Pregnancy outcomes in women with Familial Mediterranean Fever receiving colchicine: Is amniocentesis justified?




To evaluate the outcome of pregnancies in women with familial Mediterranean fever (FMF) who are taking colchicine, and to reconsider the justification for amniocentesis in these women.


The outcome of 179 pregnancies in a group of women with FMF taking colchicine was compared with the outcome of 197 pregnancies in women with FMF who did not take colchicine during pregnancy and with 312 pregnancies in another cohort of healthy pregnant women of similar age and ethnicity.


There was no difference in the 3 groups regarding early abortions, late abortions, or congenital malformations. There was a mild trend towards a better outcome for the colchicine-treated group but these results did not reach statistical significance.


Treatment with colchicine during pregnancy in patients with FMF is beneficial in controlling the disease while not affecting the outcome of the pregnancy; therefore there is no justification for recommending amniocentesis in women taking colchicine solely because of this treatment.


Familial Mediterranean fever (FMF) is a periodic disease characterized by recurrent attacks of fever accompanied by peritonitis, pleuritis, arthritis, or erysipelas-like skin lesions (1, 2). The disease affects mainly Turks, Armenians, Arabs, and non-Ashkenazi Jews. One of the most devastating complications of FMF is amyloidosis, which mainly affects the kidneys but may also involve other organs and tissues. Colchicine is the drug of choice for FMF (3) because it controls the acute attacks and prevents the development of amyloidosis. Since FMF affects women during their child-bearing years, a concern has been raised regarding the effect of colchicine on the reproductive system and on the outcome of the pregnancy (4).

Patients with FMF have a variable pregnancy course. Some enjoy an attack-free period during pregnancy, whereas others experience devastating attacks. Acute peritonitis of FMF in pregnant women may lead to premature contractions of the uterus and eventual miscarriage. Therefore, it is necessary to control FMF attacks through colchicine treatment during pregnancy.

The potential teratogenic effect of colchicine arises from its effect on microtubules and the possible interruption of spindle formation in tissues associated with high mitotic activity (5). In a study by Shoji and Makino, the effects of colchicine on mouse embryos were investigated with special regard to the frequency of fetal malformations and death (6). Mice were injected with 0.5–2.5 mg/kg of colchicine in a single subcutaneous injection, (which was equivalent to a minimum dose of 30 mg of colchicine in a patient weighing 60 kg). The frequency of dead fetuses tended to increase as the colchicine dose increased and as the embryos were younger at the time of the drug injection. Gross malformations, such as exencephaly, maxillary defects, and facial defects were noted exclusively in the colchicine-treated mice. Based on these observations, the colchicine manufacturer's brochure contains a clear warning not to use the medication during pregnancy or during breastfeeding.

Owing to the reservations regarding the safety of colchicine use during pregnancy and because of legal reasons (the manufacturer's warning), our policy has been to recommend amniocentesis at mid-second trimester to all women with FMF taking colchicine during pregnancy. Amniocentesis carries the risk of intra-amniotic infection (chorioamnionitis) and risk of fluid leak, which may lead to fetal loss at a rate of 1:200 (7). Therefore, one should seriously weigh the risk of hypothetical colchicine teratogenicity with those potentially proven complications associated with amniocentesis.

The above data and observations led us to conduct a retrospective study in which we analyzed the outcome of pregnancies in women with FMF who took colchicine during pregnancy. We compared the outcomes of that group with the outcome of pregnancies in FMF patients who did not take colchicine during pregnancy and with a control cohort of age- and ethnicity-matched healthy women. Our main goal was to look for possible differences that could justify or refute our current policy of recommending amniocentesis to patients who take colchicine during pregnancy.



Women with FMF were recruited from April 2004 to May 2008 from our Center for FMF at Hadassah University Medical Center. Only women who did not have any other concomitant disease (with the exception of FMF) and who had children were included in the study. All patients were clinically diagnosed with FMF according to the Tel Hashomer criteria (8). Some patients had undergone genetic analysis although it was not required to confirm the FMF diagnosis. The patients visited the FMF clinic once every 4–6 months for a physical examination, blood count, and urinalysis. During these visits the patients completed a questionnaire regarding their ethnic origin, general health, whether they had fertility problems, how many pregnancies they had, how many children they have, the course and outcome of the pregnancies and the newborns, whether they took colchicine during pregnancy, the doses of colchicine during these periods, if there were any early or late abortions, and whether the newborns had any congenital malformations or diseases. The patients were also asked whether they underwent amniocentesis, and if the baby was delivered vaginally or by cesarean section. The questionnaire also included questions regarding disease duration (FMF), clinical manifestations, age at onset, age at diagnosis, and age when colchicine administration was begun. Oral informed consent was obtained from each participant at the time of the questionnaire completion.

Early abortion was defined as fetal loss at <14 weeks of pregnancy, whereas late abortion was defined as fetal loss at ≥14 weeks. Malformations were detected by the treating pediatrician either at birth or at a later time. The malformation was defined as major if it was fatal, was likely to cause a serious handicap, or required surgery (9), and minor malformations did not fulfill these criteria.

Primarily, a retrospective study design was used. However, FMF patients who gave birth during the recruitment years were followed prospectively. The FMF patients were divided into 3 groups: patients with FMF who took colchicine throughout each of their pregnancies (group 1); patients with FMF who took colchicine during some of their pregnancies (group 2); and patients with FMF who did not take colchicine during any of their pregnancies, although they took the medication between the pregnancies (group 3). Patients in groups 2 and 3 who did not take colchicine during pregnancy stopped taking the medication 2–3 months before conception.

Eighty-four healthy women (312 pregnancies) served as additional control subjects. The women were recruited from the gynecology clinic by one investigator (A B-C) and were matched with the FMF patients (groups 1 and 2) by age and ethnicity. Women who were at risk for a complicated pregnancy were excluded from the study.

Statistical analysis.

The comparison of continuous variables between the study groups was carried out using analysis of variance with Dunnett's post test. The chi-square test was applied in order to compare the study groups for categorical variables. The comparison of paired categorical data within one of the study groups was performed using McNemar's test. All tests applied were two-tailed and P values less than or equal to 0.05 were considered statistically significant.


In our clinic we actively follow ∼900 patients with FMF of whom 60% are <20 years. Of the remainder, 165 women were eligible for the study. Of these 165, 28 patients could not be contacted due to a change in address or telephone number, 5 women were excluded due to concomitant diseases (2 with anti-phospholipid antibodies syndrome, 1 with systemic lupus erythematosus, 1 with diffuse amyloidosis [due to FMF], and 1 with psychiatric problems that did not allow her to participate in the study). Thus, 132 women with FMF who had a total of 480 pregnancies were recruited for this study from April 2004 to May 2008. The group of patients who took colchicine (1.0–1.5 mg daily) during all their pregnancies (group 1) consisted of 61 women. Twenty-two patients used colchicine during some of their pregnancies (group 2), and 49 patients with FMF patients did not take colchicine during the pregnancy (group 3), many due to the fear of teratogenicity. However, those in group 3 did take colchicine between pregnancies. The dose of colchicine was stable during pregnancies in most patients since their disease was controlled. All the groups were comparable regarding ethnic origin, the percentage that underwent genetic testing, and MEFV mutations distribution (data not shown). Additional patient demographic information is shown in Table 1.

Table 1. Demographic details of the 3 groups analyzed in the study*
 All pregnancies with colchicine (group 1)Some pregnancies with colchicine (group 2)No pregnancies with colchicine (group 3)Compared groups P
  • *

    Values are the mean ± SD (range) unless indicated otherwise. FMF = familial Mediterranean fever.

  • In the compared groups, the first value is by Dunnett's post test and the second value by analysis of variance.

FMF patients, no.612249   
Age at study end, years37.8 ± 8.0 (22–60)40.6 ± 8.6 (27–55)49.4 ± 11.8 (27–69)0.17< 0.00010.002
Age at FMF onset, years8.2 ± 5.0 (2–24)11.5 ± 8.4 (3–34)16.4 ± 9.5 (2–40)0.03< 0.00010.004
Age at colchicine initiation, years14.6 ± 6.0 (3–29)20 ± 9.4 (4–36)32.3 ± 10.2 (11–50)0.002< 0.0001< 0.0001
Age at first pregnancy, years24.4 ± 3.7 (18–32)23 ± 4.3 (19–36)23.2 ± 4.8 (16–40) 0.210 
Follow up      
 Age of the first child at time of study, years12.7 ± 8.5 (2–35)17.0 ± 8.1 (4–34)25.8 ± 11.6 (3–47)   
 Age of the last child at time of study, years7.0 ± 6.1 (1–23)8.5 ± 5.3 (1–20)18.0 ± 9.8 (1–37)   

There was a significant difference between the 3 groups regarding age of the patients at the end of the study, age at disease onset, and age when starting colchicine treatment. However, the age of the women with FMF at the time of their first pregnancy was similar in all 3 groups. When applying Dunnett's tests for detecting which pairs of groups mainly contributed to this difference, it was found that the mean ± SD age at the end of the study, age at FMF onset, and age at the initiation of colchicine treatment of the women in group 3 differed significantly from the patients in both groups 1 and 2. Table 1 also shows the length of the followup periods of the women and their children. The lowest average time was in group 1 (mean ± SD 7.0 ± 6.1 years), while in groups 2 and 3 this period was longer. This longer followup period allows detection and tracing of late-onset complications related to events occurring during pregnancy or delivery.

The pregnancy outcomes of the 3 groups are summarized in Table 2. The average number of pregnancies per women were quite similar between the groups, and the cesarean section rate was similar in group 1 and 3. However, the number of early abortions, late abortions, and congenital malformations were higher in groups 2 and 3 compared with group 1 but did not reach the level of statistical significance. The rate of amniocenteses was remarkably higher among the patients in group 1 because they were taking colchicine and therefore were encouraged to undergo this procedure in the second trimester of pregnancy. However, many did not undergo amniocentesis due to religious reasons. All of the results of the amniocenteses were normal and the outcome of the procedure was uneventful.

Table 2. Pregnancy outcome in the 3 groups*
 All pregnancies with colchicine (group 1)Some pregnancies with colchicine (group 2)No pregnancies with colchicine (group 3)
  • *

    All P values not significant. FMF = familial Mediterranean fever; VSD/ASD = ventricular septal defect, atrial septal defect; Fallot's = Fallot's syndrome.

FMF patients, no.612249
Pregnancies, no.179104197
Average pregnancies/patient, no.34.74
Cesarean section, no. (%)17 (9.8)5 (4.8)18 (9.1)
Early abortions, no. (%)16 (8.9)23 (22)27 (13.7)
Late abortions, no. (%)2 (1.1)3 (2.8)4 (2.0)
Major congenital malformations*1 VSD/ASD1 died, 1 cleft lip1 Fallot's, 1 cleft lip
Minor congenital malformationsNone1 hydronephrosisNone
Amniocenteses, no. (%)38 (21.2)11 (10.5)5 (2.5)
Post-procedure abortions/complications, no.000

In group 2 some of the pregnancies in the same patient were while the patient was taking colchicine, and other pregnancies were without colchicine; therefore, we tried to compare the outcome of the pregnancies in these 2 subgroups (Table 3). Analysis of the outcome according to the pregnant patients is given through a cross tabulation (Table 4). Outcome refers to either early or late abortion in each of the treatment modes. For each woman under each treatment mode, outcome was defined as positive if at least 1 early or late abortion occurred during that treatment mode. As can be seen in Table 4, in 13 of 22 women the outcome was the same whether the patient was with or without colchicine treatment (in 9 women the outcome was negative under both treatment modes and in 4 the outcome was positive under both treatment modes). Among the women in whom outcome differed with or without treatment (9 [40%]), 6 had an early or late abortion when untreated but not when treated, whereas 3 had an early or late abortion when treated but not when untreated. This difference between the 6 and 3 women was not statistically significant (P = 0.508 by McNemar's test) although a trend in favor of treatment with colchicine during pregnancy was noted.

Table 3. Analysis of group 2, women with FMF who took colchicine for some of their pregnancies*
 Some pregnancies with colchicineSome pregnancies without colchicine
  • *

    All P values not significant. FMF = familial Mediterranean fever.

Pregnancies, no.5351
Early abortions, no. (%)10 (18)12 (23.5)
Late abortions, no.12
Major congenital malformations, no.2
Minor congenital malformations, no.1
Table 4. Cross tabulation of FMF patients in group 2 (women with FMF who took colchicine for some of their pregnancies)*
 Pregnancy without treatmentTotal
Without early/late abortionWith early/late abortion
  • *

    Numbers shown are the numbers of women in group 2, not the number of pregnancies (for details see Results). FMF = familial Mediterranean fever.

  • P = 0.508 by McNemar's test.

Pregnancy with treatment   
 Without early/late abortion9615
 With early/late abortion347

The control group included 84 healthy women who had a total of 312 pregnancies. Forty-six of those pregnancies terminated in early abortion and 6 in late abortion, and 2 pregnancies resulted in infants with a major malformation (a large diaphragmatic hernia, and translocation of chromosome 3), and 1 infant with mild muscle weakness (Table 5).

Table 5. Analysis of pregnancy outcomes of group 1, subgroup 3, and the healthy control group
 All pregnancies with colchicine (group 1)No pregnancies with colchicine (subgroup 3)*Healthy controls
  • *

    Patients from group 3 who were younger than 53 years at the end of the study.

  • P values not significant.

Patients, no.612984
Age, mean ± SD years37.8 ± 8.041.7 ± 8.540.8 ± 11
Pregnancies, no.179100312
Early abortions, no. (%)16 (8.9)9 (9)46 (14)
Late abortions, no. (%)2 (1.1)3 (3)6 (2)
Major congenital malformations, no. (%)1 (0.5)2 (2)2 (0.6)
Minor congenital malformations, no.1


Two factors may theoretically affect fertility and pregnancy outcomes in women with FMF, the disease itself and colchicine treatment. Regarding the first factor, previous studies have shown that recurrent attacks of FMF caused peritoneal fibrosis leading to scarring of the salpinx, thereby leading to mechanical (secondary) infertility (10). Similarly, early studies showed a higher rate of abortion among patients with FMF compared with healthy women (11). This higher rate was also attributed to the attacks of peritonitis that the patients with FMF experienced, which led to premature uterine contractions.

In order to address the potential role of the second factor, colchicine, we investigated pregnancy outcomes in women with FMF. We evaluated the outcome of 480 pregnancies in 132 FMF patients, some of whom did not take colchicine during pregnancy. In fact, the group without colchicine treatment served as a control to the group treated with colchicine. Furthermore, we also had an additional control group of 84 healthy patients who do not have FMF and who were comparable to the study group in age and ethnic origin.

Although the age of the women at their first pregnancy was similar among the 3 groups, there were some demographic differences. Group 3 was older at the time of the study's end and their FMF diagnosis and colchicine treatment were delayed compared with the other 2 groups. One of the reasons for this delay was the lack of awareness of the diagnosis of FMF at that time. Secondly, some of the women in group 3 were diagnosed as having FMF before colchicine was introduced as treatment for this disease in 1972, explaining their older age and the lag time between diagnosis and treatment.

The outcome of pregnancies in group 1 was not inferior to that of the healthy cohort, despite colchicine treatment. The internal analysis of the patients in group 2, in whom some of the pregnancies were while the patients were receiving colchicine while others were without treatment, served as an additional built-in control. Tables 3 and 4 show that the outcome of pregnancy in patients receiving colchicine was not inferior to the pregnancies during which the same patients did not receive colchicine treatment; if anything, the outcome was even slightly better.

In a review of 116 colchicine-treated mothers with 225 pregnancies, 40 were treated during the first trimester, 91 were treated to term, and 94 were untreated (12). Neither the colchicine-treated mothers nor their children were found to be adversely affected during 10 years of followup. Spontaneous abortions were more prevalent in the untreated group (20.2%) than in the treated group (12.2%), an outcome possibly related to attacks of peritoneal irritation in untreated FMF patients. There were also 2 cases of trisomy 21 (1 live birth and 1 stillborn) in the treated patients, raising the rare possibility that it may have been related to colchicine treatment. However, due to the low number of these cases the authors considered it to be a chance coincidence. In a recent retrospective study, data were collected in delivery rooms, and the outcomes of 239 pregnancies in FMF patients with or without colchicine treatment were compared with the outcomes of pregnancies in healthy women (13). Again, no difference was observed between the 2 groups.

Our study included more than twice the number of pregnancies reported in the studies by Rabinovitch et al (12) and Ofir et al (13) and our results are quite similar. The number of early abortions was relatively higher in the untreated group as was also shown by Rabinovitch et al. The number of late abortions and congenital malformations was also higher in this group, although not of statistical significance. In none of our cohorts was trisomy 21 found. These findings suggest that colchicine is not an offending agent in FMF pregnancies and does not lead to a worse outcome compared with pregnancies in which the medication was avoided. Furthermore, colchicine treatment is beneficial in preventing FMF attacks, leading to a better quality of life during pregnancy with less episodes of peritonitis, which can cause premature contractions and abortions. This observation supports our policy to recommend treatment with colchicine during pregnancies to all FMF patients.

Previously, amniocentesis was recommended because of legal issues due to the warnings of the pharmaceutical companies and because of animal studies showing teratogenicity at a much higher dose of colchicine than is ever given to humans. In the present study, the good outcome of the pregnancies among the treated FMF patients and the lack of a higher rate of malformations among their offspring suggest that there is no need to perform amniocentesis in patients undergoing colchicine treatment. As a matter of fact, in Turkey where FMF is quite common, physicians do not recommend amniocentesis for patients taking colchicine during pregnancy (Ozdogan H: personal communication). This policy in Turkey was based upon experts' opinion rather than on firm scientific grounds; however, the result of our present study does support this approach.

The present study has a number of limitations. First, the use of a retrospective design may lead to recall bias and selection bias; however, this potential bias may be applied to all the studied groups. Furthermore, it should be emphasized that there is no way to conduct a similar prospective study since it would be unethical to discontinue colchicine in one group of FMF pregnant women while continuing it in the other group. Secondly, there is a significant difference in the age of the patients at the end of the study in group 3 compared with groups 1 and 2. This difference may raise a theoretical possibility that improvement in pregnancy care over the years is responsible for the somewhat better outcome in the treated patient rather than the colchicine itself. Regarding this point, it should be stressed that there has been no improvement in the rate of late abortions or malformations during the last few decades since changes in health care cannot affect chromosomal defects responsible for this outcome.

Furthermore, when we reanalyzed all the FMF patients in group 3 who were younger than 53 years at the end of the study, we found 29 women with a mean ± SD age that was similar to that of the patients in group 1. The patients in group 3 had a total of 100 pregnancies with 9 early abortions, 3 late abortions, and 2 major congenital defects (Table 5). There was no significant difference between this subgroup and the patients of group 1. This finding suggests that the age gap between the groups did not play a significant role. Moreover, the addition of the normal healthy control group and the analysis of group 2, which showed similar results (no drawback for colchicine treatment during pregnancy) further support our final conclusions. Finally, the results of the present study are also in accord with those of the relatively older study by Rabinovitch et al (12), and with the very recent one by Ofir et al (13). It should be stressed that the present study is a noninferiority study where we found that treatment with colchicine did not worsen the outcome of pregnancies compared with the nontreated group, and this was clearly shown.

In summary, our study supports changing our current policy. We should not recommend routine amniocentesis for FMF women taking colchicine during pregnancy since colchicine is relatively safe, beneficial, and does not change the pregnancy outcome. The decision regarding amniocentesis or other investigations should be made individually, after taking all factors contributing to pregnancy risks into account rather than colchicine treatment alone.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Eldad Ben-Chetrit had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Avraham Ben-Chetrit, Eldad Ben-Chetrit.

Acquisition of data. Eli Ben-Chetrit, Avraham Ben-Chetrit, Berkun, Eldad Ben-Chetrit.

Analysis and interpretation of data. Eli Ben-Chetrit, Avraham Ben-Chetrit, Eldad Ben-Chetrit.