UNIVERSITY OF CALGARY. Blood Volume Monitoring Guided Ultrafiltration Biofeedback on the Reduction of

Transcription

1 UNIVERSITY OF CALGARY Blood Volume Monitoring Guided Ultrafiltration Biofeedback on the Reduction of Intradialytic Hypotensive Episodes in Hemodialysis by Kelvin Cheuk-Wai Leung A THESIS SUBMITTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE GRADUATE PROGRAM IN MEDICAL SCIENCE CALGARY, ALBERTA DECEMBER, 2016 KELVIN CHEUK-WAI LEUNG 2016

2 Abstract The majority of patients with end stage renal disease rely on hemodialysis (HD) to maintain fluid balance. Unfortunately, rapid fluid removal [ultrafiltration (UF)] often results in symptomatic, intradialytic hypotension (IDH). Our objective was to perform a randomized controlled crossover trial to evaluate the effectiveness of a blood volume monitoring guided UF biofeedback in the reduction of symptomatic IDH. Over the study period, symptomatic patients first had their dialysis prescription and dry weight optimized over a four-week period before being randomized to an eight-week biofeedback intervention or standard control HD. There was a two-week washout period before patients crossed over for a second eight-week study period. There were no differences in the rate of symptomatic IDH, volume status (as measured by electrical bioimpedance), or biomarkers of cardiac stress between the two groups. ii

3 Acknowledgements I would like to thank my amazing supervisors, Dr. Jennifer MacRae and Dr. Robert Quinn for their guidance, patience, and mentorship throughout the research and thesis writing process. I truly cherish the many principles, lessons and encouragement provided throughout this project as well as in my early nephrology career. I am also thankful to my other committee members, Dr. Pietro Ravani and Dr. Henry Duff for their time, thoughtful input, and support throughout this process. I would also like to thank the hemodialysis staff and patients at Carewest Dr. Vernon Fanning Community Hemodialysis, Sheldon M. Chumir Community Hemodialysis, Sunridge Community Hemodialysis, Foothills Medical Centre Hemodialysis, and the Peter Lougheed Centre Hemodialysis for being part of the project. Finally, I would like to thank my wife, Diana for her patience and support while I completed this master s degree program. iii

4 Table of Contents Abstract... ii Acknowledgements... iii Table of Contents... iv List of Tables... vii List of Figures and Illustrations... viii List of Symbols, Abbreviations and Nomenclature... ix CHAPTER 1: INTRODUCTION Overview of Thesis Background Risks of Chronic Hypervolemia Definition and Pathophysiology of Intradialytic Hypotension Intradialytic Hypotension Related Outcomes Strategies to Prevent Intradialytic Hypotension Relative Blood Volume Use of Relative Blood Volume in Biofeedback Technology Statement of the Problem BVM Without Biofeedback BVM with Biofeedback BVM and Isolated UF Biofeedback Summary...10 CHAPTER 2: BLOOD VOLUME MONITORING GUIDED ULTRAFILTRATION BIOFEEDBACK ON THE REDUCTION OF INTRADIALYTIC HYPOTENSIVE EPISODES IN HEMODIALYSIS: STUDY PROTOCOL FOR A RANDOMIZED CONTROLLED TRIAL (MANUSCRIPT #1) Background Methods Study Design Study Setting Study Participants Definition of symptomatic IDH Interventions BVM-guided UF Biofeedback System in the Fresenius Determination of the Critical Blood Volume Weight adjustments Responding to episodes of IDH Outcomes Primary Outcome Secondary Outcome Sample Size Calculation Recruitment Assignment of Intervention, Allocation Concealment, and Implementation Blinding Data Collection...24 iv

5 2.8.1 Collection of Baseline and Run-in Period Data Collection of Primary Outcome Variables Collection of Secondary Outcome Variables Participant Retention and Follow up Statistical Analysis Ethics Discussion...27 CHAPTER 3: BLOOD VOLUME MONITORING GUIDED ULTRAFILTRATION BIOFEEDBACK ON THE REDUCTION OF INTRADIALYTIC HYPOTENSIVE EPISODES IN HEMODIALYSIS: RESULTS OF A RANDOMIZED CROSSOVER TRIAL (MANUSCRIPT #2) Background Methods Overview Study Population Study Protocol Dialysis Prescription Study Endpoints Data Collection Sample Size calculation Statistical Analysis Results Cohort Formation Cohort Characteristics Run-in Period Primary Outcome: Symptomatic IDH Secondary Outcomes Discussion Summary of Findings Discussion of Prior Studies Strengths and Limitations Implications for Clinical Care Conclusions...49 CHAPTER 4: SUMMARY OF FINDINGS, CLINICAL IMPLICATIONS AND CONCLUSION Summary of Study and Findings Clinical Implications Impact of Biofeedback Technology Learning Curve Lessons from the Implementation of Biofeedback Technology IDH was Common despite Low UFR and IDWG True Dry Weight was Difficult to Determine Use of Surrogate Outcomes Conclusions...68 REFERENCES...69 v

6 APPENDIX A: NURSING IDH MANAGEMENT ALGORITHM...82 APPENDIX B: WEEKLY DIALYSIS ORDERS (CONTROL DIALYSIS)...83 APPENDIX C: WEEKLY DIALYSIS ORDERS (BIOFEEDBACK/INTERVENTION DIALYSIS)...84 APPENDIX D: CONTROL DIALYSIS REFERENCE INSTRUCTIONS...85 APPENDIX E: BIOFEEDBACK/INTERVENTION DIALYSIS REFERENCE INSTRUCTIONS...86 APPENDIX F: REFERENCE BIOFEEDBACK/INTERVENTION DIALYSIS ALARM/TROUBLESHOOTING GUIDE...87 APPENDIX G: DIALYSIS RECOVERY TIME SURVEY...89 APPENDIX H: COPYRIGHT PERMISSION FOR MANUSCRIPT # vi

7 List of Tables Table 2 1: Data Collection Schedule Table 3 1: Baseline Patient Characteristics Table 3 2: Average Weights, Interdialytic Weight Gain (IDWG), Percentage Reaching Dry Weight, and Ultrafiltration Rate Table 3 3: Rates of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone by Period Table 3 4: Rates of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone by Intervention Table 3 5: Proportion of Hemodialysis Sessions Affected by Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone by Intervention Table 3 6: Average Fluid Content, Fluid Ratios, Brain Naturetic Peptide, Troponin, Post Dialysis Recovery Time, and Change in Standing Heart Rate vii

8 List of Figures and Illustrations Figure 2 1: Anticipated Participant Flow Diagram Figure 2 2: Depiction of blood volume monitoring (BVM) profile and the corresponding ultrafiltration profile Figure 3 1: Participant Flow Diagram Figure 3 2: Rate of Symptomatic IDH Events Per Hour, By Randomization Order and Study Period Figure 3 3: Rate of Symptomatic IDH Events Per Hour, By Intervention and Study Week Figure 3 4: Rate of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone Per Hour, in the Run-In and Study Periods Figure 3 5: Proportion of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone Per Hour, in the Run-In and Study Periods viii

9 List of Symbols, Abbreviations and Nomenclature Symbol BNP BVM CI CONSORT DRIP ECW ECW:ICW ECW:TBW ESRD HD hs-cardiac troponin ICW IDH IDWG RBV SD SPIRIT TBW UF UFR Definition Brain Natriuretic Peptide Blood Volume Monitoring Confidence Interval Consolidated Standards of Reporting Trials Dry-Weight Reduction In Hypertensive Hemodialysis Patient Extracellular Water Extracellular Water to Intracellular Water Ratio Extracellular Water to Total Body Water Ratio End Stage Renal Disease Hemodialysis High Sensitivity Cardiac Troponin Intracellular Water Intradialytic Hypotension Interdialytic Weight Gain Relative Blood Volume Standard Deviation Standard Protocol Items: Recommendations for Interventional Trials Total Body Water Ultrafiltration Ultrafiltration Rate ix

10 Chapter 1: INTRODUCTION 1

11 1.1 Overview of Thesis This manuscript-based thesis is divided into four chapters. The present chapter presents a summary of the current literature and the challenges that exist in the management of hemodialysis (HD) patients. In Chapter Two: Blood Volume Monitoring Guided Ultrafiltration Biofeedback on the Reduction of Intradialytic Hypotensive Episodes in Hemodialysis: Study Protocol for a Randomized Controlled Trial (Manuscript #1) 1, the first of two manuscripts is presented. This manuscript describes the study protocol for a randomized controlled trial assessing the use of blood volume monitoring (BVM) guided ultrafiltration (UF) biofeedback in reducing symptomatic IDH episodes. The technical aspects of the biofeedback mechanism, determination of the critical blood volume and the plan for data collection and data analysis are detailed in this chapter. In Chapter Three: Blood Volume Monitoring Guided Ultrafiltration Biofeedback on the Reduction of Intradialytic Hypotensive Episodes in Hemodialysis: Results of a Randomized Controlled Crossover Trial (Manuscript #2), the second manuscript is presented. The second manuscript reports the results of the trial, including the primary outcome of reduction in the rate of symptomatic IDH using BVM guided UF biofeedback and the secondary analyses that examines the rate and proportion of asymptomatic IDH and symptoms alone, as well as markers of cardiac stress and volume. The implications of the study as well as the study limitations are discussed. Additional information is presented in the appendices. Finally, in Chapter Four: Summary of Findings, Clinical Implication and 2

12 Conclusions, the main study is reviewed in the context of current literature and implications for clinical practices are discussed. 1.2 Background The majority of patients with end stage renal disease (ESRD) are treated with HD for regulation of fluid balance and other native kidney functions 2 4. To maintain euvolemia, HD patients typically undergo fluid removal or UF to remove one to five liters per session Risks of Chronic Hypervolemia Despite technological advancements, patients treated with HD have a five-year mortality rate of 57% 4,5, with early literature attributing deaths due to cardiac and infectious causes 4,5. A growing body of literature has emerged linking chronic fluid overload in HD to hypertension, left ventricular hypertrophy and increased all-cause mortality 6 8. HD patients whose extracellular volume is expanded by more than 15% (approximately 2.5L in a 70 kg person) experience more than a two-fold increase in the risk of all-cause mortality over a 3.5 year period 6. This is further accentuated by the presence of hypertension 9. Furthermore, hypervolemia in HD patients has been associated with an increase in both brain natriuretic peptide (BNP) and left atrial volume, which in turn, has been associated with an increased risk of mortality

13 1.2.2 Definition and Pathophysiology of Intradialytic Hypotension Rapid fluid removal on dialysis results in intradialytic hypotension (IDH) in as many as 25-50% of patients treated with HD 15 and may contribute to the excess mortality seen in chronic hypervolemia 7,16,17. Although various definitions exist 18, IDH is most commonly described as an abrupt drop in systolic blood pressure of 20mm Hg with associated symptoms of cerebral, cardiac, gastrointestinal, or musculoskeletal ischemia The pathophysiology of IDH is multifactorial, resulting from a combination of aggressive UF of large interdialytic volume gains, reduced plasma refill rate and cardiac output, and inadequate autonomic response 22. Aggressive UF causes a decrease in intravascular blood volume that exceeds the plasma refilling rates determined predominantly by plasma sodium and albumin levels. The drop in intravascular blood volume and altered venous compliance due to autonomic dysfunction (which causes veins to dilate) lead to decreased cardiac filling and decreased cardiac output. Autonomic dysfunction also decreases the vascular resistance of arteries, further contributing to hypotension 15,21, Intradialytic Hypotension Related Outcomes Multiple observational studies have shown that the presence of IDH or a drop in the postdialysis systolic blood pressure to be associated with an increase in morbidity and mortality The postulated mechanism is that reduced tissue perfusion results in ischemic damage to vital organs including the heart and brain 16, This was 4

14 demonstrated by a study showing that IDH was associated with more than a three-fold increase in the development of myocardial stunning and regional wall motion abnormalities, which over time progressed to a fixed left ventricular systolic dysfunction 16,33. Furthermore, patients with myocardial stunning are more susceptible to cardiac arrhythmias and have elevated levels of high-sensitivity cardiac troponin (hstroponin), which has been associated with increased cardiac stress and all-cause mortality 16, Strategies to Prevent Intradialytic Hypotension There is no standardized approach in clinical practice with respect to the prevention and management of IDH 21,22,36. Strategies that address the various factors that contribute to IDH have been proposed. To reduce aggressive UF, an increase in the duration or frequency of HD sessions can be employed. Some studies have assessed the use of sequential UF, which allows for plasma refilling by intermittently pausing UF 37,38. The use of sodium and UF profiling to temporarily increase dialysate sodium and UF rate at the start of dialysis (when the patient is most volume overloaded), and then progressively decrease dialysate sodium and UF rate toward the end of dialysis (when the patient is most prone to hypotension), has been studied with some success 39,40. High dialysate calcium has been used to increase cardiac contractility of patients with poor cardiac function 41,42. Cool dialysate has been used to increase peripheral vascular resistance and maintain blood pressure during dialysis via vasoconstriction 43. Convective clearance 5

15 therapies such as hemofiltration or hemodiafiltration are associated with improved hemodynamic stability as a result of peripheral vasoconstriction, either by a reduction in inflammatory mediators or via cooling effect of the replacement fluid 44,45. Intradialytic midodrine is another tool to promote vasoconstriction via its alpha-agonist effects 46. Finally, several different forms of biofeedback have been proposed 47, Relative Blood Volume The rate of fluid removal appears to be an important determinant of IDH, so a proposed strategy to prevent IDH is to adjust the UF rate (UFR) according to the patient s vascular response and their ability to refill. Relative changes in the patient s blood volume can be detected by measuring the concentration of hemoglobin or protein in the blood at the site of the arterial port using validated optical photometry or ultrasound techniques Since the total amount of hemoglobin or protein does not change in HD, any changes in hemoglobin or protein concentration are attributed to intravascular volume changes as a result of UF. In clinical studies, higher rates of UF result in faster declines in the relative blood volume (RBV) However, what has been shown is that patients that who demonstrate a steeper, more negative slope are more likely to develop symptomatic IDH 53 55,58,59,53,60. Given this finding, there have been attempts to correlate IDH and dialysis symptoms to a specific value of RBV (i.e. the critical RBV), but they have been inconclusive 54,57,58. The RBV can easily change as a result of numerous conditions aside from UF. For example, a change in position from supine to sitting will cause the RBV to 6

16 decline, whereas receiving IV medications will cause the RBV to rise 61. Furthermore, it is important to recognize that the RBV, as its name implies, is relative to the blood volume at the start of HD. Therefore, even if a patient is below his or her dry weight or intravascularly depleted at the start of dialysis, that RBV will still be used as the reference value Use of Relative Blood Volume in Biofeedback Technology The use of biofeedback technology has been proposed to reduce IDH and associated sequelae 47,48,62,63,50,51. In biofeedback if a monitored parameter is not within a prespecified range, a computer algorithm automatically adjusts variables to return that parameter of interest to the pre-specified range. In HD, biofeedback technology exists for parameters of core temperature, blood pressure, blood conductivity, or blood volume. Blood volume monitoring (BVM) guided biofeedback is based on the principle that the RBV (when compared to baseline) appears to decrease more rapidly in sessions complicated by IDH 53,60. The critical RBV is set at the beginning of HD. Given the variable nature of the critical RBV, there is no standardized method for selecting a critical RBV. One study determined the critical RBV for the cohort by averaging the previous RBV values of when symptoms occurred 64. As UF occurs, the RBV declines and the dialysis machine biofeedback can adjust either the dialysate sodium, the UF rates, 7

17 or a combination of both to compensate and prevent the RBV from reaching or passing the critical RBV Statement of the Problem Randomized studies assessing the role of BVM and BVM-guided biofeedback in the prevention of IDH are largely limited to adjustment of both dialysate sodium and UFR, an approach that has important limitations. BVM Without Biofeedback One the largest studies assessing BVM was completed without biofeedback. Over a 6- month period, Reddan et al. randomized 443 patients to BVM or conventional HD. Patients in the group with BVM had UF rates adjusted at the discretion of the bedside dialysis nurse with a suggested UF adjustment algorithm. There were no differences in patient symptoms or IDH. In fact, the patients randomized to BVM had an increased mortality, with many deaths attributed to infectious causes, as well as an increase in access-related and non-cardiovascular hospitalizations. It was not clear if the negative finding reflected a lack of adherence to the UF adjustment algorithm, or if it was related to more cautious fluid removal due to the feedback provided by the BVM 62. The intervention required manual adjustment of the UF rate and it was unclear how promptly or frequently the bedside nurses responded to changes in the RBV using the suggested/non-mandatory algorithm. 8

18 1.3.2 BVM with Biofeedback When BVM is combined with biofeedback there may be a beneficial reduction in IDH 47,48,64,65,66. A meta-analysis of six studies using BVM guided biofeedback reported a 39% overall reduction in the number of dialysis sessions complicated by IDH 65, but the current literature has important limitations. First, of the six studies assessing IDH, three did not have IDH as the primary outcome. Second, five studies employed an intervention that included both dialysate sodium and UF biofeedback 48,63, The use of the combination of dialysate sodium and UF biofeedback, compared to UF biofeedback alone, raises concerns regarding the potential for sodium loading and resultant interdialytic weight gain. Increased plasma osmolality leads to increased thirst, which in turn, results in an increased fluid intake and weight gain. Increased interdialytic weight gain means that higher UF rates are required on dialysis, thereby perpetuating the vicious cycle of IDH. Third, the duration of the intervention period in most studies ranged from two weeks to six months, and the majority of the studies used a short two to four-week intervention. As a result, it is difficult to comment on the long-term sustainability of the intervention. Fourth, three studies did not have a washout period between intervention and control groups, and one did not have a run-in period. This raises concerns about the possibility of a carryover effect 65. Fifth, the majority of the studies did not assess for changes in body fluid composition, nor did they address patient quality of life, which is an important issue in 9

19 HD 70. Finally, the available studies are generally of low quality due to poorly described randomization strategies, unclear exclusion criteria, and the inclusion of low numbers of IDH-prone patients 48,63,65, BVM and Isolated UF Biofeedback There has been only one published randomized study that examined the impact of BVMguided UF biofeedback on symptoms during dialysis 64. Twenty-six patients were randomized in a crossover fashion to either six weeks of BVM-guided UF biofeedback or conventional dialysis before crossing over to the other treatment. Dialysis symptoms and hypotension were each reduced by an absolute 8% (20% relative reduction). However, the primary outcome was dialysis symptoms alone with or without hypotension or IDH. Symptoms without hypotension may not be related to the UF process that this technology targets. Similarly, the secondary outcome of hypotension was not well defined and did not fit any previously published standard for IDH, making interpretation of the outcome difficult. There was no run-in or dialysis optimization period prior to randomization, raising concerns about the possibility of inappropriate dry weights prior to study enrollment. Finally, the study lacked blinding given its open label nature. 1.4 Summary Despite advances, the risk of IDH is still a common and serious complication of dialysis therapy 71. A rigorously designed clinical trial, using internationally recognized 10

20 definitions of IDH is needed to test whether biofeedback technology incorporating BVM guided UF adjustments results in a reduction of IDH and patient symptoms. 11

21 Chapter 2: BLOOD VOLUME MONITORING GUIDED ULTRAFILTRATION BIOFEEDBACK ON THE REDUCTION OF INTRADIALYTIC HYPOTENSIVE EPISODES IN HEMODIALYSIS: STUDY PROTOCOL FOR A RANDOMIZED CONTROLLED TRIAL (MANUSCRIPT #1) 12

22 2.1 Background The majority of patients with end stage renal disease (ESRD) are treated with hemodialysis (HD) for regulation of fluid balance and other native kidney functions 2,3. Patients treated with HD have a five-year mortality rate of 57%, with the majority of deaths due to cardiac and infectious causes 4,5. A growing body of literature has emerged linking chronic fluid overload in HD to hypertension, left ventricular hypertrophy and increased all-cause mortality 6 8. To prevent volume overload and maintain fluid balance, one to five liters of fluid must be removed during each HD treatment. Rapid fluid removal, or ultrafiltration (UF), during a short period of time can lead to intradialytic hypotension (IDH) in as many as 25-50% of patients treated with HD 15. IDH is most commonly defined as an abrupt drop in systolic blood pressure of 20mm Hg accompanied by symptoms of cerebral, cardiac, gastrointestinal, or musculoskeletal ischemia Observational studies have shown that the presence of IDH or a drop in the post-dialysis systolic blood pressure are associated with an increase in morbidity and mortality 16, Ultrafiltration on HD leads to a fall in the patient s blood volume. The blood volume can be measured by tracking the changes in hemoglobin or protein concentration at the arterial port during dialysis using optical photometry or ultrasound techniques In clinical studies, higher rates of UF lead to faster declines in the relative blood volume 13

23 (RBV) and as a result, a steeper decline in the blood volume monitoring (BVM) curves BVM-guided biofeedback is based on the principle that the RBV (compared to baseline) appears to decrease more rapidly in sessions complicated by IDH 53,60. As a result, the use of BVM-guided UF biofeedback, whereby the dialysis machine automatically reduces the rate of UF prior to reaching the patient s critical blood volume, has been proposed for the prevention of IDH and its sequelae 47,48,50,51,62,63. We will conduct a randomized controlled clinical trial to test whether biofeedback technology incorporating BVM-guided UF adjustments alone (without adjustments of dialysate sodium concentration), in addition to best clinical practice, results in a reduction in the frequency of symptomatic IDH episodes and patient symptoms compared to best clinical practice alone. Secondary outcomes include frequency of IDH-related interventions, dialysis-related symptoms, dialysis adequacy, volume control, biomarkers of volume overload and cardiac stress, blood pressure, and quality of life. 2.2 Methods Study Design This is a 22-week randomized crossover trial. During the first part of the study (Part 1 - Run-In/Dialysis Optimization Period), eligible participants will undergo a four-week runin period. In this period, all participants will undergo a comprehensive clinical review including clinical weight assessment, antihypertensive medication review, and 14

24 standardization of the dialysis prescription. At the end of the run-in period, participants that still meet eligibility criteria will enter the randomized crossover period. In part two (Randomized Crossover Period), participants are randomized to regular best clinical practice HD (without BVM-guided UF biofeedback; control period) or to best clinical practice plus BVM-guided UF biofeedback (intervention period) for an eight-week period. This will be followed by a two-week washout period and then participants will be crossed over for a second eight-week period. The participant flow chart and timeline is shown in Figure 2 1. The study will be conducted and reported following the Consolidated Standards of Reporting Trials (CONSORT) 2010 guidelines 72. The study protocol was approved by the University of Calgary Conjoint Health Research Ethics Board (Ethics ID: REB ) Study Setting The clinical trial will be held at two tertiary care (Foothills Medical Centre and Peter Lougheed Centre) and three community HD units (Fanning Centre, Sheldon Chumir Centre and Sunridge Centre) in the Southern Alberta Renal Program, Calgary, Alberta Canada Study Participants All participants treated with hemodialysis for more than three months will be screened for eligibility. To be eligible for the study, participants must be >18 years of age, 15

25 medically stable, undergo HD three to four times per week for a minimum of three hours per session, and have >30% of their HD sessions in the preceding eight weeks complicated by symptomatic IDH. Participants with serum sodium of 133mmol/L 73, hemoglobin <80g/L, active malignancy, a history of blood transfusion or hospitalization in the preceding four weeks, routine use of diuretics for volume management, a history of ongoing urine output estimated at greater than or equal to 250ml (one cup) per day, or a planned change in the renal replacement modality during the study period will be excluded. Informed consent will be obtained from each individual that agrees to participate in the study Definition of symptomatic IDH Symptomatic IDH is defined as a drop in systolic blood pressure of 20mm Hg from baseline with associated symptoms 19,20. Symptoms include sudden-onset headache, dizziness, unconsciousness, thirst, dyspnea, angina, muscle cramps, or vomiting. 20, Interventions Part 1 -- Run-In Period: Following enrollment, participants will undergo a four-week run-in period to optimize their dialysis weight, dialysis prescriptions, and to determine critical RBV values (BVM will be enabled). During the first two weeks of the run-in period, participants will undergo medication review and dry weight reduction based on a 16

26 modified protocol from the Dry-Weight Reduction In Hypertensive Hemodialysis Patient (DRIP) trial. 71,74 Part 2 Crossover Study Period: Participants that continue to have more than 30% of their sessions complicated by symptomatic IDH during the Run-In Period are randomized to either eight-weeks of best clinical practice (without biofeedback; control period) or best clinical practice plus BVM-guided UF biofeedback (intervention/biofeedback period), followed by a two-week washout period (using control period HD prescription), before crossing over to the other study period for a second eight-week block. Participants will have clinical assessments of their dry weight at the beginning of each week. Control period: All study participants will be dialyzed with the Fresenius 5008 HD machine (Fresenius Medical Care, Bad Homburg, Germany) using high flux dialyzers. Participants in the best clinical practice (control) period will use the same prescription used during the run-in period: dialysate sodium of 138 mmol/l, dialysate calcium of 1.25 mmol/l, dialysate temperature of 36 o C, and a constant UF rate. Biofeedback will be disabled in this group. Intervention/Biofeedback period: Participants in the BVM-guided UF biofeedback group will have the same prescription as the control group, but will also have the ultrafiltration 17

27 rate automatically adjusted by the Fresenius 5008 HD machine based on the changes in RBV BVM-guided UF Biofeedback System in the Fresenius 5008 The Fresenius 5008 uses an ultrasound monitor incorporated into the machine to detect ultrasonic velocity changes to derive the total protein concentration, which is a sum of plasma proteins and hemoglobin. A temperature monitor is also incorporated to correct for temperature related changes in ultrasound velocity 49,75. Since the total protein does not change, any changes in its concentration is attributed to blood volume changes (Figure 2 2). The RBV is calculated by dividing the initial concentration of total protein by the total protein concentration at any given time, multiplied by This method for measuring RBV has been previously validated with both optical and laboratory hemoglobin techniques 49,52. The HD software in the Fresenius 5008 HD machine adjusts the UF rate based on the critical blood volume entered at the beginning of the dialysis session for each individual patient. There is no adjustment to the dialysate sodium concentration. To determine the actual UF rate, the HD software first calculates the maximum UF rate. The maximum UF rate is two times the total UF divided by the remaining time. The actual UF rate is the maximum UF rate multiplied by the UF co-efficient, which is a number between zero and one. To allow for maximum UF at the onset of HD, the UF co-efficient is one at the start 18

28 of HD and continues to remain at one until the RBV is halfway towards the critical RBV. At the halfway point between the start and the critical RBV, the UF co-efficient decreases linearly. Once the critical RBV is reached, the UF co-efficient becomes zero, resulting in cessation of UF 49,52, Determination of the Critical Blood Volume There is currently no standardized method for determining critical RBV. To standardize this process, a single investigator (KL) will use the following method/algorithm: 1. Identify the most recent episode of symptomatic IDH as per the study definition. 2. The critical blood volume will be equal to the RBV recorded immediately prior to the episode of symptomatic IDH. 3. The critical blood volume will be reassessed weekly Weight adjustments During the crossover period, participants will have weekly assessments of their dry weight by a single study investigator (KL). The rounding dialysis physician will be encouraged to discuss any weight adjustments with study personnel Responding to episodes of IDH In the event that an IDH episode occurs, the bedside dialysis nurse will follow a predefined IDH algorithm (Appendix A: Nursing IDH Management Algorithm) 71. Following 19

29 resolution of an IDH episode as defined in the primary outcome section, UF will resume. In the control group, a constant UF rate will be reset to meet the UF goal. In the intervention group, BVM-guided UF biofeedback will be re-enabled. 2.3 Outcomes Primary Outcome The primary outcome is the rate of symptomatic IDH. The number of symptomatic IDH episodes along with the duration of each dialysis treatment will be captured. The rate of IDH for each session will be calculated by dividing the number of episodes by the duration of the session in hours. The rate of IDH will be calculated for every dialysis treatment. The rate of symptomatic IDH will be measured during each period of the study. By using the rate of symptomatic IDH, rather than frequency or number of IDH episodes per session as done in previous studies, we will be more sensitive to meaningful changes in IDH episodes as multiple episodes of symptomatic IDH can occur in a single HD session 47,48,64. Given the subjective nature of IDH-related symptoms and intervention, in-services will be provided to all nurses at participating dialysis centers to ensure that all events are recorded consistently. 20

30 2.3.2 Secondary Outcome The secondary outcomes of interest are the proportion of HD sessions with symptomatic IDH; rate and proportion of HD sessions with asymptomatic IDH; rate and proportion of HD sessions with dialysis symptoms alone; dialysis adequacy, as measured by single session Kt/V; total body water, extra cellular water (ECW), intracellular water (ICW), and the ECW:ICW as determined by electrical bioimpedance; changes in serum brain natriuretic peptide (BNP); changes in high sensitivity-cardiac troponin (hs-cardiac troponin); changes in antihypertensive medications; and patient reported interdialytic symptom survey. Dialysis adequacy, measured by single-pool Kt/V, will be measured and recorded (usual care) at the end of each dialysis session on the session sheet. Single-pool Kt/v will be calculated using the previously validated, online clearance measurement method, which detects changes in conductivity in the dialysate to reflect the clearance of serum electrolytes and urea 77. Whole body and segmental bioimpedance analysis has been validated for the determination of fluid composition in HD patients Electrical bioimpedance will be performed during the mid-week HD session of weeks 1, 4, 8, 12, 14, 18, and 22 of the study to determine the total body water, ECW, ICW, and the ECW:ICW ratio. 21

31 Biomarkers of cardiac stress, BNP and high sensitivity cardiac troponin; as well as the number and class of anti-hypertensive medication use will be recorded at the mid-week HD session of weeks 1, 4, 8, 12, 14, 18, and 22 of the study. Diuretic use will not be recorded as participants with significant residual renal function are excluded. A survey provided at the beginning of each dialysis will be administered and completed by the patient alone, or with the aid of the dialysis nurse. This validated survey inquiries about the time it took the patient to recover to baseline following a dialysis session (interdialytic period) Sample Size Calculation We have previously estimated that 23% of our HD population have IDH 71. To estimate the sample size, we conservatively assumed that only one IDH episode would occur per run (although it is likely that more than one episode will occur per run, increasing study power). With this approach, and using a 2x2 crossover design, we estimated that a sample size of 20 participants would provide a power of 90% to demonstrate a 30% reduction in the rate of IDH under the biofeedback versus the control, with a two-sided alpha of We will enroll 32 participants to guard against dropout, which can be as high as 30% 64. These estimates are based on simulation studies assuming only one IDH episode per run and two Poisson processes with average rates of IDH equal to 7 (control) and 5 (intervention) episodes over 24 runs. We used R Foundation for Statistical Computing, 22

32 Vienna, Austria. (URL for simulations, and the package clusterpower in R. 2.5 Recruitment Recruitment will take place sequentially at the five participating HD units. From pilot data, we anticipate that approximately 20% of the 400 HD patients in participating units will meet the criteria for IDH 71, and at least half of the eligible participants (40 participants) will be willing to participate in our study 84, Assignment of Intervention, Allocation Concealment, and Implementation Randomization of allocation sequence will be done using computer-generated random numbers under the supervision of a statistician in the Department of Medicine, University of Calgary. Subsequently, the allocation sequence will be inserted into sequentially numbered, opaque, sealed envelopes. Following study consent and enrollment, the envelope containing the allocation sequence will be provided to the patient s HD nurse, who will implement the intervention according to the protocol. 2.7 Blinding Given the nature of our intervention, and the inability to disable the BVM-guided UF biofeedback screens, prompts, and alarms, we will not be able to blind the study 23

33 personnel (KL) or the bedside dialysis nurse. The trial participants will be blinded to the intervention. To reduce bias, the HD nurse will be encouraged to document all intradialytic symptoms and interventions, as well as not to reveal the allocation sequence to the trial participant. In addition, study personnel (KL) will perform dry weight assessments and administer surveys uniformly. 2.8 Data Collection Collection of Baseline and Run-in Period Data Baseline patient demographics (age, gender, HD vintage, race), comorbidities (i.e. congestive heart failure, diabetes mellitus, peripheral vascular disease), cause of renal disease, medications (number and class of anti-hypertensive), laboratory investigations (i.e. serum electrolytes, complete blood count, albumin), HD prescription (i.e. dialyzer type, composition, temperature, blood and dialysate flow rates, dry weight, anticoagulation, use of sodium and/or UF profiles), and number of episodes of IDH over the preceding eight-weeks and run-in period will be extracted from HD charts and local electronic health records by the study investigator Collection of Primary Outcome Variables Both sitting and standing blood pressure will be measured in a standardized fashion at the beginning and end of each dialysis session 86,87. Intradialytic blood pressures will be 24

34 measured in a sitting position every 30 minutes and at the time of IDH related symptoms, as defined in the outcomes section. Sitting blood pressures will be measured with patient seated with feet flat on the floor/foot rest, back against chair with his or her bare arm resting on a support, whereby the midpoint of the upper arm is at the level of the heart. Standing blood pressures will be measured with the patient standing feet flat on the floor, with his or her bare arm resting on a support, whereby the midpoint of the upper arm is at the level of the heart. An appropriate-sized blood pressure cuff, where the cuff bladder length is 80% of the arm circumference, will be used. A Trendelenburg or supine position blood pressure will be accepted if the patient is unable to sit during the pre-specified blood pressure measurement or is experiencing IDH-related symptoms. All blood pressures will be measured using an automated cuff attached to the dialysis machine. Manual blood pressure will be accepted if an automated blood pressure cannot be obtained. Blood pressure, method of measurement, patient position, IDH-related symptoms, and nursing interventions will be recorded by the bedside dialysis nurse Collection of Secondary Outcome Variables Dialysis adequacy, bioimpedance, BNP, hs-cardiac troponin, interdialytic symptom survey, and anti-hypertensive medication use will be collected by study personnel at their pre-specified time frames (Table 2 1). 25

35 2.8.4 Participant Retention and Follow up Following enrollment, every effort will be made to follow up participants until the end of the study period. Participants moving to different dialysis sites across the city of Calgary will be followed unless they move to a site not using the Fresenius 5008 machines at which point they will be censored. Participants who move out of the city will be censored. 2.9 Statistical Analysis We will use mixed-effects regression to test the null hypothesis that the ratio of the rate of IDH episodes (main model exposure) during the intervention period over the rate of IDH episodes during the control period will range between 0.7 and 1.3 (i.e., alternative two-sided hypothesis is that the incidence rate ratio will be 0.7 or lower). We will study the effect of the exposure (biofeedback or control HD) as fixed effect and account for the correlation in the data due to subject using random effects. The multi-level model will have errors due to within-subject variation (within subject variance) and due to betweencluster variation (between subject variance). In the event that the study generates overdispersed data we will use negative binomial regression with the approach as described above. We will also explore whether participants with larger drops in systolic blood pressure (> 30 mmhg) benefit from the planned intervention. They will be analyzed as a subgroup. 26

36 2.10 Ethics Ethics approval has been granted from the University of Calgary Research Ethics Board. The research coordinator will determine patient eligibility, obtain consent from participants for participation in the study, and access the sealed envelope containing allocation details. The coordinator will obtain the HD run sheets and submit them to an independent data entry clerk who is blinded to allocation. All data is kept strictly confidential. The principal investigator (JM) and study personnel (KL) are responsible for coordination of the study. The trial is registered with Clinicaltrials.gov (NCT ) Discussion Randomized studies assessing the role of BVM and BVM-guided UF biofeedback in the prevention of IDH are limited and of low quality. One the largest studies assessing BVM without biofeedback by Reddan et al., randomized 443 participants to BVM or conventional HD over a six-month period 62. Participants in the group with BVM had UF rates adjusted at the discretion of the bedside dialysis nurse with a suggested UF algorithm. Due to the need for manual adjustment of the UF rate, it is unclear if the nurses promptly responded to changes in RBV using the suggested algorithm. In fact, the participants randomized to BVM had increased hospitalization and mortality, which may have reflected a lack of adherence to the UF algorithm, or alternatively could have been related to more cautious fluid removal due to the feedback provided by the BVM device. 27

37 However, when BVM is combined with biofeedback there may be a beneficial reduction in IDH and blood pressure 47,48,64,88. In a 16-week randomized crossover study by Dasselaar et al, 28 hypertensive patients were randomized to either BVM-guided sodium and UF biofeedback or standard HD for blood pressure and volume control 88. Despite a significant reduction in both blood pressure (22.5/8.3mm Hg) and extracellular water to body volume ratios in the BVM group when compared to the standard HD group, the overall weight did not change. Although the values were not reported, this study also found a statistically significant reduction in dialysis hypotension. A meta-analysis of six studies using BVM guided biofeedback reported a 39% overall reduction in the number of dialysis sessions complicated by IDH 65. Of the six studies assessing IDH, three did not have IDH as the primary outcome. Five studies employed an intervention that included both dialysate sodium and UF biofeedback 48,63, The use of the combination of dialysate sodium and UF biofeedback over UF biofeedback alone raises concerns regarding the potential for sodium loading and resultant interdialytic weight gain through increased plasma osmolality and thirst, perpetuating the vicious cycle of IDH. In addition, the duration of the intervention period in these studies ranged from two weeks to six months, with the majority using a short two to four-week intervention, putting the long-term sustainability of the intervention into question. The majority of the studies did not assess for changes in body fluid composition, nor did they address patient quality of life. In addition, the available studies are generally of low quality due to unclear 28

38 randomization, unclear exclusion criteria, and the inclusion of low numbers of IDH-prone participants 65. There was only one randomized crossover study (26 HD participants) that examined the impact of BVM-guided UF biofeedback on patient symptoms during dialysis. The primary outcome was an 8% absolute risk reduction in dialysis symptoms. There was also a statistically significant reduction in hypotension (the definition of which was unclear). 64 This industry-funded trial had several important weaknesses including its short duration (six-weeks), lack of a run-in period or washout period, and the use of a non-standardized definition of IDH making the interpretation of the results difficult. Currently there is a prospective, multi-center, triple-arm, parallel group, crossover, randomized controlled trial in progress that is comparing BVM-guided UF biofeedback and blood thermal monitoring temperature biofeedback, to BVM-guided UF and sodium biofeedback, to standard HD in effectively reducing excess fluid in fluid overloaded HD patients 89. The primary outcome in this study is the proportion of HD sessions that are complicated with intra- and post-dialytic symptoms related to ultrafiltration, irrespective of blood pressure; the secondary outcome includes both symptomatic and asymptomatic IDH, using a non-standard definition of >40mm Hg drop in systolic blood pressure. In contrast, our study focuses enrolling symptomatic IDH patients, the use of the standard 29

39 IDH definition of 20mm Hg, and studying only BVM guided UF biofeedback without another accompanying biofeedback technology. We have described our methods for recruitment, randomization, allocation concealment, dialysis intervention, outcome assessment, and data collection methods in detail. The study protocol was developed according to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) ; and will be conducted, with results reported following the Consolidated Standards of Reporting Trials (CONSORT) statement

40 Figure 2 1: Anticipated Participant Flow Diagram 31

41 Figure 2 2: Depiction of blood volume monitoring (BVM) profile and the corresponding ultrafiltration profile 32

42 Randomization Crossover and Washout Table 2 1: Data Collection Schedule Run-In Period Week Week 1 4 Period 1 Period 2 Week Week Every Session Week 8 Week 22 HD Run-Sheet X Kt/V X Electrical Bioimpedance X X X X X X BNP X X X X X X hs-cardiac Troponin X X X X X X Interdialytic Symptom Survey X X X X X X Medication Review X X X X X X HD= Hemodialysis; BNP= brain natriuretic peptide; hs-cardiac troponin=high sensitivity cardiac troponin 33

43 Chapter 3: BLOOD VOLUME MONITORING GUIDED ULTRAFILTRATION BIOFEEDBACK ON THE REDUCTION OF INTRADIALYTIC HYPOTENSIVE EPISODES IN HEMODIALYSIS: RESULTS OF A RANDOMIZED CROSSOVER TRIAL (MANUSCRIPT #2) 34

44 3.1 Background Patients treated with hemodialysis (HD) typically remove one to five liters of fluid per session [ultrafiltration, (UF)] to maintain euvolemia. Rapid UF results in intradialytic hypotension (IDH) in as many as 25-50% of patients treated with HD 15. Despite advances in various dialysis technologies, the risk of IDH is still a common and serious complication of dialysis therapy 71 associated with an increase in morbidity and mortality 16, Some have suggested that the development of symptoms and IDH correlate with patientspecific thresholds of relative blood volume (RBV) 53,60. The RBV is measured by tracking changes in total protein or hemoglobin concentration at the arterial port during HD, using optical or ultrasound techniques. The RBV decreases with UF and higher UF rates resulting in steeper declines in the blood volume monitoring (BVM) curve This resulted in the development of BVM-guided biofeedback technology, whereby the UF rate (or in some systems dialysate sodium) is automatically adjusted to minimize these outcomes 54,57,58. Although BVM-guided biofeedback may be effective in reducing IDH, the existing literature has important limitations and there are important gaps in knowledge 47,48,64,65,66. The majority of the available studies used a form of BVM-guided biofeedback that adjusts both dialysate sodium and UF 47,48,63, This has the potential to lead to positive 35

45 sodium balance and volume overload. As a result, there is interest in the use of BVMguided UF biofeedback without adjusting the dialysate sodium. With this approach, the dialysis machine automatically reduces the rate of UF prior to reaching the patient s critical blood volume to prevent the development of IDH and its sequelae 47,48,50,51,62,63. We conducted a randomized controlled trial to study whether biofeedback technology incorporating BVM-guided UF adjustments, resulted in a reduction in the rate of symptomatic IDH episodes when compared to best clinical practice alone. 3.2 Methods Overview We performed a randomized, single blind, crossover trial in maintenance HD patients. We accrued patients from five centers (two hospital based HD units the Foothills Medical Centre and the Peter Lougheed Centre; and three community HD units Carewest Dr. Vernon Fanning Community HD, Sheldon M. Chumir Community HD, and Sunridge Community HD) in Calgary, Alberta, Canada between June 2014 and May The study protocol was approved by the University of Calgary Conjoint Health Research Ethics Board and was conducted in accordance with the Helsinki Declaration (Ethics ID: REB , clinicaltrials.gov: NCT ). 36

46 3.2.2 Study Population All patients treated with HD for more than three months were screened for eligibility. Patients older than 18 years of age, who were medically stable, undergoing HD three to four times per week for a minimum of three hours per session, and who had 30% of their HD sessions in the preceding eight-weeks complicated by symptomatic IDH were eligible. Patients with serum hemoglobin <80g/L, serum sodium of 133mmol/L, those who had an active malignancy, a history of blood transfusion or hospitalization in the preceding four-weeks, had ongoing urine output estimated 250ml (one cup) per day or routinely used diuretics for volume management, had a planned change in the renal replacement modality during the study period, or were unable to provide written informed consent were excluded Study Protocol The original design of the study has been previously described 1. Eligible participants underwent a four-week run-in period during which a standardized clinical assessment of dry weight, anti-hypertensive medications, and dialysis prescription was performed. Participants who met study eligibility criteria at the end of the run-in period were randomized to best clinical practice (control) or to best clinical practice plus BVMguided UF biofeedback (biofeedback intervention) for an eight-week period. This was followed by a two-week washout period, prior to crossing over to the other arm for a second, eight-week study period. Randomization of the allocation sequence was done 37

47 using computer-generated random numbers under the supervision of a statistician in the Department of Medicine, University of Calgary. The allocation sequence was inserted into sequentially numbered, opaque, sealed envelopes Dialysis Prescription All study participants were dialyzed using the Fresenius 5008 HD machine (Fresenius Medical Care, Bad Homburg, Germany) with high flux dialyzers. With the exception of the intervention period, a standardized hemodialysis prescription was used with a dialysate sodium of 138 mmol/l, dialysate calcium of 1.25 mmol/l, dialysate temperature of 36 o C, and constant UF rate (BVM-guided UF biofeedback disabled). Those receiving the biofeedback intervention in the randomization period had the same prescription, but had their ultrafiltration rate automatically adjusted during each dialysis session based on the changes in RBV (BVM-guided UF biofeedback enabled). BVM technique validation 49,52 and description of the BVM-guided UF biofeedback mechanism has been previously published 49,52,76. A single investigator (KL) reviewed and adjusted the patient-specific critical RBV and clinical dry weight weekly. The most recent episode of symptomatic IDH was identified as per study definition and the critical blood volume was set equal to the blood volume recorded immediately prior to that episode. 38

48 3.2.5 Study Endpoints The primary outcome was the rate of symptomatic IDH, calculated by dividing the number of symptomatic IDH episodes by the duration of each dialysis session in hours (hrs). Pre-specified subgroup analyses were conducted using a 30 mm Hg drop in systolic blood pressure with associated symptoms as the definition of symptomatic IDH, and in patients with starting systolic blood pressures of <100 mm Hg. The secondary outcomes of interest were the proportion of HD sessions affected by symptomatic IDH; the rate and proportion of asymptomatic IDH; rate and proportion of dialysis symptoms; dialysis adequacy as measured by single session Kt/V; total body water (TBW), extra cellular water (ECW), intracellular water (ICW), the ECW:ICW ratio, and ECW:TBW ratio as determined by electrical bioimpedance; changes in brain naturetic peptide (BNP); changes in high sensitivity-cardiac troponin (hs-cardiac troponin); changes in antihypertensive medications; and the time to recover from the last HD session (the interdialytic symptom survey) Data Collection Symptomatic IDH was defined as a drop in systolic blood pressure of 20mm Hg from baseline with associated symptoms 19,20. Symptoms included sudden-onset headache, dizziness, loss of consciousness, thirst, dyspnea, angina, muscle cramps, or vomiting 20,21. Dialysis adequacy measured by single-pool Kt/V, blood pressures, and pre- and post- 39

49 dialysis weights were recorded (usual care) at the end of each dialysis session on the session sheet. Electrical bioimpedance, BNP, hs-cardiac troponin, and interdialytic symptoms surveys were collected during mid-week HD session in weeks 1, 4, 8, 12, 14, 18 and 22. The interdialytic survey is provided at the beginning of the dialysis session, and asked patients to record the time it took them to recover to baseline following the last dialysis session Sample Size calculation Previously published data estimated that 23% of our HD population were affected by IDH 71. We conservatively assumed that only one IDH episode would occur per session. Using a two-by-two crossover design, we estimated that a sample size of 20 participants would provide a power of 90% to demonstrate a 30% reduction in the rate of IDH using the biofeedback intervention compared to the control, with a two-sided alpha of We aimed to enroll 32 participants to guard against dropout, which can be as high as 30% Statistical Analysis Baseline characteristics are presented as means and 95% confidence intervals (CI). The primary outcome was analyzed using a mixed-effects regression model with either biofeedback intervention or control HD as the main exposure, the participants as random effects, the baseline IDH rate as a covariate, and the rate of symptomatic IDH (events per 40

50 hour) as the outcome. The presence of carry-over effect was assessed by interaction terms combining randomization order or study week with the exposure of interest prior to direct comparison of the treatments (biofeedback or control HD). Secondary outcomes were compared with either a Students t-test or chi-square test, as appropriate. Percent hydration was calculated as (Overhydration)/ECW * 100 = (ECW- (0.63*ICW))/ECW * ,91,92. All statistical analysis was performed with Stata Statistical Software, Version 11 (College Station, Texas, USA). 3.3 Results Cohort Formation A total of 420 patients were screened for trial eligibility during the study period, of which 385 patients were excluded (Figure 3 1). Thirty-five patients meeting all inclusion criteria were entered into the four-week run-in period. These patients had an average of 38% HD sessions affected by IDH in the eight-weeks prior to entering the study. By the end of the run-in period, three patients were excluded because less than 30% of their dialysis sessions were complicated by symptomatic IDH. A total of 32 patients were randomized, 16 to each randomization order. In the first randomization order (biofeedback followed by control period), two patients were removed from the study, one during the intervention period due to hospitalization from 41

51 gastrointestinal bleed and the other during the washout period due to voluntary withdrawal from dialysis. In the second randomization order (control followed by biofeedback period), four patients were removed from the study: one each for sepsis requiring hospitalization, changed dialysis modality, renal recovery, and voluntary withdrawal from dialysis. A total of 32 patients were included in the intent-to-treat analysis Cohort Characteristics Baseline characteristics are shown in Table 3 1. Our cohort had an average age of 67 years (standard deviation [SD] 13 years); 13 (41%) of these patients were above the age of 70; 17 were female (53%). At the beginning of the study, patients had been on dialysis for an average of 3.65 years (SD 3.50 years), 43% had been treated with dialysis for less than two years, and 25% of the patients treated with dialysis for more than five years. The majority (65%) of patients were overweight (BMI>25 kg/m 2 ) and more than 30% had a BMI above 30 kg/m 2. The cause of end stage renal disease (ESRD) was predominately diabetes and hypertension. A total of seven patients (22%) had systolic heart failure, 12 (38%) had coronary artery disease, nine (28%) had peripheral vascular disease, and 24 (75%) were diabetic [15 (63%) were treated with insulin]. Lower limb amputations were present in two (9%) patients. 42

52 All patients were treated with standard, thrice-weekly dialysis using high-flux dialyzers. Body temperature monitoring (BTM) was used in the majority (88%) of patients with the remainder using cold dialysate ( C) to prevent IDH. Sodium profiling and UF profiling were used in five (16%) and six (19%) patients, respectively. Eleven (34%) were treated with a high dialysate sodium bath of 140mmol/L or higher; and five (16%) were treated with a dialysate calcium of 1.50 mmol/l or higher Run-in Period During the four-week run-in period, the dialysis prescription was standardized and dry weights were optimized (Table 3 2). The dry weights and the UFR did not change during the run-in period. Interdialytic weight gain (IDWG) increased from 1.7 kg to 2.6 kg (p=0.01) and the number of antihypertensive medications was constant throughout the run-in period (1.6/patient) Primary Outcome: Symptomatic IDH The rate of symptomatic IDH by study period and randomization order is shown in Table 3 3 and Figure 3 2. There were no interactions between the primary outcome and the randomization order. 43

53 Analysis by Treatment Received The primary outcome by treatment received is shown in Table 3 4 and Figure 3 3. Although there was a significant reduction in the rate of symptomatic IDH from run-in to the eighth week in the biofeedback period (p=0.01), there were no differences between the two treatments at the eighth week (p=0.41). Furthermore, there were no differences in the degree of change from either the run-in to the eighth week (p=0.48) or from the first week to the eighth week between the two treatments (p=0.14). When all eight weeks of each treatment period was combined together, the rate of symptomatic IDH did not differ between biofeedback and control period (p=0.29). There was a significant decline in the rate of symptomatic IDH from the run-in period to the control period as a whole (50.8% decline, p=0.01) which was not seen for the biofeedback period. Sensitivity Analysis In the pre-specified sensitivity analysis using an IDH definition of 30 mm Hg decline in the presence of symptoms, our results were consistent with the main analysis. Unfortunately, there were not enough events for the second, pre-specified sensitivity analysis. The proportion of patients starting HD with a systolic blood pressure of less than 100 mm Hg was too low for a meaningful assessment. Of the patients who began the HD session with a systolic blood pressure of less than 100 mmhg the mean event rate was between zero to 0.01 throughout the study period. 44

54 3.3.5 Secondary Outcomes The results of the analysis for rate of asymptomatic IDH (p=0.64) and symptoms alone (p=0.37, Table 3 4 and Figure 3 4) as well as for proportion of HD sessions with symptomatic IDH (p=0.52), asymptomatic IDH (p=0.67), and symptoms alone (p=0.96) remained consistent with the primary analysis (Table 3 5 and Figure 3 5). The remainder of the secondary outcomes are presented in Table 3 2 and Table 3 6. There were no differences in the prespecified secondary outcome over the study period or between the two treatments (p>0.10). Of note, there was a 0.7 kg increase in IDWG (p=0.01) during the run-in period, but there were no differences in IDWG from the first to eighth week in both treatments (biofeedback p=0.54, control p=0.90). As well, our cohort demonstrated evidence of excess ECW on the basis of an elevated ECW:ICW, ECW:TBW, and percent hydration throughout the study. 3.4 Discussion Summary of Findings We performed a randomized, single blind, crossover trial in maintenance HD patients to determine the effects of BVM-guided UF biofeedback on symptomatic IDH, as compared to standard best clinical practice alone. We found that the use of BVM-guided UF 45

55 biofeedback did not reduce the rate of symptomatic IDH, asymptomatic IDH, or symptoms alone. In addition, we were unable to appreciate any significant differences in biomarkers of cardiac stress (BNP and hs-cardiac troponins), body water composition (ECW:ICW ratios, ECW:TBW ratios, percent hydration), interdialytic weight gains, and patient post-dialysis recovery time Discussion of Prior Studies Our study findings differ from prior studies that show a reduction in IDH with the use of BVM biofeedback 47,48, However, our study has important differences that should be highlighted. First and most importantly, our study only adjusted the UF rate, whereas the majority of biofeedback studies adjusted both dialysate sodium and UF rate 48,63, The adjustment or increase of dialysate sodium to prevent hemodynamic instability allows for an relatively rapid increase in intravascular refilling from the extracellular space, thus preventing volume related blood pressure drops or intradialytic symptoms 73, The biofeedback system in our study only adjusted the UF rate, thus potentially reducing its effectiveness and may explain the difference in observed results. Second, most of the studies that employed the dialysate sodium and UF rate biofeedback interventions had event reductions over a shorter period of time (two to four weeks) as compared with our study. The consequences of elevated dialysate sodium start to appear after approximately two weeks and it is unclear when they would peak If the duration of these studies 46

56 were extended, we may see a gradual rise in event rates nullifying the early reduction in the intervention period as a result of sodium loading and hypervolemia. Lastly, the study population and measured outcomes were different than in our study. Three of the studies included patients not prone to some form of symptomatic IDH 48,63,67, three had unclear exclusion criteria 64,68,69, and the outcomes in five studies 48,63,64,67,68 were either not restricted to one of the few recognized definitions of symptomatic IDH 18,19 or were combined with asymptomatic IDH. These differences make interpretation of the conflicting results difficult. Only one other study has assessed the impact of a BVM-guided UF biofeedback alone, without the adjustment of dialysate sodium and found a significant reduction of 8% in each of dialysis symptoms (with or without hypotension) and hypotension (a 20% relative reduction) 64. Assuming a four-hour dialysis session in their study, their reported rates of symptoms (with or without hypotension) were slightly higher when compared to our study (~ /hr versus /hr symptomatic IDH in ours). Their rate of hypotension alone was quite a bit lower than ours (~ /hr versus /hr asymptomatic IDH in ours). These differences may be due to a few reasons. The first is the differences in the outcome definition. Our study used a standardized and specific definition of IDH, requiring an abrupt drop in blood pressure in the presence of symptoms. Whereas in the Gabrielli study, the primary outcome was intradialytic morbid 47

57 event, which was defined as any symptom during dialysis with or without hypotension. This allowed any symptom to be recorded as an event. Similarly, their secondary outcome of hypotension did not specify the degree or abruptness of the blood pressure decline. Secondly, the absence of a run-in period in the previous study raises concerns about proper dry weight optimization. If the dry weight or UF goals are inappropriately set, it could result in hemodynamic instability and symptoms. This is evident in our study where 9% of patients were excluded during the run-in period due to a lack of symptomatic IDH after the dry weight was appropriately adjusted (also reflected in the rise in IDWG in the run-in period). As a result, our remaining cohort may have IDH due to other non-uf related factors such as autonomic dysfunction, decreased vascular compliance and reserve, and poor cardiac function, thus potentially minimizing the effect of the intervention Strengths and Limitations Our study has several strengths. Although we had small study numbers, we were adequately powered to detect a clinically significant difference. We also had a dedicated run-in period and a long enough washout period to reduce the potential of a carryover effect from one period to another. However, given the nature of the study intervention, only the patients could be blinded. All attempts were made to ensure nursing protocols for the management and documentation of IDH were consistent. Although the rounding 48

58 physicians were not blinded to the study intervention, they were not required to manage intervention alarms or adjust dialysis prescriptions in an attempt to decrease the possibility of unintentional un-blinding. In addition, the BVM technology can be affected by changes other than UF. For example the patient s blood volume changes with shifts in the patient s position or other activities during HD, such as infusion of any IV medications or food intake 22,100. Finally, the critical RBV and dry weights were only reassessed on a weekly basis. However, our experience suggests that there was only minute variation between these values and weights throughout the study period Implications for Clinical Care The role of BVM-guided UF biofeedback in the prevention and management of IDH remains unclear. The technology has a strong theoretical rationale, but we were unable to demonstrate an impact on important clinical parameters. There are unresolved issues about how best to implement this technology, such as the method of selection of critical RBV or the selection of patients most likely to benefit, that need to be better understood in order to definitively determine its clinical usefulness, if any Conclusions In conclusion, the use of BVM-guided UF did not result in a reduction in the rate of symptomatic IDH, asymptomatic IDH, or symptoms alone. We also did not appreciate an 49

59 improvement in volume control or biomarkers of cardiac stress. Further studies are required to clarify the role of this and other related BVM-guided biofeedback ultrafiltration systems in clinical practice. 50

60 Table 3 1: Baseline Patient Characteristics Patient characteristics (N=32) Age (years), mean (SD) 67.2 (13.1) Female n (%) 17 (53%) Dialysis Vintage (years), mean (SD) 3.65 (3.50) Weight (kg), mean (SD) 78.4 (21.0) BMI (kg/m 2 ), mean (SD) 29.5 (7.4) Race Caucasian, n (%) 16 (50%) Asian, n (%) 10 (31%) Other, n (%) 6 (19%) Etiology of ESRD Diabetes and Hypertension, n (%) 23 (72%) Glomerulonephritis, n (%) 3 (9%) Obstructive, n (%) 3 (9%) Other, n (%) 3 (9%) Comorbidities Coronary Artery Disease, n (%) 12 (38%) Systolic Heart Failure, n (%) 7 (22%) Atrial Fibrillation, n (%) 4 (13%) Diabetes, n (%) 24 (75%) Diabetes with Insulin use, n (%) 15 (63%) Peripheral Vascular Disease, n (%) 9 (28%) Stroke, n (%) 4 (13%) Hypertension, n (%) 28 (88%) Lower Limb Amputation, n (%) 3 (9%) Vascular Access CVC, n (%) 20 (62.5) AVF, n (%) 12 (37.5) Dialysis Prescription HD Duration of 4 hours, n (%) 28 (88%) Dialysate Sodium, mmol/l, mean (SD) (1.6) 140 mmol/l, n (%) 11 (34) Dialysate Calcium 1.5 mmol/l, n (%) 5 (16%) Sodium Profile, n (%) 5 (16%) UF Profile, n (%) 6 (19%) BTM Enabled, n (%) 28 (88%) 51

61 Table 3 2: Average Weights, Interdialytic Weight Gain (IDWG), Percentage Reaching Dry Weight, and Ultrafiltration Rate Run-In Period Biofeedback Period Control Period Start End Start End Start End Weight (kg) ( ) ( ) ( ) ( ) ( ) ( ) IDWG (kg) ( ) ( ) ( ) ( ) ( ) ( ) Ultrafiltration Rate (ml/kg/hr) 7.43 ( ) 6.02 ( ) 7.29 ( ) 5.87 ( ) 7.01 ( ) 4.80 ( ) Percent Reaching Dry Weight (%) Values are expressed as mean (95% CI) unless otherwise indicated. 52

62 Table 3 3: Rates of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone by Period Period Order Symptomatic IDH, events/hour Mean rate (95% CI) Asymptomatic IDH, events/hour Mean rate (95% CI) Symptoms Alone, events/hour Mean rate (95% CI) Run-in 0.15 ( ) 0.31 ( ) 0.07 ( ) Period 1 A 0.13 ( ) 0.38 ( ) 0.05 ( ) B 0.07 ( ) 0.29 ( ) 0.07 ( ) Washout A 0.08 ( ) 0.31 ( ) 0.03 ( ) B 0.05 ( ) 0.30 ( ) 0.07 ( ) Period 2 A 0.08 ( ) 0.31 ( ) 0.05 ( ) B 0.06 ( ) 0.26 ( ) 0.04 ( ) A= Biofeedback in period 1 followed by control in period 2. B= Control in period 1 followed by biofeedback in period 2. Values are expressed as mean (95% CI) unless otherwise indicated. 53

63 Table 3 4: Rates of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone by Intervention Symptomatic IDH, events/hour Mean rate (95% CI) Asymptomatic IDH, events/hour Mean rate (95% CI) Symptoms Alone, events/hour Mean rate (95% CI) Run-In Period 0.15 ( ) 0.31 ( ) 0.07 ( ) Biofeedback Period 0.10 ( ) 0.33 ( ) 0.05 ( ) Start of Period 0.15 ( ) 0.36 ( ) End of Period 0.07 ( ) 0.26 ( ) Control Period 0.07 ( ) 0.30 ( ) 0.06 ( ) Start of Period 0.09 ( ) 0.32 ( ) End of Period 0.11 ( ) 0.33 ( ) Values are expressed as mean (95% CI) unless otherwise indicated 54

64 Table 3 5: Proportion of Hemodialysis Sessions Affected by Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone by Intervention Symptomatic IDH Mean proportion (95% CI) Asymptomatic IDH Mean proportion (95% CI) Symptoms Alone Mean proportion (95% CI) Run-In Period 0.38 ( ) 0.60 ( ) 0.18 ( ) Biofeedback Period 0.21 ( ) 0.54 ( ) 0.23 ( ) Control Period 0.18 ( ) 0.51 ( ) 0.23 ( ) Values are expressed as mean (95% CI) unless otherwise indicated. 55

65 Table 3 6: Average Fluid Content, Fluid Ratios, Brain Naturetic Peptide, Troponin, Post Dialysis Recovery Time, and Change in Standing Heart Rate Start of Run-In p- p- End of Run-In Period Biofeedback Period Control Period Period value value ECW (L) ( ) ( ) ( ) ( ) ICW (L) ( ) ( ) ( ) ( ) TBW (L) ( ) ( ) ( ) ( ) ECW:ICW Ratio 0.86 ( ) 0.86 ( ) ( ) 0.86 ( ) 0.39 ECW:TBW Ratio 0.46 ( ) 0.46 ( ) ( ) 0.46 ( ) 0.49 Hydration (%) ( ) ( ) ( ) ( ) 0.47 BNP (ng/l) 5916 ( ) 6441 ( ) ( ) 6333 ( ) 0.66 hs-cardiac troponin (ng/l) 84 (63-105) 74 (57-91) (55-91) 70 (46-93) 0.85 Kt/v (95% CI) 1.25 ( ) 1.30 ( ) ( ) 1.27 ( ) 0.56 Recovery Time, hours (95% CI) 8.27 ( ) 9.31 ( ) ( ) 8.65 ( ) 0.91 Change in Standing Heart Rate (Post-Pre) ( ) ( ) ( ) ( ) 0.84 Values are expressed as mean (95% CI) unless otherwise indicated. 56

66 Figure 3 1: Participant Flow Diagram 57

67 Run-In Period 1 Washout Period 2 A B Figure 3 2: Rate of Symptomatic IDH Events Per Hour, By Randomization Order and Study Period In randomization order (A), biofeedback in period 1 followed by control in period 2. In randomization order (B), control in period 1 followed by biofeedback in period 2 58

68 Figure 3 3: Rate of Symptomatic IDH Events Per Hour, By Intervention and Study Week 59

69 Events per hour Run-In Control Biofeedback Intervention Symptomatic IDH rate Asymptomatic IDH rate Symptoms Alone rate Figure 3 4: Rate of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone Per Hour, in the Run-In and Study Periods 60

70 Proportion of HD Sessions Run-In Control Biofeedback Intervention Symptomatic IDH Proportion Asymptomatic IDH Proportion Symptoms Alone Proportion Figure 3 5: Proportion of Symptomatic IDH, Asymptomatic IDH, and Symptoms Alone Per Hour, in the Run-In and Study Periods 61

71 Chapter 4: SUMMARY OF FINDINGS, CLINICAL IMPLICATIONS AND CONCLUSION 62

72 4.1 Summary of Study and Findings A single previous study described improvements in symptoms and hypotension with the use of BVM-guided UF adjustments alone, but the use of a non-standardized outcome definition and dialysis prescriptions raised concerns for the interpretation of the results 64. We performed a randomized controlled clinical trial that addressed these shortcomings to test whether biofeedback technology incorporating BVM-guided UF biofeedback resulted in a reduction in the rate of symptomatic IDH episodes in an IDH prone population, when compared to best clinical practices alone. BVM-guided UF biofeedback does not appear to result in a reduction in the rate of symptomatic IDH, asymptomatic IDH, or symptoms alone. In addition, biomarkers of cardiac stress, volume control, and patient reported symptoms were not different in the intervention group. 4.2 Clinical Implications Impact of Biofeedback Technology Learning Curve With biofeedback, the rate of symptomatic IDH did not begin to trend downward until week six during the intervention period, whereas the rate decreased in the control group in week two and then plateaued. This might suggest that there is an operator learning curve with the biofeedback technology. As new technology is introduced into the HD unit, both physicians and nurses need to be familiar with the technology before its full potential can be realized. While the determination of the critical RBV has been standardized in our study, initial unfamiliarity with the technology may have resulted in this value being set inappropriately low. This may have led to a higher initial UFR and 63

73 potential for more IDH events during the first few weeks of the study period before a more appropriate critical RBV was set. From the nursing aspect, there was a machine feature whereby the dialysis machine automatically adjusts the critical RBV. While the study instructions asked for this to be disabled, in certain cases it was accidentally left the machine on and reduced the critical RBV, which may have increased the UFR and potentially IDH events. Similarly, there have been situations where the nursing staff accidentally forgot to initiate the biofeedback technology until late into the HD session. Depending on the patient and biofeedback parameters, this can either result in an abrupt increase in the UFR and thus IDH events at the end of the HD session, or alternatively, it can trigger alarm warnings. It took a few HD sessions for any given bedside nurse to be familiar with the setup and proper programming of the biofeedback technology. As the nursing staff rotated between HD sessions it meant that this familiarization period was extended to several weeks. This may explain the observation that, while the randomization order was not important, it appeared that patients randomized to start with biofeedback technology first trended towards higher rates of symptomatic IDH than those that had the biofeedback technology second. This may also be an explanation for the relatively larger confidence intervals around our initial point estimates that seem to become narrower during the second period of biofeedback intervention Lessons from the Implementation of Biofeedback Technology Although in theory, the use of biofeedback technology appears intuitive, there were a few lessons learned from its implementation in the study. The first consideration was the use 64

74 of the concept of critical RBV. Unfortunately the RBV does not always correlate with patient symptoms 53 55,58,59 and can vary from factors aside from UF, such as infusion of IV medication or position changes 22,100. Furthermore, the interpretation of RBV values or the BVM curves have not been standardized. Thus, the use of critical RBV may be a flawed approach. The second is the process of determining the critical RBV. In our study, the critical RBV was the RBV value immediately prior to the patient s last symptomatic IDH episode. We developed our own method for selecting the critical RBV as there was no established protocol or method for selecting a patient specific critical blood volume. Our critical RBV selection algorithm was designed in the hopes of reducing symptomatic IDH events, therefore it favored an increase in the critical RBV value. If the critical RBV is set too high, the machine algorithm may prevent the dry weight being reached due to automatic reductions in the UF. Over time, this can lead to volume overload and its harmful consequences 6. It is unclear if a modified protocol allowing for a gradual critical RBV lowering in or if an universal value applied to all patients (similar to the Gabrielli study 64 ) would have been more effective in preventing symptomatic IDH. Furthermore, we do not know the optimal frequency to adjust the critical RBV or if different values should be used for varying interdialytic periods. Third, the optimal length of therapy or time-dependent effect with biofeedback technology is unclear. In our study, the rates of IDH in those treated with biofeedback intervention appear to decrease and plateau after week five. Although less likely, it is unclear if additional study weeks may have shown a further decline in event rates. 65

75 Until further research is available, caution should be applied before widespread implementation of this technology. Along the same lines, further research will be required to determine if the use of a critical RBV is clinically the most sensitive and specific method for predicting episodes of IDH. Currently other uses of the noninvasive BVM technology such as slopes of the RBV 59 or the use of absolute blood volume 54,101 have been proposed as tools to guide fluid removal IDH was Common despite Low UFR and IDWG Our cohort had relatively high IDH rates despite having relatively low UFR and IDWGs. Typically, higher UFR 18,102 (i.e. >10-13 ml/kg/hr) have been associated with an increased risk of negative outcomes and thought to be a key factor in the development of IDH 17, However, in our study population, the rate of IDH was high despite relatively low UFR s and low IDWG s. The average UFR in our cohort was between ml/kg/hr and the IDWG was approximately 2 kg across both run-in and study periods. One potential explanation for this observation is that high UFRs are not the only factor in the development of IDH, especially in our patient population. There may be other risk factors for IDH, such as autonomic dysfunction (diabetes), decreased vascular compliance and reserve (lower limb amputations), and poor cardiac function (systolic heart failure) that we have not assessed. In fact, given the multifactorial nature of IDH, attributing IDH to high UFRs alone may be an oversimplification of the pathophysiology

76 4.2.4 True Dry Weight was Difficult to Determine It is possible that our patients had an inappropriately high estimated dry weight in relation to their true dry weight. The bioimpedance and biomarker results indicate that the patients were in fact volume overloaded 6,80,91,92, although comparable to other hemodialysis and peritoneal dialysis populations 106,107. Clinically, the study patients were thought to be only mildly volume overloaded when they came in for their HD sessions since their clinical exam was not consistent with excess body water, and they typically had very low IDWG and low UFR. The discrepancy between the bioimpedance values and the clinically derived dry weight highlights the difficulty in determining true dry weight. Our cohort s higher BMI reflects a degree of obesity which may have complicated the volume assessment. Aside from the mechanical aspect of differentiating between adipose and edema, obese patients are more likely to have venous insufficiency and resultant edema which may or may not be related to volume excess 108,109. Although an inappropriately high or elevated dry weight may have reduced the symptomatic IDH event rates (by not removing so much fluid), the same error in dry weight determination occurred in both groups, thus making it unlikely to have changed our result. Despite this, dry weight determination in future research should incorporate a standardized tool such as bioimpedance Use of Surrogate Outcomes In our study, we chose to use symptomatic IDH as the primary outcome. Although symptomatic IDH is both a clinically and patient relevant outcome, it is a surrogate 67

77 outcome. To date, various definitions of IDH has been associated with mortality 27 29, but has not been demonstrated to be causal in experimental studies. This is important as there are many therapies aimed at reducing or preventing symptomatic IDH 21,22,36, while it is unclear if these have an effect in reducing mortality. This is especially of concern given the multifactorial nature of IDH 22. Although unlikely, it is certainly possible that symptomatic IDH is not directly related to but is only a marker for a higher presence comorbidities and mortality. Given the lack of definite causal effect, more research needs to be done to establish this link. One such possibility is the use of long term follow up studies that explore the effect of biofeedback on mortality. 4.3 Conclusions Our study showed that the use of blood volume monitoring guided ultrafiltration did not result in a reduction in the rate of symptomatic IDH, asymptomatic IDH, or symptoms alone. Until further studies are performed, the role of BVM-guided ultrafiltration in clinical practice is unclear. 68

78 REFERENCES 1. Leung, K. C., Quinn, R. R., Ravani, P. & MacRae, J. M. Ultrafiltration biofeedback guided by blood volume monitoring to reduce intradialytic hypotensive episodes in hemodialysis: study protocol for a randomized controlled trial. Trials 15, 483 (2014). 2. U.S. Renal Data System. USRDS 2012 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States. (2012). 3. U.S. Renal Data System. USRDS 2011 Annual Data Report: Atlas of End-Stage Renal Disease in the United States. Natl. Institutes Heal. Natl. Inst. Diabetes Dig. Kidney Dis. Bethesda, MD (2011). 4. Canadian Institute for Health Information. Treatment of End-Stage Organ Failure in Canada CORR 2012 Annual Report (2012). 5. Mailloux, L. U. et al. Mortality in dialysis patients: analysis of the causes of death. Am J Kidney Dis. 18, (1991). 6. Wizemann, V. et al. The mortality risk of overhydration in haemodialysis patients. Nephrol Dial Transplant. 24, (2009). 7. Kalantar-Zadeh, K. et al. Fluid retention is associated with cardiovascular mortality in patients undergoing long-term hemodialysis. Circulation 119, (2009). 8. Velasco, N. et al. Optimal fluid control can normalize cardiovascular risk markers and limit left ventricular hypertrophy in thrice weekly dialysis patients. Hemodial Int. 16, (2012). 9. Chazot, C. et al. Importance of normohydration for the long-term survival of 69

79 haemodialysis patients. Nephrol Dial Transplant. 27, (2012). 10. Agarwal, R., Bouldin, J. M., Light, R. P. & Garg, A. Inferior vena cava diameter and left atrial diameter measure volume but not dry weight. Clin J Am Soc Nephrol. 6, (2011). 11. Tripepi, G. et al. Left atrial volume monitoring and cardiovascular risk in patients with end-stage renal disease: a prospective cohort study. J Am Soc Nephrol. 18, (2007). 12. Tripepi, G. et al. Left atrial volume in end-stage renal disease: a prospective cohort study. J Hypertens. 24, (2006). 13. Zoccali, C. et al. Cardiac natriuretic peptides are related to left ventricular mass and function and predict mortality in dialysis patients. J Am Soc Nephrol. 12, (2001). 14. Lee, S. W., Song, J. H., Kim, G. a, Lim, H. J. & Kim, M.-J. Plasma brain natriuretic peptide concentration on assessment of hydration status in hemodialysis patient. Am J Kidney Dis. 41, (2003). 15. Henrich, W. L. Hemodynamic instability during hemodialysis. Kidney Int. 30, (1986). 16. Burton, J. O., Jefferies, H. J., Selby, N. M. & McIntyre, C. W. Hemodialysisinduced cardiac injury: determinants and associated outcomes. Clin J Am Soc Nephrol. 4, (2009). 17. Flythe, J. E., Kimmel, S. E. & Brunelli, S. M. Rapid fluid removal during dialysis is associated with cardiovascular morbidity and mortality. Kidney Int. 79, (2011). 70

80 18. Flythe, J. E., Xue, H., Lynch, K. E., Curhan, G. C. & Brunelli, S. M. Association of mortality risk with various definitions of intradialytic hypotension. J. Am. Soc. Nephrol. 26, (2015). 19. Kooman, J. et al. EBPG guideline on haemodynamic instability. Nephrol. Dial. Transplant 22 Suppl 2, ii22-44 (2007). 20. National Kidney Foundation. K/DOQI clinical practice guidelines for cardiovascular disease in dialysis patients. Am J Kidney Dis. 45, S1-153 (2005). 21. Schreiber, M. J. Setting the stage. Am J Kidney Dis. 38, S1 S10 (2001). 22. Davenport, A. Can advances in hemodialysis machine technology prevent intradialytic hypotension? Semin Dial. 22, (2009). 23. Daugirdas, J. T. Pathophysiology of dialysis hypotension: an update. Am J Kidney Dis. 38, S11-7 (2001). 24. Daugirdas, J. T. Dialysis hypotension: a hemodynamic analysis. Kidney Int. 39, (1991). 25. Perazella, M. a. Review Articles: Approach to Patients with Intradialytic Hypotension: A Focus on Therapeutic Options. Semin. Dial. 12, (2002). 26. van der Sande, F. M. et al. Control of core temperature and blood pressure stability during hemodialysis. Clin. J. Am. Soc. Nephrol. 4, 93 8 (2009). 27. Caplin, B., Kumar, S. & Davenport, A. Patients perspective of haemodialysisassociated symptoms. Nephrol. Dial. Transplant. 26, (2011). 28. Zager, P. G. et al. U curve association of blood pressure and mortality in hemodialysis patients. Kidney Int. 54, (1998). 29. Shoji, T., Tsubakihara, Y., Fujii, M. & Imai, E. Hemodialysis-associated 71

81 hypotension as an independent risk factor for two-year mortality in hemodialysis patients. Kidney Int. 66, (2004). 30. Yoshimitsu, T. et al. Cerebral ischemia as a causative mechanism for rapid progression of brain atrophy in chronic hemodialysis patients. Clin Nephrol. 53, (2000). 31. Ishida, I. et al. Hemodialysis causes severe orthostatic reduction in cerebral blood flow velocity in diabetic patients. Am J Kidney Dis. 34, (1999). 32. Breidthardt, T. et al. Troponin T for the detection of dialysis-induced myocardial stunning in hemodialysis patients. Clin J Am Soc Nephrol. 7, (2012). 33. Burton, J. O., Jefferies, H. J., Selby, N. M. & McIntyre, C. W. Hemodialysis- Induced Repetitive Myocardial Injury Results in Global and Segmental Reduction in Systolic Cardiac Function. Clin. J. Am. Soc. Nephrol. 4, (2009). 34. Selby, N. M. & McIntyre, C. W. The acute cardiac effects of dialysis. Semin. Dial. 20, (2007). 35. Burton, J. O., Korsheed, S., Grundy, B. J. & McIntyre, C. W. Hemodialysis- Induced Left Ventricular Dysfunction Is Associated with an Increase in Ventricular Arrhythmias. Ren. Fail. 30, (2008). 36. Davenport, A. Using dialysis machine technology to reduce intradialytic hypotension. Hemodial. Int. 15 Suppl 1, S37-42 (2011). 37. Rouby, J. J. et al. Hemodynamic changes induced by regular hemodialysis and sequential ultrafiltration hemodialysis: A comparative study. Kidney Int. 17, (1980). 38. Chazot, C. & Jean, G. The advantages and challenges of increasing the duration 72

82 and frequency of maintenance dialysis sessions. Nat Clin Pr. Nephrol. 5, (2009). 39. Song, J. H. et al. Effect of sodium balance and the combination of ultrafiltration profile during sodium profiling hemodialysis on the maintenance of the quality of dialysis and sodium and fluid balances. J. Am. Soc. Nephrol. 16, (2005). 40. Zhou, Y. L. et al. Impact of sodium and ultrafiltration profiling on haemodialysisrelated hypotension. Nephrol Dial Transplant. 21, (2006). 41. Maynard, J. C., Cruz, C., Kleerekoper, M. & Levin, N. W. Blood pressure response to changes in serum ionized calcium during hemodialysis. Ann Intern Med. 104, (1986). 42. van der Sande, F. M., Cheriex, E. C., van Kuijk, W. H. & Leunissen, K. M. Effect of dialysate calcium concentrations on intradialytic blood pressure course in cardiac-compromised patients. Am. J. Kidney Dis. 32, (1998). 43. Selby, N. M. & McIntyre, C. W. A systematic review of the clinical effects of reducing dialysate fluid temperature. Nephrol Dial Transplant. 21, (2006). 44. van der Sande, F. M., Kooman, J. P., Konings, C. J. & Leunissen, K. M. Thermal effects and blood pressure response during postdilution hemodiafiltration and hemodialysis: the effect of amount of replacement fluid and dialysate temperature. J Am Soc Nephrol. 12, (2001). 45. Donauer, J., Schweiger, C., Rumberger, B., Krumme, B. & Böhler, J. Reduction of hypotensive side effects during online-haemodiafiltration and low temperature haemodialysis. Nephrol Dial Transplant. 18, (2003). 73

83 46. Prakash, S., Garg, A. X., Heidenheim, a P. & House, A. a. Midodrine appears to be safe and effective for dialysis-induced hypotension: a systematic review. Nephrol Dial Transplant. 19, (2004). 47. Ronco, C. et al. Impact of biofeedback-induced cardiovascular stability on hemodialysis tolerance and efficiency. Kidney Int. 58, (2000). 48. Santoro, A. et al. Blood volume controlled hemodialysis in hypotension-prone patients: a randomized, multicenter controlled trial. Kidney Int. 62, (2002). 49. Johner, C., Chamney, P. W., Schneditz, D. & Krämer, M. Evaluation of an ultrasonic blood volume monitor. Nephrol Dial Transplant. 13, (1998). 50. Mann, H., Stiller, S., Schallenberg, U. & Thömmes, A. Optimizing dialysis by variation of ultrafiltration rate and sodium concentration controlled by continuous measurement of circulating blood volume. Contrib Nephrol. 74, (1989). 51. Mancini, E. et al. Continuous on-line optical absorbance recording of blood volume changes during hemodialysis. Artif Organs. 17, (1993). 52. Dasselaar, J. J., Huisman, R. M., DE Jong, P. E. & Franssen, C. F. M. Relative blood volume measurements during hemodialysis: comparisons between three noninvasive devices. Hemodial Int. 11, (2007). 53. de Vries, J. P. et al. Non-invasive monitoring of blood volume during hemodialysis: its relation with post-dialytic dry weight. Kidney Int. 44, (1993). 54. Booth, J., Pinney, J. & Davenport, A. Do changes in relative blood volume monitoring correlate to hemodialysis-associated hypotension? Nephron Clin Pr. 74

84 117, c (2011). 55. Agarwal, R., Kelley, K. & Light, R. P. Diagnostic utility of blood volume monitoring in hemodialysis patients. Am J Kidney Dis. 51, (2008). 56. Sinha, A. D., Light, R. P. & Agarwal, R. Relative plasma volume monitoring during hemodialysis AIDS the assessment of dry weight. Hypertension 55, (2010). 57. Krepel, H. P., Nette, R. W., Akçahüseyin, E., Weimar, W. & Zietse, R. Variability of relative blood volume during haemodialysis. Nephrol Dial Transplant. 15, (2000). 58. Andrulli, S. et al. The role of blood volume reduction in the genesis of intradialytic hypotension. Am J Kidney Dis. 40, (2002). 59. Seibert, E. et al. Slope analysis of blood volume and calf bioimpedance monitoring in hemodialysis patients. Nephrol Dial Transplant. 27, (2012). 60. Steuer, R. R., Leypoldt, J. K., Cheung, A. K., Senekjian, H. O. & Conis, J. M. Reducing symptoms during hemodialysis by continuously monitoring the hematocrit. Am. J. Kidney Dis. 27, (1996). 61. Dasselaar, J. J. et al. Relative blood volume changes underestimate total blood volume changes during hemodialysis. Clin J Am Soc Nephrol. 2, (2007). 62. Reddan, D. N. et al. Intradialytic blood volume monitoring in ambulatory hemodialysis patients: a randomized trial. J Am Soc Nephrol. 16, (2005). 63. Nesrallah, G. E., Suri, R. S., Thiessen-Philbrook, H., Heidenheim, P. & Lindsay, R. M. Can extracellular fluid volume expansion in hemodialysis patients be safely reduced using the hemocontrol biofeedback algorithm? A randomized trial. ASAIO 75

85 J. 54, (2008). 64. Gabrielli, D. et al. Improved intradialytic stability during haemodialysis with blood volume-controlled ultrafiltration. J. Nephrol. 22, (2008). 65. Nesrallah, G. E. et al. Biofeedback dialysis for hypotension and hypervolemia: a systematic review and meta-analysis. Nephrol Dial Transplant. 28, (2013). 66. Gil, H.-W. et al. Efficacy of hemocontrol biofeedback system in intradialytic hypotension-prone hemodialysis patients. J. Korean Med. Sci. 29, (2014). 67. Bégin, V., Déziel, C. & Madore, F. Biofeedback regulation of ultrafiltration and dialysate conductivity for the prevention of hypotension during hemodialysis. ASAIO J. 48, (2002). 68. Déziel, C., Bouchard, J., Zellweger, M. & Madore, F. Impact of hemocontrol on hypertension, nursing interventions, and quality of life: a randomized, controlled trial. Clin J Am Soc Nephrol. 2, (2007). 69. Selby, N. M., Lambie, S. H., Camici, P. G., Baker, C. S. & McIntyre, C. W. Occurrence of regional left ventricular dysfunction in patients undergoing standard and biofeedback dialysis. Am J Kidney Dis. 47, (2006). 70. Murtagh, F. E. M., Addington-Hall, J. & Higginson, I. J. The Prevalence of Symptoms in End-Stage Renal Disease: A Systematic Review. Adv Chronic Kidney Dis. 14, (2007). 71. Tai, D. J., Conley, J., Ravani, P., Hemmelgarn, B. R. & MacRae, J. M. Hemodialysis prescription education decreases intradialytic hypotension. J Nephrol. 26, (2012). 72. Schulz, K. F., Altman, D. G. & Moher, D. CONSORT 2010 Statement: updated 76

86 guidelines for reporting parallel group randomised trials. Trials 11, 32 (2010). 73. Penne, E. L. & Sergeyeva, O. Sodium gradient: a tool to individualize dialysate sodium prescription in chronic hemodialysis patients? Blood Purif. 31, (2011). 74. Agarwal, R., Alborzi, P., Satyan, S. & Light, R. P. Dry-weight reduction in hypertensive hemodialysis patients (DRIP): a randomized, controlled trial. Hypertension 53, (2009). 75. Schneditz, D., Pogglitsch, H., Horina, J. & Binswanger, U. A blood protein monitor for the continuous measurement of blood volume changes during hemodialysis. Kidney Int. 38, (1990). 76. Chapdelaine, I., Déziel, C. & Madore, F. in Progress in Hemodialysis - From Emergent Biotechnology to Clinical Practice (ed. Carpi, A.) (InTech, 2011). at < Automated_blood_volume_regulation_during_hemodialysis.pdf> 77. Kuhlmann, U. et al. Accuracy and safety of online clearance monitoring based on conductivity variation. Nephrol Dial Transplant. 16, (2001). 78. Lukaski, H. C., Johnson, P. E., Bolonchuk, W. W. & Lykken, G. I. Assessment of fat-free mass using bioelectrical impedance measurements of the human body. Am J Clin Nutr. 41, (1985). 79. Chan, C., McIntyre, C., Smith, D., Spanel, P. & Davies, S. J. Combining nearsubject absolute and relative measures of longitudinal hydration in hemodialysis. Clin J Am Soc Nephrol. 4, (2009). 80. Chamney, P. W. et al. A whole-body model to distinguish excess fluid from the 77

87 hydration of major body tissues. Am J Clin Nutr. 85, 80 9 (2007). 81. Wabel, P., Chamney, P., Moissl, U. & Jirka, T. Importance of whole-body bioimpedance spectroscopy for the management of fluid balance. Blood Purif. 27, (2009). 82. Wabel, P. et al. Towards improved cardiovascular management: the necessity of combining blood pressure and fluid overload. Nephrol Dial Transplant. 23, (2008). 83. Lindsay, R. M., Heidenheim, P. a, Nesrallah, G., Garg, A. X. & Suri, R. Minutes to recovery after a hemodialysis session: a simple health-related quality of life question that is reliable, valid, and sensitive to change. Clin J Am Soc Nephrol. 1, (2006). 84. MacRae, J. M., Ahmed, S. B., Atkar, R. & Hemmelgarn, B. R. A randomized trial comparing buttonhole with rope ladder needling in conventional hemodialysis patients. Clin J Am Soc Nephrol. 7, (2012). 85. Tai, D. J., Ahmed, S. B., Palacios-Derflingher, L., Hemmelgarn, B. R. & MacRae, J. M. Pneumatic compression devices during hemodialysis: a randomized crossover trial. Nephrol Dial Transplant. 28, (2013). 86. CHEP. The 2013 Canadian Hypertension Education Program Recommendations. (2013). 87. Reeves, R. A. The rational clinical examination. Does this patient have hypertension? How to measure blood pressure. JAMA. 273, (1995). 88. Dasselaar, J. J., Huisman, R. M., de Jong, P. E., Burgerhof, J. G. M. & Franssen, C. F. M. Effects of relative blood volume-controlled hemodialysis on blood 78

88 pressure and volume status in hypertensive patients. ASAIO J. 53, (2007). 89. Hecking, M. et al. Blood volume-monitored regulation of ultrafiltration in fluidoverloaded hemodialysis patients: study protocol for a randomized controlled trial. Trials 13, 79 (2012). 90. Chan, A., Tetzlaff, J. M., Altman, D. G., Dickersin, K. & Moher, D. SPIRIT 2013: new guidance for content of clinical trial protocols. Lancet 381, 91 2 (2013). 91. Wang, Z. et al. Hydration of fat-free body mass: new physiological modeling approach. Am. J. Physiol. 276, E995 E1003 (1999). 92. van Biesen, W. et al. Fluid status in peritoneal dialysis patients: The European body composition monitoring (EuroBCM) study cohort. PLoS One 6, 1 9 (2011). 93. Hecking, M. et al. Significance of interdialytic weight gain versus chronic volume overload: Consensus opinion. Am. J. Nephrol. 38, (2013). 94. Cybulsky, A. V, Matni, A. & Hollomby, D. J. Effects of high sodium dialysate during maintenance hemodialysis. Nephron 41, (1985). 95. Sang, G. L., Kovithavongs, C., Ulan, R. & Kjellstrand, C. M. Sodium ramping in hemodialysis: a study of beneficial and adverse effects. Am. J. Kidney Dis. 29, (1997). 96. Mc Causland, F. R. & Waikar, S. S. Optimal dialysate sodium-what is the evidence? Semin. Dial. 27, (2014). 97. Moret, K. et al. The effect of sodium profiling and feedback technologies on plasma conductivity and ionic mass balance: a study in hypotension-prone dialysis patients. Nephrol. Dial. Transplant 21, (2006). 98. de Paula, F. M. et al. Clinical consequences of an individualized dialysate sodium 79

89 prescription in hemodialysis patients. Kidney Int. 66, Keen, M. L. & Gotch, F. A. The association of the sodium setpoint to interdialytic weight gain and blood pressure in hemodialysis patients. Int. J. Artif. Organs 30, (2007) Dasselaar, J. J., Huisman, R. M., de Jong, P. E. & Franssen, C. F. M. Measurement of relative blood volume changes during haemodialysis: merits and limitations. Nephrol Dial Transplant. 20, (2005) Thijssen, S., Kappel, F. & Kotanko, P. Absolute blood volume in hemodialysis patients: Why is it relevant, and how to measure it? Blood Purif. 35, (2013) Flythe, J. E., Curhan, G. C. & Brunelli, S. M. Disentangling the ultrafiltration ratemortality association: the respective roles of session length and weight gain. Clin. J. Am. Soc. Nephrol. 8, (2013) Saran, R. et al. Longer treatment time and slower ultrafiltration in hemodialysis: associations with reduced mortality in the DOPPS. Kidney Int. 69, (2006) Movilli, E. et al. Association between high ultrafiltration rates and mortality in uraemic patients on regular haemodialysis. A 5-year prospective observational multicentre study. Nephrol. Dial. Transplant 22, (2007) Assimon, M. M., Wenger, J. B., Wang, L. & Flythe, J. E. Ultrafiltration Rate and Mortality in Maintenance Hemodialysis Patients. Am. J. Kidney Dis. 68, (2016) Yilmaz, Z. et al. Evaluation of fluid status related parameters in hemodialysis and peritoneal dialysis patients: Clinical usefulness of bioimpedance analysis. Med. 50, (2014). 80

90 107. Ronco, C. et al. Baseline hydration status in incident peritoneal dialysis patients: the initiative of patient outcomes in dialysis (IPOD-PD study). Nephrol. Dial. Transplant 30, (2015) van Rij, A. M. et al. Obesity and impaired venous function. Eur. J. Vasc. Endovasc. Surg. 35, (2008) Willenberg, T. et al. Impact of obesity on venous hemodynamics of the lower limbs. J. Vasc. Surg. 52, (2010). 81

91 APPENDIX A: NURSING IDH MANAGEMENT ALGORITHM BP = blood pressure; HD = hemodialysis; HR = heart rate; IV = intravenous; LOC = level of consciousness; NP = nasal prongs; PRN = as needed; Q5 = every 5; SOB = short of breath; TW = target weight; UF = ultrafiltration; RBV = relative blood volume. 82