To the Editor—We thank Bénet et al.Reference Bénet, Ecochard and Vanhems ¹ for their letter discussing the difficulties in evaluating the infectious risk linked to performing low-level disinfection (LLD) of endovaginal and transrectal ultrasonography (US) probes used with disposable probe covers. As indicated in our article, we used a modeling approach to approximate this infectious risk for some pathogens, and agreed that our modeling had limitations, closely related to the assumptions we used as inputs.Reference Leroy, M’Zali, Kann, Weber and Smith ²

However, we do not completely agree with all the concerns raised by Bénet et al.Reference Bénet, Ecochard and Vanhems ¹ The probability (Pr2b, in our article) that the probe might remained contaminated after cleaning and LLD was estimated from a meta-analysis of 2 cohort studies that quantified the efficacy of LLD on bacterial agents: the studies of Buffet-Bataillon et al.Reference Buffet-Bataillon, Vallee, Lebrun, Cormier, Poulian and Jolivet-Gougeon ³ and Kac et al. (2007),Reference Kac, Gueneret and Rodi ⁴ and not the Kac et al. article published in 2010.Reference Kac, Podglajen, Si-Mohamed, Rodi, Grataloup and Meyer ⁵ We agreed that assuming LLD efficacy is similar for bacterial and viral agents is arguable, as discussed in our paper, but we made the assumption because of the lack of data available on viruses.Reference Thorpe, Ouellet and Hershow ⁶ Moreover, if human immunodeficiency virus (HIV) is a fragile virus as discussed in our paper, other non-enveloped viruses such as human papillomavirus are relatively resistant to commonly used clinical disinfectants.Reference Meyers, Ryndock, Conway, Meyers and Robison ⁷ We also agreed that the probability of probe transmission from an HIV-infected patient varies with viral load and the mucosal breach, and choosing sexual transmission as a proxy for both probabilities of transmission from an infected patient to the probe as well as of transmission from the infected probe to the next patient could be arguable. However, given the lack of specific data, it appears to us to be the more realistic assumption to make, better than data from drug users that supposes percutaneous injections, or accidental contaminated blood exposures. Sensitivity analysis would have provided an estimation of the impact of the parameters’ variability on the modeling results. This is problematic as well, since we would have had to make compounding assumptions on these parameters’ variability, especially to numerically approximate how far our proxy was from the reality of probe-to-patient and patient-to-probe transmission. Interestingly, our modeling was checked by an “empirical modeling” performed for human papillomavirus, in which Casalegno et al.Reference Casalegno, Le Bail Carval and Eibach ⁸ empirically estimated Pr1a to Pr2b. The results of the empirical and baseline modeling were similar with a slight underestimation from the baseline modelling, indicating that our assumptions for Pr1a to Pr2b were likely realistic. Of note, pooled prevalences were calculated by using a random effects model with inverse variance weighting using the Der Simonian and Laird method, referencing to our previous work in which pooled estimates were clearly calculated taking into account both sample size and data dispersion.Reference Leroy ⁹

Beyond these technical points, the question raised is how can we estimate the infectious risk related to LLD performed on endovaginal and transrectal US probes used with disposable covers, and how can we provide quantitative estimates to guide public health decision-making using relevant and robust scientific evidence. One approach is modeling given the lack of comprehensive data, with the limitations due to the model assumptions. Bénet et al.Reference Bénet, Ritter and Vanhems ¹⁰ proposed another approach based on a secondary analysis of a large hospital-based cohort study and compared their results with ours to conclude on the safety of endocavitary US regarding HIV and hepatitis C virus. However, their results appeared to us to have some limitations as well. Patients were initially part of a cohort study designed for another purpose and selected for that secondary analysis on having 2 HIV and hepatitis C virus serologies, but no complementary information on the other transmission risk factors was collected even though it might be of interest for adjustment in the analysis. Then patients who underwent endocavitary US were identified by searching in the French procedures classification (Classification Commune des Actes Médicaux). Therefore, since patients were selected on the basis of 2 repeated serologies, they could present with a particular status regarding the risk of viral transmission that could limit the external validity of their results beyond this particular cohort study.

One can ask whether for some pathogens, such as HIV, hepatitis B virus, and hepatitis C virus, that lead to rare or very rare transmission but with dramatic consequences, a simple case report but with a very well-documented causality relationship would be enough to alert public health services, as it happened in the United Kingdom. ¹¹ Another observation reported hepatitis C virus transmission to patients in the in vivo fecundation process, likely through healthcare workers during routine procedures such as daily US.Reference Lesourd, Izopet and Mervan ¹² For more common viruses such as human papillomavirus related to an increased risk of cervical cancer, quantitative studies may be valuable. Lastly, we agreed with the conclusions of Bénet et al.Reference Bénet, Ecochard and Vanhems ¹ that appropriate investigations are warranted in order to guarantee patients’ safety regarding these semicritical devices. We find ourselves wondering why LLD is still recommended for routine vaginal/rectal US in France whereas high-level disinfection is encouraged in its European neighbors, Australia, North America, and Japan. If waiting for more convincing evidence is required by public health services, a precautionary principle can be applied as an easy and efficient start point and patients may be asked for their serologic HIV/hepatitis C virus/hepatitis B virus status before endovaginal/transrectal US. In case of positivity, disinfection procedures can be switched to high-level disinfection.