Objectives To characterize post-conization time-varying HPV positivity trajectories in a real-world setting with irregular follow-up, and to assess whether the association between baseline infection multiplicity (single vs. multiple) and postoperative HPV positivity is robust to testing intensity and potential informative visiting. Methods In this retrospective cohort of 872 women who underwent conization and completed ≥1 postoperative HPV test, we modeled HPV positivity (yes/no) over continuous months since conization using generalized additive mixed models with subject-specific random effects. Time-varying between-group differences were evaluated using a group-specific time-varying parameter. To address irregular follow-up and potential informative visiting, we performed prespecified sensitivity analyses by (i) restricting to women with ≥3 postoperative HPV tests and (ii) additionally adjusting for testing intensity (number of HPV tests during follow-up). Models adjusted for age, menopausal status, admission type, health insurance type, margin status, preoperative cytology, gravidity/parity, and postoperative hysterectomy. Results In the overall cohort, postoperative HPV positivity decreased over time in both groups (per 4 months: single infection OR = 0.66, 95% CI 0.63–0.70; multiple infection OR = 0.71, 95% CI 0.65–0.78; both P 0.0001). Multiple infection was associated with a persistently higher time-varying trajectory (OR = 1.22, 95% CI 1.12–1.32; P 0.0001), with differences concentrated in the early-to-mid follow-up period. Findings were consistent in sensitivity analyses restricting to women with ≥3 tests and after adjustment for testing intensity. Conclusions Post-conization HPV testing occurs under irregular, potentially informative follow-up in real-world practice. Accounting for testing intensity, the association between baseline multiple infection and less favorable time-varying postoperative trajectories remained robust, supporting risk-stratified surveillance while highlighting the need to explicitly consider follow-up patterns when estimating postoperative HPV dynamics.
Zhou et al. (Wed,) studied this question.