Antimicrobial Resistance Across Clinical Settings

Evidence

Two recent studies mapped antimicrobial resistance in sharply different clinical systems. One examined methicillin-resistant Staphylococcus aureus (MRSA) submitted to a public-health referral laboratory in western Ethiopia. The other examined Neisseria gonorrhoeae cases managed across six Swedish university-hospital sexually transmitted infection clinics. The organisms, specimen pathways, and patient populations are not interchangeable. Their value as a pair lies in showing how strongly a resistance map depends on where samples come from and how cases are detected.

The Ethiopian study was a retrospective cross-sectional analysis covering five years at the Nekemte Public Health Research and Referral Laboratory Center. Among 545 S. aureus isolates, 67.2% were classified as MRSA. The distribution changed significantly over the study period. Most MRSA isolates came from hospital-associated submissions, and outpatient departments were the leading patient location. Middle-ear discharge accounted for 67% of MRSA samples, making this a particularly specimen-weighted dataset rather than a general survey of all MRSA disease in the region.

Resistance was extensive within that isolate collection. The abstract reports that 93.7% of MRSA isolates met its multidrug-resistant definition. Penicillin resistance was 96.7%, while the highest reported susceptibility was 72.1% for gentamicin. Multidrug-resistant MRSA prevalence also varied significantly by health facility, patient location, and specimen type. Those internal differences are as important as the overall percentages: they indicate that even within one regional laboratory network, the apparent resistance burden shifted with the source of the samples.

The Swedish study retrospectively reviewed 1,060 NAAT-confirmed gonorrhea cases from 2022. The six participating university hospitals contributed about one-third of Sweden’s reported cases that year and spanned multiple geographic regions. Culture was used after positive nucleic-acid testing to assess susceptibility, but culture recovery varied by anatomical site. Positive culture occurred in 68.7% of analyzed urethral samples, 54.4% of vaginal or cervical samples, 54.0% of rectal samples, and 33.4% of pharyngeal samples. That variation means susceptibility results were drawn from only part of the NAAT-confirmed case set.

The Swedish resistance pattern was uneven rather than uniformly high. Ciprofloxacin resistance was reported in 43.8% of cases with the study’s tabulated denominator, and azithromycin resistance in 14.2%. One ceftriaxone-resistant case was recorded, while no cefixime-resistant case was found among those tested. The dataset also showed why case ascertainment matters: only 48% of patients were tested because of symptoms, while check-ups after unprotected sex and contact tracing identified many others. Resistance surveillance in this setting was therefore embedded in testing and tracing systems that included infections not first detected through symptoms.

Analysis

Taken together, these studies show why an antimicrobial-resistance percentage is not a portable property of “bacteria” in general. It is a measurement produced by a particular organism, specimen stream, patient population, laboratory method, and health system. The Ethiopian dataset was dominated by MRSA recovered from middle-ear discharge and mostly hospital-linked submissions, whereas the Swedish dataset began with NAAT-confirmed gonorrhea cases across several anatomical sites and depended on successful culture for susceptibility testing. Their numerical resistance estimates therefore answer different questions.

The cross-study inference is that surveillance architecture matters alongside resistance biology. A high multidrug-resistance burden in one referral-laboratory network and uneven resistance across drugs in one national STI-clinic sample both favor granular, continuously refreshed maps over broad assumptions. That is an analysis of the shared pattern, not evidence that the two organisms spread similarly or respond to the same control measures. A useful next research step would be standardized reporting that preserves organism, specimen, anatomical site, care setting, time period, and testing completeness, allowing trends to be compared without erasing the context that generated them.

The studies also converge on a narrower operational point: aggregate results can conceal meaningful variation. In Ethiopia, multidrug-resistant MRSA prevalence differed by facility, patient location, and specimen. In Sweden, the probability of obtaining a culture—and therefore an isolate for phenotypic susceptibility testing—differed substantially by anatomical site. In both cases, the surveillance denominator affects what the resistance map can represent. This does not invalidate either study; it defines the boundary of each finding and cautions against treating a regional estimate as a universal rate.

Limitations

Both studies were retrospective and observational, so neither establishes why resistance patterns developed. The Ethiopian evidence available here is abstract-only. It supports the reported isolate counts, percentages, setting, and associations, but it does not provide enough ingested detail to audit laboratory procedures, missing data, antibiotic panels, or the exact multidrug-resistance definition. Because middle-ear discharge dominated the sample mix, the findings may not describe bloodstream, respiratory, wound, or community MRSA in western Ethiopia. A clinical isolate also does not by itself prove invasive infection; the abstract does not fully resolve infection versus colonization for every submission.

The Swedish study covered six university-hospital clinics and one calendar year. Although the cases represented about one-third of Sweden’s 2022 total, they were not a population-random sample. Culture recovery was incomplete and lowest for pharyngeal samples, creating a possible selection effect in susceptibility measurement. Some resistance denominators differed because not every isolate had every test result. The work characterized gonorrhea care and resistance in a well-organized Swedish clinic network; it cannot be used to infer MRSA patterns, Ethiopian conditions, or resistance rates elsewhere.

Neither study tested an intervention or compared surveillance designs prospectively. The synthesis therefore supports context-aware interpretation of resistance data, not a treatment rule, a patient-level prediction, or a claim that one health system’s findings transfer to another.