Impact of loop electrosurgical excision procedure for cervical intraepithelial neoplasia on HIV-1 genital shedding: a prospective cohort study: population and statistical queries



Our departmental journal club enjoyed discussing the recent paper entitled ‘Impact of loop electrosurgical excision procedure for cervical intraepithelial neoplasia on HIV-1 genital shedding: a prospective cohort study’.[1] The study has highlighted that loop electrosurgical excision procedure (LEEP) in a cohort of HIV-positive women at high risk of cervical cancer appears to increase HIV-1 viral shedding 2 weeks following the procedure, regardless of highly active antiretroviral therapy (HAART) status. We would like to raise some concerns regarding the study population and statistical analysis performed.

Although the eligibility criteria for this study were clearly defined, the generalisability of the study cannot be fully appreciated. We wonder whether women attending cervical screening centres in Kenya receive ‘usual care’ or comprise a unique group, for example of high social class, poor general health status, or screened as part of HIV care, which may affect their response to LEEP? The success of the recruitment process was not described, leaving it ambiguous as to what extent participants differed from those who chose not to take part.

As this was an uncontrolled cohort study measuring levels of HIV-1 viral shedding over a period of 14 weeks within the same individuals, the use of a chi-square test to compare proportions of viral detection at each time point is inappropriate, given that the samples were not independent of each other.

In relation to the power calculation, it was not clear what level of power, accepted risk of type-II error, estimated effect size, and estimated variance the authors incorporated. In addition to the acknowledged possibility of a type-II error in the cohort not on HAART, we question the validity of results from a mixed-effects model fitted on such a small number of participants (n = 10).

Huchko et al. compared baseline characteristics between those on HAART and those not on HAART. This seems somewhat peculiar when the study did not compare women on HAART with those not on HAART. The methods and main results suggest that the two groups were treated as entirely separate cohorts.

Huchko et al. explained that they did not adjust for baseline clinical or demographic characteristics within their mixed-effects model because of the fact that measures were taken from within individuals, but we wonder whether possible confounding factors such as WHO stage of HIV, time since HIV diagnosis, and time since HAART may partially explain the relationship between LEEP and viral shedding?

For the HAART cohort, a significant increase of approximately four-fold in concentration of cervical viral load at 2 weeks compared with baseline was described, with this mean concentration returning to baseline at 4 weeks; however, it slowly rises again to approximately 2.5 times the baseline concentration at 14 weeks, but this is not statistically significant. Could the non-significance of this effect be due to the small sample size, and if so, what is the clinical significance of this?

Finally, we thank Huchko et al. for an interesting article confirming that current guidance regarding the suggested period of abstinence from sexual intercourse appears adequate in relation to the theoretical risk of increased HIV infectivity post-procedure.