Perinatal & Pediatric
The Role of Gravity and Informative Censoring in Explaining the Association Between Pregnancy Loss and Subsequent Hypertension Sabrina Chiodo* Sabrina Chiodo Chiodo Chiodo Division of Epidemiology, Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada
Background: Pregnancy loss is linked to hypertension risk, but prior studies have not accounted for the potential influence of reproductive history or informative censoring. We examined how accounting for gravidity and informative censoring impacts the association between pregnancy loss and the five-year incident hypertension risk.
Methods: This population-based retrospective cohort study used administrative health data from British Columbia, Canada (1994–2016). Participants were nulliparous women (15–49 years) with a recorded pregnancy and no prior hypertension. Pregnancy loss (miscarriage/stillbirth) was identified via ICD codes. We used inverse probability of treatment and censoring weights to account for confounding and informative censoring. Time to incident hypertension was estimated using Weibull accelerated failure time models under three analytic strategies: (1) gravidity excluded from weighting/outcome models; (2) gravidity in the treatment weighting model only; and (3) gravidity in both the weighting and outcome models.
Results: Among 309,608 pregnancies (27,044 losses), losses were linked to older age, higher BMI, and higher gravidity. Without accounting for gravidity, pregnancy loss was associated with a 15% shorter time to hypertension (Time Ratio [TR] 0.85, 95% CI 0.80, 0.90). When gravidity was included in the treatment weighting model only, pregnancy loss remained associated with earlier hypertension onset (TR 0.91, 95% CI 0.86, 0.97), though this association was attenuated after additionally accounting for informative censoring (TR 0.96, 95% CI 0.90, 1.03). When gravidity was included in both the weighting and outcome models, pregnancy loss was no longer associated with hypertension (TR 0.95, 95% CI 0.90, 1.02).
Conclusion: The observed association between pregnancy loss and earlier hypertension may largely be driven by reproductive history and informative censoring. Future studies must explicitly account for these factors to avoid biased inference.