Candidaemia in patients with dialysis-dependent acute renal failure: aetiology, predisposing and prognostic factors

Background. Infections remain the major cause of death among patients with acute renal failure (ARF), especially in severe ARF necessitating dialysis therapy (ARF(d)). Although the clinical features and outcomes of candidaemi a in various patient populations have been described, data concerning candi daemic episodes among patients with ARF(d) are scarce. This study investiga ted the aetiology, predisposing, and prognostic factors for candidaemia in the ARF(d) patient population. Three patient groups were investigated in th is study. Methods. During an 8-year study period from January 1992 to December 1999, 37 candidaemic episodes that developed among 653 ARFd patients were assigne d to ARF(d) candidaemic group, and 170 candidaemic episodes developing in p atients without ARF(d) or chronic uraemia as the non-ARF(d) candidaemic gro up, and 28 matched ARF(d) patients without candidaemia were assigned to the ARF(d) control group. Among these groups, clinical characteristics in ARF( d) candidaemia patients, predisposing factors, and outcomes were compared. Four management strategies including central catheter removal, anti-fungal therapy, both, or neither were applied. The prognostic factors for attribut able death were evaluated by univariate analysis followed by the multivaria te logistic regression analysis. Results. The proportion of ARF(d) patients with candidaemia was significant ly higher than in patients who had no ARF(d) or chronic uraemia (5.7% vs 0. 15%, P<0.001). Compared with the non-ARF(d) candidaemic group, systemic lup us erythematosus (SLE), administration of corticosteroid, and central venou s catheter-associated candidaemia were more common in the ARF(d) candidaemi c group (P<0.05). In matched case-control study, multiple antibiotic usage was shown to be a predisposing factor for developing candidaemia in patient s with ARF(d), and corticosteroid therapy has a marginal significance (P=0. 059). The occurrence of candidaemia increased the mortality rate of ARF(d) (71% vs 39.2% in ARF(d) control group, P < 0.05). By multivariate logistic analysis, the variables associated with attributable death in ARF(d) candid aemic group were identified to be an APACHE II score of <greater than or eq ual to> 18, and anti-fungal therapy for > 48 h. Central venous catheters we re removed in 32 (86.5%) of the 37 ARF(d) candidaemic patients, among whom the 18 patients who had received anti-fungal therapy for > 48 h had a lower attributable death rate than those patients who had not (27.8% vs 64.3%, P < 0.05). Of the remaining five patients who did not have their catheter re moved, three patients subsequently died and two patients improved only afte r catheter removal. Conclusions. The higher prevalence of candidaemia in ARF(d) patients is due to their underlying illnesses and multiplicity of predisposing factors, ra ther than ARF and dialysis therapy per se. Predisposing factors include SLE , indwelling central venous catheter, multiple antibiotic usage, and cortic osteroid therapy. Prompt anti-fungal therapy and catheter removal should be mandatory for ARF(d) patients with candidaemia.