Table 1

Association between seroconversion of anti-human papillomavirus (HPV)-16 IgG and viral and demographic factors: univariate Cox proportional hazards regression analysis

No. seroconverted/ no. person-yearsRelative risk (95% confidence interval) for seroconversionP
All subjectsa
 Age (yrs)
  <2017/516.51
  ≥208/3000.79 (0.34–1.83)0.578
 Race
  White, Asian, and others18/620.51
  Hispanic2/105.40.64 (0.15–2.74)0.543
  African American5/90.61.92 (0.71–5.17)0.199
 HPV types detected ≤ time tb
  HPV negative3/493.41
  Other types2/221.51.57 (0.26–9.40)0.622
  HPV-16-related: 31, 33, 35, 52, or 583/60.38.88 (1.79–44.13)0.008
  HPV-1617/39.473.84 (21.57–252.77)<0.001
 HPV types detected ≤ time tb
  HPV negative or other types5/714.91
  HPV-16 or HPV-16-related (31, 33, 35, 52, or 58)
   Viral load
    Low4/40.614.44 (3.87–53.84)<0.001
    High16/59.141.47 (15.15–113.53)<0.001
   Type-specific persistence for ≥2 visits
    No12/54.632.43 (11.40–92.24)<0.001
    Yes8/44.028.13 (9.17–86.27)<0.001
 Anti-HPV-16 IgA at time tb
  No20/802.71
  Yes5/13.815.14 (5.62–40.83)<0.001
 Anti-HPV-16 IgA absorbance level at time tb,c
  No IgA20/802.71<0.001d
  <0.23192/9.58.99 (2.09–38.71)
  ≥0.23193/4.328.33 (8.21–97.75)
Subjects with HPV-16 incident infectione
 Number of visits in which HPV-16 DNA was detected ≤ time tb
  17/5.91
  ≥26/7.61.31 (0.4–4.13)0.641
 HPV-16 viral load when the incident infection was first detected
  Low load2/7.31
  High load11/6.25.67 (1.23–26.18)0.026
  • a The entry time was at baseline.

  • b The independent variable was analyzed as a time-varying covariate. The sample size presented has the number of seroconverters in the numerator. The person-years contributed by each individual at risk were partitioned by the exposure categories. The denominator for a particular exposure category is the sum of person-years from individuals who were at some point in time being exposed to that category. Because the variable was time dependent, an individual could contribute person-years in more than one exposure category. It is erroneous to directly calculate and compare relative risks based on these person-years. These numbers are shown merely to reflect the sample size.

  • c Median absorbance among samples positive for IgA to HPV-16 virus-like particles.

  • d P for linear trend.

  • e The entry time was at the follow-up visit when HPV-16 DNA was first detected.