Ultrafiltration in hemodialysis - is 13mL/kg/hr UF rate the safest approach?
Updated: Aug 12
By: Gerren Hobby MD (@ghobby)
Adequacy and volume control are the two pillars of maintenance dialysis prescriptions and standards for both have been established. To manage volume, a dry weight is established at which signs and symptoms of fluid overload are absent while avoiding hypotension and cramping during dialysis. Although this is empirically derived, the ultrafiltration rate (UFR) at which we move towards a patient's dry weight has been studied in depth.
High ultrafiltration rates make patients more prone to intradialytic hypotension and potentially transient ischemia due to hypoperfusion in the heart, gut, and brain. Volume depletion from ultrafiltration reduces coronary blood flow leading to cardiac ischemia which is manifested by myocardial wall stunning on echocardiography. Over time, this process can lead to ventricular remodeling and subsequently, heart failure and arrhythmia. Ultrafiltration-induced intestinal ischemia additionally leads to bacterial translocation into the bloodstream. This causes increased blood endotoxin levels that can result in chronic inflammation. Lastly, repeated ultrafiltration-induced ischemia is associated with white matter changes in the brain, and dementia in dialysis patients. The end result is that higher ultrafiltration rates are associated with a host of biological consequences.
The importance of these physiological effects spurred analyses of the association of UFR and mortality. Despite the seemingly straight-forwardness of this endeavor, gaps in the literature exist and analyses have been laden with confounding variables which obscure the direct association between UFR and morbidity and mortality. The examination of the UFR and mortality began in 2006 with a paper by Saran et al. At that time it was known that excessive interdialytic weight gain, and its associated higher UFR, was an independent predictor of mortality. This paper studied the relationship between treatment time and ultrafiltration rate with patient outcomes in the Dialysis Outcomes and Practice Pattern Study (DOPPS). Ultrafiltration rates above 10 mL/h/kg were associated with a higher mortality risk (risk ratio of 1.09) as well as a higher incidence of intradialytic hypertension. A year later, in a paper by Movilli et al, patients with an ultrafiltration rate of less than 12.37 mL/h/kg were found to have better survival in maintenance hemodialysis patients as compared to those with a higher ultrafiltration rate.
The third major paper published in 2011 by Flythve, et al showed that an ultrafiltration rate greater than 13 mL/h/kg was associated with increased all-cause cardiovascular mortality as compared to those with an ultrafiltration rate of 10 mL/h/kg or less. Subsequent papers have confirmed this association. As a result, the Centers for Medicaid and Medicare Services established a quality measure to limit ultrafiltration rates to 13 mL/h/kg.
The interesting thing is that these major studies scaled ultrafiltration to body mass without first assessing a simpler, unscaled ultrafiltration rate. At its core, ultrafiltration rate during a particular dialysis session is primarily determined by interdialytic weight gain and treatment time. Patients with greater amounts of volume overload have been found to be associated with higher mortality rates. On the other hand, longer treatment time is associated with a reduced mortality. Its easy to see that a longer treatment time allows for a lower UFR. All of this is intuitive so far, but the plot thickens when you consider all factors that affect interdialytic weight gain and treatment time.
Interdialytic weight gain is impacted by fluid intake and residual kidney function. On the one hand, fluid intake could be increased by non-compliance with fluid restriction or conversely decreased due to poor nutritional intake. Although these two states tilt the scale in different directions, both malnutrition and volume overload increase mortality. Residual kidney function reduces interdialytic weight gain, but also introduces a confounder as residual kidney function confers a mortality benefit in and of itself.
The second major factor in ultrafiltration rate, prescribed dialysis treatment time, also has potential confounding variables. Longer treatment times have the potential to be prescribed due to low adequacy, recirculation in dialysis access, higher phosphorus levels, or potentially intradialytic hypotension that necessitate lower ultrafiltration rates. Each of these independent variables and their prescribed treatments will affect mortality in unique ways.
Despite the formidable number of variables mentioned above, scaling ultrafiltration to body mass adds yet another layer of complexity. Improved nutrition will lead to a higher body mass, but also increased sodium intake and thus increased interdialytic weight gain. Malnourished patients will likely have lower interdialytic weight gains resulting in a lower UFR, but also a low body mass which also affects the denominator of scaled UFR.
In sum, multiple layers of complexity exist when evaluating something as seemingly simple as associating scaled UFR and mortality. Although past literature has attempted to minimize the effect of these confounding factors on ultrafiltration rate, no papers to date have examined the association of mortality and ultrafiltration rate, which are not scaled to body mass. The recent paper by Raimann, et al. addresses this gap in the literature. Take a look at the amazing visual abstract by @CTeodosiu below.
This was an observational cohort study of 2,542 patients on thrice-weekly in-center hemodialysis between January 1, 2014 and October 31, 2018. The mean age of the patients was 62 years. 42% were female, 40% were African American, and 39% had a history of diabetes mellitus. During the study period, a total of 494 patients died.
Unscaled ultrafiltration rates were evaluated for their association with mortality. Next, ultrafiltration rates scaled to body mass were examined over a wide range of body mass categories. It was found that an unscaled ultrafiltration rate of greater than 1,000 mL/h was associated with a mortality hazard ratio at 1.5. This relationship was independent of post-dialysis body weight between a body mass of 80kg to 140 kg. A slightly lower ultrafiltration rate of 900 milliliters per minute was associated with an expectedly lower mortality hazard ratio of 1.3.
Interestingly, when scaled ultrafiltration rates were viewed in patients with different body mass, a trend was found that a single ultrafiltration rate, scaled to body mass, was associated with higher mortality rates in larger patients. An ultrafiltration rate of 13 mL/h/kg had a mortality hazard ratio (MHR) 1.2 for a body mass of 60 kg, a MHR 1.45 for a body mass of 80 kg and a MHR greater than 2.0 for a body mass of 100 kg.
In summary, this paper was the first to examine the association of unscaled and scaled ultrafiltration with mortality. It found that a higher ultrafiltration rate is indeed associated with an increased mortality as all of the past literature has shown, but it did show that an ultrafiltration rate of greater than 13mL/h/kg is associated with a higher risk mortality in those with a higher body mass compared to those of a lower body mass. This means that the long held quality measure of a UFR of 13mL/h/kg is disadvantageous to higher body mass patients. This paper calls into question the suitability of scaled ultrafiltration rates, as well as providing a basis for much needed future prospective trials evaluating an optimal UFR.