Risk factors for candidaemia in hospitalized patients with liver cirrhosis: a multicentre case–control–control study

Abstract

Objectives

The aim of this study was to evaluate the risk factors for candidaemia in patients with liver cirrhosis.

Methods

This was a case–control–control (1:2:2) study performed in four Italian tertiary centres from 2006 to 2015. Cases were patients with liver cirrhosis developing candidaemia. For every case of candidaemia we enrolled two additional patients undergoing blood cultures for suspected infection yielding isolation of a bacterial pathogen (control A) and two additional patients undergoing blood cultures for suspected infection yielding negative results (control B). Patients were matched according to age, sex and model for end stage liver disease at hospital admission.

Results

During the study period 90 cases, 180 controls A and 180 controls B were included. At multivariate analysis assessed by means of multinomial conditional regression models, factors independently associated with candidaemia were previous (<30 days) acute-on-chronic liver failure (relative risk ratio (RRR) 2.22 (95% confidence interval (CI) 1.09–4.54), p = 0.046), previous(<30 days) gastrointestinal endoscopy (RRR 2.38 (95% CI 1.19–4.78) p = 0.014), previous(<30 days) antibiotic treatment for at least 7 days (RRR 2.74 (95% CI 1.00–7.48), p = 0.049), presence of central venous catheter (RRR 2.77 (95% CI 1.26–6.09, p = 0.011), total parenteral nutrition (RRR 3.90 (95% CI 1.62–9.40), p = 0.002) at infection onset and length of in-hospital stay >15 days (RRR 4.63 (95% CI 2.11–10.18), p <0.001] Conversely, rifaximin treatment was associated with lower rate of candidaemia (RRR 0.38 (95% CI 0.19–0.77), p = 0.007). Multivariable analysis for 30-day mortality showed that patients with isolation of Candida spp. from blood cultures had worse outcome when compared with controls even though the difference did not reach a statistical significance (hazard ratio 1.64 (95% 0.97–2.75) p = 0.06).

Conclusions

We identified previous antibiotic use, gastrointestinal endoscopy or acute-on-chronic liver failure and presence of central venous catheter especially for parenteral nutrition as independent factors associated with candidaemia. Surprisingly, chronic rifaximin use was a protective factor.