A second generation cervico-vaginal lavage device shows similar performance as its preceding version with respect to DNA yield and HPV DNA results

BACKGROUND: Attendance rates of cervical screening programs can be increased by offering HPV self-sampling to non-attendees. Acceptability, DNA yield, lavage volumes and choice of hrHPV test can influence effectiveness of the self-sampling procedures and could therefore play a role in recruiting non-attendees. To increase user-friendliness, a frequently used lavage sampler was modified. In this study, we compared this second generation lavage device with the first generation device within similar birth cohorts. METHODS: Within a large self-sampling cohort-study among non-responders of the Dutch cervical screening program, a subset of 2,644 women received a second generation self-sampling lavage device, while 11,977 women, matched for age and ZIP-code, received the first generation model. The second generation device was different in shape, color, lavage volume, and packaging, in comparison to its first generation model. The Cochran’s test was used to compare both devices for hrHPV positivity rate and response rate. To correct for possible heterogeneity between age and ZIP codes in both groups the Breslow-Day test of homogeneity was used. A T-test was utilized to compare DNA yields of the obtained material in both groups. RESULTS: Median DNA yields were 90.4 μg/ml (95% CI 83.2-97.5) and 91.1 μg/ml (95% CI 77.8-104.4, p= 0.726) and hrHPV positivity rates were 8.2% and 6.9% (p= 0.419) per sample self-collected by the second - and the first generation of the device (p= 0.726), respectively. In addition, response rates were comparable for the two models (35.4% versus 34.4%, p= 0.654). CONCLUSIONS: Replacing the first generation self-sampling device by an ergonomically improved, second generation device resulted in equal DNA yields, comparable hrHPV positivity rates and similar response rates. Therefore, it can be concluded that the clinical performance of the first and second generation models are similar. Moreover, participation of non-attendees in cervical cancer screening is probably not predominantly determined by the type of self-collection device.