Peritoneal Dialysis International, Vol. 29, pp. 542–547 Printed in Canada. All rights reserved. 0896-8608/09 $3.00 + .00 Copyright © 2009 International Society for Peritoneal Dialysis 542 RETROPERITONEAL LEAKAGE AS A CAUSE OF ACUTE ULTRAFILTRATION FAILURE: ITS ASSOCIATED RISK FACTORS IN PERITONEAL DIALYSIS Man Fai Lam, 1 Wai Kei Lo, 2 Kai Chung Tse, 1 Terrence P.S. Yip, 2 Sing Leung Lui, 2 Tak Mao Chan, 1 and Kar Neng Lai 1 Nephrology Division, 1 Department of Medicine, Queen Mary Hospital; Nephrology Division, 2 Department of Medicine, Tung Wah Hospital, University of Hong Kong, Hong Kong Correspondence to: K.N. Lai, Department of Medicine, Uni- versity of Hong Kong, Queen Mary Hospital, No. 102, Pokfulam Road, Hong Kong. knlai@hkucc.hku.hk Received 3 September 2008; accepted 29 October 2008. ♦ Background: Ultrafiltration failure is an important clini- cal problem in patients on maintenance peritoneal dialysis (PD) and is associated with high morbidity and mortality. Acute ultrafiltration failure (AUFF) is usually secondary to mechanical problems with the peritoneal catheter or peri- toneal leakage. Retroperitoneal leakage (RPL) is an impor- tant cause of AUFF and often poses diagnostic difficulty. Herein we analyze the incidence of AUFF secondary to RPL in our centers and study its associated risk factors. ♦ Methods: After excluding causes due to mechanical prob- lems with the peritoneal catheter, patients complicated by AUFF underwent computerized tomographic peritoneog- raphy (CTP) or magnetic resonance imaging of the perito- neal cavity (MRP) to determine any RPL. Other patients on maintenance PD without RPL served as controls for com- parison of risk factors. Demographic and peritoneal mem- brane characteristics, including history of hernia and pleuroperitoneal leakage, were analyzed. ♦ Results: During the 5-year study period, 36 patients in a cohort of 743 patients on maintenance PD developed AUFF. 23 of these 36 patients were found to have RPL, which was confirmed by either CTP (n = 16) or MRP (n = 7). The duration of PD at the time of RPL and the dialysate-to-plasma ratio of creatinine at 4 hours were 49.3 ± 24.5 (range 0.5 – 87.9) months and 0.70 ± 0.09 respectively. Incidences of hernia (52.2%) and pleuroperitoneal communication (34.8%) were significantly higher than in PD patients without RPL (13% and 7% respectively, p = 0.001). Logistic regression analysis identified hernia and pleuroperitoneal communi- cation as the risk factors for RPL. The odds ratios for RPL with hernia and pleuroperitoneal communication were 6.62 [95% confidence interval (CI) 2.35 – 18.69, p < 0.001] and 6.23 (95% CI 1.83 – 21.19, p = 0.003) respectively. ♦ Conclusion: RPL was not uncommon in patients with AUFF. A high index of suspicion for RPL is needed in the man- agement of patients with history of hernia or pleuroperi- toneal communication presenting with AUFF. Perit Dial Int 2009; 29:542–547 www.PDIConnect.com KEY WORDS: Ultrafiltration failure; retroperitoneal leakage; computerized tomographic peritoneography; magnetic resonance peritoneography. U ltrafiltration (UF) failure is an important clinical problem in patients on maintenance peritoneal di- alysis (PD). It may lead to severe fluid overload and sub- sequently congestive heart failure. Fast peritoneal transport (1), loss of peritoneal surface membrane, and high lymphatic absorption (2) are main causes of UF fail- ure; other anatomical or mechanical etiologies are oc- casionally seen. A patient may present with a sudden reduction in UF volume or with acute ultrafiltration fail- ure (AUFF) in which a mechanical or an anatomical prob- lem has to be seriously considered. Peritoneal leakage due to a tear of the peritoneal membrane (3) is a dis- tinctive clinical entity that can lead to AUFF. These leak- ages are usually associated with localized subcutaneous edema although selected conditions, such as pleuro- peritoneal leakage, are difficult to detect clinically. In addition, AUFF can arise from retroperitoneal leakage (RPL) in patients on PD (4). Herein we analyze the inci- dence of AUFF secondary to RPL in our centers and study its associated risk factors. METHODS AND PATIENTS We studied our patients on long-term PD who were complicated by fluid overload or UF failure between Oc- tober 2002 and September 2007. Recent history of peri- tonitis or any abdominal pathology during the onset of UF failure was assessed and any subcutaneous edema or pericatheter leakage was examined. Following confirma-