Impact of Age on Peritonitis Risk in Peritoneal Dialysis Patients: An Era Effect

Abstract

Background and objectives: Despite reductions in the frequency of peritoneal dialysis (PD)-related infectious complications over time, peritonitis and catheter infection remain important causes of morbidity and mortality. Given the increasing number of elderly patients reaching end-stage renal disease, making informed decisions about PD utilization is contingent on an understanding of the infectious complications of PD in this population. We therefore studied the impact of age on infection rates, organisms and outcomes.

Design, setting, participants and measurements: On the basis of data collected from 1996 to 2005 in the multicenter Baxter Peritonitis Organism Exit sites Tunnel infections database, the study population included 4247 incident Canadian PD patients: 1265 patients aged ≥70 yr and 2982 patients aged <70 yr. We defined two eras of PD initiation: 1996 to 2000 and 2001 to 2005.

Results: In a negative binomial model, older age was independently associated with a higher peritonitis rate (rate ratio [RR] 1.06 per decade increase; 95% CI 1.01 to 1.10; P = 0.008). However, this association was present only among those who initiated PD at an earlier time (RR 1.13 per decade increase; 95% CI 1.07 to 1.20; P < 0.001 in 1996 to 2000 versus 1.01 per decade increase; 95% CI 0.95 to 1.06; P = 0.81 in 2001 to 2005). Catheter-related infections were less frequent with increasing age regardless of era (RR 0.93 per decade increase; 95% CI 0.89 to 0.97).

Conclusions: The higher peritonitis rate observed in elderly patients may represent an era effect, as age was not associated with peritonitis among patients initiating PD between 2001 and 2005. In addition, catheter infection was less frequent with increasing age.

Despite several advances that have led to important reductions in the frequency of peritoneal dialysis (PD)-related infectious complications, peritonitis remains a significant cause of morbidity and mortality (1–4), and infection is the most common reason for transfer to hemodialysis (HD) (5,6).

As life expectancy increases worldwide, more elderly patients are reaching end-stage renal disease (ESRD), requiring initiation of dialysis (7). Given that older dialysis patients have a higher comorbidity burden than do younger dialysis patients (7), older patients may not be similar to their younger counterparts in the propensity to develop PD-related infections, and the outcome of these infections may also differ. Although data on infectious complications in the general PD population are relatively robust, there are fewer data characterizing infectious complications of PD in the elderly patient.

In a study comparing PD-related infection rates among 63 nondiabetic patients older than 70 yr and 86 PD patients aged 40 to 60, there was a significantly higher peritonitis rate in the elderly, but no difference in exit site infection rate (8). In contrast, other studies have found similar peritonitis rates in older versus younger PD patients (9,10,11). Even among those with similar peritonitis rates, the spectrum of organisms has been reported to be different, with one study demonstrating increased Gram negative peritonitis (9), and another showing more Staphylococcus epidermidis peritonitis in the elderly (10). The latter study also found a lower exit site infection rate in the elderly.

The primary objective of the current study was to examine the impact of age on infection rates, infecting organisms, and outcomes of PD peritonitis and PD catheter infection. The secondary objective was to assess for an era effect with regard to the relationship between age and PD-related infections.

Materials and Methods

Patients

The study included PD patients from 25 centers across Canada for whom data were available through the Peritonitis Organism Exit sites Tunnel infections (POET) database. The data from all Canadian PD centers using the POET clinical monitoring system software (Baxter Healthcare) were collected as described previously (12). The database includes prospectively collected data on incident PD patients who initiated dialysis between 1996 and 2005, as well as data on prevalent patients from as early as 1990 that were retrospectively entered into the database when their center started using the POET software. Only prospectively collected data on incident patients were used for this study. Information contained within the POET database includes patient demographics, cause of infection, catheter complications, and therapy transfers.

For the purpose of providing demographic characteristics, patients were categorized as younger or older using a threshold of 70 yr (<70 yr versus ≥70 yr). Demographic data available for the current study include age, gender, race, cause of ESRD, diabetes status, modality before PD start (new to dialysis, transfer from HD, failed transplant, other/unknown) and PD modality (continuous ambulatory peritoneal dialysis [CAPD] versus automated peritoneal dialysis). For the latter variable, the first PD modality used, as well as the PD modality at the time of death, transfer to HD, transplantation or censoring were documented.

Given that the prospective cohort included patients who initiated PD over a 10-yr period, we defined two eras of patients to assess for an era effect: an earlier cohort consisting of those who initiated PD between 1996 and 2000, and a more recent cohort consisting of those who initiated PD between 2001 and 2005. The era cutoff was chosen based on a prehoc hypothesis that peritonitis rates would decrease around the time that mupirocin use was adopted as routine exit site care in Canada. This hypothesis was tested by determining the association between age and peritonitis rates for each year of PD initiation during the study.

Infectious Complications

The infectious complications studied were peritonitis and catheter infection. Catheter infection was defined as a PD catheter exit site infection and/or tunnel infection.

Infection Outcomes

Possible outcomes of infection in the database included death, catheter removal, resolution, cuff removal, other failure, and ”no outcome listed.” To ensure data validity, only data from centers with at least 95% of peritonitis outcomes reported were included in the outcome analyses.

Statistical Analyses

Continuous variables are reported as mean ± SD, and were compared between younger and older patients by means of the t test. Categorical variables are reported as percentages and were compared between groups using the χ² test. The association between age and the rate of peritonitis or catheter infection was tested in a univariate negative binomial regression model, as well as in a multivariable negative binomial model that included gender, race, diabetic status, cause of ESRD, modality before PD start, and PD modality as covariates. A negative binomial model was also used to study the relationship between organism-specific infection rates and age. The association between age and infection outcomes was tested with a logistic regression model. To assess for an era effect of age, we used an age × era interaction term as an initial screening; if found to be statistically significant, subsequent analyses were performed for each of the two eras. Statistical significance was defined as a P value of <0.05. All statistical analyses were performed using SAS (version 9.1).

Results

Of the 6544 patients in the database, there were 4247 incident PD patients in whom data were collected prospectively, and 2297 prevalent patients in whom data were entered retrospectively. The study sample was restricted to the 4247 incident PD patients, of whom 1265 were ≥70 yr of age and 2982 were <70 yr of age. The study sample included a total of 7319 yr of follow-up: 2041 yr of follow-up in those patients ≥70 yr and 5278 yr of follow-up in those patients <70 yr old. Demographic characteristics of patients in each age group are presented in Table 1. There were significant differences between younger and older PD patients: Elderly patients were more likely to be male (58% versus 54%; P = 0.015) and Caucasian (88% versus 80%; P < 0.001), and were less likely to be diabetic (35% versus 43%; P < 0.001) compared with younger patients. The distribution of the causes of ESRD was different between the age groups, with diabetes, glomerulonephritis, and cystic disease less likely to be the cause of ESRD in the older cohort compared with the younger cohort (P < 0.001). CAPD was the initial PD modality used in the majority of patients, with a higher proportion of older patients subsequently remaining on CAPD as compared with younger patients (56% versus 50%, P < 0.001).

Table 1.

Demographic characteristics

Characteristic	Age ≥70 (n = 1265)	Age <70 (n = 2982)	P
Age (mean, years)	76 ± 5	51 ± 13	<0.001
Gender (% male)	57.9	53.8	0.015
Race (%)
Caucasian	88.3	79.5	<0.001
African Canadian	0.5	2.5	<0.001
Asian	6.8	6.2	0.47
other	4.4	11.8
Modality (% on CAPD)
initial	74.0	73.9	0.94
most recent	56.1	49.7	<0.001
Modality before PD start (%):
new to dialysis	61.9	55.7	<0.001
transfer from HD	22.9	24.0.0	0.43
failed transplant	0.3	3.6	<0.001
other/unknown	14.9	16.7
Cause of ESRD (%)
diabetes mellitus	27.6	38.4	<0.001
hypertension	30.9	10.9	<0.001
glomerulonephritis	9.8	17.8	<0.001
cystic kidney disease	3.1	6.3	<0.001
other	28.6	26.6
Diabetic	35.4	42.5	<0.001

CAPD, continuous ambulatory peritoneal dialysis; PD, peritoneal dialysis; HD, hemodialysis; ESRD, end-stage renal disease.

To determine when the relationship between age and peritonitis rate started to change, we ran an analysis for each PD start year. The year-by-year analysis revealed a loss of significance for the association between age and peritonitis after the year 2000 (data available in Appendix 1). On the basis of these data, the eras for the subsequent analyses were defined as 1996 to 2000 and 2001 to 2005.

Infectious complications of PD were divided into peritonitis and catheter infection. The peritonitis data are presented in Table 2. A total of 3058 episodes of peritonitis occurred in 1605 patients. The remaining 2642 patients did not have any peritonitis episodes. In the univariate negative binomial regression model, increasing age was associated with a higher peritonitis rate (rate ratio 1.06 per decade; 95% CI 1.02 to 1.10; P = 0. 004). This effect persisted after adjustment for gender, race, diabetes, cause of ESRD, modality before PD start, and PD modality (rate ratio 1.06 per decade; 95% CI 1.01 to 1.10; P = 0.008).

Table 2.

Association between age and peritonitis rate

Sample	Rate ratio (per decade increase in age)	95% CI	p value
Overall (n = 3058 episodes in 4247 patients)	1.06	1.01 to 1.10	0.008
1996 to 2000 (n = 1517 episodes in 1494 patients)	1.13	1.07 to 1.20	<0.001
2001 to 2005 (n = 1541 episodes in 2753 patients)	1.01	0.95 to 1.06	0.81

CI, confidence interval.

Initial screening for an era effect revealed a significant interaction between age and era (P = 0.006). We subsequently performed covariate-adjusted analyses for patients who initiated PD between 1996 and 2000 (n = 1494) and for those initiating PD between 2001 and 2005 (n = 2753). In the earlier era, the rate ratio associated with each decade increase in age was significantly higher (rate ratio 1.13; 95% CI 1.07 to 1.20; P < 0.001), whereas an association between age and peritonitis rate was not seen among patients initiating dialysis between 2001 and 2005 (rate ratio 1.01; 95% CI 0.95 to 1.06; P = 0.81).

When peritonitis rates for individual organism categories were assessed overall, increasing age was associated with a higher Gram positive peritonitis rate (rate ratio 1.05 per decade increase; 95% CI 1.00 to 1.10, P = 0.043) and a higher coagulase negative Staphylococcus (CNS) peritonitis rate (rate ratio 1.07; 95% CI 1.01 to 1.14; P = 0.032). Although age did not affect Escherichia coli peritonitis in the overall analysis, an era effect was observed, with a higher E. coli peritonitis rate with increasing age from 1996 to 2000 (rate ratio 1.37 per decade increase; 95% CI 1.07 to 1.74; P = 0.011). but no association between E. coli peritonitis and age from 2001 to 2005 (rate ratio 0.94 per decade increase; 95% CI 0.74 to 1.18; P = 0.58). There was also a trend toward a lower S. aureus peritonitis rate with increasing age in the contemporary cohort (rate ratio 0.84; 95% CI 0.68 to 1.03; P = 0.096). The associations between age and the organisms causing peritonitis by era are described in Table 3.

Table 3.

Association between age and peritonitis rate for each organism category by era

	1996 to 2000 (695 incident patients with 1517 peritonitis episodes)			2001 to 2005 (910 incident patients with 1541 peritonitis episodes)
Category	Rate ratio (per decade increase in age)	95% CI	P	Rate ratio (per decade increase in age)	95% CI	P
Gram positive	1.06	0.99 to 1.13	0.098	1.04	0.98 to 1.11	0.22
Gram negative	1.11	0.96 to 1.29	0.17	0.97	0.87 to 1.07	0.54
Culture negative	0.96	0.85 to 1.09	0.55	1.00	0.89 to 1.12	0.98
CNS	1.09	0.99 to 1.19	0.084	1.06	0.97 to 1.16	0.20
S. aureus	1.05	0.86 to 1.29	0.62	0.84	0.68 to 1.03	0.096
Streptococcus	0.90	0.78 to 1.04	0.14	1.08	0.94 to 1.24	0.26
Pseudomonas	1.05	0.97 to 1.14	0.21	1.00	0.92 to 1.10	0.93
E. coli	1.37	1.07 to 1.74	0.011	0.94	0.74 to 1.18	0.58
Yeast	1.09	0.93 to 1.28	0.30	1.05	0.93 to 1.18	0.46

CNS, coagulase negative staphylococcus; CI, confidence interval.

Peritonitis outcomes were studied in a subgroup of patients from centers that had at least 95% of outcomes reported, which included 921 episodes in 453 patients. Death resulting from peritonitis occurred with a higher frequency in patients ≥70 yr of age (4.9% versus 2.2%, P = 0.039). When age was assessed as a continuous variable, the odds ratio (OR) for peritonitis-related death was 1.37 per decade increase in age (95% CI 1.02 to 1.83; P = 0.035). There was no association between age and catheter removal after peritonitis (P = 0.31) nor between age and resolution of peritonitis (P = 0.11). There was a significant interaction between age and era for peritonitis-related death (P = 0.037). When outcome was assessed in each era, the higher frequency of peritonitis-related death with increasing age disappeared among those initiating PD between 2001 and 2005 (Table 4).

Table 4.

Association between age and peritonitis outcomes by era

	1996 to 2000 (n = 561)			2001 to 2005 (n = 360)
Outcome	OR (per decade increase in age)	95% CI	P	OR (per decade increase in age)	95% CI	P
Death	2.33	1.41 to 3.89	0.001	1.09	0.67 to 1.77	0.73
Catheter removal	1.02	0.83 to 1.26	0.84	1.00	0.80 to 1.26	0.97
Resolution	0.88	0.74 to 1.05	0.16	0.93	0.77 to 1.13	0.48

Data are reported for subgroup of patients from centers with >95% of outcomes reported. OR, odd ratio; CI, confidence interval.

With respect to catheter infection, in both negative binomial models (with and without covariates), increasing age was associated with a lower catheter infection rate (rate ratio 0.92 per decade; 95% CI 0.87 to 0.98; P = 0.004). There was no era effect for the association between age and catheter infection (P = 0.85). When catheter infection rates for individual organism categories were assessed, increasing age was associated with a lower Gram positive infection rate (rate ratio 0.93 per decade increase; 95% CI 0.88 to 0.98; P = 0.008) and a lower S. aureus infection rate (rate ratio 0.89 per decade increase; 95% CI 0.82 to 0.96; P = 0.003). All other organism infection rates, including those for Gram negative organisms, CNS, Streptococcus species, pseudomonas, E. coli and yeast, did not vary with increasing age (Table 5). Era did not influence the association between age and catheter infection organisms. There was no relationship between age and catheter infection outcomes.

Table 5.

Association between age and catheter infection rate by organism

	Rate ratio (per decade increase in age)	95% CI	P
Gram positive	0.93	0.88 to 0.98	0.008
Gram negative	1.01	0.90 to 1.14	0.85
Culture negative	1.01	0.79 to 1.28	0.96
CNS	1.01	0.90 to 1.15	0.83
S. aureus	0.89	0.82 to 0.96	0.003
Streptococcus	1.01	0.96 to 1.07	0.67
Pseudomonas	1.02	0.88 to 1.47	0.82
E. coli	1.00	0.95 to 1.07	0.90
Yeast	1.02	0.97 to 1.08	0.43

CNS, coagulase negative staphylococcus; CI, confidence interval.

Discussion

In this study, we have identified important associations between age and PD-related infectious complications, as well as the impact of the era of PD initiation on these findings. Although increasing age was associated with a higher peritonitis rate overall, this association was not present in the subgroup of patients who initiated dialysis in more recent years. There was a higher Gram positive peritonitis rate with increasing age, largely accounted for by a higher rate of CNS peritonitis. Peritonitis-related mortality was more common with increasing age among those initiating PD in 1996 to 2000 but not in 2001 to 2005. Finally, catheter infections were less common with increasing age.

The existing literature on the association between age and peritonitis rate is inconsistent (8,9,10,11,13). Several factors may be responsible for this variability. First, many of the studies that have looked at the effect of age on peritonitis have been small, single-center studies with limited statistical power. Second, different results may reflect the varying age cutoffs used to define “elderly” in the studies. Third, the era in which the patients received dialysis is quite variable, with some studies reporting on patients who were on PD in the late 1980s, and others reporting on more contemporary PD cohorts. The importance of the latter issue relates to the major advances in PD connectology (14–20) and exit site care (21–24) that occurred over this time period. Our study found that increasing age was associated with a higher peritonitis rate among patients initiating PD between 1996 and 2000, with the association disappearing among patients initiating PD in a more recent era. The lack of association between increasing age and peritonitis in recent years may reflect the fact that the “flush before fill” technique and the use of topical antibacterial agents provide an added “safety net” against contamination of the system in elderly patients who may have impaired vision or dexterity.

Surprisingly, increasing age was associated with a lower S. aureus catheter infection rate and a trend toward a lower S. aureus peritonitis rate in the more contemporary cohort. A lower incidence of S. aureus catheter infections among patients older than 60 has been previously reported (10). The basis for elderly patients having fewer S. aureus infections is unclear. Unfortunately, we do not have data on S. aureus nasal carriage, which has been linked to subsequent catheter infection and peritonitis (25–27). It is known that older patients are more compliant with some aspects of their PD (28), and it is plausible that they may also be more compliant with application of topical intranasal or exit-site ointments as prophylaxis against S. aureus infection.

In addition to the difference in S. aureus peritonitis, we found that increasing age was a risk factor for E. coli peritonitis among those initiating dialysis between 1996 and 2000. This is in keeping with a study by Kadambi et al. that reported a higher proportion of Gram negative peritonitis in older patients (9). It was thought that the higher incidence of constipation and diverticular disease in elderly patients may have predisposed these patients to Gram negative peritonitis by increasing translocation of organisms across the bowel wall. Although biologically plausible, the data to support the role of constipation and diverticulosis in development of peritonitis with enteric organisms are conflicting (29–33). Furthermore, the basis for the disappearance of the association between increasing age and E. coli peritonitis after the year 2000 is unclear. Because E. coli is not a common exit site organism, the change cannot be ascribed to introduction of exit site ointments.

The lower catheter infection rate among elderly patients has been reported previously (10). This may reflect more meticulous exit site care in this patient population, as it has been shown that patients under 55 yr of age are more likely to require retraining for breaks in exit site care protocols (34). Alternatively, it is plausible that the lower level of physical activity in this group may provide less opportunity for contamination.

On the basis of previous studies, the incidence of death due to peritonitis ranges from 2.2% to 5.9% (1,4,35–38). This is similar to the overall peritonitis-related mortality of 2.9% in our study. Interestingly, the higher risk of peritonitis-related death with increasing age disappeared in the more contemporary cohort, resulting in similar infection outcomes regardless of age among those who initiated PD after 2000. Although the basis for the improved peritonitis outcomes in elderly PD patients is unclear, one possibility is that the treatment of infection has become more standardized in recent years, and that this standardization of care may have preferentially benefited older patients, who were more susceptible to adverse outcomes with suboptimal infection management.

As with all large datasets, our study has several limitations. The data included in the database have not been validated against patient charts. The completeness of data entry varied across centers, and important variables such as markers of malnutrition and inflammation were not available. To limit errors associated with incomplete reporting of peritonitis outcome data from some centers, we limited the outcome analyses to a subset of patients from centers with complete data. Recognizing this as a limitation, we suggest these data are still of clinical importance, particularly because the number of patients and infection episodes included remain significantly higher than previous reports. Finally, we acknowledge limitations due to a lack of consensus for the definition of death attributable to peritonitis. We would, however, expect consistency in the reporting of death due to peritonitis among all ages within any given center, and as a result, believe this would not have affected our results.

In conclusion, our study demonstrates that although increasing age was associated with higher peritonitis rates among patients who initiated PD in the 1990s, age was not associated with peritonitis in a more contemporary PD cohort. Furthermore, catheter infection is less common with increasing age. Finally, no association between age and more frequent adverse outcomes of peritonitis was seen among those who initiated PD after the year 2000.

Disclosures

SJN received a 1-year educational fellowship from Baxter Healthcare in 2006. JMB has received speaker honoraria from Baxter Healthcare. KS is an employee of Baxter Healthcare. SVJ has held an investigator-driven grant from OrthoBiotec, has received speaker and consulting fees from Amgen Canada and OrthoBiotec, and has received speaker fees from Pfizer within the last 5 years.

Appendix 1

Association between age and peritonitis for each PD start year

1996 (n = 1): too few patients for analysis

1997 (n = 5): too few patients for analysis

1998 (n = 365): rate ratio 1.10 (95% CI 0.99 to 1.23, P = 0.084)

1999 (n = 539): rate ratio 1.17 (95% CI 1.06 to 1.29, P = 0.002)

2000 (n = 584): rate ratio 1.15 (95% CI 1.04 to 1.27, P = 0.006)

2001 (n = 685): rate ratio 0.97 (95% CI 0.89 to 1.06, P = 0.52)

2002 (n = 635): rate ratio 1.08 (95% CI 0.98 to 1.19, P = 0.13)

2003 (n = 646): rate ratio 1.02 (95% CI 0.90 to 1.17, P = 0.72)

2004 (n = 591): rate ratio 0.88 (95% CI 0.76 to 1.02, P = 0.84)

2005 (n = 196): rate ratio 1.03 (95% CI 0.91 to 1.17, P = 0.61)