In which some New Englanders (diplomatically) tell off the CDC/USPHS PrEP guidelines. I guess one of the nice things about working at Harvard is that you can tell other medical professionals that the CDC is wrong and expect them to listen. This editorial focuses on a few specific shortcomings of the 2017 CDC/USPHS Clinical Practice Guidelines (available here), and I’m sharing it because I thought the comments were mostly well-considered. Let’s walk through their points one by one:
(1) The CDC guidelines recommend clinicians bring up PrEP selectively based on an assessment of patient’s risk, rather than discussing PrEP with everyone. This should be a no-brainer. Sex is stigmatized in the US, especially when the people having that sex aren’t cisgender and heterosexual. The same is true of injection drug use. These aren’t things it’s reasonable to expect every patient to disclose, particularly if they don’t have a longstanding therapeutic relationship with the person seeing them. So, unless your clinic is staffed by Lyta Alexander and Lwaxana Troi, it’s unreasonable to expect your healthcare providers to accurately identify everyone who needs to hear about PrEP (or sexual health education in general). Instead, the clinician ought to bring up these topics with every patient.
(2) The CDC guidelines aren’t consistent about for whom they recommend PrEP. For example, the guideline’s summary states that women who engage in sexual activity high-risk for HIV acquisition (e.g. infrequent condom use, high partner number) qualify for PrEP, while the recommended indication text boxes only suggest PrEP for women with STIs or partners who have or are known to be at high risk for HIV. This isn’t nitpicking - it’s the difference between >80% of women presenting to Planned Parenthood being eligible for PrEP versus only 2% being eligible!
(3) The CDC’s three HIV exposure risk categories – MSM, heterosexual men and women, and PWID – are needless complicated and muddle together sexual orientation, gender identity, and risky behavior. I have some sympathy for the guideline writers on this one – of course we all want guidelines based in data, but let a few guidelines panelists be overly dogmatic about having the data write the guidelines and your recommendations will come out unhelpful, stupid, or both. Not all people who have penises are men, and many people who have sex with men and/or women aren’t heterosexual, and all of these folks might or might not be at elevated risk of acquiring HIV depending on their individual behaviors. The authors argue (and I agree) that the CDC’s recommendations should be simplified and based on patient’s behaviors re: sex and drugs without stratification based on their gender identity or sexual orientation. I think this is probably the most compelling point in the paper - anyone who can’t navigate gender & sexuality 101 isn’t equipped to take care of people with or at risk of acquiring HIV, and so these HIV risk categorizations and their limitations really aren’t an acceptable misstep.
(4) The CDC’s criteria for PrEP fail to consider HIV risk factors beyond the individual (e.g. epidemiologic / community-level risk factors). This passage didn’t work for me. I read it as implying that the authors think the CDC should have recommended being black as an indication for MSM to receive PrEP, yet this isn’t stated outright in the proposed revisions. I understand where they’re coming from data-wise, but the idea of racially-stratified public health recommendations makes me pretty uncomfortable, and by the way the authors write this section I wonder if they felt the same.
(5) The CDC presents “HIV-positive partner” as a sexual risk factor without qualifying the partner’s viral suppression status. Another no-brainer in the era of U=U, though I think it’s reasonable to include the caveats that patient might still benefit from PrEP if they’re concerned about the veracity of their partner’s viral suppression status, or if they have other undisclosed sex partners (hence my rule of thumb: anyone who wants PrEP needs PrEP).
(6) The CDC guidelines aren’t consistent about the relevance of condoms to PrEP eligibility. On the one hand, the guidelines suggest PrEP isn’t warranted if patients use condoms correctly and consistently; on the other hand, the guidelines state that patients on PrEP should be counseled to use condoms and that PrEP was approved for use in combination with other safe sex practices. So which is it: are PrEP and condoms an unnecessary overlap, or two great tastes that taste great together? The authors suggest the message should be that PrEP is highly effective at preventing HIV whether or not condoms are used, though condom use is still valuable for preventing other STIs and pregnancy.
Link to the article: 30689766
Sertraline failed to improve outcomes of cryptococcal meningitis in a new double-blind RCT. Alas. The in vitro data for sertraline as an antifungal was reasonably promising across a wide spectrum of fungi, and in an observational study that used a historical cohort of patients with cryptococcal meningitis as the comparator group, addition of high-dose sertraline seemed to be associated with faster clearance of CSF fungal cultures. But, sometimes a beautiful theory is ruined by data.
This trial, out of Uganda, randomized adults with HIV and cryptococcal meningitis 1:1 to receive either standard therapy (in this case, Ampho deoxycholate 0.7-1mg/kg/day for 1-2 weeks and then stepdown to 800mg/day fluconazole – unfortunately these folks don’t have access to either liposomal amphoB or adjunctive flucytosine) or standard therapy plus 400mg/day sertraline that was then tapered over a 4-month period. The primary outcome of interest was 18-week mortality.
A total 460 patients were enrolled in the trial before it was stopped for futility. The mean age was 35, half had a baseline Glasgow coma score <15, half were on ART, and about 10% were also being treated for TB. The groups were well-balanced in terms of demographics and other variables. Death by week 18 occurred in 52% of the patients treated with sertraline versus 46% of the patients treated with placebo (HR 1.2 with 95% CI 0.9-1.6). No benefit to sertraline was observed in any of the subgroup analysis, and in contrast to the previous human study, sertraline did not appear to speed clearance of fungal cultures. Neurocognitive scores were no different between groups among the survivors, but depression scores were modestly lower in the sertraline-treated group (CES-D score of 13.5 vs 16.6 with p=0.052).
So, sertraline is an effective treatment for people who are depressed about having cryptococcal meningitis. Great. 31345462
While we’re on the subject of great ideas that ended with a sad trombone noise, another Lancet paper out this month report the efficacy of a tenofovir-DF impregnated intravaginal ring for HIV pre-exposure prophylaxis. This was a phase-1 trial conducted in 17 sexually active, HIV-negative women in 2017. The primary endpoint was the safety of the intravaginal ring (i.e. incidence of grade 2+ adverse events) over a 3-month period; in addition, the authors studied the concentrations of tenofovir achieved in cervicovaginal fluid and plasma, and also the levels of active drug metabolite in cervical tissue and dried blood. The women were randomized 3:1 to receive either the drug-impregnated ring or a placebo ring. Of the 12 women given the drug-impregnated right, eight had to discontinue therapy due to development of ulcerations around the ring. Whoops! Drug-eluting intravaginal rings for HIV prevention have a lot to recommend them – they may be easier to adhere to than a pill that needs to be taken every day, they are discrete, and they can be co-formulated with contraceptives if desired. The dapirivine ring, an NNRTI-eluting device, actually has good data for safety and reasonable data for efficacy. But it seems tenofovir DF is the not the drug to use. 31320290
Looking for big (CD4) gains? Patients who have recently started ART have occasionally asked me something along the lines of “how high is my CD4 count supposed to get?” This study in AIDS sheds some prognostic light. The authors reviewed HIV treatment data from the Athens Multicenter AIDS Cohort Study (AMACS) database, which includes all but two of the HIV clinics in Greece and has been enrolling patients for over 20 years. Specifically, they enrolled all patients starting a NNRTI, PI, or INSTI-based combination ART regimen from 2003-present who had at least 6 months of followup with serial HIV viral load and CD4 measurements. The primary outcome of interest was the trajectory and endpoint of the CD4 count after initiation of ART, stratified by the baseline CD4 count.
A total 3,405 patients were included in the analysis. Most patients received either an NNRTI (51%) or a PI (42%) as INSTIs have only been available in recent years. The cohort was mostly male (86%), white (92%), and able to achieve a undetectable viral load for 90%+ of the duration of followup after starting ART (77%). The median duration of followup was 3.9 years. Figure 1 shows the growth in CD4 count after starting ART, stratified by baseline CD4. In short, folks who started off with fewer than 500 cells/ul had about a 400 cell/ul total increase in CD4 count over a 5-year period, most of which they accumulated in the first 1-2 years after starting ART – though the lower the baseline CD4 count, the slower the regained those counts on ART. Folks who started off with a CD4 count of at least 500 cells/ul had about a 200 cell/ul total increase in CD4 over the same period, and they accumulated essentially all of those extra CD4 cells in the first 2 years after starting therapy. Finally, those with lower viral loads off ART enjoyed faster gains in CD4 count after starting ART, and those on ART who spent a greater proportion of time with an undetectable viral load gained CD4 cells faster as well.
Long story short: your patients can expect to see the CD4 increase on ART for about 5 years, though most of the movement will be in the first 2 years, and most folks will (eventually) be able to get at least into the 400-600 cells/ul range, though people who start with lower counts may take a lot longer to get there. All the more reason to start HIV treatment at diagnosis! 31305332
Ocular syphilis is increasing in incidence, and is often a manifestation of early syphilis and syphilis in HIV. The authors conducted a case-control study at the University of British Columbia, matching patients with ocular syphilis diagnosed over an 8-year period 1:4 with controls taken from the pool of all syphilis diagnoses made in BC during the same period. The study used the CDC’s case definition of ocular syphilis, which is compatible ocular disease (e.g. anterior uveitis, uveitis, panuveitis, optic neuropathy, interstitial keratitis, retinal vasculitis, or diminished visual acuity – for this study, these had to be attributed to syphilis by the treating physician) plus syphilis at any stage. Controls were matched to cases on the bases of age, gender, and date of diagnosis.
A total 6,716 people in British Columbia received a diagnosis of syphilis between 2010-2018. Of these, ocular syphilis represented 66 cases (1%) – though of note the incidence almost doubled over the course of the study period. The median age of the cases was 49, 88% of subjects were men, 55% were MSM, and 91% had no known prior diagnosis of syphilis. Thirty-eight percent of subjects with ocular syphilis were diagnosed with primary or secondary syphilis, and half were HIV-positive (median CD4 470; 45% with a viral load <50 copies/ml). The RPR titer was >1:32 in 88% of cases (and among the HIV-positive patients, 100% had an RPR of >1:32), but the only CSF abnormality seen in a majority of cases was WBCs >5 (in 65%). The most common ophthalmologic involvements were panuveitis (see in 42%), optic neuritis (in 33%), and retinitis (in 32%), and of the 45 patients who had document visual impairments and known followup, 41 (91%) had improvement in their visual symptoms after treatment. When comparing the cases and controls in multivariate regression analysis, ocular syphilis was associated with primary or secondary stage syphilis (OR 5.0 with 95% CI 1.9-13.2), early latent syphilis (OR 4.3 with 95% CI 1.6-11.3), and HIV-positive serostatus (OR 2.2 with 95% CI 1.1-4.1).
To summarize all that – ocular syphilis seems to be most common in the early (primary and secondary) stages of syphilis, is more common in people living with HIV - but not necessarily those with a low CD4 count - and is associated with a high RPR titer and a variety of ophthalmologic diagnoses. 31420644
Speaking of syphilis, a retrospective study of 559 MSM treated at the Melbourne Sexual Health Center found that MSM who solely practice insertive anal intercourse (colloquially, “exclusive tops”) were more likely to present with primary rather than secondary syphilis than MSM who engaged in receptive anal intercourse (colloquially, “bottoms”) (OR 3.4 with p<0.001). This makes sense when you look at how the cases of primary syphilis were diagnosed: 73% presented with a penile chancre, versus only 23% with an anal chancre, and consistent use of condoms with insertive anal intercourse was associated with a lower odds of being diagnosed with primary syphilis (aOR 0.5 with p=0.02). Basically – primary syphilitic chancres in the anal canal are likely to be missed by patients, clinicians, or both, and so patients who bottom are more likely to have their syphilis missed in the early stages. So I guess the take home message is, know your patient’s sexual history and screen high-risk folks for syphilis regularly even in the absence of genital symptoms. 31420649