Background/Purpose: SLE women with unplanned pregnancies might be at
increased risk of both disease and pregnancy complications. Until now, there
has been no population-based estimate of the induced abortion (IA) rate in
women with SLE, and no one has explored disease-related factors that might
affect IA rates. Thus, we determined the rate of IA in women with SLE and compared
this with general population rates. In addition, in women with SLE, we
investigated potential disease-related predictors of IA.

Methods: We identified women with SLE using Quebec’s universal
healthcare databases (01/1996-12/2011). All
women with SLE, 15-45 years, were identified based on ³1 hospitalization with a SLE diagnosis, or
³2 physicians’ claims for SLE within any 2-to-24-month period. We
determined the number of IA during the interval, as defined by procedure codes
or physicians’ claims for IA. We applied age-specific and relevant calendar-period
IA rates to the observed years of follow-up to determine the expected number of
IA. We then calculated the standardized incidence ratio (SIR) of observed to expected IA.

            To
investigate predictors of IA in SLE women, we conducted a nested-case control
analysis, where SLE women with an IA (i.e. cases) were matched to ³1 control SLE women (i.e. without an IA at
the index date) for age, calendar time, and cohort entry. Within this
nested case-control cohort, we performed a multivariate conditional logistic
regression including teratogenic (i.e. mycophenolate mofetil, methotrexate, leflunomide,
cyclophosphamide) and non-teratogenic (i.e. azathioprine, cyclosporine,
sulfasalazine) immunosuppressive exposure, corticosteroid use, and hospitalization
for SLE in the previous year. Only SLE women covered by the public drug plan
for ³16 weeks prior to the index
date were included in this multivariate analysis.

Results: We observed 293 IA among 2508 women with SLE,
yielding an incidence rate of 17.1 IA per 1000 person-years (95% CI 15.2, 19.2).  Compared to the general population, we
were unable to detect a difference among women with SLE in the number of IA
(SIR 1.10; 95% CI 0.98, 1.24). In the multivariate analysis, including 78 cases
and 1066 corresponding SLE controls, we did not see higher rates of IA among
women exposed to teratogenic immunosuppressives (RR 0.37; 95% CI 0.13,
1.07) and those using corticosteroids
(RR 0.67;
95% CI 0.39, 1.16).  Results were inconclusive for the effect
of non-teratogenic immunosuppressive use (RR 1.12; 95% CI 0.49, 2.57)
and prior hospitalization for SLE (RR 1.32; 95%
CI 0.80,
2.17) on IA.

Conclusion: Our findings suggest that women with SLE have a similar
rate of IA compared to the general population.  Although teratogenic drug exposure was not
necessarily associated with a higher IA rate, the numbers do indicate that some
unplanned pregnancies occur in women exposed to teratogenic immunosuppressives.
Our results should prompt further research on contraceptive counseling and practices
in women with SLE.