Ultrafiltration Rate and Residual Kidney Function Decline: Yet Another Good Reason to Ask About Urine

Epidemiologic studies examining ultrafiltration rate (UFR) and adverse outcomes in thrice-weekly hemodialysis patients have linked rapid UFRs to higher risks of mortality,^1,2 mesenteric ischemia,³ and new-onset de mentia.⁴ Data from mechanistic studies support these findings, demonstrating that hemodialysis-induced circulatory stress can cause ischemic injury to end-organs including the heart, gut, brain, and kidneys.⁵ Based on consistent observational data and strong biological plausibility, the United States (US) Centers for Medicare & Medicaid Services incorporated a UFR reporting metric into the 2020 End-Stage Renal Disease Quality Incentive Program, heightening the dialysis community’s awareness of fluid removal practices.

In this issue of AJKD, Lee et al provide additional data suggesting that exposure to higher UFRs associates with adverse clinical outcomes among individuals receiving maintenance hemodialysis at a frequency of twice-weekly or less.⁶ Using electronic health care record (EHR) data from a large US dialysis provider, the authors evaluated the association between UFRs and mortality among 1,524 patients receiving hemodialysis twice-weekly or less. After excluding patients without baseline UFR or renal urea clearance (KRU) data, they found that exposure to more rapid UFRs was associated with an incrementally higher risk of mortality. Among individuals with greater baseline residual kidney function (RKF), as determined by KRU, the association between UFR and mortality was attenuated. In addition, Lee et al assessed the association between UFR and RKF decline in a subset of 749 patients who survived a year after starting less-frequent hemodialysis and had at least 1 KRU measurement after their baseline measurement. The authors defined rapid RKF decline as a 1-year reduction in KRU greater than the cohort’s median value. Exposure to more rapid UFRs was associated with an incrementally higher odds of rapid RKF decline.

The study by Lee et al extends the evidence base by demonstrating that the UFR-mortality association also exists among individuals receiving hemodialysis on a less-frequent basis, and that rapid fluid removal may contribute to RKF decline. Delivered UFR is calculated based on ultrafiltration volume, dialysis treatment time, and postdialysis weight. Interdialytic weight gain strongly informs the prescribed ultrafiltration volume and is influenced by the extent of RKF. Larger interdialytic weight gains and/or shorter dialysis treatments result in more rapid UFRs. One plausible mechanism for the UFR-RKF association observed by Lee et al is that overly rapid UFRs may trigger hemodynamic compromise, leading to kidney ischemia and a subsequent decrement in RKF. Although this study did not show an association between the occurrence of intradialytic hypotension and RKF decline, others have.⁷ Moreover, intradialytic imaging studies of the heart⁸ and brain⁹ suggest that clinically significant end-organ ischemia can occur during hemodialysis, even in the absence of frank hypotension.

The strengths of the investigation by Lee et al include the relatively large cohort of patients receiving less-frequent hemodialysis; use of an EHR database, enabling the authors to consider important biochemical and dialysis treatment-related parameters; and the inclusion of multiple sensitivity analyses. Interpretation of the study’s findings must be considered in the context of the data vintage (2007–2011, an era of higher UFRs) and other limitations. Fluid removal practices in the US have changed in recent years, with mean UFRs declining from 9.2 mL/h/kg in 2011 to 7.8 mL/h/kg in 2019.¹⁰

Beyond the issue of present-day generalizability, the authors acknowledged several limitations, including possible underestimation of RKF by KRU; a cohort comprising a mixture of incident and prevalent patients, including individuals who initiated maintenance hemodialysis on a less-frequent regimen (11%) and those who switched from another dialysis modality or treatment schedule (89%); and the presence of residual confounding due to unmeasured factors such as the clinical indication for starting a less-frequent regimen and objective volume status. However, other clinically relevant unmeasured confounders may have influenced results. For example, diuretic use was not considered despite its association with lower interdialytic weight gain¹¹ (a strong correlate of the exposure, UFR) and improved survival¹² (a study outcome).

Of importance, the study findings rest on the accuracy of EHR-derived RKF data. As any clinician who makes dialysis rounds knows, interdialytic urine collections are challenging. So much so that one study found that less than 5% of hemodialysis patients have recorded RKF information.¹³ Thus, it is likely that RKF data are not missing at random in EHRs, and when present. may be unreliable. For example, 24-hour urine collections are more common in individuals with greater RKF, and greater RKF is associated with better outcomes. Although comparisons of patients with and without RKF data are not presented by Lee et al, it is plausible that individuals with routine RKF assessments are healthier than those without.

RKF assessment remains a mainstay in peritoneal dialysis care, where both dialysis and native kidney clearance contribute to urea-based dialysis dose calculations.¹⁴ Among individuals receiving peritoneal dialysis, greater RKF (defined using urea metrics and urine volume) associates with improved survival.^15,16 The benefits of preserved RKF likely stem from improved volume control, greater middle-molecule and protein-bound solute clearance, and reduced inflammation,¹⁷ among other effects. In peritoneal dialysis, routine RKF assessment informs diuretic dosing and the number, volume, and composition of exchanges, facilitating individualized therapy.

In theory, RKF assessment should serve an analogous role in hemodialysis care. The 2015 KDOQI guideline for hemodialysis dose recommends, albeit at an ungraded level, that weekly dialysis dose calculations should include contributions from dialysis and native kidney urea clearance among individuals on other than thrice-weekly schedules.¹⁸ Routine RKF assessment in hemodialysis would inform diuretic prescribing, ultrafiltration goals, and dose assessments, with the latter identifying individuals appropriate for less than thrice-weekly schedules and those who need to transition to more frequent schedules.

One barrier to incorporating RKF into hemodialysis decision-making is the lack of a universally accepted measure of RKF (Table 1). Second, although equations incorporating both native kidney and intermittent dialyzer solute clearance into weekly dialysis dose assessments exist, they are somewhat cumbersome and poorly understood.¹⁹ Third, urine volume, serum urea, and serum creatinine fluctuate over the interdialytic period. To estimate RKF most accurately, it would be necessary to collect urine across the entire interdialytic interval and measure serum urea and serum creatinine at the interval’s flanking hemodialysis treatments. Unfortunately, data extrapolation from less daunting, shorter urine collections (eg, 8 or 12 hours) to the entire interdialytic period is likely innaccu-rate due to the non–steady-state nature of the interdialytic interval. Finally, urine collections, regardless of duration, are difficult to complete, as underscored by their high degree of missingness in EHRs. Patients must collect their urine in bulky jugs—a challenging task for many, and an impossible one for those with dexterity or vision limitations. Moreover, the test may be disagreeable to patients since it requires urine jug storage in home refrigerators and transport when leaving home.

Table 1.

Select Measures of Residual Kidney Function

Measure	Comments
Plasma urinary inulin clearance	Gold standard Impractical
GFR estimated by creatinine and/or ureaa	Recommended by European Best Practice Guidelines for hemodialysis Not validated in the dialysis population Associates with nutritional status When creatinine is used: overestimation of true GFR due to tubular secretion When urea is used: underestimation of true eGFR due to tubular reabsorption Does not capture middle-molecule and protein-bound solute clearance
Urinary urea clearancea,b	Recommended by KDOQI and the International Society of Peritoneal Dialysis Associates with nutritional status Underestimates true clearance Does not capture middle-molecule and protein-bound solute clearance
Urinary creatinine clearancea	Overestimates true clearance Does not capture middle-molecule and protein-bound solute clearance
Plasma biomarkers (eg, cystatin C, β2-microglobulin, β2-trace protein)	Used to develop new estimating equations for RKF Impractical (not widely available) and additional research is needed
Measured urine volumea	No estimate of solute clearance Not linearly correlated with solute clearance
Patient-reported urine volumea	Least burdensome (no urine collection) No estimate of solute clearance Not linearly correlated with solute clearance

Abbreviations: eGFR, estimated glomerular filtration rate; KDOQI, Kidney Disease Outcomes Quality Initiative; RKF, residual kidney function.

^aAssociates with clinical outcomes in observational studies.

^bMeasurement used by Lee et al.⁶

While these barriers are surmountable, we may be overcomplicating RKF assessment. A recent DOPPS analysis found that an affirmative response to the simple question, “Did the patient have urine output > 200 mL/day or 1 cup/day at the time of enrollment?” was associated with lower mortality.²⁰ This supports a previous study showing that patient-reported 24-hour urine volume of “at least 1 cup” is associated with improved survival.²¹ Although patient-reported urine volume does not provide insight into solute clearance, it can guide volume management decisions and, in some cases, identify patients for whom formal timed urine collections are warranted. For instance, a decline in patient-reported urine volume in an individual on a less-frequent hemodialysis regimen could prompt earlier investigation for RKF decline, thereby yielding a transition to more frequent dialysis before overt complications arise.

To meet the growing calls to individualize dialysis²² by expanding the use of less-frequent hemodialysis as detailed in the article by Lee et al and other tailored regimens, we must use the plethora of data points available to inform decision making. Unfortunately, RKF, even in a simple patient-reported form, is not a consistent data point. Longitudinal RKF assessment would allow clinicians to monitor and proactively protect RKF (to the degree possible), titrate treatment length and frequency, and facilitate the much-needed study of RKF preservation strategies. Although some predictors of RKF decline are not modifiable (eg, older age, diabetes, and heart failure), other predictors are (eg, intradialytic hypotension).¹⁷ Findings from Lee et al add higher UFRs among individuals receiving less-frequent hemodialysis to the list of modifiable factors. However, to make these findings actionable, we must prioritize routinely asking our patients the simple but illuminating question, “How much urine do you make in a day?” until we have more readily available, accurate measures of RKF.