Volume, Blood Pressure & Hemodynamic Tolerance in Maintenance HD

What This Guide Is — and What It Is Not

This is a clinician-only reference for the chronic problems of volume, BP and treatment tolerance in adults on maintenance in-center HD. It is written for nephrologists, fellows, dialysis physicians and nurses, renal pharmacists and renal dietitians. It is meant for desktop rounds and bedside use, and it prints as a reference copy. It complements two existing guides on this site: the practical prescription detail lives in Iso-UF, Sodium & UF Ramping, and the patient-facing explanation of target weight lives in Finding Your True Dry Weight.

Three things are deliberately outside scope. Acute kidney injury (AKI), intensive care, and continuous kidney replacement therapy have different physiology and their own guide (CRRT). Peritoneal dialysis (PD), pediatric dialysis, pregnancy and kidney transplant recipients get boundary notes only — no numeric rule in this guide should be carried across to them. Isolated heart-failure UF is a cardiology topic with its own trials.

How claims are labeled

Every recommendation here is attributed to its source and kept in the source's own grading language. The UK Kidney Association (UKKA) 2025 guideline grades are quoted as UKKA grades (for example "2B"); the Kidney Disease: Improving Global Outcomes (KDIGO) 2020 conference report is consensus and carries no grade; the Kidney Disease Outcomes Quality Initiative (KDOQI) 2015 adequacy guideline is foundational but older. These systems are never converted into one another. Verbs follow the strength of evidence: "reduced" and "did not reduce" are reserved for outcomes in randomized controlled trials (RCTs); "is associated with" marks observational data; "plausibly" marks mechanism; "simulated" marks in silico work. Where this guide offers its own synthesis — the review lens and the seven pathways — it says so.

Where the Fluid Is, and What Keeps BP Up While It Leaves

The intuition first. Think of the plasma as a small bucket sitting inside a much larger, waterlogged sponge. The dialyzer can only drain the bucket. Every liter that leaves the body during a session has to pass through the bucket on its way out, and the bucket is topped up only as fast as the sponge can push water back in. If you drain faster than the sponge refills, the bucket runs low — even though the sponge is still soaked. That is the whole story of "congested but hypotensive," and the rest of this section replaces the analogy with the real physiology.

1 · The compartments

Total body water divides into intracellular water and extracellular water; extracellular water divides again into the interstitial fluid that bathes cells and the plasma that circulates. The plasma is the smallest of these and is the only compartment the dialyzer touches. Congestion — peripheral edema, lung water, an elevated jugular venous pressure — is largely interstitial and pulmonary-extravascular, and it is this excess that bioimpedance spectroscopy (BIS) and lung ultrasound try to estimate. So when a clinician says a patient is "overloaded," they are usually describing the sponge; when a patient's BP falls during UF, the problem is usually the bucket.

2 · Plasma refill is the rate-limiting step

During UF, removing plasma water raises plasma oncotic pressure and lowers capillary hydrostatic pressure, so Starling forces favor movement of fluid from the interstitium back into the capillaries. How fast that happens depends on interstitial pressure and compliance (a patient with a lot of excess interstitial fluid generally refills more readily than one near euvolemia), plasma albumin, capillary permeability (inflammation and sepsis change it), and time. Refill is not instantaneous, and its rate varies between patients and within the same patient across a session. The ultrafiltration rate (UFR) is the variable the team controls; refill is the variable it can only estimate. When UFR exceeds refill for long enough, plasma volume falls — this is what a falling relative blood volume (RBV) trace shows. Note the word relative: an RBV curve reports a percentage change from the start of the session, not an absolute blood volume (ABV).

3 · What defends BP when plasma volume falls

A shrinking plasma volume reduces venous return and cardiac filling. Whether BP holds depends on compensation. The ventricle matters first: many patients on HD have a thickened left ventricle (LV) — left ventricular hypertrophy (LVH) — and diastolic stiffness, which put them on the steep part of the Frank–Starling curve, where a small fall in filling causes a large fall in stroke volume. Arteriolar constriction and the heart-rate response then defend pressure through the autonomic nervous system, which is often blunted by diabetes, age and uremia. Venous capacitance matters too: splanchnic and cutaneous veins can hold a large, mobile reserve of blood, and warming or vasodilation pools blood there and cuts preload further — the physiologic basis for cooling the dialysate. Arrhythmia, myocardial ischemia, infection, bleeding and drugs can each break this chain independently of any volume error.

4 · Why the consequence is more than a low reading

Recurrent falls in perfusion during HD are plausibly linked to injury in vulnerable vascular beds — the myocardium (regional wall-motion abnormalities, sometimes called myocardial stunning), the brain, the gut and the residual kidney. A 2024 review synthesizes this "circulatory stress" model and the microvascular endothelial dysfunction that both drives and results from it.⁸ This is mechanistic evidence: it explains why intradialytic hemodynamics matter, but it does not by itself prove that any single intervention improves survival. The KDIGO conference made the same point about UFR — higher rates are associated with mortality, biologic plausibility supports an ischemic link, but no randomized trial has shown that lowering UFR improves outcomes.¹

Words That Are Often Used as Synonyms — and Are Not

Much of the confusion in volume management is linguistic. Two clinicians can agree that a patient is "at dry weight" while meaning a prescription, an estimate and a physiologic state respectively. KDIGO draws the key distinction explicitly: the target weight is the prescription and can vary from treatment to treatment, while the estimated dry weight is a clinical estimate of the weight at euvolemia; in acute illness or with severe symptoms it may be reasonable to keep a patient slightly above the estimated dry weight, but the long-term risks of chronic overload must be weighed.¹

Are the Numbers Good Enough to Act On?

Before reviewing any pathway, ask whether the observations are reliable. Volume decisions are made on weights, BP readings and session logs that are each prone to systematic error, and a precise calculation built on a wrong weight is still wrong. The UKKA guideline is explicit that routine dialysis-unit BP readings should inform the safety of delivering a session rather than hypertension management, because they agree poorly with interdialytic ambulatory readings.²

Which Question Is the BP Reading Answering?

A BP number in HD can answer three different questions: does this patient have hypertension, is there an intradialytic pattern, and is it safe to continue this session. Different measurements suit each. Interdialytic ambulatory blood pressure monitoring (ABPM) is the reference standard for diagnosis; home blood pressure monitoring (HBPM) or standardized out-of-unit clinic readings suit ongoing management; routine pre-, intra- and postdialysis readings are best used for session safety.² Observational data showing a U- or J-shaped relationship between routine predialysis systolic BP and mortality are confounded by illness and measurement error, whereas out-of-unit readings generally show a more linear relationship with outcomes — another reason not to treat the nadir of a routine-BP curve as a target.²

The only randomized test of BP targets in HD that the guidelines could draw on is the Blood Pressure in Dialysis pilot, which compared an intensive predialysis systolic goal of 110–140 mmHg with a standard goal of 155–165 mmHg and was designed to assess feasibility, not outcomes.¹ Any target displayed at the bedside should therefore carry its source, its measurement method and its grade; a "BP controlled" badge based on a single routine session reading is not meaningful.

Three Axes, Kept Separate

Instead of a single "volume status" label, hold three independent questions. Congestion evidence: present, absent on the assessment available, or uncertain. UF tolerance in reviewed sessions: adequate, recurrently impaired, or uncertain. RKF contribution: measured (with date and method), urine output only, or unknown. Then attach modifiers that shift interpretation without being scored: cardiac reserve, autonomic dysfunction, arrhythmia, frailty, vascular disease, nutrition and body-composition change, acute illness, the current prescription, medications, and the patient's own account of treatment burden.

This is a teaching lens authored for this guide, not a validated classification. Its purpose is to stop an uncertain case being forced into a confident branch. "Congested and hypotensive" is a first-class state, and "information needed" is a legitimate result.

Seven Review Pathways

Each pathway structures a review: identify the pattern, check the data, consider competing mechanisms, list options with their tradeoffs, and decide how to reassess. None outputs a target weight, a UFR or a drug order, and a patient can — and often should — enter more than one. They are editorial syntheses of the sources cited, offered for local adaptation and clinical judgment.

Ultrafiltration Rate, Treatment Time and Frequency

UFR is simply volume divided by time, which is why it is the most modifiable lever in the field: for a fixed volume, adding time lowers the rate in direct proportion. Removing 3.5 L from a 70 kg patient over 4 hours is 12.5 mL/kg/h; over 4.5 hours it is 11.1 mL/kg/h. The physiology says this should help, because a lower rate gives refill more time to keep pace. The evidence is real but observational, and it is worth reading carefully.

In a secondary analysis of the Hemodialysis (HEMO) Study, 1,846 patients on thrice-weekly HD with UFR above 13 mL/kg/h had higher all-cause and cardiovascular mortality than those at or below 10 mL/kg/h (adjusted hazard ratios 1.59 and 1.71); rates of 10–13 mL/kg/h were not associated with excess mortality overall, although spline analysis suggested risk began to rise above 10 mL/kg/h.⁵ In 396,358 United States patients, the UFR associated with a given excess mortality risk depended on body weight but not in a one-to-one ratio, and differed by sex, age-weight combination and dialysis vintage.⁶ Larger IDWG is likewise associated with higher cardiovascular death in a cohort of 34,107 patients.⁷ KDIGO concluded that, in the absence of randomized evidence that lowering UFR improves outcomes, a single UFR threshold for all patients at all times is likely inappropriate, and that decisions may differ treatment to treatment.¹

Does adding time or sessions help? What the trials show

The Time to Reduce Mortality in End-Stage Renal Disease (TiME) trial tried to answer this pragmatically: 266 units randomized incident patients to a default session of at least 255 minutes or usual care. Uptake was too low — mean session duration differed by only about 9 minutes (216 vs 207) — and the trial stopped at a median 1.1 years with no difference in mortality or hospitalization.²⁷ That is a failed test of the question, not evidence that time does not matter. The Frequent Hemodialysis Network (FHN) daily trial randomized 245 patients to six versus three in-center sessions per week; frequent HD improved the coprimary composites of death or change in left ventricular mass (hazard ratio 0.61) and death or change in physical-health score (0.70), improved hypertension and phosphate control, and increased vascular-access interventions (hazard ratio 1.71).²⁸ The KDOQI 2015 adequacy update added explicit emphasis on volume and BP control and on flexibility in frequency, duration and UFR.³ UKKA suggests extended hours for patients who fail to achieve BP control or experience IDH on standard thrice-weekly HD where resources allow (2A).²

Dietary Sodium, Fluid Intake and Dialysate Sodium

Water follows sodium. A sodium load raises plasma osmolality, triggers thirst, and the water the patient drinks to correct it is held in the extracellular space until the next session — so IDWG is largely a sodium problem disguised as a fluid problem. That is why fluid restriction without salt restriction tends to fail. UKKA recommends dietary salt reduction to a maximum of 5 g/day (1B) and suggests individualized fluid restriction together with salt reduction, considering urine output, gains and UF (2D).² Note the units: 5 g of sodium chloride contains about 2 g of sodium, because sodium is roughly 39% of the mass of salt.

The dialysate is the second sodium source. If dialysate sodium exceeds plasma sodium, diffusion adds sodium during the session, which can stabilize intradialytic BP but drive post-session thirst. The 2024 Cochrane review (17 randomized studies, 509 patients, mostly short crossover trials, several using outdated practice) found that low dialysate sodium (<138 mM) reduces IDWG (mean difference −0.36 kg, high certainty) and antihypertensive use, and probably reduces predialysis mean arterial pressure and left ventricular mass index — but probably increases IDH (risk ratio 1.58; 13 studies, 15,764 sessions) and cramps (risk ratio 1.84), both moderate certainty.¹⁶ Those IDH counts are sessions, not people, and effects on death and hospitalization remain uncertain.

A large pragmatic trial of dialysate sodium on hard outcomes, RESOLVE, is registered; this guide makes no statement about its status or results until they are published in a peer-reviewed source.

Temperature and Other Prescription Levers

Cooling. Warmth vasodilates cutaneous and splanchnic beds and pools blood away from the heart; cooler dialysate counteracts this. For the hypotension-prone patient, UKKA suggests lowering dialysate temperature to reduce IDH (2A).² That is a different question from the one tested in the Personalised Cooler Dialysate (MyTEMP) trial: 84 Ontario centers serving 15,413 patients were randomized to a center-wide policy of dialysate 0.5–0.9 °C below each patient's body temperature (floor 35.5 °C) or a standard 36.5 °C. Achieved temperatures were 35.8 vs 36.4 °C, and the composite of cardiovascular death or hospitalization for myocardial infarction, ischemic stroke or heart failure did not differ (adjusted hazard ratio 1.00, 96% confidence interval 0.89–1.11); the mean intradialytic systolic drop was almost identical.¹⁵ MyTEMP argues against cooling everyone for cardiovascular protection; it does not settle targeted cooling for a symptomatic individual, which remains a reasonable trial with attention to thermal discomfort.

Buffer, magnesium and calcium. UKKA recommends bicarbonate rather than acetate dialysate to reduce IDH risk (1B) and suggests avoiding low dialysate magnesium (≤0.25 mmol/L), particularly with dialysate calcium of 1.25 mmol/L, in patients at risk of IDH (2D).² Specific dialysate concentrations for your unit need review against full sources and local equipment.

Profiling and sequential UF. Isolated (sequential) UF, UF profiling and sodium profiling are distinct interventions with distinct evidence and tradeoffs. They are covered with prescription worksheets in Iso-UF, Sodium & UF Ramping and should not be assumed protective for all patients.

BP-Lowering Drugs, Timing and Pressors

Three questions keep medication decisions honest: what is the indication (BP lowering, cardioprotection, heart-failure therapy, rate control), what happens to the drug during HD (dialyzability and timing), and what is the monitoring burden. Doses are deliberately absent from this guide; use a current, locally verified prescribing reference.

What the UKKA recommends (graded, search to 2022)

*The published wording of Rec 30 appears to be missing a word; read it with its second sentence, which favors consistent evening dosing over pre-session omission when drugs are implicated in IDH.²

What the trials add

Timing. In the cluster-randomized TAKE-HOLD trial (10 units, 131 participants, 4 weeks), taking versus holding BP drugs before HD did not meet noninferiority for the primary outcome of asymptomatic IDH (difference 8%, 95% confidence interval −3% to 19%), but taking them reduced uncontrolled predialysis hypertension (difference −15%).¹⁷ Failed noninferiority is not proof that holding is better; it means the trade-off is real and individual.

Choice of agent. In the Hypertension in Hemodialysis Patients treated with Atenolol or Lisinopril (HDPAL) trial, 200 hypertensive patients with LVH were randomized to thrice-weekly post-dialysis atenolol or lisinopril. Ambulatory BP improved similarly, but serious cardiovascular events were more frequent with lisinopril (incidence rate ratio [IRR] 2.36).¹⁹ The trial was open-label and selected; it informs, but does not prove, a class effect.

Mineralocorticoid receptor antagonists — the newest and most important correction. Two large placebo-controlled trials published in 2025, after UKKA's search window, examined spironolactone for cardiovascular protection. ACHIEVE screened 3,689 patients, enrolled 3,565 into an open-label run-in, and randomized 2,538 who tolerated it; the composite of cardiovascular death or heart-failure (HF) hospitalization did not differ (hazard ratio 0.92, 95% confidence interval 0.78–1.09), all-cause death did not differ, and the monitoring committee stopped the trial for futility at a planned interim analysis.²¹ ALCHEMIST randomized 644 high-risk HD patients after run-in; the primary cardiovascular composite was identical in both arms (hazard ratio 1.00, 0.73–1.36), and the trial stopped early because sponsor funding ended — not for futility.²² Together they mean spironolactone should not be started for cardiovascular protection in maintenance dialysis. They do not test BP lowering in an individual with resistant hypertension, which is the narrower indication UKKA Rec 29 addresses — and the potassium cost still applies.

Midodrine. The 2004 systematic review behind its use pooled small, methodologically weak studies showing higher BP during midodrine exposure; its title overstates certainty, and no hard-outcome benefit has been shown.²⁰ Albumin infusion reduced IDH in a crossover trial of 65 hospitalized hypoalbuminemic patients (249 sessions); that inpatient result should not become an outpatient rescue protocol.³²

Changing the Target Weight — What Supports It, What to Watch

Target weight is a dynamic prescription. It should move when tissue mass changes (weight lost from illness or poor intake will otherwise be "removed" as fluid), when urine output changes, and when examination, BP pattern and symptoms change. The best randomized evidence for probing it is the Dry-weight Reduction in Hypertensive Hemodialysis Patients (DRIP) trial: 150 hypertensive patients, with 100 assigned to additional UF without extra time. Post-weight fell about 1 kg by 8 weeks and 44-hour ambulatory systolic BP fell 6.6 mmHg (95% confidence interval −12.2 to −1.0) relative to control.¹⁸ That is a BP endpoint over 8 weeks, not survival, and probing increased intradialytic hypotensive symptoms. DRIP supports careful, monitored probing in hypertensive patients; it does not support a fixed decrement per session.

Residual Kidney Function and Incremental Treatment

Residual function does two jobs in volume management: it removes fluid and sodium continuously, which lowers IDWG and the UFR needed, and it adds solute clearance. It is also vulnerable — hypotension, nephrotoxins and aggressive UF can hasten its loss. That is why the incremental approach (starting with fewer or shorter sessions while RKF is substantial) is attractive, and why its evidence must be read by design. Urine output is useful but is not clearance; a decision about frequency should rest on a dated, timed collection with urea clearance (Kru), alongside potassium, fluid and nutritional control. KDOQI 2015 allows dose and frequency to account for measured RKF.³

The defensible practice is nephrologist-led selection of patients with measured, adequate RKF, explicit escalation criteria agreed before starting, and scheduled re-measurement — not automatic frequency reduction from urine output.

Instrument-Assisted Volume Assessment

Each tool sees one compartment, at one moment, with its own assumptions. None returns "liters to remove." UKKA concluded that there is inadequate evidence to recommend one technology as superior to another or to clinical assessment (2C).²

What the outcome trials and meta-analyses show

The Lung water by Ultra-Sound guided Treatment (LUST) trial enrolled 367 high-cardiovascular-risk HD patients. A lung ultrasound–guided strategy relieved congestion more often (78% vs 56%) but did not reduce the primary composite of death, nonfatal myocardial infarction or decompensated heart failure (hazard ratio 0.88, 0.63–1.24); recurrent heart-failure findings were post hoc.⁹ BISTRO did not show slower loss of RKF with bioimpedance added to a structured protocol.¹⁰ The older Crit-Line Intradialytic Monitoring Benefit (CLIMB) trial (443 patients, one device and protocol) found more hospitalizations and higher mortality with blood-volume monitoring — a historical warning that a monitoring protocol can harm, not a verdict on modern devices.¹⁴

Pooled analyses point in different directions because they pool different things. A 2019 meta-analysis of 12 trials found no mortality reduction (rate ratio 0.92, 0.57–1.51) but a modest systolic BP reduction.¹¹ A 2023 bioimpedance meta-analysis (15 trials, 2,763 patients, HD and PD) found a pooled mortality reduction (0.71, 0.51–0.99) that was not significant in the HD-only subgroup (0.72, 0.42–1.22).¹³ A 2025 meta-analysis of 21 trials and 4,239 patients found fewer cardiovascular events with technology-guided dry-weight adjustment (relative risk 0.79, 0.71–0.88), a bioimpedance-subgroup mortality reduction (0.67), and more muscle cramps (rate ratio 1.09).¹² These reviews share trials, so their patient numbers must not be added; endpoint definitions and risk of bias differ. The honest summary: tools help confirm or refute clinical suspicion and may reduce some events, the hard-outcome signal is inconsistent, and harms such as cramps are real. Neither "technology has never helped" nor "technology proves survival benefit" is supported.

Hemodiafiltration, Extended Hours and Other Settings

In the CONVINCE trial, 1,360 patients who were candidates for high-dose HDF (convection volume ≥23 L per session) were randomized to high-dose HDF or high-flux HD; death from any cause was lower with HDF (17.3% vs 21.9%; hazard ratio 0.77, 0.65–0.93) over a median 30 months.²³ The population was selected for access and capability to reach that convection volume. Convection volume is not net fluid removal, and CONVINCE was not a trial of UF intolerance. UKKA suggests online HDF could be trialed in symptomatic IDH (2B) but only after alternative causes have been addressed (2C).² Readers wanting the mechanics should see Hemodiafiltration Deep Dive.

Home, nocturnal and extended-hours HD lower UFR by design and need their own surveillance; numeric rules from in-center HD should not be transposed to them, nor to PD, AKI, intensive care or continuous therapies.

Symptoms, Recovery and Reassessment

The patient's report is clinical data. Map symptoms across the whole cycle: before the session (dyspnea, edema, orthopnea after the long interval), during (cramps, dizziness, nausea, chest discomfort), after (post-dialysis fatigue and recovery time), and between sessions (thirst, function). KDIGO emphasizes building symptoms into prescription discussions and communicating risk–benefit trade-offs clearly.¹ Any named patient-reported outcome instrument should be validated and licensed for your population before it enters routine use. Neither "feels better" nor "BP improved" alone shows that both congestion and tolerance are controlled — track a benefit measure and a balancing measure together. For the fatigue side, see Post-Dialysis Fatigue.

Four Transparent Arithmetic Tools

These tools do arithmetic and show their working. They do not classify a result as safe, prescribe a weight or a rate, or store anything: values stay in this browser tab, are not sent anywhere, and clear when you press Reset or leave the page. Every result is an arithmetic estimate that needs clinical interpretation.

Eight Fictional Teaching Cases

All cases are invented for teaching. Each ends with what is still unknown and a reassessment plan, because in real practice the right answer is usually a sequence of reviews, not a single order.

Research Frontier — Interesting, Not Yet Actionable

Adaptive and closed-loop UF. A 2026 paired in silico study replayed 119,472 historical treatments from 11,246 patients through an adaptive UF controller guided by continuous RBV. Simulated time within a predefined RBV target region rose from 36.3% to 56.4% and improved in 90.0% of patients; UF volume rose on average but changed in both directions.²⁹ Several authors reported industry affiliations. This is a surrogate simulation — not an outcomes trial and not a validated prescribing engine.

Absolute blood volume–guided management. A 2026 two-center prospective observational cohort of 600 patients described more favorable 12-week renin–angiotensin–aldosterone and BP patterns in the group managed with ABV monitoring than with clinical assessment.³¹ Group assignment was not randomized, so confounding cannot be excluded.

Education tools. A 12-week randomized study of 80 patients found that a "traffic light" fluid diary improved knowledge and intermediate fluid measures.³⁰ It is a small, short, context-specific study, and color-coded adherence scores should not migrate into clinician judgments about patients.

Watch list. Conference-only results (for example, small spironolactone studies in incremental HD presented as abstracts) and registered trials without published results are tracked here and are not used to support any recommendation in this guide.

Using This Guide in a Dialysis Unit

A unit can adopt the review structure without turning it into protocol medicine. A practical approach is a monthly multidisciplinary review (nephrologist, charge nurse, dietitian, pharmacist where available) of patients flagged by recurrent IDH, repeated high UFR on delivered time, persistent missed targets, or out-of-unit hypertension. Each review records the three axes, the data gaps, the options considered, the plan and the reassessment date.

Local policies can add reviewed instructions — for example, your unit's IDH first-response steps — but should record the institution, author, approver, version, effective date and the reason for any divergence from the sources here. No claim is made that using this guide reduces mortality, and unadjusted unit outcomes should not be benchmarked against one another. For the quality-system wrapper — quality assessment and performance improvement (QAPI) versus continuous quality improvement (CQI) — see the dialysis QAPI vs CQI guide.

Key Studies at a Glance

Primary endpoint first, with its effect estimate and interval as reported. Abbreviations in the table: randomized controlled trial (RCT), meta-analysis (MA), hazard ratio (HR), risk ratio (RR), odds ratio (OR), incidence rate ratio (IRR), confidence interval (CI), cardiovascular (CV). "Read it as" states the main limit on inference. Numbers come from the source abstracts or full text indexed in PubMed; extraction depth beyond that is noted where it matters.