September 2019: HIV and STDs

What predicts death in patients who develop IRIS after treatment for HIV?  The authors conducted a prospective observational study in Kenya, Thailand, and the US that enrolled adults with HIV who had a CD4 count <100 cells/ul and were starting ART for the first time.  They followed the participants biweekly for the first three months after initiating ART, then out to week 48, and observed the incidences of IRIS and mortality, respectively.  Their criteria for IRIS was increase in CD4 count either >50 cells/ul or by more than 2-fold, OR a >0.5 log reduction in HIV plasma titer, accompanied by a new clinical presentation of infection or inflammation without a more compelling alternate etiology.

The cohort enrolled from 2006 through 2013, and in total the authors enrolled 506 participants.  The median age was 37, 39% were women, the median CD4 count was 29 cells/ul, and the media plasma viral load was 5.3 log10 copies/ml (about 200,000 copies/ml).  Most patients started NNRTI-based ART; integrase inhibitors saw use exclusively in the US, and there were the minority (15%) of regimens.  This is potentially relevant because INSTIs are really potent drugs that suppress viremia as fast or faster than other agents, and it seem plausible that they might be associated with more frequent or more severe IRIS.  Anyway, 97 (19%) of subjects developed IRIS, all within the first 6 months of therapy.  The most common underlying cause was TB (33%), followed by MAC (18%), VZV (16%), and Cryptococcus (8%).  In multivariate regression analysis, the only risk factor associated with IRIS was hemoglobin (HR 0.74; 95% CI 0.64-0.85), though D-dimer approached association (HR 1.8; 95% CI 1.0-3.3).  Thirty-one (6% of) participants died within 6 months of starting ART; of these, only seven (23% of the deaths; 1% of the cohort) died with a diagnosis of IRIS, and nearly all of these participants had multiple AIDS-related comorbidities contributing to their deaths.  In multivariate Cox regression analysis, IRIS was associated with death (HR 3.2; 95% CI 1.1-9.2); other factors associated with death included male gender (HR 0.3; 95% CI 0.1-0.7), BMI (HR 0.8; 95% CI 0.7-0.9), WBC count (HR 1.3; 95% CI 1.1-1.5), and D-dimer (HR 3.5; 95% CI 1.00-11.9).

What’s the take-home?  Well, the authors actually performed a decision-tree analysis on their data  (figures 4 and 5, worth checking out) that shows how hemoglobin neatly stratifies risk of IRIS (from 52% in those with a Hgb <8.5 to 10% in those with a Hgb >11.3) and how the combination of CRP, BMI, and D-dimer stratify risk of death after initiating ART (from 43% in those with a CRP >106 ug/ml to 1% in those with a CRP <106 ug/ml, D-dimer <3.89 ug/ml, and a BMI >15.6).  So, perhaps a markedly low hemoglobin and/or BMI and markedly elevated CRP and/or D-dimer should prompt a search for treatable opportunistic infections (and particularly, TB) prior to initiation of ART in patients a CD4 count less than 100 cells/ul.  31504347 

Speaking of folks with low CD4 counts and high viral loads, what’s the optimal initial ART regimen in this population?  A group out of Italy retrospectively examined patients in the ICONA cohort who started ART with a baseline CD4 count <200 and a baseline HIV plasma viral load >5 log10 copies/ml.  They stratified this group based on the type of ART regimen initiated (NNRTI-based, boosted PI-based, or INSTI-based) and examined the composite primary endpoint of treatment failure - defined as two consecutive HIV viral loads >50 copies/ml after 6 months of treatment - discontinuation of the anchor drug class (i.e. switch from NNRTI to PI-based therapy; within-class switches did not count), or death.

A total 1195 patients met the study inclusion criteria (696 starting PI-based therapy, 315 starting INSTI-based therapy, and 184 starting NNRTI-based therapy).  Over a mean 2.3 years of followup, 642 patients experienced a treatment failure; compared with NNRTI-based regimens, treatment failure was more likely with PIs (aIRR 1.6 with 95% CI 1.3-2) and less likely with INSTIs (aIRR 0.7 with 95% CI 0.5-1.0).  I think there’s some bias here because we preferentially give PIs to people who are deemed at higher risk of treatment failure (owing to either known medication nonadherence or HIV resistance mutations, which are often associated with prior medication nonadherence); that said, these data suggest we ought to prefer INSTIs and/or PIs to NNRTIs in patients with more advanced HIV disease.  31173639

Transgender men who have sex with men (MSM) are at high risk for acquiring HIV and have low rates of PrEP uptake.  This is a study out of the Fenway clinic, a Boston community health center that became one of the world’s premier providers of healthcare to LGBT+ folks (if you have a couple of free minutes, check out their history here).  The authors surveyed transgender MSM (people assigned female at birth who identify as men or masculine and who have sex with other men- in this case, specifically transgender MSM who had had sex with a cisgender man in the past 6 months) about their identities, their sexual histories, and whether they were on PrEP.  This was an online survey and recruited trans MSM from across the US.  The authors recruited for the study using peer-to-peer networks (i.e. word of mouth), advertising on the dating/hookup app Grindr, Facebook, and other forms of social media, partnering with community organizations that serve trans MSM, and advertising at the 2017 Trans Wellness Conference.  They included transgender MSM in the study design and authorship of the survey.

In total, 857 people participated in the survey; two-thirds were under 30 years of age, 70% were white, and about 20% identified as Latinx.  Just over 70% identified with a binary gender (i.e. as male / a man rather than genderqueer or non-binary), and one-third identified as gay.  Of the respondents, 80% were receiving testosterone surgery, half had undergone top surgery (i.e. breast reduction or mastectomy), and about 20% had undergone bottom surgery (i.e. phalloplasty).   90% of respondents had health insurance, 80% had had an HIV test, and 1.6% knew themselves to be HIV seropositive.

With regards to PrEP, 84% of the sample had heard about HIV pre-exposure prophylaxis, but only a third had ever taken PrEP and only 22% were taking it currently.  What proportion of the sample had an indication for PrEP depended on which CDC algorithm you applied.  Using the flowchart for MSM, 55% of the sample had an indication (50% due to non-monogamous sex and condomless receptive sex in the past 6 months, 5% due to non-monogamous sex and a bacterial STI diagnosis in the past 6 months).  Applying the CDC’s criteria for heterosexual women, only 31% of the respondents would have met criteria for PrEP; the differences here owe to the heterosexual algorithm being more complicated and requiring an understanding of the partner’s risk of HIV seropositivity (perhaps an argument for just using the MSM criteria for everyone!).  In regression analysis, the factors associated with transgender MSM having an indication for PrEP included NOT meeting sex partners online, not having sex exclusively with cisgender men, greater number of sexual partners, receiving high versus no stigma from sex partners, and a high self-perceived risk of acquiring HIV.

So what is the bottom line? Transgender MSM’s need for PrEP is high, and clinicians would substantially underestimate that risk by assessing these patients using criteria for cisgender heterosexual women rather than MSM.  Moreover, transgender MSM’s uptake of PrEP falls significantly short of their need, and this survey likely OVERestimates the true uptake of PrEP in this population, as its recruitment strategy was biased toward folks with high health literacy (e.g. attendees of a trans wellness conference, people engaged with community organizations that serve transgender MSMS) and unusually good access to care (e.g. 90% of respondents had health insurance and 80% received gender-affirming medications and/or surgery). 31536171

Routine anal pap screening is associated with a lower incidence of invasive anal cancers in people living with HIV.  This isn’t a huge surprise given the efficacy of screening for cervical cancer, another HPV-mediated malignancy, but it’s nice to have more data for a practice that some institutions have already adopted on the basis of rather marginal evidence.  The authors performed a single-center prospective cohort study enrolling patients at an HIV clinic in Spain who either accepted & followed up for at least 6 months with anal cytology screening or declined anal cytology screening; enrollment began in 2005 and this analysis included data gathered through the end of 2016.  The screening consisted of anal cytologies (cytobrush collection from the anal canal) performed every 6-12 months. Abnormal results were followed by high-resolution anoscopy, and either infrared coagulation or surgery (for condylomata) or biopsy (for everything else).  The primary outcome was the incidence of invasive anal cancer during the period of followup, which for each patient included the period from their first visit after Jan 1st 2005 to their last encounter with the clinic.

Between 2005 and 2016, the authors included a total 3111 PLWH (1596 MSM, 888 men who have sex with women [MSW], and 627 women) into the cohort.  Of these, 54% enrolled in the anal screening program.  The folks who enrolled in screening were more likely to be MSM vs MSW, had higher baseline CD4 counts (572 vs 443 cells/ul), and had higher CD4 nadirs (276 vs 215 cells/ul).  At baseline, 44% of subjects had a normal anal cytology; during followup, 17% of the screened subjects developed HSIL and 54% LSIL; while rates of HSIL did not differ between groups, LSIL was more common in MSM than MSW or women (37% vs 12% and 16%; p<0.001).  In total, the screenings led to 1288 high-resolution anoscopies and 744 lesion biopsies.  A total 10 patients were diagnosed with invasive anal squamous cell carcinoma: 2 in the screening group and 8 in the group who declined screening.  All patients had a CD4 nadir under 150 cells/ul, though half had recovered their counts to >350 at the time of cancer diagnosis.  In a regression model of freedom from invasive anal cancer diagnosis, being enrolled in the screening program was protective (HR 0.2; 95% CI 0.4-1.0).

So, it took 1288 anoscopies and 744 biopsies to prevent ~6 cancers among ~1500 people over a decade.  Not bad.  An RCT seems warranted, and in fact is already underway (the ANCHOR study, due to conclude around 2022).  The big question this study brought up for me, though, was exactly who needs this screening.  If everyone who developed invasive anal SCC had a CD4 nadir less than 150 cells/ul, it seems like we could avoid a lot of unnecessary screenings and resultant medical misadventures by only screening the people actually at risk for the disease. 31504329