Fellow's quick read · Infectious Disease
Skin & Soft Tissue Infections
IDSA SSTI 2014 (Stevens et al., CID 2014;59:e10–52) · local Michigan & our institution CPGs · reviewed 2026-06-23
Personal study digest for a new ID fellow. Recommendations are from the cited guideline(s) and local CPGs; the orange "What's changed since 2014" box is reviewer synthesis of newer evidence, each claim cited. Not a substitute for the full guideline.
In one line
Sort every SSTI on two axes — PURULENT vs NONPURULENT and mild / moderate / severe — because purulence points to S. aureus (drain it, cover MRSA) and bland cellulitis is β-hemolytic strep (a β-lactam), while severity sets oral-vs-IV and how broad you go. The one diagnosis you cannot miss is a necrotizing infection — that's a surgical emergency, not an antibiotic decision.
When to suspect / classify
- Purulent (abscess, furuncle, carbuncle) → S. aureus, MSSA and MRSA. Nonpurulent (cellulitis, erysipelas) → β-hemolytic streptococci.
- Severity (drives oral vs IV, narrow vs broad): mild = local only; moderate = systemic signs; severe = failed oral therapy/I&D, or SIRS (T >38 °C, HR >90, RR >24, WBC >12 000 or <400), or immunocompromised, or signs of deep infection (bullae, skin sloughing, hypotension, organ dysfunction).
- Red flags for necrotizing infection: pain out of proportion, rapid progression, bullae/skin necrosis, crepitus, hypotension/toxicity. Pain beyond the visible margin in a sick patient = explore.
Workup the fellow drives
- Purulent → incision & drainage is the primary treatment (strong, high). Send pus for Gram stain & culture; tissue/aspirate beats swabs.
- Simple cellulitis: blood cultures and routine aspirate/biopsy not recommended (strong, moderate). Culture blood/tissue only in systemic illness, neutropenia/immunocompromise, immersion injury, or animal bites.
- Don't over-image bland cellulitis. Reserve MRI for pyomyositis or suspected necrotizing disease — and never let imaging delay surgical exploration.
- Consider pseudocellulitis when "bilateral lower-leg cellulitis" is really stasis dermatitis, lipodermatosclerosis, or contact dermatitis — true cellulitis is almost always unilateral.
Empiric & definitive therapy
| Scenario | Drug / dose / route | Notes & strength |
|---|---|---|
| Nonpurulent, mild | Cephalexin 500 mg PO q6h (local: 1000 mg PO TID) or dicloxacillin 500 mg PO q6h | Cover strep; outpatient (strong, moderate) |
| Nonpurulent, moderate | Cefazolin 1–2 g IV q8h or ceftriaxone 1 g IV q24h | IV β-lactam; step down when improving |
| Nonpurulent, severe / deep concern | Vancomycin + piperacillin-tazobactam 4.5 g IV q6h | Broad until necrotizing excluded (strong, mod); narrow fast |
| Purulent, mild | I&D alone (± adjunctive TMP-SMX or clindamycin — see box) | Drainage is the treatment (strong, high) |
| Purulent, moderate | I&D + TMP-SMX 1–2 DS PO BID or doxycycline 100 mg PO BID | Empiric MRSA; avoid empiric clindamycin locally (resistance) |
| Purulent, severe | I&D + vancomycin IV (alt: daptomycin, linezolid) | Tailor to culture; MSSA → cefazolin/nafcillin |
| Necrotizing fasciitis | Emergent surgery + vancomycin (or linezolid) + pip-tazo or carbapenem; add clindamycin for toxin suppression | Documented group A strep → penicillin + clindamycin (strong, low) |
| Animal bite | Amoxicillin-clavulanate 875 mg PO BID (IV: ampicillin-sulbactam) | Must cover Pasteurella + anaerobes; cephalexin misses Pasteurella |
Duration & stopping
- Uncomplicated cellulitis: 5 days, extended only if not improved by then (strong, high). Most mild-moderate SSTI: 5–7 days.
- Necrotizing infection: antibiotics continue until no further debridement is needed and the patient is clinically stable off pressors — duration is driven by source control, not the calendar.
- Resolution of erythema lags clinical cure (esp. with lymphedema); spreading redness isn't automatically treatment failure.
Key decisions a fellow owns
- Drain or drug? Purulent collection → I&D first; antibiotics are the adjunct, not the primary therapy.
- When to add MRSA coverage to cellulitis: purulence, penetrating trauma, known MRSA, injection drug use, or SIRS — not reflexively for every cellulitis.
- Recognize necrotizing disease early and call surgery now; antibiotics and imaging must not delay the OR.
- Prevent recurrence: treat tinea pedis, edema, and toe-web maceration; consider prophylactic penicillin for ≥3–4 episodes/year.
- Local stewardship: our institution/Michigan MRSA is ≥30% clindamycin-resistant — don't use clindamycin for empiric MRSA coverage; TMP-SMX and doxycycline keep the lowest resistance.
Special populations
- Immunocompromised / neutropenic: broaden empirically (add gram-negative ± antifungal cover), culture aggressively, biopsy atypical lesions early — odd organisms and rapid deterioration are the rule.
- Diabetic foot infection: a distinct entity — multidisciplinary (podiatry/vascular/wound) and now governed by the IDSA/IWGDF 2023 update, not this guideline.
- β-lactam allergy: verify it (allergy/BLAES consult); most "penicillin allergy" tolerates cefazolin and is worth de-labeling.
- Renal function: dose-adjust vancomycin and TMP-SMX; watch TMP-SMX hyperkalemia in CKD.
What's changed since 2014
Reviewer synthesis — not the guideline. Each claim cited.
- Adjunctive antibiotics after I&D of even SMALL abscesses now help — overturning the 2014 "I&D alone" default. TMP-SMX beat placebo after drainage (cure 80.5% vs 73.6%, +6.9 pp) Talan, NEJM 2016; PMID 26962903, and in abscesses ≤5 cm clindamycin or TMP-SMX both beat placebo (~83% vs 69%), with clindamycin giving fewer 1-month recurrences (6.8% vs 13.5%) but more diarrhea Daum, NEJM 2017; PMID 28657870. Benefit limited to S. aureus.
- For NONPURULENT cellulitis, adding MRSA coverage does not improve cure — cephalexin + TMP-SMX = cephalexin alone (per-protocol 83.5% vs 85.5%, P=.50) Moran, JAMA 2017; PMID 28535235. Reinforces β-lactam monotherapy for bland cellulitis.
- Five new ABSSSI agents (2014–2018) the 2014 text barely anticipated: dalbavancin, oritavancin, tedizolid (all 2014), delafloxacin (2017), omadacycline (2018). The long-acting lipoglycopeptides give SINGLE-dose IV MRSA therapy (oritavancin 1200 mg; dalbavancin 1500 mg), enabling admission-free treatment. DOTS tested 2-dose dalbavancin in complicated S. aureus bacteremia: not superior by DOOR but noninferior on clinical efficacy Turner, JAMA 2025;334:866–877.
- Necrotizing infection — clindamycin's grip is loosening. It remains the toxin-suppression standard, but group A strep clindamycin resistance now reaches 30–50% at some US centers, so linezolid is the emerging anti-toxin alternative CID 2023;76:346 focused debate; CDC STSS guidance 2025. IVIG for streptococcal toxic shock has not shown a clear mortality/quality-of-life benefit in RCTs.
- Still current: 5-day duration for uncomplicated cellulitis holds (strong, high) — the post-2014 push is to actually stop at 5–7 days. No newer IDSA SSTI guideline exists (2014, erratum 2015); only diabetes-related foot infection got its own 2023 update.
Anki cards minted this run
- PURULENT vs NONPURULENT as the first empiric branch point (Staph + drain vs strep + β-lactam).
- Nonpurulent cellulitis — adding TMP-SMX to cephalexin gives no benefit (Moran, JAMA 2017).
- Drained abscess ≤5 cm — clindamycin vs TMP-SMX adjunct: similar cure, clinda fewer recurrences but more diarrhea (Daum, NEJM 2017).
- Long-acting lipoglycopeptides (dalbavancin/oritavancin) for single-dose ABSSSI; DOTS 2025 in complicated SAB.
Held (dedup): adjunctive-antibiotic-for-abscess concept, necrotizing-fasciitis empiric regimen + clindamycin/linezolid toxin note, and animal-bite → amoxicillin-clavulanate (Pasteurella) are already in the deck.
Sources: Stevens DL et al. Practice Guidelines for the Diagnosis and Management of Skin and Soft Tissue Infections: 2014 Update by the IDSA. Clin Infect Dis 2014;59(2):e10–52 (doi:10.1093/cid/ciu296). Michigan Medicine SSTI CPG (rev. 11/2024). Our institution SSTI CPG (rev. 10/2021). Talan DA et al. NEJM 2016;374:823–32 (doi:10.1056/NEJMoa1507476; PMID 26962903). Daum RS et al. NEJM 2017;376:2545–55 (doi:10.1056/NEJMoa1607033; PMID 28657870). Moran GJ et al. JAMA 2017;317:2088–96 (doi:10.1001/jama.2017.5653; PMID 28535235). Turner NA et al. (DOTS) JAMA 2025;334:866–877 (doi:10.1001/jama.2025.12543). Linezolid-vs-clindamycin focused debate, Clin Infect Dis 2023;76:346. CDC streptococcal toxic shock syndrome clinical guidance, 2025. PubMed metadata via the PubMed MCP.