Quantitative Assessment of Hypermucoviscosity and Siderophore Production in Clinical Isolates of Hypervirulent Klebsiella pneumoniae under Iron-Restriction Stress: Virulence Genotype, Capsular Serotype, Carbapenemase Carriage and Clinical Outcome in Al-Anbar, Iraq.

Abstract

      Background: Hypervirulent Klebsiella pneumoniae (hvKp) is still identified at the bench by the string test, a binary reading of a continuous phenotype. Whether quantitative measures of mucoviscosity and iron acquisition define the pathotype more faithfully than that single threshold has not been settled in Middle Eastern settings, where carbapenem resistance is endemic.

Methods: Two hundred and forty consecutive clinical K. pneumoniae isolates recovered at a tertiary referral centre in Al-Anbar, Iraq, were cultured under iron-restricted and iron-saturated conditions. Hypermucoviscosity was quantified as string-test length and capsular uronic acid concentration; siderophore output was measured by the chrome azurol S assay in µg/mL and percentage siderophore units. Relative expression of iucA and iroB was determined by 2^−ΔΔCt after reference-gene stability testing. Five virulence markers, capsular serotype and blaNDM, blaOXA-48 and blaOXA-51 were detected by PCR. Death was modelled by binary logistic regression.

Results: All four quantitative measures were mutually correlated (Spearman ρ = 0.783, 0.566, 0.528, 0.505, 0.444 and 0.357; all P < 0.001) and differed significantly by capsular type (all Kruskal–Wallis P < 0.001). Siderophore output followed a threshold rather than a dose–linear pattern: median CAS concentrations were 15.03, 14.54 and 16.91 µg/mL at 0, 1 and 2 virulence markers, rising to 52.64, 47.65 and 44.73 µg/mL at 3, 4 and 5 markers. A marker count ≥3 predicted the published CAS cut-off of 30 µg/mL with AUC 0.964 (sensitivity 94.5%, specificity 97.8%). CAS concentration discriminated genotypic hypervirulence better than the string test (AUC 0.989 vs 0.926; ΔAUC 0.063, 95% CI 0.027–0.106; P < 0.001), the loss residing in string-test specificity (62.5% vs 98.4%) rather than sensitivity (91.1% vs 92.9%; κ = 0.383). Iron restriction induced iucA 4.62-fold (IQR 3.84–5.31) and iroB 5.23-fold (IQR 4.24–5.93); both P < 0.001, with a stable rpoB (P = 0.665). In the adjusted model, only virulence-marker burden independently predicted death (OR 1.39 per additional marker, 95% CI 1.02–1.88; P = 0.036; model AUC 0.755).

Conclusions: In this population the hypervirulent phenotype is internally coherent and threshold-defined. A genotype of ≥3 markers coupled with quantitative CAS measurement at ≥30 µg/mL outperforms the conventional string test, and should anchor the operational definition of hvKp in comparable settings.