Home About us Current issue Ahead of Print Back issues Submission Instructions Advertise Contact Login   

Search Article 
Advanced search 
Saudi Journal of Kidney Diseases and Transplantation
Users online: 1056 Home Bookmark this page Print this page Email this page Small font sizeDefault font size Increase font size 

ORIGINAL ARTICLE Table of Contents   
Year : 2003  |  Volume : 14  |  Issue : 1  |  Page : 23-29
The Predictors of Early Mortality in Patients Starting Chronic Hemodialysis

1 Service de Néphrologie et Hémodialyse, Centre Hospitalo-universitaire Fattouma, Bourguiba, Monastir, Tunisia
2 Département de Statistique Médicale, Faculté de Médecine de Monastir, Tunisia

Click here for correspondence address and email


To evaluate the predictors of early mortality in patients on chronic hemodialysis, we reviewed the records of 192 patients starting chronic hemodialysis at our centre between January 1996 and September 1999. The overall incident mortality within 90 days was 32 (16.7%) patients. The cardiovascular causes accounted for 50% of all the causes of mortality. By using multivariate stepwise logistic regression analysis, early mortality rate was not significantly increased in the comparison of age or gender groups but increased in patients with diabetes mellitus, as well as those with reduced dialysis frequency. The most powerful predictor of survival was serum albumin level of less than 30 g/l. Thus, the survival rates in patients with serum albumin less than 30 g/l and those with serum albumin equal to or greater than 30 g/l were 67.8% and 90.2%, respectively, (p<0.001). The odds ratio was 4.68. We conclude that these findings suggest that the important predictors of early mortality in the first 90 days of starting hemodialysis include the presence of diabetes mellitus, the decreased frequency of dialysis sessions and the presence of low serum albumin. The low serum albumin below 30 g/l was the strongest predictor of early mortality.

Keywords: Renal replacement therapy, Serum albumin, Early mortality, Prognosis, Survival.

How to cite this article:
Gmar-Bouraoui S, Skhiri H, Achour A, Frih A, Dhia N B, Hammami S, El May M. The Predictors of Early Mortality in Patients Starting Chronic Hemodialysis. Saudi J Kidney Dis Transpl 2003;14:23-9

How to cite this URL:
Gmar-Bouraoui S, Skhiri H, Achour A, Frih A, Dhia N B, Hammami S, El May M. The Predictors of Early Mortality in Patients Starting Chronic Hemodialysis. Saudi J Kidney Dis Transpl [serial online] 2003 [cited 2022 Aug 7];14:23-9. Available from: https://www.sjkdt.org/text.asp?2003/14/1/23/33084

   Introduction Top

Over the years selection criteria for acceptance on renal replacement therapy (RRT) became more liberal leading to an increase in the number of patients receiving this modality of treatment all over the world. [1] This expansion was associated with increased proportion of elderly patients and those with co-morbidity.

Previous studies showed that early mortality occurring in the first 90 days after starting RRT represented a high percentage of the first-year mortality rate, (ranged from 12.6% [2] to 32% [3],[4] ) . Short-term prognosis studies had variable results due to different methodo­logies. [5] The Canadian studies considered the early period as the first six months after starting RRT, [6],[7] while others excluded mortality that occurred in the first month. [8] The United States Renal Data System (USRDS) registry excluded all patients who did not survive beyond 90 days and only mortality that occurred in patients aged more than 65 years were taken into account in the statistical analysis. [9] The European Renal Association (ERA-EDTA) registry included mortality within 90-days but suffers from incomplete data particularly since 1992 [10] and early mortality was probably undere­stimated in this registry. [2]

Hypoalbuminemia has been demonstrated by most studies to be an adverse prognostic indicator in the long-term survival on RRT. [11],[12],[13]

However, an inverse correlation with early mortality was found by some studies [2],[6] but not by others. [7]

In this study we attempt to evaluate the predictive value of the serum albumin of early mortality in patients on chronic hemo­dialysis for less than 90 days.

   Patients and Methods Top

We reviewed the records of 345 new patients started on renal replacement therapy at the University Hospital Fattouma Bourguiba between January 1st, 1996 and September 30th, 1999.

We excluded patients with acute renal failure, rapidly progressive renal failure, acute on chronic renal failure, chronic renal failure with recovering residual renal function after a short period of RRT, patients with extrarenal diseases conferring poor prognosis independent of dialysis effect, such as neoplasia (primary or metastatic) and myeloma, and those with incomplete data especially serum albumin. There were 192 patients for the evaluation of mortality in the first 90 days of dialysis.

The reviewed clinical data included: age, gender, body mass index (BMI), cause of renal disease, comorbidity, first mode of hemo­dialysis (emergency or programmed dialysis), dialysis frequency (two times or three times a week), mortality within the first 90 days, and causes and duration to mortality.

The laboratory parameters were recorded at starting hemodialysis and included: blood urea nitrogen (BUN), creatinine clearance (Clcr) using Cockroft-Gault formula, total serum cholesterol, triglycerides, serum CO2, potassium, calcium, phosphorus, alkaline phosphatase, as well as hemoglobin level, leucocytes count and urinalysis.

   Statistical Analysis Top

In order to determine which parameters were associated with early mortality, patients who died within the first 90 days were compared to those who survived. We used BMDP computer software for the statistical analysis. Univariate statistical tests using chi-square for binary variables and t-test for con­tinuous variables were performed. Variables statistically associated with early mortality were introduced into a multivariate stepwise logistic regression with p<0.25 to enter. Risk coefficients for the best predictor parameters intercepted by this model were noted.

We also performed multivariate survival analysis using the Cox-proportional hazards method in which time to mortality was the dependent variable. When a significant predi­ctor was selected, the difference in survival was tested according to the presence of this parameter. The statistical significance was set at p < 0.05.

   Results Top

The Clinical parameters of the study patients are summarized in [Table - 1]. The average age was 48.6 ± 16.9 years. Out of 192 available for evaluation, 32 (16.7%) died in the first 90 days after starting chronic hemodialysis.

Cardiovascular causes accounted for 50% of the reported causes of mortality and included pericarditis, severe cardiac failure, myocardial infarction and cerebrovascular accident [Table - 2].

Age was not different between the deceased and survivors; the mean age was 49.8 ± 17.4 and 48.4 ± 17 years, respectively. Furthermore, there was no significant difference in the mortality between patients aged less than 65 and those aged 65 years or more (16.9% and 15.6%, respectively). There was also no significant difference of early mortality between males and females, 20.5% Vs 10.7%, respectively. Other variables that were not associated with early mortality included: BMI, regular nephrological follow-up, first dialysis (emergency or not), ischemic heart disease, cardiac failure, cholesterol level, Clcr, BUN, use of temporary access, systemic sepsis, coma and ventilator dependency.

[Table - 3], shows several clinical and bio­logical parameters that were associated with poor prognosis in the short term using the univariate analysis. We found that patients with normal or low blood pressure had worse prognosis compared to those with chronic hypertension; rate of early mortality was 23.9% Vs 12.8%, respectively, (p<0.05). The presence of diabetes mellitus or amyloidosis was associated with high rate of early mortality, 29.7% and 33.3%, respectively. The patients with urinary tract infection had significantly higher early mortality rate than those without infection 29.7% Vs 13.5% respectively. The patients treated on hemodia­lysis two times a week had significantly higher mortality than those treated three times a week, 24.6% Vs 13.3%, respectively. Hypo­albuminemia less than 30 g/l was associated with poorer prognosis than the patients with serum albumin concentration of 30 g/l or more, 32.2% vs 9.8%, respectively (p<0.0001).

Other biochemical parameters were signi­ficantly different between the deceased and surviving groups including triglycerides, hemoglobin and elevated leukocyte count.

[Table - 4] shows the significant results of the multivariate analysis of the clinical and bio­chemical parameters that correlated with early mortality. Hypoalbuminemia had the strongest predictive value of early mortality with an odds ratio of 4.68, followed by diabetes mellitus, decreased frequency of dialysis, and low hemoglobin level.

The Cox-proportional hazard analysis confirmed the results of the logistic regres­sion, but with a higher degree of signifi­cance for diabetes mellitus, treatment frequency and hypoalbuminemia. However, the low hemoglobin which was the fourth predicting factor in the logistic regression did not reach the level of statistical signi­ficance in the Cox-proportional hazards method.

   Discussion Top

This study found an early mortality rate of 16.7% in our population of chronic dialysis, which was higher than the rate reported elsewhere. [10] The estimated early mortality rate varied in the medical literature from 11% to 26%. [2],[3],[5],[7],[10],[14] This variation between centers is due to different methodologies used in the collection and analysis of data.

There was no influence of advanced age on early mortality in our study compared to other studies. [2],[7],[10] This may be due to our inclusion criteria that excluded patients with neoplasia and myeloma, which are potentially of poor prognosis at short term and occur mostly in the older patients.

Our study showed that early mortality was not statistically different between males and females. Some studies reported lower risk of mortality in females in a one-year survival analysis. [12]

The cardiovascular causes and withdrawal from dialysis were the most common causes of early mortality particularly in diabetic and elderly patients. [12],[15],[16] Though the cardiovascular comorbidity was the most common cause of death in our study but there were no withdrawals from dialysis.

Early nephrological referral seemed to be beneficial in the short-term prognosis in patients starting RRT. [15] Early mortality was more frequent in the group of late referral. [17] There was no correlation between early mortality and regular nephrological follow­up in our study. Nevertheless we did not calculate the follow-up period in order to evaluate the effect of early nephrological follow-up on the outcome.

In our study, dysrhythmia was associated with poor prognosis, while ischemic heart disease and cardiac failure were not, mostly because patients with severe cardiac failure were admitted on peritoneal dialysis in our centre, and patients treated by this kind of RRT were excluded from the study. Further­more, we did not find coma and ventilatory dependence associated with early mortality as in previous studies. [7]

One of the major prognostic indicators in our study was the primary renal disease. Diabetes mellitus and amyloidosis were associated with the highest mortality rates. Diabetes mellitus is well known in long­term survival studies to carry poor progno­sis. [3],[16],[18] It has also been suggested as a risk factor associated with adverse prognosis in a large scale study [10] but was not in another. [7]

Mortality in amyloidosis, may be due to Addisonian crisis secondary to adrenocortical insufficiency related to amyloid deposits in the adrenal glands. [19] Detecting this disorder systematically at starting chronic hemodialysis may be indicated. In this study, glomerulo­nephritis and polycystic kidney disease seemed to be associated with low risk of mortality, a finding that was noted in many other series. [9],[10],[12]

Chronic hemodialysis is associated with a typical atherogenic profile consisting mainly of hypertriglyceridemia with increased very low-density lipoprotein (VLDL) and low­ density lipoprotein (LDL). [20],[21] This explains why cardiovascular etiolgies remained the most frequent cause of mortality in patients on RRT over the years. [22] In our study, the univariate analysis showed that hyper­triglyceridemia was associated with excess of mortality.

Low systolic blood pressure as a marker of severe cardiac failure has been reported as an adverse prognostic indicator in acute renal failure. [23] This was also demonstrated in a prospective study in the chronic dialysis population. [24]

Up to 29.7% of our patients were treated twice a week. This group was at higher risk of early mortality. The adverse effect of shorter treatment time on mortality was also demonstrated in long-term survival analysis. [12] Low delivered dialysis dose is well known to increase mortality. [25]

One of the powerful independent predictors of mortality in patients on RRT is hypo­albuminemia; an indicator of malnut­ration [11],[12],[26],[27] serum albumin was found to be a predictive factor of hospitalization and mortality. [28],[29] In our study, patients with low serum albumin had a high mortality rate. However, the low serum albumin did not correlate with the adequacy of dialysis estimated by the urea reduction rate. [27],[30]

Furthermore, C-reactive protein and serum amyloid A were suggested as determinants of serum albumin concentration in hemo­dialysis patients. [30] Hypoalbuminemia was suggested as a component of the acute­ phase response. [31]

We conclude that our study findings suggest that the important predictors of early mortality in the first 90 days of start on hemodialysis include the presence of diabetes mellitus, the decreased frequency of dialysis sessions and the presence of hypoalbuminemia. The low serum albumin below 30 g/l was the strongest predictor of early mortality.

   References Top

1.Berthoux F, Jones EH, Gellert R, et al. Epidemiological data of treated end-stage renal failure in the European Union (EU) during the year 1995: report of the European Renal Association Registry and the National Registries. Nephrol Dial Transplant 1999; 14:2332-42.  Back to cited text no. 1    
2.Khan IH, Catto GR, Edward N, MacLeod AM. Mortality during the first 90 days of dialysis: a case control study. Am J Kidney Dis 1995;25:276-80.  Back to cited text no. 2  [PUBMED]  
3.Wright LF. Survival in patients with end­stage renal disease. Am J Kidney Dis 1991;17:25-8.  Back to cited text no. 3  [PUBMED]  
4.Soucie JM and McClellan WM. Early mortality in dialysis patients: risk factors and impact on incidence and mortality rates. J Am Soc Nephrol 1996;7:2169-75.  Back to cited text no. 4    
5.Khan IH, Campbell MK, Cantarovich D, et al. Survival on renal replacement therapy in Europe: is there a center effect'? Nephrol Dial Transplant 1996;11:300-7.  Back to cited text no. 5  [PUBMED]  [FULLTEXT]
6.Barett BJ, Parfrey PS, Morgan J, et al. Prediction of early mortality in end-stage renal disease patients starting dialysis. Am J Kidney Dis 1997;29:214-22.  Back to cited text no. 6    
7.Foley RN, Parfrey PS, Hefferton D, Singh I, Simms A, Barrett BJ. Advance prediction of early mortality in patients starting maintenance dialysis. Am J Kidney Dis 1994;23:836-45.  Back to cited text no. 7  [PUBMED]  
8.Churchill DN, Taylor DW, Cook RJ, et al. Canadian hemodialysis morbidity study. Am J Kidney Dis 1992;19:214-34.  Back to cited text no. 8  [PUBMED]  
9.United States Renal Data System: USRDS annual data report. Patient mortality and survival. Am J Kidney Dis 1999;34(Suppl 1): S74-S86.  Back to cited text no. 9    
10.Tsakiris D, Jones EH, Briggs JD, et al. Mortalities within 90 days from starting renal replacement therapy in the ERA­EDTA Registry between 1990 and 1992. Nephrol Dial Transplant 1999;14:2343-50.  Back to cited text no. 10  [PUBMED]  [FULLTEXT]
11.Iseki K, Kawazoe N, Fukiyama K. Serum albumin is a strong predictor of mortality in chronic dialysis patients. Kidney Int 1993;44:115-9.  Back to cited text no. 11  [PUBMED]  
12.Lowrie EG, Lew NL. Mortality risk in hemodialysis patients: the predictive value of commonly measured variables and an evaluation of mortality rate differences between facilities. Am J kidney Dis 1990; 15:458-82.  Back to cited text no. 12  [PUBMED]  
13.Spiegel DM, Breyer JA. Serum albumin: a predictor of long term-outcome in perito­neal dialysis patients. Am J Kidney Dis 1994;23:283-5.  Back to cited text no. 13    
14.United States Renal Data System: USRDS annual data report. Survival probabilities and causes of mortality. Am J Kidney Dis 1991;18(Suppl 2):S49-S60.  Back to cited text no. 14    
15.Innes A, Rowe PA, Burden RP, Morgan AG. Early mortality on renal replacement therapy: the need for early nephrological referral. Nephrol Dial Transplant 1992; 7:467-71.  Back to cited text no. 15  [PUBMED]  [FULLTEXT]
16.United States Renal Data System: USRDS annual data report. Causes of mortality. Am J Kidney Dis 1997;30(Suppl 1):S107-S17.  Back to cited text no. 16    
17.Jungers P, Skhiri H, Zingraff J, et al. Benefits of early nephrological manage-ment of chronic renal failure. Presse Med 1997;26:2-5.  Back to cited text no. 17  [PUBMED]  
18.Lowrie EG, Lew NL, Huang WH. Race and diabetes as mortality risk predictors in hemodialysis patients. Kidney Int Suppl 1992;38:S22-S31.  Back to cited text no. 18  [PUBMED]  
19.Danby P, Harris KP, Williams B, Feehally J, Walls J. Adrenal dysfunction in patients with renal amyloid. Q J Med 1990;76:915-22.  Back to cited text no. 19  [PUBMED]  [FULLTEXT]
20.Koniger M, Quaschning T, Wanner C, Schollmeyer P, Kramer-Guth A. Abnor­malities in lipoprotein metabolism in hemodialysis patients. Kidney Int Suppl 1999;71:S248-50.  Back to cited text no. 20    
21.Samuelsson O, Attman PO, Knight-Gibson C et al. Lipoprotein abnormalities without hyperlipidemia in moderate renal insuffici­ency. Nephrol Dial Transplant 1994; 9:1580-5.  Back to cited text no. 21    
22.Herzog CA. Acute myocardial infarction in patients with end-stage renal disease. Kidney Int Suppl 1999;71:S130-3.  Back to cited text no. 22  [PUBMED]  
23.Druml W, Lax F, Grimm G, Schneeweiss B, Lenz K, Laggner AN. Acute renal failure in the elderly 1975-1990. Clin Nephrol 1994;41:342-9.  Back to cited text no. 23  [PUBMED]  
24.Foley RN, Parfrey PS, Harnett JD, Kent GM, Murray DC, Barre PE. Impact of hyper-tension on cardiomyopathy, morbidity and mortality in end-stage renal disease. Kidney Int 1996; 49:1379-85.  Back to cited text no. 24  [PUBMED]  
25.Held PJ, Port FK, Wolfe RA, et al. The dose of hemodialysis and patient mortality. Kidney Int 1996;50:550-6.  Back to cited text no. 25    
26.Goldwasser P, Mittman N, Antignani A, et al. Predictors of mortality in hemodialysis patients. J Am Soc Nephrol 1993;3:1613-22.  Back to cited text no. 26  [PUBMED]  
27.Owen WF Jr, Lew NL, Liu Y, Lowrie EG, Lazarus JM. The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemo­dialysis. N Engl J Med 1993;329:1001-6.  Back to cited text no. 27    
28.Marcen R, Teruel JL, De La Cal MA, Gi mez C. Spanish Cooperative Study of Nutrition in Hemodialysis. The impact of malnutrition in morbidity and mortality in stable haemodialysis patients. Nephrol Dial Transplant 1997;12:2324-31.  Back to cited text no. 28    
29.De Lima JJ, Sesso R, Abensur H, et al. Predictors of mortality in long-term hemo­dialysis patients with a low prevalence of comorbid conditions. Nephrol Dial Trans­plant 1995;10:1708-13.  Back to cited text no. 29    
30.Kaysen GA, Stevenson FT, Depner TA. Determinants of albumin concentration in hemodialysis patients. Am J Kidney Dis 1997;29:658-68.  Back to cited text no. 30  [PUBMED]  
31.Kaysen GA, Rathore V, Shearer GC, Depner TA. Mechanisms of hypoalbuminemia in hemo-ialysis patients. Kidney Int 1995; 48:510-6.  Back to cited text no. 31  [PUBMED]  

Correspondence Address:
S Gmar-Bouraoui
Service de Néphrologie et Hémodialyse, Centre Hospitalo-universitaire, Fattouma Bourguiba, Monastir
Login to access the Email id

Source of Support: None, Conflict of Interest: None

PMID: 17657086

Rights and PermissionsRights and Permissions


  [Table - 1], [Table - 2], [Table - 3], [Table - 4]


    Similar in PUBMED
    Search Pubmed for
    Search in Google Scholar for
  Related articles
    Email Alert *
    Add to My List *
* Registration required (free)  

    Patients and Methods
    Statistical Analysis
    Article Tables

 Article Access Statistics
    PDF Downloaded403    
    Comments [Add]    

Recommend this journal