July 2019: HIV and STDs

Genotype-guided drug reduction is safe for treatment-experienced patients who have achieved viral suppression on complex antiretroviral regimens.  For what it’s worth, the roadblock I ran into most often when trying to reduce these patients’ pill burdens was that they didn’t want me to. After long struggles with adherence and drug resistance that I, an ID fellow breezing in and out of their lives, hadn’t been party to, most of the patients I saw in this situation were inclined to stay on what was working for them, thank you very much. And hey, more power to them!  But for those patients on salvage regimens who’d rather not be taking eight pills a day, is it safe to “de-escalate” ART once they’ve achieved viral suppression?

The ECOVIR study enrolled patients who had been virally suppressed for at least 48 weeks on a regimen involving 4+ antiretrovirals.  The investigators used genotype data to identify the least active drugs in each patient’s regimen, then stopped those drugs to reach a regimen with three or fewer components.  Specifically, they used the French ANRS resistance algorithm, and sought to reach a genotypic susceptibility score (GSS) of >2, indicated either two fully active drugs or one fully active and two partially active drugs.  After the switch (if approved by the patient and their HIV clinician), the researchers followed the patient’s viral load monthly for the first six months, then every 3 months thereafter. The primary outcome was maintenance of viral suppression through week 24.

Eighty-nine patients participated in the study.  These folks had been on ART a median 24 years, had mean CD4 counts of 538 (current) and 75 (nadir), and on average had been virally suppressed for 8 years.  Most patients were taking 4 (84%) or 5 (15%) antiretrovirals; one patient was taking six.  The resistance algorithm suggested a reduced regimen with a GSS >2 for 97% of patients. Of these, 15 (17%) did not switch to the reduced regimen; the most common reason for not switching was the physician’s preference (n=9), followed by the patient’s preference (n=2), viral rebound (n=2), and death or loss to followup (n=1).  The reduced regimens mostly included discontinuation of NRTIs (from 93% to 21%), followed by stopping PIs (75% to 52%) and stopping maraviroc (32% to 17%).  Three quarters of patients ended up on a two-drug regimen, most often raltegravir plus etravirine, and at week 24 67/71 (94%) had maintained viral suppression. Of the four patients who had viral rebound, one occurred at week 4, two at week 12, and one at week 24, and in all cases the discarded drug had been considered inactive (specifically: etravirine with multiple NNRTI mutations and emtricitabine with M184V).  No additional resistance mutations were observed in the patients who had viral rebound; after switching back to their original regimens, 3 re-achieved viral suppression and one had persistent low-level viremia.  In extended followup to week 48, there was one additional viral rebound at week 48, producing an overall success rate of 92%.

So, genotype-guided reduction of ART regimens to include two active drugs (or one active and two partially active drugs) is highly successful in reducing the number of antiretrovirals prescribed in patients with extensive treatment histories who have achieved viral suppression.  I’m not too surprised to see some viral rebounds after withdrawing lamivudine/emtricitabine in the setting of M184V, as it’s still doing the work of at least half a drug just by maintaining selection for that fitness-reducing mutation; I was more surprised by etravirine, and wonder if the ARNS algorithm matches up with the traditional Tibotec scoring system used to evaluate etravirine’s activity in the face of multiple mutations.  Anyway, I think the take-home is that two fully active drugs (or the cobbled-together equivalent of partially active drugs) is probably adequate for maintenance of viral suppression even in patients with complex genotypes.  Though I wonder if medications with high barriers to resistance and convenient dosing might be more successful in dual regimens (e.g. dolutegravir and doravirine vs raltegravir and etravirine?) 31273376

 Specific vaginal flora are associated with increased risk of T.vaginalis acquisition in women.  Previous research has shown that normal flora in the GI tract play an important role in excluding pathogens by taking up space and nutrients and also by providing antigens that stimulate mucosal immune responses.  These data suggest that similar associations may be present in the female GU tract.

The authors used clinical samples from the Mombasa Cohort, a prospective cohort study of female sex workers that involved monthly screenings for HIV and other STI screening as well as GU microbiome sequencing.  They compared the pre-infection microbiota of women who developed trichomoniasis (TV) to the vaginal microbiota of women who were not infected.  They categorized patients based on menstrual status (i.e. having regular menses or either amennorheic/pregnant/menopausal) and whether or not they had bacterial vaginosis by Nugent scoring to be able to adjust for these important potential confounders.  Samples from visits at which the patients had recently taken antibiotics or were menstruating (which is associated with a change in the nature & diversity of GU flora) were not included.  The methods section is quite extensive (this is a JID paper) should you want to know more.

Anyway, eighteen participants with a total 25 episodes of TV were matched to 50 controls who did not develop TV during a similar period of exposure.  The mean age was 36 years; women who acquired TV were less likely to be on hormonal contraception (12 vs 34%) or to be amenorrheic (20% vs 44%); otherwise the groups were similar; notably, rates of BV were similar between groups (40% vs 34%), as were rates of self-reported unprotected sex (24% vs 16%).  Of the clinical variables, only amenorrhea (RR 0.39; 95% CI 0.12-1.23) was associated with TV acquisition.

Total vaginal bacterial concentrations at the pre-TV visit were similar between groups, which remained true after adjusting for menstrual status.  The researchers observed increased species richness in the patients who acquired TV (Chao-1 index of 80.12 vs 69.85 with RR 1.02 and p=0.02; I don’t have the stats background to fully explain what that means, but my understanding is that Chao-1 index is an estimate for the abdundance/diversity of organisms present in small numbers).  More critically, increasing abundance of Bulleidia sp. were associated with TV acquisition, (OE 2.18; 95% 1.0-4.7) and several other species had associations with similar effect sizes but wider confidence intervals, including Veillonella, Prevotella, Anaerococcus, and Bacteroidales (notably, all anaerobes). Prevotell amnii was also detected with greater frequency in the vaginal secretions of women who went on to develop TV (64% vs 34%; RR 2.21 with 95% CI 1.12-4.38), as was Snethia sanguinegens (80% vs50%; RR 2.58 with 95% CI 1-6.62). 

Interestingly, no associations were found between Lactobacillus and TV; this may be because the Mombasa cohort is enrolled out of Kenya, and Lactobacillus-dominated vaginal microbiota appear to be more common in women of European and Asian vs African descent (see: 28700747 and 28234976), meaning there may be ethnic or geographic differences in the normal host GU flora and their interactions with pathogens.  In any case, this paper suggests that the vaginal microbiota may mediate susceptibility to TV infection. I wonder whether microbiota transplants or intravaginal probiotics might have some role for patients with recurrent TV?  Lactobacillus-based intravaginal probiotics have already been tried for bacterial vaginosis, with limited evidence suggesting modest benefit (31299136); as best as I can tell, “VMT” has been suggested but not performed. 31287879

Being on “high CNS penetration” antiretrovirals doesn’t seem to prevent HIV-associated neurocognitive impairment.  How depressing.  This conclusion comes from the Neurocognitive Assessment in the Metabolic and Aging Cohort (NAMACO), which enrolled patients >45 years old over a three-year period who were on antiretrovirals and had achieved viral suppression.  The researchers subjected participants to neurocognitive assessments, calculated the CNS penetration effectiveness (CPE) scores of the patients’ antiretroviral regimens, and evaluated CPE scores as predictors of neurocognitive impairment.  The authors performed two analyses: one considering the CPE scores of the regimens outright, and another considering “cumulative exposure” to CPE (i.e. with patients accumulating CPE score based on their regimen per day on that regimen). They also considered a thorough list of potential confounders, including age, sex, education level, % of time spent with viral suppression, current and nadir CD4 counts, etcetera. Regarding the CPE score: this is a previously established scoring system that rates individual drugs from 0 (no CNS penetration, e.g. cobicistat) to 4 (highest possible CNS penetration, e.g. dolutegravir). The total CPE score of the regimen is the sum of the CPE score of each drug, and the score is stratified into low (<5), medium (6-8), and high (>9)

The full study population included 909 patients; most were white (92%) men (80%), the median age was 53, and they had been on ART for a median 12 years. Forty percent had some degree of neurocognitive impairment, and 27% were diagnosed with an HIV-associated neurocognitive impairment (i.e. no other clear diagnosis was present).  The patients’ mean daily CPE score was 6.7 (IQR 5.7-7.4). In both univariate and multivariate analyses, higher CPE score (by either mode of analysis) was not associated with the presence of neurocognitive impairment or HIV-associated neurocognitive impairment.  This remained true even when they changed CPE to a categorical variable (i.e. low/medium/high CPE) and only compared the low vs high categories.

The NAMACO study is ongoing, with planned followup at 2 and 4 years after enrollment, so perhaps we’ll discover that high-CPE regimens can mitigate the progression of HIV-associated neurocognitive impairment once it occurs.  But I’m not getting my hopes up. 31304188


Does your patient on efavirenz have neuropsychological symptoms?  Unfortunately, the data says that switching off efavirenz won’t help.  This study out of UPMC examined the MACS cohort, which enrolls MSM with or at high risk of acquiring HIV and follows them every six months to assess for depression severity and neuropsychological symptoms using a number of validated tools.  They classified the patients’ neurocognitive impairment as asymptomatic, minor, or HAND (HIV-associated dementia). For this study, the authors selected participants with HIV who either were never started on efavirenz (NoEFV), were on efavirenz and did not switch off (NoSwitch), or who were on efavirenz and then switched to something else (SwitchOff), and compared their neurocognitive outcomes. 

The analysis included 1989 participants (NoEFV group, n=1675; NoSwitch, n=44; SwitchOff, n=270).  The median age was 37, participants had a median 3 years of follow-up, and the median CD4 count was 442 cells/ul.  Of the patients on combination ART, 69% had achieved viral suppression.  Baseline rates of hand in the NoEFV, NoSwitch, and SwitchOff groups were 3.2%, 4.1%, and 2.8%. 

First, the authors considered just the SwitchOff group, looking at their neuropsychological testing 2 years prior and 4 years after the switch off of efavirenz.  Scores within each neuropsychological domain tested were no different for these patients after the switch.  This remained true even after adjusting for depression scores, and also when restricted to the patients who had achieved viral suppression before the switch off of efavirenz.  Next, they compared depression scores and neuropsychological impairments between the three groups.  Again, they found no differences in testing overall or in the scores of the individual neuropsychological domains tested.

So, switching of efavirenz doesn’t seem to do anything for patients’ depression or neurocognitive impairments.  These data concur with earlier studies that are cited in the discussion, most notably the recent data from Uganda indicating that efavirenz was no more associated with depression or suicidality than nevarapine, bringing the purported psychiatric side effects of efavirenz into question. 30932965