Version May 2024
OVERVIEW — Although maintenance dialysis prevents death from uremia, mortality among patients with end-stage kidney disease (ESKD) remains high. Patients on dialysis have a substantially higher multivariable-adjusted mortality than patients not on dialysis who have cancer, diabetes, or cardiovascular disease [1].
However, mortality rates among patients with ESKD have improved over time. According to the United States Renal Data System (USRDS) report in 2022, all-cause mortality in patients with ESKD declined steadily between 2010 and 2019, before sharply increasing in 2020, in association with the COVID-19 pandemic [2]. Between 2001 and 2016, adjusted relative mortality rates reported by USRDS declined by 29 percent [1]. Over this period, the adjusted relative reduction in mortality was 28 percent among patients on hemodialysis (HD) and 43 percent among patients on peritoneal dialysis (PD).
It is well established that suboptimal replacement of native kidney function has an adverse effect on patient survival. This has important implications since more intensive dialysis, particularly above a certain threshold value, may improve survival, such as with short daily or nocturnal HD.
Additional factors associated with the dialysis procedure may also correlate with decreased survival. As examples:
●Dialysis vintage is associated with an enhanced risk of death, with each additional year of dialysis treatment associated with an increase in the risk of death of approximately 6 percent [3].
●Potassium levels <4.0 or >5.6 mEq/L are associated with increased mortality in patients on HD compared with serum levels between 4.6 and 5.3 mEq/L [4].
●In the United States, some, but not all, studies have reported that mortality may be lower among patients on dialysis being treated in not-for-profit facilities compared with those undergoing dialysis in for-profit facilities [5-9].
●In one study in the United States, a high nephrologist caseload was associated with a higher patient mortality rate compared with that of patients under the care of nephrologists with lower caseloads [10].
●The size of the dialysis facility may affect mortality. One study of 385,074 patients on dialysis suggested that a higher mortality was observed among patients dialyzed at facilities with <15 stations [11].
A review of those factors associated with mortality among patients on dialysis, with an emphasis upon those characteristics related to dialysis itself, will be presented here [12-14].
CAUSES OF DEATH — Before discussing the different factors underlying mortality in patients on dialysis, it is helpful to briefly review the major causes of death in this patient population: cardiovascular disease, infection, and withdrawal from dialysis [15-18].
●Cardiovascular disease accounts for approximately 50 percent of deaths. While a decline in cardiovascular deaths has occurred in the general population, a similar trend has not been observed in patients on dialysis [15]. This discrepancy is in part due to the demographics of patients about to be started on dialysis: Approximately 40 percent are diabetic; the average age is approximately 60 years, and approximately 20 percent are >75 years of age; many patients have underlying cardiac disease.
●Infections, which are the second most common cause of death, are usually due to common organisms (such as Staphylococcus aureus) and are frequently related to the hemodialysis (HD) vascular access [19]. Among patients on dialysis, infection-related mortality is approximately 20 times higher than in the general population [20]. (See "Non-access-related infections in patients on chronic dialysis".)
●Withdrawal from dialysis accounts for approximately 15 to 25 percent of patient deaths. (See "Kidney palliative care: Withdrawal of dialysis".)
RISK FACTORS NOT RELATED TO DIALYSIS — A large number of risk factors that are unrelated to the dialysis procedure have been associated with decreased survival among patients on dialysis.
Comorbid conditions — The presence of comorbid disease is an increasingly common problem, much more prevalent in new patients started on dialysis today than previously [15,18,21-27]. As noted in the 2007 United States Renal Data System (USRDS) Annual Report, diabetes mellitus was the leading cause of end-stage kidney disease (ESKD) in those initiating hemodialysis (HD) (44 percent) [15]. Extended survival at 10 years appears to be less likely for patients with diabetes than those without (4 versus 11 to 14 percent). The 2022 USRDS report observed that approximately 60 percent of patients on dialysis in 2020 had diabetes [2].
Heart disease is also very common in the dialysis population. USRDS data suggest that approximately 28 percent of patients with incident ESKD have a history of congestive heart failure [2]. Approximately 80 percent of all patients enrolled in the Hemodialysis (HEMO) Study were noted to have some form of heart disease, with nearly 40 percent having ischemic heart disease [28]. It has also been estimated that only 27 percent of patients about to enter a dialysis regimen have a normal echocardiogram, while 19 percent already have severe left ventricular hypertrophy [29,30]. In addition, as many as 75 percent of the total ESKD population have at least a 50 percent narrowing of at least one coronary artery. Even patients with mild to moderate chronic kidney failure are at markedly increased cardiovascular risk, with chronic kidney disease alone considered a coronary heart disease risk equivalent. (See "Risk factors and epidemiology of coronary heart disease in end-stage kidney disease (dialysis)" and "Chronic kidney disease and coronary heart disease".)
In addition to antecedent coronary artery disease, additional factors common in patients on dialysis may promote the development of coronary disease and enhanced cardiovascular mortality:
●Hypertension, which is present in approximately 80 percent of patients at the onset of dialysis and, with effective fluid control, approximately 25 to 30 percent at the end of the first year. (See "Hypertension in patients on dialysis".)
●Metabolic abnormalities, particularly hyperphosphatemia, an elevated calcium phosphorous product, and increased parathyroid hormone (PTH) levels. These extremely common abnormalities are increasingly recognized as possibly important unique risk factors for patients on dialysis. (See "Risk factors and epidemiology of coronary heart disease in end-stage kidney disease (dialysis)" and "Vascular calcification in chronic kidney disease" and 'Disorders of mineral metabolism' below.)
●Left ventricular hypertrophy, due in part to hypertension, chronic anemia, and perhaps decreased kidney function. (See "Overview of screening and diagnosis of heart disease in patients on dialysis".)
●Hyperlipidemia. (See "Lipid abnormalities in nephrotic syndrome" and "Chronic kidney disease and coronary heart disease".)
●Diabetes mellitus, which is a major risk factor for coronary artery disease (in patients with and without kidney failure) and is common among patients requiring dialysis. (See "Prevalence of and risk factors for coronary heart disease in patients with diabetes mellitus".)
The concurrent presence of other life-threatening conditions will also affect overall survival [25,27,31-33]. In an extensive analysis of Dialysis Outcomes and Practice Patterns Study (DOPPS) data, 17 comorbid conditions accounted for nearly all of the total variance that predicted survival among patients on HD [27].
Underlying kidney disease — Five-year survival among patients on dialysis is best with chronic glomerular diseases and polycystic kidney disease (PKD), intermediate with hypertension-induced kidney disease, and worst with diabetic nephropathy [15,34]; five-year survival of the patient with diabetic nephropathy is only 20 percent [15,34].
Age — Survival declines with increasing age, with patients <45 years of age doing best [15,21,34,35]. Older patients with renal vascular disease appear to have the worst prognosis, with such patients having 5- and 10-year survival rates of only approximately 15 and 5 percent, respectively. In one prospective study from Canada, these older patients were frequently not offered dialysis as possible therapy [36,37].
Country — Survival among patients on dialysis has been substantially lower in the United States than in Europe and Japan [21,25,35,38,39]. Some evidence suggests that this is due in part to general differences in mortality (particularly that due to cardiovascular disease) among the general population in various countries [40,41]. In one European study that adjusted for age, sex, and diabetes, for example, approximately 25 percent of the difference in mortality among patients on dialysis between the northern and southern areas of Europe could be attributed to differences in mortality in the general population of these areas [41].
Differences in mortality in various countries and geographic regions therefore cannot be entirely explained by differences among patients that are unrelated to dialysis itself. These include variations among comorbid conditions (with increased incidence of diabetes, psychiatric diseases, and cardiovascular disease [heart failure and peripheral vascular disease] in the United States); patient age (which tends to be higher in the United States); racial or genetic differences; other demographic factors; and differences in transplantation rate (low rate in Japan) [25,39].
Other factors that may contribute to the differences in survival include variations in underlying kidney disease (such as increased incidence of immunoglobulin A [IgA] nephropathy in Japan) and differences in thresholds for accepting patients for dialysis [39]. Another factor may be differences in adherence to the dialysis schedule; in some countries such as Japan, nonmedical reasons for a missed treatment are highly uncommon, compared with the United States [42]. A decline in the relative mortality rates in the United States may reflect the consistent widespread improvement in the overall quality of the national dialysis prescription [15]. (See 'Improved survival with more intensive dialysis' below.)
Race — As a group, African Americans and Asian Americans have a lower mortality rate than White patients [15,43-45]. In one study, the survival rates of Black patients, White patients, and patients of other races at five years were 35, 25, and 32 percent, respectively. Similar relative results were observed in a single-center study, which reported 47 and 36 percent survival rates at five years for Black patients and White patients, respectively [46]. In initial studies, this survival advantage appeared to persist after adjustment for patient characteristics, comorbidities, and laboratory abnormalities [43,47].
However, in subsequent studies, the survival advantage was markedly attenuated or lost in multivariate analyses that adjusted for other patient variables [48,49]. As an example, in an analysis of outcomes of 1,330,007 patients identified in the USRDS, although the survival of Black patients as a group was better compared with White patients, this advantage was modified by age [49]. Whereas a survival advantage persisted for Black patients >50 years of age, Black patients <50 years of age had higher mortality compared with White patients, particularly when transplantation was factored in as a competing risk.
The underlying reasons for increased survival among older Black patients are unclear. One possibility is that mortality among Black patients with predialysis chronic kidney disease (CKD) is much higher than among White patients with predialysis kidney disease, resulting in the selection of relatively "healthier" patients who eventually require dialysis. This, in turn, would result in an apparent survival advantage among Black patients with ESKD. Support for this hypothesis was provided by a study of participants of the Third National Health and Nutrition Examination Survey (NHANES), in which a higher risk for death in the early stages of CKD was observed among Black patients [50]. Another possibility is that Black patients on dialysis may have a relatively greater use of activated vitamin D therapy, a modality indirectly linked with improved survival [51].
Psychosocial factors — Psychosocial factors also appear to significantly affect mortality, independent of the presence of other comorbid conditions. Increased levels of social support, enhanced behavioral compliance, and positive perceptions of the effects of illness are all associated with a decreased risk of dying [52]. Their effect on mortality appears to be equivalent to those of medical risk factors. (See "Psychiatric illness in adults receiving maintenance dialysis".)
Location of residence may alter survival. Residence in areas with higher median household income has been associated with improved survival and, among Black patients, residence in highly segregated areas with increased mortality [53].
Nutrition — There is greater medical risk and increased mortality in undernourished patients, particularly those with hypoalbuminemia. Although these observations have primarily been made in patients on maintenance HD and are related to the dialysis dose (see 'Malnutrition' below), the presence of malnutrition prior to the initiation of dialysis is strongly predictive of increased mortality with the onset of dialysis [54].
Salt intake — A post-hoc analysis of the HEMO Study including 1770 patients on HD has suggested that higher reported dietary sodium intake and the ratio of sodium to calorie or potassium intake are associated with increased all-cause mortality and a slightly increased ultrafiltration requirement [55]. Reported sodium intake was not associated with predialysis systolic blood pressure. Prescribed sodium restriction was not associated with mortality, systolic blood pressure, or ultrafiltration requirement in this study. This study is somewhat limited by the use of dietary recall, which may have underestimated both sodium and calorie intake [56].
Residual kidney function — Residual kidney function is associated with improved survival of patients on dialysis. Among patients on peritoneal dialysis (PD), the beneficial effect of continued urine output on survival has been shown in multiple studies. (See "Prescribing peritoneal dialysis".)
The following studies suggest that preserved kidney function is associated with improved survival among HD patients [57-60]:
●In a retrospective study of 650 patients on incident HD, mortality was significantly lower at 6, 12, and 24 months after dialysis was initiated among those with residual kidney function versus those without; this benefit was maintained after correcting for albumin, age, comorbidities, hemodiafiltration, and underlying kidney disease [59].
●In the Netherlands Cooperative Study on the Adequacy of Dialysis among 740 patients on HD, survival was directly related to preserved kidney function [58].
●Among 734 patients on incident HD, the presence of urine output at one year after dialysis initiation, compared with those without residual function, was significantly associated with a lower all-cause mortality (hazard ratio [HR] 0.70, 95% CI 0.52-0.93) and a trend toward decreased cardiovascular mortality (HR 0.69, 95% CI 0.45-1.05) [60].
The benefit of residual kidney function on cardiovascular survival is likely multifactorial. It may be via improved chronic regulation of fluid and electrolyte balance. Patients who have a preserved urine output may accumulate less fluid in the interdialytic period and, therefore, will have less time spent in a state of volume overload. They also will require less fluid removal during dialysis, which at times may occur too rapidly. Another possibility is that residual kidney function allows increased removal of protein-bound uremic toxins, some of which have a potential for adverse cardiovascular effects [61].
Others — The effect on mortality of a number of additional patient characteristics has been assessed, primarily in observational cohort studies; to what extent there is a direct cause-and-effect relationship between these characteristics and mortality amongst ESKD patients is largely unknown.
●Inflammation – Patients on dialysis with evidence of an ongoing inflammatory state have an enhanced mortality. This is discussed in detail separately. (See "Inflammation in patients with kidney function impairment" and "C-reactive protein in cardiovascular disease".)
●Serum cardiac enzymes – Elevations of cardiac troponin T (cTnT) among asymptomatic patients on chronic dialysis are associated with decreased survival. The exact source of cTnT in this setting and the reasons for the correlation between elevated levels and a poor prognosis are unknown. This is discussed in detail separately. (See "Cardiac troponins in patients with kidney disease" and "Clinical manifestations and diagnosis of coronary artery disease in end-stage kidney disease (dialysis)".)
●Sleep disorder – The presence of a sleep disorder enhances the risk of death in patients without kidney failure (see "Clinical presentation and diagnosis of obstructive sleep apnea in adults"). A paucity of data exists concerning their effect upon survival in patients undergoing maintenance dialysis. Limited observational evidence suggests that the presence of periodic limb movements in sleep, the total number of arousals, and a decline in sleep quality during the first year on dialysis may strongly and independently correlate with near-term mortality [62,63].
●Predialysis care – An increasing number of studies, although all observational and retrospective, suggest that patients referred late to a nephrologist for predialysis medical care, compared with those referred early in the disease course, have an enhanced mortality risk once dialysis is initiated. (See "Overview of the management of chronic kidney disease in adults".)
There is marked variability in the quality of predialysis care based upon the medical center and geographic area. In the United States, there also appears to be geographic clustering of centers with increased mortality and less-than-adequate predialysis care [64].
●Hemoglobin levels – A discussion of the role of low hemoglobin levels in association with mortality among patients with kidney disease is discussed in detail separately. (See "Treatment of anemia in patients on dialysis".)
There is also some evidence that variability in hemoglobin levels increases mortality in ESKD [65-67]. This was shown in a retrospective cohort study of 34,963 patients on HD in which survival analysis found that each 1 g/dL increase in hemoglobin variability (defined by the residual standard deviation) was associated with a 33 percent increase in the death rate [65].
●Fractures – Patients with ESKD have an increase in risk of bone fractures, with fractures of pelvis/hip being most common [68,69]. Among such patients, hospitalizations for fractures (particularly vertebral, shoulder/upper arm, or pelvis/hip) are associated with a relatively high one-year mortality rate [69-71]. In a study from the USRDS, patients on dialysis with a hip fracture had a twofold higher risk of death than those matched controls without a fracture [71]. Risk factors for falling include older age, diabetes, high number of prescribed oral drugs, antidepressant use, and inability to walk without significant help [70].
●Frailty – Frailty is commonly defined by the presence of three or more of the following: weight loss, muscle weakness, fatigue or exhaustion, low physical activity, and slow gait. Independent of age, increased mortality is noted among frail patients on dialysis. As an example, based upon data from 2275 adults from the dialysis Wave II study, a higher risk of death was noted among those defined as frail (adjusted HR 2.24, 95% CI 1.60-3.15) [72]. A subsequent meta-analysis also reported that frailty was associated with an increased risk of mortality [73].
●Nonadherence – Nonadherence to the dialysis regimen, which can be defined in part by regularly skipping HD sessions and/or poor adherence to dietary restrictions, is not uncommon [42]. It is associated with increased mortality [74]. In a study of 739 patients, of whom 67 were classified as skippers (absent >3 percent of treatments), increased mortality was associated with skipping (HR 1.69, 95% CI 1.23-2.3) [74].
●Physical activity – Habitual physical activity has been associated with decreased mortality among patients undergoing maintenance HD [75-78]. As an example, among 202 outpatients on HD, at a median follow-up of 45 months, mortality risk decreased with increased physical activity, as quantitated by an accelerometer (HR per 10 min/day increase in physical activity 0.78, 95% CI 0.66-0.92) [77]. A similar observation was made in the Comprehensive Dialysis Study, in which, among 1554 ambulatory dialysis patients, lower physical activity, estimated using the 94-item Human Activity Profile (HAP), was associated with increased mortality [78].
●Additional features – A large number of additional patient characteristics and/or agents are reportedly associated with mortality. These include periodontal disease, rosiglitazone, warfarin, clopidogrel, aspirin, and high blood lead levels [79-82].
The remainder of this topic review will focus upon those factors largely unique to the dialysis procedure that affect patient mortality.
ADEQUACY OF DIALYSIS — The higher mortality rate in the United States was thought to be related at least in part to inadequate dialysis. The weekly hemodialysis (HD) times declined progressively in the United States from 25 to 40 hours in the 1960s, to 12 to 15 hours in the 1970s and 1980s, to as low as 7 to 8 hours in the 1990s [83,84]. This decreasing time likely contributed to higher mortality, as suggested by the following:
●Studies in both Germany and the United States have documented the relationship between shorter dialysis time and poorer outcome [85,86]. Patients dialyzed <3.5 hours three times per week have approximately twice the mortality risk compared with patients dialyzed four or more hours three times per week [85].
●Historically, the group from Tassin, France has reported one of the best dialysis survival estimates of any program or registry, with a 15-year survival rate of 65 percent in 445 patients followed from 1970 through 1990 (figure 1) [35]. These patients, who were dialyzed very intensively (Kt/V of 1.67), also had a high incidence of full rehabilitation, and almost all patients were rendered normotensive on no antihypertensive medications.
●Similar benefits with longer times on dialysis have been observed among patients undergoing nocturnal HD and short daily HD. (See "Outcomes associated with nocturnal hemodialysis" and "Short daily hemodialysis".)
Historic overview of survival and Kt/V — The utilization of urea kinetic modeling and the Kt/V index was in part responsible for the decreasing HD time and concomitant increase in mortality in the United States. Urea kinetic modeling became widely applied after the results of the National Cooperative Dialysis Study (NCDS) were published [87-89]. The NCDS found that a Kt/V urea >0.9 with thrice weekly treatment provided an "adequate" dialysis prescription in that the patient failure rate of 13 percent was substantially below that in patients with a lower Kt/V. (See "Prescribing and assessing adequate hemodialysis".)
Although the NCDS included no patients with diabetes or those with other important comorbid conditions, a Kt/V of 1.0 was widely accepted as representing adequate dialysis. Shorter dialysis times were permitted as long as the goal Kt/V was attained. However, this goal was quite arbitrary since no attempt was made in the NCDS to determine whether a higher Kt/V might be associated with further clinical, biochemical, or rehabilitative improvement.
Improved survival with more intensive dialysis — Subsequent observational data suggested that a higher Kt/V was beneficial. Excellent survival rates have been reported in Tassin, France (Kt/V >1.6) and Minnesota, United States (Kt/V >1.3), where patients, even those in high-risk groups, were more intensively dialyzed [35,90]. Urea kinetic modeling was utilized, but attention was paid to all the details involved in delivering a more optimal dialysis prescription. This improvement in survival occurred despite a striking increase in comorbidity, which itself would have decreased survival [90].
Similar findings were reported in a retrospective analysis of 13,473 patients, in which patient outcome was compared with dialysis intensity. Patient survival was significantly reduced at urea reduction ratios (URRs) <60 percent (Kt/V <1.20) [91]. A definition of the URR can be found elsewhere. (See "Prescribing and assessing adequate hemodialysis".)
A number of observational studies were compatible with the hypothesis that more intensive dialysis leads to improved survival:
●In one study, increasing the mean Kt/V from 0.82 (pre-1988) to 1.33 led to a reduction in the gross mortality rate from 22.8 down to 9.1 percent per year [92]. This was associated with increases in protein catabolic rate from 0.83 to 1.0 (an index of protein intake (see "Protein intake in patients on maintenance hemodialysis")) and in plasma albumin concentration from 3.5 to 3.9 g/dL (35 to 39 g/L). Thus, enhanced nutrition probably contributed to the improvement in survival (see 'Malnutrition' below). However, increasing the degree of dialysis was associated with the use of biocompatible HD membranes; it is therefore possible that the observed benefits were due at least in part to avoidance of bioincompatible membranes. (See 'Hemodialysis membranes' below.)
●In another report, increasing the mean Kt/V from 1.18 (pre-1989) to 1.46 (and the URR from 61 to 70 percent) led to a reduction in the gross mortality rate from 22.5 down to 18.1 percent per year [93]. This improvement was achieved with standard cellulosic bioincompatible membranes.
●A national sample in the United States suggested a 7 percent reduction in mortality for each 0.1 unit increase in Kt/V (figure 2) [94]. Although no statistically significant benefit was noted above a Kt/V of 1.3, a projected trend favoring an increased benefit with increasing dialysis was observed above this dialysis level.
A difficulty concerning the optimal dialysis dose was that some [95,96], but not all [97], studies observed an increased relative risk (RR) of death among patients with extremely high values for Kt/V (>1.6) or the URR (URR between 75 and 79 percent). Since these values reflect increased urea clearance and/or an extremely low body volume (which may represent poor nutrition), the increased death rate may have been due to the effects of marked malnutrition.
Support for this hypothesis comes from a study that evaluated the correlation between mortality, dialysis dose, and body volume in over 3000 patients on dialysis [96]. Consistent with previous reports, patients with a Kt/V or URR value in the highest (fifth) quintile had an increased RR of death (19 percent greater than those in the third quintile). However, a higher Kt alone was not associated with an increased RR of death, as there was a 2 percent decrease in mortality risk per liter increase in clearance.
These observations, in combination with the finding that patients in the fifth quintile suffered from more severe protein-calorie malnutrition, suggest that the increased mortality associated with very high Kt/V or URR values may reflect poor nutrition. Such findings support the use of alternative measures of dialysis adequacy, such as the Kt alone. (See "Prescribing and assessing adequate hemodialysis".)
Based upon these observational studies, the optimal level of HD and the best measure of HD adequacy were therefore undefined. Statistical analysis of the previous studies, as well as others by Gotch et al, suggested that faulty methodology may have supported the conclusion that increasing dialysis equals increasing survival [98]. The above studies were also retrospective in design and ignored the wide variation in dialysis delivery.
Limitations of Kt/V — Urea kinetic modeling is the preferred method for measuring the dialysis dose. However, despite its value as a practical guide to dialysis treatment, Kt/V does not account for the following factors that appear to affect survival (see "Prescribing and assessing adequate hemodialysis", section on 'Limitations'):
●Clearance of different-sized solutes (such as middle molecules). (See "Uremic toxins".)
●Rapidity of fluid removal during HD. (See 'Rapid ultrafiltration' below.)
●Residual kidney function in patients on HD. In contrast to patients on peritoneal dialysis (PD), in whom residual kidney function is considered when assessing dialysis adequacy, most nephrologists do not include urinary clearance of urea when determining Kt/V. However, this may be changing. (See "Incorporating residual kidney function into the dosing of intermittent hemodialysis".)
HEMO Study — In an attempt to overcome these difficulties concerning survival and dialysis dose (as well as survival and low- and high-flux dialysis membranes), a large, prospective, clinical trial called the Hemodialysis (HEMO) Study was performed [99]. In this study, 1846 patients were randomly assigned to a standard or high dose of dialysis and a low- or high-flux dialyzer. The standard-dose goal was an equilibrated Kt/V of 1.05, which is equivalent to a URR of 65 percent or a single-pool Kt/V of 1.25. The high-dose goal was an equilibrated Kt/V of 1.45, which is the same as a ratio of 75 percent or a single-pool Kt/V of 1.65. The primary outcome was death from any cause, while the main secondary outcomes were the rate of all hospitalizations (but excluding those related to access) and the composite outcomes of the first hospitalization for a cardiac problem or death from any cause, the first hospitalization for an infectious cause or death, and the first decline of >15 percent of the serum albumin from baseline value or death.
The following results concerning dialysis dose were reported at a mean follow-up of 4.5 years:
●Impressive separation was obtained. The standard-dose group received an equilibrated Kt/V of 1.16, a single-pool Kt/V of 1.32, and a URR of 66.3 percent, while the high-dose group had values of 1.53, 1.71, and 75.2 percent, respectively.
●The risk of death from any cause, the primary outcome, was the same in the high- and standard-dose groups (RR 0.96 for high versus standard dose, CI 0.84-1.10).
●The risk of the main secondary outcomes (as previously delineated) was also the same for both dialysis doses.
●Subgroup analysis revealed a significant survival benefit for women receiving a high-dialysis dose (19 percent lower risk of death than women in the standard-dose group). However, men receiving high-dose dialysis had a 16 percent higher risk of death than those receiving standard-dose dialysis.
Subsequent analysis found that differences in dialysis dose also had no effect upon the incidence of infection-related deaths [100].
Thus, patients on HD on a three treatment per week schedule do not appear to derive a major benefit from a higher than standard dialysis dose (in terms of Kt/V urea), as recommended by the Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines. (See "Prescribing and assessing adequate hemodialysis" and 'Summary and recommendations' below.)
Some data suggest that mortality among patients in dialysis centers with the highest percentage of patients who exceed KDOQI guidelines (URRs of >65 percent) is lower than in those with fewer patients who achieve these ratios. As an example, based upon data from 2858 United States dialysis facilities from 1999 to 2002, those centers in the lowest quintile for the percentage of patients with higher than recommended URRs had 22 percent higher mortality rates than those in the highest quintile [101].
Peritoneal dialysis — Among patients treated with PD, a similar correlation between the degree of solute clearance (measured weekly either by Kt/V or the creatinine clearance) and survival was also postulated to exist. This hypothesis was supported by the results of a prospective, cohort, multicenter study in Canada and the United States (CANUSA) that evaluated the relationship of the adequacy of dialysis and nutritional status to mortality, morbidity, and technique failure.
However, survival in this study was related to residual kidney function (which changed over time) and not to peritoneal clearance alone (which did not change). Confirmation that an increase in peritoneal clearance alone (beyond a minimum threshold value) is not a principal factor underlying survival was provided by the Adequacy of Peritoneal Dialysis in Mexico (ADEMEX) trial. Instead, subsequent analysis of this trial found that measures of health-related quality of life significantly predicted patient survival [102]. (See "Prescribing peritoneal dialysis".)
HEMODIALYSIS MEMBRANES — Survival, dialysis adequacy, and the biocompatibility of hemodialysis (HD) membrane may all be closely interrelated. This is discussed separately. (See "Clinical consequences of hemodialysis membrane biocompatibility".)
Differences in survival may also be due to variations in the ability of HD membranes to remove molecules larger than urea, the so-called middle molecules. High-flux membranes are better able to remove such substances, compared with that observed with low-flux membranes. Whether survival differs upon exposure to high- and low-flux membranes was also evaluated in the Hemodialysis (HEMO) Study, in which nearly 2000 patients were randomly assigned to a standard or high dose of dialysis and a low- or high-flux dialyzer (based on clearance of beta-2-microglobulin) [99]. The primary outcome was death from any cause, while the main secondary outcomes were the rate of all hospitalizations (but excluding those related to access) and the composite outcomes of the first hospitalization for a cardiac problem or death from any cause, the first hospitalization for an infectious cause or death, and the first decline of >15 percent of the serum albumin from baseline value or death. The results and discussion relating to survival and dialysis dose can be found elsewhere in this topic review. (See 'HEMO Study' above.)
With respect to dialysis membranes, the following results were reported at a mean follow-up of 4.5 years:
●Impressive separation was obtained as the low- and high-flux groups had clearances of beta-2-microglobulin of 3 and 34 mL/min, respectively.
●The risk of death from any cause, the primary outcome, was the same in the high- and low-flux groups (relative risk [RR] 0.92).
●The risk of the main secondary outcomes (as previously delineated) was also the same for both flux groups. In the high-flux group, there were significant reductions, compared with the low-flux group, in the risk of death from cardiac causes and the combined outcome of first hospitalizations or death from a cardiac cause [28]. However, as previously mentioned, total mortality was the same in both flux groups.
●Subgroup analysis revealed a significant survival benefit for patients with >3.7 years of dialysis receiving high-flux dialysis (32 percent lower risk than the low-flux group). However, men receiving high-dose dialysis had a 16 percent higher risk of death than those receiving standard-dose dialysis, and the strength of the interaction between flux and years of dialysis was weakened when years of dialysis was treated as a continuous variable.
Subsequent analysis found that differences in dialysis membrane also had no effect upon the incidence of infection-related deaths [100].
Thus, as noted in the HEMO Study, survival was the same in patients chronically dialyzed with either high- or low-flux dialysis membranes. This was also found in the Membrane Permeability Outcome (MPO) study, in which 738 patients on incident HD, separated by serum albumin levels ≤4 and >4 g/dL, were randomly assigned to high- or low-flux dialysis membranes [103]. Overall, there was no difference in survival with either membrane, although there was a nonsignificant trend for increased survival with high flux (RR 0.76, 95% CI 0.56-1.04). Among the subgroup of patients with serum albumin ≤4 g/dL, survival was significantly higher in the high-flux group (hazard ratio [HR] 0.63, 95% CI 0.45-0.90). This is less likely to be a spurious finding, given that this was a prespecified subgroup [104]; nevertheless, this result was not observed in the HEMO trial and should be confirmed by additional studies.
Although not directly applicable to the chronic setting, the RR for mortality (and recovery of kidney function) also appears to be the same with both high- and low-flux membranes among those with acute kidney injury [105].
HEMODIALYSIS SESSION LENGTH — The effect of hemodialysis (HD) session length on mortality independent of conventional markers of dialysis adequacy is unclear among patients undergoing standard, three times per week dialysis therapy. Some observational studies of large databases suggest that increased session length improves survival [12,106,107]:
●Based upon data from the Dialysis Outcomes and Practice Patterns Study (DOPPS) of 22,000 patients on HD, treatment time that was >240 minutes per session was significantly associated with decreased risk of mortality (relative risk [RR] 0.81) [12]. In addition, a 7 percent lower RR of death was observed with each additional one-half-hour time on dialysis.
●In a study of 6593 patients in the Australian/New Zealand registry, the lowest mortality risk was observed in patients with HD session lengths of 4.5 to 4.9 hours [106].
These studies did not control for body size, which is usually a determinant of session length. However, support for a role of session length independent of body size was provided by an analysis of data from a large, outpatient dialysis organization that showed that treatment time <240 minutes per session was associated with increased mortality independent of body size (hazard ratio [HR] 1.26) [108].
A discussion of the effect of prolonged dialysis session length on outcomes, including survival, with nocturnal HD is presented separately. (See "Outcomes associated with nocturnal hemodialysis".)
CONTROL OF FLUID BALANCE AND HYPERTENSION — It is important to determine an accurate dry weight and to achieve this target weight. Chronic fluid overload is associated with increased mortality [109-113].
A large observational study showed significant associations between both baseline and one-year cumulative fluid overload and increased mortality among outpatients on incident hemodialysis (HD) [110]. Patients were stratified by systolic blood pressure (<130, 130 to 160, and >160 mmHg), and bioimpedance spectroscopy was used to assess fluid status. Fluid overload was defined as ≥15 percent fluid excess for men and ≥13 percent fluid excess for women, and the first measurement was obtained within three months of starting HD. Both fluid overload at initiation and one-year cumulative fluid overload predicted increased mortality across all blood pressure groups. Cumulative fluid overload had the larger effect on mortality (for patients with systolic blood pressure <130 mmHg, hazard ratio [HR] 1.94; for patients with systolic blood pressure 130 to 160 mmHg, HR 1.51; for patients with systolic BP >160 mmHg, HR 1.62). The effect of fluid overload was similar in all groups as stratified by age, sex, body mass index (BMI), and comorbidities including diabetes and heart disease.
Although this study does not suggest a causal relationship between fluid overload and mortality (cause-specific mortality was not determined), this study and others support a detrimental effect of fluid overload [114].
Adequate fluid removal depends upon the technical ability to remove excess fluid, the patient's tolerance to the fluid removal rate, and the patient's adherence to fluid-intake restrictions. Shorter dialysis necessitates more rapid fluid removal, which is more likely to induce episodes of hypotension. Hypotension is associated with transient myocardial ischemia and subsequent cardiac stunning that then needs to be treated with fluid administration. The net effect is that true dry weight is never achieved, leading to persistent volume expansion, hypertension, and possible deterioration in left ventricular function. These complications may contribute to the high cardiac mortality in patients on dialysis. (See "Intradialytic hypotension in an otherwise stable patient".)
Dry weight has been defined as not merely the absence of edema but the minimal body sodium content and volume of body water that can be tolerated without inducing hypotension. Gradual fluid removal with longer dialysis permits a greater degree of total fluid loss and more likely attainment of dry weight. As an example, slow dialysis, such as with nocturnal HD, commonly results in normotensive patients who do not require the use of antihypertensive medications. The ability to discontinue drug therapy may be an important component of this regimen since it further minimizes the risk of hypotension as the excess fluid is removed.
The ability to remove adequate fluid also depends upon the ability of the patient to adhere to fluid-intake restrictions. Some data suggest that psychologic interventions may improve the ability to adhere to such restrictions, although the quality of the medical literature is generally poor [115].
Inadequate ultrafiltration — Inadequate total body fluid removal may increase the overall risk of death. In one study of over 3000 patients on HD, the relative risk (RR) of death was correlated with ultrafiltration status [116]. In this study, soft-tissue hydration was determined via bioelectric impedance analysis and impedance vector length; a shorter vector length, which was associated with inadequate ultrafiltration, was associated with an increased RR of death. A similar correlation between survival and ultrafiltration was observed in a study from the Netherlands [58].
Better volume management via intradialytic blood-volume monitoring may improve outcomes, including mortality:
●In a six-month study in which 443 patients were randomly assigned to an intradialytic blood-volume monitoring device and conventional monitoring, mortality and non-access-related hospitalizations were significantly higher with the device [117]. This result may have been due in part to the atypically low hospitalization and mortality rates for the conventionally monitored group.
●The assessment of volume status by bioimpedance spectroscopy was compared with that by routine clinical practice (ie, control group) among 156 randomly assigned patients [118]. At one year, the left ventricular mass index, the left atrial volume index, and blood pressure values decreased in the bioimpedance spectroscopy group but not in the control group.
Rapid ultrafiltration — High ultrafiltration rates have been associated with increased cardiovascular mortality [13,119,120]. All studies found a higher mortality in patients undergoing high rates of ultrafiltration (eg, an ultrafiltration volume >5.7 percent of postdialysis weight [120] or ultrafiltration rates >13 mL/hour/kg [13,119]) (see "Evaluation of sudden cardiac arrest and sudden cardiac death in patients on dialysis"). This may be due, in part, to repeated episodes of transient myocardial ischemia/stunning that can occur even in patients with normal coronary arteries [13,14]. Based on these mortality data, the Centers for Medicaid and Medicare Services adopted a quality measure targeting an ultrafiltration rate ≤13 mL/hour/kg. This ultrafiltration rate target may complicate the treatment of patients who gain excessive fluid between treatments but has been associated with fewer episodes of intradialytic hypotension [121].
Additional study is required to determine the optimal ultrafiltration strategy for thrice weekly HD. Ultrafiltration rate–associated harm may occur at levels less than 13 mL/hour/kg [12,119,122], and observational data suggest that ultrafiltration rate thresholds scaled to weight may lead to worse outcomes in patients with higher body weight [123,124]. (See "Prescribing and assessing adequate hemodialysis", section on 'Patient-specific parameters'.)
INTERDIALYTIC INTERVAL — Long interdialytic intervals that occur over a weekend among patients who are dialyzed on a three times weekly schedule likely contribute to mortality and morbidity. This was suggested by a retrospective analysis of 32,065 patients on hemodialysis (HD) who participated in the End-Stage Renal Disease Clinical Performance Measures Project [125]. This analysis demonstrated an increase in all-cause mortality on the day following the long interval, compared with other days (22.1 versus 18.0 deaths per 100 person-years). After the long interval, there was also an increase in the following (per 100 patient-years):
●Mortality from cardiac causes (10.0 versus 7.5)
●Mortality from infection (2.5 versus 2.1)
●Mortality from cardiac arrest (1.3 versus 1.0)
●Mortality from myocardial infarction (6.3 versus 4.4)
●Admissions for myocardial infarction (6.3 versus 3.9)
●Congestive heart failure (29.9 versus 16.9)
●Stroke (4.7 versus 3.1)
●Dysrhythmia (20.9 versus 11.0)
●Any cardiovascular event (44.2 versus 19.7)
The authors were unable to determine whether adverse events occurred before, during, or after the first weekly dialysis sessions.
These data suggest that the two-day interval is a time of increased risk for cardiovascular and noncardiovascular mortality and morbidity for patients who are on a three times weekly dialysis schedule. Potential reasons for the increased risk were not addressed in this study, but fluid accumulation, hemodynamic changes, and electrolyte abnormalities or changes may contribute, at least to the cardiovascular events.
MALNUTRITION — Studies have demonstrated greater medical risk and increased mortality in undernourished and small patients on dialysis (as noted by a low body mass index [BMI]), with hypoalbuminemia and lower plasma levels of urea nitrogen and creatinine than expected from the intensity of dialysis [91,126-128]. As an example, a plasma albumin concentration <4 g/dL (40 g/L) is the single laboratory finding most closely associated with an increased probability of death. The increase in risk is dose dependent, rising modestly at a plasma albumin concentration of 3.5 to 3.9 g/dL (35 to 39 g/L), but being much greater at values <3.0 g/dL (30 g/L) (figure 3) [91,126,127].
These observations have primarily been made in patients on maintenance hemodialysis (HD) and may overestimate the effect of hypoalbuminemia since some measurements taken in the weeks to months before death were included. However, a similar relationship between hypoalbuminemia and increased mortality has been demonstrated with peritoneal dialysis (PD) and, although less pronounced than in the figure, with plasma albumin concentrations measured within six weeks of starting dialysis (figure 4) [129]. (See "Nutritional status and protein intake in patients on peritoneal dialysis".)
The presence of malnutrition prior to the initiation of dialysis is also strongly predictive of increased mortality [54]. Despite this, the presence of associated conditions, particularly comorbid conditions, significantly contributes to mortality, independent of nutritional status. In one study, for example, the outcomes of a total of 153 consecutive patients on PD were analyzed based upon the presence or absence of malnutrition, with or without comorbid conditions [130]. The combined presence of malnutrition plus a comorbid condition was associated with a high mortality, while there is a trend with malnutrition alone for an increase in mortality.
Conversely, measures of good nutritional status, such as high BMI in combination with normal or high muscle mass, are associated with increased survival [131,132]. As an example, in a study of patients on incident PD, the best survival was observed in those with high BMI and normal or high muscle mass [131].
As with hypoalbuminemia, a similar incidence of undernutrition has been noted with the use of creatinine production to assess lean body mass in stable patients on maintenance dialysis. In one study, estimated lean body mass was below normal in 47 and 66 percent of patients on HD and PD, respectively [133]. There was a significant correlation between lean body mass and the plasma albumin concentration in the group as a whole; however, there was substantial variability in individual patients.
Another marker of malnutrition, a low ratio of body weight for height, also predicts mortality in patients on HD. In an analysis of data from nearly 13,000 patients, there was a highly significant inverse relationship between mortality rates and the ratio of body weight for height among those with ratios below the 50th percentile [134].
The association between markers of malnutrition and reduced patient survival does not prove a cause-and-effect relationship. Nevertheless, it constitutes the rationale for attempting to improve nutrition via diet and more intensive dialysis. (See "Pathogenesis and treatment of malnutrition in patients on maintenance hemodialysis".)
The meaning of a low plasma albumin concentration may also differ among patients on HD and PD due in part to variations in volume status. The serum albumin among patients on HD is often measured predialysis, a time when volume expansion is maximal; by comparison, the concentration of albumin in the PD population should not vary acutely. Differences in mortality among these two modalities that appear to correlate with albumin levels may therefore instead be due to variations in intravascular volume.
Although poor nutrition is a common cause of hypoalbuminemia in patients on dialysis, a number of chronic inflammatory conditions can also reduce the serum albumin concentration [135]. This effect appears to be mediated by cytokines, such as interleukin-1 and tumor necrosis factor-alpha, which suppress hepatic albumin synthesis [136]. Thus, the presence of an untreated, underlying inflammatory process may contribute to the decreased survival observed among patients on chronic HD with hypoalbuminemia. (See "Inflammation in patients with kidney function impairment".)
Unlike the general population, some, but not all, studies have shown that being overweight (perhaps a surrogate for "overnutrition") may enhance survival among patients on HD [132,137-143]. In one study of over 1300 patients on HD, every one-unit increase in the BMI was associated with a reduction of 30 percent in the relative risk (RR) of dying [137]. This result may be confounded in part by race and sex since African American females have the highest BMI [144].
The apparent survival benefit conferred by higher body weight may be limited to older adults (>65 years of age). This was suggested by a prospective analysis of 1749 patients who were stratified by age and BMI [145]. Among 984 patients <65 years of age, the age-standardized mortality rate was 1.7 times higher for patients with a BMI of ≥30 kg/m2 compared with those with a normal (20 to 24 kg/m2) BMI. By contrast, among 765 patients >65 years of age, there was no difference in mortality between individuals in these BMI groups.
In addition, increasing height itself may contribute to mortality. A prospective study that was conducted from 2001 until 2006 and included 117,644 patients on HD with BMI values ≥12 to ≤60 kg/m2 suggested that height, adjusted for weight, was directly associated with all-cause mortality and with cardiovascular, infection-related, and cancer mortality [146].
Obesity has been associated with worse survival outcomes among patients on PD in other studies [140,147].
CHRONIC HEMODIALYSIS ACCESS — Arteriovenous fistulas have advantages over arteriovenous grafts and central venous catheters, with some evidence suggesting a survival benefit with fistulas. (See "Approach to the adult patient needing vascular access for chronic hemodialysis".)
DISORDERS OF MINERAL METABOLISM — A number of reports have delineated an increased risk of all-cause and cardiovascular mortality in patients with disorders of mineral metabolism [148-152]. Although not found in all studies, the association with decreased survival primarily involves increased phosphate, calcium, calcium x phosphate product, and/or parathyroid hormone (PTH) levels. These in turn may be associated with accelerated atherosclerosis, arterial calcification, and an increased risk of adverse cardiovascular outcomes and death.
This was shown in a retrospective analysis of a database of over 40,000 patients on hemodialysis (HD) in which the relative risks (RRs) of death were determined for multiple categories of serum phosphate, calcium, calcium x phosphate product, and PTH [148]. After multivariate adjustments, the lowest RRs of death were observed among patients with serum phosphate concentrations between 3 and 5 mg/dL (0.97 and 1.61 mmol/L), serum calcium concentrations <8.0 mg/dL (2.0 mmol/L), a calcium x phosphate product of <45 to 50 mg2/dL2, and intact PTH <600 pg/mL (600 ng/L).
Concentrations of these substances that were above each of these levels were significantly associated with decreased survival. Compared with serum phosphate concentrations between 4.0 and 5.0 mg/dL (1.29 and 1.61 mmol/L), for example, an increased RR of death was noted for serum phosphate levels >5.0 mg/dL (1.61 mmol/L): The RR increased gradually from 1.07 to 2.02 at serum phosphate levels of 5.0 to 6.0 and ≥9.0 mg/dL, respectively.
Given that the calcium result may have been confounded by the tendency of calcium levels to increase or decrease with reciprocal changes in serum phosphate, the risk of death was also determined for calcium levels measured within serum phosphate categories of 4 to 5, 5 to 6, 6 to 7, and 7 to 8 mg/dL. Within each phosphate category, higher calcium levels were associated with significant increases in mortality risk.
Analyses of similar large databases have also found that increased levels of calcium, phosphate, and/or PTH had significant adverse effects upon survival [149-151,153,154]. As examples:
●Among 25,588 patients participating in the Dialysis Outcomes and Practice Patterns Study (DOPPS), the highest mortality was noted for those with calcium levels >10.0 mg/dL, phosphorus levels >7.0 mg/dL, and PTH levels >600 pg/mL [154].
●Among 1629 patients on incident HD and peritoneal dialysis (PD) from the Netherlands, phosphate and calcium x phosphate levels above Kidney Disease Outcomes Quality Initiative (KDOQI) guideline levels were associated with increased mortality risks in both patient groups [151].
However, other studies have noted that low PTH and calcium levels are associated with an increase or no effect on mortality [155-158]. A possible explanation for these disparate findings includes variations in study design, differences in population evaluated, and the use of only single measurements of calcium, phosphate, and PTH.
To better assess the association between mortality and changes in calcium, phosphate, and PTH over time, two studies have been performed to assess survival compared with differences over time in changes in bone mineral elements [159,160]. One prospective study, for example, was performed of 1006 patients on incident HD and PD in whom longitudinal changes in bone mineral parameters were assessed [159]. PTH levels decreased over the course of one year, while calcium and phosphate levels and the calcium x phosphate product increased over the first six months and then remained stable. After adjustment for comorbidity and other confounding variables, an increased risk of death was observed with the highest quartile for phosphate levels, using both baseline and time-dependent values (hazard ratio [HR] 1.57, 95% CI 1.07-2.30). By comparison, an increased mortality risk with the highest quartile for calcium, calcium x phosphate product, and PTH levels was only observed with the time-dependent values.
In addition to phosphate, calcium, calcium x phosphate product, and PTH levels, additional serum laboratory values related to bone disease are associated with survival among dialysis:
●Fibroblast growth factor-23 (FGF-23) is an osteoblastic hormone involved in the regulation of phosphate and vitamin D. With increasing phosphate levels in those with increasing kidney dysfunction, elevations in FGF-23 levels stimulate the renal excretion of phosphate and inhibit the synthesis of 1,25-dihydroxyvitamin D. As with increased phosphate levels, elevated FGF-23 levels are associated with increased mortality among patients with end-stage kidney disease (ESKD). This was shown in a nested case-control study of 200 patients who died and 200 patients who survived the first year of HD [161]. Based upon multivariable analysis, increasing FGF-23 levels were associated with an increasing risk of death when analyzed via either a continuous scale or quartiles. As an example, the risk of death was significantly higher in patients with the highest FGF-23 levels (quartile 4) compared with the group with the lowest FGF-23 levels (quartile 1) (odds ratio [OR] 5.7, 95% CI 2.6-12.6).
FGF-23 has also been associated with cardiovascular events and mortality in predialysis chronic kidney disease (CKD) patients [162,163], kidney transplant patients [164], and non-CKD patients [165]. Further studies are needed to define the role of FGF-23.
●Increased mortality is described in patients on dialysis with elevated serum alkaline phosphate levels [166,167]. In one study, levels of alkaline phosphatase ≥120 units/L were associated with an increased risk of death (HR 1.25, 95% CI 1.21-1.29) [166].
These observational studies suggest that the many interventions utilized to maintain target calcium, phosphate, and PTH levels must be adequately assessed in rigorous trials to help determine their effect upon survival in patients with ESKD. These interventions include low-phosphate diets, calcium- and non-calcium-based phosphate binders, vitamin D analogs, and/or calcimimetics.
Some of these issues are discussed in detail separately:
●(See "Management of hyperphosphatemia in adults with chronic kidney disease".)
●(See "Management of secondary hyperparathyroidism in adult patients on dialysis".)
●(See "Vascular calcification in chronic kidney disease".)
ACHIEVING MULTIPLE CLINICAL PERFORMANCE MEASURES — Meeting specific goals for some clinical parameters has a beneficial impact upon survival in dialysis patients. Prominent examples include attaining and maintaining target levels for either hemodialysis (HD) dose or hemoglobin concentration. (See 'Adequacy of dialysis' above.)
Practically all studies that have examined the effect upon survival of achieving clinical performance goals have evaluated benefits due to a single measure of clinical care. There is evidence that achieving multiple clinical targets may be associated with superior survival outcomes versus those observed with attaining fewer clinical targets or only one [168-173]. As examples:
●Among nearly 200,000 patients, the risk of death during the first year of HD correlated with having met zero, one, two, or three of the following guideline goals at initiation of HD: use of an arteriovenous fistula or graft at initiation, hemoglobin level ≥11 g/dL, and goal albumin level [173]. Compared with those who met no goal, the risk of death was significantly lower among those who met one (0.81, 0.80-0.83), two (0.53, 0.41-0.56), or three goals (0.34, 0.30-0.39). A lower mortality rate was also observed in those who met each individual goal.
●During a 12-month follow-up period in a retrospective study of 15,287 patients on HD, death or hospitalization were compared with achieving target levels for hemoglobin (≥11 g/dL), serum albumin levels (3.7 g/dL by bromocresol purple laboratory method), use of a fistula for vascular access, and a single-pool Kt/V urea of ≥1.2 [168]. Seven, 24, 39, 24, and 6 percent of patients met four, three, two, one, and zero targets, respectively. There was a significantly increased risk of death and hospitalization among those in whom targets were not met, which increased with each specific target that was not achieved. Compared with meeting all four targets:
•Adjusted hazard ratios [HRs] for death were 1.9 (95% CI 1.4-2.6), 2.6 (95% CI 1.9-3.5), 3.5 (95% CI 2.6-4.7), and 4.6 (95% CI 3.3-6.4) for those who met three, two, one, and zero targets, respectively.
•Adjusted HRs for hospitalization were 1.1 (95% CI 0.98-1.3), 1.3 (95% CI 1.1-1.5), 1.5 (95% CI 1.3-1.7), and 1.6 (95% CI 1.4-1.9), respectively.
●A retrospective analysis of 13,792 patients on incident HD evaluated the correlation between survival and achieving Kidney Disease Outcomes Quality Initiative (KDOQI) clinical practice guidelines for single-pool Kt/V, hematocrit, serum albumin, calcium, phosphate, and parathyroid hormone (PTH) level [170]. Compared with not achieving any one target, the adjusted HRs for mortality were 0.11 (95% CI 0.06-0.19), 0.19 (95% CI 0.15-0.25), 0.30 (95% CI 0.24-0.37), 0.39 (95% CI 0.31-0.49), 0.52 (95% CI 0.42-0.65), and 0.74 (95% CI 0.49-0.93) for those who met six, five, four, three, two, and one targets, respectively. By comparison, an increased mortality was associated with achieving the blood pressure goal (predialysis <140/90 mmHg; 1.90, 95% CI 1.73-2.10).
These studies are limited in part because of their design, underreporting of comorbid conditions, and unknown incidence of disease severity. However, these results are generally consistent with other, primarily observational studies showing increased risk of death and hospitalization with decreased dialysis dose or hemoglobin levels, low serum albumin concentration, abnormal mineral metabolism, and HD with a vascular access other than an arteriovenous fistula. This suggests that quality improvement measures should be implemented to increase the number of those who meet multiple performance measures.
MODALITIES OF KIDNEY REPLACEMENT THERAPY
Hemodialysis versus peritoneal dialysis — The data described above were almost entirely from patients treated with maintenance hemodialysis (HD). Survival of patients on HD versus peritoneal dialysis (PD) is presented in detail separately. (See "Dialysis modality and patient outcome".)
With PD, there is increasing evidence that increased survival is associated with centers that are relatively more experienced with PD [174-176]. In one observational study of outcomes in large dialysis organizations, an increased mortality risk was observed in patients being treated in organizations with the lowest rates of PD use [176].
Home hemodialysis — Home HD is generally associated with higher survival rates compared with in-center HD. Improved survival is noted despite adjustments for age, diagnosis, comorbidity, and sex compared with in-center HD. (See "Choosing home hemodialysis for end-stage kidney disease".)
Alternative forms of hemodialysis — Although undergoing intense study, no significant conclusions can be drawn concerning the effect on survival of nocturnal HD or short daily HD. (See "Technical aspects of nocturnal hemodialysis" and "Short daily hemodialysis".)
Hemodiafiltration — HDF is not commonly used in the United States but is readily available in some countries. In contrast to conventional HD, HDF is a form of kidney replacement therapy that predominantly uses convective clearance along with some diffusive clearance to remove retained solutes. As such it has the potential to remove more middle molecules. The mortality associated with HDF compared with HD is discussed in detail separately. (See "Alternative kidney replacement therapies in end-stage kidney disease", section on 'Mortality'.)
SUMMARY AND RECOMMENDATIONS
●Overview – Although maintenance dialysis prevents death from uremia, mortality among patients with end-stage kidney disease (ESKD) remains high. Patients on dialysis have a substantially higher multivariable-adjusted mortality than patients not on dialysis who have cancer, diabetes, or cardiovascular disease. However, mortality rates among patients with ESKD have improved over time. (See 'Overview' above.)
●Causes of death – Among patients with ESKD, cardiovascular disease accounts for approximately 50 percent of deaths. Infections are the second most common cause of death, followed by withdrawal from dialysis. (See 'Causes of death' above.)
●Risk factors not related to dialysis – A large number of risk factors that are unrelated to the dialysis procedure have been associated with decreased survival among patients on dialysis. These include the following:
•Comorbid conditions such as diabetes and cardiovascular disease (see 'Comorbid conditions' above)
•Type of underlying kidney disease (see 'Underlying kidney disease' above)
•Lack of residual kidney function (see 'Residual kidney function' above)
•Malnutrition (see 'Malnutrition' above)
•Frailty (see 'Others' above)
•Disorders of mineral metabolism (see 'Disorders of mineral metabolism' above)
●Risk factors related to dialysis – Factors associated with the dialysis procedure also correlate with decreased survival. These factors include:
•Inadequate dialysis (see 'Adequacy of dialysis' above)
•Shorter hemodialysis (HD) session time (see 'Hemodialysis session length' above)
•Inadequate total body fluid removal (see 'Inadequate ultrafiltration' above)
•Rapid ultrafiltration rates (see 'Rapid ultrafiltration' above)
•Long interdialytic intervals (see 'Interdialytic interval' above)
●Benefit of multiple clinical performance measures – Most studies of patients on dialysis that have examined the effect upon survival of achieving clinical performance goals have evaluated benefits due to a single measure of clinical care. There is evidence that achieving multiple clinical targets may be associated with superior survival outcomes versus those observed with attaining fewer clinical targets or only one. (See 'Achieving multiple clinical performance measures' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges William Henrich, MD, MACP and Lionel U Mailloux, MD, FACP, who contributed to earlier versions of this topic review.
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟
