‘Time and Time Again’: Oscillatory and Longitudinal Time Patterns in Dialysis PatientsKooman J.P.a · Usvyat L.b · van der Sande F.M.a · Thijssen S.b · Levin N.b · Leunissen K.M.a · Kotanko P.b
aUniversity Hospital Maastricht, Maastricht, The Netherlands; bRenal Research Institute, New York, N.Y., USA Corresponding Author
Oscillatory and longitudinal time patterns play a major role in human physiology. In chronic hemodialysis patients, abnormalities in both time patterns have been observed, while time patterns can also influence the response of patients to the treatment. Abnormal oscillatory patterns have been observed for ultradian rhythms (cycle time <20 h), such as an impaired heart rate variability and circadian rhythms, as reflected by reduced day-night blood pressure differences. Conversely, the circadian rhythm of body temperature may influence the hemodynamic tolerance to the dialysis treatment. With regard to infradian (cycle time >28 h) rhythms, large seasonal differences in mortality, but also in blood pressure and interdialytic weight gain, have been observed in dialysis patients. The most important longitudinal pattern is the general reduction of life span in dialysis patients. One explanation of this phenomenon relates to the concept of accelerated aging in dialysis patients, for which there are various supportive arguments. From a phenomenological point of view, this concept translates into the high prevalence of frailty, even in young dialysis patients. A multidimensional approach appears necessary to adequately address this problem. In this review, the relevance of disturbed time patterns in dialysis patients is discussed. The changes may reflect an impairment or reduction in homeostatic/homeodynamic control in dialysis patients and also may have important prognostic and therapeutic implications.
© 2012 S. Karger AG, Basel
In human life, time passes both in longitudinal or oscillatory patterns. The longitudinal, or ‘hourglass’, pattern is unidirectional and reflects life as passing through different stages. Birth, puberty, adulthood, menopause, ageing, and death are examples of the hourglass pattern in a human life . These patterns are reflected both in human physiology, psychology, and in passage rites universally present in human societies.
The longitudinal aspect of time is also deeply rooted in our concept of causality. Although it may be sometimes problematic to translate physical concepts to biology [2,3], several authors have described a relation between the aging process and the second law of thermodynamics [4,5]. Whereas order in the organism is maintained by incorporation of free energy from the environment and the generation of high entropy waste, during the ageing process, an increase in entropy in the organism is reflected by cross-linking of proteins, oxidative damage of biomolecules, somatic mutations, and a reduced functional reserve [6,7]. Also, the aging process has been defined as a loss in complexity of the organism, leading to a reduction in the degrees of freedom to respond to external stressors, resulting in impaired homeostatic control mechanisms, a reduced functional reserve, and an increased susceptibility to disease [4,8,9].
On the other hand, in living organisms as well as in geophysical cycles, time is also defined by an oscillatory (cyclical) pattern. Before the industrial era, life of the community was built on solar, lunar, and seasonal rhythms. However, in modern times, due to societal changes, a detachment between intrinsic and environmental cycles has occurred .
Cyclical rhythms play a key role in human physiology and are of critical importance in phasing the function of the body to the varying demands of the environment . Thus, homeostasis does not only reflect the static capacity of the organism to maintain a constant internal environment, but also a successful dynamic adaptation to environmental changes. Cyclical patterns also serve to reset the organism to its initial conditions . The evolutionary advantage of cyclical patterns is shown by its preservation from unicellular to multicellular life .
Both longitudinal and oscillatory timing patterns can be altered in patients with chronic diseases. These alterations may be related to a disturbed homeodynamic control. This concept has been described as an extension of the homeostasis concept and reflects the ability of an organism to respond dynamically to the demands of a changing environment [4,5]. Fluctuating and oscillatory responses are important characteristics of the homeodynamic concept.
Recent studies have shed more light on the pathophysiological role of alterations in temporal patterns in dialysis patients. The aim of the present review is to discuss recent developments in the understanding of alterations in cyclical and longitudinal rhythms in dialysis patterns, and their relationship with pathophysiology and outcome in this population (table 1).
|Table 1. Examples of oscillatory and longitudinal rhythms with findings in dialysis patients, and potential therapeutic implications|
Biological rhythms can be distinguished according to their period, i.e. the duration of a complete cycle. Circadian rhythms have a period of around 24 h, whereas the periods of ultradian and infradian rhythms are respectively shorter (<20 h) or longer (>28 h) than circadian rhythms. Whereas all oscillatory rhythms reflect the organisms’ self-organization in a temporal pattern, circadian and infradian patterns are in general more linked to the passage of time, and as such to the periodic nature of the external environment (day-night cycle, passing of the seasons); however, this is less evident for ultradian rhythms, such as those of neural activity, heart rate, and respiration [13,14].
The ultradian pattern which has received most attention in dialysis patients is heart rate variability (HRV). HRV, the variation in the beat-to-beat interval of the heart is a physiological phenomenon, influenced by the autonomic nervous system. Without autonomic control, the sinus node has an intrinsic regular rhythm of approximately 110 beats/min . HRV is strongly influenced by the autonomic nervous system and, in general, HRV is regarded as a sign of functionally efficient autonomic control .
While there are multiple ways to describe HRV quantitatively, frequency domain analysis is preferred by many investigators since it relates certain frequency bands to the autonomous nervous system. The very-low-frequency (VLF; <0.04 Hz) band is thought to be influenced by thermoregulation and possibly the renin-angiotensin system ; a low-frequency (LF; 0.04–0.15 Hz) component is influenced by the baroreceptor reflex, and is affected by the sympathetic and possibly the parasympathetic nervous system; and the high-frequency (HF; 0.15–0.40 Hz) band which is related to respiration and predominantly considered to be under parasympathetic control [15,16,17]. The LF/HF ratio is considered as an index of sympathovagal balance. Although the exact nature of the disturbance may differ between studies (differences in both the LF and HF spectrum have been observed in dialysis patients as compared to controls), abnormalities in HRV are frequently observed in dialysis patients with reported prevalence rates of 53% , climbing as high as 82% . Next to diabetes,  other possible contributing factors include LVH, fluid overload, hyperparathyroidism, and myocardial ischemia [21,22]. Abnormalities in HRV have been established as a significant independent risk factor for higher mortality and sudden cardiac death in cardiovascular disease and healthy populations [16,23]. Also in chronic hemodialysis patients, a decreased HRV was found to be an independent predictor of mortality .
Although frequency domain analysis is the most frequently used method in the literature on HRV in dialysis, other methods are also available to assess HRV. A fairly recent development is the analysis of nonlinear dynamics in HRV and complexity analysis . Here self-similarity of HRV variability over different time periods is assessed, in analogy to spatial fractal analysis .
Roughly speaking, greater complexity is related to a higher information content of the signal and a greater robustness towards external stressors as compared to a completely regular rhythm . It is important to distinguish this phenomenon, which is also known under the name chaos, from randomness, as observed in ventricular fibrillation . To date, few studies have used nonlinear and complexity analysis in assessing HRV in dialysis patients [27,28]. One study showed a relation between adverse prognosis reduced complexity of HRV in dialysis patients .
Apart from these prognostic effects, several studies have shown that abnormalities in HRV might affect the tolerance to dialysis treatment and increase the risk of dialysis hypotension . Reciprocally, the dialysis treatment can also influence HRV. An improvement in HRV was observed the first day after dialysis treatment, whereas with nocturnal dialysis, a reduced sympathetic drive was observed . Also other treatment-related factors, such as the dialysate glucose concentration, were shown to influence HRV . In addition, it has been shown that other interventions such as exercise training and renal transplantation may also improve HRV [31,32].
In summary, HRV is impaired in dialysis patients, which may have prognostic implications, but also be relevant for a reduced hemodialysis tolerance. Impaired HRV variability may not only reflect impaired autonomic control, but also reduced ability of the heart to respond to environmental stressors. From the data available, it appears to be that abnormal HRV can be at least partly reversible. Whether this translates into an improved prognosis of dialysis patients remains to be determined.
Cyclical rhythms can be defined by distinct parameters, such as the period (duration of a complete cycle), mesor (an estimate of the mean level), amplitude [difference between peak (or through) and mesor], and the acrophase (the time at which the peak of a rhythm occurs)  (fig. 1).
|Fig. 1. Characteristics of cyclical rhythms (from Usvyat et al.  with permission). Mesor = Estimate of central tendency of the distribution of values; Peak = the point of culmination of an oscillatory function; Nadir = lowest value of an oscillatory function; Amplitude = difference between the peak and the mean value of a wave; Acrophase = time at which the peak of a rhythm occurs.|
The most apparent oscillatory rhythm in human physiology is the circadian rhythm. However, as will be discussed later, seasonal rhythmical variations may also be of importance for human pathophysiology (fig. 2).
|Fig. 2. Examples of relevant cyclical rhythms in the pathophysiology of dialysis patients.|
Circadian rhythms display cyclical patterns with periods between 20 and 28 h. Recently, the insight into the molecular regulation of circadian patterns in organisms has greatly expanded. In the suprachiasmatic nucleus, the expression of the transcription factors CLOCK and BMAL1 stimulate transcription of Period (per) and Cryptochrome (Cry) genes whose products (PER and CRY) in turn repress the transcription of BMAL 1 and CLOCK . While it is not the aim of this review to discuss these mechanisms in detail, it is important to note that the effectiveness of the circadian clock depends on an interplay between the ‘master clock’ located in the suprachiasmatic nucleus (SCN) of the hypothalamus , and the peripheral clocks in individual cells of target organs [36,37]. In short, these peripheral clocks modulate local physiological rhythms . Peripheral clocks are found in many organs, among which is the kidney, and are an important factor in day-night differences in glomerular filtration and sodium reabsorption . Synchronization between the master clock and the neuroendocrine and autonomous nervous system is further achieved by interaction between the SCN and the paraventricular nucleus of the hypothalamus, with an important role for vasopressin [40,41]. The different components of the circadian system form an interconnected system of various oscillators with multiple feedback loops, and as such are a classical example of system biology [38,42,43].
Apart from endogenous clocks, also external stimuli, most notably the environmental light-dark cycle mediated through the retinohypothalamic tract, influence the circadian clock, acting as synchronizers (or zeitgebers) . Disruption of endogenous rhythms and synchronizers is, for instance, causal in jet lag, but also possibly related to health problems in subjects on shift work . Moreover, a desynchronization between the internal clock and external stimuli has been observed in patients with hypertension and diabetes . Recently, a disturbed rhythmicity of melatonin, which has anti-inflammatory and antioxidant properties, have been implicated in metabolic and epigenetic dysregulation .
Uremia, as well as the dialysis treatment itself can strongly interact with the physiologic circadian rhythms of the body, further contributing to a loss of homeostasis in these patients. Dialysis may disturb the synchronization between endogenous rhythms and synchronizers, among others due to the immobilization of patients during daytime at times of normal activity, changes in food intake, and the effect of dialysis on body temperature. As not all disturbances in diurnal rhythms can be addressed, this review focuses primarily on three exemplary rhythms with special relevance for dialysis patients, i.e. temperature, sleep-wake, and blood pressure rhythms.
The circadian rhythm of temperature is mediated through the SCN, possibly by a direct effect on the hypothalamic thermoregulatory center . The decline in temperature appears to be preceded (‘phase-advanced’) by vasodilatation of the skin, inducing heat loss and a decline in core temperature [45,46]. Melatonin, which is synthesized in the pineal gland and strongly responsive to the light-dark cycle but which also obtains input from the SCN , also contributes to peripheral vasodilatation. There is still discussion whether these circadian temperature changes are mediated through the hypothalamic thermoregulatory set-point, or by a direct connection between the SCN and the peripheral thermoregulatory system [11,45].
Although no studies are available in which the circadian temperature rhythm is compared between dialysis patients and healthy subjects during a complete daily cycle, the dialysis procedure can interact with the temperature rhythm. The dialysis procedure itself has a strong effect on thermoregulation, tending to increase core temperature even in the absence of net heat transfer from the extracorporeal system to the patient . The mechanism behind this phenomenon has not fully been elucidated yet. The relation between heat accumulation and ultrafiltration-induced blood volume suggests that initial vasoconstriction due to hypovolemia reduces heat loss from the skin (peripheral shell mechanism), whereas the resulting increase in core temperature leads to secondary peripheral vasodilatation and a decline in blood pressure [49,50]. However, the increase in core temperature has also been observed during dialysis without ultrafiltration . Whatever the mechanism, the resulting alterations in core temperature and the peripheral vasodilatation may interfere with the sleep-wake cycle of the dialysis patient, as will be discussed in the following paragraph.
Although chronic diseases , as well as ageing , may affect the amplitude of the circadian temperature rhythm, and uremic hypothermia  may decrease the mesor of the temperature cycle, no study has compared the amplitude of the circadian temperature rhythm in dialysis patients with control subjects. Conversely, the degree of change in body temperature during dialysis also appears related to the timing of the dialysis shift . In phase with the circadian temperature rhythm, predialytic core temperature was higher in patients treated during an evening session, when the increase in body temperature was lower. As for yet unknown reasons, these differences were more pronounced during summer as compared to the winter. Although the intradialytic increase in core temperature was related to the decline in blood pressure during dialysis, controlled studies are needed to elucidate whether circadian changes in core temperature affect the hemodynamic response in dialysis patients.
In up to 30–80% of dialysis patients, the quality and/or quantity of sleep are reduced, and REM sleep is reduced. Many factors likely contribute to this problem, such as the use of sleep-influencing drugs, stress associated with the dialysis procedure, reduced physical activity during the daytime and the sleep apnea syndrome . This section primarily focuses on the interaction of the sleep-wake cycle with the circadian temperature rhythm and the influence of hemodialysis, as well as abnormalities in the diurnal melatonin rhythm.
The peripheral vasodilatation associated with the increase in core temperature during hemodialysis discussed in the previous paragraph is analogous to the normal vascular response preceding sleep. The resulting skin warming stimulates warm-sensitive neurons in the preoptic anterior hypothalamus involved in sleep regulation , inducing daytime sleepiness, which may in turn reduce sleep quality during nighttime. In a controlled trial, cooling of the dialysate, which prevents peripheral vasodilatation, was associated with improved sleep quantity and quality as compared to standard dialysis temperature [54,56]. Notably, the dialysis condition even had an effect on peripheral skin temperature measured in the early morning of the next day after dialysis. Peripheral skin temperature was lower in the early morning following the standard dialysis as compared to the cool dialysis session. This is possibly related to phase advancing of the normal circadian vascular rhythm, in which the increase in core temperature during the morning is preceded by peripheral vasoconstriction .
Nocturnal hemodialysis also improves sleep quality, despite the possibility of alarms and the attachment to a dialysis module. This might be related to improved volume and metabolic control, but possibly also to the fact that the vasoactive response to the dialysis procedure coincides with the normal circadian vasomotor response .
Although reported daytime values vary between different reports, there is agreement about the attenuation of the diurnal melatonin rhythm in dialysis patients, likely resulting in an adverse effect on sleep quality and quantity . It has been hypothesized that one of the causes behind this disturbed rhythm is an impairment of β-receptor responsiveness, which negatively influences the synthesis of N-acetyltransferase, an important enzyme in melatonin synthesis. However, the disturbed sleep-wake itself could inversely affect melatonin production at nighttime .
The possible implications of these findings are that both cool dialysis and nocturnal dialysis might improve sleep quality and quantity [56,59]. Nocturnal dialysis was also associated with a partial restoration of the diurnal melatonin rhythm . In addition, exogenous melatonin might be helpful in restoring the sleep-wake cycle, as 3 mg of melatonin administered at 22:00 h improved sleep quality and quantity in a randomized trial in dialysis patients .
The diurnal blood pressure rhythm is strongly influenced by external stimuli such as the sleep/wake state physical activity, but is also dependent upon the interplay between central and peripheral circadian molecular clock mechanisms, mediated, among others, by the autonomic nervous and renin-angiotensin-aldosterone system [61,62,63]. In healthy individuals, the blood pressure pattern displays a clear difference between day and night (i.e. the dipper pattern; defined as a reduction of systolic blood pressure of >10% during night compared with daytime). In dialysis patients, this pattern is clearly disturbed: studies showed that approximately 80% of patients could be classified as nondippers . Different explanations have been proposed for the increase in the nighttime/daytime blood pressure ratio, such as sympathetic overactivity, a high prevalence of the sleep apnea syndrome, and physical inactivity. However, clinical evidence for the relative importance of these different mechanisms in dialysis patients is limited, given the fact that most of these phenomena have been studied in patients at earlier stages of renal failure .
A role for fluid overload is suggested by the observation that in contrast to healthy subjects, who had a lower fluid status during nighttime, in dialysis patients no day-night differences in fluid state were observed. Interestingly, these diurnal volume differences are preserved in dialysis patients with urine output larger than 500 ml/day and in predialysis patients . However, volume reduction by ultrafiltration did not result in improvement of the day-night blood pressure rhythm , although it is unclear whether in this study normovolemia was actually reached. However, in many patients, circadian blood pressure rhythm is already disturbed relatively early in the course of renal failure when volume excess is less likely .
An interesting recent observation is the possible relation between disturbances in the diurnal melatonin rhythm and nondipper hypertension . Also, in nonuremic patients with hypertension, melatonin intake at nighttime reduced nocturnal blood pressure .
In addition to its pathophysiologic interest, the nondipping status appears to be of major clinical significance. In nonuremic hypertensive subjects, as well as patients with chronic kidney disease not on dialysis , the nondipping pattern was found to be related to mortality. However, in a recent meta-analysis in hypertensive subjects, the prognostic value of the nondipping pattern was not superior to that of the overall 24-hour ambulatory blood pressure values . Still, in dialysis patients, the night-to-daytime systolic blood pressure ratio was the only 24-hour ambulatory BP parameter associated with all cardiovascular mortality, showing a strong interaction with left ventricular hypertrophy .
The nondipping blood pressure pattern also may have therapeutic implications. In hypertensive patients with type II diabetes, the prescription of antihypertensive agents at bedtime significantly reduced cardiovascular mortality and decreased 24-hour blood pressure [71,72]. The effect of timing of antihypertensive treatment on cardiovascular outcome has not yet been studied in dialysis patients. It is also not known whether other interventions, such as nocturnal dialysis, might be helpful in restoring the day-night blood pressure rhythm. If indeed melatonin plays an important role in nocturnal blood pressure regulation, the fact that nocturnal dialysis was able to partially restore the diurnal melatonin rhythm might possibly contribute to an improvement in nightly blood pressure control. To our knowledge, this subject has not yet been addressed in the literature in a controlled study. In a study by Chan et al. , 24-hour ambulatory blood pressure was reduced by nocturnal dialysis. However, in a study of 24 patients on nocturnal dialysis, only 25% displayed a normal dipping pattern .
Next to the circadian blood pressure rhythm, ultradian blood pressure rhythms, with 6-, 8-, and 12-hour cycles have also been observed. The causes and consequences of these mechanisms have not yet been fully elucidated. However, blunted ultradian rhythms have been observed in patients with chronic kidney disease which correlated better with renal dysfunction as compared with circadian patterns . Further research is necessary to address the clinical relevance of these observations in dialysis patients.
Mortality and Timing of the Day
In the general population, cardiovascular mortality is highest during the morning period . This appears to be related to an increased oxygen demand of the heart, coinciding with a reduction in coronary blood flow. Moreover, possibly under influence of sympathetic activation, platelet adhesion increases during the morning, in combination with an increase in coagulation and a reduction in fibrinolytic activity . Loss of synchronization between internal peripheral clocks and external stimuli may further contribute to this mismatch (fig. 3) .
|Fig. 3. Desynchronization between internal clock and external stimuli in organ damage (from Takeda and Maemura  with permission).|
No study has addressed the relation between mortality and time of day in dialysis patients. In contrast, the relation between the timing of the shift and mortality has been addressed in recent studies. One study found lower cardiovascular mortality in elderly patients in a morning dialysis session as compared to patients treated in other shifts . Data from the HEMO study  also described a lower cardiovascular mortality in patients treated during the morning shift, but no difference in all-cause mortality after adjustment for confounders. As no study randomly assigned patients to different shifts, it remains difficult to elucidate to what extent case-mix is responsible for the observed differences between shifts.
Infradian rhythms are rhythms with periods longer than 28 h. In the general population, seasonal rhythms in particular are related to factors such as birth, death, and various physiological and laboratory parameters. There appears to be no solid evidence for an important role of lunar parameters on human (patho)physiology .
In the following paragraphs, infradian rhythms related to the dialysis procedure and seasonal variations in dialysis patients are discussed.
Intradialytic Interval and Mortality
An important infradian rhythm is imposed by the timing of hemodialysis treatment itself. Given the fact that hemodialysis treatment is performed discontinuously at varying intervals, the risk of hypervolemia and hyperkalemia is greater towards the end of the longer interdialytic interval (48–60 h), whereas the risk of hypovolemia and hypokalemia is greatest at the end of the dialysis procedure. Although not formally tested, these are possible reasons behind the increased risk of sudden cardiac death after the long (weekend) interdialytic interval, as well during the immediate postdialytic period [79,80]. In the study of Bleyer et al.  concerning sudden cardiac death within 12 h after dialysis, 25% of patients had predialytic plasma potassium below 4 mmol/l. In a study of sudden cardiac death in dialysis facilities, the highest risk was observed in the Monday session. Potentially modifiable predictive factors appeared to be, among others, a low-dialysate potassium bath and the use of central venous catheters. Sixteen percent of patients experienced intradialytic hypotension before the event .
The higher mortality risk after the long interdialytic interval also translates to an increased all-cause mortality related to death from a variety of cardiovascular causes and intriguingly, even from infections . Although in the general population increased mortality is also observed on Mondays , in dialysis patients this phenomenon appears to be related to the interdialytic interval. Depending on the timing of the dialysis session, the risk was either increased on Monday or Tuesday . It would be interesting to compare the data in hemodialysis patients to those of peritoneal dialysis patients, but to our knowledge, these data are not available for the latter group of patients.
The practical implications of these findings are that, at least theoretically, this risk could be reduced by increasing the frequency of the dialysis sessions. Moreover, special attention to derangements in fluid status as well as prevention of hyper- and hypokalemia appears warranted. Discussion of the availability of automatic external defibrillators in a dialysis unit or the role implantable cardioverter-defibrillators falls beyond the scope of this review.
In preindustrial societies, the seasons had a strong effect on all aspects of human life, including birth rate. This trend has clearly decreased with the start of industrialization . However, in the general population, important physiological parameters, such as blood pressure and body temperature, as well as phasing of endogenous rhythms are different between seasons [85,86,87]. Blood pressure was higher during the winter period in a large cohort of elderly persons . Various factors, such as an increased vascular tone, as well as lower vitamin D3 levels have been implicated in these differences.
Seasonal differences in blood pressure, as well as various laboratory parameters are also evident in dialysis patients. Different groups have shown higher blood pressure during the winter as compared to the summer period. The same holds true for differences in interdialytic weight gain [88,89], being lower in the summer, possibly related to increased perspiration and insensible fluid loss . The use of hypertonic peritoneal glucose solutions was also shown to be higher with lower environmental temperatures . However, not all differences in blood pressure between seasons appear to be explained by volume factors. In the single study addressing this subject, no difference in extracellular volume assessed by bioimpedance was observed between seasons , suggesting that other factors, such as differences in adrenergic activity, vascular tone or differences in vitamin D3 levels also play a role in the differences in blood pressure between seasons [93,94,95]. In addition, other physiologic parameters, such as predialytic body temperature, are seasonally variable, as recently shown by Usvyat [unpubl. obs.] in a RRI database study containing 15,056 dialysis patients (fig. 4). Although the clinical implications of these findings have not yet been investigated, they might be of relevance given the significant relation between predialytic body temperature and risk of intradialytic hypotension .
|Fig. 4. Relation between season and predialytic body temperature (pre-BT), corrected for timing of treatment shift.|
In the general population, mortality varies between seasons. Both excessively high temperatures during the summer and cold temperatures during the winter are related to increased mortality . The increased winter mortality is both due to cardiovascular as well as infection-related causes. The effect of colder temperatures on cardiovascular mortality has been attributed to various factors, such as higher coagulability or increased vascular tone. Somewhat paradoxically, the negative effect of cold on cardiac events appears to be highest in warmer climates, probably because of better protection to cold temperatures in regions used to harsher climates [98,99].
In the single study which addressed this issue in dialysis patients, we also observed significantly increased mortality during the winter period in US dialysis patients, irrespective of the climate region  (fig. 5).
|Fig. 5. Seasonal variations in mortality calculated as deaths per 100 patient years (from Usvyat et al.  with permission).|
The increased mortality was mainly related to cardiovascular causes. In contrast to the general population, in whom especially elderly persons appear at risk for seasonally related mortality, seasonal differences in mortality were also clearly pronounced in younger dialysis patients, possibly related to a frail phenotype (see discussion).
The implications of these findings remain to be determined. However, apart from pathophysiological interest, the relatively large differences in mortality between seasons need to be taken into account when interpreting or designing trials. Whether the excess winter mortality in dialysis patients is modifiable should be considered in future studies. At present, it primarily confirms the frailty and vulnerability of our population to external stimuli.
Next to cyclical patterns, longitudinal time patterns may also be altered in patients with ESRD. In the following paragraphs, the relation between ESRD and life span, as well as accelerated ageing, will be discussed. This will be followed by a discussion on new developments in dynamic epidemiologic patterns in dialysis patients.
Patients with renal failure often follow a prolonged disease course, starting with chronic kidney disease of gradually increasing severity, followed by dialysis, and finally transplantation. After a failed transplant, patients return to dialysis.
This course is generally characterized by a shortening of life time. Life expectancy is, however, highly variable, primarily depending on age, comorbidity, and the presence of a functioning renal transplant . The survival benefit of renal transplantation becomes dramatically apparent in data from patients who started dialysis in childhood who were followed from their 18th birthday. In this group, average life expectancy for those patients with a functioning renal transplant was 63 versus 38 years for those remaining on dialysis . The survival benefit of renal transplantation is highest in the younger population, but also apparent in the elderly [102,103].
These pronounced differences in survival indicate that present-day dialysis techniques are insufficiently able to correct the uremic state. It is of interest that a recent observational study showed comparable outcomes between patients treated with extended hemodialysis and patients after cadaveric renal transplantation . Although these results might be influenced by case-mix, it is tempting to speculate that the improved outcomes after nocturnal hemodialysis are related to a better ability to maintain homeostasis as compared to conventional dialysis techniques, as shown by an improvement in fluid status, toxin removal, mineral metabolism, and other parameters . However, despite the benefits of extended dialysis, the outcome is inferior as compared to living-related transplantation, since it is still associated with dialysis-related morbidity and huge healthcare costs. Preemptive renal transplantation, i.e. transplantation before the start of dialysis which is associated with excellent long-term outcomes , provides at present the best treatment for patients with end-stage renal disease.
The mechanisms behind the reduced life span in dialysis patients have not yet been completely elucidated. However, it is clinical acumen that dialysis patients appear to age more rapidly as compared to nonuremic subjects. However, there are relatively few citations on this subject in dialysis patients [107,108].
This clinical observation can be defined more properly when the biological phenotypic characteristics of ageing, i.e. an exponential increase in mortality, functional decline due to physiological alterations, and increased susceptibility to disease are compared with the phenotype of uremic patients (http://www.senescence.info/aging definition.html). Another indicator of accelerated ageing is the high prevalence of frailty, a geriatric syndrome which is also very common in younger dialysis patients and which is significantly related to mortality . Frailty is defined by symptoms of muscle weakness, slow walking speed, poor endurance, physical inactivity, and unintentional weight loss. Thus, sarcopenia and malnutrition, which are also common in dialysis patients, are an important hallmark of this syndrome. From a pathophysiologic point of view, the frailty syndrome is characterized by reduced functional reserve and an increased tendency towards failure of homeostatic mechanisms [110,111,112]. This would explain the increased susceptibility of dialysis patients to external stressors. Of concern is that 44% of dialysis patients under 40 were classified as frail . It is relevant that a small but significant percentage (21%) of octogenarian patients did not conform to the criteria of frailty . Thus, treatment decisions based on age alone may be harmful, as some younger patients may have a higher biological age, as characterized by frailty symptoms, as compared to relatively healthy elderly patients. Another important component of accelerated aging in dialysis patients concerns the cardiovascular system: there are important similarities between both functional and structural vascular alterations in uremic and ageing subjects [107,113].
The pathophysiology of these phenotypical changes is likely multifactorial. Several pathogenetic factors related to ageing have been identified in dialysis patients, such as reduced telomere length, accumulation of advanced glycation end-products, systemic inflammation, and increased oxidative stress [114,115,116]. In analogy to other diseases such as cancer, systemic inflammation appears to play a central role in the pathogenesis of the accelerated aging process [114,117]. Of interest is that also in nonuremic subjects, aging is associated with low-grade inflammation . Other factors, however, may also be involved. An intriguing recent hypothesis attributes accelerated aging in uremic patients to disturbances in phosphate metabolism . In renal failure, a reduction in Klotho expression combined with resistance to fibroblast growth factor 23, in combination with vascular calcification is observed, analogous to findings in Klotho-deficient mice .
Of importance is that recent observations also identify physical inactivity not only as a consequence, but also as a potential culprit in an accelerated ageing process . One of the mechanisms behind this relation is a reduction in the physiological reserve of the patient and thus increasing vulnerability to external stressors. Moreover, there is a strong relation between sarcopenia and physical inactivity in renal patients . It has been hypothesized that physical inactivity promotes storage of metabolically active central adipocytes, resulting in increased production of proinflammatory cytokines .
It is tempting to speculate whether clinical syndromes such as malnutrition-inflammation-atherosclerosis syndrome, which combines pathophysiologic elements with phenotypical alterations, might actually be part of the larger umbrella of a multidimensional accelerated ageing process in uremia, in which metabolic alterations, systemic inflammation, oxidative stress, sarcopenia, functional deterioration, and end-organ disease reinforce each other in a vicious circle. In the terminology of reliability theory, this will result in a reduction in the redundancy of homeostatic/homeodynamic compensatory mechanisms, leading to system failure [120,122], or, alternatively stated, in a reduction in the degrees of freedom in which the human organism can respond to external stressors . This is analogous to the concept of time, represented as a biological helix by F. Eugene Yates  in which the biological time is expressed as a daily cycle of metabolic action, which is diminished during the aging process, until an organism becomes fragile and less tolerant to previously tolerable fluctuations (external stressors), due to a reduced metabolic reserve and impairment of compensatory mechanisms (fig. 6).
|Fig. 6. Homeodynamic time, in which the biological time is expressed as a daily cycle of metabolic action, which is diminished during the aging process, until an organism becomes fragile and less tolerant to previously tolerable fluctuations (external stressors; see text for explanation; from Yates  with permission).|
If the concept of accelerated aging in uremia is to be developed further, it is necessary to address it in a quantitative way. One recent study, applying the derivative of the Gompertz function, a model used to describe the exponential increase in mortality rate with ageing, found evidence for increased senescence rates in dialysis patients . Comparing physiological parameters, dialysis patients with healthy controls of older age classes could further quantify the concept of accelerated ageing in dialysis patients. It should be stated that accelerated aging is not unique to uremia, and has been described in other chronic diseases such as COPD and AIDS [124,125,126]. Future research, e.g. using cluster analysis, should identify whether there are specific phenotypical changes related to this concept in uremic patients as compared to patients with other organ diseases .
Given its systemic nature, the therapeutic approach to this accelerated aging process in dialysis patients is multifactorial. Next to optimal metabolic control, dialysis prescription, and nutritional support, active rehabilitation is likely of great importance. In this respect, it is of great importance that renal patients show a favorable response to exercise training, psychosocial support, and other interventions based on active rehabilitation [128,129].
Prognostic studies in dialysis patients have mainly been based on a cross-sectional approach. Recently, the dynamics of important clinical and laboratory parameters before death have been studied, using the timing of the event as the starting point, whereas the trend in the predictive variable is analyzed using a backward approach. From 40 weeks before death, an accelerated decline in vital parameters such as albumin, systolic blood pressure, and body weight has been observed  (fig. 7).
|Fig. 7. Linear split splines of serum albumin before death. Knot point at 3 months before death. From Kotanko et al.  with permission.|
These trends were recently confirmed in a large international database (MONDO consortium; submitted data). The relevance of this approach resides in its potential ability to construct dynamic risk models, which serve as an alert system to the treating physician . Future studies should address whether easily applicable dynamic risk models can be developed which are applicable to individual patients.
In summary, both oscillatory and longitudinal time patterns may be altered in dialysis patients. The alterations may have diagnostic, therapeutic, and prognostic implications. However, these alterations are also reflective of a deeper disturbance in homeostatic/homeodynamic control in this patient population. Further research is needed and may lead not only to clearer quantitative indicators but also to new or extensions of current dialysis processes and practices, as well as to the development of pharmacological and rehabilitative approaches to therapy.
Jeroen P. Kooman, MD, PhD
Department of Internal Medicine, Division of Nephrology
University Hospital Maastricht
PO Box 5800, NL–6202 AZ Maastricht (The Netherlands)
Published online: August 8, 2012
Number of Print Pages : 15
Number of Figures : 7, Number of Tables : 1, Number of References : 131
Kidney and Blood Pressure Research
Vol. 35, No. 6, Year 2012 (Cover Date: February 2013)
Journal Editor: Lang F. (Tübingen)
ISSN: 1420-4096 (Print), eISSN: 1423-0143 (Online)
For additional information: http://www.karger.com/KBR