- Email: [email protected]
Survival on hemodialysis and peritoneal dialysis over 12 years with emphasis on nutritional parameters
Survival on Hemodialysis and Peritoneal Dialysis Over 12 Years With Emphasis on Nutritional Parameters Morrell M. Avram, MD, FACP, Rajanna Sreedhara, MD, FACP, Paul Fein, MD, FACP, Kyin Kyin Oo, MD, Jyoti Chattopadhyay, PhD, and Neal Mittman, MD, FACP ● We analyzed the prognostic importance of nutritional markers and mortality data in 537 hemodialysis (HD) and 422 peritoneal dialysis (PD) patients followed for up to 12 years. Patients on HD had a 44% lower risk of mortality than did those treated with PD (P < 0.0001). The difference in mortality between the modalities was even more striking among diabetics but less striking among younger patients. Over a 12-year period, survival of dialysis patients with lower enrollment levels of albumin, creatinine, and parathyroid hormone (PTH) were significantly lower. In multivariate Cox’s proportional hazards models, serum prealbumin and enrollment PTH level of <65 pg/mL were independent predictors of mortality both in HD and PD patients. In conclusion, HD patients had higher cumulative survival than PD patients over a 12-year period. Nutritional markers at enrollment continue to be strong predictors of mortality for up to 12 years. © 2001 by the National Kidney Foundation, Inc. INDEX WORDS: Albumin; hemodialysis; nutrition; parathyroid hormone (PTH); peritoneal dialysis; prealbumin.
V
ARIOUS INVESTIGATORS have reported inconsistent results in their comparisons of outcomes of patients treated with hemodialysis (HD) and peritoneal dialysis (PD). Lowrie et al1 and Bloembergen et al2 have reported higher mortality risk in PD. On the contrary, Collins et al3 and Fenton et al4 have reported better survival in PD. Other investigators have reported identical survival in patients treated with HD and PD.5,6 Most of these studies differ in their methodology and, more importantly, had relatively short follow-up time. In the present study, we compared the mortality rates between HD and PD patients followed for up to 12 years. We and others have reported that lower levels of nutritional indices such as albumin, prealbumin, and creatinine are associated with risk of early death in patients with end-stage renal disease (ESRD).1,7-9 This association of nutritional indices and mortality is strongly predictive and consistent across studies. In addition, several surveys of patients with ESRD confirm a malnourished state. In this study, we compared nutritional indices in HD and PD patients and their predictive value on survival over 12 years. METHODS Patients Five hundred thirty-seven HD patients and 422 PD patients treated at the Long Island College Hospital’s outpatient dialysis facilities were enrolled in this study from 1987 to 1999. For the prealbumin study, 126 HD and 116 PD patients were enrolled beginning June 1991. The patients were followed up to January 1, 2000. Follow-up was censored on transplantation, transfer of dialysis modality, or
transfer to another dialysis center. On enrollment, the following demographic and clinical data were collected: age, race, etiology of ESRD, diabetic status, hypertension, and seropositivity to human immunodeficiency virus (if known).
Laboratory Data Venous blood was drawn at routine monthly visits for multiphasic biochemical screen, including albumin, creatinine, blood urea nitrogen, and total cholesterol levels. Prealbumin level was measured by rate nephelometry on a Beckman Array Protein System (Beckman Instruments Inc, Brea, CA). Serum intact PTH was measured by radioimmunoassay (Nichols Institute Diagnostics, San Juan Capistrano, CA). “Enrollment” value was the first value of a given nutritional marker for the patient obtained at or immediately after enrollment. “Mean follow-up” value was the mean of all subsequent values of the given nutritional marker for the patient.
Description of Dialysis Prescription The HD patients were treated with bicarbonate-based dialysate and volumetric ultrafiltration control. Non-reprocessed cellulose-based membranes were used for more than 98% of all treatments (predominantly cellulose acetate and cellulose triacetate). The target urea reduction ratio was ⬎60%, but subsequently increased to ⬎65%. Continuous
From the Avram Division of Nephrology, Avram Center for Kidney Diseases, Long Island College Hospital, Brooklyn, NY. Supported by National Kidney Foundation of New York/ New Jersey and Nephrology Foundation of Brooklyn. Address reprint requests to Morrell M. Avram, MD, FACP, Avram Division of Nephrology, Long Island College Hospital, 339 Hicks Street, Brooklyn, NY 11201. E-mail: [email protected] © 2001 by the National Kidney Foundation, Inc. 0272-6386/01/3701-0216$3.00/0 doi:10.1053/ajkd.2001.20754
American Journal of Kidney Diseases, Vol 37, No 1, Suppl 2 (January), 2001: pp S77-S80
S77
S78
AVRAM ET AL Table 1.
Demographics, Clinical Characteristics, and Biochemical Parameters of Hemodialysis and Peritoneal Dialysis Patients
Variable
Enrollment age (Years) Race (%) White Black Hispanic Sex (%) Male Female Diabetes (%) Enrollment albumin (g/dL) Enrollment creatinine (mg/dL) Enrollment total cholesterol (mg/dL) Enrollment prealbumin (mg/dL) Enrollment parathyroid hormone (pg/mL) Mean albumin (g/dL) Mean creatinine (mg/dL) Mean total cholesterol (mg/dL) Mean parathyroid hormone (pg/mL) Months on dialysis at enrollment
Hemodialysis
Peritoneal Dialysis
P
60 ⫾ 16
55 ⫾ 16
⬍0.0001
25 59 16
23 61 16
NS
47 53 48 3.67 ⫾ 0.49 11.1 ⫾ 4.7 180 ⫾ 53 26.5 ⫾ 6.4 355 ⫾ 450 3.66 ⫾ 0.47 10.9 ⫾ 4.0 177 ⫾ 50 364 ⫾ 436 23.2 ⫾ 39.3
48 52 42 3.41 ⫾ 0.68 10.8 ⫾ 4.5 205 ⫾ 61 35.0 ⫾ 9.8 347 ⫾ 345 3.34 ⫾ 0.64 11.1 ⫾ 4.3 206 ⫾ 58 348 ⫾ 335 3.7 ⫾ 8.4
NS 0.029 ⬍0.0001 NS ⬍0.0001 ⬍0.0001 NS ⬍0.0001 NS ⬍0.0001 NS ⬍0.0001
Values shown are mean ⫾ SD.
ambulatory peritoneal dialysis prescription was 2-L Dianeal solution (Baxter Healthcare Corporation, McGraw Park, IL), 4 exchanges/day in most patients before dialysis adequacy measurements were performed (before 1992). After 1992, PD prescription was guided to achieve a target Kt/V of 1.7, later increased to 2.1. Dextrose concentration in the dialysate varied with the need for ultrafiltration.
Statistical Analysis For selected comparisons between two group means, parametric (t-test) or non-parametric (Mann-Whitney test) tests were used. Correlations were reported as either the Pearson correlation coefficient or the Spearman rank correlation coefficient. Observed survival of HD and PD patients was computed by Kaplan Meier method. Survival was evaluated by first univariate and then by multivariate Cox’ proportional hazards model using age, race, sex, diabetes, prior months on dialysis, comorbidity and nutritional status as independent variables. Calculations were performed using SPSS for Windows 9.0.1 (SPSS Inc, Chicago, IL).
RESULTS Patient Characteristics
Table 1 describes the demographic, clinical, and biochemical characteristics of HD and PD patients. The causes of renal failure were similar in the two dialysis modalities except for a higher proportion of human immunodeficiency virus (HIV) infection (7.9% v 1.7%) in PD patients.
Dialysis Modality and Patient Survival Cumulative observed survival over 12 years was significantly higher in HD patients (P ⬍ 0.0001). Even after adjustment for demographic and clinical and biochemical factors, HD patients had better survival (P ⬍ 0.0001). Table 2 shows the Cox’s proportional hazards analysis for independent predictors of mortality in total ESRD patient population. The HD patients had a 44% lower risk of mortality than did those treated with PD (P ⬍ 0.0001). Age, diabetic status, HIV Table 2. Independent Predictors of Mortality in Hemodialysis and Peritoneal Dialysis Patients
Variables
Age (y) Diabetes (yes v no) Enrollment albumin (g/dL) Human immunodeficiency virus (yes v no) Months on dialysis Dialysis modality (hemodialysis v peritoneal dialysis)
Relative Risk
P
1.03 1.4 0.52
⬍0.0001 0.001 ⬍0.0001
1.72 1.004
0.013 0.009
0.56
⬍0.0001
Other non-significant variables included in the model were sex, race, creatinine, and total cholesterol. N ⫽ 888 (Cox’s proportional hazards model).
NUTRITION AND MORTALITY IN DIALYSIS
S79
status, and enrollment serum albumin level were also independent predictors of mortality in this model. Non-diabetic HD patients had 29% (P ⫽ 0.028) and diabetic HD patients had 56% (P ⬍ 0.0001) lower adjusted mortality risk than similar PD patients (P ⫽ 0.028). In patients ⬍59 years of age, the difference in adjusted mortality risk between HD and PD patients became marginal (RR ⫽ 0.72, P ⫽ 0.07). Over 12 years of observation, the survival of diabetic patients was significantly lower in either dialysis modality (P ⬍ 0.0001 for both).
P ⬍ 0.0001 and 597 v 280 pg/mL in PD, P ⬍ 0.0001). Both prealbumin and PTH correlated significantly with other nutritional indices (data not shown). Diabetic patients had lower levels of enrollment serum prealbumin and PTH. Survival analyses over 12 years by Kaplan Meier method showed significantly lower survival in dialysis patients with lower enrollment levels of albumin, creatinine, prealbumin, and PTH. After adjustment for confounding variables (Cox regression), the results were similar.
Biochemical Markers and Patient Survival By univariate Cox’s regression analysis, survival in HD and PD was favorably influenced by higher enrollment levels of biochemical markers of nutrition, such as albumin (RR ⫽ 0.41 in HD, P ⬍ 0.001 and RR ⫽ 0.46 in PD, P ⬍ 0.0001), prealbumin (RR ⫽ 0.92 in HD, P ⬍ 0.001 and RR ⫽ 0.96 in PD, P ⫽ 0.004), and creatinine (RR ⫽ 0.94 in HD, P ⬍ 0.001 and RR ⫽ 0.88 in PD, P ⬍ 0.0001). Patients with serum prealbumin levels ⱕ30 mg/dL had higher mortality risk (HD, RR ⫽ 1.9; PD, RR ⫽ 2.1) than those with prealbumin level ⬎30 mg/dL. In Cox’s multivariate regression model, age (RR ⫽ 1.03, P ⫽ 0.004), diabetes (RR ⫽ 1.88, P ⫽ 0.015), and lower serum prealbumin levels (⬍30 v ⬎30 mg/dL, RR ⫽ 2.06, P ⫽ 0.008) were significant independent risk factors for mortality in HD patients. In PD patients, diabetes (RR ⫽ 2.03, P ⫽ 0.03) and prealbumin levels (⬍30 v ⬎30 mg/dL, RR ⫽ 1.98, P ⫽ 0.025) were independent predictors of mortality. Serum prealbumin levels in patients who survived through the entire study period were significantly higher than in those who died before the end of the study period (29.3 v 25.3 mg/dL in HD, P ⫽ 0.006; 39.7 v 31.8 mg/dL in PD, P ⬍ 0.0001). In a separate Cox’s multivariate model, enrollment serum PTH was also a significant independent predictor of mortality in HD and PD patients. Enrollment PTH level of ⬍65 pg/mL was associated with 65% higher mortality risk in HD (P ⬍ 0.0001) and triple the mortality risk in PD (P ⬍ 0.0001) of those with PTH ⬎200 pg/mL. Enrollment serum PTH level in patients who survived during the entire study period was significantly higher than those who died before the end of the study period (480 v 316 pg/mL in HD,
We have shown that HD patients had 44% lower mortality risk than those patients treated with PD. Our results are in agreement with those of Lowrie et al1 and Bloembergen et al.2 The mortality risk for diabetic PD patients was more than twice that of diabetic HD patients (P ⬍ 0.0001) which is in agreement with findings by Held et al.10 Our study is unique in that we have followed a large number of patients in a single center for up to 12 years. The most likely explanation for excess mortality in PD may be dialysis adequacy. Many PD patients may have received inadequate dialysis before 1992, when there was no accepted target Kt/V and dialysis adequacy measures were not routinely monitored. Secondly, delivered dialysis Kt/V may not be equivalent to urea clearance obtained with residual renal function (ie, loss of residual renal Kt/V may not be made up by an equal increase in dialysis Kt/V). Thus, once the residual renal function has decreased to an insignificant level, delivered dialysis Kt/V may need to be even greater than the current target. Thirdly, decreased oral intake in PD patients due to mechanical reasons or early satiety (caloric load from glucose in dialysate) may lead to protein malnutrition with associated excess mortality risk. Other possible causes include worse lipemic profiles and excess cardiovascular morbidity, as well as higher infection risk in PD patients. Finally, a selection bias cannot be ruled out, as patients with poor hemodynamic features may have been entered into the PD program. Nutritional status of dialysis patients is one of the most important risk factors affecting survival. A lower enrollment serum albumin level in PD patients would indicate a poorer nutritional status in these patients compared with HD pa-
DISCUSSION
S80
AVRAM ET AL
tients. In addition, the decrease in albumin level during follow-up measurements in PD patients and not in HD patients would also suggest that the nutritional state deteriorated while patients were on PD (Table 1). In this study, we have shown that prealbumin is an independent predictor of mortality in HD as well as PD patients over 8 years of follow-up confirming our earlier findings over a shorter duration of follow-up.8,11 Serum PTH level was also an independent predictor of mortality in HD and PD patients, though PTH was long thought to contribute to the toxicity of uremic syndrome.12 Several animal and human studies indicate a direct relationship between higher dietary protein/phosphorus intake and higher level of serum PTH, and in this study we found that PTH level was directly correlated to dietary protein intake in PD patients (r ⫽ 0.34, P ⬍ 0.0001). We hypothesize that low PTH level in the setting of ESRD reflects a poor nutritional state and thus predicts lower survival. Recently, other investigators have confirmed our work.13,14 In conclusion, HD is associated with decreased mortality risk relative to PD, the effect being greatest in diabetic patients. A major limitation of our retrospective cohort study is that we have not adjusted for dialysis dose in comparisons between the two modalities. We also could not adjust for comorbid conditions other than diabetes and HIV infection/acquired immune deficiency syndrome (AIDS). Significant differences in the preexisting comorbid conditions may have affected patient survival comparisons. We have also confirmed our previous findings regarding the prognostic importance of enrollment nutritional markers (albumin, creatinine, prealbumin, and PTH) in patients on renal replacement therapy. REFERENCES 1. Lowrie EG, Huang WH, Lew NL: Death risk predictors among peritoneal dialysis and hemodialysis patients: A preliminary comparison. Am J Kidney Dis 26:220-228, 1995 2. Bloembergen W, Port F, Mauger E, Wolfe RA: Com-
parison of mortality between patients treated with hemodialysis and peritoneal dialysis. J Am Soc Nephrol 6:177-183, 1995 3. Collins AJ, Hao W, Xia HEJP, Ebben JP, Everson SE, Constantini EG, Ma JZ: Mortality risks of peritoneal dialysis and hemodialysis. Am J Kidney Dis 34:1065-1074, 1999 4. Fenton S, Schaubel D, Desmeules M, Morrison HI, Mao Y, Copleston P, Jeffery JR, Kjellstrand CM: Hemodialysis versus peritoneal dialysis: A comparison of adjusted mortality rates. Am J Kidney Dis 30:334-342, 1997 5. Murphy SW, Foley RN, Barrett BJ, Kent GM, Morgan J, Barre P, Campbell P, Fine A, Goldstein MB, Handa SP, Jindal KK, Levin A, Mandin H, Muirhead N, Richardson RM, Parfrey PS: Comparative mortality of hemodialysis and peritoneal dialysis in Canada. Kidney Int 57:1720-1726, 2000 6. Maiorca R, Brunori G, Zubani R, Cancarini GC, Manili L, Camerini C, Movilli E, Pola A, d’Avolio G, Gelatti U: Predictive value of dialysis adequacy and nutritional indices for mortality and morbidity in CAPD and HD patients. A longitudinal study. Nephrol Dial Transpl 10:2295-2305, 1995 7. Avram MM, Mittman N, Bonomini L, Chattopadhyay J, Fein PA: Markers for survival in dialysis: A seven year prospective study. Am J Kidney Dis 26:209-219, 1995 8. Avram MM, Goldwasser P, Erroa M, Fein PA: Predictors of survival in continuous ambulatory peritoneal dialysis patients: the importance of prealbumin and other nutritional and metabolic markers. Am J Kidney Dis 23:91-98, 1994 9. Avram MM, Sreedhara R, Avram DK, Mushnick RA, Fein PA: Enrollment parathyroid hormone level is a new marker of survival in hemodialysis and peritoneal dialysis therapy for uremia. Am J Kidney Dis 28:924-930, 1996 10. Held PJ, Port FK, Turenne MN, Gaylin DS, Hamburger RJ, Wolfe RA: Continuous ambulatory peritoneal dialysis and hemodialysis: comparison of patient mortality with adjustment of comorbid conditions. Kidney Int 45:11631169, 1994 11. Sreedhara R, Avram MM, Blanco M, Batish R, Avram MM, Mittman N: Prealbumin is the best nutritional predictor of survival in hemodialysis and peritoneal dialysis. Am J Kidney Dis. 28:937-942, 1996 12. Avram MM, Feinfeld DA, Huatuco AH: Search for the uremic toxin: Decreased motor-nerve conduction velocity and elevated parathyroid hormone in uremia. N Engl J Med 298:1000-1003, 1978 13. Akizawa T, Kinugasa E, Kurihara S, Tsukamoto Y, Fukagawa M, Kawaguchi Y, Kurokawa K: Parathyroid hormone deficiency is an indicator of poor nutritional state and prognosis in dialysis patients. J Am Soc Nephrol 10:615A615A, 1999 (abstr) 14. Heaf JG, Lokkegard H: Parathyroid hormone during maintenance dialysis: Influence of low calcium dialysate, plasma albumin and age. J Nephrol 11:203-210, 1998
V
ARIOUS INVESTIGATORS have reported inconsistent results in their comparisons of outcomes of patients treated with hemodialysis (HD) and peritoneal dialysis (PD). Lowrie et al1 and Bloembergen et al2 have reported higher mortality risk in PD. On the contrary, Collins et al3 and Fenton et al4 have reported better survival in PD. Other investigators have reported identical survival in patients treated with HD and PD.5,6 Most of these studies differ in their methodology and, more importantly, had relatively short follow-up time. In the present study, we compared the mortality rates between HD and PD patients followed for up to 12 years. We and others have reported that lower levels of nutritional indices such as albumin, prealbumin, and creatinine are associated with risk of early death in patients with end-stage renal disease (ESRD).1,7-9 This association of nutritional indices and mortality is strongly predictive and consistent across studies. In addition, several surveys of patients with ESRD confirm a malnourished state. In this study, we compared nutritional indices in HD and PD patients and their predictive value on survival over 12 years. METHODS Patients Five hundred thirty-seven HD patients and 422 PD patients treated at the Long Island College Hospital’s outpatient dialysis facilities were enrolled in this study from 1987 to 1999. For the prealbumin study, 126 HD and 116 PD patients were enrolled beginning June 1991. The patients were followed up to January 1, 2000. Follow-up was censored on transplantation, transfer of dialysis modality, or
transfer to another dialysis center. On enrollment, the following demographic and clinical data were collected: age, race, etiology of ESRD, diabetic status, hypertension, and seropositivity to human immunodeficiency virus (if known).
Laboratory Data Venous blood was drawn at routine monthly visits for multiphasic biochemical screen, including albumin, creatinine, blood urea nitrogen, and total cholesterol levels. Prealbumin level was measured by rate nephelometry on a Beckman Array Protein System (Beckman Instruments Inc, Brea, CA). Serum intact PTH was measured by radioimmunoassay (Nichols Institute Diagnostics, San Juan Capistrano, CA). “Enrollment” value was the first value of a given nutritional marker for the patient obtained at or immediately after enrollment. “Mean follow-up” value was the mean of all subsequent values of the given nutritional marker for the patient.
Description of Dialysis Prescription The HD patients were treated with bicarbonate-based dialysate and volumetric ultrafiltration control. Non-reprocessed cellulose-based membranes were used for more than 98% of all treatments (predominantly cellulose acetate and cellulose triacetate). The target urea reduction ratio was ⬎60%, but subsequently increased to ⬎65%. Continuous
From the Avram Division of Nephrology, Avram Center for Kidney Diseases, Long Island College Hospital, Brooklyn, NY. Supported by National Kidney Foundation of New York/ New Jersey and Nephrology Foundation of Brooklyn. Address reprint requests to Morrell M. Avram, MD, FACP, Avram Division of Nephrology, Long Island College Hospital, 339 Hicks Street, Brooklyn, NY 11201. E-mail: [email protected] © 2001 by the National Kidney Foundation, Inc. 0272-6386/01/3701-0216$3.00/0 doi:10.1053/ajkd.2001.20754
American Journal of Kidney Diseases, Vol 37, No 1, Suppl 2 (January), 2001: pp S77-S80
S77
S78
AVRAM ET AL Table 1.
Demographics, Clinical Characteristics, and Biochemical Parameters of Hemodialysis and Peritoneal Dialysis Patients
Variable
Enrollment age (Years) Race (%) White Black Hispanic Sex (%) Male Female Diabetes (%) Enrollment albumin (g/dL) Enrollment creatinine (mg/dL) Enrollment total cholesterol (mg/dL) Enrollment prealbumin (mg/dL) Enrollment parathyroid hormone (pg/mL) Mean albumin (g/dL) Mean creatinine (mg/dL) Mean total cholesterol (mg/dL) Mean parathyroid hormone (pg/mL) Months on dialysis at enrollment
Hemodialysis
Peritoneal Dialysis
P
60 ⫾ 16
55 ⫾ 16
⬍0.0001
25 59 16
23 61 16
NS
47 53 48 3.67 ⫾ 0.49 11.1 ⫾ 4.7 180 ⫾ 53 26.5 ⫾ 6.4 355 ⫾ 450 3.66 ⫾ 0.47 10.9 ⫾ 4.0 177 ⫾ 50 364 ⫾ 436 23.2 ⫾ 39.3
48 52 42 3.41 ⫾ 0.68 10.8 ⫾ 4.5 205 ⫾ 61 35.0 ⫾ 9.8 347 ⫾ 345 3.34 ⫾ 0.64 11.1 ⫾ 4.3 206 ⫾ 58 348 ⫾ 335 3.7 ⫾ 8.4
NS 0.029 ⬍0.0001 NS ⬍0.0001 ⬍0.0001 NS ⬍0.0001 NS ⬍0.0001 NS ⬍0.0001
Values shown are mean ⫾ SD.
ambulatory peritoneal dialysis prescription was 2-L Dianeal solution (Baxter Healthcare Corporation, McGraw Park, IL), 4 exchanges/day in most patients before dialysis adequacy measurements were performed (before 1992). After 1992, PD prescription was guided to achieve a target Kt/V of 1.7, later increased to 2.1. Dextrose concentration in the dialysate varied with the need for ultrafiltration.
Statistical Analysis For selected comparisons between two group means, parametric (t-test) or non-parametric (Mann-Whitney test) tests were used. Correlations were reported as either the Pearson correlation coefficient or the Spearman rank correlation coefficient. Observed survival of HD and PD patients was computed by Kaplan Meier method. Survival was evaluated by first univariate and then by multivariate Cox’ proportional hazards model using age, race, sex, diabetes, prior months on dialysis, comorbidity and nutritional status as independent variables. Calculations were performed using SPSS for Windows 9.0.1 (SPSS Inc, Chicago, IL).
RESULTS Patient Characteristics
Table 1 describes the demographic, clinical, and biochemical characteristics of HD and PD patients. The causes of renal failure were similar in the two dialysis modalities except for a higher proportion of human immunodeficiency virus (HIV) infection (7.9% v 1.7%) in PD patients.
Dialysis Modality and Patient Survival Cumulative observed survival over 12 years was significantly higher in HD patients (P ⬍ 0.0001). Even after adjustment for demographic and clinical and biochemical factors, HD patients had better survival (P ⬍ 0.0001). Table 2 shows the Cox’s proportional hazards analysis for independent predictors of mortality in total ESRD patient population. The HD patients had a 44% lower risk of mortality than did those treated with PD (P ⬍ 0.0001). Age, diabetic status, HIV Table 2. Independent Predictors of Mortality in Hemodialysis and Peritoneal Dialysis Patients
Variables
Age (y) Diabetes (yes v no) Enrollment albumin (g/dL) Human immunodeficiency virus (yes v no) Months on dialysis Dialysis modality (hemodialysis v peritoneal dialysis)
Relative Risk
P
1.03 1.4 0.52
⬍0.0001 0.001 ⬍0.0001
1.72 1.004
0.013 0.009
0.56
⬍0.0001
Other non-significant variables included in the model were sex, race, creatinine, and total cholesterol. N ⫽ 888 (Cox’s proportional hazards model).
NUTRITION AND MORTALITY IN DIALYSIS
S79
status, and enrollment serum albumin level were also independent predictors of mortality in this model. Non-diabetic HD patients had 29% (P ⫽ 0.028) and diabetic HD patients had 56% (P ⬍ 0.0001) lower adjusted mortality risk than similar PD patients (P ⫽ 0.028). In patients ⬍59 years of age, the difference in adjusted mortality risk between HD and PD patients became marginal (RR ⫽ 0.72, P ⫽ 0.07). Over 12 years of observation, the survival of diabetic patients was significantly lower in either dialysis modality (P ⬍ 0.0001 for both).
P ⬍ 0.0001 and 597 v 280 pg/mL in PD, P ⬍ 0.0001). Both prealbumin and PTH correlated significantly with other nutritional indices (data not shown). Diabetic patients had lower levels of enrollment serum prealbumin and PTH. Survival analyses over 12 years by Kaplan Meier method showed significantly lower survival in dialysis patients with lower enrollment levels of albumin, creatinine, prealbumin, and PTH. After adjustment for confounding variables (Cox regression), the results were similar.
Biochemical Markers and Patient Survival By univariate Cox’s regression analysis, survival in HD and PD was favorably influenced by higher enrollment levels of biochemical markers of nutrition, such as albumin (RR ⫽ 0.41 in HD, P ⬍ 0.001 and RR ⫽ 0.46 in PD, P ⬍ 0.0001), prealbumin (RR ⫽ 0.92 in HD, P ⬍ 0.001 and RR ⫽ 0.96 in PD, P ⫽ 0.004), and creatinine (RR ⫽ 0.94 in HD, P ⬍ 0.001 and RR ⫽ 0.88 in PD, P ⬍ 0.0001). Patients with serum prealbumin levels ⱕ30 mg/dL had higher mortality risk (HD, RR ⫽ 1.9; PD, RR ⫽ 2.1) than those with prealbumin level ⬎30 mg/dL. In Cox’s multivariate regression model, age (RR ⫽ 1.03, P ⫽ 0.004), diabetes (RR ⫽ 1.88, P ⫽ 0.015), and lower serum prealbumin levels (⬍30 v ⬎30 mg/dL, RR ⫽ 2.06, P ⫽ 0.008) were significant independent risk factors for mortality in HD patients. In PD patients, diabetes (RR ⫽ 2.03, P ⫽ 0.03) and prealbumin levels (⬍30 v ⬎30 mg/dL, RR ⫽ 1.98, P ⫽ 0.025) were independent predictors of mortality. Serum prealbumin levels in patients who survived through the entire study period were significantly higher than in those who died before the end of the study period (29.3 v 25.3 mg/dL in HD, P ⫽ 0.006; 39.7 v 31.8 mg/dL in PD, P ⬍ 0.0001). In a separate Cox’s multivariate model, enrollment serum PTH was also a significant independent predictor of mortality in HD and PD patients. Enrollment PTH level of ⬍65 pg/mL was associated with 65% higher mortality risk in HD (P ⬍ 0.0001) and triple the mortality risk in PD (P ⬍ 0.0001) of those with PTH ⬎200 pg/mL. Enrollment serum PTH level in patients who survived during the entire study period was significantly higher than those who died before the end of the study period (480 v 316 pg/mL in HD,
We have shown that HD patients had 44% lower mortality risk than those patients treated with PD. Our results are in agreement with those of Lowrie et al1 and Bloembergen et al.2 The mortality risk for diabetic PD patients was more than twice that of diabetic HD patients (P ⬍ 0.0001) which is in agreement with findings by Held et al.10 Our study is unique in that we have followed a large number of patients in a single center for up to 12 years. The most likely explanation for excess mortality in PD may be dialysis adequacy. Many PD patients may have received inadequate dialysis before 1992, when there was no accepted target Kt/V and dialysis adequacy measures were not routinely monitored. Secondly, delivered dialysis Kt/V may not be equivalent to urea clearance obtained with residual renal function (ie, loss of residual renal Kt/V may not be made up by an equal increase in dialysis Kt/V). Thus, once the residual renal function has decreased to an insignificant level, delivered dialysis Kt/V may need to be even greater than the current target. Thirdly, decreased oral intake in PD patients due to mechanical reasons or early satiety (caloric load from glucose in dialysate) may lead to protein malnutrition with associated excess mortality risk. Other possible causes include worse lipemic profiles and excess cardiovascular morbidity, as well as higher infection risk in PD patients. Finally, a selection bias cannot be ruled out, as patients with poor hemodynamic features may have been entered into the PD program. Nutritional status of dialysis patients is one of the most important risk factors affecting survival. A lower enrollment serum albumin level in PD patients would indicate a poorer nutritional status in these patients compared with HD pa-
DISCUSSION
S80
AVRAM ET AL
tients. In addition, the decrease in albumin level during follow-up measurements in PD patients and not in HD patients would also suggest that the nutritional state deteriorated while patients were on PD (Table 1). In this study, we have shown that prealbumin is an independent predictor of mortality in HD as well as PD patients over 8 years of follow-up confirming our earlier findings over a shorter duration of follow-up.8,11 Serum PTH level was also an independent predictor of mortality in HD and PD patients, though PTH was long thought to contribute to the toxicity of uremic syndrome.12 Several animal and human studies indicate a direct relationship between higher dietary protein/phosphorus intake and higher level of serum PTH, and in this study we found that PTH level was directly correlated to dietary protein intake in PD patients (r ⫽ 0.34, P ⬍ 0.0001). We hypothesize that low PTH level in the setting of ESRD reflects a poor nutritional state and thus predicts lower survival. Recently, other investigators have confirmed our work.13,14 In conclusion, HD is associated with decreased mortality risk relative to PD, the effect being greatest in diabetic patients. A major limitation of our retrospective cohort study is that we have not adjusted for dialysis dose in comparisons between the two modalities. We also could not adjust for comorbid conditions other than diabetes and HIV infection/acquired immune deficiency syndrome (AIDS). Significant differences in the preexisting comorbid conditions may have affected patient survival comparisons. We have also confirmed our previous findings regarding the prognostic importance of enrollment nutritional markers (albumin, creatinine, prealbumin, and PTH) in patients on renal replacement therapy. REFERENCES 1. Lowrie EG, Huang WH, Lew NL: Death risk predictors among peritoneal dialysis and hemodialysis patients: A preliminary comparison. Am J Kidney Dis 26:220-228, 1995 2. Bloembergen W, Port F, Mauger E, Wolfe RA: Com-
parison of mortality between patients treated with hemodialysis and peritoneal dialysis. J Am Soc Nephrol 6:177-183, 1995 3. Collins AJ, Hao W, Xia HEJP, Ebben JP, Everson SE, Constantini EG, Ma JZ: Mortality risks of peritoneal dialysis and hemodialysis. Am J Kidney Dis 34:1065-1074, 1999 4. Fenton S, Schaubel D, Desmeules M, Morrison HI, Mao Y, Copleston P, Jeffery JR, Kjellstrand CM: Hemodialysis versus peritoneal dialysis: A comparison of adjusted mortality rates. Am J Kidney Dis 30:334-342, 1997 5. Murphy SW, Foley RN, Barrett BJ, Kent GM, Morgan J, Barre P, Campbell P, Fine A, Goldstein MB, Handa SP, Jindal KK, Levin A, Mandin H, Muirhead N, Richardson RM, Parfrey PS: Comparative mortality of hemodialysis and peritoneal dialysis in Canada. Kidney Int 57:1720-1726, 2000 6. Maiorca R, Brunori G, Zubani R, Cancarini GC, Manili L, Camerini C, Movilli E, Pola A, d’Avolio G, Gelatti U: Predictive value of dialysis adequacy and nutritional indices for mortality and morbidity in CAPD and HD patients. A longitudinal study. Nephrol Dial Transpl 10:2295-2305, 1995 7. Avram MM, Mittman N, Bonomini L, Chattopadhyay J, Fein PA: Markers for survival in dialysis: A seven year prospective study. Am J Kidney Dis 26:209-219, 1995 8. Avram MM, Goldwasser P, Erroa M, Fein PA: Predictors of survival in continuous ambulatory peritoneal dialysis patients: the importance of prealbumin and other nutritional and metabolic markers. Am J Kidney Dis 23:91-98, 1994 9. Avram MM, Sreedhara R, Avram DK, Mushnick RA, Fein PA: Enrollment parathyroid hormone level is a new marker of survival in hemodialysis and peritoneal dialysis therapy for uremia. Am J Kidney Dis 28:924-930, 1996 10. Held PJ, Port FK, Turenne MN, Gaylin DS, Hamburger RJ, Wolfe RA: Continuous ambulatory peritoneal dialysis and hemodialysis: comparison of patient mortality with adjustment of comorbid conditions. Kidney Int 45:11631169, 1994 11. Sreedhara R, Avram MM, Blanco M, Batish R, Avram MM, Mittman N: Prealbumin is the best nutritional predictor of survival in hemodialysis and peritoneal dialysis. Am J Kidney Dis. 28:937-942, 1996 12. Avram MM, Feinfeld DA, Huatuco AH: Search for the uremic toxin: Decreased motor-nerve conduction velocity and elevated parathyroid hormone in uremia. N Engl J Med 298:1000-1003, 1978 13. Akizawa T, Kinugasa E, Kurihara S, Tsukamoto Y, Fukagawa M, Kawaguchi Y, Kurokawa K: Parathyroid hormone deficiency is an indicator of poor nutritional state and prognosis in dialysis patients. J Am Soc Nephrol 10:615A615A, 1999 (abstr) 14. Heaf JG, Lokkegard H: Parathyroid hormone during maintenance dialysis: Influence of low calcium dialysate, plasma albumin and age. J Nephrol 11:203-210, 1998