Bacterial Skin Infections: From Impetigo to Cellulitis and MRSA

Two organisms, one broken barrier

The skin's barrier is the whole story — breach it, and the commensals move in. As the Foundations chapter sets out, intact skin is a formidable wall: a dry, acidic, tightly-keyed surface colonised by harmless commensals. Bacterial infection almost always begins with a breach — a cut, an abrasion, a burn, a bite, or the cracked, weeping skin of eczema. Through that gap step the two organisms behind the overwhelming majority of skin and soft-tissue infections: Staphylococcus aureus, a Gram-positive cluster-forming coccus that many of us carry harmlessly in the nose, and Streptococcus pyogenes (group A streptococcus), a Gram-positive chain-former. Staphylococcus tends to cause localised, pus-forming (purulent) disease — the boil, the abscess. Streptococcus tends to spread through tissue planes as a diffuse, non-purulent redness. Knowing which of the two you are likely fighting shapes every drug choice that follows.

The backbone drug for both organisms is a narrow-spectrum penicillinase-resistant penicillin — flucloxacillin in the UK and much of Europe, dicloxacillin or a first-generation cephalosporin such as cephalexin elsewhere. As the Antimicrobials chapter explains in full, these beta-lactams kill by binding the penicillin-binding proteins (PBPs), the enzymes that cross-link the bacterial cell wall; without a wall, the organism bursts under its own osmotic pressure. Flucloxacillin's bulky side chain protects the beta-lactam ring from the staphylococcal penicillinase that destroys plain penicillin, which is exactly why it — not amoxicillin — is the anti-staphylococcal workhorse. Streptococcus pyogenes, notably, has never developed penicillin resistance, so it remains exquisitely sensitive to penicillin itself.

Impetigo: the most superficial infection

The honey-crusted rash of childhood — a surface infection with a surface treatment. Impetigo sits in the very top layer of the epidermis, which is why it forms fragile blisters and the classic golden, honey-coloured crust rather than deep swelling. It is highly contagious, common in children, and usually caused by Staphylococcus aureus (sometimes Streptococcus pyogenes). Because it is so superficial, the treatment is driven by how much skin is involved, not by reaching for systemic drugs. For a small, localised patch, topical therapy is enough: a topical antibiotic such as mupirocin or fusidic acid, or — for very limited non-bullous impetigo — topical hydrogen peroxide 1% cream, an antiseptic that avoids antibiotic exposure altogether and is now a recommended first step in several guidelines precisely to spare antibiotics.

When impetigo is widespread, bullous, or accompanied by systemic upset, it graduates to an oral antibiotic — flucloxacillin first-line, with clarithromycin or another macrolide for the penicillin-allergic. One resistance lesson deserves emphasis: topical fusidic acid is effective but its heavy, unregulated use for minor skin complaints has driven rising staphylococcal resistance in several countries. That is a textbook illustration of stewardship — the more casually we deploy a topical antibiotic across trivial rashes, the faster we lose it. Reserving fusidic acid, favouring mupirocin or antiseptics for the mildest disease, and using oral drugs only when the surface approach cannot reach, are all the same principle applied at different depths.

Folliculitis, furuncles and carbuncles

The next tier is infection of the hair follicle, almost always staphylococcal. Folliculitis is superficial inflammation of the follicle — tiny pustules pierced by a hair, often needing nothing more than an antiseptic wash. When infection deepens into the follicle and surrounding dermis it becomes a furuncle (a boil): a tender, red, walled-off collection of pus. Several adjacent boils that coalesce into a single deep, multi-headed mass form a carbuncle, typically on the thick skin of the neck or back. The pharmacological pearl here is that pus needs drainage, not just a prescription: a fluctuant boil or carbuncle is treated first by incision and drainage, with an oral anti-staphylococcal antibiotic (flucloxacillin) reserved for surrounding cellulitis, systemic symptoms, or immunocompromise. Recurrent boils should raise the question of Staphylococcus aureus carriage and the possibility of a resistant strain.

Cellulitis and erysipelas: infection goes deep

The hot, spreading red leg — where empirical cover and the decision to admit both matter. Cellulitis is infection of the deep dermis and subcutaneous fat: a warm, tender, poorly-demarcated spreading redness, usually of a lower leg, often with fever and malaise. Erysipelas is its more superficial cousin — brighter, raised, with a sharply-defined border, classically on the face or shin and classically streptococcal. Because both are driven by Streptococcus pyogenes and Staphylococcus aureus, empirical therapy must cover both, and flucloxacillin does exactly that first-line (some add or substitute a penicillin where pure streptococcal erysipelas is likely). For patients who are systemically well, oral therapy at home is appropriate; the key skill is knowing when to escalate. Marking the leading edge of the redness with a pen and reviewing lets you see, objectively, whether the antibiotic is winning.

The penicillin-allergic patient forces the alternatives to the front, and they draw directly on the Antimicrobials chapter's protein-synthesis inhibitors. Clarithromycin (a macrolide) and clindamycin (a lincosamide) both bind the bacterial 50S ribosomal subunit and cover Gram-positive skin flora well; clindamycin has the added value of switching off toxin production in severe streptococcal disease. Doxycycline (a tetracycline binding the 30S subunit) is another oral option and, usefully, also covers community MRSA. Admission for intravenous therapy is warranted when there is rapid spread, high fever or sepsis, immunocompromise, significant comorbidity such as poorly-controlled diabetes, involvement of the face/orbit, or failure of oral treatment — and, above all, at any hint of the necrotising infection described below.

MRSA: when flucloxacillin simply fails

One acquired gene rewrites the target — and the whole beta-lactam class stops working. Meticillin-resistant Staphylococcus aureus is not resistant because it destroys the drug — it is resistant because it has changed the target. As the Antimicrobials chapter details, MRSA carries the mecA gene, which produces an altered penicillin-binding protein, PBP2a. Beta-lactams work by fitting into the PBP active site; PBP2a has such low affinity for them that flucloxacillin, meticillin, and essentially every penicillin and cephalosporin can no longer bind. The cell wall keeps being built despite the drug. This is why the failure is class-wide: it is not a stronger penicillin that is needed but a drug that attacks the wall — or the ribosome — by a route PBP2a does not guard.

For community-acquired MRSA skin infection, several oral agents still work and are chosen by local sensitivity: doxycycline (30S ribosome), co-trimoxazole — trimethoprim–sulfamethoxazole, which blocks two sequential steps of bacterial folate synthesis — and clindamycin (50S ribosome), with the caveat of testing for inducible resistance. For serious or hospital-acquired MRSA the workhorses are intravenous: vancomycin, a glycopeptide that binds the D-alanyl-D-alanine terminus of the cell-wall precursor and blocks cross-linking upstream of the PBPs entirely (which is why the altered PBP2a is irrelevant to it), and linezolid, an oxazolidinone that binds the 50S subunit to shut down protein synthesis and has excellent oral bioavailability. Both are covered in depth in the Antimicrobials chapter's glycopeptide and protein-synthesis-inhibitor sections; vancomycin's narrow therapeutic window and need for level monitoring are the classic teaching points.

Necrotising fasciitis: the surgical emergency

At the far, dangerous end of the spectrum lies necrotising fasciitis — infection racing along the fascial planes beneath the skin, killing tissue faster than it visibly reddens. The warning signs are pain out of all proportion to the visible findings, rapid progression, systemic toxicity, and later skin changes (dusky discolouration, blistering, crepitus). The single most important message is that this is a surgical emergency, not a pharmacological one: survival depends on immediate, aggressive surgical debridement to remove dead tissue, with broad-spectrum intravenous antibiotics (typically covering Gram-positives including MRSA, Gram-negatives and anaerobes, often with clindamycin added specifically to suppress streptococcal toxin) as vital support, never a substitute. Antibiotics alone do not cure necrotising fasciitis; the surgeon does. Recognising it early and escalating without delay is the whole task.

Stewardship: when NOT to give antibiotics

The hardest prescribing decision is often the decision not to prescribe. Every bacterium on this page also lives, harmlessly, on healthy skin — which is exactly why the crucial distinction is colonisation versus infection. A wound swab growing Staphylococcus aureus in a patient with no redness, no warmth, no pain and no systemic upset is telling you the organism is present, not that it is causing disease; treating a positive swab in the absence of clinical infection breeds resistance for no benefit. Eczema is the classic trap, linking to the Eczema chapter: almost all eczematous skin is colonised by Staphylococcus aureus, and colonisation can worsen the eczema, but this is not the same as clinical infection. Weeping, pustules, honey-crusting and a sudden flare point to true infection worth treating; ordinary dry, itchy, chronic eczema does not, and its mainstay remains emollients and topical anti-inflammatories, not antibiotics. Reserving antibiotics for genuine infection, choosing the narrowest effective agent, draining pus, and repairing the barrier are the four habits that keep these drugs working for the next patient.