Emerging imipenem resistance in Nocardia infections in Taiwan: a retrospective cohort study at a tertiary teaching hospital.

BACKGROUND

Nocardia species are emerging pathogens of increasing significance, particularly in immunocompromised hosts. Due to diverse clinical manifestations and resistance patterns, accurate identification is critical. This study characterized species distribution in Taiwan, evaluated identification methodologies, and identified mortality risk factors.

METHODS

We analyzed 99 Nocardia isolates from 76 patients at a Taiwanese tertiary center (2011-2025). Identification via MALDI-TOF MS and concatenated gene sequencing (16S rRNA, secA1, hsp65) was compared against whole-genome sequencing (WGS). Susceptibility was tested using broth microdilution. Potential mortality-associated variables were first evaluated in univariate analyses, and the final multivariable model was constructed using Firth's penalized logistic regression.

RESULTS

N. brasiliensis was most prevalent (27/76; 35.5%), primarily causing cutaneous infections, while N. farcinica (25/76; 32.9%) was linked to pulmonary cases and higher mortality. Firth's penalized logistic regression identified immunosuppressant use (OR 18.44; 95% CI, 3.62-93.89) and higher Charlson Comorbidity Index (CCI) (OR 1.70; 95% CI, 1.23-2.35) as independent predictors of mortality. High rates of imipenem non-susceptibility were observed in N. farcinica (85.7%) and N. cyriacigeorgica (83.3%). In addition, the proportion of imipenem non-susceptible isolates increased significantly during 2020-2025 compared with 2011-2019 (94.1% vs. 67.1%, p = 0.024). Regarding species identification, MALDI-TOF MS demonstrated acceptable performance (PPA = 0.89), whereas concatenated gene sequencing achieved perfect agreement with WGS (PPA = 1.00).

CONCLUSIONS

Accurate species-level identification is crucial for guiding antimicrobial therapy, especially given the emerging burden of imipenem non-susceptibility in clinically important Nocardia species. Immunosuppressant use and a higher CCI were independently associated with mortality, highlighting the importance of early recognition and appropriate management in high-risk patients. Concatenated multilocus sequence analysis based on 16S rRNA, secA1, and hsp65 provides a practical and accurate alternative to WGS for routine species identification in clinical microbiology laboratories.