HIV and STDs: September 2018

Gonorrhea susceptibility to azithromycin follows a seasonal pattern that tracks with the prevalence of macrolide use in the community. It makes sense that exposing patients to macrolides to treat winter respiratory infection would select for azithromycin-nonsusceptible strains of gonorrhea in those patients, but what an elegant study! Olesen and colleagues created a mathematic model of the effects of fluctuating selection pressure for macrolide resistance in N. gonorrhoeae based on seasonal macrolide use on the MIC distribution among N. gonorrhoeae isolates 3 months later. Then they compared their model’s predictions to observed data using A) data on the seasonality of population-wide macrolide use from an insurance claims database and B) data on the seasonality of N. gonorrhoeae azithromycin MIC distributions from a bank of 62500 isolates collected by the CDC as part of a decade-long surveillance project.

After some mathematical futzing (adjusting for between-clinic variations in MICs, year to year variations in reporting protocols, etcetera), they produced the beautiful Figure 2, which shows that N. gonorrhoeae azithromycin MICs indeed fluctuate seasonally as though under the influence of population-wide macrolide use, with seasonal MICs peaking 3 months after the peak of macrolide use in March/April. They were even able to back calculate the “MIC cost” of excess macrolide prescription. This work, besides just being really cool, demonstrates that antibiotic prescribing, appropriate or otherwise, has off-target costs. 30239814

PrEP clinics have an engagement in care problem. So says data from the Public Health Seattle / King County STD clinic, whose PrEP program targets MSM and transgender women at high risk for acquiring HIV infection. The authors assessed adherence to their PrEP program between 2014 and 2016 using pharmacy and clinic records, including patient’s personal reports of stopping and restarting PrEP between visits. The analysis population included 307 patients followed for a median 12 months. Of these, 133 (43%) disengaged from the PrEP program: 26% discontinued PrEP after starting the medication and 17% never started PrEP. Figure 2 in the manuscript presents a Kaplan-Meier curve showing that among those who started PrEP, the discontinuation rate was greatest in the first 3 months but continued steadily thereafter.

The authors were able to provide patients’ reasons for stopping PrEP in 28 cases. The most frequent reason given (46%) was having entered into a monogamous relationship with a HIV-negative partner; 14% stated they were no longer at high risk for acquiring HIV for other reasons, and the remainder discontinued therapy due to medication side effects. In the discussion section, the authors quote several other studies of PrEP nonadherence/disengagement reporting rates of 20-38% over similar periods of followup, and suggest that their drop-out rate may have been higher because their STD clinic actively promotes PrEP to high-risk patients who come in for other reasons (e.g. the patient population is not entirely self-referrals for PrEP). 29485544

Speaking of PrEP, is it cost-effective for serodiscordant couples seeking to conceive? A Boston University-based research group developed a markov simulation model to answer this question, comparing the cost-effectiveness of administering PrEP versus giving cART for the HIV-infected partner alone versus use of assisted reproduction technologies. For serodiscordant couple in whom the HIV-infected partner is virologically suppressed, condomless intercourse at ovulation with cART alone was the most cost-effective strategy; however, PrEP plus cART became cost-effective when the male partner had HIV-infection with a low probability of virologic suppression and PrEP was available at generic pricing. In the era of U = U and highly potent single tablet antiretroviral regimens, one hopes this scenario will be rare enough that it won’t be worth quibbling over cost when it does come up. 30234602

Bariatric surgery significantly improves comorbidities in morbidly obese people living with HIV infection. Sharma et al performed a retrospective analysis of a national data set (the United States Nationwide Inpatient Sample database) to compare patients with HIV infection and morbid obesity hospitalized between 2004 and 2014 who did or did not undergo bariatric surgery (all determined by the presence of relevant ICD-9 codes). Of the database’s 7803 patients with HIV and morbid obesity, 346 (4%) underwent bariatric surgery. Younger age, male gender, and history of hypertension and hyperlipidemia all predicted referral for bariatric surgery. In multivariate analysis, bariatric surgery was associated with decreased risk of renal failure (IRR 0.05; p = 0.004), respiratory failure (IRR 0.17; p = 0.02), and sepsis (0.23; p = 0.04) during hospitalization, but increased risk of gastrointestinal strictures (IRR 2.5; p = 0.001). Consider, though, that some of the differences in end-organ disease may relate to selection bias in who was cleared for surgery in the first place.

Mortality risk was not significantly decreased in surgery recipients, but since the mortality rate in the nonsurgical patients was only 1.4% and the data set’s duration of follow-up was not given, I’m not sure that tells us anything about whether bariatric surgery confers these patients long-term survival benefit. Since HIV infection significantly increases cardiovascular risk, and bariatric surgery can mitigate obesity-related comorbidities that also increase cardiovascular risk, my guess is that it does. 30157083

Here’s data showing why “having had an undetectable viral load for at least six months” may be an important part of the U = U definition. Undetectable equals untransmissable, a statement endorsed by multiple major health organizations, means that a person with HIV infection who is taking antiretroviral therapy and has had an undetectable viral load for at least six months is at zero risk of transmitting HIV to their sex partners. (What counts as undetectable? Conservatively, <200 copies/ml if you look at the studies on which U=U is based; thanks to Dr. Sax for doing the homework)

So why does the patient have to be undetectable for at least six months – why can’t we confidently state the transmission risk is zero as soon as the patient’s serum viral load is undetectable? Well, infectious virus may be eliminated from different body fluids and body spaces at different rates. Here’s more evidence suggesting this is the case. Lahiri and colleagues collected paired serum samples and rectal tissue biopsies from twelve patients with HIV infection starting dolutegravir-based therapy over a three month period. All patients had undetectable serum viral loads by week six. However, 11 of 12 patients had detectable RNA in rectal tissue at the start of therapy, only 3 of these (27%) attained viral suppression in rectal tissue by three months, and these three patients started out with a 10-fold lower viral titer than their nonsuppressing peers (p = 0.02). No correlation between rectal tissue dolutegravir levels and viral suppression was found. While this work does not prove that these patients could still transmit HIV to a sex partner, it shows that HIV RNA replication may persist elsewhere after HIV RNA has been eliminated from serum, making HIV transmission from a recently undetectable sex partner at least biologically plausible. 30005011