A retrospective assessment of risk factors for peritonitis among an urban CAPD population

Peritonitis is a major reason why patients transfer from peritoneal dialysis (PD) to hemodialysis. We evaluated the peritonitis infection rates in 146 peritoneal dialysis patients who underwent dialysis at our facility between 1 January 1981 and 31
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
    Peritoneal Dialysis International, Vol. 13, pp 126-131 0896-8608/93 $300 + 00 Printed in Canada All rights reserved. Copyright &copy; 1993 International Society for Peritoneal Dialysis A RETROSPECTIVE ASSESSMENT OF RISK FACTORS FOR PERITONITIS AMONG AN URBAN CAPD POPULATION Stephen M. Korbet, Edward F. Vonesh, and Catherine A. Firanek Section of Nephrology, Department of Medicine, Rush Medical College, Chicago, Illinois, and Baxter Health Care, Round Lake, Illinois, U.S.A. Peritonitis is a major reason why patients transfer from peritoneal dialysis (PD) to hemodialysis. We evaluated the peritonitis infection rates In 146 peritoneal dialysis patients who underwent dialysis at our facility between 1 January 1981 and 31 December 1989. PeritonitIs was the primary cause for changing treatment, with 24 (16.4%) of the patients transferring because of this complication. This represented 54.5% of all patients discontinuing CAPD due to method failure. A gamma-Poisson regression analysis was performed in an attempt to Identify potential risk factors associated with an increased Incidence of peritonitis. The results indicated that race, education level, and PD system used were significantly associated with the rate at which peritonitis occurred in our patient population. There was an almost twofold increase in the rate of peritonitis among blacks as compared to whites (2.2 vs 1.2 episodes/patient year). The level of education completed at the start of dialysis had a negative correlation with peritonitis rates. Patients with &le;8, 9-12, and &ge;13 years of education had peritonitis rates of 2.4, 1.8, and 1.2 episodes/patient year, respectively. Finally, the system used had a significant effect with our patients on CCPD having lower peritonitis rates as compared to patients on either a connect or disconnect system (0.6 vs 2.5 vs 1.8 episodes/patient year, respectively). Recognizing potential risk factors for peritonitis will help us better understand and address this significant problem in our PD programs. Reducing peritonitis rates should facilitate a decrease in patient transfer due to method failure. KEY WORDS: Peritonitis; risk factors. p eritonitis is a major complication associated with chronic peritoneal dialysis (PD) (1-6). It has been shown to be a significant risk factor associated with the success or failure of peritoneal dialysis as a therapy (1,5). By better understanding the distribution of and potential risk factors for peritonitis, a more concentrated and informed approach can be Correspondence to: Stephen M. Korbet, 1653 W. Con gress Parkway, Chicago, Illinois, U.S.A. 60612. Received 23 March 1992; accepted 1 June 1992. taken towards the prevention of peritonitis. Ideally this would result in improved technique survival for peritoneal dialysis. In this paper the peritonitis experience of an urban population of PD patients is examined. A gamma-Poissonregression analysis (3,4) is performed for the purpose of assessing the effects of sex, race, age, type ofPD system, diabetes mellitus (DM), cardiovascular disease (CVD), education level, year of training, and PD training time on the rate of peritonitis. METHODS PATIENT POPULATION Between 1 January 1981 and 31 December 1989 a total of 146 patients began treatment for end-stage renal disease at the Rush-Presbyterian-St. Luke's Medical Center in Chicago, Illinois. Patients began treatment using either continuous ambulatory peritoneal dialysis (CAPD, 112 patients) or continuous cyclic peritoneal dialysis (CCPD, 34 patients). In addition to the type of system used and the dates the patient started and ended the system, information was also collected on the patient's sex, race, age (at the start of treatment), years of education (at the start of treatment), PD training time, the presence or absence of diabetes mellitus and cardiovascular disease, and the reason the patient ended PD. Patients whose history included the diagnosis of cardiovascular disease, angina, myocardial infarction and/or cerebral vascular accident were reported as having cardiovascular disease. The patient and technique survival of this patient population was previously examined (6). Since peritonitis was a major cause of method failure in our patient population, we evaluated our peritonitis experience and the possible risk factors for this complication. Peritonitis was defined by the presence of abdominal pain or cloudy dialysate and a dialysate white-cell count of &ge;100 leukocytes/ &mu;L with more than 50% polymorphonuclear leukocytes. A relapse of peritonitis was defined when the same organism was identified within two weeks of completing antibiotic therapy for the previous infection.  b  y on J   ul   y1  3  ,2  0 1 1  w w w . p d i   c  onn e  c  t   . c  omD o wnl   o a  d  e  d f  r  om     STATISTICAL METHODS For a given patient on a given system, the "observed" rate of peritonitis is given by the number of infections divided by the total time on the system. For a given group of patients the "observed" group rate of peritonitis is the total number of infections divided by the total time. By total number of infections and total time we mean the sum of all infections/ patient and the sum of all the time each patient has spent on the system. In relating the type of PD system as well as patient-related prognostic factors such as sex, race, age, education, etc., to peritonitis, a gamma-Poisson regression model was used (3,4). This model, also known as the negative binomial regression model, assumes the rate of peritonitis over time is constant for an individual patient on a particular PD system. While constant within a patient, the model allows the rate of peritonitis to vary from one patient to another even if the two patients have the exact same set of prognostic characteristics. These assumptions have been thoroughly tested and documented elsewhere (3,4,7). Under this model a patient who switches to a second PD system (i.e., from a connect to a disconnect CAPD system) would have to be treated as a "new" patient in order to reflect the effect of the new system on the individual's rate of peritonitis. In addition, if a patient left treatment for transplantation or other reasons and reentered after a minimum of four to eight weeks, that patient would also be treated as a "new" patient. Of the 146 patients an additional 24 "new" patients would be generated in accordance with the preceding arguments, thus yielding a total of 170 "patients" in the study. Reported here are the results based solely on the first system used by the 146 patients. The results remained unchanged when the 24 "new" patients were included in the analysis. In performing a gamma Poisson regression analysis, an overall model incorporating the effects of sex, race, age, type ofPD system, education, PD training time, year trained, DM, and CVD was fit to the observed numbers of infections and times on PD. Two-way interactions between race, system, and education were included in this overall model in order to determine if a change in one of these factors affects any change observed in a second factor. A likelihood ratio test was carried out to select a final model that would incorporate only those factors that are significantly associated with the rate of peritonitis (4). The results of this analysis are summarized below. RESULTS A total of146 patients (67 females, 79 males) were trained for chronic peritoneal dialysis. The majority ofpatients were black [78] with 55 white patients and 13 patients classified as other (7 Hispanic, 6 Asian). The average age was 48.5&plusmn;15 years (range 11-83 years). There were 41 (28%) patients with diabetes and 52 (35.6%) patients with CVD [not mutually exclusive: 24 (16.4%) patients had both diabetes and CVD]. The average level of education in our patients was 12.5&plusmn;3.3 years (range 3-20 years), and patients averaged 6.4&plusmn;3.4 days of training (range 1-26). The patient demographics are shown in Table 1. The beta  b  y on J   ul   y1  3  ,2  0 1 1  w w w . p d i   c  onn e  c  t   . c  omD o wnl   o a  d  e  d f  r  om     cap system was used in 48 (79%) of the patients on a disconnect system with the remaining 13 patients (all of whom started after 1 February 1989) on a Y-connector system (9 O-set, 4 ultraset). The distribution ofrace, age, education, presence or absence of diabetes, and CVD were all significantly different between the three systems (p<0.05, chi-square test). Only sex was similarly distributed between the three systems (p = not significant). Patient outcomes are provided in Table 2. Peritonitis (Figure 1) was the primary reason for method failure (i.e., leaving treat-ment for reasons other than death, transplantation, and transferring to another dialysis facility on PD). Of the 44 patients with method failure, 24 (54.5%) failed PD because of peritonitis. Table 3 provides a detailed breakdown of peritonitis rates for selected groups ofpatients. The effects ofrace, education, and system used on the overall peritonitis rates are readily apparent. Beginning with a full gamma-Poisson regression model, the combined effects of sex, race, age, system, education, training time, DM, CVD, and year of training on peritonitis were examined. Subsequent models with fewer factors were then examined in order to determine only those factors that show a significant association with peritonitis. The final model found race, education, and type of system as the only significant factors associated with peritonitis. A full description of the model is provided in the Appendix. Table 4 summarizes the results obtained from the model in terms of rate ratios for the various factors investigated. The predicted peritonitis rates (PPR) can be determined by multiplying the predicted peritonitis rate for the control group (3.38 episodes/patient year) by the rate ratio's (RR) correspondingto the factors that comprise the select group of interest if different fram the cantrol group (white, 0 education, connect system). For example, the average peritonitis rate for black patients (RR=1.629) with 12 years of education (RR=0.507) using a disconnect system (RR=0.938) is predicted to be The average peritonitis rate for white patients with 8 years of education (RR=0.636) using a connect system is predicted to be The combined effects of race (white vs black) and education on peritonitis rates, regardless of the type ofPD system utilized, are demonstrated in Figure 2. DISCUSSION Peritonitis continues to be the single most common complication faced by patients on chronic peritoneal dialysis. As such, it has become a leading reason for technique failure among chronic peritoneal dialysis patients accounting for the transfer from CAPD/CCPD in 27%-56% of patients (5,6,8,9). A  b  y on J   ul   y1  3  ,2  0 1 1  w w w . p d i   c  onn e  c  t   . c  omD o wnl   o a  d  e  d f  r  om     better understanding of the risk factors associated with peritonitis is essential if we are to improve the technique survival of our patients on peritoneal dialysis. In the present evaluation a full gamma-Poisson regression model was used to assess the combined effects of sex, race, age, system, education, training time, DM, CVD, and year of training on peritonitis rates in an urban population of peritoneal dialysis patients. In the final model race, level of education, and system were found to be the factors significantly associated with peritonitis. The National CAPD Registry evaluated risk factors associated with the discontinuation of CAPD and found race to have a significant impact with black patients having a higher risk of transfer than white patients (5). In their evaluation, peritonitis rates were significantly higher in patients discontinuing CAPD as compared to those not discontinuing, and the peritonitis rates in black patients discontinuing were significantly higher than those ofwhite patients who discontinued. We also found that the race factor has a significant effect on technique failure, since our black patients had a relative risk of technique failure over twice that of our white patients (6). This may be explained by the present finding that our black patients have almost twice the peritonitis rate as our white patients. In a recent report of peritonitis rates and risk factors in CAPD patients from the United States Renal Data System, race was again found to be a significant factor, with black patients having a higher peritonitis risk than white patients (10). The racial difference in peritonitis rates observed in bur study cannot be accounted for based on differences in age, years of education, PD training time, or the presence or absence of diabetes mellitus and/or cardiovascular disease, because these were taken into account in the statistical model. Furthermore, we previously evaluated these parameters and found them not to be statistically different between our black and white patients (6). The effect of other psychosocial or economic differences was not evaluated in our study and could be a significant factor to account for the overall differences observed. The possibility also exists that the increased peritonitis rates observed in the black patient population may be the result of racial differences in immunologic response. It is well known that certain bacterial infections occur with increased frequencyin the black population (11-13). The increased susceptibility may be the result of inherent differences in the antibody response to bacterial antigens (14). Further evaluation in this area is needed. The relationship between the level of education and peritonitis has not been a factor previously evaluated in studies on peritoneal dialysis. In a recent report often years' experience at one facility, Rubin and Hsu (8) found that increased educational achievement was associated with a prolonged peritoneal dialysis technique survival; however, the effect did not reach statistical significance (p<0.06). Unfortunately, they did not evaluate the effect of education on peritonitis rates. We were previously unable to demonstrate an independent effect of education on the technique survival of our patients (6); however, in the present evaluation we found a significant independent effect of education on peritonitis. This becomes an important issue, since peritonitis was the primary reason for method failure in our patients. Based on our findings, one could speculate that the more educated a patient, the more compliant, and this would lead to fewer episodes of peritonitis and vice versa. Although this is possible,  b  y on J   ul   y1  3  ,2  0 1 1  w w w . p d i   c  onn e  c  t   . c  omD o wnl   o a  d  e  d f  r  om     it is not always the case, and intelligence  per se may not always be linked to the level of education one has attained. In fact, the level of education may be a marker for some other psychosocial factor influencing peritonitis that we have not taken into account. In any event, we do not advocate that the availability of peritoneal dialysis to patients be limited on the basis of the level of education, attained. Instead, we suggest that for those patients with lower levels of education additional support or assistance (such as more frequent follow-up and review of the patient's CAPD technique) may be required to prevent an increased incidence of peritonitis and/or technique failure. Further evaluation of the impact of education and possibly other psychosocial parameters on peritonitis/technique survival is warranted. The effect of the peritoneal dialysis system used on peritonitis rates has long been of interest due to technical advancements that have lead to the development of new exchange sets and connection and sterilization devices. The peritonitis rates of patients on CAPD and CCPD were evaluated by the N ational Registry and found to be similar (1.4 vs 1.3 episodes/ patient year, respectively) (5). There have been several reports, however, that have shown substantially better peritonitis rates for patients on CCPD (0.20.66 episodes/patient year) as compared to those reported with standard CAPD (15-18). The differences in peritonitis rates between these modalities may be the result of selection bias; however, in the study by Holley et al. (15), using matched controls, the peritonitis rate in the CCPD patients was 0.3 episodes/patient year as compared to 1.3 episodes/ patient year for patients on the CAPD spike system. In a prospective randomized study, de Fijter et al. (18) observed a remarkable difference in the peritonitis rates between CCPD and CAPD (0.46 vs 0.9 episodes/patient year, respectively). They found the time to the first episode of peritonitis to be significantly longer in their CCPD patients (> 12 months) as compared to their CAPD patients (3 months, p<0.05). Our gamma-Poisson analysis (adjusting for age, race, sex, etc.) demonstrated that the use of CCPD in our patients had an independent and significant effect on peritonitis rates. The improved peritonitis rates in CCPD patients is felt to be the result of fewer connections and disconnections between catheter and transfer set, an initial flushing of the system prior to filling, and lower catheter-related infection rates (15,16). The development of the Y-connector system with the flush-before-fill concept in CAPD has lead to a substantial improvement in peritonitis rates with rates of 0.43-0.8 episodes/patient year being reported (10,19-21). This has lead to a significant decrease in relative risk to technique failure in comparison to patients maintained on CAPD with the standard connection technique. Based on the gamma-Poisson model, we found that the system used had a significant influence on peritonitis rates with patients on CCPD showing the better outcome. Unfortunately, we did not have enough experience with the y -connector system to separately evaluate its effect in our patient population. From the data available it is apparent that the peritoneal dialysis system used can have a significant effect on peritonitis rates and, therefore, impacts on technique failure. The evaluation for risk factors of peritonitis in our chronic peritoneal dialysis populations can provide invaluable information. If we are to alter the course of this problem and improve the technique failure that undermines the viability of peritoneal dialysis as a chronic modality of dialytic therapy, then we must have an understanding of the patients at greatest risk and the treatment options that offer the greatest likelihood for success. ACKNOWLEDGMENT The authors wish to thank Dr. E.J. Lewis for his support and helpful comments. APPENDIX The final gamma-Poisson regression model is expressed as Estimates of the parameters &alpha;, T, &beta;o, &beta;1, &beta;2, &beta;3, and &beta;4 and their standard errors are summarized in Table 5. To illustrate their use, the predicted rate of peritonitis among white PD patients using a disconnect CAPD system and who completed 12 years of education would be  b  y on J   ul   y1  3  ,2  0 1 1  w w w . p d i   c  onn e  c  t   . c  omD o wnl   o a  d  e  d f  r  om 
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks