Correspondence *University of Toronto, Suite 8N-840, 200 Elizabeth Street, Toronto, ON M5G 2C4, Canada

Email joanne.bargman@uhn.on.ca

Original article

Jaar BG et al. (2005) Comparing the risk for death with peritoneal dialysis and hemodialysis in a national cohort of patients with chronic kidney disease. Ann Intern Med 143: 174–183   PubMed

Synopsis

Background

It has not been definitively established whether dialysis modality affects survival in patients with end-stage renal disease.

Objective

To compare the risk of death between hemodialysis and peritoneal dialysis patients.

Design and intervention

Between October 1995 and June 1998, the US CHOICE study prospectively recruited incident dialysis patients aged >17 years. These patients were followed for up to 7 years. Demographic, clinical and laboratory data were recorded at baseline, and predialysis glomerular filtration rate was estimated using the MDRD EQUATION. Baseline dialysis technique was considered to be that used 4 weeks after recruitment. Conversion to another dialysis modality was defined as a technique switch if patients remained on the new modality for 30 days. Cox regression analysis was used to compare the RELATIVE RISK (RR) of death between hemodialysis and peritoneal dialysis patients. The main analysis was by intention to treat.

Outcome measure

The primary endpoint was death.

Results

Patients (n = 1,041) from 19 states were enrolled a median of 45 days after starting dialysis, and were followed up for a median of 2.4 years. The 274 patients on peritoneal dialysis at baseline were significantly younger than the 767 on hemodialysis, and had a better case-mix profile. The risk of death in peritoneal dialysis patients relative to hemodialysis patients was nonsignificantly greater in the unadjusted analysis (1.10), but increased to 1.61 (95% CONFIDENCE INTERVAL [CI] 1.13–2.30) after correction for baseline demographic, clinical and laboratory variables. When each year of follow-up was analyzed separately, the adjusted RR of death remained considerably higher in peritoneal dialysis patients in the second year (2.34; 95% CI 1.19–4.59), but not in the first year (1.39; 95% CI 0.64–3.06). When patients were divided into tertiles according to their propensity to be initiated on peritoneal dialysis, the adjusted RR of death compared with patients on hemodialysis was substantially increased only in the intermediate tertile (2.21; 95% CI 1.07–4.59). The risk of death for peritoneal dialysis patients relative to hemodialysis patients appeared to be higher in patients with cardiovascular disease than in patients without cardiovascular disease (RR 2.10, 95% CI 1.36–3.25 vs RR 0.83, 95% CI 0.38–1.81); however, no significant interaction between cardiovascular disease, age or diabetes and the RR of death for peritoneal dialysis versus hemodialysis was observed. Overall, 24.8% of peritoneal dialysis patients and 5% of hemodialysis patients switched technique. When patients were analyzed according to their current type of dialysis (disregarding any previous switches), those on peritoneal dialysis were at higher risk of death than those on hemodialysis (RR 1.41; 95% CI 1.00–1.98). When switches were counted as treatment failures, the RR of death for patients on peritoneal dialysis increased further (2.81; 95% CI 2.04–3.86).

Conclusion

The adjusted risk of death in peritoneal dialysis patients is comparable to that in hemodialysis patients during the first year of treatment, but is higher in the second year.

Commentary

Although the CHOICE study was an observational study intended to compare survival in hemodialysis and peritoneal dialysis patients, its design makes it difficult to draw conclusions about the mortality rate in these two populations. Patients were enrolled a median of 45 days after commencing dialysis; the method of patient selection is not described in the current paper, but it is clear that a much higher proportion of peritoneal dialysis patients than hemodialysis patients were enrolled. Patients who died before recruitment or who were too sick for recruitment would have been excluded. These systematic exclusions favor a positive outcome for hemodialysis patients, because the mortality rate is much higher in hemodialysis patients than in peritoneal dialysis patients in the first few months on dialysis when comorbidity is controlled for.¹ Moreover, most patients receiving hemodialysis in the CHOICE study began treatment with a dialysis catheter (66%).² Starting dialysis using a catheter, as opposed to an arteriovenous fistula, more than doubles the risk of death in the first 90 days;³ therefore, excluding the first 45 days of dialysis from the study introduces a bias toward hemodialysis.

Furthermore, the analysis adjusted for factors that were affected by the choice of dialysis modality, such as employment status and level of C-reactive protein (CRP). The ability to work is dependent on the type of dialysis used, and this adjustment might therefore have affected the outcome of the CHOICE study. CRP, measured a median of 5 months from the initiation of dialysis, was significantly higher in the hemodialysis patients than in the peritoneal dialysis patients (3.92 mg/l vs 3.18 mg/l; P <0.005). CRP levels are increased by hemodialysis and are a risk factor for death.⁴ Controlling for CRP thus statistically negates the effect of inflammation caused by the procedure. Residual renal function was determined by asking patients whether their urine volume was at least 250 ml per day, rather than by actually measuring glomerular filtration rate. Because most peritoneal dialysis centers regularly collect 24 h urine samples, peritoneal dialysis patients are much more aware of their urine output than hemodialysis patients. Based on the above definition, in the present study a higher proportion of peritoneal dialysis patients than hemodialysis patients reported residual renal function (88% vs 81%; P = 0.015). This was one of the clinical parameters controlled for in the analysis, but it is important to note that the data could have been inaccurate.

It is ironic that the name of this study is 'CHOICE', given that many of the patients enrolled were probably not given a choice of dialysis modality.⁵ Patients, if educated about their options, often have strong preferences for one technique or the other, and more than 30% of patients would not switch modality even if it increased survival by 100%.⁶ Patients should be given sufficient information on the choices available, and the decision on dialysis method should be left to them. Furthermore, not every choice in life has to be made to maximize survival, or we would probably never leave the house, cross the street, or drive on the freeway.

Acknowledgments

The synopsis was written by Chloe Harman, Associate Editor, Nature Clinical Practice.

References

1. Fenton SS et al. (1997) Hemodialysis versus peritoneal dialysis: a comparison of adjusted mortality rates. Am J Kidney Dis 30: 334–342 | PubMed | ISI | ChemPort |
2. Astor BC et al. (2005) Type of vascular access and survival among incident hemodialysis patients: the Choices for Healthy Outcomes In Caring for ESRD (CHOICE) Study. J Am Soc Nephrol 16: 1449–1455 | Article | PubMed |
3. Xue JL et al. (2003) The association of initial hemodialysis access type with mortality outcomes in elderly Medicare ESRD patients. Am J Kidney Dis 42: 1013–1019 | PubMed |
4. Korevaar JC et al. (2004) Effect of an increase in C-reactive protein level during a hemodialysis session on mortality. J Am Soc Nephrol 15: 2916–2922 | Article |
5. Bass EB et al. (2004) How strong are patients' preferences in choices between dialysis modalities and doses? Am J Kidney Dis 44: 695–705 | PubMed |
6. Rubin HR et al. (2004) Patient ratings of dialysis care with peritoneal dialysis vs hemodialysis. JAMA 291: 697–703 | Article | PubMed | ISI | ChemPort |

Contact the journal about this article

Subject areas under which this article appears: Dialysis (hemodialysis, peritoneal dialysis, continuous renal replacement)