Figure 3.

Subgroup analyses of association between human papillomavirus (HPV) infection and preterm birth. *, Subgroup odds ratios (ORs) are size weighted and pooled individual study ORs, using random-effects model including studies that adjusted for at least the subgroup characteristic. Not adjusted subgroup ORs are size weighted and pooled individual study ORs using random-effects model including studies that did not adjusted for the subgroup characteristic. ¹, Only for prospective cohorts and prospective case-controls studies. ², Human papillomavirus exposure may have been measured during pregnancy in some studies, but details were not provided. ³, Refer to Supplementary Table 1 for more details on laboratory methods used in each study. ⁴, Refer to Supplementary Table 2 for more details on adjustment methods used and variables adjusted for in each study. I², percentage of the variability in ORs that is due to true heterogeneity across subgroup’s studies. CI, confidence interval; NE, not estimable.