Carbapenem Resistance in an ICU
An ICU in a 120-bed hospital records 14 patients colonized with Klebsiella pneumoniae. Meropenem is started for three patients with fever and rising oxygen needs. Admission surveillance cultures show meropenem MICs of 0.5, 1, and 2 mg/L for patients A, B, and C. After 10 days, repeat cultures show MICs of 32, 16, and 2 mg/L. Patient B never received a carbapenem and was in the bed next to A; hand-hygiene audits show 58% compliance. The infection-control team can either restrict carbapenems, improve hand hygiene, or both. Budget permits one intensive intervention for 6 months. The hospital director asks for a recommendation before the next outbreak review.
1. Which single intervention should the team prioritize, and what causal evidence would support or weaken your choice?
Read the worked answer
Prioritise hand hygiene, but treat the choice as provisional. Patient B's rise from 1 to 16 mg/L without carbapenem exposure points to transmission of an already resistant strain or plasmid, not merely selection within B. Patient A's larger rise is compatible with selection under meropenem. With one intensive intervention, hand hygiene addresses the demonstrated breach at 58% compliance and can reduce new colonisations; carbapenem restriction would reduce selection pressure but would not by itself stop bed-to-bed spread. Supporting evidence would be identical isolates or plasmid types from A and B, and a fall in new colonisations after compliance rises above 90%. Weakening evidence would be distinct strains in A and B, or resistance emerging in C despite low carbapenem use. If both were affordable, combine them.
2. Patient C's MIC remains 2 mg/L. Does this prove meropenem is ineffective for C? Explain using exposure and breakpoint reasoning.
Read the worked answer
No. An MIC of 2 mg/L is an in vitro inhibitory concentration, not a clinical outcome. Whether meropenem works depends on free-drug exposure relative to the organism's MIC at the infection site, the dose and infusion strategy, protein binding, and the local clinical breakpoint. If the breakpoint is 8 mg/L and a high-dose prolonged infusion maintains free concentrations above 2 mg/L for enough of the dosing interval, C may still respond. Conversely, a susceptible MIC does not guarantee cure in pneumonia if lung penetration is poor. The correct judgement needs C's clinical trajectory, infection site, dose, and local breakpoint, not the MIC alone.
3. Design a 6-month measure that could distinguish selection from transmission, naming the data and a comparison group.
Read the worked answer
Run prospective surveillance on two comparable ICUs or wards, one assigned to intensified hand hygiene and one to carbapenem restriction, and collect admission and weekly rectal cultures. The primary outcome is new colonisation with meropenem-resistant K. pneumoniae. Add whole-genome sequencing of paired isolates to identify shared strains and plasmids. If resistant isolates cluster by bed proximity and decline only in the hygiene ward, transmission dominates. If resistance emerges in patients with high carbapenem exposure and diverse genetic backgrounds, selection dominates. Matching by staffing, case mix, and length of stay reduces confounding; randomisation is preferable if feasible.