CJASN
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published ahead of print on November 29, 2006
Clin J Am Soc Nephrol 2: 89-99, 2007
© 2007 American Society of Nephrology
doi: 10.2215/CJN.01170905
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
CJN.01170905v1
2/1/89    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bradbury, B. D.
Right arrow Articles by Gillespie, B. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bradbury, B. D.
Right arrow Articles by Gillespie, B. W.
Related Collections
Right arrowRelated Article

Epidemiology and Outcomes

Predictors of Early Mortality among Incident US Hemodialysis Patients in the Dialysis Outcomes and Practice Patterns Study (DOPPS)

Brian D. Bradbury*, Rachel B. Fissell{dagger}, Justin M. Albert{ddagger}, Mary S. Anthony*, Cathy W. Critchlow*, Ronald L. Pisoni{ddagger}, Friedrich K. Port{ddagger}, and Brenda W. Gillespie§

* Department of Global Epidemiology, Amgen, Inc., Thousand Oaks, California; {dagger} Department of Internal Medicine, Veteran’s Affairs Medical Center, {ddagger} Dialysis Outcomes and Practice Patterns Study (DOPPS), University Renal Research and Education Association and § Department of Biostatistics, University of Michigan, Ann Arbor, Michigan

Address correspondence to: Dr. Brian D. Bradbury, 1 Amgen Center Drive, MS: 24-1-C, Thousand Oaks, CA 91320; Phone: 805-313-4343; Fax: 805-447-1984; E-mail: bradbury{at}amgen.com


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosures
 References
 
Mortality risk among hemodialysis (HD) patients may be highest soon after initiation of HD. A period of elevated mortality risk was identified among US incident HD patients, and which patient characteristics predict death during this period and throughout the first year was examined using data from the Dialysis Outcomes and Practice Patterns Study (DOPPS; 1996 through 2004). A retrospective cohort study design was used to identify mortality risk factors. All patient information was collected at enrollment. Life-table analyses and discrete logistic regression were used to identify a period of elevated mortality risk. Cox regression was used to estimate adjusted hazard ratios (HR) measuring associations between patient characteristics and mortality and to examine whether these associations changed during the first year of HD. Among 4802 incident patients, risk for death was elevated during the first 120 d compared with 121 to 365 d (27.5 versus 21.9 deaths per 100 person-years; P = 0.002). Cause-specific mortality rates were higher in the first 120 d than in the subsequent 121 to 365 d for nearly all causes, with the greatest difference being for cardiovascular-related deaths. In addition, 20% of all deaths in the first 120 d occurred subsequent to withdrawal from dialysis. Most covariates were found to have consistent effects during the first year of HD: Older age, catheter vascular access, albumin <3.5, phosphorus <3.5, cancer, and congestive heart failure all were associated with elevated mortality. Pre-ESRD nephrology care was associated with a significantly lower risk for death before 120 d (HR 0.65; 95% confidence interval 0.51 to 0.83) but not in the subsequent 121- to 365-d period (HR 1.03; 95% confidence interval 0.83 to 1.27). This care was related to approximately 50% lower rates of both cardiac deaths and withdrawal from dialysis during the first 120 d. Mortality risk was highest in the first 120 d after HD initiation. Inadequate predialysis nephrology care was strongly associated with mortality during this period, highlighting the potential benefits of contact with a nephrologist at least 1 mo before HD initiation.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosures
 References
 
Mortality rates among hemodialysis (HD) patients exceed 20% per year (1), and a higher mortality rate within the first year after initiation of HD has been described (2). Identifying the period of highest risk for death after initiation of HD and factors that are associated with this higher risk are important to the care of patients who are new to HD (incident). Observational studies among prevalent HD patients have identified patient characteristics that are associated with greater mortality risk, including white race, older age, low serum albumin levels, low and elevated serum phosphorus levels, anemia, and cardiovascular disease (traditional risk factors) (3–14), as well as other nontraditional risk factors, including C-reactive protein and IL-6 levels (15,16). Studies also have supported the importance of early nephrology referral in the predialysis period for reducing mortality after HD initiation (17–19). Because these studies typically have included prevalent rather than incident patients, only limited information is available concerning mortality rates and factors that influence mortality immediately after HD initiation. One reason for this may be the 90-d Medicare entitlement period (coordination of benefits) during which coverage is not guaranteed (1). The few studies that have assessed mortality rates or risk predictors in the period immediately after HD initiation (2,20–24) suggest an elevated mortality risk in the first 90 d, but it is unclear whether that elevation is limited to the first 90 d. One population-based study (2) found that 6% of HD patients died within the first 90 d, accounting for nearly 35% of deaths within the first year.

Studies that quantify mortality rates and identify mortality predictors among incident HD populations are useful for understanding the influence of predialysis care and length of time on dialysis. Data from the 2004 Annual Dialysis Report (1) indicate higher mortality rates for patients who have received HD for >5 yr as compared with <2 yr, suggesting that length of time on HD modifies mortality risk. Most patients begin HD with several comorbid conditions that may worsen or develop additional comorbidities with continued dialysis. We hypothesized that the influence on mortality of specific comorbid conditions may change over time, whereas the influence of others remains constant. This study identified a period of elevated mortality risk among incident HD patients and examined the magnitude of associations between various patient characteristics and mortality during this period of elevated risk and throughout the first year on HD using a representative sample of US incident HD patients from the Dialysis Outcomes and Practice Patterns Study (DOPPS).


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Conclusion
 Disclosures
 References
 
Data Source
This analysis used data from DOPPS phases I (1996 through 2001) and II (2002 through 2004). As described previously (25,26), DOPPS is an international cohort study of practice patterns and outcomes among HD patients. In DOPPS I, dialysis facilities from Europe, Japan, and the United States contributed data. DOPPS II added more countries, including Australia, New Zealand, and Canada, and other European countries. Most dialysis facilities that participated in DOPPS I were asked to continue in DOPPS II. Those that chose not to participate were replaced with other facilities. The same study design was used in DOPPS I and II. Facilities that treat 25 or more patients were eligible for selection. A randomized, stratified selection method was used to identify facilities for participation from the pool of eligible facilities. Census information, including basic demographic characteristics and mortality, was collected on all HD patients who were treated in each of the selected facilities. Within each of the selected dialysis facilities, 20 to 40 HD patients who were 18 yr and older were randomly selected for participation in the DOPPS sample. Site investigators abstracted patient information from medical records, including demographic characteristics, predialysis medical care, and laboratory and medical history information for patients who were enrolled in the study. Throughout follow-up, site investigators collected hospitalization, mortality, and loss-to-follow-up information, including dates. Primary and secondary diagnoses and causes of death, as well as reasons for loss to follow-up, were recorded. The DOPPS study received institutional review board approval.

Study Population
Information was collected at the time of enrollment on 15,500 and 13,000 dialysis patients in DOPPS I and II, respectively. For this analysis, only US patients who began dialysis <30 d before enrollment and had at least 1 d of follow-up were included (n = 4802). Limiting the study to incident patients reduced the amount of missing information on predialysis care and permitted analyses of mortality risk immediately after dialysis initiation. Demographic characteristics, medical history, predialysis medical care, and laboratory data that were collected at study enrollment were used for this analysis. Patients were followed from enrollment until the first of the following: Death, transplant, transfer out of the facility, switch to peritoneal or home dialysis, withdrawal from dialysis, or end of the study period. Patients who withdrew from HD were followed for an additional 60 d to capture patients who chose to stop dialysis and died within 2 mo of their last HD.

Outcomes
The primary outcomes in this analysis were all-cause and cause-specific mortality. Causes of death were grouped as follows: Cardiovascular; vascular; infection; a combined category of liver disease, gastrointestinal, and other known causes; and a final category of unknown cause. In addition, deaths that occurred subsequent to withdrawal, irrespective of the immediate cause of death, were examined. Deaths during the first year of follow-up were categorized as deaths that occurred within versus after the identified early period of elevated mortality risk. In analyses of deaths that occurred during this early period, patients who survived were censored at the end of the early period. Analyses of deaths that occurred beyond this early period of elevated risk included only patients who survived beyond this period; those who survived beyond 365 d were censored at 366 d.

Covariates
Patient information included four types: Demographics and dialysis care, laboratory values, medical history, and predialysis medical care. Demographic characteristics included age (18 to 44, 45 to 64, 65 to 74, 75+ yr), gender, race (white, nonwhite), body mass index ([kg/m2]; <20, 20 to <25, 25 to <30, 30+), and primary cause of ESRD (diabetes, hypertension, other [includes glomerulonephritis; secondary glomerulonephritis-vasculitis; interstitial nephritis; polycystic kidney disease, congenital; and renal tumor] and missing); dialysis care was measured by vascular access type in place at dialysis initiation (arteriovenous fistula, graft [bovine, synthetic], catheter [cuffed percutaneous, temporary], other, missing). Laboratory values were categorized according to the Kidney Disease Outcomes Quality Initiative (K/DOQI) guidelines (27–29) and include the following: Albumin (<3.5, ≥3.5 g/dl, unknown), albumin-corrected calcium (<8.4, 8.4 to 9.5, >9.5 g/dl, unknown), hemoglobin (<11.0, 11.0 to <12.0, 12.0+ g/dl, unknown), phosphorus (<3.5, 3.5 to 5.5, >5.5 g/dl, unknown), and calcium-phosphorus product (<55, >55 mg2/ml2, unknown). The effect of parathyroid hormone (PTH) was not assessed because of large amount (>50%) of missing information. Composite variables that reflect patient medical history characteristics were described previously (25); each variable (dichotomized as yes/no) was captured at enrollment. Visit to a nephrologist >1 mo before enrollment in DOPPS (pre-ESRD nephrology care; yes/no) was used to assess pre-ESRD medical care.

Statistical Analyses
All analyses were conducted on the combined DOPPS I and DOPPS II samples unless stated otherwise. Descriptive statistics characterized the study population. Life-table analyses were used to examine the estimated hazard function, by month, for patients who were captured in the census (all HD patients in the facility) and separately for patients who were enrolled in the DOPPS I and II samples to identify a period of elevated risk. Discrete survival analysis using logistic regression with intervals of 1 mo was used to test for differences in mortality rates for early versus late periods (e.g., 0 to 90 versus 91 to 365, 0 to 120 versus 121 to 365, 0 to 150 versus 151 to 365). The t test and {chi}2 analyses were used to compare characteristics of patients who were identified in the census but not included in the DOPPS II sample (nonparticipants) with those who were included in the DOPPS II sample (participants). This subanalysis was limited to DOPPS II patients because date of first dialysis was not collected in the DOPPS I census, and, therefore, incident patients, other than those in the DOPPS I sample, could not be identified. Cox proportional hazards regression was used to generate crude and adjusted hazard ratios (AHR) and 95% confidence intervals (CI) for the association between baseline patient characteristics and mortality in different periods during the first year of follow-up. The full models adjusted for all patient characteristics. Two-sided P < 0.05 was considered statistically significant. Time-dependent Cox regression also was used to examine the interaction between patient characteristics and time at risk for mortality, controlling for all other patient characteristics. In particular, tests for differences in the HR for a particular covariate before and after the period of elevated mortality were performed. All covariates initially were assessed for time-dependent effects using plots of ln(–ln[S(t)]) versus time for each covariate in turn, with S(t) estimated from models that were adjusted for other baseline predictors but stratified on the covariate of interest. The final model included the significant time-dependent interaction terms that were identified in previous models (P < 0.05) along with all other main effects. Analyses were conducted using SAS 9.1 (SAS Institute, Cary, NC).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Conclusion
 Disclosures
 References
 
A total of 4156 incident US HD patients were identified in the DOPPS II census, and of those, 1025 (25%) were included in the DOPPS II sample. Study participants (n = 1025) were similar to nonparticipants (n = 3131) with respect to race (P = 0.83) and gender (P = 0.20); however, nonparticipants tended to be slightly older (64.2 versus 62.6 yr; P = 0.02). From the DOPPS I sample, 3777 incident US dialysis patients were identified. There were no material differences between DOPPS I and DOPPS II sample participants with regard to demographic characteristics except that DOPPS II patients were more likely to have diabetes as their primary cause of ESRD (53.3 versus 47.3%; P = 0.0006).

Among the 4802 patients in the combined DOPPS I and DOPPS II sample, 4292 (89%) enrolled in DOPPS within 14 d of dialysis initiation; 1977 (41%) enrolled at their first dialysis session. The median time between first dialysis and DOPPS enrollment was 3 d. The distribution of patient characteristics for the population overall and according to mortality during the first year is presented in Table 1. Overall, 51% of the patients were 65 yr or older, 56% were male, and 65% were white.


View this table:
[in this window]
[in a new window]

  		Table 1. Demographic characteristics for the patients overall, for those who died within 120 d and 121 to 365 d after initiating HD, and for those who were censored or survived to 365 d, DOPPS I and II, 1996 to 2004a

 
In the first 365 d of follow-up, 841 (17.5%) deaths occurred. Examination of the estimated hazard functions among all patient groups (DOPPS I and II samples, separately and combined, and DOPPS II census) revealed an elevated hazard of death during the first 120 d, with this hazard being most pronounced among patients in the DOPPS II census (Figure 1). Although we tested other time-period comparisons (e.g., 0 to 90 versus 91 to 365, as stated in the Materials and Methods section), in the combined DOPPS I and II samples, mortality risk was highest in the first 120 d (27.5 deaths per 100 person-years) compared with 121 to 365 d (21.9 deaths per 100 person-years; P = 0.002).


Figure 1
View larger version (15K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Figure 1. Estimated hazard function for the Dialysis Outcomes and Practice Patterns Study (DOPPS) II Census (n = 4156), DOPPS I sample (n = 3777), DOPPS II sample (n = 1025), and DOPPS I and II samples (n = 4802).

 
Cause-specific death rates by time period and the proportion of deaths that occurred subsequent to patient withdrawal from HD are presented in Figure 2. Cardiovascular causes accounted for the most deaths in the first year, and the rates were significantly higher in the first 120 d compared with the subsequent 121 to 365 d (10.5 versus 6.8 per 100 person-years; P = 0.008). The causes of cardiac death that contributed most to this effect were cardiac arrest, cause unknown; acute myocardial infarction; cardiac arrhythmia; and atherosclerotic heart disease (death rates 5.2 versus 3.7, 1.7 versus 1.2, 1.5 versus 0.8, and 1.2 versus 0.8 per 100 person-years, respectively). Vascular causes also were significantly more common in the first 120 d than during days 121 to 365.


Figure 2
View larger version (20K):
[in this window]
[in a new window]
[as a PowerPoint slide]
  		Figure 2. (Left) Cause-specific mortality rates and 95% confidence intervals for the <120- and 121- to 365-d periods. (Right) Percentage of all deaths during the <120 and 121- to 365-d periods that occur subsequent to withdrawal.

 
Overall, 151 patients withdrew from dialysis during the first year and subsequently died within 60 d. The median time between withdrawal and death was 7 d. Deaths that occurred subsequent to patient withdrawal accounted for 20% of deaths in the first 120 d and 15% in the subsequent 121 to 365 d. The primary cause of death was unknown in 44% of deaths among patients who withdrew from HD compared with fewer than 18% among those who did not withdraw. If patients had been censored at the time of withdrawal, rather than followed for up to 60 d after withdrawal, the mortality rate in the first 120 d after HD initiation would have been 20.6 deaths per 100 person-years compared with 15.1 deaths per 100 person-years between 121 and 365 d (P < 0.0001).

HR estimates for the association between each predictor and death up to 120 d, 121 to 365 d, and during the first 365 d overall, adjusted for other patient characteristics, are provided in Table 2. Patient characteristics that were significantly associated with an increased risk for mortality during the first 120 d included age >75 yr (AHR 2.49), white race (AHR 1.40), catheter vascular access (AHR 1.71), albumin levels <3.5 g/dl (AHR 1.57), phosphorus levels <3.5 mg/dl (AHR 1.47), HIV/AIDS (AHR 2.85), history of cancer (AHR 1.41), history of lung disease (AHR 1.33), history of neurologic disease (AHR 1.50), and history of a psychiatric disorder (AHR 1.35). Factors that were associated with a significantly lower mortality risk during the first 120 d included the diagnosis of hypertension (AHR 0.55) and pre-ESRD nephrology care >1 mo before initiation of dialysis (AHR 0.65). Cause-specific death rates (per 100 person-years) were lower in the first 120 d among those who saw a nephrologist >1 mo before ESRD compared with those who did not (cardiac 9.24 versus 17.7; infection 2.2 versus 3.5; vascular 1.8 versus 3.2; liver/gastrointestinal/other 3.2 versus 6.3; unknown 5.3 versus 7.9 [data not shown]). Similarly, deaths subsequent to withdrawal were lower among those who saw a nephrologist >1 mo before ESRD compared with those who did not (4.5 versus 8.2 per 100 person-years, respectively).


View this table:
[in this window]
[in a new window]

  		Table 2. AHR and adjusted 95% CI for the association between demographic, laboratory, pre-ESRD treatment, and comorbid characteristics and death ≤120, 121 to 365, and ≤365 d after initiation of HD among incident HD patients (n = 4802), DOPPS 1996 to 2004a

 
Patient characteristics that were significantly associated with an increased risk for mortality between 121 and 365 d included age >75 yr (AHR 2.45), white race (AHR 1.39), catheter vascular access (AHR 1.42), calcium levels >9.5 (AHR 1.30), HIV/AIDS (AHR 7.6), history of neurologic disease (AHR 1.83), and history of lung disease (AHR 1.46). Patients with calcium levels <8.4 were at significantly lower risk for mortality (AHR 0.65), and those with hypertension remained at lower risk for mortality (AHR 0.67), but the protective effect of pre-ESRD nephrology care no longer was evident.

Finally, we evaluated whether the association between each predictor and mortality differed between the <120-d and 121- to 365-d periods using predictor by time interaction terms while simultaneously adjusting for the other main effects. Of variables that were significantly associated with death in either the <120-d or 121- to 365-d time period, only serum calcium <8.4 mg/L (test for interaction with time, P = 0.009), HIV/AIDS status (P = 0.02), and pre-ESRD nephrologist care (P = 0.003) were observed to have a statistically significant different impact on mortality risk in the first 120-d than in the subsequent 121- to 365-d period.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Conclusion
 Disclosures
 References
 
In this, the largest population-based sample of incident US HD patients examined to date, we found that the period of elevated mortality risk extended through the first 120 d. In previous studies, mortality risk was assessed during the first 90 d (2,20,21), whereas this analysis compared risks during various time intervals during the first year to identify the period of elevated risk. In addition, this study provides a detailed examination of the association between various comorbid and clinical characteristics that were collected at HD initiation and mortality risk during the first year of HD, which may help physicians to prioritize attention to specific health issues. Although various time periods were examined, the largest difference (both in magnitude and in statistical significance) was observed for the first 120 d versus the subsequent 121 to 365 d after HD initiation. Cardiovascular causes accounted for the largest percentage of deaths during the first year, but the rate of cardiovascular-related deaths was considerably higher in the first 120 d. In addition, the proportion of deaths that occurred subsequent to patient withdrawal was considerably higher in the first 120 d. Patients who were older or white; patients with low serum albumin levels at baseline, a catheter access in place at first dialysis, HIV/AIDS, a history of congestive heart failure (CHF), cancer, lung disease, neurologic disease or a psychiatric disorder; or patients who did not visit a nephrologist at least 1 mo before initiating HD were at significantly elevated risk for mortality within 120 d of initiating HD. In the subsequent 121 to 365 d, older age, white race, history of lung disease or a psychiatric disorder, and HIV/AIDS remained strongly predictive of mortality. Of the predictors examined, only pre-ESRD nephrology care, HIV/AIDS status, diabetes status, and calcium levels <8.4 g/dl were found to have a statistically different impact on the risk for mortality during the first year of follow-up.

Few studies have assessed mortality rates and mortality predictors in incident HD patients (2,20,21). Evaluations that use large population-based databases such as Medicare are difficult because of the 90-d entitlement period (1), which precludes complete ascertainment of patients with ESRD during this period. Analyses of mortality during this initial period are prone to bias, particularly given the elevated mortality risk during the first 120 d that we observed, unless conducted in populations in which all patients who initiate dialysis can be identified and are guaranteed continued coverage. Studies that were conducted in non–population-based samples of incident patients have reported 90-d mortality rates ranging from 12 (20–22) to 16% (23). A Michigan study of nearly 2400 patients with ESRD reported deaths within 90 d of initiation of dialysis in 10% of patients (24). Most recently, Soucie and McClellan (2) reported 6% of HD patients in an ESRD registry in the southeastern United States died within 90 d after dialysis initiation. In a subanalysis (data not shown), we found that 5.6% of deaths occurred within 90 d of HD initiation; 8% of patients died within the first 120 d, representing 46% of all deaths during the first year of follow-up. Both our study and that of Soucie and McClellan (2) were population based, minimizing concerns about selection bias, and both include relatively recent data from patients who initiated HD. The latter point may account for the lower mortality rates as compared with findings from earlier studies, potentially reflecting improvements in pre-ESRD treatment, including earlier referral to a nephrologist, which we and others (30,31) suggest improves survival once on dialysis. It is interesting that patients with diabetes were more likely than patients without diabetes (78.6 versus 72.6%; P < 0.0001) to receive predialysis care from a nephrologist, but this greater access to care did not seem to explain the unexpected lower 120-d mortality rate that was observed among patients with diabetes.

Our findings with respect to factors that are associated with elevated mortality risk are consistent with previous studies among prevalent HD patient populations, which are composed of patients with a range of times since dialysis initiation. The strength of older age and white race as mortality risk factors throughout the first 365 d found in our study among incident patients and in other studies of prevalent populations (10,11) suggests a consistent effect of these factors on mortality risk. We also found that factors such as low serum albumin and phosphorus levels, catheter vascular access, and absence of referral to a nephrologist >1 mo before HD initiation were associated strongly with early mortality. These results support the importance of the ongoing recommendations for arteriovenous fistula placement in the predialysis period and avoidance of temporary or permanent catheter placement (27). Furthermore, these results suggest that greater medical attention before ESRD onset may lead to improved survival (20,31). Our findings are consistent with those from the RightStart Program (32), which demonstrated significantly lower mortality rates for patients who were new to dialysis when specific disease management protocols were used before ESRD. Moreover, our results remained largely unchanged even after adjustment for pre-ESRD erythropoietin (EPO) use, which reflects one component of pre-ESRD care. Previous studies have reported a reduction in mortality for patients who were treated with EPO before initiating HD (33), but in this analysis, when pre-ESRD nephrologist care and EPO use were included in the same model, only pre-ESRD nephrologist care remained strongly protective in the first 120 d. In the 121- to 365-d period, when both pre-ESRD nephrologist care and EPO use were included in the same model, the effect of pre-ESRD nephrologist care was completely attenuated.

We found that specific chronic comorbid conditions, including history of CHF, lung and neurologic disease (e.g., dementia, depression, peripheral neuropathy), and HIV/AIDS, were associated with increased mortality risk throughout the first year of follow-up, although the magnitude of their effects was not constant. With the exception of CHF, the strength of these mortality risk factors increased with longer follow-up. History of CHF was more strongly associated with death within the first 120 d, emphasizing the prognostic importance of cardiovascular disease in patients who initiate HD (34).

In our analyses, the crude mortality rates for both the first 120 d and the subsequent 121 to 365 d were substantially lower when patients who withdrew from dialysis were censored at the time of their last HD (were not included as deaths in the estimation of incidence rates), reflecting the greater numbers of deaths captured during the 60 d after withdrawal. Although the majority of withdrawals occurred within the first 120 d and a larger proportion of deaths subsequent to withdrawal occurred during this period, when patients who withdrew were censored at the time of withdrawal, the resulting effect estimates remained largely unchanged (data not shown), indicating that mortality risk factors were not simply predictors of withdrawal. However, these findings do provide further evidence that previous estimates of mortality risk among prevalent patient populations may be underestimates for incident patients, particularly given the incomplete ascertainment of HD patients within the first 90 d of dialysis initiation.

This study should be evaluated in light of the following limitations. First, the risk for mortality among the entire HD patient population (DOPPS II census) is slightly higher than among study participants, especially in the first 120 d after HD initiation, suggesting possible selection bias. In fact, the death rates in the first 120 d (versus 121 to 365 d) among patients who were randomly selected for inclusion in the sample were 28.9 versus 21.0 per 100 person-years, respectively, compared with the analogous death rates among those who actually were included in the sample (25.4 versus 19.8 deaths per 100 person-years, respectively). Although the 121- to 365-d death rates are similar, the early death rates are quite different, showing an underascertainment of deaths during the early period, consistent with the comparison of the DOPPS II census versus the sample in Figure 1. One explanation for this underascertainment is that in many cases, hospital records for patients who died were archived, which made accessing these records difficult at the time of study form completion. In addition, the Health Insurance Portability and Accountability Act, which was enacted in April 2003 during the DOPPS II period, required patient consent for any new DOPPS II patients before enrollment. Consequently, patients who were close to death at the time of HD initiation may not have been approached by a study coordinator or agreed to participate, which could have resulted in a slightly healthier population being enrolled. Therefore, the 120-d mortality rate that we report may be a slight underestimation of the true mortality rate in the incident patient population.

Second, the time-varying influence of mortality risk factors that was observed in this study may reflect depletion of patients with specific comorbidities over time. Because we assessed only the influence of patient characteristics that were collected at baseline and included only patients who were new to dialysis, our effect estimates may underestimate the association between mortality and characteristics that often change after beginning dialysis (e.g., levels of hemoglobin, calcium-phosphorus product). As such, a single measurement that is taken at baseline might not characterize adequately a patient’s relevant exposure. The choice to analyze only baseline characteristics was influenced partially by the frequency of DOPPS data collection (approximately every 4 mo). Patients who died within 120 d of initiating HD would have no available follow-up information, but those who survived beyond 120 d potentially would have additional time points at which data could be collected. More important, however, the intent of this analysis was to identify patient characteristics at dialysis initiation that predict mortality over some early period of follow-up. Patient information that is collected subsequent to starting dialysis, particularly laboratory values, is sensitive to concomitant treatments and may be modified by increased contact with a nephrologist, obscuring associations with specific characteristics at HD initiation.

The strengths of this study also should be acknowledged. This analysis represents the largest population of incident HD patients examined to date. Because data collection for patients who were enrolled in DOPPS was not connected to Medicare reimbursement, underascertainment of the population at risk as a result of the 90-d coordination-of-benefits period was not a concern. In addition, because most patients were enrolled before patient consent was required, consent issues were minimized. Substantial information on patient medical history as well as laboratory parameters was collected during the baseline assessment, facilitating a more comprehensive investigation of mortality risk factors.


   Conclusion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
Disclosures
 References
 
This study suggests that a period of elevated mortality risk for patients who initiate HD extends to approximately 120 d after dialysis initiation. Excess early mortality was observed for all causes of death except infection, with cardiac causes showing the largest increase in early mortality. In addition, death rates after withdrawal from dialysis were substantially larger in the early period. Most covariates had consistent effects during the entire period. However, during the first 120 d on HD, patients who were under the care of a nephrologist at least 1 mo before dialysis initiation had significantly reduced mortality. This care was related to approximately 50% lower rates of both cardiac deaths and withdrawal from dialysis during the first 120 d.


   Disclosures
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosures
References
 
B.D.B., M.S.A., and C.W.C. are employees of Amgen, Inc.


   Acknowledgments
 
The Dialysis Outcomes and Practice Patterns Study (DOPPS) is supported by research grants from Amgen and Kirin without restrictions on publications.


   Footnotes
 
Published online ahead of print. Publication date available at www.cjasn.org.

See the related editorial, "Predialysis Nephrology Care Improves Dialysis Outcome: Now What? Or Chapter Two," on pages 143–145.

Received September 29, 2005. Accepted September 26, 2006.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosures
 References
 

  1. US Renal Data System: USRDS 2004 Annual Data Report: Atlas of End-Stage Renal Disease in the United States, Bethesda, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases,2004
  2. Soucie JM, McClellan WM: Early death in dialysis patients: Risk factors and impact on incidence and mortality rates. J Am Soc Nephrol7 :2169 –2175,1996[Abstract]
  3. Goldwasser P, Mittman N, Antignani A, Burrell D, Michel MA, Collier J, Avram MM: Predictors of mortality in HD patients. J Am Soc Nephrol3 :1613 –1622,1993[Abstract]
  4. Avram MM, Mittman N, Bonomini L, Chattopadhyay J, Fein P: Markers for survival in dialysis: A seven-year prospective study. J Kidney Dis26 :209 –219,1995
  5. Leavey SF, Strawderman RL, Jones CA, Port FK, Held PJ: Simple nutritional indicators as independent predictors of mortality in HD patients. Am J Kidney Dis31 :997 –1006,1998[Medline]
  6. Fernandez-Reyes MJ, Alvarez-Ude F, Sanchez R, Mon C, Iglesias P, Diez JJ, Vazquez A: Inflammation and malnutrition as predictors of mortality in patients on HD. J Nephrol15 :136 –143,2002[Medline]
  7. Cooper BA, Penne EL, Bartlett LH, Pollock CA: Protein malnutrition and hypoalbuminemia as predictors of vascular events and mortality in ESRD. Am J Kidney Dis43 :61 –66,2004[CrossRef][Medline]
  8. Herzog CA, Li S, Weinhandl ED, Strief JW, Collins AJ, Gilbertson DT: Survival of dialysis patients after cardiac arrest and the impact of implantable cardioverter defibrillators. Kidney Int68 :818 –825,2005[CrossRef][Medline]
  9. Pisoni RL, Bragg-Gresham JL, Young EW, Akizawa T, Asano Y, Locatelli F, Bommer J, Cruz JM, Kerr PG, Mendelssohn DC, Held PJ, Port FK: Anemia management and outcomes from 12 countries in the dialysis outcomes and practice patterns study (DOPPS). Am J Kidney Dis44 :94 –111,2004[CrossRef][Medline]
  10. Ma JZ, Ebben J, Xia H, Collins AJ: Hematocrit level and associated mortality in HD patients. J Am Soc Nephrol10 :610 –619,1999[Abstract/Free Full Text]
  11. Goodkin DA, Bragg-Gresham JL, Koenig KG, Wolfe RA, Akiba T, Andreucci VE, Saito A, Rayner HC, Kurokawa K, Port FK, Held PJ, Young EW: Association of comorbid conditions and mortality in HD patients in Europe, Japan, and the United States: The Dialysis Outcomes and Practice Patterns Study (DOPPS). J Am Soc Nephrol14 :3270 –3277,2003[Abstract/Free Full Text]
  12. Block GA, Klassen PS, Lazarus JM, Ofsthun N, Lowrie EG, Chertow GM: Mineral metabolism, mortality, and morbidity in maintenance HD. J Am Soc Nephrol15 :2208 –2218,2004[Abstract/Free Full Text]
  13. Kalantar-Zadeh K, Kuwae N, Regidor DL, Kovesdy CP, Kilpatrick RD, Shinaberger CS, McAllister CJ, Budoff MJ, Salusky IB, Kopple JD: Survival predictability of time-varying indicators of bone disease in maintenance hemodialysis patients. Kidney Int70 :771 –780,2006[CrossRef][Medline]
  14. Young EW, Albert JM, Satayathum S, Goodkin DA, Pisoni RL, Akiba T, Akizawa T, Kurokawa K, Bommer J, Piera L, Port FK: Predictors and consequences of altered mineral metabolism: The Dialysis Outcomes and Practice Patterns Study. Kidney Int67 :1179 –1187,2005[CrossRef][Medline]
  15. den Elzen P, Manen JG, Boeschoten EW, Krediet RT, Dekker FW: The effect of single and repeatedly high concentrations of C-reactive protein on cardiovascular and non-cardiovascular mortality in patients starting with dialysis. Nephrol Dial Transplant21 :1588 –1595,2006[Abstract/Free Full Text]
  16. Honda H, Qureshi AR, Heimburger O, Barany P, Wang K, Pecoits-Filho R, Stenvinkel P, Lindholm B: Serum albumin, C-reactive protein, interleukin 6, and fetuin a as predictors of malnutrition, cardiovascular disease, and mortality in patients with ESRD. Am J Kidney Dis47 :139 –148,2006[CrossRef][Medline]
  17. Arora P, Obrador GT, Ruthazer R, Kausz AT, Meyer KB, Jenuleson CS, Pereira BJ: Prevalence, predictors, and consequences of late nephrology referral at a tertiary care center. J Am Soc Nephrol10 :1281 –1286,1999[Abstract/Free Full Text]
  18. Kessler M, Frimat L, Panescu V, Briancon S: Impact of nephrology referral on early and midterm outcomes in ESRD: EPidemiologie de l’Insuffisance REnale chronique terminale en Lorraine (EPIREL): Results of a 2-year, prospective, community-based study. Am J Kidney Dis42 :474 –485,2003[CrossRef][Medline]
  19. Lin CL, Chuang FR, Wu CF, Yang CT: Early referral as an independent predictor of clinical outcome in end-stage renal disease on HD and continuous ambulatory peritoneal dialysis. Ren Fail26 :531 –537,2004[CrossRef][Medline]
  20. Innes A, Rowe PA, Burden RP, Morgan AG: Early deaths on renal replacement therapy: The need for early nephrological referral. Nephrol Dial Transplant7 :467 –471,1992[Abstract/Free Full Text]
  21. Khan IH, Catto GR, Edward N, MacLeod AM: Death during the first 90 days of dialysis: A case control study. Am J Kidney Dis25 :276 –280,1995[Medline]
  22. Garcia-Garcia G, Deddens JA, D’Achiardi-Rey R, First MR, Samuels SJ, Kant S, Pollak VE: Results of treatment in patients with end-stage renal disease: A multivariate analysis of risk factors and survival in 341 successive patients. Am J Kidney Dis5 :10 –18,1985[Medline]
  23. Wright LF: Survival in patients with end-stage renal disease. Am J Kidney Dis17 :25 –28,1991[Medline]
  24. Weller JM, Port FK, Swartz RD, Ferguson CW, Williams GW, Jacobs JF Jr: Analysis of survival of end-stage renal disease patients. Kidney Int21 :78 –83,1982[Medline]
  25. Pisoni RL, Gillespie BW, Dickinson DM, Chen K, Kutner MH, Wolfe RA: The Dialysis Outcomes and Practice Patterns Study (DOPPS): Design, data elements, and methodology. Am J Kidney Dis44[Suppl 3] :7 –15,2004
  26. Young EW, Goodkin DA, Mapes DL, Port FK, Keen ML, Chen K, Maroni BL, Wolfe RA, Held PJ: The Dialysis Outcomes and Practice Patterns Study (DOPPS): An international HD study. Kidney Int74[Suppl] :S74 –S81,2000[CrossRef]
  27. NKF-K/DOQI Clinical Practice Guidelines: Update 2000. Am J Kidney Dis37[Suppl 1] :S1 –S235,2001
  28. National Kidney Foundation Kidney Disease Outcomes Quality Initiative (NKF K/DOQI): Bone metabolism and disease in chronic kidney disease. Available at: http://www.kidney.org/professionals/kdoqi/guidelines.cfm. Accessed November 15,2006
  29. K/DOQI clinical practice guidelines for nutrition in chronic renal failure. Am J Kidney Dis35[Suppl 2] :S17 –S104,2000
  30. Biesenbach G, Hubmann R, Janko O, Schmekal B, Eichbauer-Sturm G: Predialysis management and predictors for early mortality in uremic patients who die within one year after initiation of dialysis therapy. Ren Fail24 :197 –205,2002[CrossRef][Medline]
  31. Khan SS, Xue JL, Kazmi WH, Gilbertson DT, Obrador GT, Pereira BJ, Collins AJ: Does predialysis nephrology care influence patient survival after initiation of dialysis? Kidney Int67 :1038 –1046,2005[CrossRef][Medline]
  32. Wingard RL, Pupim L, Ikizler TA, Thomas JE, Hakim R: Reduction of 90-day mortality in incident ESRD patients: The RightStart Program [Abstract]. J Am Soc Nephrol16 :96A ,2005
  33. Fink J, Blahut S, Reddy M, Light P: Use of erythropoietin before the initiation of dialysis and its impact on mortality. Am J Kidney Dis37 :348 –355,2001[Medline]
  34. Harnett JD, Foley RN, Kent GM, Barre PE, Murray D, Parfrey PS: Congestive heart failure in dialysis patients: Prevalence, incidence, prognosis and risk factors. Kidney Int47 :884 –890,1995[Medline]

Related Article

Predialysis Nephrology Care Improves Dialysis Outcomes: Now What? Or Chapter Two
Thomas A. Golper
Clin. J. Am. Soc. Nephrol. 2007 2: 143-145. [Full Text] [PDF]



This article has been cited by other articles:


Home page
 Nephrol Dial TransplantHome page
M. K. Sigrist, A. Levin, L. Er, and C. W. McIntyre
Elevated osteoprotegerin is associated with all-cause mortality in CKD stage 4 and 5 patients in addition to vascular calcification
Nephrol. Dial. Transplant., June 2, 2009; (2009) gfp253v1.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
I-W. Wu, S.-Y. Wang, K.-H. Hsu, C.-C. Lee, C.-Y. Sun, C.-J. Tsai, and M.-S. Wu
Multidisciplinary predialysis education decreases the incidence of dialysis and reduces mortality--a controlled cohort study based on the NKF/DOQI guidelines
Nephrol. Dial. Transplant., June 2, 2009; (2009) gfp259v1.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
C. Couchoud, M. Labeeuw, O. Moranne, V. Allot, V. Esnault, L. Frimat, B. Stengel, and for the French Renal Epidemiology and Information
A clinical score to predict 6-month prognosis in elderly patients starting dialysis for end-stage renal disease
Nephrol. Dial. Transplant., May 1, 2009; 24(5): 1553 - 1561.
[Abstract] [Full Text] [PDF]

Home page
 J. Am. Soc. Nephrol.Home page
W. M. McClellan, H. Wasse, A. C. McClellan, A. Kipp, L. A. Waller, and M. V. Rocco
Treatment Center and Geographic Variability in Pre-ESRD Care Associate with Increased Mortality
J. Am. Soc. Nephrol., May 1, 2009; 20(5): 1078 - 1085.
[Abstract] [Full Text] [PDF]

Home page
 CJASNHome page
B. D. Bradbury, M. D. Danese, M. Gleeson, and C. W. Critchlow
Effect of Epoetin Alfa Dose Changes on Hemoglobin and Mortality in Hemodialysis Patients with Hemoglobin Levels Persistently below 11 g/dL
Clin. J. Am. Soc. Nephrol., March 1, 2009; 4(3): 630 - 637.
[Abstract] [Full Text] [PDF]

Home page
 CJASNHome page
M. B. Patwardhan, K. Kawamoto, D. Lobach, U. D. Patel, and D. B. Matchar
Recommendations for a Clinical Decision Support for the Management of Individuals with Chronic Kidney Disease
Clin. J. Am. Soc. Nephrol., February 1, 2009; 4(2): 273 - 283.
[Abstract] [Full Text] [PDF]

Home page
 NDT PlusHome page
F. Carrera and M. Burnier
Use of darbepoetin alfa in the treatment of anaemia of chronic kidney disease: clinical and pharmacoeconomic considerations
NDT Plus, January 1, 2009; 2(suppl_1): i9 - i17.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
CJN.01170905v1
2/1/89    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bradbury, B. D.
Right arrow Articles by Gillespie, B. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bradbury, B. D.
Right arrow Articles by Gillespie, B. W.
Related Collections
Right arrowRelated Article

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS