 88 controls), significantly more stillbirths and pre-term births were observed in children born to women who had undergone renal transplantation, and who had been treated with prednisone, together with azathioprine, ciclosporin, or both. However, the relative risks were not estimated, there was no opportunity for confounder control and there was no information on whether the children with adverse birth outcomes, such as stillbirth and pre-term birth, were born to women exposed to ciclosporin or azathioprine. The relative risk of pre-term birth was not estimated in the abstract of the study by Francella et al., but given the figures in the abstract, the calculated crude OR is 1.3 in patients who conceived whilst taking mercaptopurine.30
The risk of malformations has not been estimated previously in human controlled studies, and therefore our finding of an increased risk is new. One child in our study was born with aphakia, a malformation that, at least theoretically, is in keeping with ocular anomalies which have been found in animal studies.6, 10 To our knowledge, multiple malformations, as seen in the second child in our study, have not been found previously. The risk of malformation was not estimated in the study by Sgro et al.27 The risk of malformations amongst live births can be calculated from the figures in the abstract of the study by Francella et al.: the crude OR is 1.6 in patients who conceived whilst taking mercaptopurine.30
Our more than three-fold increased risk of low birth weight corresponds with the results of other studies, which have shown a positive association between maternal exposure to azathioprine (or other immunosuppressive drugs) and low birth weight/intrauterine growth retardation.6, 9, 12, 13, 27
Several review articles have addressed the issue of the safety of azathioprine and mercaptopurine during pregnancy.5–8 These articles have suggested, in the main, that the use of these drugs during pregnancy is safe, although they admit that the amount of evidence available in this field remains limited.
The women receiving azathioprine or mercaptopurine during pregnancy in this study had chronic, severe disorders of various kinds. There were probably two main reasons for these women to maintain such medication during pregnancy: (i) to decrease disease activity; and (ii) to maintain an already stable condition. Therefore, the use of azathioprine or mercaptopurine during pregnancy does not necessarily reflect severe disease activity, but may be an indication of maintenance therapy. We therefore believe that the exposed women in our study had severe disorders, with either ongoing disease activity or a potential threat of disease activity. In this type of study, the influence of the disease itself on the risk of adverse birth outcomes is difficult to distinguish from a possible adverse drug reaction, because drug use (which may reflect some level of disease activity) and the underlying disease are closely correlated, and, furthermore, disease activity and medical treatment often fluctuate over time. In an attempt to take into account the influence of the underlying disease, it may be most appropriate to examine the results of the analyses using the third control group, and to compare women with the same types of underlying disease. Using this third control group, the risk of low birth weight, pre-term birth, perinatal mortality and malformation still remained increased, although none of the associations reached statistical significance (in keeping with our small sample size).
In conclusion, our data suggest that the use of azathioprine or mercaptopurine during pregnancy may not be as safe as previously expected. Information about drug use may be difficult to interpret, because we do not have sufficient data to study the influence of disease activity and co-medication, or to stratify according to the type of underlying disease. Therefore, more data are needed to determine whether these associations are causal or occur through confounding.
The study was supported by the Western Danish Research Forum for Health Sciences.