ABSTRACT
Purpose
The aim of the study was to assess the risk of pregnancy termination and major congenital anomalies with tramadol by comparing women exposed to tramadol during pregnancy with women exposed to codeine and unexposed women, using the French EFEMERIS database.
Methods
The study was based on the EFEMERIS (Évaluation chez la Femme Enceinte des MÉdicaments et de leurs RISques) database, which contains reimbursed medications prescribed and dispensed to pregnant women, dates of conception and pregnancy outcome, and data on the children (congenital anomalies, neonatal diseases, etc.). Women were considered to be exposed if they were prescribed and dispensed tramadol or codeine at least once during pregnancy. Women who had a pregnancy outcome between July 1, 2004 and December 31, 2020 and living in Haute‐Garonne (south‐west France) were included in this study.
Results
We compared 1602 (1.0%) pregnancies exposed to tramadol (including 1628 fetuses/newborns due to multiple pregnancies) with 6311 (3.8%) exposed to codeine (including 6406 fetuses/newborns) and 158 426 (95.2%) unexposed to tramadol and codeine (including 160 784 fetuses/newborns). The rate of pregnancies exposed to tramadol increased sevenfold between 2004 and 2020. The study showed an increase in the rate of spontaneous pregnancy termination in women exposed to tramadol compared with women exposed to codeine (aHR [95% CI] = 2.23[1.64–3.03]) and unexposed women (aHR [95% CI] = 1.86[1.46–2.37]), after adjustment for maternal age, dispensation of folic acid, teratogenic drugs, NSAIDs, presence of hypertension, diabetes, and long‐term Illness. Multivariate analysis did not show an increased rate of major congenital anomalies in fetuses/newborns exposed in utero to tramadol during the first trimester of pregnancy compared with fetuses/newborns exposed to codeine (ORa [95% CI] = 1.03 [0.67–1.57]) and those not exposed to these medications (ORa [95% CI] = 1.24 [0.86–1.79]).
Conclusion
This study found an association between tramadol use and the risk of spontaneous pregnancy termination. We cannot conclude that there is a causal link because of a possible indication bias. No association between prescription and dispensation of tramadol during the first trimester of pregnancy and an increased risk of major congenital anomalies has been found.
Keywords: congenital anomalies, EFEMERIS database, opioid, pregnancy, pregnancy termination, tramadol
Summary.
Data on the risks of tramadol use during pregnancy are limited. Thisstudy provides additional data on this topic.
This study showed that the use of tramadol in pregnant women is increasing in France (7times more in 15 years).
This study found an association between tramadol use and the risk of spontaneouspregnancy termination. We therefore questioned whether there was an indication bias. Indeed, tramadol may have been prescribed and dispensed for some women in anticipation of abortion pain.
Further studies, using different methods, are needed to confirm or refute the risk of spontaneous pregnancy termination.
Two previously published studies have reported conflicting data regarding the risk of congenital anomalies associated with tramadol.
This study found no evidence of an increased risk of congenital anomalies following exposure to tramadol during the first trimester of pregnancy.
1. Introduction
Tramadol is an opioid analgesic indicated for moderate to severe pain [1]. This medication is a nonselective agonist of morphine receptors, with a higher affinity for mu receptors. The peculiarity of this opioid is that it also has serotonergic and noradrenergic effects [1]. Tramadol inhibits norepinephrine reuptake and increases serotonin release. The available data on the risk of use during pregnancy remain limited and controversial. Two studies evaluating the risk of congenital anomaly with tramadol have been published. The first, from 2015 [2] and carried out using the Swedish Medical Birth Registry, compared women exposed to tramadol during the first trimester of their pregnancy with women who were not exposed. This study found a small increased risk of congenital anomalies (OR = 1.30 [1.06–1.69]), particularly cardiovascular anomalies (OR = 1.56 [1.04–2.29]) and talipes equinovarus (OR = 3.63 [1.66–6.89]). The second, from 2022 [3] is a cohort study including all pregnancies registered in Denmark between 1997 and the end of 2016. Each tramadol‐exposed pregnancy was matched with four unexposed pregnancies using a propensity score. The authors did not find an increased risk of congenital anomaly in pregnant women exposed to tramadol (OR = 1.04 [0.87–1.24]). Thus, the currently published studies do not allow reaching a conclusion on the risk of congenital anomalies with the use of tramadol during pregnancy.
The aim of this study was therefore to assess the risk of pregnancy termination and major congenital anomalies with tramadol by comparing women exposed to tramadol during pregnancy with women exposed to codeine, as well as women unexposed to tramadol and codeine, using the French EFEMERIS database.
2. Methods
This study was based on the EFEMERIS (Évaluation chez la Femme Enceinte des MÉdicaments et de leurs RISques) database, which has been reported on elsewhere [4]. Data in EFEMERIS come from three sources. First, data on reimbursed medications prescribed and dispensed to patients, as well as the dates of conception and pregnancy outcome, were obtained from the French health insurance system. For all pregnancies, a document containing the date of conception is completed by the physician and must be sent to the French health insurance system for reimbursement of health care costs. Second, data on the mother including socio‐professional characteristics and medical history (diabetes, hypertension), and data on the children (congenital anomalies, neonatal diseases, etc.) were obtained from the children's medical certificates completed during the compulsory medical examinations performed at 8 days, 9 months, and 2 years. These certificates were recorded by the Mother and Child Protection Centre. Finally, data on hospital admissions and pregnancy terminations were obtained from the Toulouse University Hospital Centre and the French health insurance system.
The women included in the study were all residents of Haute‐Garonne (South‐West France) and were covered by the general health insurance system, which covers about 86% of the general population. This study focused on pregnancies between July 1, 2004 and December 31, 2020.
This observational study compared the risk of pregnancy termination and major congenital anomalies in pregnant women who had a prescription and a dispensation for tramadol at least once during pregnancy with pregnant women who had a prescription and a dispensation of codeine and with unexposed women. The codeine‐exposed group was chosen so as to have a control group with comparable pathologies; codeine being a weak opioid analgesic like tramadol. Women were considered to be exposed if they were prescribed and dispensed tramadol or codeine at least once during pregnancy (between the date of the last menstrual period and the date of the pregnancy outcome).
We first conducted a descriptive analysis. We calculated the rates of women who were prescribed and dispensed tramadol at least once during pregnancy, of women who were prescribed and dispensed codeine at least once during pregnancy, and of women who were not prescribed and dispensed tramadol or codeine during pregnancy. We then determined the annual trend in the rate of pregnant women exposed to tramadol during pregnancy. We calculated the rate of tramadol prescription and dispensation by prescriber specialty. We also determined the rate of pregnancies exposed to tramadol according to pregnancy trimester. We then compared the characteristics (sociodemographic, disease) of the three groups of women: “exposed to tramadol,” “exposed to codeine” and “unexposed” as well as the medications prescribed and dispensed during their pregnancy. Finally, we described pregnancy outcomes (live births and terminations) between the three groups and the health status of the children. We calculated the rate of fetuses/newborns with at least one major congenital anomaly. The different types of major congenital anomaly were described within the three groups. In EFEMERIS, as in EUROCAT, congenital anomalies are classified according to the International Classification of Diseases (ICD10, Chapter XVII “Congenital malformations and chromosomal anomalies,” codes Q00‐Q99). The denominator used to calculate the prevalence of major congenital anomalies in the three groups was the number of live births, fetal deaths (after 22 weeks' gestation), and medical terminations of pregnancy (MTP), outcomes for which the congenital anomaly status was known.
Second, we compared the risk of pregnancy termination based on exposure to tramadol. This analysis only included spontaneous pregnancy terminations (miscarriages and stillbirths). In this study, women exposed to tramadol, whatever the trimester, were considered as exposed. Thus, this design may introduce an immortal time bias. To reduce this bias, we used Cox proportional hazard models with the pregnancy duration as the time variable. We then considered a woman as exposed to tramadol (or codeine) from the first date of tramadol or codeine dispensation until her pregnancy outcome. Before the first date of dispensation, she was considered “unexposed”. We also performed a sensitivity analysis by excluding women who were dispensed one or more nonsteroidal anti‐inflammatory drugs during their pregnancy because these medications may be the cause of an increased risk of spontaneous miscarriage and stillbirth.
We also assessed the risk of major congenital anomalies based on tramadol exposure. For this analysis, we selected pregnancy outcomes for which it was possible to identify a congenital anomaly (live births, fetal deaths (after 22 weeks of amenorrhea), and medical terminations of pregnancy), and pregnancies exposed to tramadol or codeine, at least during the first trimester of pregnancy, corresponding to the period of maximal teratogenic risk. Women exposed to tramadol in the first trimester of pregnancy were compared with women exposed to codeine in the first trimester and with unexposed women, using multivariate logistic models. Only congenital anomalies considered major were retained. The Guide available on the EUROCAT website was used for this selection (https://eu‐rd‐platform.jrc.ec.europa.eu/eurocat). Two sensitivity analyses were performed. In the first, we excluded fetuses/newborns with genetic disorders that were often unrelated to drug exposure. In the second, only fetuses/newborns exposed to tramadol and those exposed to codeine during at least the first 56 days of pregnancy (period of organogenesis) were included.
For each of the multivariate analyses, the crude and adjusted odds ratios (OR) were estimated with their 95% confidence intervals. Confounders were identified using a directed acyclic graph (DAG).
Descriptive and multivariate analyses were performed using RStudio software, version 4.3.2 (October 31, 2023, UCRT). A two‐sided 5% significance level was used.
3. Results
Our cohort included 166 339 pregnancies, of which 1.0% (n = 1602) exposed to tramadol (including 1628 fetuses/newborns due to multiple pregnancies), 3.8% (n = 6311) exposed to codeine (including 6406 fetuses/newborns) and 95.2% (n = 158 426) not exposed to tramadol and codeine (including 160 784 fetuses/newborns) (Figure 1).
FIGURE 1.
Flow chart.
The proportion of pregnancies exposed to tramadol increased sevenfold between 2004 and 2020 (Figure 2).
FIGURE 2.
Change in tramadol prescriptions and dispensations to pregnant women over time, N = 1602 pregnancies. *Number of pregnancies that received at least one prescription and dispensation of tramadol compared to the total number of pregnancies in the year (year of delivery).
The majority of tramadol prescriptions came from general practitioners (84.0%), obstetrician gynecologists (4.4%) and dental surgeons (1.5%). The majority of pregnant women received tramadol in the first trimester of pregnancy (71.2%), 18.9% in the second, and 14.7% in the third trimester (some women may have received tramadol in more than one trimester).
3.1. Characteristics of Women (Table 1)
TABLE 1.
Comparison of the characteristics of the women in the three groups, N = 166 339.
| Pregnancies exposed to tramadol (N = 1602) | Pregnancies exposed to codeine (N = 6311) | Unexposed pregnancies (N = 158 426) | ||||
|---|---|---|---|---|---|---|
| Mean or number | Standard deviation [min–max] or % | Mean or number | Standard deviation [min–max] or % | Mean or number | Standard deviation [min–max] or % | |
| Socio‐professional characteristics | ||||||
| Age of mothers (in years) | 31.0 | 5.5 [18–47] | 30.9 | 5.2 [15–51] | 30.7 | 5.0 [13–55] |
| Level of education (N = 91 532) | N = 941 a | N = 4208 a | N = 86 383 a | |||
| Primary school | 31 | 3.3% | 101 | 2.4% | 1385 | 1.6% |
| Secondary/middle school | 202 | 21.5% | 665 | 15.8% | 10 333 | 12.0% |
| Baccalaureate/high school level | 296 | 31.5% | 1176 | 27.9% | 19 890 | 23.0% |
| 1 or 2 years higher education | 348 | 37.0% | 2002 | 47.6% | 47 440 | 54.9% |
| 3 and more years higher education | 64 | 6.8% | 264 | 6.3% | 7335 | 8.5% |
| Women in work (N = 67 849) | N = 642 a | N = 3061 a | N = 63 805 a | |||
| 439 | 68.4% | 2157 | 70.5% | 47 385 | 74.3% | |
| Long‐term adverse health conditions (N = 156 304) | N = 1568 a | N = 6176 a | N = 148 560 a | |||
| 76 | 4.8% | 225 | 3.6% | 3181 | 2.1% | |
| Obstetric disease and history | ||||||
| Hypertension during pregnancy (N = 166 339) | N = 1602 a | N = 6311 a | N = 158 426 a | |||
| 97 | 6.1% | 357 | 5.7% | 5785 | 3.7% | |
| Gestational diabetes (N = 166 339) | 69 | 4.3% | 238 | 3.8% | 3387 | 2.1% |
Number of women for whom data is available.
Regarding the socio‐professional characteristics of the women, there was a statistically significant difference between the three groups, in particular the level of education, with a lower level in the tramadol group. In addition, women exposed to tramadol more often had “long‐term adverse health conditions” (30 long‐term and serious health conditions for which patients are fully covered by the French healthcare scheme).
The mean age was similar in the three groups. Women exposed to tramadol and exposed to codeine had more gestational diabetes (4.3%, compared to 3.8% in the codeine‐exposed group and 2.1% in the unexposed group) and hypertension (6.1%, compared to 5.7% in the “codeine‐exposed” group and 3.7% in the “unexposed” group) during their pregnancy than women not exposed.
Women exposed to tramadol and those exposed to codeine had more different medications prescribed and dispensed during their pregnancy (14.0 ± 8.1 and 14.2 ± 7.3, respectively compared with 9.2 ± 6.2 in unexposed women) excluding tramadol and codeine. They also received more folic acid supplementation before pregnancy and in the first trimester (37.6% and 40.2%, respectively compared with 31.6% in unexposed patients).
3.2. Pregnancy Outcomes
Of the 1602 pregnancies exposed to tramadol, 11.0% (n = 177) resulted in a nonlive birth (miscarriage, stillbirth, ectopic pregnancy, medical termination or elective termination), which is approximately twice the rate in “codeine‐exposed” and “unexposed” groups (5.1% and 5.7%, respectively) (Table 2). Considering only spontaneous pregnancy termination, the rate in the “tramadol” group remained twice that of the “codeine” and “unexposed” groups (5.6% vs. 2.4% and 2.9%, respectively). A Cox proportional risk model was used to measure the association between tramadol and the risk of spontaneous termination of pregnancy. This model showed an increase in the spontaneous pregnancy termination rate in women exposed to tramadol compared with women exposed to codeine (aHR [95% CI] = 2.23[1.64–3.03]) and unexposed women (aHR [95% CI] = 1.86[1.46–2.37]), after adjustment for maternal age, dispensation of folic acid, teratogenic drugs, NSAIDs, presence of hypertension, diabetes, and long‐term adverse health conditions. Sensitivity analysis excluding women who received one or more nonsteroidal anti‐inflammatory drugs during pregnancy showed similar results with an aHR [95% CI] of 1.83[1.19–2.81] when comparing women exposed to tramadol with those exposed to codeine and 2.49[1.48–4.2] in comparison with the “unexposed” group.
TABLE 2.
Pregnancy outcomes of women in the 3 groups, N = 166 339 pregnancies.
| Pregnancies exposed to tramadol (N = 1602) | Pregnancies exposed to codeine (N = 6311) | Unexposed pregnancies (N = 158 426) | ||||
|---|---|---|---|---|---|---|
| Number | % | Number | % | Number | % | |
| Type of pregnancy outcome | ||||||
| Live births | 1425 | 89.0 | 5998 | 95.0 | 149 328 | 94.2 |
| Terminations | 177 | 11.0 | 317 | 5.1 | 9162 | 5.7 |
| Early terminations a (< 22 weeks of amenorrhea) | 160 | 10.0 | 275 | 4.4 | 7972 | 5.0 |
| Fetal deaths (≥ 22 weeks of amenorrhea) | 7 | 0.4 | 19 | 0.3 | 321 | 0.2 |
| Medical terminations of pregnancy | 10 | 0.6 | 23 | 0.4 | 869 | 0.5 |
Early terminations included miscarriages, ectopic pregnancies and elective terminations.
3.3. Characteristics of Fetuses/Neonates
In our study population, there were 156 751 pregnancies that resulted in live birth with a total of 159 146 neonates, including 0.9% (n = 1451) in the tramadol‐exposed group, 3.8% (n = 6087) in the codeine‐exposed group, and 95.3% (n = 151 608) in the unexposed group. Children exposed in utero to tramadol were more often born preterm than those in the other two groups (8.5% compared with 6.7% for codeine exposure and 6.2% for unexposed children) (Table 3).
TABLE 3.
Comparison of characteristics of newborns from the three groups, N = 159 146.
| Newborns exposed to tramadol (N = 1451) | Newborns exposed to codeine (N = 6087) | Newborns exposed to tramadol vs. codeine (N = 7524) | Unexposed neonates (N = 151 608) | Newborns exposed to tramadol vs. not exposed (N = 152 793) | ||||
|---|---|---|---|---|---|---|---|---|
| Number | % | Number | % | p | Number | % | p | |
| N = 1446 a | N = 6078 a | N = 7524 a | N = 151 347 a | N = 152 793 a | ||||
| Male sex | 742 | 51.3 | 3132 | 51.5 | 0.91 | 77 368 | 51.1 | 0.90 |
| Preterm | 122 | 8.5 | 412 | 6.7 | 0.06 | 9334 | 6.2 | 0.002 |
| Extremely (< 28 weeks of amenorrhea) | 1 | 0.1 | 14 | 0.2 | 265 | 0.2 | ||
| Very (28–32 weeks of amenorrhea) | 11 | 0.8 | 38 | 0.6 | 890 | 0.6 | ||
| Moderate to late (≥ 32 weeks of amenorrhea) | 110 | 7.6 | 360 | 5.9 | 8179 | 5.4 | ||
| Small for gestational age (N = 140 553) | N = 1257 a | N = 5353 a | N = 6610 a | N = 133 943 a | N = 135 200 a | |||
| 133 | 10.7 | 549 | 10.3 | 0.73 | 14 037 | 10.5 | 0.91 | |
Number of newborns for whom data is available.
Of the 999 fetuses/newborns (pregnancies of ≥ 22 weeks of amenorrhea duration and medical terminations) who were exposed to tramadol in the first trimester of pregnancy, 30 or 3.0% developed a major congenital anomaly compared with 79 or 2.9% for codeine exposures in the first trimester of pregnancy (n = 2675) and 3719 or 2.4% for unexposed (n = 151 811). The types of congenital anomalies are described in Table 4.
TABLE 4.
Major congenital anomalies in fetuses/newborns “exposed” to tramadol during the first trimester of pregnancy, codeine during the first trimester of pregnancy, and “unexposed”, N = 156 485.
| Major congenital anomalies according to EUROCAT | Fetuses/newborns exposed to tramadol in T1 (N = 999a) | Fetuses/newborns exposed to codeine in T1 (N = 2675a) | Unexposed fetuses/neonates (N = 152 811) | |||
|---|---|---|---|---|---|---|
| Number | % | Number | % | Number | % | |
| Overall rate of major congenital anomalies | 30 | 3.0 | 79 | 3.0 | 3719 | 2.4 |
| Nervous system | 5 | 0.5 | 4 | 0.1 | 360 | 0.2 |
| Eye | 1 | 0.1 | 1 | 0 | 104 | 0.1 |
| Ear, face and neck | 1 | 0.1 | 0 | 0 | 48 | 0 |
| Circulatory system | 6 | 0.6 | 15 | 0.6 | 677 | 0.4 |
| Respiratory system | 1 | 0.1 | 2 | 0.1 | 67 | 0 |
| Cleft lip/cleft palate | 2 | 0.2 | 5 | 0.2 | 146 | 0.1 |
| Digestive system | 1 | 0.1 | 1 | 0 | 171 | 0.1 |
| Abdominal wall defect | 1 | 0.1 | 3 | 0.1 | 57 | 0 |
| Urinary system | 9 | 0.9 | 16 | 0.6 | 783 | 0.5 |
| Genital System | 1 | 0.1 | 5 | 0.2 | 370 | 0.2 |
| Musculoskeletal system | 3 | 0.3 | 11 | 0.4 | 466 | 0.3 |
| Other anomalies/syndromes | 3 | 0.3 | 6 | 0.2 | 478 | 0.3 |
| Genetic disorders | 3 | 0.3 | 15 | 0.6 | 448 | 0.3 |
Exclusion of fetuses exposed only during T2 and/or T3 to tramadol (and codeine) and those for which the diagnosis of a possible congenital anomaly cannot be made (spontaneous miscarriage, ectopic pregnancies, etc.).
Multivariate analysis did not show an increase in the rate of major congenital anomaly in fetuses/newborns exposed in utero to tramadol in the first trimester of pregnancy compared to fetuses/newborns exposed to codeine (ORa [95% CI] = 1.03 [0.67–1.57]) and those not exposed to these drugs (ORa [95% CI] = 1.24 [0.86–1.79]), after adjusting for maternal age, dispensation of folic acid, teratogenic drugs, history of diabetes, gestational diabetes, and long‐term adverse health conditions.
We performed 2 sensitivity analyses: one excluding genetic disorders and one including fetuses/newborns exposed to tramadol and codeine at least during the first 56 days of pregnancy (organogenesis period). The results of these analyses were similar to those found in the main analysis.
4. Discussion
In this study, comparing women exposed to tramadol with women exposed to codeine and women not exposed to tramadol and codeine during pregnancy, we found an association between prescription and dispensation of tramadol and pregnancy termination. However, no association between prescription and dispensation of tramadol in the first trimester of pregnancy and the occurrence of major congenital anomalies was observed.
Although tramadol is poorly evaluated in pregnant women, in EFEMERIS we find a sevenfold increase in the proportion of pregnancies exposed to tramadol between 2004 and 2020. This trend is reflected in the general population, with tramadol use increasing by 68% between 2006 and 2017 [5]. The withdrawal of dextropropoxyphene/paracetamol from the market in France in March 2011 has contributed in particular to this increase [6, 7]. In fact, the increase in tramadol use is not specific to the population of pregnant women.
In our analyses, we compared women exposed to tramadol during pregnancy with women exposed to codeine in order to have two populations as comparable as possible, codeine being, like tramadol, a weak opioid analgesic. These two groups are indeed comparable with regard to certain characteristics such as hypertension, diabetes, and occupational category, but we nevertheless find a higher rate of long‐term adverse health conditions and a lower level of education in the group exposed to tramadol. These differences are much more pronounced when comparing the tramadol‐exposed group with the unexposed group. The differences in the rates of women with a long‐term adverse health conditions among those exposed to tramadol and codeine compared with unexposed women may be explained by the fact that certain long‐term conditions such as diabetes, rheumatoid arthritis, etc. can cause pain and require the use of analgesics. In addition, women exposed to tramadol and codeine have a lower level of education and are more often unemployed. This may be explained by the fact that lower socio‐economic levels are associated with a higher risk of chronic disease [8] and pain.
We found a spontaneous pregnancy termination rate twice as high among women who were exposed to tramadol, especially early miscarriage (≤ 22 weeks of amenorrhea). The 2022 study, based on the Danish national birth and prescription registries and including 36 467 pregnancies, including 7310 exposed to tramadol in the first trimester, did not find an association between tramadol use during pregnancy and the risk of spontaneous abortion [3]. We therefore questioned whether there was an indication bias. Indeed, tramadol may have been prescribed and dispensed for some women in anticipation of abortion pain. Although paracetamol or ibuprofen are prescribed and dispensed in the first‐line setting, the subsistence of pain sometimes requires the use of stronger analgesics such as codeine or tramadol [9, 10]. We also investigated the use of nonsteroidal anti‐inflammatory drugs (NSAIDs) in our exposed population and their potential role in this increased risk of abortion. Studies have reported an increased risk of spontaneous miscarriage in women exposed to NSAIDs early in pregnancy [11, 12]. We therefore performed additional analyses by excluding the women who were dispensed NSAIDs during their pregnancy in the three groups. We still found a higher rate of spontaneous termination of pregnancy in women exposed to tramadol and not to NSAIDs compared with women exposed to codeine and not to NSAIDs and unexposed women. The population of women exposed to NSAIDs was probably underestimated because we cannot exclude the use of these drugs for self‐medication. Furthermore, these results must be interpreted with caution because in EFEMERIS, as in many medical and administrative databases, miscarriages are not recorded exhaustively (very early miscarriages are often not identified).
Animal studies have not shown any teratogenic effects associated with tramadol. However, very few studies evaluating this association in humans have been published. The two publications found in the literature show divergent results, which led us to conduct this study. Kallen's study, published in 2015 [2] showed an association between tramadol use during the first trimester of pregnancy and the risk of congenital anomalies, particularly cardiovascular anomalies and talipes equinovarus. Conducted from the Swedish Medical Birth Registry, it included 1 682 846 pregnant women, of whom 1751 had used tramadol during the first months of their pregnancy. However, whether or not to take tramadol was based on self‐reported questionnaires completed by the pregnant women at the end of the first trimester, which can lead to a recall bias (women who have had a malformed child have a better recall of drugs taken during pregnancy). In our study, we did not find an increased risk of skeletal and cardiovascular anomalies as observed in this study. Like our work, the second study in 2022 [3], which was carried out using the Danish national drug prescription database and including 18 907 pregnancies, 3796 of which were exposed to tramadol, found no association between tramadol use and risk of congenital anomaly.
Our study has limitations, particularly those related to the use of medical and administrative databases. Health Insurance does not collect information on medicines taken during hospitalizations, or medicines that are dispensed but not reimbursed. In addition, we cannot measure self‐medication, which may lead us to underestimate the number of medications taken during pregnancy. In addition, EFEMERIS allows us to know which medicines have been prescribed and dispensed, but we cannot be sure that they have been taken and when. Finally, some information about the pregnant women is missing, such as alcohol consumption, which is a known risk factor for congenital anomalies, drug indications, or family history of congenital anomalies. In fact, we were not able to adjust for these confounding factors in the analyses. Finally, the EFMERIS population includes pregnant women in Haute‐Garonne. However, the characteristics of the women and neonates in this database are comparable to those of the general French population [4].
Among the strengths of our study, we would like to point out that EFEMERIS is a specific database for the assessment of the teratogenic risk of medications, as it links data on drug prescriptions and dispensations during pregnancy with pregnancy outcomes [4]. The drug prescription and dispensation data provided by the French National Health Insurance Agency make it possible to obtain reliable and accurate information on medications dispensed during pregnancy and to avoid the recall bias that can occur when patients' medication use is collected by questionnaire. In EFEMERIS, we also find information about the women such as age, occupation, and level of education, as well as some diseases occurring during pregnancy, which may be risk factors for congenital anomalies or pregnancy termination and which were taken into consideration in our multivariate statistical analyses. The EFEMERIS population includes around 86% of children born in Haute‐Garonne, which represents a sample of sufficient size to perform studies evaluating the congenital anomaly risk of drugs [4]. Finally, we compared pregnancies exposed to tramadol with those exposed to codeine in order to have comparable populations, which is a strong point.
5. Conclusion
This study found an association between tramadol use and the risk of spontaneous pregnancy termination. We cannot conclude that there is a causal relationship because of a possible indication bias. This finding needs to be confirmed by specific additional studies using different methods and data sources.
However, no association was observed between tramadol prescription and dispensation during the first trimester of pregnancy and an increased risk of major congenital anomalies. It therefore provides additional data on the risk of congenital anomalies of tramadol in pregnant women.
Ethics Statement
The EFEMERIS cohort was approved by the French Data protection Agency on April 7, 2005 (authorization number 05‐1140). This study was performed on anonymized patient data.
Conflicts of Interest
The authors declare no conflicts of interest.
Acknowledgments
We thank our data providers who made anonymized data available for our research institution: the Haute‐Garonne health insurance system, the Haute‐Garonne mother and child welfare service, the Prenatal Diagnosis Centre, and the Haute‐Garonne hospital medical information system of Toulouse University Hospital.
Funding: The present study is part of the REproduction Gestation And Reproduction (REGARDS) research program funded by the French National Agency for the Safety of Medicines and Health Products (Agence Nationale de Sécurité du Médicament et des Produits de Santé, ANSM). This program aims to improve drug monitoring in pregnant women in France. Allocation of funding did not impact the study design and conduct, or the collection, management, analysis, and interpretation of data, or the preparation, review, and approval of the manuscript.
References
- 1. Beakley B. D., Kaye A. M., and Kaye A. D., “Tramadol, Pharmacology, Side Effects, and Serotonin Syndrome: A Review,” Pain Physician 18, no. 4 (2015): 395–400. [PubMed] [Google Scholar]
- 2. Källén B. and Reis M., “Use of Tramadol in Early Pregnancy and Congenital Malformation Risk,” Reproductive Toxicology 58 (2015): 246–251. [DOI] [PubMed] [Google Scholar]
- 3. Sørensen A. M. S., Noergaard M. M., Gotfredsen D. R., et al., “Exposure to Tramadol During Early Pregnancy and Risk of Spontaneous Abortion or Major Congenital Malformations,” Obstetrics and Gynecology 139, no. 4 (2022): 545–553. [DOI] [PubMed] [Google Scholar]
- 4. Lacroix I., Hurault C., Sarramon M. F., et al., “Prescription of Drugs During Pregnancy: A Study Using EFEMERIS, the New French Database,” European Journal of Clinical Pharmacology 65 (2009): 839–846. [DOI] [PubMed] [Google Scholar]
- 5. Monzon E. and Richard N., Etat des lieux de la consommation des antalgiques opioïdes et leurs usages problématiques (Agence Nationale de Sécurité du Médicament, 2019), https://www.omedit‐idf.fr/wp‐content/uploads/2020/01/Rapport_Antalgiques‐Opioides_Fev‐2019_3.pdf_2019‐03‐06.pdf. [Google Scholar]
- 6. Commission de la Transparence , Avis tramadol (chlorhydrate de) (Haute Autorité de Santé, 2016), https://www.hassante.fr/upload/docs/evamed/CT15342_CONTRAMALMONOTRAMAL_PIS_RI_Avis2_CT15342.pdf. [Google Scholar]
- 7. Bismuth S., Leng E. L., Oustric S., Montastruc J. L., and Lapeyre‐Mestre M., “Quel(s) antalgique(s) après le retrait du dextropropoxyphène ? Enquête auprès de médecins généralistes de la région Midi‐Pyrénées,” Thérapie 66, no. 1 (2011): 25–28. [DOI] [PubMed] [Google Scholar]
- 8. Pathirana T. I. and Jackson C. A., “Socioeconomic Status and Multimorbidity: A Systematic Review and Meta‐Analysis,” Australian and New Zealand Journal of Public Health 42, no. 2 (2018): 186–194. [DOI] [PubMed] [Google Scholar]
- 9. Kemppainen V., Mentula M., Palkama V., and Heikinheimo O., “Pain During Medical Abortion in Early Pregnancy in Teenage and Adult Women,” Acta Obstetricia et Gynecologica Scandinavica 99, no. 12 (2020): 1603–1610. [DOI] [PubMed] [Google Scholar]
- 10. Khazin V., Weitzman S., Rozenzvit‐Podles E., et al., “Postoperative Analgesia With Tramadol and Indomethacin for Diagnostic Curettage and Early Termination of Pregnancy,” International Journal of Obstetric Anesthesia 20, no. 3 (2011): 236–239. [DOI] [PubMed] [Google Scholar]
- 11. Nielsen G. L., Sorensen H. T., Larsen H., and Pedersen L., “Risk of Adverse Birth Outcome and Miscarriage in Pregnant Users of Non‐steroidal Antiinflammatory Drugs; Population Based Observational Study and Case‐Control Study,” BMJ 322 (2001): 266–270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Li D.‐K., Liu L., and Odouli R., “Exposure to Non‐Steroidal Anti‐Inflammatory Drugs During Pregnancy and Risk of Miscarriage: Population‐Based Cohort Study,” BMJ 327 (2003): 368–370. [DOI] [PMC free article] [PubMed] [Google Scholar]