'The obesity paradox': a reconsideration of obesity and the risk of preterm birth.

OBJECTIVE: The association between obesity and spontaneous preterm births (sPTBs) has been shown to be influenced by obesity-attendant comorbidities. Our objective was to better understand the complex relationship of obesity and its attendant comorbidities with sPTBs.

STUDY DESIGN: A retrospective analysis utilizing maternally linked hospital and birth certificate records of 2 049 196 singleton California deliveries from 2007 to 2011. Adjusted relative risks (aRRs) for sPTBs were estimated using multivariate Poisson regression modeling.

RESULTS: Obese women had higher aRRs for sPTBs than their normal body mass index (BMI) controls. aRRs (95% confidence interval) increased with increasing BMI category: Obese I=1.10 (1.08 to 1.12); Obese II=1.15 (1.12 to 1.18); and Obese III=1.26 (1.22 to 1.30). When comparing only obese women without comorbidities to their normal BMI controls, aRRs reversed, that is, obese women had lower aRRs of sPTBs: Obese I=0.96 (0.94 to 0.98), Obese II=0.95 (0.91 to 0.98); and Obese III=0.98 (0.94 to 1.03). This same reversal of aRR direction was also observed among women with comorbidities: 0.92 (0.89 to 0.96); 0.89 (0.85 to 0.93); and 0.89 (0.85 to 0.93), respectively. Increasing BMI increased the aRRs for sPTBs among patients with gestational diabetes (P<0.05), while decreasing the risk among patients with chronic hypertension and pregnancy-related hypertensive disease (P<0.05).

CONCLUSIONS: The obesity and preterm birth paradox is an example of what has been described as 'Simpson's Paradox'. Unmeasured confounding factors mediated by comorbidities may explain the observed protective effect of obesity upon conditioning on the presence or absence of comorbidities and thus resolve the paradox.