Page 67 - Fertility in Women with Rheumatoid Arthritis Vruchtbaarheid van vrouwen met reumatoïde artritis

The DAS28 tended to be higher in women who miscarried (mean ± SD 3.92 ± 0.94 versus 3.59 ± 1.17 in women who had not miscarried; P = 0.166), and more women with a miscarriage had received MTX therapy in the past (82% versus 68%; P = 0.174). There were no signi cant differences in RF positivity, periconception use of NSAIDs, sulfasalazine use, prednisone use, and median cumulative number of DMARDs.
Table 2 - Univariate analysis of covariates in patients with rheumatoid arthritis who did and those who did not experience miscarriage*
Miscarriages in RA
Variable
Age, mean ± SD years Missing,no.
RF positive, no. (%) Missing,no.
ACPA positive, no. (%) Missing,no.
DAS28, mean ± SD Missing,no.
Medication use
MTX in past, no. (%)
NSAIDs periconception, no. (%) Sulfasalazine periconception, no. (%) Prednisone periconception, no. (%) Cumulative no. of DMARDs, median (IQR)§
Miscarriage
Yes (n = 28) No (n = 134) P
33.9 ± 3.9 32.0 ± 3.8 0.022† 0 1 –
22 (79) 92 (69) 0.478‡ 4 - 1 –
23 (82) 80 (60) 0.058‡ - 1 –
3.92 ± 0.94 3.59 ± 1.17 0.166† - 1 –
23 (82) 6 (21) 10 (36) 11 (39) 2 (1–3)
91 (68) 41 (31) 44 (33) 49 (37) 2 (1–3)
0.174‡ 0.371‡ 0.827‡ 0.473‡ 0.461¶
* RF = rheumatoid factor; ACPA = anti-citrullinated protein antibody; DAS28 = Disease Activity Score in 28 joints; MTX = methotrexate; NSAIDs = nonsteroidal antiinflammatory drugs; DMARDs = disease-modifying antirheumatic drugs; IQR = interquartile range.
† By Student’s t-test.
‡ By Fisher’s exact test.
§ Including biologic agents. ¶ By Mann-Whitney U test.
Multivariate analysis
We performed logistic regression analysis with the occurrence of miscarriage as the dependent variable and maternal age, ACPA positivity, DAS28 and previous MTX use as independent variables. Two patients who did not miscarry had missing variables and were excluded from the analysis. None of the variables showed a statistically signi cant change in the risk of miscarriage. Two variables (DAS28 and past MTX use) had a subsequent P value for association with miscarriage of >0.10, and were therefore omitted from the model. After omitting the DAS28 (P = 0.227) and subsequently excluding past MTX use (P = 0.113) as independent variables, none of the remaining variables (increasing age and presence of ACPAs) showed any statistically signi cant effect (P = 0.065 and P = 0.092, respectively) (Table 3).
65