Fig. 3.

Trends in estimates of association between baseline HPV serostatus and type-specific HPV incident detection by HPV antibody concentration level, stratified by sex (females, males, or both). Seven publications measured antibody levels in terms of ELISA units per mL (EU/mL) [14, 19, 21, 22, 24, 25, 41], and one, in Luminex units per mL (LU/mL) [15]. Five studies reported estimates by tertiles of antibody levels [14, 15, 22, 24, 41], two by quartiles [21, 24], and one dichotomized antibody levels (low versus high concentration) [19]. T1, T2, T3 refer to the lowest, middle, and upper tertiles of antibody levels used by each publication, respectively; Q1, Q2, Q3, Q4 refer to the first, second, third, and fourth quartiles, respectively; Low and High refer to low and high antibody concentrations, respectively. Horizontal dashed lines indicate the null value of the estimate of the association (1 on a relative scale). The y-axis indicates the risk of incident HPV detection on the relative scale. For instance, a blue dot with the highest y-value in the top right panel indicates a risk ratio of incident detection between individuals with no HPV antibodies versus individuals with naturally acquired HPV antibodies in the lowest tertile. The lower end of antibody level for the lowest tercile/quartile used by publications was between 7–8 EU/mL or 0.2–0.28 LU/mL, and the higher end of between 40–64 EU/mL or 0.46–0.55 LU/mL