Adverse Effects of Conventional Thrice-Weekly Hemodialysis: Is It Time to Avoid 3-Day Interdialytic Intervals?

Dialysis treatment is scheduled thrice weekly ever since the inception of hemodialysis as a form of renal replacement therapy for patients with end-stage renal disease (ESRD). Due to reasons not related to health and disease status, but rather to convenience, hemodialysis patients remain without dialysis for a longer interval (∼3 days) prior to the first dialysis session of the week. A growing body of observational data supports the hypothesis that morbid and mortal events occur more frequently during the last day of the long interdialytic interval than during any other day of the week [1,2,3]. If this risk is causally related to the 3-day interdialytic interval, then there is a need to evaluate timing and frequency of prescribed hemodialysis regimens [4].

We review the evidence associating the long interdialytic interval of thrice-weekly hemodialysis with heightened risk of morbidity and mortality and summarize the pathophysiology of this potentially harmful effect. To mitigate the risk of long interdialytic intervals and the potential benefits of their implementation in clinical practice, we discuss alternative dialysis practices. We conclude with some recommendations for research.

Large registry-based cohort studies [1,2,3,5,6,7,8] have consistently linked intermittent dialysis schedules with elevated risk of death and cardiovascular-related complications (table 1). In the first study, Bleyer et al. [5] retrospectively evaluated mortality data of the USA dialysis population during 1977-1997 (i.e. patients either on hemodialysis or peritoneal dialysis), enrolled in the United States Renal Data System (USRDS), as well as data from 7,096 patients participating in the USRDS Case Mix Adequacy Study (CMAS). The investigators observed that among hemodialysis patients, Mondays were associated with increased risk of sudden and cardiac death (18.4% of sudden deaths and 18.1% of cardiac deaths occurred on Mondays relative to 14.3% expected, p < 0.00001) [5]. The next most common day of occurrence of sudden and cardiac deaths was Tuesday (15.4% relative to 14.3% expected, p < 0.00001). In contrast, deaths from non-cardiac causes in hemodialysis patients and deaths from both cardiac and non-cardiac causes in patients receiving peritoneal dialysis were evenly distributed throughout the week. When analysis was restricted to the CMAS participants, for whom data on dialysis schedules were available, it was shown that for patients being on a Monday-Wednesday-Friday schedule, compared to 14.3% expected deaths, deaths were more on Mondays: sudden deaths were 20.8% and cardiac deaths 20.2% [5]. A similar trend was found on Tuesdays for patients being on a Tuesday-Thursday-Saturday (TTS) schedule.

Three subsequent cohort studies evaluating morbidity and mortality data from various populations provide additional support to the potential link of 3-day interval with heightened cardiovascular risk [1,2,3]. A retrospective analysis of 32,065 hemodialysis patients participating in the End-Stage Renal Disease Clinical Performance Measures Project in the USA during 2004-2007 [1] showed that, compared to any other day of the week, the first dialysis day of the week was associated with higher risk of all-cause mortality (22.1 vs. 18.0 deaths per 100 person-years of follow-up, p < 0.001) and cardiovascular mortality (10.2 vs. 7.5, p < 0.001) [1]. A similar day-of-week pattern was evident for cardiovascular-related hospital admissions (rates per 100 person-years: 44.2 vs. 19.7, p < 0.001).

A post hoc analysis of data from 22,163 hemodialysis patients participating in Dialysis Outcomes and Patterns Practice Study (DOPPS) [3] allowed comparisons from around the world. In the USA, Mondays were associated with 41% higher risk of death relative to the 7-day average (p < 0.001); in Europe, the death risk was 34% higher (p = 0.001). In contrast, for Japanese patients, a 27% higher risk of death was noted on Mondays, but this did not reach statistical significance compared to the 7-day average (p = 0.154) [3]. A similar trend for the day-of-week mortality was also evident in patients who were on a TTS dialysis schedule, as Tuesdays were associated with a significant 39, 22 and 41% higher risk of all-cause mortality relative to the 7-day average in the USA, European and Japanese patients, respectively. In comparison to the non-cardiovascular deaths, the association between day-of-week mortality and dialysis schedule was stronger in the case of cardiovascular mortality.

Using the Australian and New Zealand Dialysis and Transplant (ANZDATA) Registry, the day-of-week and mortality relationship was explored in 14,636 ESRD patients who died between 1999 and 2008 [2]. This study showed that cardiac deaths were more likely to occur on Mondays than on any other day of the week in hemodialysis patients receiving ≤3 dialysis treatments per week (odds ratio (OR) 1.26, 95% confidence intervals (CI) 1.14-1.40, p < 0.001). Similarly, Mondays were also associated with 89% higher risk of death related to hyperkalemia (OR 1.89, p = 0.008) and 17% higher risk of sudden cardiac death (OR 1.17, p = 0.004) compared to any other day of the week. The next most common day of mortality occurrence was Tuesday (OR 1.08, 95% CI 0.97-1.20, p = 0.10). In contrast, mortal events were evenly distributed across the week in patients receiving >3 dialysis treatments per week, in patients receiving home hemodialysis and in patients receiving peritoneal dialysis [2].

Many of the above studies were limited by the fact that the exact timing of events was not available. This did not allow study investigators to examine whether deaths and cardiovascular complications had occurred before, during or after the first weekly dialysis session. A retrospective analysis of 80 cases of sudden cardiac death during 1995-2003 in 5 dialysis units in the USA addressed this issue; the last 12-hour period of the 3-day interval (i.e. the period of the highest accumulation of metabolic and volume-related parameters) was associated with a 3-fold elevated risk of sudden death [6].

A limitation of all the above studies is that no cause-and-effect association between day-of-week mortality and dialysis scheduling is possible, given their retrospective and observational design.

The above findings are in line with the ‘Peak Concentration Hypothesis', which was introduced by Keshaviah et al. [9] in 1989. This hypothesis suggested that the peaks in concentrations of urea may raise the risk of overt uremic toxicity in patients receiving intermittent thrice-weekly hemodialysis, despite the fact that both hemodialysis and peritoneal dialysis patients may have similar time-averaged urea concentration for the same dialysis adequacy [9]. Despite absence of direct evidence that urea per se is toxic [10], the principles of this theory may be relevant to shifts in metabolic and volume-related parameters between dialysis treatments, generating hypotheses for the mechanistic background of increased risk of morbidity and mortality associated with the long interdialytic interval. The most important of these mechanisms are depicted in figure 1 and are summarized below.

Among hemodialysis patients, elevated blood pressure (BP), particularly when recorded outside of the dialysis unit either with home or with ambulatory BP monitoring, is a strong cardiovascular risk factor [11,12]. Unlike the general hypertensive population with normal renal function or people with early stage chronic kidney disease, patients on maintenance hemodialysis display markedly impaired patterns and rhythms of BP during the interdialytic interval. Among hemodialysis patients, BP steadily increases during the interdialytic interval and the rate of BP increment is directly proportional to the interdialytic weight gain [13,14]. Notably, the impact of interdialytic weight gain is superimposed on the circadian variation of BP. In a study investigating the chronobiology of BP rhythms over the 3-day interdialytic interval with the use of home BP monitoring, it was shown that interdialytic increase in systolic and diastolic BP was not linear and that it reached a plateau after the first 48 h elapsed from dialysis [15]. In addition, the increment in systolic BP was greater than that in the diastolic BP, resulting in progressive amplification of pulse pressure during the interdialytic interval. This time-dependent amplification of pulse pressure suggests that patients receiving conventional thrice-weekly hemodialysis may accumulate a potent additional cardiovascular risk factor during the interdialytic interval that would be most pronounced during the third day after dialysis [15].

In addition to elevated interdialytic BP, excess volume accumulation over the long interdialytic interval also influences hemodynamic stability during dialysis [16]. This is due to the need for more aggressive ultrafiltration during the first dialysis session of the week, resulting in greater shifts of volume and pressure within dialysis. Apart from raising the risk of symptomatic intradialytic hypotension, associated with excess mortality [17], aggressive ultrafiltration is proposed to promote sub-clinical cardiac stunning and cardiomyocyte ischemia, factors that may also worsen left ventricular function [18].

Potassium

Both hyper and hypokalemia are well-known risk factors of cardiac arrhythmias and sudden death [19]. Fluxes in potassium levels during the intra and interdialytic periods represent another important mechanism that could explain the day-of-week pattern of mortality. In the study by Bleyer et al. [6], patients who died during the last 12-hour period of the long interdialytic interval had higher average serum potassium levels than that of the survivors. Genovesi et al. [7] showed that hemodialysis patients with pre-dialysis serum potassium levels >6.0 mmol/l had a 2.7-fold elevated risk of sudden cardiac death during the 3-year follow-up. In addition, Krishnasamy et al. [2] showed that among patients receiving ≤3 dialysis treatments per week, deaths related to hyperkalemia occurred more frequently on Mondays than on any other day of the week. In addition, high potassium fluxes during the first weekly dialysis session may also be of major clinical significance. Karnik et al. [8] reported that use of low potassium concentrations (i.e. 0-1 mEq/l) in the dialysate was associated with increased risk of sudden cardiac death. In a subsequent study, Pun et al. [20] confirmed that performing dialysis with a low potassium bath against elevated serum potassium leads to high imbalances of the potassium gradient between the intracellular and extracellular body compartments, raising the risk of serious arrhythmias and sudden death. However, all the studies mentioned above were observational; cause-and-effect relationship cannot be established.

Acid-Base Imbalances

Hemodialysis patients have significantly lower serum bicarbonate levels and lower pH after the long interdialytic period [21]. The relationship of pre-dialysis serum bicarbonate levels to subsequent mortality is U-shaped [22]. Whether worsened acid-base balance exerts an effect on the excess mortality during the 3-day interdialytic interval remains unknown.

Calcium and Phosphate

Ionized calcium regulates myocardial contractility and vascular tone [23]. Serum concentrations of calcium and phosphate were significantly higher after the long interdialytic interval relative to the regular 2-day interdialytic interval [24]. In the long term, elevated calcium phosphate product is involved in the pathogenesis of vascular calcification, which impairs arterial wall structure [25,26]. Since impaired mineral metabolism is a strong and independent predictor of mortality in hemodialysis [27], the intermittently higher calcium phosphate product during the long interdialytic interval may worsen the outcomes in these patients. Furthermore, recent studies have shown that high dialysate calcium concentrations acutely raise BP and worsen arterial compliance during dialysis [28]. Prospective studies support that session-to-session variability in serum calcium and phosphate levels is a predictor of mortality in hemodialysis [29,30]. Thus, shifts in calcium phosphate product may be also implicated in elevated risk of mortality associated with the long interdialytic interval.

Hemodialysis acutely reduces the augmentation component of central systolic pressures, resulting probably from decrease in the amount of the pulse wave reflected at peripheral sites, but has no effect on arterial compliance [31]. Furthermore, indices of wave reflections are increased during 3-day interdialytic period (30% more than during a regular 2-day interdialytic interval), and these changes are directly related to the interdialytic weight gain [32]. The increase in peak and end systolic pressures in the ascending aorta raises myocardial pressure afterload and myocardial oxygen demand; reduced sub-endocardial blood flow during diastole favors the occurrence of acute ischemic cardiac events [33]. In the long term, this intermittent effect would further promote left ventricular hypertrophy (LVH), raising the risk of cardiac arrhythmias and sudden death.

The development and progression of LVH represents a strong predictor of cardiovascular mortality in hemodialysis [34]. A higher left ventricular pre-load and the consequent higher mechanical stretch of myocardial fibers during the period of the greatest fluid accumulation (i.e. the last 12-hour period of the long interdialytic interval) could be a precipitating factor for acute pulmonary edema and sudden cardiac death. This is supported by the results from the aforementioned study by Foley et al. [1] in which hospital admissions for congestive heart failure and cardiac arrhythmias were shown to occur more frequently on the days following the long interdialytic interval. Cardiac arrhythmias among hemodialysis patients may be provoked by capillary rarefaction and cardiomyocyte ischemia, which are typical features of uremic cardiomyopathy [35].

In the general population, Mondays are associated with higher risk of acute myocardial infraction and sudden cardiac death than any other day of the week [36,37]. This day-of-week pattern has been attributed to changes in physiologic circadian patterns of various parameters, including the temporal plasma catecholamine release and the peaks and falls in plasma cortisol levels [36,37,38]. Another contributing factor is proposed to be the overactivity of the sympathetic nervous system. Hemodialysis patients maybe particularly susceptible to the adverse cardiovascular effects of circadian and temporal changes in hormonal status and the activity of sympathetic nervous system, as majority of these patients have documented cardiovascular disease and impaired compensatory vascular mechanisms [39].

The impact of 6 times per week in-center hemodialysis versus conventional thrice-weekly hemodialysis was examined by the 12-month Frequent Hemodialysis Network (FHN) randomized trial [40]. This study showed that enhanced-frequency hemodialysis was associated with 39% reduction in the risk of death or change in left ventricular mass index (LVMI), as assessed by MRI (hazard ratio (HR) 0.61, 95% CI 0.46-0.82), and with 30% reduced risk of death or change in the physical health composite score of the RAND 36-item health survey (HR 0.70, 95% CI 0.53-0.92) [40]. The major finding of the FHN trial was the improvement in LVH with short sessions of daily hemodialysis; notably, LVH regression has been clearly associated with reduced mortality in hemodialysis patients [34]. Other benefits of enhanced-frequency dialysis regimens include improvement in phosphorus homeostasis, BP control, arterial stiffness, total body water and self-reported quality of life [41,42,43,44,45].

Effects of enhanced-frequency hemodialysis on ‘hard' outcomes or mortality were evaluated in recent retrospective cohort studies. The following 2 studies showed an improvement in mortality. In a retrospective cohort study including 1,873 patients who were receiving daily home hemodialysis and 9,365 matched controls receiving thrice-weekly hemodialysis selected from the USRDS database during 2005-2008, Weinhandl et al. [46] showed that daily home hemodialysis was associated with 13% lower risk of all-cause mortality than that of in-center thrice-weekly hemodialysis (HR 0.87, 95% CI 0.78-0.97). Another retrospective study, enrolling 338 patients from the International Quotidian Dialysis Registry who received intensive home hemodialysis between January 2000 and August 2010 and another 1,388 matched controls participating in the DOPPS study who received thrice-weekly hemodialysis during the same time-period, showed that intensive home hemodialysis is associated with 45% reduction in the risk of all-cause mortality (HR 0.55, 95% CI 0.34-0.87) [47].

In contrast to the above, survival benefits of enhanced-frequency hemodialysis were not confirmed in the study by Suri et al. [48], which showed that during 1,382 patient-years of follow-up, 318 patients receiving short sessions of daily hemodialysis (>5 times per week) had 60% higher risk of death than 575 matched controls receiving thrice-weekly hemodialysis (HR 1.60, 95% CI 1.1-2.3). This excess mortality may be due to incomparability of groups at study initiation with respect to several parameters, such as comorbidities, type of dialysis access, BP control and residual diuresis.

These modalities were also related to harms, including elevated risks of vascular access complications, worsened residual renal function and increased burden of therapy, resulting in higher rates of method discontinuation among patients scheduled for enhanced-frequency dialysis [4,49,50]. Vascular access events emerge as the most important issue; although plausible, the currently available evidence is inconclusive, as in only one of the two randomized studies in the field frequency of vascular access interventions was more common in patients receiving frequent dialysis [50]. Deterioration of residual renal function was noted only in post hoc analyses of the main FHN trial [49], but also in the primary results [40], while burden of therapy and method discontinuation rates should be considered ‘softer' end points influenced by several confounding parameters. Although these complications cannot be neglected, they should not overshadow the currently available evidence, which is at least suggestive of a beneficial impact of enhanced-frequency hemodialysis on intermediate cardiovascular end points and mortality.

Another treatment option toward attenuation of the mortality risk attributed to conventional thrice-weekly dialysis regimens could be extended-time, home or in-center nocturnal hemodialysis, which although not more frequent, reduces the duration of the long interdialytic period. Ok et al. [51] prospectively followed 247 patients assigned to 8-hour in-center nocturnal hemodialysis and 247 matched controls assigned to conventional thrice-weekly hemodialysis for 12 months. Extended-time nocturnal hemodialysis was associated with 72% reduced risk of all-cause mortality relative to conventional hemodialysis. A subsequent prospective cohort study compared risk of all-cause mortality during a mean follow-up of 24 months in 746 patients who initiated extended-time, in-center, nocturnal hemodialysis in 77 dialysis centers of a large dialysis-care provider in the USA during 2006-2007 and in 2,062 controls who remained on conventional thrice-weekly hemodialysis [52]. Nocturnal hemodialysis was associated with 25% reduction in the risk of all-cause mortality (HR 0.75, 95% CI 0.61-0.91); furthermore, no day-of-week effect was observed on occurrence of mortal events during the follow-up [52]. However, once again, these observations have to be confirmed in future randomized clinical trials; cause-and-effect relationships cannot be assured.

In a pilot study, 9 prevalent hemodialysis patients were switched to alternate-day hemodialysis for 12 months and another 9 patients remained on conventional thrice-weekly hemodialysis [53] Improvements in LVMI, in BP control and in several other biochemical factors were observed, but these results are only preliminary.

The benefit of enhanced-frequency and/or extended-time dialysis regimens on hard outcomes and cost effectiveness of these modalities remain unclear. Until such data become available, physicians may need to develop clinically feasible dialysis practices in order to attenuate the risk of long interdialytic intervals for patients who receive standard thrice-weekly hemodialysis. Interventions could include some of the following. First, reduction of the long interdialytic period, with prescription of an additional dialysis treatment on Saturdays or Mondays in patients carrying a high risk of adverse cardiovascular events during the long interdialytic interval (i.e. patients with severe heart failure, high interdialytic weight gain and frequent episodes of intradialytic hypotension) [54]. In addition, treatment interventions during the first dialysis session of the week, after the long interdialytic interval, may also be of benefit. Such interventions include extending the duration of this session in order to reduce the rate of volume withdrawal and individualized prescription of dialysate potassium to avoid high potassium shifts [54]. Last, interventions that target reduction of the risk of hyperkalemia during the long interdialytic interval, such as dietary consultation of non-adherent patients and administration of drugs, to reduce absorption of dietary potassium may improve potassium homeostasis during the long interdialytic period [54,55]. Diets avoiding animal protein in preference for plant protein may mitigate acidosis; such approaches are testable in pilot clinical trials.

All the above therapeutic approaches may not be relevant for some specific sub-populations of ESRD patients, that is, patients who initiate renal replacement therapy or prevalent dialysis patients with preserved residual renal function. High residual renal function offers several advantages to dialysis patients, including higher fluid and sodium removal, more effective potassium excretion, higher middle-molecule clearance and reduced LVH [56]. On the basis of a relatively higher capacity to maintain homeostasis of volume and metabolic parameters over a longer out-of-dialysis period, twice-weekly hemodialysis is suggested to be a sufficient dialysis dose for such patients [57]. This notion is supported by recent observational studies that show that twice-weekly hemodialysis is associated with a slower reduction of residual renal function, fewer episodes of intradialytic hypotension, improved health-related quality of life and higher survival rates as compared with conventional thrice-weekly hemodialysis [58,59,60]. However, these potential benefits need to be tested in properly designed randomized trials.

Although a cause-and-effect relationship between day-of-week mortality and dialysis scheduling remains unproven, if a cause-and-effect relationship exists, it may be due to magnified fluctuations in volume status and BP, electrolyte imbalances and changes in arterial wall parameters during the long interdialytic interval. Clinical studies evaluating potential harms and benefits of practices with altered timing and frequency of prescribed hemodialysis therapy have provided promising results on intermediate outcomes. Additional randomized clinical trials evaluating hard end points are urgently warranted to elucidate the impact of these alternative dialytic modalities on survival. In the meantime, individualized prescription of the dialysis frequency and duration along with close monitoring of our patients appear to be the most appropriate therapeutic approaches to lessen the risk of adverse events during the long interdialytic interval and improve the way we deliver dialysis.

Authors have no conflicts of interest to disclose.