Abstract
In 2023, New Zealand (NZ) reformed its cervical cancer screening protocol to use HPV testing as the primary screening test, prioritising colposcopy referral for instances of specific high-risk HPV type detection (HPV16 and HPV18). However, a lack of HPV prevalence studies conducted in NZ means this protocol is based on research conducted on populations of unproven generalisability. Type-specific prevalence differs geographically and by differing availability of vaccination and cervical screening programmes. The NZ cervical screening programme may benefit from re-evaluation of current protocols if more recent, local data were available.
The primary aim of this study was to describe type-specific HPV prevalence in NZ people with cervices aged 25–69, with results for HPV16, 18, 31, 45, 51, 52, and pooled results of HPV33/58, 56/59/66 and 35/39/68. Secondary aims included:
• estimating odds of high-grade histology diagnosis per type detected.
• investigating the value of sample viral quantity (estimated using HPV test critical threshold (CT) values) as indicators of cervical lesion risk.
• investigating agreement between HPV results obtained from simultaneously collected vaginal and cervical samples.
In order to provide useful comparators to NZ, a meta-analysis of recently (since 2013) published studies reporting type-specific HPV prevalence in countries with nationwide HPV vaccination and active invitation to nationwide cervical screening was conducted. High heterogeneity was observed amongst the results of the 42 papers fulfilling these criteria, predominantly attributable to differences in methodology. This illustrates the need for more standardised reporting. There is also a relative lack of type-specific prevalence reporting in many countries.
The aim of describing type-specific HPV prevalence in NZ was achieved through quantitative statistical analysis of three datasets: one pilot program trialling HPV tests as a primary screening method; one larger dataset of routine screening data collected following the nationwide implementation of the new screening protocol in 2023, and one dataset including HPV tests performed according to pre-2023 screening guidelines. Compared with the results of the meta-analysis all three cohorts demonstrated significant differences, in particular lower prevalence of HPV16 and higher prevalence of HPV52.
This NZ data was paired with histology results (where available) to estimate the HPV type-specific odds of a high-grade cervical lesion detection by fitting a logistic regression model. This demonstrated odds of a high-grade histology diagnosis to be greatest for HPV16, 18 and 31, although these findings were almost certainly affected by ascertainment bias due to the prioritisation of colposcopy for HPV16 and HPV18.
Further logistic regression models were constructed to investigate the clinical significance of viral quantity in screening samples. These modelled type-specific CT values as predictors of high-grade histology results, and demonstrated likely correlation for HPV16 and 31, with inconclusive results for HPV18.
Investigation of HPV test agreement between simultaneously collected vaginal swab samples and clinician-collected cervical samples was also conducted using Gwet’s AC1. This demonstrated a high degree of agreement between the two, with deviations mostly being the result of positive results from vaginal samples without corresponding positive results from the cervical samples.
These results suggest amending current screening protocols in NZ to include HPV31 detection to indicate immediate colposcopy referral is advisable (as with HPV16 and 18). The evidence of risk associated with the viral quantity of HPV16 and 31 is promising, but should be investigated further before any potential role in screening triage can be recommended.