Background: Gram-negative complicated urinary tract infections (cUTIs) can have serious consequences for patients and hospitals. Aim: To examine the clinical and economic burden attributable to Gram-negative carbapenem-non-susceptible (C-NS; resistant/intermediate) infections compared with carbapenem-susceptible (C-S) infections in 78 US hospitals. Methods: All non-duplicate C-NS and C-S urine source isolates were analysed. A subset had principal diagnosis ICD-9-CM codes denoting cUTI. Collection time (<3 vs 3 days after admission) determined isolate classification as community or hospital onset. Mortality, 30day re-admissions, length of stay (LOS), hospital cost and net gain/loss in US dollars were determined for C-NS and C-S cases, with the C-NS-attributable burden estimated through propensity score matching. Three subgroups with adequate patient numbers were analysed: cUTI principal diagnosis, community onset; other principal diagnosis, community onset; and other principal diagnosis, hospital onset. Findings: The C-NS-attributable mortality risk was significantly higher (58%) for the other principal diagnosis, hospital-onset subgroup alone (odds ratio 1.58, 95% confidence interval 1.14e2.20; P < 0.01). The C-NS-attributable risk for 30-day re-admission ranged from 29% to 55% (all P < 0.05). The average attributable economic impact of C-NS was 1.1 e3.9 additional days LOS (all P < 0.05), US$1512e10,403 additional total cost (all P < 0.001) and US$1582e11,848 net loss (all P < 0.01); overall burden and C-NS-attributable burden were greatest in the other principal diagnosis, hospital-onset subgroup. Conclusion: Greater clinical and economic burden was observed in propensity-score-matched patients with C-NS infections compared with C-S infections, regardless of whether cUTI was the principal diagnosis, and this burden was most severe in hospital-onset infections.

Aim We aimed to evaluate the association between selective COX‐2 inhibitors (coxibs) and the risk of colorectal neoplasms and vascular events with and without low‐dose aspirin. Method We searched for randomized controlled trials and comparative studies in PubMed, EMBASE and Cochrane Library databases using pertinent key terms. Risk ratios (RRs) were calculated for each study with a fixed‐ or random‐effects model. Results Eight clinical studies with 44 566 subjects were eligible. The use of coxib significantly reduced the overall risk of colorectal neoplasms by 21% (RR = 0.79, 95% CI 0.70−0.89; P = 0.000). The chemopreventive effect of coxibs was beneficial in the first year (RR = 0.74, 95% CI 0.58–0.94; P = 0.013), marginal in the third year (RR = 0.79, 95% CI 0.63−1.01; P = 0.059) and counterproductive in the fifth year (RR = 1.65, 95% CI 1.23−2.21; P = 0.001). Compared with the use of aspirin alone, combined use of coxib and aspirin for 3 years increased the risk of a colorectal neoplasm by 80% in the fifth year (RR = 1.80, 95% CI 1.22−2.66; P = 0.003) but decreased by 79% and 30%, respectively, the risks of cardiovascular thromboembolic events (RR = 1.79, 95% CI 1.33−2.41; P = 0.0001) and renal impairment/hypertension (RR = 1.30, 95% CI 1.09−1.54; P = 0.003) caused by coxib use alone. Conclusion Coxibs may reduce the overall risk of colorectal neoplasms, but the chemopreventive effects are attenuated over time. When participants take low‐dose aspirin simultaneously, coxibs may not be useful for chemoprevention of colorectal neoplasm.