Mortality attributable to third-generation cephalosporin resistance in Gram-negative bloodstream infections in African hospitals: a multi-site retrospective study
Dramowski A., Ong’ayo G., Rehman AM., Whitelaw A., Labi A-K., Obeng-Nkrumah N., Ndir A., Magwenzi MT., Onyedibe K., Wolkewitz M., de Kraker MEA., Scott JAG., Aiken AM., Mujuru H., Odipo E., Mwarumba S., Mturi N., Gordon NC., Diop A., Ndiaye AT., Shwe D., Diala UM., Okolo M., Grundmann H.
Abstract Background Bloodstream infections (BSI) caused by Enterobacteriaceae show increasing frequency of resistance to third-generation cephalosporin (3GC) antibiotics on the African continent but the mortality impact has not been quantified. Methods We used historic data from six African hospitals to assess the impact of 3GC resistance on clinical outcomes in Escherichia coli and Klebsiella pneumoniae BSI. We matched each bacteraemic patient to two uninfected patients. We compared outcomes between 3GC-susceptible and 3GC-resistant BSI and their respective uninfected controls using Cox regression models. Results For 1431 E. coli BSI patients, we matched 1152 (81%) 3GC-susceptible and 279 (19%) 3GC-resistant cases to 2263 and 546 uninfected inpatient controls. For 1368 K. pneumoniae BSI patients, we matched 502 (37%) 3GC-susceptible and 866 (63%) 3GC-resistant cases to 982 and 1656 uninfected inpatient controls. We found that 3GC-resistant E. coli had similar hazard ratios (HRs) for in-hospital mortality over their matched controls as compared to susceptible infections over their controls (ratio of HRs 1.03, 95% CI 0.73–1.46). Similarly, 3GC-resistance in K. pneumoniae BSI was not associated with mortality (ratio of HR 1.10, 95% CI 0.80–1.52). Estimates of mortality impact varied by site without a consistent pattern. Conclusions In a retrospective analysis, including the use of matched uninfected patients, there did not appear to be an impact of 3GC-resistance on mortality in E. coli or K. pneumoniae BSI in African hospitals, as compared with susceptible BSI with equivalent species. Better information on the actual use of antibiotics in treating infections in African hospitals would improve these impact estimates.