hVISA infections in MRSA patients treated with vancomycin

Source article: Vancomycin treatment outcomes and machine learning-based identification of heterogeneous vancomycin-intermediate Staphylococcus aureus infections

Objectives Heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA) infections remain underdiagnosed with unclear clinical impact. This multicenter study evaluated vancomycin treatment outcomes and developed a machine learning (ML)-based diagnostic model. Methods We prospectively enrolled 200 patients with MRSA infections from 20 Chinese hospitals between February 2012 and July 2021. Population analysis profile-area under the curve (PAP-AUC) testing was performed on frozen isolates to confirm hVISA ph…

By the Truvace deskOctober 18, 20262 min read1 sourcePeer-reviewedTRV-2026-1854· cite & verify

Vancomycin treatment outcomes and machine learning-based identification of heterogeneous vancomycin-intermediate Staphylococcus aureus infections — "Work Bench - Microbiology Laboratory, Spotswood Sewerage Pumping Station, Victoria, circa 1939" is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/.

Trace impact readingNegative state

P 73

Both sides are scored from claims and sources, not community votes.

G 67

In brief

Researchers enrolled 200 patients with MRSA infections across 20 Chinese hospitals and found that most, 117, carried heterogeneous vancomycin-intermediate Staphylococcus aureus. They tested isolates in the lab, sequenced confirmed strains, tracked vancomycin levels, and trained machine learning models to predict which patients had hVISA.

hVISA is underdiagnosed, and the results suggest it undermines standard vancomycin therapy, so a practical prediction tool could help clinicians act sooner. The model was evaluated on a held-out test set, but the evidence comes from one study and does not show that model-guided dosing improves patient outcomes.

Main points

117 of 200 MRSA patients (58.5%) had hVISA, including 56.0% of adults and 76.0% of pediatric patients.
Adults with hVISA on vancomycin monotherapy had a 34.6% treatment failure rate versus 5.0% for VSSA.
ST5, ST239 and ST764 were the predominant hVISA clones, and ST5 and ST239 had higher failure rates than ST764.
The best of 10 ML models used vancomycin MIC, tracheal intubation and prior penicillin use as key predictors.

The gain

A LASSO-logistic regression model built on clinical predictors identified hVISA infection in MRSA patients with a test-set AUC of 0.834, which could support earlier detection and tailored vancomycin dosing.

The problem

In adults on vancomycin monotherapy, hVISA infection was linked to much higher treatment failure than VSSA infection, even though hVISA patients received higher doses and still reached lower PK/PD indices.

The rundown

The prospective cohort was enrolled between February 2012 and July 2021. Frozen isolates underwent PAP-AUC testing to confirm the hVISA phenotype, and confirmed strains were whole-genome sequenced. Vancomycin therapeutic drug monitoring and PK/PD analysis were also performed.

Despite receiving higher vancomycin doses, adults with hVISA reached lower PK/PD indices than those with vancomycin-susceptible S. aureus. The authors link this to the observed failure gap in monotherapy and call for early identification.

Ten machine learning models were compared, and LASSO-logistic regression had the best test-set AUC of 0.834 (95% CI 0.711-0.930). The authors say it supports phenotype- and sequence type-guided personalized dosing.

What this doesn’t fix

The model's test-set confidence interval is fairly wide (0.711-0.930), and the study reports prediction performance rather than evidence that model-guided dosing improves outcomes.