De-escalating broad-spectrum therapy in ESBL E. coli bacteraemia with a debrided foot infection
A 63-year-old man with type 2 diabetes (HbA1c 10.2%) is admitted with a deep diabetic foot infection and sepsis. He underwent surgical debridement on day 1. Empiric therapy was meropenem 1 g IV every 8 hours, vancomycin 1 g IV every 12 hours and metronidazole 500 mg IV every 8 hours. On day 5 he is afebrile, haemodynamically stable, and the wound is clean with healthy granulation tissue. Admission blood cultures grew Escherichia coli that is ESBL-producing, meropenem-susceptible (MIC 0.5 mg/L), ertapenem-susceptible (MIC 0.25 mg/L), ceftriaxone-resistant and ciprofloxacin-resistant. Deep wound cultures grew the same E. coli plus Enterococcus faecalis; there was no growth of anaerobes or MRSA. White cell count has fallen from 21 to 9.8 x 10^9/L and procalcitonin from 12 to 1.4 ng/mL. Serum creatinine is 1.1 mg/dL and the vancomycin trough is 14 mg/L.
1. Which of the three antimicrobials can be stopped now, and on what evidence? Be explicit about what you would continue and how you would handle the enterococcus.
Read the worked answer
Stop vancomycin. There is no MRSA in blood or deep wound cultures, no Gram-positive cocci in clusters on Gram stain, and the patient is improving. A trough of 14 mg/L is "therapeutic" but irrelevant without a target organism; continuing exposes him to nephrotoxicity and cost.
Stop metronidazole. No anaerobes were isolated from deep tissue and meropenem already has reliable anaerobic cover. The fear of anaerobes in diabetic foot infection applies mainly to ischaemic, necrotic or poorly debrided wounds; this one has been debrided and is granulating.
Continue meropenem 1 g IV every 8 hours. It covers the ESBL E. coli at MIC 0.5 mg/L, and renal function (creatinine 1.1 mg/dL) does not require adjustment (no reduction until CrCl below 50 mL/min).
Enterococcus faecalis: meropenem does not reliably cover E. faecalis. In a polymicrobial, adequately debrided foot infection with clear clinical improvement, specific anti-enterococcal therapy is usually unnecessary when it is a co-isolate rather than the dominant organism. If the patient deteriorates, or if the wound becomes limb-threatening again, add ampicillin 2 g IV every 6 hours after confirming susceptibility. Do not add it reflexively just because the organism was reported - that is how clinicians end up back on three drugs.
2. Should therapy be changed from meropenem to ertapenem because the isolate is ertapenem-susceptible and ertapenem is cheaper? Justify your answer with pharmacology and clinical evidence.
Read the worked answer
No - keep meropenem. Ertapenem is a carbapenem but is not the preferred carbapenem for ESBL-producing Enterobacterales bacteraemia:
- It is about 95% protein-bound, so free-drug exposure at the site of infection is lower than with meropenem.
- An inoculum effect and clinical failures have been described with ertapenem in ESBL bacteraemia, particularly with a high bacterial load or an undrained source.
- Outcome data are observational and weaker than for meropenem or imipenem; no randomised trial supports ertapenem for ESBL bacteraemia.
- Cost saving on the vial is trivial compared with the cost of treatment failure in a bacteraemic diabetic patient.
Ertapenem would be defensible only for a non-bacteraemic, well-controlled infection with a low MIC and guaranteed source control, or when meropenem is genuinely unavailable and the patient is stable. That is not this patient: he had documented bacteraemia and presented with sepsis.
Also resist the temptation to use meropenem at a lower dose or as a prolonged infusion purely to save stock - 1 g every 8 hours as a 30-minute infusion is adequate when the MIC is 0.5 mg/L and renal function is normal.
3. Define the total duration of therapy, the endpoint criteria and when you would extend treatment. Include the oral switch decision.
Read the worked answer
Duration: 7 days of effective therapy is sufficient for uncomplicated gram-negative bacteraemia when source control has been achieved, based on randomised evidence of non-inferiority of 7 versus 14 days. Count from the day source control was achieved (debridement on day 1) or from the first negative blood culture, whichever is later; do not restart the clock for every afebrile day.
Extend to 10-14 days if: blood cultures remain positive at 48-72 hours, there is undrained abscess or underlying osteomyelitis, the patient is neutropenic, or there is endovascular infection (endocarditis, infected aneurysm).
Endpoint criteria: afebrile for at least 48 hours, improving wound, down-trending white cell count and procalcitonin, no new metastatic focus, and stable renal function. Repeat blood cultures are not routinely needed in a responding patient but are mandatory if fever recurs.
Oral switch: not possible here because the isolate is ciprofloxacin-resistant. If earlier discharge is desired, check trimethoprim-sulfamethoxazole susceptibility, but do not use oral therapy while the patient is still bacteraemic. Arrange podiatry follow-up, offloading, and optimisation of glycaemic control - uncontrolled diabetes is the reason this infection will recur, and it is part of the antimicrobial stewardship decision.