Safety of hydroxychloroquine in pregnant patients with connective tissue diseases: A study of one hundred thirty-three cases compared with a control group
The use of hydroxychloroquine (HCQ) in pregnancy remains controversial. The recent demonstration that HCQ passes across the placenta, with cord blood concentrations nearly identical to those found in maternal blood, emphasizes the need for careful evaluation of pregnancies in women receiving HCQ. However, only small series of HCQ-treated pregnant women have been reported, and most of these studies had no control group. We now report our experience with 133 pregnancies in women being treated with HCQ, resulting in 117 live births. Results in the HCQ group are compared with those in a control group.
One hundred thirty-three consecutive pregnancies in 90 women treated with 200 mg of HCQ either twice daily (122 pregnancies) or once daily (11 pregnancies) were studied. These pregnancies were compared with 70 consecutive pregnancies in 53 women with similar disorders who did not receive HCQ. Electrocardiography was performed in 47 children of mothers treated with HCQ and in 45 children in the control group.
Eighty-eight percent of pregnancies in the HCQ group and 84% of those in the control group ended successfully with a live birth. The outcomes of pregnancy were not statistically different between groups. One child in each group died of causes related to prematurity. Three malformations were observed in the HCQ group (1 hypospadias, 1 craniostenosis, and 1 cardiac malformation) versus 4 in the control group. On the electrocardiograms, the PR interval and the corrected QT interval were not statistically different between groups. No visual, hearing, growth, or developmental abnormalities were reported in any of the children at the last follow-up (ages 12–108 months; mean age 26 months).
Our findings support preliminary evidence for the safety of HCQ therapy during pregnancy. This treatment probably should be maintained throughout pregnancy in patients with systemic lupus erythematosus.
Hydroxychloroquine (HCQ), an antimalarial drug containing the 4-aminoquinoline radical, is widely used in the treatment of systemic lupus erythematosus (SLE). Results of a randomized, controlled trial performed in nonpregnant patients showed that discontinuation of HCQ doubles the risk of SLE flare during the next 6 months (1). It is generally agreed that pregnancy per se increases disease activity in patients with SLE. Therefore, withdrawal of HCQ at the onset of pregnancy may result in exacerbation of SLE. Additionally, discontinuation of the drug might not prevent side effects in fetuses because of its very long terminal elimination half-life, which is estimated to be 1–2 months (2). Consequently, even though use of HCQ during pregnancy remains controversial (3), Parke (4, 5) and other investigators (6–9) have proposed continuation of this treatment throughout gestation.
We recently demonstrated that HCQ passes across the placenta, with cord blood concentrations nearly identical to those found in maternal blood, suggesting that during pregnancy the level of exposure to HCQ in mother and fetus is similar (10). These results raise the issue of potential teratogenic and toxic effects of HCQ and emphasize the need for careful evaluation of pregnancies in women being treated with HCQ. However, only small series of such pregnancies have been described, with a maximal sample size of 35 live births (7) and a total of 101 pregnancies resulting in live births (4–9). Most of these studies had no control group.
We now report our experience concerning HCQ treatment of women with SLE, at a single center, throughout 133 pregnancies that resulted in 117 live births. The results in patients receiving HCQ treatment are compared with those in a control group. Because the terminal elimination half-life of HCQ is very long (2), and because there is no information concerning the metabolism of HCQ in newborns, the children of mothers enrolled in the study were followed up for a mean of 26 months.
PATIENTS AND METHODS
All of the patients had been monitored regularly at the Pitié-Salpêtrière Hospital before becoming pregnant.
Hydroxychloroquine group (group A).
Between January 1993 and March 2002, 133 consecutive pregnancies occurred in 90 women who had been treated with HCQ sulfate (Plaquenil; Sanofi-Winthrop, Gentilly, France) for at least 6 months prior to pregnancy. In this group, HCQ treatment continued throughout gestation. Indications for HCQ treatment were SLE as defined by the American College of Rheumatology criteria (11) (n = 69), miscellaneous or unclassified connective tissue disease (n = 13), or primary Sjögren's syndrome as defined by the European criteria (12) (n = 8). Additionally, 20 patients with SLE fulfilled the Sapporo criteria for definite antiphospholipid syndrome (13). Thirty-nine patients demonstrated positivity for anti-SSA/Ro antibodies, and 16 showed positivity for anti-SSB/La antibodies.
All but 4 of the pregnancies occurred spontaneously. Fifty-seven women had 1 pregnancy, 25 patients had 2 successive pregnancies, 6 patients had 3 successive pregnancies, and 2 patients had 4 successive pregnancies. At study entry, 64 women were nulliparous, and 26 were multiparous. The mean age of patients at the time of delivery was 31 years (range 21–43 years).
In 122 pregnancies women received HCQ 200 mg twice daily, and in 11 pregnancies women received HCQ 200 mg once daily. Other treatments included prednisone (108 pregnancies), aspirin 100 mg/day (112 pregnancies), low molecular weight heparin (35 pregnancies), azathioprine (2 pregnancies), and intravenous immunoglobulin (2 pregnancies).
Control group (group B).
Between January 1993 and December 2001, 70 consecutive pregnancies occurred in 53 women who had not been treated with HCQ for at least 6 months prior to conception. These women had disorders similar to those among women in the active-treatment group, and they did not receive HCQ during gestation. Abstention from HCQ therapy was essentially attributable to the physician's or patient's personal decision and, in some cases, was due to an ocular contraindication. Forty-one patients had SLE, 8 had primary Sjögren's syndrome, and 4 had miscellaneous or unclassified connective tissue disease. Additionally, 8 patients had antiphospholipid syndrome. Twenty-eight patients were positive for anti-SSA/Ro antibodies, and 13 were positive for anti-SSB/La antibodies.
All but 1 pregnancy occurred spontaneously. At study entry, 36 women were nulliparous, and 17 were multiparous. Thirty-eight women had 1 pregnancy, 13 women had 2 successive pregnancies, and 2 patients had 3 successive pregnancies. The mean age at the time of delivery was 32 years (range 21–42 years). Treatment in these women included prednisone (56 pregnancies), aspirin 100 mg/day (54 pregnancies), low molecular weight heparin (13 pregnancies), and azathioprine (2 pregnancies).
During their pregnancies, patients were closely followed up by both an internist and a high-risk pregnancy obstetric team, monthly until 32 weeks' gestation and then every 2 weeks. Instrumental monitoring included repeated fetal electrocardiography and Doppler velocimetry. Each child was carefully examined during the neonatal period. Electrocardiography was performed on the third day of life in 47 children of mothers in group A and in 45 children of mothers in group B. Pediatric cardiologists who were blinded as to maternal treatment measured the PR and corrected QT (QTc) intervals for all children in whom electrocardiography was performed.
Data on the health of each child were gathered from the mothers, general practitioners, and/or pediatricians when children were a mean age of 26 months (median age 24 months; range 12–108 months). During followup visits, each mother was asked about the health of her child. Each mother was also asked if results of a medical examination of her child were normal, especially the complete medical examinations that were systematically performed at 4 months, 9 months, and 24 months in France. Such medical appointments are mandatory and include evaluations of hearing and vision. In some rare cases, when women were lost to followup, general practitioners or pediatricians were telephoned and asked to provide information about the child.
Gestational age was based on the date of the last menstrual period and on ultrasonographic findings during the first trimester of pregnancy. Spontaneous abortion was defined as spontaneous termination of pregnancy prior to 20 weeks' gestation; fetal death was defined as spontaneous termination of pregnancy after 20 weeks' gestation; premature birth was defined as termination of pregnancy with a live birth before 37 weeks' gestation; and full-term birth was defined as termination of pregnancy with a live birth at or after 37 weeks' gestation.
Student's t-test was used for continuous variables, and the chi-square test or Fisher's exact test was used when required for dichotomous variables. To calculate its statistical power, this study was considered to be an equivalence study. Epsilon (the greatest acceptable difference between rates within which rates are considered equivalent) was set at 2%.
Two groups of pregnant women with connective tissue disease were compared: patients treated with HCQ (group A) and control patients who did not receive HCQ (group B). No difference was found between the 2 groups for age, disease, and other treatments received during pregnancy (Table 1). Parity among patients in group A (64 nulliparous, 26 multiparous) was similar to that among patients in group B (36 nulliparous, 17 multiparous).
Table 1. Characteristics of study groups*
|Mothers|| || || |
| n||90||53|| |
| Age, mean (range) years||31 (21–43)||32 (21–42)||0.21|
| Disease|| || || |
| Systemic lupus erythematosus||69 (77)||41 (77)||0.92|
| Miscellaneous/UCTD||13 (14)||8 (15)||0.92|
| Primary Sjögren's syndrome||8 (9)||4 (8)||1.00|
| Antiphospholipid syndrome||20 (22)||8 (15)||0.30|
|Pregnancies|| || || |
| Number||133||70|| |
| Treatment|| || || |
| Prednisone||108 (81)||56 (80)||0.84|
| Aspirin, 100 mg/day||112 (84)||54 (77)||0.22|
| Low molecular weight heparin||35 (26)||13 (19)||0.22|
| Azathioprine||2 (2)||2 (3)||0.61|
| Intravenous immunoglobulin||2 (2)||0 (0)||0.55|
Eighty-eight percent of pregnancies in group A (117 of 133) and 84% of those in group B (59 of 70) ended successfully with a live birth (Table 2). In group A, the outcomes of the 133 pregnancies included 15 spontaneous abortions, 1 fetal death, 33 premature births, and 84 full-term births. The 117 live births in this group occurred at a mean gestational age of 37.1 weeks (range 26–41 weeks). The mean weight of the 119 newborns (2 twin pregnancies) was 2,754 gm (range 500–4,300). No differences between group A and group B were observed, except for a slightly lower mean gestational age in group A (Table 2).
Table 2. Pregnancy outcomes*
|Spontaneous abortion||15 (11.3)||7 (10)||0.78|
|Fetal death||1 (0.8)||2 (2.9)||0.27|
|Therapeutic abortion||0 (0)||2 (2.9)||0.12|
|Live birth†||117 (88)||59 (84.3)||0.46|
|Premature birth||33 (28)||12 (17)||0.21|
|Full-term birth||84 (72)||47 (67)||0.21|
|Gestational age, mean (range) weeks||37.1 (26–41)||38.1 (29–41)||<0.02|
|Weight, mean (range) grams||2,754 (500–4,300)||2,897 (1,200–4,250)||0.15|
Deaths and malformations.
In group A (HCQ pregnancies), 1 premature twin died of pulmonary hypoplasia at 15 days of life. Three malformations were observed: hypospadias (n = 1), craniostenosis (n = 1), and transposition of the great arteries with interventricular communication and pulmonary stenosis (n = 1). The child with transposition of the great arteries underwent successful surgical correction at 10 days of life. In 1 child, idiopathic thrombocytopenic purpura developed at age 2 years; this child is still being treated with intravenous immunoglobulin at age 3 years.
In group B (control), 1 child who was born prematurely died of complications of prematurity. Four malformations were observed: hypospadias (n = 1), aplasia cutis congenita of the scalp (n = 1), ulnar polydactyly (n = 1), and severe cardiac malformation that required therapeutic abortion (aplastic right heart) (n = 1).
Electrocardiograms (EKGs) were obtained in 47 children of women in group A and in 45 children of women in group B. The mean (±SD) PR intervals were 91.4 ± 12.7 msec and 92.1 ± 11.9 msec, respectively (P = 0.92). The mean (±SD) QTc intervals were 406 ± 34 msec and 407 ± 32 msec, respectively (P = 0.40). The EKG findings were otherwise normal for age.
Followup of children.
Data for each child were collected at a mean age of 26 months (median 24 months, range 12–108 months). No visual, hearing, growth, or developmental abnormalities were reported by the mothers, general practitioners, or pediatricians.
In women treated with HCQ, 133 pregnancies leading to 117 live births were followed up and compared with those in a control group. Malformations were carefully recorded, and all pregnancies were evaluated, including therapeutic abortion. Three malformations were observed in children of mothers in the HCQ group: hypospadias, craniostenosis, and cardiac malformation. Four malformations were observed in children of women in the control group (70 pregnancies). For comparison, the malformation rate in the general population, according to the European registry EUROCAT (European Concerted Action on Congenital Anomalies and Twins) (minor birth defects are excluded) is equal to 2.3% of all births (including fetal deaths and induced abortion) (14). Thus, the results in our study are equivalent to what would be expected in a large population of pregnancies, with the statistical power of our study being >90%.
Additionally, these results may not be imputable to HCQ, especially because of the diversity of malformations observed. In previous reports, the only congenital abnormalities observed in children of mothers who were treated with HCQ during their pregnancies were patent ductus arteriosus with atrial septal defect (8) and Down's syndrome (6). However, because these malformations are not exceptional in a general population (prevalence of 0.1% and 0.2%, respectively) (14), no firm conclusions can be drawn regarding the role of HCQ in these birth defects.
Concerning the potential toxic effects of HCQ, both retinal toxicity and ototoxicity have been reported in children born to women treated with chloroquine (15–17). These complications have never been reported in association with HCQ. Due to the difficulties encountered in assessing vision in young children, ophthalmologic examinations were not systematically performed in our series. However, although the children were not routinely checked on a prospective basis, data for each of them were carefully collected from mothers, general practitioners, and/or pediatricians. Because the terminal elimination half-life of HCQ is very long (2), clinical data on the children were collected at a mean age of 26 months. No clinical vision or hearing abnormalities have been observed, and the growth and development of these children appear to be normal. Our clinical data are in accordance with those reported by Klinger et al (8), who found that results of ophthalmologic examinations and tests were normal in 14 children who were exposed to HCQ in utero and were assessed at a mean age of 1.9 years. These results have been confirmed in 26 additional children (7, 9).
Because cardiac conduction abnormalities have been reported in adults who received long-term treatment with HCQ (18), we obtained EKGs in the children of mothers enrolled in this study. No difference was found between the HCQ group and the control group with regard to duration of the PR or QTc intervals, and the EKG findings were considered normal for age.
The limitations of our study should be noted. First, we did not systematically perform complete ophthalmologic examinations and tests in all children. Second, even though the number of pregnancies in this study is substantial in comparison with previous data, it may be insufficient to detect a discrete increase in birth defects. If so, a multicenter study including complete ophthalmologic examinations and long-term followup of the children would be recommended.
Nonetheless, our findings support preliminary evidence of the safety of HCQ use during pregnancy (4–9) and are concordant with the experience of selected experts, as reported in a national survey concerning the use of antimalarial drugs in lupus pregnancy (19). Indeed, according to results of that survey, 69% of respondents continued to prescribe antimalarial drugs sometimes, often, or always during pregnancy, and this was particularly true for those who treated a larger number of pregnant lupus patients per year (P < 0.01). Additionally, none of the respondents reported having seen any fetal toxicity related to the use of antimalarials. In conclusion, in the absence of HCQ-related adverse effects documented in pregnancy, and because SLE flares are frequently observed following HCQ discontinuation, we think this treatment should probably be maintained throughout pregnancy in patients with SLE.
The authors thank Dr. Janine Goujard for assistance in providing data from EUROCAT.