GI, Hepatobiliary, STI & Skin Syndromes

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

◆A febrile patient with polycystic liver disease — which cyst is infected, and how do you treat it?

In a liver studded with cysts, the diagnostic problem is not whether there is a cyst but which one is infected, and ordinary cross-sectional imaging can’t reliably separate an infected cyst from a bland or hemorrhagic one because in cystic (polycystic) liver disease they all look complicated. The optimal study is FDG–PET-CT: it lights up the single metabolically active cyst against a background of quiet ones, localizing the target and directing therapy. (The transplant/endocarditis use of PET-CT is a different indication for the same tool, covered in the companion Transplant & CMV digest.) Once the culprit is identified, the preferred antimicrobial classes are fluoroquinolones or third-generation cephalosporins — fluoroquinolones are favored partly because they penetrate cyst fluid well — and the recommended duration is a prolonged 4–6 weeks, reflecting a walled-off focus that behaves like an abscess. The usual organisms are enteric gram-negatives, several of which carry an inducible chromosomal AmpC; that organism list and the de-repression logic live in the Antimicrobial digest and are not repeated here.

◆Viral hepatitis a fellow should know cold: which HCV regimen, and who gets screened for hepatitis D?

For a treatment-naïve adult without cirrhosis or with compensated cirrhosis, two pangenotypic direct-acting antiviral regimens cover essentially everyone, so you no longer need the genotype to choose between them: sofosbuvir–velpatasvir (Epclusa), 400/100 mg once daily for 12 weeks, or glecaprevir–pibrentasvir (Mavyret), three 100/40 mg tablets (300/120 mg) once daily with food for 8 weeks. The discriminators a vignette hangs on: Mavyret is the 8-week course and must be taken with food; Epclusa is the 12-week course. The “compensated” qualifier also matters — the protease-inhibitor–containing Mavyret is avoided in decompensated (Child-Pugh B/C) disease, which is outside the naïve/compensated scenario tested here.

The hepatitis B pearl to apply reflexively is about its most dangerous travelling companion. Hepatitis D (delta) is a defective virus that cannot infect anyone who is not already HBsAg-positive, and when it superinfects chronic HBV it produces the most aggressive form of viral hepatitis — cirrhosis develops in roughly 75% of HDV patients within 15 years. Because a positive HBsAg is the only prerequisite, the rule is simple and universal: per EASL, screen every HBsAg-positive individual for HDV at least once. The screening and treatment landscape is expanded in the enrichment box below.

◆“Cellulitis” that isn’t: the soft-tissue look-alikes

Several classic vignettes present as “cellulitis” that fails antibiotics because it was never an infection. The erysipelas-versus-stasis-dermatitis split turns on laterality plus tempo. Erysipelas is an invasive Streptococcus pyogenes infection — unilateral, acute, febrile, sharply demarcated, and the patient looks genuinely ill. Venous stasis dermatitis is a chronic, bilateral, low-grade or afebrile process driven by transudation of protein-rich fluid into the interstitium; the legs are edematous with a brown (hemosiderin) hue, loose weeping blebs that ooze after a day on the feet, and only minimal tenderness. Bilateral + chronic + afebrile = stasis, not cellulitis — the single most common reason a “cellulitis” doesn’t respond to amoxicillin-clavulanate (the two only rarely coexist).

Erythrasma is another mimic: Corynebacterium minutissimum in intertriginous skin (classically the groin, mimicking tinea cruris), giving well-demarcated, dry, asymptomatic patches that fluoresce coral-red under a Wood lamp (porphyrin production); treat with oral erythromycin or topical clindamycin/erythromycin. The non-infectious mimic to know is relapsing polychondritis: recurrent, unilateral “cellulitis” of the ear that responds only sluggishly to antibiotics is the classic trap. Because it is inflammation of cartilage, auricular chondritis spares the fleshy lobule (which contains no cartilage) — the pathognomonic clue — and over time yields floppy ears and a saddle-nose deformity; nasal, ocular, joint, and airway cartilage are also targets, so a history of chronic nasal stuffiness is a tell. Discriminate it from the other saddle-nose causes: syphilis and granulomatosis with polyangiitis (GPA) can flatten the bridge but do not inflame the external ear, and lepromatous leprosy can deform the pinna, but through nodular subcutaneous thickening rather than cartilage loss.

◆A child with eczema erupts in vesicles — herpeticum or coxsackium?

When a child with atopic dermatitis suddenly erupts in vesicles, the fork in the road is HSV versus Coxsackievirus, because one needs admission and antivirals and the other needs supportive care and time. Eczema herpeticum is HSV superinfecting eczematous skin (usually atopic dermatitis, occasionally another inflammatory dermatosis): monomorphic, grouped, “punched-out” vesicles and shallow erosions concentrated where the eczema is worst. It can progress to dangerous disseminated infection, so it warrants hospitalization and antiviral therapy. Eczema coxsackium is the Coxsackievirus counterpart (an atypical hand-foot-mouth picture): polymorphous vesicles and bullae that are larger, more irregular, and acral — hands, feet, arms, legs, and nails — and it is managed supportively. When the morphology is ambiguous, PCR settles it. The discriminator: monomorphic punched-out vesicles over eczema point to HSV (treat); polymorphous acral vesicles and bullae point to Coxsackie (support).

◆Antimicrobials near the end of life: cure, comfort, or a time-limited trial?

Near the end of life, antibiotics stop being an automatic good and become one more intervention to weigh against the goals of care. The useful reframe is to ask what the antibiotic is for: sometimes it is for comfort — a symptomatic urinary tract infection causing dysuria or delirium may be worth treating precisely because treatment relieves symptoms — and sometimes the honest answer is that it offers little beyond the burdens of IV access, monitoring, drug interactions, C. difficile risk, and the illusion of doing something. The structuring tool is the time-limited trial: rather than starting antibiotics open-endedly, the team and family agree in advance on a defined duration and explicit, observable endpoints — is the patient more comfortable, more alert, less symptomatic? — and then reassess at that mark, continuing only if the agreed goal is being met and stopping without a sense of failure if it is not. It converts an emotionally loaded all-or-nothing choice into a bounded, revisitable plan aligned with what the patient actually wants.

◆Where the neighboring STI and genitourinary syndromes live

This digest’s title promises STIs, but several of those syndromes are worked up in sibling digests to avoid duplication. Pelvic inflammatory disease and its perihepatic extension, the Fitz-Hugh–Curtis syndrome — right-upper-quadrant pain from gonococcal or chlamydial perihepatitis that can masquerade as a hepatobiliary problem — are covered in the Bacteriology digest. Chronic bacterial prostatitis (fluoroquinolones as the oral treatment of choice when susceptible) and the arrival of oral gepotidacin for uncomplicated urogenital gonorrhea (the EAGLE-1 non-inferiority data against ceftriaxone plus azithromycin) are handled in the Antimicrobial digest. They are flagged here only so the map is complete.

Sources: AASLD/IDSA HCV Guidance: Recommendations for Testing, Managing, and Treating Hepatitis C (Clin Infect Dis 2023; doi:10.1093/cid/ciad319); EASL Clinical Practice Guidelines (hepatitis B / hepatitis delta; HBsAg-positive HDV screening); Hepatitis D Virus Infection — review (N Engl J Med 2023; doi:10.1056/NEJMra2212151; PMID 37407002).