Interpregnancy Interval and Adverse Pregnancy Outcomes: An Analysis of Successive Pregnancies
Firstly, as mentioned in the discussion, intervening pregnancies ending before 20 weeks of gestation were not included. This leads to a systematic overestimation of interpregnancy interval for any woman who had a loss before 20 weeks. It could be expected that 10% or more of the women would have had an intervening loss.2 This higher risk group of pregnancies (pregnancy after a loss) thus can have had unrecognized short interpregnancy intervals, given the tendency for women to replace a perinatal loss with a new pregnancy (selective fertility).3 Because the births included in the analysis had to be 20 weeks or longer, women who had unrecognized losses were virtually excluded from the shortest interpregnancy interval category of 0–5 months. The matched analysis directly compares any pregnancy with an unrecognized intervening loss with either a preceding or subsequent birth at 20 or more weeks of gestation to the same mother. Our own work, which did take into account losses at less than 20 weeks of gestation, found that an interpregnancy interval of less than 3 months after a second-trimester loss was significantly associated with a recurrent loss.4
Secondly, to address generalizability, the parity sequence of the study births needs to be clarified. Readers will assume that the three successive births for each woman were in a parity sequence of 0, 1, and 2. The terminology of first, second, and third birth that is used in some places in the article is ambiguous. It could mean the first available birth record during the study period regardless of parity. The methods state that the analyses included all women with at least three singleton deliveries, with no initial parity stipulation. The recorded parity of each pregnancy obviously would provide a useful check that pregnancies were consecutive.