Revisiting Filtration Fraction As an Index of the Risk of Hemofilter

Revisiting Filtration Fraction as an Index of the Risk of Hemoﬁlter Clotting in Continuous Venovenous Hemoﬁltration

Parta Hatamizadeh,1 Ashita Tolwani,2 and Paul Palevsky3 CJASN 15: 1660–1662, 2020. doi: https://doi.org/10.2215/CJN.02410220

1Division of Introduction calculated filtration fraction will increase, while there is Nephrology, Hypertension and Hemofilter clotting is a major challenge in continuous no evidence to suggest that increasing the prefilter re- fi Renal Transplantation, venovenous hemo ltration (CVVH). Experts recommend placement rate increases the risk of hemofilter clotting. University of Florida, keeping filtration fraction, the fraction of plasma water Therefore, using filtration fraction as an index for risk Gainesville, Florida 2 that is removed by ultrafiltration, below 20% or 30% to stratification of hemofilter clotting has serious limitations, Division of Nephrology, avoid clotting (1). Clotting risk is believed to markedly particularly in the prefilter replacement setting. fi University of Alabama increase when the post lter hematocrit (Hct) exceeds at Birmingham, 0.4–0.5, and a filtration fraction of 30%–50% is pre- Birmingham, Alabama sumably associated with an increase in the postfilter The Origin of Applying a Single Universal 3Renal-Electrolyte Hct to this range (2). Filtration Fraction Cutoff Point to Prevent Division, University of Pittsburgh, Pittsburgh, However, a key question is whether a single value Hemofilter Clotting Pennsylvania of maximal filtration fraction is appropriate in all The application of filtration fraction for risk strat- fi fi situations, independent of other parameters such as i cation of hemo lter clotting and the assignment of a Correspondence: Hct and concentration of plasma proteins. Addition- single cutoff point as the maximum allowable filtra- Dr. Parta ally, the role of prefilter as compared with postfilter tion fraction in all clinical scenarios has been endorsed Hatamizadeh, administration of replacement fluids contributes to by many authors without providing strong evidence Division of Nephrology, additional confusion and uncertainty regarding the apart from referring to prior publications, which fi fi University of Florida, interpretation of ltration fraction. Given the de nition of themselves have not provided any compelling evi- 1600 SW Archer Road, filtration fraction (FF), when replacement fluids are dence (1,4). It appears that these anecdotes originate Room CG-98, P.O. administered postfilter, it is calculated as below: from some publications from the early 1980s, where Box 100224, their observed filtration fraction in continuous arteriove- Gainesville, FL 32610. QUF 2 net 1 QRF 2 post fi Email: hatamizadehp@ FF5 (1) nous hemo ltration (CAVH) was mainly between 20% ufl.edu QB 3 ðÞ1 2 Hct and 30% (7,8). However, those publications report experience with CAVH, and not CVVH, which is the Where QUF-net is fluid removal rate, QRF-post is postfilter modality that is currently used. Unlike CVVH, CAVH fluid replacement rate, and QB is blood flow rate. does not allow for control of the filtration rate because For the calculation of filtration fraction when replace- it is a function of the existing arteriovenous hydrostatic ment fluid is administered prefilter, most experts propose pressure gradient, oncotic pressures, and membrane adding the prefilter replacement fluid rate (QRF-pre)to and circuit specifications. Furthermore, these publica- both numerator and denominator, as in Equation 2 (3,4): tions simply report the observed filtration fraction achieved in the CAVH setting, the importance of which Q 1 Q FF5 UF 2 net RF 2 pre (2) was about attainment of convective clearance rather than Q 3 ðÞ1 2 Hct 1 Q B RF 2 pre hemofilter clotting. In fact, they have not reported any correlation between filtration fraction and hemofilter According to Equations 1 and 2, with fixed QB and clotting. Therefore, these studies cannot be extrapolated QUF-net, at any fluid replacement rate, the filtration to CVVH and provide no information regarding the fraction is higher in the postfilter compared with the relationship between filtration fraction and hemofilter prefilter replacement setting. Therefore, when com- clotting; hence, they should not have been the basis for paring pre- and postfilter replacement settings, as- defining a single filtration fraction value as the maximum suming that a lower filtration fraction is associated allowable filtration fraction in CAVH. with a lower risk of hemofilter clotting would be consistent with these calculations because studies have shown a lower risk of clotting when using the prefilter A Better Alternative than Filtration Fraction for compared with the postfilter fluid replacement setting Determining the Risk of Hemofilter Clotting (5,6). However, according to Equation 2, in the prefilter In 1992, Jenkins et al. (9) showed in an experimental replacement setting, by increasing the QRF-pre, the studythatataHctcloseto0.45,“operational

Figure 1. | Postfilter hematocrit rises exponentially by increasing postfilter replacement rate but is not affected by alterations of prefilter replacement rate. In postfilter replacement setting, with incremental increase in replacement fluid rate, the postfilter hematocrit increases exponentially; whereas in prefilter replacement setting, the postfilter hematocrit remains constant irrespective of prefilter replacement rate, provided that all other parameters remain constant. The graph shows changes in postfilter hematocrit in pre- and postfilter replacement settings in a patient with hematocrit of 0.25, blood flow rate of 200 ml/min, and no fluid removal. In practice, postfilter hematocrits .0.5 may not be achievable in most circumstances because blood will clot and the filter will stop working when the hematocrit reaches such a high level. The x axis represents replacement fluid rate (ml/min) and the y axis shows the postfilter hematocrit. Hct, hematocrit.

instability” of the hemofiltration system and impending Hctmax 2 Hctpatient FFmax5 À Á (6) circuit clotting occur. Hctmax 3 1 2 Hctpatient Because hemoconcentration, not filtration fraction per se, fi fi de nes the risk of hemo lter clot formation, calculating Equation 6 reveals the significant effect of the patient’s Hct fi end-of-hemo lter hemoconcentration is likely to be more on FFmax For example, assuming Hctmax50.4, the FFmax for fi . useful than the ltration fraction. Using Hct as a surrogate a Hct of 0.35 is 19%, whereas for a Hct of 0.25, it is 50%. fi for hemoconcentration, we can calculate the ltration For practical purposes, calculating the maximum allow- fi fraction that can lead to a certain end-of- lter Hct, depend- able ultrafiltration rate (QUF-total, which is QUF-net1QRF-post) ’ ing on the patient sHct. is useful. Using Equation 3, replacing Qpost with QB2 According to basic mathematical concepts, we know: QUF-total and solving the equation for QUF-total result in Equation 7: Hctpre 3 Qpre5Hctpost 3 Qpost (3) Hct 2 Hct Q 5Q 3 post pre (7) Where Hctpre is prefilter Hct, Qpre is prefilter flow (which is UF-total B Hctpost QB), Hctpost is postfilter Hct, and Qpost is postfilter flow. In postfilter replacement, Qpost can be calculated by sub- The maximum allowable ultrafiltration rate can then be tracting the ultrafiltration flow from the prefilter flow which, calculated using Equation 8: according to Equation 1, will result in the following equation:

Hctmax 2 Hctpatient Qpost5QB 2 QUF 2 net 2 QRF 2 post QUF-total-max5QB 3 (8) À Â ÃÁ Hctmax 5QB 2 FF 3 QB 3 1 2 Hctpre (4) Therefore, the higher the Hct and the lower the QB, the lower By replacing Qpost in Equation 3, according to Equation 4, the QUF-total should be set to avoid hemofilter clotting. and solving it for FF, the following formula ensues: Equations 4–8 are applicable only in the postfilter replacement setting. For the prefilter replacement setting, Hctpost 2 Hctpre FF5 À Á (5) in the absence of any net ultrafiltration (when QUF-total5QRF-pre), Hctpost 3 1 2 Hctpre thesamevolumethathasbeengivenastheprefilter replacement fluid will be removed from it within the By plugging in the patient’s Hct for Hctpre and the maximum hemofilter, so the end-of-hemofilter Hct returns to the end-filter Hct threshold (Hctmax) for Hctpost, one can calculate predilution Hct, i.e.,thepatient’sHct.Therefore,witha the maximum allowable filtration fraction (FFmax) to avoid fixed fluid removal rate, any increase or decrease in QRF-pre reaching a Hct level at which the risk of filter clotting is increases or decreases the convective clearance, respec- thought to be unacceptably high (Equation 6). tively; however, it does not change end-of-hemofilter Hct 1662 CJASN

(Figure 1). Therefore, it should not affect the likelihood of Acknowledgments filter clotting and should not be factored in calculations of The content of this article does not reflect the views or opinions CJASN FFmax or QUF-max. of the American Society of Nephrology (ASN) or .Re- Thus, for the calculation of FFmax and QUF-max in prefilter sponsibility for the information and views expressed herein lies replacement setting, the same formulas as those of the entirely with the author(s). postfilter replacement setting (Equations 6 and 8, respec- tively) may be used with the understanding that, in the References prefilter replacement setting, filtration fraction represents 1. Tolwani A: Continuous renal-replacement therapy for acute the filtered fraction of plasma (ignoring the prefilter fluid kidney injury. N Engl J Med 367: 2505–2514, 2012 2. Pedrini LA: On-line hemodiafiltration: Technique and efficiency. replacement rate) and QUF-max refers to QUF-net-max (maxi- fl J Nephrol 16[Suppl 7]: S57–S63, 2003 mum uid removal rate). 3. MacEwen C, Watkinson P,Winearls C: Circuit life versus bleeding risk: The impact of achieved activated partial thromboplastin time versus achieved filtration fraction. Ther Apher Dial 19: 259–266, Conclusion 2015 Although end-of-filter Hct, as an index of hemoconcen- 4. Neri M, Villa G, Garzotto F,Bagshaw S, Bellomo R, Cerda J, Ferrari tration, is more relevant than filtration fraction in de- F, Guggia S, Joannidis M, Kellum J, Kim JC, Mehta RL, Ricci Z, fi Trevisani A, Marafon S, Clark WR, Vincent JL, Ronco C; No- termination of the risk of hemo lter clotting, it still menclature Standardization Initiative (NSI) alliance: Nomen- overlooks other contributing factors such as the circuit clature for renal replacement therapyin acute kidney injury: Basic shape and design and the coagulability state of the patient. principles. Crit Care 20: 318, 2016 Nevertheless, reliance on a single arbitrary maximum 5. Uchino S, Fealy N, Baldwin I, Morimatsu H, Bellomo R: Pre- dilution vs. post-dilution during continuous veno-venous he- allowable filtration fraction, as an index for risk stratifica- fi mofiltration: Impact on filter life and azotemic control. Nephron tion of hemo lter clotting, lacks accuracy and validity. Clin Pract 94: c94–c98, 2003 End-of-hemofilter Hct, which is a contributing factor to 6. van der Voort PH, Gerritsen RT, Kuiper MA, Egbers PH, Kingma hemofilter clotting, is a more meaningful index. We pro- WP,Boerma EC: Filter run time in CVVH: Pre- versus post-dilution and nadroparin versus regional heparin-protamine anti- pose using the above-mentioned formulas to adjust QUF to fi coagulation. Blood Purif 23: 175–180, 2005 maintain the calculated end-of-hemo lter Hct below a 7. Kramer P,Kaufhold G, Gro¨ne HJ, Wigger W, Rieger J, Matthaei D, certain number which, based on currently available data Stokke T, Burchardi H, Scheler F: Management of anuric (9,10), appears to be approximately 0.4. intensive-care patients with arteriovenous hemofiltration. Int Although conceptually sound, these equations need to be J Artif Organs 3: 225–230, 1980 8. Lauer A, Saccaggi A, Ronco C, Belledonne M, Glabman S, Bosch validated in clinical studies. JP: Continuous arteriovenous hemofiltration in the critically ill patient. Clinical use and operational characteristics. Ann Intern Disclosures Med 99: 455–460, 1983 P. Hatamizadeh has a patent titled renal replacement therapy 9. Jenkins RD, James EF, Chen B, Golper TA: Operational instability machine. P. Palevsky reports receiving personal fees from Baxter, in extracorporeal filtration of blood. Blood Purif 10: 292–308, grants from Dascena, and grants from BioPorto, outside the 1992 10. Cinar Y, Demir G, PaçM, Cinar AB: Effect of hematocrit on blood submitted work. The remaining author has nothing to disclose. pressure via hyperviscosity. Am J Hypertens 12: 739–743, 1999

Funding Published online ahead of print. Publication date available at None. www.cjasn.org.