31/12/19

Variation of effect estimates in the analysis of mortality and length of hospital stay in patients with infections caused by bacteria-producing extended-spectrum beta-lactamases: a systematic review and meta-analysis

Shamsrizi P, Gladstone BP, Carrara E, Luise D, Cona A, Bovo C, Tacconelli E.

BMJ Open. 2020 Jan 20;10(1):e030266. doi: 10.1136/bmjopen-2019-030266., 01/2020.

Objective: To  assess the variation of effect estimates in the analysis of mortality  and length of stay (LOS) in patients with infections caused by  extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae.


Design: Systematic  review and meta-analysis METHODS: Literature search for clinical  studies from 1 January 1960 to 1 October 2018 was conducted in PubMed.  Primary outcomes were risk ratios (RRs) of all-cause and attributable  mortality and weighted mean differences (WMDs) in LOS in patients with  bloodstream infections (BSIs) and non-invasive infections. Any change in  the effect estimates was assessed by grouping studies according to  design, setting, economy-based country classification, reporting period,  microbiological aetiology, infection type and adjustment for  appropriateness of empirical treatment. The impact of ESBL production  was calculated using random-effect meta-analysis and heterogeneity was  evaluated by I2 statistics and metaregression.


Results: Eighty-four  studies including 22 030 patients and 149 outcome measures were  included in the meta-analysis. Most studies were retrospective cohorts  from high-income countries, providing unadjusted estimates. ESBL  production in patients with BSIs (56 studies) increased the RR for  all-cause mortality by a factor of 1.70 (95% CI 1.52 to 1.90;  p<0.001), attributable mortality (16 studies) by 1.75 (95% CI 1.448  to 2.108; p<0.001) and WMD in the intensive care unit by 3.07 days  (95% CI 1.61 to 4.54; p<0.001). WMD in hospital LOS was significantly  higher in BSIs (4.41 days; 95% CI 3.37 to 5.46; p<0.001) and  non-invasive (2.19 days; 95% CI 1.56 to 2.81; p<0.001). Subgroup  analyses showed variation of estimates by study design, population,  strain and assessment of appropriateness of empiric treatment. High  heterogeneity was observed in all analyses.


Conclusions: Current  evidence of the clinical burden of infections caused by ESBL-producing  bacteria is highly heterogeneous and based mainly on unadjusted  estimates derived from retrospective studies. Despite these limitations,  ESBL production in strains causing BSIs seems associated with higher  all-cause and attributable mortality and longer hospitalisation.


Keywords: bloodstream infection; extended-spectrum beta-lactamase; length of stay; meta-analysis; mortality; systematic review.