Chronic kidney disease (CKD) is a health problem that in Spain affects more than 60,000 patients undergoing renal replacement therapy, including dialysis and kidney. transplantation.1 The survival of people with CKD, especially those on dialysis, is lower than that of the general population.2,3 The main cause of the high mortality is the increased risk of cardiovascular disease,4,5 which is due both to the high prevalence of classic cardiovascular risk factors (cardiovascular disease, hypertension, diabetes mellitus),6 and to specific factors of uremia, known as emerging, among which are anemia,7 bone mineral disease and loss of residual kidney function, as well as complications associated with factors of the dialysis technique itself, among which we find excessive ultrafiltration (UF) and intradialysis hypotension.8,9

The cardiovascular risk factors related to the hemodialysis technique itself are potentially modifiable. Although urea kinetics (KT/Vurea) is a marker of adequate dialysis, in some centers it conditions the duration of each therapy,10 which could lead to short sessions if based only on this indicator. This model is common in the United States where by following a correct KT/Vurea, short dialysis sessions are frequent increasing cardiovascular risk,11 among other reasons, short dialysis requires high hourly UFs that may increase the risk of hypotension.12

UF is a modifiable parameter and it has been described that it has a direct impact on the survival of the patient.13 It is determined by the amount of fluid removed or ultrafiltered, and varies with the interdialytic weight gain (IWG) and the rapidity in which it is removed. A high hourly weight-adjusted ultrafiltration rate (UFR/W) increases the probability of arterial hypotension and the risk of cardiovascular mortality in conventional hemodialysis.14,15,16–18 An increase in mortality has been described in patients with an hourly ultrafiltration ratio (UFR) of 6 mL/kg/h, and the increased mortality is being significant with UFR above 10 mL/kg/h.19

Episodes of hypotension intradialysis are associated with higher cardiovascular mortality,20 arrhythmias, myocardial microinfarctions21 which are the cause of “myocardial stunning”,22 in addition to cerebral microinfarctions. The greater susceptibility to myocardial ischemia is due to the high prevalence of coronary atheromatosis and ventricular hypertrophy in these patients,23 even in the absence of significant coronary lesion.24,25 Hypotension is also associated with an increased risk of thrombosis of the vascular access.26

Presently, the UFR has been adjusted to the patient's weight, and it is expressed as mL/kg/h26 and different studies have established a threshold of UFR/W of 10–13 mL/kg/h as the maximum rate, more fluid removal will require an increased in the duration of dialysis sessions, since above these limits there has been described an association with excess mortality.19,27,28 However, many of these studies come from populations where the standards of health care (dialysis guidelines, access to health care and medication) or the characteristics of the population are not equivalent to Spain. Therefore, we have addressed the relationship between UFR/W and mortality in a dialysis unit in Spain.

Given that the results obtained on the relationship between UFR/W and mortality do not coincide with the previous reports in the literature, we analyzed whether there were differences in the characteristics of the different populations being evaluated.

Assimon et al.19 concluded that a cut-off point of UFR/W > 13 mL/kg/h is associated with higher mortality, despite the fact that, like us, they observed higher UFR/W in younger patients with lower weight. However, comparing our results with those of this study, we found that our patients with UFR > 13 mL/kg/h in ≥50% of the sessions had lower dry weight (50.9 vs. 73.5 kg; p < 0.001), lower IWG (2.61 vs. 3.6 kg; p = 0.001) and a not so high mean UFR (13.41 mL/kg/h vs. 16.0 mL/kg/h; p < 0.001). There were no differences regarding age or duration of the sessions (Table 6).

We also compare our results with those of Flythe et al.29 and we observed that our patients had lower IWG (2. 6 vs. 3.6 kg; p < 0.001) and lower UFR (13.4 vs. 16.8 mL/kg/h; p < 0.001) (Table 6).

Regarding the use of drugs, we compared our results with those of the Flythe study.29 It is noteworthy that our patients, globally, have more use of beta-blockers, but this significant difference disappears in the group with ≥ 50% sessions with UFR > 13 mL/kg/h, perhaps due to our low “n” value of 19 subjects. There are no differences in the use of angiotensin converting enzyme inhibitor/angiotensin II receptor antagonists (ACEI/ARA2) or calcium antagonists (Table 6). Our patients had a higher prevalence of cardioprotective medication, specifically beta-blockers, which reduce mortality on dialysis,30 and improve the control of blood pressure: our mean systolic blood pressure (SBP) was 144 ± 19.3 mmHg in the group of ≥ 50% of sessions with UFR/W > 13 mL/kg/h, while in the other groups with higher UFR, 49.1% had SBP > 151 mmHg and 40.9% had SBP > 160 mmHg.19,29

In the present study, we did not find statistically significant differences in mortality in relation to the UFR/W. As expected, we observed an association between mortality and the presence of more und basic comorbidities. The with mortality is higher in older patients, patients with a history of ischemic heart disease and peripheral arterial disease, use of anticoagulants, a reduced LVEF, longer time of permanence in chronic hemodialysis, lower pre and post-dialysis SBP, poor nutritional status, lower weight, lower percentage weight reduction during dialysis and high potassium (Table 4).

High UFRs are often an obligatory consequence of high IWG and short hemodialysis sessions, with a relationship between high UFR and cardiovascular mortality.14–19 By increasing the duration of hemodialysis sessions and decreasing the IWG, a reduction of UFRs could be achieved. Currently, strategies based on long dialysis sessions for better blood volume control have fallen into disuse in certain countries, although not in ours B. Thus are remarkable results of the study by the group of Tassin in which, with long hemodialysis sessions (eight hours) achieve beneficial effects, accomplishing a tight control of blood volume, with reduced antihypertensive drugs and gradual reduction of body weight.31,32 By increasing the time or frequency of hemodialysis sessions (dialysis doses), the aim is to reduce the risk of arterial hypotension. The most important determinant of intradialysis hypotension is the decrease in plasma volume below a critical threshold as a result of an imbalance between UF and refilling (vascular filling with water from the extravascular space).

Factors that predispose to intradialysis hypotension include diastolic dysfunction, atrial fibrillation, baroreceptor dysfunction, low sodium and calcium concentrations in the dialysate, high dialysate temperature, and hypotensive drugs, among others. The classic strategy to reduce the number of intradialysis hypotensive episodes is to adjust or decrease the UFR/W. UFR/W rates greater than 10 mL/kg/h are associated with an increased risk of intradialysis hypotension and mortality.33 Some authors suggest that using low UFR/W (3–4 mL/kg/h) should allow a safe reduction in body weight and good blood pressure control even without drugs.34 The UFR/W rate has been defined in different ways, which means that the results of the different studies cannot be extrapolated. Based on previous literature, a high UFR/W rate can be defined as the mean over a period or a percentage of sessions with a high UFR/W, usually greater than 10 mL/kg/h, as discussed below.

Among the main studies analyzing the relationship between mortality and UF are four observational studies. In the Dialysis Outcomes and Practice Patterns Study (DOPPS), Saran et al.33 demonstrated an association between UFR/W > 10 mL/kg/h and total mortality (hazard ratio [HR] 1.09, p = 0.02) in 22,000 hemodialysis patients. In this study, longer sessions and higher Kt/Vurea were associated with lower mortality, while UFR/W > 10 mL/kg/h were associated with higher mortality. Hemodialysis sessions of more than 240 min were significantly associated with a lower relative risk (RR) of mortality (every 30 min of the hemodialysis session decreased the RR of mortality by 7%). Furthermore, they suggest that long hemodialysis sessions could reduce cardiovascular morbidity and mortality. In addition, Movilli et al.9 conducted a prospective study in 287 hemodialysis patients, and described that survival decreases with high UF rates, especially with UFR/W greater than 12.4 mL/kg/h, for which they recommend increasing the time or frequency of hemodialysis sessions to avoid high UFR/W and improve survival. In a post hoc analysis of 1,846 patients from the HEMO study (Hemodialysis Study), Flythe et al.29 found that UFR/W > 10 mL/kg/h were associated with an increase in cardiovascular and all-cause mortality, although this effect was more marked with UFR/W > 13 mL/kg/h (mortality of any cause HR 1.59, 95% CI 1.29–1.96; cardiovascular mortality HR 1.71, 95% CI 1.23–2.38). Furthermore, the work of Assimon et al.,19 who analyzed UF in 118,394 hemodialysis patients in a retrospective cohort study, observed that UFR/W > 13 mL/kg/h during hemodialysis sessions of the first 30 days of follow-up was associated with higher mortality, and that an increase in UFR/W of 1 mL/kg/h was associated with an increased risk of death of 3%.

In our work, we studied the mean and the percentage of sessions that reached an excessive UF, defined as UFR/W > 10 mL/kg/h, during the 30 months of the study. Our mean total UFR/W was 9.0 mL/kg/h, but 59% (127) of the patients had a UFR/W > 10 mL/kg/h in more than 25% of the sessions and a 8.8% (19) had a UFR/W > 13 mL/kg/h in ≥50% of the dialysis sessions (Tables 2 and 3). Analysis of these data leads us to conclude that we could have increased the duration or frequency of hemodialysis sessions to avoid high rates of UF and improve survival. This data allows to estimate that, by adjusting the UFR in each session for each patient according to their dry weight, IWG and baseline comorbidities, the potential individual risk associated with the most demanding UFR/W could be reduced. In our sample, patients with higher UFR/W were younger, with lower dry weight, higher IWG, lower initial and final weight, higher UF volume and greater percentage of weight reduction in hemodialysis. The characteristics of this population with higher UFR/W may have influenced the low mortality observed since, even when analyzing separately the group with the highest cardiovascular risk and the oldest, we did not notice statistically significant differences. However, we cannot rule out that they may suffer imperceptible damage that increases mortality in a long follow-up. The high UFR/W in patients with lower dry weight are explained by the fact that, in these people, small IWG in absolute values ​​(kg) represent important increases in blood volume with respect to weight and lead to higher UFR/W. To avoid high UFR/W in patients with lower dry weight, it would be necessary to consider increasing the time and/or frequency of hemodialysis sessions.

Probably, the small sample size of our study compared to that of Assimon (19 patients vs. 21,735 with UFR/W > 13 mL/kg/h) represents an important limitation. Mortality may not have statistical significance in our group with the highest UFR/W because, being people with less weight (dry weight 51.4 kg) and younger, they have a lower cardiovascular risk compared to the American population studied by Assimon, which had a dry weight 22.1 kg higher than ours. As compared to the Assimon study, our patients with a higher UFR/W had a lower IWG and, therefore, a lower UFR/W. This is compounded by a probable lower cardiovascular risk due to being underweight and younger than the individuals in the Assimon study. We consider that this may be the explanation why this group, despite having UFR/W > 13 mL/kg/h in ≥50% of hemodialysis sessions, was not associated with a significant increase in mortality. Our mean follow-up was 2.3 years, this is like in the Assimon study, but it would be interesting to analyze the data regarding mortality in the longer term. Contrasting our results with those of the Assimon and Flythe studies, we found that individuals with higher UFR/W are, in general, younger in all three studies. However, even the highest mean UFR/W and IWG (13.41 mL/kg/h and 2.61 kg) in our patients are lower than those of the other two studies, a fact that we consider has significantly influenced the results of our study, since people with higher UFR/W did not have a significant increase in mortality during the period of follow-up. The differences related to weight and age are probably related to the demographic and socioeconomic characteristics of the different populations, especially weight, given the higher prevalence of obesity in the United States.

An elevated UFR/W promotes the development of intradialysis hypotension, loss of residual renal function, ischemic damage, and possible vascular access thrombosis. In addition, high UFR/W contributes to maintain appropriate euvolemia, which is associated with favorable cardiovascular outcomes, including the improvement in left ventricular hypertrophy35 and prevention of heart failure decompensations, with a better survival.36–39 Therefore, it is important to implement individualized strategies in order to guarantee a specific UFR/W for the hemodynamic and cardiovascular situation of each patient, but not exceeding 13 mL/kg/h.

Intradialysis echocardiographic assessment and measurement of the changes in troponin induced by hypotension have shown that cardiac ischemia is a potential mediator of mortality related with UF.40 Excessive UFR/W should be avoided to allow appropriate refilling in each patient, for which different solutions are proposed. Daugirdas has analyzed how UFR in hemodialysis could be optimized41 and suggests that it might be more appropriate to scale UFR rates in relation to body surface area rather than body weight. Another alternative would be to calculate the “ideal” dry weight of the patient by bioimpedance, and made an individual adjustment of the UFR.42 Other authors are performing UF profiles to assess whether UFR can be intensified at the beginning of the hemodialysis session.43

In conclusion, in the patient population of our dialysis unit, probably representative of Spanish, it was not observed a relationship between UFR/W and mortality. These results differ from those reported in the literature. The differences could be explained by a better baseline situation of the patients with the highest UFR/W, as well as differences in the baseline characteristics of the population and the magnitude of the UFR/W versus previous studies that did associate UFR/W with mortality. Obese patients on hemodialysis (HD) tend to have lower mortality due to the reverse epidemiology, while the thinner patients may be malnourished. This t means that, possibly, extreme weights have to be eliminated from the analysis, since they may aggravate or reduce mortality. Obese people will have a lower rate of UF per weight and it also protects them from mortality. In the United States, obese patients are frequent and they may have an impact on the results. Thus, in our population of patients with an appropriate pharmacological protection of cardiovascular system and in which the upper extremes of UFR/W are avoided, we did not observe a negative impact of UFR/W on mortality. Another possible explanation for not finding differences in mortality between the groups, may be a greater care for more vulnerable patients, who would be treated with longer dialysis and a lower rate of UF/hour/weight, in an attempt to improve tolerance; this strategy would approximate their mortality to that of the healthier group. Given that the sample size was relatively small, it would be interesting to propose a Spanish multicenter study to establish cut-off points for UFR/W of increased risk of death adapted to the Spanish reality.

Project financed by the health research fund ISC-III and co-financed with FEDER funds PI16/01298.

The authors have no conflicts of interest to declare.