Intra-dialytic hypotension and blood volume and blood temperature monitoring


Dr Kevan Polkinghorne, Department of Nephrology, Monash Medical Centre, 246 Clayton Road, Clayton, Vic. 3168, Australia. Email:


Intra-dialytic hypotension (IDH) is a common problem affecting haemodialysis patients. Its aetiology is complex and influenced by multiple patient and dialysis factors. IDH occurs when the normal cardiovascular response cannot compensate for volume loss associated with ultrafiltration, and is exacerbated by a myriad of factors including intra-dialytic fluid gains, cardiovascular disease, antihypertensive medications and the physiological demands placed on patients by conventional haemodialysis. The use of blood volume monitoring and blood temperature monitoring technologies is advocated as a tool to predict and therefore prevent episodes of IDH. We review the clinical utility of these technologies and summarize the current evidence of their effect on reducing the incidence of IDH in haemodialysis population.


Intra-dialytic hypotension (IDH) is one of the most common problems affecting chronic haemodialysis (HD) patients. It is defined as a fall in systolic or mean arterial pressure of more than 20 mmHg that results in clinical symptoms,1 and occurs in 20–30% of treatments.2 Its aetiology is still incompletely understood. However, it is likely to be multifactorial and include a combination of patient and dialysis factors such as poor cardiac function, inter-dialytic fluid gains, incorrect ideal body weight (IBW), excessive ultrafiltration (UF) and the short duration of conventional HD. Recurrent episodes of IDH are associated with significant morbidity as well as mortality.3,4

It is hypothesized that IDH develops when the normal cardiovascular responses (vasoconstriction and increase in cardiac output) cannot compensate for large volume losses that can occur with UF5. This happens when the UF rate exceeds the plasma refilling rate and persists for long enough to reach a critical threshold in the reduction of blood volume (BV).6 This critical threshold of BV differs in individual patients and is influenced by the integrity of the compensatory cardiovascular response.7 An impaired response may lead to cardiac under-filling, activation of the simpatico-inhibitory cardiopressor reflex and sudden hypotension.8 The rise in temperature observed in conventional dialysis opposes the normal cardiovascular response to volume loss, contributing further potential for cardiovascular instability.

Intra-dialytic hypotension is commonly associated with minor symptoms such as cramps, nausea and vomiting. Recurrent episodes of IDH cause frequent interruptions to HD, the inability to attain IBW and consequently result in fluid overload. Chronic fluid overload can lead to hypertension and increased cardiac output, resulting in left ventricular hypertrophy. This increases the risk of cardiovascular mortality and morbidity.9 IDH also causes a reduction in diastolic blood pressure and decreased cardiac perfusion, which can lead to myocardial ischaemia.10 Long-term IDH has been linked to the development of cardiac fibrosis, which predisposes to reduced left ventricular compliance and arrhythmias.11 Sudden cardiac death is a major cause of mortality (up to 15%) in long-term HD patients.12

Given the large impact of IDH on HD patients, research has focused on ways to identify patients at risk, and predict and prevent future episodes. Simple strategies such as to minimizing sodium and fluid intake to prevent excessive inter-dialytic fluid gains, regular review of medications and frequent assessment of IBW are important in reducing IDH, but alone are often insufficient to prevent IDH. The last two decades have seen the introduction of dialysis machine-based technology aimed at reducing or predicting IDH. The focus of these machine modules has been on the monitoring and modulation of blood volume (BVM) or blood temperature (BTM) with real-time feedback that can be manual or automated.13 BVM techniques use changes in haematocrit to provide a measure of the relative change in BV. BTM allows for the modulation of temperature during dialysis in order to improve existing cardiovascular responses during dialysis. Here we review the clinical data on the utility of such techniques in predicting and preventing IDH.


In renal failure sodium retention and the subsequent increase in total body sodium leads to an expansion of the extracellular volume. Fluid overload is defined as the excess in extracellular volume above that is found in normal subjects.14 The extracellular fluid is predominantly located in the interstitial and intravascular compartments. Removal of fluid during HD occurs from the intravascular compartment through UF. This results in a loss of BV that is offset by the plasma refilling rate, which occurs via capillary uptake from the interstitium.14

There is a strong association between high UF rates and the incidence of IDH.15 High UF rates are often the product of short dialysis times restricting conventional HD. They are further exacerbated by patient comorbidities, cardiovascular disease and autonomic instability, high intra-dialytic weight gain and the prescription of multiple antihypertensive medications. The importance of the UF rate in the aetiology of IDH is highlighted by the lower incidence of IDH observed in short daily and nocturnal home HD patients.16 More frequent treatments result in lesser intra-dialytic weight gains and therefore a lower rate of UF per treatment. This avoids the excessive falls in plasma volume associated with higher UF rates.

The dry weight or IBW can be simply defined as the lowest weight tolerated by the patient without manifesting any symptoms, and is in theory analogous to the patient's normal physiological weight. In clinical practice IBW and the target UF volume are usually determined by the clinical assessment of fluid status and degree of inter-dialytic weight gain. While clinical assessment is adequate in determining the IBW in most situations, it is unable to predict which patients will develop IDH and the onset of episodes in these patients.

Modulation of blood volume has been developed to allow better assessment of IBW and to predict and prevent episodes of IDH. BVM devices (such as Crit-line® or Hemoscan®) use light to continuously measure haematocrit or haemoglobin values. A reduction in BV results in a greater concentration of haematocrit or haemoglobin and a lesser passage of light.17,18 The relative blood volume (RBV) is a measure of the BV at a given time and is expressed as a percentage of the volume at the commencement of treatment.19 With volume overload, there is a relatively small change in RBV with fluid removal and therefore fluid removal is usually well tolerated. As the patient approaches IBW, there are more significant changes in RBV with equivalent UF prescriptions. It is the slope of the RBV curve rather the absolute value that can provide information about the patient's haemodynamic stability.20 The concept of a critical RBV that predicts IDH was found to vary markedly from patient to patient, and between treatments in the same patient.21 Early studies demonstrated that the RBV curve decreases more rapidly in dialysis sessions with IDH,22 and that changes in RBV can be used to predict and therefore prevent episodes of IDH.23,24,25

Several small studies have suggested BVM devices may be useful to predict IDH and allow intervention to prevent subsequent episodes (Table 1).27,28,30 In a prospective, randomized cross-over trial of 12 IDH-prone patients, BVM was compared with conventional dialysis monitoring.28 The incidence of IDH in patients having dialysis sessions using BVM was 33.3%, compared with 81.9% in patients who were using conventional monitoring. A randomized cross-over trial of 36 hypotension-prone dialysis patients comparing BVM and conventional dialysis showed a 30% reduction in the incidence of IDH when patients received treatment with BVM.27 This finding was more pronounced in patients with symptomatic IDH and the absence of inter-dialytic hypotension. In a multicentre prospective study BVM was used to assess RBV reduction during HD and to establish clinical predictive factors.21 123 HD patients were divided into IDH-prone, normotensive and hypertensive groups. There was no difference in the RBV curves among the three groups and no critical RBV level for predicting IDH was identified.

Table 1.  Summary of randomized BVM trials, characterized by study population
StudyPatient noStudy designIntervention (control)Primary end pointOutcome
  1. BVM, modulation of blood volume; ECV, extracellular volume; HBS, Hemo-biofeedback systems; IDH, intra-dialytic hypotension; RCT, randomized controlled trial; RR, risk ratios.

Non-IDH-prone population     
 Reddan (2005)26443RCTCrit-line (conventional monitoring)MorbidityIncreased hospitalization in BVM group, Adjusted RR 1.61 (95% CI 1.15–2.25)
 Nersallah (2008)1960RCTHemo-biofeedback systems (conventional monitoring)Change in ECV at 6 monthsLower IDH in HBS group (0.13 vs 0.31) P = 0.04)
IDH-prone population     
 Santaro (2002)2736Cross-over RCTHemoscan (conventional haemodialysis)IDH reduction30% reduction in IDH, (P = 0.004) sessions in BVM group
 Ronco (2000)2812Cross-over RCTHemoscan (conventional haemodialysis)IDH reductionLess IDH in BVM group (24 vs 59 HDx sessions P ≤ 0.001)
 Gabrielli (2009)2926Cross-over RCTFresenius 4008HD (conventional monitoring)Intra-dialytic morbidityReduction in IDH in BVM group (24% vs 32%, P = 0.04)

The effect of BVM on morbidity and hospitalization rates in HD was assessed in 443 HD patients randomized to 6 months of BVM (n = 227) or conventional monitoring (n = 216).26 In contrast to most previous studies, the patients were not selected on the basis to being prone to IDH. More non-access-related hospitalizations were seen in the BVM compared with conventional monitoring groups (120 vs 81 episodes). The unadjusted and adjusted risk ratios for non-access-related hospitalizations were 1.49 (95% CI, 1.07–2.08, P = 0.017) and 1.61 (95% CI, 1.15–2.25. P = 0.01), respectively. The adjusted risk ratios for cardiovascular admissions was 1.85 (95% CI, 1.19–2.86, P = 0.006). Mortality at 6 months was greater in the BVM than the conventional monitoring group (8.7% and 3.3%, respectively; P = 0.021 by log–rank test). The results of this study, the largest prospective, randomized trial published, conflict with previous smaller studies. Possible explanations offered for the increased rate of hospital admissions observed in the BVM group were increased vigilance and subsequent interventions to improve outcomes. This was contradicted by the increased mortality in the BVM group. It was noted that the conventional monitoring group had a lower than expected mortality and hospitalization rate, which may have exacerbated the differences between the two groups. However, the biggest determinant and likely explanation is that unlike previous trials the study population was not limited to those with clinical issues of volume management and haemodynamic instability.

In addition, recent work has also examined the assumption the relationship between the afferent haemoconcentration, observed RBV and the total blood volume (TBV). The RBV measurements determined by the haemoconcentration of afferent blood can adequately represent the TBV only if there is uniform mixing of plasma and erythrocytes throughout the different vascular beds of the circulation.31 The authors demonstrate that this assumption is incomplete as the whole-body haematocrit is lower than the haematocrit of arterial or venous blood and that this ratio also changes during HD.32 The observed RBV will therefore differ significantly from the TBV and therefore introduce errors in the assessment of the patients risk of IDH.33

BVM- derived changes in RBV values have been shown to variably over and underestimate the RBV when compared with laboratory-derived values.34 The three most commonly used BVM devices, Crit-line, Haemoscan® and Fresenius® BVM, were compared with each other and to laboratory-derived BV changes (based on changes in haemoglobin).32 All three devices yielded values different from the laboratory-derived values and there was also significant variability between the three devices. This possibly reflects the different methods by which the changes in BV are acquired.

Modulation of blood volume has been used to assess the different rates of UF on RBV. UF profiles and rates vary from constant, high at onset and isolated pulses. The highest rate of IDH was found in dialysis sessions where UF occurred in pulses or steps.35 Attempts have been made to measure the changes in RBV over a series of sessions and store this in the dialysis machine so that UF can be adjusted once the RBV reaches a patient-specific threshold. However, the RBV adjusted for UF varies greatly between dialysis sessions reflecting different UF requirements.36 The more fluid overloaded a patient, the smaller the decrease in RBV per unit of UF volume.36,37

This technology has been expanded to create a preferred UF profile for an individual patient based on stored RBV measurements obtained from these patients. During HD the dialysis machine checks the RBV measurement against the stored profile and adjusts the UF rate and dialysate sodium concentration accordingly. This uses fuzzy logic principles, which aim to derive a definite conclusion from what is often imprecise or ambiguous data. This aims to mimic human decision making allowing a degree of flexibility not possible with mathematical modelling.38 After an initial successful single centre experience39 the biofeedback system technology has been shown to reduce the incidence of IDH in several randomized trials.19,29,40 A recent study aimed to assess to utility of UF index (UF rates divided by post-dialysis weight), RBV slopes and volume index (RBV slopes adjusted for UF rate and weight) in determining BV status in 150 difficult patients.41 While these were shown to be possible markers of volume status they did not predict the onset or frequency of IDH. The use of RBV slopes has been shown to be useful in the assessment of IBW in hypertensive HD patients.42

Various BVM technologies are now readily available; however, their utility in IDH remains unclear. BVM devices (especially with the addition of fuzzy logic systems) decrease the incidence of IDH in a at risk population; however, there is limited evidence that BVM can predict IDH in individual patients or that there is a long-term morbidity and mortality benefit, especially in the wider HD population. The technology is undergoing constant refinement, as is the interpretation and analysis of the RBV curves in relation to the other parameters such as weight, UF rate and sodium concentration. An integrated device may enable better prediction and prevention of IDH in individual patients. BVM may have a beneficial role in the assessment of IBW.


The effects of temperature on HD stability were first observed in the 1980s43 with the recognition that body temperature rises during dialysis.44 This is believed to be secondary to the compensatory response to loss of plasma volume, resulting in a reduction in blood flow to the skin and an increase in the total peripheral resistance leading to vasoconstriction and heat retention.45 Additional mechanisms include heat transfer from the dialysate to the patient, and a possible inflammatory response from the interaction of blood and extracorporeal circuit.15 The rise in temperature interferes with the normal response to UF by causing concurrent vasodilatation, which opposes the normal cardiovascular response to fluid removal. This contributes to haemodynamic instability, the threshold for which differs in individual patients.46 Multiple studies have shown that cool dialysis with a dialysate temperature of 34–35°C has confers greater cardiovascular stability than a dialysate temperature of 37°C or higher.44,45,47–51

Biofeedback devices have been developed to measure the BTM in the arterial and venous circuits (which allow for recirculation) and feedback the information to arterial and venous thermostats in the machine, allowing for modulation of the dialysate temperature. The machine can be programmed to allow for a constant body temperature and a negative overall energy transfer termed isothermic HD.52 This is contrasted with thermoneutral HD, which aims to prevent energy transfer between the dialysate and extracorporeal blood.53

One of the first large trials to show a benefit of isothermic dialysis over thermoneutral dialysis was the European Randomized Clinical Trial during which 116 hypotension-prone dialysis patients were randomized in a cross-over design, comparing isothermic dialysis with thermoneutral dialysis.53 A median of 6 of 12 dialysis sessions in the thermoneutral group, compared with 3 of 12 in the isothermic group, were complicated by IDH (P < 0.001). The observed body temperature nadir was higher than observed in other studies and this may have contributed to the overall favourable tolerance of the intervention. There were no significant side effects or discontinuation of dialysis due to cold or shivering.

Selby and colleagues performed a systematic review assessing the clinical effects of reducing dialysate temperature.2 A total of 22 randomized studies (the majority were blinded and unblinded cross-over designs) in 408 patients were examined. Sixteen studies (235 patients) assessed a fixed empirical reduction in temperature while the remaining 6 (173 patients) examined isothermic cooling or programmed cooling with BTM. In the fixed temperature group the standard dialysate temperature varied between 36.5°C and 38.5°C with the majority using 37.5°C. In the intervention group the cool dialysate temperature varied between 34°C and 35.5°C with the majority of studies using 35°C or 35.5°C. In the BTM group patients underwent programmed cooling or isothermic dialysis, the temperature in the intervention group that underwent programmed cooling varied between 35.3°C and 35.7°C. The stability of the patients during HD also varied with a mixture of stable and unstable patients studied. A total of eight studies addressed the issue of IDH and cool temperature dialysis either using a fixed temperature reduction (6) or BTM (2).45,53–57

The overall rate of IDH was 7.1 times greater than in conventional dialysis (95% CI, 6.7–12.4) compared with thermo-regulated HD. In studies examining fixed temperature reduction the rate of IDH was 9.5 times less compared with the control while for those studies comparing isothermic cooling or programmed cooling the rate was 2 times less. When the data were adjusted for studies that had no IDH in the intervention group,45,56 the overall rate of IDH in cool dialysis was 2.6 times less compared with conventional dialysis (95% CI, 1.5–3.8). There was also a benefit on blood pressure post dialysis, with the higher values observed in cool dialysis, attributed to increased total peripheral resistance. There were no differences in symptoms as reported by the patients.

The issue of the optimal magnitude of temperature decrease was addressed in a recent trial (not included in the systematic review).58 Fourteen patients with a history of IDH were studied in a cross-over randomized trial. Isothermic dialysis was compared with ‘cooling’ dialysis (decrease core temperature by 0.5°C), with thermoneutral dialysis used as the control. The nadir of systolic blood pressure (SBP) during isothermic and thermoneutral dialysis was lower than during ‘cooling dialysis’ suggesting that greater stability is conferred by a small decrease in core body temperature.

Temperature control can improve blood pressure stability in a IDH-prone population without causing discomfort or morbidity. The procedure is simple, safe and efficient to use. The early concerns regarding dialysis quality have not materialized; however, long-term prospective validation is lacking. The precise temperature at which the benefit is derived needs to be balanced with symptoms of hypothermia. It is also likely that individual patients have a different temperature threshold at which a benefit to haemodynamic stability is conferred. More studies using the BTM devices are needed to further establish its role, especially in the adjustment of core body temperature based on the individual patient susceptibility to IDH. This would ideally occur in the form a randomized trial comparing fixed temperature reduction, isothermic dialysis and dialysis with a small decrease in core body temperature. Future studies of temperature controlled dialysis need to show a reduction in morbidity and mortality as well as a cost benefit in reducing hospitalization rates.


Despite its multifactorial aetiology the main cause of IDH is the decrease in BV due to the UF rate exceeding the capillary plasma refilling rate. It is unlikely that any single treatment option will significantly alter patient outcomes, but rather incremental gains will be achieved with an integrated, multidisciplinary approach. BVM devices have had a moderate effect on the reduction of the incidence of IDH; however, its effects are limited to an at-risk population. The expansion and integration of these technologies to create an individual patient dialysis profile may prove more successful. The role of cool temperature dialysis shows greater promise in reducing IDH; however, there is still uncertainty about the necessary reduction in temperature to achieve optimal results. With the technologies available today, BTM technology is more mature and offers a relatively simple and effective means of combating IDH in susceptible patients. The widespread use of BVM and BTM monitoring in the general HD population, not prone to IDH, cannot be supported with the evidence currently available. Ultimately, these technologies will need to be trialled in combination, in a way that demonstrates a mortality and morbidity benefit, and to effectively allow the determination of an individualized HD profile that can account for the multitude of dialysis and patient factors that contribute to IDH.