Levels of Serum Anti–Müllerian Hormone, a Marker for Ovarian Reserve, in Women With Rheumatoid Arthritis


Erasmus MC, Room Na609, PO Box 2040, 3000 CA Rotterdam, The Netherlands. E-mail: j.brouwer.1@erasmusmc.nl



Fertility is reduced in women with rheumatoid arthritis (RA), even before diagnosis. This may be due to a diminished ovarian reserve. The current study examined serum levels of anti–Müllerian hormone (AMH), the most reliable endocrine marker for ovarian reserve, in early RA patients and the influence of disease activity and methotrexate (MTX) use on AMH concentrations.


Serum AMH levels were measured in 72 women with recent-onset RA ages 18–42 years and compared to 509 healthy women. The association between AMH and rheumatoid factor (RF), anti–cyclic citrullinated peptide (anti-CCP), erosions, C-reactive protein (CRP) level, disease activity (Disease Activity Score in 28 joints [DAS28]), and use of MTX was assessed.


At diagnosis, age-adjusted serum AMH levels did not differ significantly between patients and controls (P = 0.254). AMH levels were not related to the presence of RF (P = 0.487), anti-CCP (P = 0.686), or erosions (P = 0.350), and showed no significant correlation with CRP levels (r = −0.207, P = 0.083) or disease activity scores (DAS28; r = 0.007, P = 0.955). After 6 months of treatment, AMH levels in patients (n = 53) were lower than at the time of diagnosis (P < 0.001), but did not differ from controls (P = 0.741). There was no significant difference in AMH values after 6 months of treatment between patients who did (n = 31) or did not (n = 22) receive MTX (P = 0.287).


AMH levels in women with early RA are comparable to those of healthy controls, indicating that the reduced fertility in this patient group is not caused by diminished ovarian reserve. AMH levels are not affected either by disease activity or by short-term MTX use.


Rheumatoid arthritis (RA) can already be manifest in women during their reproductive years and might therefore reduce their ability to have children. Female RA patients experience more fertility problems compared to healthy controls, resulting in a longer mean time to pregnancy ([1-3]). Reduced fertility seems to be present before diagnosis in female RA patients ([1]). Since on average women with RA appear to reach menopause at an earlier age compared to controls ([1]), the question is raised whether reduced fertility in women with RA is related to a compromised ovarian reserve.

In fertility clinics, the ovarian reserve, which constitutes the number of primordial follicles, is estimated using serum anti–Müllerian hormone (AMH) levels. AMH is a member of the transforming growth factor β family and is produced in the ovary by granulosa cells of early developing follicles ([4]). In both healthy and subfertile women, there is a strong correlation between serum AMH levels and the number of developing follicles in the ovaries, which declines with advancing age until the follicle pool is nearly depleted and the woman enters menopause. Due to this gradual decrease of quantity, which seems to go hand in hand with a decrease in quality of the oocytes harbored by the ovarian follicles, a woman will become infertile approximately 10 years before she will enter menopause. The age at which a woman reaches menopause, as well as the preceding period of decreased fertility and infertility, varies greatly among women. At present, serum AMH is the most reliable predictor for the age at which a woman will enter menopause ([5]). Little is known about the effect of disease activity, parameters of inflammation, or use of antirheumatic drugs on measurement of serum AMH levels.

Since patients with RA who try to conceive take adjusted, and often less effective, medication, they frequently have increased disease activity. Thus far, no reports on the effect of RA disease activity on serum measurements of AMH are known.

Preferably, women with RA who want to conceive should be screened for increased risk of subfertility before alteration or cessation of medication, so they can be referred to a gynecologist in due time and the time to pregnancy will be as short as possible. Since many of these patients take methotrexate (MTX), currently the first-choice medication in newly diagnosed RA, the effect of this potentially harmful medication on AMH levels should be studied.

Therefore, the aim of this study was to compare serum AMH concentrations between women with early RA and healthy controls, and to assess the influence of parameters of disease activity and the use of MTX on serum AMH levels in women of childbearing age with RA.

Box 1. Significance & Innovations

  • Anti-Müllerian hormone (AMH) levels in women with early rheumatoid arthritis (RA) are comparable to healthy controls, indicating that the reduced fertility in this group is not caused by a diminished ovarian reserve.
  • AMH levels in women with early RA are not affected by disease activity.
  • AMH levels in women with early RA are not affected by short-term use of methotrexate

Patients and methods


Patient data were derived from the Rotterdam Early Arthritis Cohort (REACH), a prospective cohort study in the greater Rotterdam area that was started in July 2004 and is still ongoing ([6]). Enrolled patients had joint symptoms for less than 12 months that were not explained by trauma or overexertion. They did not receive any antirheumatic medication before enrollment. The Erasmus MC medical ethical review board approved the study protocol and written informed consent was acquired before inclusion. Patients were examined at several visits, including the first visit (T0) and 6 months later (T6). Patient characteristics, general history, disease activity scores, and medications prescribed were registered. When available, serum samples for each visit were stored. Radiographs of the hands and feet were made at baseline and assessed for bony erosions. For the current study, all women ages 18–42 years fulfilling the 2010 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) RA classification criteria were included ([7]). Serum samples from T0 and T6 visits were used.


An existing control population of healthy women was used ([8]). From this population, all women ages 18–42 years were selected (n = 509). All controls had regular menstrual cycles ranging from 25–35 days or were proven fertile. Twenty-nine percent of the controls had ever been pregnant (unknown for 60%). No additional fertility assessments were available. The controls did not receive any hormones or oral contraceptives during the 3 months prior to blood sampling. In case of proven fertility, blood sampling was at least 6 months after delivery. Sampling was performed randomly during the menstrual cycle ([8]).


Disease activity was scored using the Disease Activity Score in 28 joints (DAS28). Serum samples were stored at −80°C. Laboratory measurements included IgM rheumatoid factor (RF; by enzyme-linked immunosorbent assay [ELISA]), anti–cyclic citrullinated peptide (anti-CCP; by EliA CCP on ImmunoCAP 250, Phadia), C-reactive protein (CRP) levels (mg/liter, by local standards), and AMH levels (μg/liter, by in-house double-antibody ELISA; commercially available as Gen II Beckman Coulter). The range of the AMH standards used in the in-house AMH assay was 0.037–5 ng/ml ([8]).

Statistical analysis

Differences between the groups were calculated using Student's t-test or the Mann-Whitney U test for continuous variables and Fisher's exact test for categorical variables. Univariate analysis of covariance was performed to adjust for age differences between the groups. Correlations between AMH and CRP level and DAS28 were assessed by the partial correlation coefficient, correcting for age. Within-subject differences between T0 and T6 were studied using the paired-samples t-test or the Wilcoxon signed rank test. The proportional change in AMH levels between T0 and T6 was calculated as AMH (T6)/AMH (T0). Linear regression was performed to study the effect of different variables on proportional AMH change. Reference curves for serum AMH levels in the controls as a function of age were calculated from a linear regression model using a natural cubic spline fitted on log-transformed AMH values of the control group ([8]).

Values are shown as the mean ± SD or the median (interquartile range [IQR]). Two-sided P values less than 0.05 were considered significant. The statistical package used was SPSS Statistics, version 18.0.0 (IBM).


Patient data

Between July 2004 and February 2011, 174 women ages 18–42 years were enrolled in the REACH study. Of the enrolled women, 95 (55%) fulfilled the 2010 ACR/EULAR RA classification criteria. Three of the women were diagnosed with psoriatic arthritis, but still fulfilled the classification criteria and were included in the analysis. One patient had a history of unilateral ovariectomy because of severe endometriosis and was excluded from the study. The first (T0) and second (T6) visits were included in the analysis. Serum samples from T0 were available for 72 women, while 53 also had serum stored at T6. The median time between the 2 visits was 182 days (IQR 175–189 days).

The patients with available serum at T0 did not differ from the patients without available serum (n = 22) concerning age, body mass index (BMI), symptom duration, CRP level, or the proportion of patients positive for RF, anti-CCP, or erosions. At T0, included patients had a significantly lower mean ± SD DAS28 of 4.52 ± 1.29 compared to 5.18 ± 1.07 in patients with no serum available (P = 0.032) (Table 1). Information on obstetric history was available in 50% of the patients. Twenty-five percent had ever been pregnant. No additional data on fertility problems were available.

Table 1. Patient characteristics at the time of diagnosis*
 Patients for analysis (n = 72)Patients without available serum (n = 22)aP
  1. Values are the median (interquartile range) unless otherwise indicated. BMI = body mass index; RF = rheumatoid factor; anti-CCP = anti–cyclic citrullinated peptide; CRP = C-reactive protein; DAS28 = Disease Activity Score in 28 joints.
  2. aPatients without available serum at the first visit were excluded from further analyses.
  3. bDifferences between the 2 groups as calculated by the Mann-Whitney U test.
  4. cDifferences between the 2 groups as calculated by Fisher's exact test.
  5. dDifferences between the 2 groups as calculated by Student's t-test.
Age, years35.1 (30.1–39.2)34.4 (29.6–38.7)0.681b
BMI, kg/m224.2 (21.9–28.3)23.7 (22.1–32.7)0.531b
Missing, no.22 
Duration of symptoms, days115 (67–182)124 (75–306)0.500b
RF positive, no. (%)33 (46)8 (47)1.000c
Missing, no.5 
Anti-CCP positive, no. (%)39 (54)7 (41)0.422c
Missing, no.5 
CRP level, mg/liter7.0 (3.3–16.0)11.0 (4.0–20.0)0.167b
DAS28, mean ± SD4.52 ± 1.295.18 ± 1.070.032d
Positive erosions, no. (%)10 (14)2 (9)0.726c

AMH at the time of diagnosis

At T0, the median age of the patients was higher than that of the controls (35.1 years [IQR 30.1–39.2 years] and 29.4 years [IQR 23.8–34.3 years], respectively; P < 0.001). BMI did not differ significantly (24.2 kg/m2 [IQR 21.9–28.3] and 22.9 kg/m2 [IQR 21.2–25.7], respectively; P = 0.063). Median AMH values were 1.71 μg/liter (IQR 0.81–4.39) in the patients and 2.82 μg/liter (IQR 1.64–4.38) in the controls. When adjusted for age, AMH values did not differ significantly between the patients and controls (P = 0.254). Eight patients (11%) had AMH levels below the tenth percentile of the controls (Figure 1A).

Figure 1.

A, Serum anti–Müllerian hormone (AMH) levels in women with recent-onset rheumatoid arthritis (RA) at the time of diagnosis are comparable to those in a healthy control population. B, Serum AMH levels in women with recent-onset RA 6 months after diagnosis differentiating between patients who did or did not receive methotrexate (MTX) treatment during these 6 months. The lines show the 10th, 50th, and 90th percentiles of predicted AMH values in healthy controls.

When adjusted for age, there was no significant difference in AMH levels when comparing RF-positive with RF-negative women (P = 0.487). No differences in AMH concentrations were found between anti-CCP–positive and anti-CCP–negative women (P = 0.686). AMH levels were not significantly different when erosions were present (P = 0.350). Finally, there was no significant correlation between AMH and either CRP level (r = −0.207, P = 0.083) or DAS28 (r = 0.007, P = 0.955).

AMH 6 months after diagnosis

Blood samples were available for both visits in 53 patients. The mean ± SD DAS28 value decreased significantly during 6 months from 4.54 ± 1.29 to 3.02 ± 1.38 (P < 0.001), as did CRP levels (median 7.0 mg/liter [IQR 4.0–18.0] to 5.0 mg/liter [IQR 2.0–9.4]; P = 0.008).

Median AMH levels at T6 were significantly lower than at T0 (1.92 μg/liter [IQR 0.84–3.75] and 2.57 μg/liter [IQR 0.90–5.30], respectively; P < 0.001). AMH levels at T6 in patients were not different from those in controls (P = 0.741). Five women (9%) had AMH levels below the tenth percentile of controls. At T6, there was neither a significant correlation between CRP levels and AMH (r = −0.170, P = 0.247) nor between the DAS28 and AMH (r = 0.084, P = 0.563).

Thirty-one (58%) of 53 women were prescribed MTX at the first visit in dosages of 7.5–25 mg/week (mean dosage 20 mg/week). These women had significantly higher CRP levels (median 8.0 mg/liter [IQR 5.0–26.0]; P = 0.022) and DAS28 values (mean ± SD 4.91 ± 1.38; P = 0.009) at T0 than those who did not receive MTX (median CRP level 4.5 mg/liter [IQR 2.0–11.0] and mean ± SD DAS28 4.04 ± 0.96). The proportion of RF-positive women was higher in the MTX group (68% versus 23%; P = 0.002). At T0, age (P = 0.139), BMI (P = 0.804), symptom duration (P = 0.346), and age-adjusted AMH levels (P = 0.229) showed no significant difference between the 2 groups.

After 6 months of treatment, age-adjusted AMH levels did not differ significantly between those receiving MTX and patients not receiving MTX (P = 0.287). The proportional change in AMH levels did not differ between the 2 groups (P = 0.422) (Figure 2) and was not dependent on age, DAS28, presence of RF or anti-CPP, and MTX use. AMH levels did not differ significantly either between patients receiving MTX and controls (P = 0.394) or between patients without MTX and controls (P = 0.657) (Figure 1B).

Figure 2.

The individual change in serum anti–Müllerian hormone (AMH) levels from the time of diagnosis (T0) to 6 months later (T6) in A, rheumatoid arthritis (RA) patients who did not take methotrexate (MTX) and B, RA patients who took MTX during the 6 months following diagnosis. The proportional change in time of AMH levels between the 2 groups shows no significant difference (P = 0.422).


To our knowledge, this is the first study reporting on AMH serum levels in patients with RA. Our results show that serum AMH concentrations in women with early RA are comparable to those in healthy controls. Serum AMH levels in early RA are neither influenced by parameters of disease activity nor by use of MTX. This study hereby provides an answer to Clowse et al, who suggest AMH as a marker to detect subclinical damage to the ovary during and after cyclophosphamide therapy, but also point out that the impact of rheumatic disease and antirheumatic drugs on AMH expression is unknown ([9]).

Subfertility might already be present in RA patients at the time of diagnosis ([1]). However, according to current results, this does not appear to be due to a reduced ovarian reserve. Since earlier menopause has been reported especially in RF-positive patients with RA ([1]), autoimmunity could play an important role in the early exhaustion of the primordial follicle pool. In RA, autoantibodies such as RF and anti-CCP can be found years before the disease becomes manifest ([10]). These autoantibodies are thought to represent a subgroup of RA patients with the most severe disease and the highest chance for extraarticular manifestations. Therefore, in case a decreased ovarian reserve might already be present early after diagnosis, it is most likely to be found in this subgroup of patients. The results of this study, however, do not confirm the hypothesis of a decreased ovarian reserve at the time of diagnosis in either anti-CCP–positive or RF-positive women.

Another possible explanation for early menopause is MTX, the first-choice medication for newly diagnosed RA patients. Women with RA who want to become pregnant may have been taking MTX for a considerable amount of time. However, there is a lack of studies on the effect of long-term low-dose MTX treatment on ovarian reserve in humans. In rats, daily administration of low- (0.05 mg/kg) or high-dose (0.15 mg/kg) MTX for 20 days showed a dose-dependent loss of vaginal cyclicity and hormonal changes toward postmenopausal values. Ovarian preantral and antral follicle growth was reduced. This was already apparent in low-dose MTX treatment and became even more distinct in rats that received higher doses ([11]). However, in our study, there was no effect of MTX on AMH levels, suggesting that short-term MTX does not affect ovarian reserve. Further research is needed to elucidate the effect of long-term low-dose MTX treatment on fertility in women with RA. Unfortunately, the number of premenopausal women in the REACH study with a followup of 1 or more years is too small to study the long-term effects of MTX treatment on fertility in this group.

In the current study, detailed information on obstetric and gynecologic history was available for a limited number of patients. Since it is reported that serum AMH levels are not altered by use of hormonal contraception or parity, it is highly unlikely that these missing data would affect the outcome of this study ([12, 13]). Timing of blood sampling in the menstrual cycle was not available, but it has been shown that serum AMH levels do not differ significantly throughout the menstrual cycle ([14]).

In conclusion, the current study shows that AMH levels in patients with RA are not influenced by disease activity and use of MTX. Furthermore, the results do not confirm reduced fertility based on a demise of the primordial follicle pool at the time of diagnosis in women with RA, since the ovarian reserve, as measured by AMH levels, is comparable to the control group. Long-term followup of female RA patients should clarify the role of ovarian reserve and other clinical factors in subfertility in this patient group.


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 published. Dr. Brouwer 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. Brouwer, Laven, Hazes, Schipper, Dolhain.

Acquisition of data. Brouwer, Laven, Hazes, Dolhain.

Analysis and interpretation of data. Brouwer, Laven, Hazes, Schipper, Dolhain.