HIV, ART & Opportunistic Infections

Educational study digest pitched at a new ID fellow. Clinical recommendations trace to the cited guidelines and trials; the enrichment box is reviewer synthesis of additional evidence, each claim cited. Not medical advice, and not a substitute for the primary guidelines.

◆The kidney that failed on tenofovir — injury or illusion?

Two very different creatinine stories share a drug class. Tenofovir disoproxil fumarate (TDF) is directly toxic to the proximal tubule: it accumulates in tubular mitochondria and knocks out reabsorption of phosphate, glucose, amino acids, and bicarbonate. The bedside signature is a Fanconi syndrome — a rising creatinine alongside hypophosphatemia, normoglycemic glucosuria, proteinuria, and a non–anion-gap metabolic acidosis. Boosting agents make it worse: co-administered ritonavir or cobicistat raise tubular tenofovir exposure. The injury is usually reversible if caught early, and the move is to switch to tenofovir alafenamide (TAF), which delivers far lower plasma tenofovir and spares the kidney and bone.

Contrast that with the far more common — and entirely benign — bump seen with integrase inhibitors. Dolutegravir and bictegravir (like cobicistat and trimethoprim) inhibit the tubular transporters OCT2/MATE that secrete creatinine. Serum creatinine ticks up by <0.3 mg/dL within weeks and then plateaus, with a completely normal phosphate, glucose, and bicarbonate and no true fall in GFR. The discriminator a fellow is tested on: an isolated small creatinine rise with a clean tubular panel is transporter competition, not nephrotoxicity — don’t stop the INSTI. (For orientation, HIV-associated nephropathy is a third pattern entirely: heavy proteinuria, rapidly falling GFR, and large echogenic kidneys in someone with uncontrolled viremia.)

◆The abacavir reaction a normal gene test won’t catch

Abacavir hypersensitivity is the reaction fellows both over- and under-diagnose. It is a multi-organ syndrome that appears within the first six weeks and requires at least two of fever, rash, gastrointestinal symptoms, constitutional complaints, and respiratory symptoms. Its defining behavior is that it worsens with each successive dose — the opposite of a self-limited virus, which is exactly the distinction a vignette hangs on. Pharmacogenomic screening for HLA-B*5701 (validated by the PREDICT-1 trial) has sharply reduced immunologically confirmed cases and is standard before any abacavir-containing regimen. The trap is treating the test as a rule-out: a negative HLA-B*5701 lowers the risk but does not exclude clinical hypersensitivity. If the syndrome is clinically suspected, the answer is always the same — stop abacavir and never rechallenge. Rechallenge after a genuine reaction can precipitate fatal hypotension within hours, so “stop and try again later” is never correct.

◆The asthma inhaler that made Cushing’s

A patient on a ritonavir- or cobicistat-boosted regimen who develops moon facies, central weight gain, and striae after starting an inhaled or intranasal corticosteroid has a drug interaction, not a new endocrine disease. Ritonavir is a potent CYP3A4 inhibitor, and most inhaled steroids — fluticasone above all — are CYP3A4 substrates cleared by that pathway. Block it, and systemic fluticasone exposure rises roughly 350-fold, producing iatrogenic (exogenous) Cushing syndrome. The confirmatory clue is a low ACTH: the exogenous glucocorticoid suppresses the axis, so this is secondary adrenal suppression, not the high-ACTH picture of a primary adrenal problem. The fix is to avoid the combination in the first place — beclomethasone is the inhaled steroid of choice with a boosted regimen because it is far less CYP3A4-dependent. The same caution applies to intra-articular triamcinolone and to budesonide.

◆Long-acting and switch therapy — and the inducer that breaks it

Long-acting injectable cabotegravir plus rilpivirine (CAB/RPV) is a maintenance option, not an induction one. It is FDA-approved as a switch for a virologically suppressed patient with no history of treatment failure and no resistance to either drug (established by the ATLAS and FLAIR trials, with every-two-month dosing validated in ATLAS-2M). The engineering that makes it convenient also makes it unforgiving: cabotegravir is cleared by UGT1A1 and rilpivirine by CYP3A4, and the depot half-lives are long — roughly 6–12 weeks for cabotegravir and 12–28 weeks for rilpivirine. That pharmacologic “tail” means any drug that drives those enzymes will hold both agents at subtherapeutic levels for months, selecting two-class resistance. Hence the board-defining rule: a potent enzyme inducer — rifampin or carbamazepine — is an absolute contraindication to CAB/RPV.

Rifamycins recur across every HIV interaction question, so it is worth holding the whole grid in one place:

The same “know the prior agent” logic governs ART after PrEP breakthrough. Always send a genotype first; if the breakthrough happened on injectable cabotegravir PrEP, also send an integrase genotype, because cabotegravir’s tail keeps selecting INSTI resistance long after the last shot. While results are pending, bridge by what the patient was taking: continue oral TDF/FTC or TAF/FTC and add dolutegravir or bictegravir; after a cabotegravir breakthrough, avoid another integrase inhibitor and start a boosted darunavir-based regimen; after lenacapavir PrEP, standard first-line therapy is appropriate. For patients whose problem is adherence rather than suppression, the LATITUDE trial extended the evidence for long-acting CAB/RPV into the non-suppressed, adherence-challenged population under close case management — a use the current panel recommendations now recognize in selected patients.

◆Opportunistic and HIV-associated infections

Coccidioidomycosis in an endemic-area patient with advancing immunosuppression turns on the difference between screening and disease. A patient whose coccidioidal serology converts from negative to positive while the CD4 is <250, and who has no active disease after a thorough evaluation, should start preemptive fluconazole 400 mg daily, continued until the CD4 climbs above 250 and the virus is suppressed. Two distinctions are heavily tested: primary prophylaxis is not recommended even at low CD4 in endemic regions, and itraconazole’s niche is bone disease, not routine coccidioidomycosis.

Severe mucocutaneous HSV in advanced HIV — the enlarging, necrotic, sometimes verrucous anogenital ulcers that can mimic squamous cell carcinoma — is treated with IV acyclovir 5 mg/kg every 8 hours, stepping down to oral once lesions regress and continuing until healed. The detail that separates a strong answer from a weak one is to send both HSV PCR and a viral culture with susceptibility testing: acyclovir-resistant, thymidine-kinase–deficient HSV is a real problem in this population and mandates a switch to foscarnet. Biopsy to exclude carcinoma is reasonable but must never delay antivirals.

Advanced HIV also raises susceptibility to invasive enteric bacteria — Shigella, Campylobacter, and nontyphoidal Salmonella — not through stomach acid but through disruption of mucosal immunity and intestinal barrier integrity; rates run at least tenfold higher than in the general population and rise as CD4 falls below 200. (Shigella is intrinsically acid-resistant, which is why the “low gastric acid” distractor is wrong.) And CMV end-organ disease in HIV is predominantly retinitis, clustering at CD4 <50 — the lower the count, the greater the tendency to cross into ocular and neurologic tissue. The full CMV workup and therapy live in the companion Transplant & CMV digest.

◆When the lymphoma lives in a body cavity

A patient with advanced HIV, pleural or peritoneal effusions but no mass or lymphadenopathy, and cytology showing large plasmablastic cells has primary effusion lymphoma (PEL) until proven otherwise. PEL is driven by HHV-8 (KSHV) and carries a deliberately confusing “null” B-cell immunophenotype: it is CD20-, CD19-, CD79a- yet CD138+, CD30+, and HHV-8 LANA-positive, and it is EBV-coinfected in roughly 80% of cases. HHV-8 LANA positivity is required for the diagnosis. Two management points follow directly: first-line therapy is ART plus EPOCH, and rituximab is omitted because the tumor is CD20-negative. Keep the wider HHV-8 family in mind — PEL, Kaposi sarcoma, multicentric Castleman disease, HHV-8–positive DLBCL, and KSHV inflammatory cytokine syndrome (KICS) — and don’t confuse PEL with plasmablastic lymphoma, which is EBV-driven, starts in the oral cavity, and has a different immunophenotype.

◆Prevention: choosing a PrEP agent

The efficacy conversation has moved decisively toward long-acting agents, and the deciding variable is usually adherence. Twice-yearly subcutaneous lenacapavir, a capsid inhibitor, is the most effective option demonstrated to date: in PURPOSE 1, it produced zero infections among 2,134 cisgender women against a background incidence of 2.41 per 100 person-years, and across PURPOSE 1 and 2 it prevented roughly 99.9% of infections. Injectable cabotegravir every two months outperformed daily oral TDF/FTC in HPTN 083 and 084. Among oral options, daily TDF/FTC covers all exposures; TAF/FTC (Descovy) is not approved for receptive vaginal sex; and on-demand “2-1-1” dosing is validated only for cisgender men. One safety rule underlies all of them: acute HIV must be excluded before any long-acting PrEP, since dosing into undiagnosed infection is functional monotherapy and selects resistance.

Sources: DHHS/OARAC Guidelines for the Use of Antiretroviral Agents in Adults and Adolescents with HIV (2026); NIH/CDC/IDSA Guidelines for the Prevention and Treatment of Opportunistic Infections in Adults and Adolescents with HIV (2026); IDSA Coccidioidomycosis (Clin Infect Dis 2016; doi:10.1093/cid/ciw360); ANCHOR anal HSIL treatment (NEJM 2022); Anal cancer screening in younger MSM with HIV (Open Forum Infect Dis 2026; doi:10.1093/ofid/ofag418); PURPOSE 1 (NEJM 2024; doi:10.1056/NEJMoa2407001; PMID 39046157); PURPOSE 2 (NEJM 2024; doi:10.1056/NEJMoa2411858); ATLAS / FLAIR (NEJM 2020; doi:10.1056/NEJMoa1904398); LATITUDE (NEJM 2026; doi:10.1056/NEJMoa2508228); PREDICT-1 HLA-B*5701 (NEJM 2008); ritonavir–corticosteroid interaction (J Antimicrob Chemother 2019).