Response to: Fitzpatrick C, Lowe M, Richardson D. Sexually transmitted infection testing and rates in men who have sex with men (MSM) using HIV pre‐exposure prophylaxis. HIV Medicine 2019; https://doi.org/10.1111/hiv.12736 (published Online First: 16 March 2019)

Dear Professors Gazzard and Lundgren, We were interested to read the Short Communication exploring attendances and sexually transmitted infection (STI) diagnoses among HIV pre-exposure prophylaxis (PrEP)-using and -nonusing men who have sex with men (MSM) [1]. While we acknowledge the importance of and huge interest in understanding STI outcomes among PrEP users, analyses of this nature are extremely challenging and complex. We would like to highlight some of these complexities and suggest that the data reported by Fitzpatrick et al. [1] do not necessarily support the authors’ claim that ‘MSM using PrEP. . . had significantly higher rates of chlamydia, gonorrhoea (particularly rectal chlamydia and gonorrhoea) and infectious syphilis, in keeping with other published data’. To assess differences in STI risk according to PrEP use status in an unbiased manner, routinely collected and established sociodemographic and behavioural indicators of confounding should have been taken into account, such as age, ethnicity, number of previous clinic attendances, number of prior HIV tests, or history of bacterial STIs. The use of non-PrEP users as a comparator group also introduces bias as PrEP users will be, by being eligible for PrEP, more likely to have condomless sex and/or previous STIs. Hence, PrEP users are a ‘higher risk’ group for HIV and STI acquisition and any unadjusted comparison to non-PrEP users would be expected to show a positive association between STI acquisition and PrEP use. Not accounting for STI testing frequency further overestimates any association. The data presented do not provide a sufficient picture of the incidence of bacterial STIs, which may initially increase, as a consequence of the effect of increased testing in PrEP users, then stabilize or decrease over time. While the data presented may indicate that the prevalence of STIs is higher in MSM using PrEP, as time has not been included in the analysis, it would be incorrect to suggest an increased rate or incidence of infection. The meta-analysis by Traeger and colleagues found some association between any STI diagnosis and PrEP use [pooled odds ratio (OR) 1.24; 95% confidence interval 0.99–1.54; P = 0.059]; however, this was only statistically significant for any rectal STI and rectal chlamydia diagnoses. Increases in rates of syphilis, chlamydia or gonorrhoea at any anatomical site were statistically nonsignificant [2]. Experience from the Scottish PrEP programme has found a similar rate of bacterial STI in MSM using PrEP compared to those never prescribed PrEP, with the authors rightly acknowledging that it is too soon to draw conclusions about the impact of PrEP on STI rates [3]. Recent analyses of STI incidence in MSM before and after starting PrEP show a mixed picture. In the Australian Prelude study of PrEP implementation, no significant increases in STI incidence were seen and there was a decline in gonorrhoea incidence after 6 months [4]. A borderline significant increase for any STI and for chlamydia was seen from 1 year pre-enrolment to followup among participants in the Australian PrEPX study [5]. A secondary objective of the English PrEP Impact Trial is to explore STI incidence among participants [6]. The use of the GUMCAD STI surveillance data set will help to Correspondence: Dr John Saunders, HIV & STI Department, Public Health England, 61 Colindale Avenue, NW9 5EQ London, UK. Tel: +44 2083277372; e-mail: john.saunders@phe.gov.uk