Three CRE Isolates in a Medical ICU: Rate, Case Definition, and Investigation
A 24-bed medical intensive care unit in a 480-bed community teaching hospital admits a 68-year-old man transferred from a long-term acute care hospital (LTACH) with a sacral pressure injury. On hospital day 6 his wound culture grows carbapenemase-producing Klebsiella pneumoniae (NDM). Over the next five weeks, two more ICU patients grow carbapenem-resistant Enterobacterales (CRE) from clinical cultures collected on days 9 and 17 of their ICU stays. All three had indwelling devices and were cared for with a shared portable glucose meter and a bladder scanner. The ICU logged 1,050 patient-days during the six-week period. The laboratory performs the modified carbapenem inactivation method (mCIM) on all carbapenem-resistant Enterobacterales but does not screen for CRE. Nursing reports that contact precautions for the index patient began only after the culture result returned, three days after collection. The hospital does not report CRE laboratory-identified (LabID) events to NHSN and has no CRE admission flags.
1. Calculate the CRE incidence density for the six-week period and state the surveillance case definition you would apply. Should the index patient's culture count as a hospital-onset event?
Read the worked answer
Denominator: 1,050 ICU patient-days. Numerator depends on the definition you state out loud.
Under NHSN MDRO/CDI LabID logic, onset is assigned by specimen collection date relative to facility admission date (day 1 = day of admission): day 1–3 = community-onset, day 4 or later = hospital-onset, with duplicate events for the same patient within 14 days counted once.
- Index patient: culture on day 6 → hospital-onset by definition, even though acquisition almost certainly occurred at the LTACH. The LabID definition is calendar-based, not causal; that mismatch is a known limitation and must be discussed, not hidden.
- Rate using all three: 3 ÷ 1,050 × 1,000 = 2.86 per 1,000 patient-days.
- Rate if the index case is attributed to the transferring facility (ICU-acquired only): 2 ÷ 1,050 × 1,000 = 1.90 per 1,000 patient-days.
Key reasoning: with no historical baseline and no prior CRE, a rate of 2.86 is not by itself interpretable - the trigger is the temporal cluster plus shared equipment, not the number. Report the definition used, because a reader comparing you to a benchmark must know whether imported cases are in the numerator. Also note that mCIM positivity (NDM metallo-β-lactamase) matters clinically: NDM producers are not covered by ceftazidime-avibactam and require aztreonam-containing regimens, which affects how long patients remain colonized and shedding.
2. Outline the investigation you would launch in the first two weeks. What data would you collect, what hypothesis would you test first, and what role do surveillance cultures and molecular typing play?
Read the worked answer
Verify before investigating: confirm susceptibility and mCIM/PCR results with the lab, confirm the three isolates are distinct patients, and check whether any earlier CRE were missed (look back at 12 months of susceptibility data).
Line list fields: admission/transfer dates and source facility, ICU room and bed moves, device days (central line, urinary catheter, ventilator), procedures, wound care, shared equipment exposure (glucose meter, bladder scanner, rectal thermometer, ECG leads), staff assignments, antibiotic exposure, and outcome.
Hypothesis to test first: point-source/contact transmission via shared, inadequately cleaned equipment plus a three-day delay in contact precautions - that is the most parsimonious explanation and the most fixable. Test it by observing glucose meter cleaning against the manufacturer's instructions for use, auditing hand hygiene and PPE doffing, and reviewing whether the meter was dedicated or shared across rooms.
Surveillance cultures: a point-prevalence rectal/perirectal survey of ICU patients estimates the hidden colonization reservoir; CRE outbreaks are frequently driven by colonized, unrecognized patients. Do not expect surface cultures to be useful - they are low-yield and rarely actionable.
Molecular typing: whole-genome sequencing or rep-PCR on the three isolates. Clonal isolates support transmission and a common vehicle; unrelated isolates suggest repeated importation and shift the intervention toward admission screening and interfacility communication. Alert the LTACH and any receiving facilities. Audit carbapenem use - stewardship changes colonization pressure. Set a defined review date (e.g., two weeks) with pre-agreed triggers for escalation.
3. What surveillance and reporting infrastructure changes would you recommend to prevent a recurrence and to make the next cluster visible sooner?
Read the worked answer
Make the problem measurable. Enroll CRE in the NHSN MDRO LabID module so hospital-onset events are captured with a standard definition, and add CRE to the facility's notifiable-condition list if the state health department requires it (many do for carbapenemase-producing organisms).
Shorten the detection gap. The three-day delay in contact precautions is the most direct failure. Institute a daily review of all positive microbiology with same-day precaution initiation, plus an electronic admission flag for any patient with prior CRE, and an alert for transfers from LTACHs and other high-acuity facilities that prompts a risk assessment at admission.
Feed data back where decisions are made. Report unit-level CRE rates and device days monthly to ICU leadership and the infection prevention committee, with process measures alongside outcome measures: hand hygiene compliance, equipment cleaning audits, time from culture collection to precaution initiation.
Consider admission surveillance cultures for patients from high-risk facilities. Evidence for universal CRE screening is not strong enough to mandate it; frame it as a risk-assessed, time-limited intervention during an outbreak, with a defined stop date and an evaluation plan.
Do not overpromise. Surveillance changes make cases visible; they do not by themselves reduce transmission. Pair every new screening test with the action it triggers, or you generate cost and alert fatigue without benefit.