Dialysis adequacy

The optimal outcome for animals with acute renal failure is survival until renal function has recovered, but secondary goals may vary qualitatively depending on the nature of the underlying disease. An optimal outcome additionally should promote physiologic and metabolic stability to facilitate recovery and promote an acceptable quality of life while minimizing secondary injury to the recovering kidneys or other organs (heart, lungs, gut, brain). As an outcome, survival is multifactorial and predicated on the diversity of the underlying cause and comorbidities, in addition to the delivered therapy. As such, outcome assessment by survival alone may be disassociated from recovery of renal function or adequate delivery of dialysis.^34,149 Consequently, more sensitive and predictive outcome measures should be considered for assessment of dialysis adequacy, including recovery of renal function, improvement of the systemic manifestations of uremia, and reduction of complications attending uremia or its therapy.¹⁹

Survival is the optimal outcome for animals with end-stage renal disease because there is no prospect for recovery of renal function. Realistic outcomes for these animals that are treated with hemodialysis vary depending on age, chronicity of the disease, comorbidities, and residual renal function. Appropriate markers for dialysis adequacy include length of survival, owner perceived quality of life (e.g., activity, social interaction, appetite), elimination of uremic symptomatology (hypertension, hyperphosphatemia, anemia), nutritional adequacy, and elimination of dialysis-associated complications.

For both acute and chronic dialysis, survival is a difficult outcome parameter to correlate specifically to dialytic interventions. Yet, despite these constraints, the kinetically modeled dose of dialysis (Kt/V) has been shown to correlate independently with survival as an outcome in humans undergoing maintenance hemodialysis,^{73,106,120,124} and it is likely to be linked similarly to the success of dialysis in animals. The empirical use of proven standards of dialysis adequacy and clinical experience in human patients are useful first approximations for appropriate veterinary standards of dialysis adequacy until evidence-based standards are determined in animals.^34,59

Quantification of hemodialysis delivery and urea kinetic modeling

The delivery (dose) and efficacy of hemodialysis can be expressed in a variety of ways with differing degrees of complexity and utility. Predialysis and immediate postdialysis concentrations of routine serum chemistries (e.g., urea nitrogen, creatinine, phosphorus, bicarbonate, electrolytes) are the simplest expression of efficacy and can be applied similarly to their use in conventional therapy (Figures 29-1 and 29-2).^{48,105,120,172,183} Although useful to document the instantaneous outcome of the treatment, these assessments do not facilitate the uniform prescription of dialysis to animals of differing size or metabolic status. Nor do they help to clarify the excretory impact of the therapy beyond the intradialytic interval. The predialysis and postdialysis concentrations of plasma urea and creatinine can be expressed further as reduction ratios (URR and CrRR, respectively), which are used routinely to evaluate the intensity of therapy (Appendix, Equations 2 and 3).* Urea reduction ratio (URR) can be expressed either as the fractional or percent in change in urea during the treatment and is the most universally used predictor of adequacy for dialysis sessions in animals (Figures 29-3 and 29-4; Table 29-2). The average dialysis treatment in cats and most dogs will achieve a URR approaching 95%. This high level of treatment intensity is due to a combination of relative long treatment time and relatively smaller patient size (5 to 40 kg) compared with humans in which a URR target is 60% to 65%. In very large animals (50 to 70 kg), this degree of treatment intensity is often difficult to obtain, and a URR of 80% to 85% is typical.

Figure 29-2 Changes in BUN during and after 5 hours of hemodialysis treatments in a 33-kg dog presented for acute antifreeze poisoning at varying degrees of residual urea clearance. The predialysis and immediate postdialysis BUN concentrations reflect a simple assessment of treatment intensity (dose). The eKt/V (approximately 2.9 per session) for the dialysis treatments was identical for each level of urea clearance, yet the rate of increase and the equilibrated BUN concentration after stopping dialysis increased inversely with residual urea clearance.

Figure 29-3 Prediction of dialysis treatment intensity (urea reduction ratio [URR]) as a function of the volume of blood processed in 72 dogs undergoing hemodialysis. URR was computed from predialysis and postdialysis BUN concentration (Appendix, Equation 2). The volume of blood processed (Q_b × T_d) was indexed to body weight to compare dogs of different sizes. The relation (modeled as URR = 1−e^{−a(Q_b × T_d/BW)}, r² = 0.69) is displayed as a thick solid line with its 95% confidence interval (CI; thin lines). To achieve a low-efficiency treatment with URR equal to 30%, a volume of 0.3 L of blood/kg of body weight must be processed during the treatment (e.g., 6 L in a 20-kg dog). The variation in resulting URR (95% CI, 15% to 45%) underscores the necessity for close monitoring of the delivered (and not prescribed) dose of dialysis for each treatment. Similarly, a URR of 80% is obtained with 1.4 L (95% CI, 0.9 to 2.9) of blood processed per kilogram of body weight (e.g., 28 L in a 20-kg dog).

Figure 29-4 Prediction of dialysis treatment intensity (urea reduction ratio [URR]) as a function of the volume of blood processed in 66 cats undergoing hemodialysis. Other conventions are as described for Figure 29-4. The closer correlation (r² = 0.85) between volume of blood processed and URR in cats compared with dogs is probably because of the more uniform body shape and sizes in this species.

Table 29-1 Recommended Extracorporeal Volumes and Characteristics of Hemodialyzers Used for Hemodialysis in Dogs and Cats

Reduction ratios are convenient for clinical assessment but do not account for all aspects of solute transfer. Uremic toxicity and patient well-being are not predicted necessarily by the highest or lowest concentration or the intermittent change of specific retained uremia solutes.⁶⁵ The integrated exposure to uremia toxins over time is considered by some a more realistic determinant of well-being and therapeutic adequacy.^{63,104,106,115,143} For urea, this is expressed as the time-averaged concentration (TAC_urea), which is calculated as the area under the BUN profile (curve) divided by the duration of the dialysis cycle (Figure 29-1; Appendix, Equation 1). TAC_urea has been shown to predict morbidity and outcome in human patients undergoing hemodialysis and provides an integrated overview of urea dynamics (and presumably uremia toxicity) during a single or over multiple dialysis cycles. It has been highly predictive of dialysis adequacy and outcome for survival but remains nonspecific and fails to distinguish the multifactorial contributions to urea metabolism during the dialysis cycle, including dialysis dose, urea generation, nutritional adequacy, residual clearance, and distribution volume.^{48,98,104,107,120}

At face value, neither predialysis BUN nor TAC_urea are adequate surrogates to characterize the adequacy of dialytic therapy or urea metabolism. An animal with a low predialysis BUN can represent effective dialysis (high dialysis delivery), recovering renal function (increased residual renal clearance), inadequate nutrition (low urea generation rate or PCR), or volume overload (expanded urea distribution volume). Conversely, under dialysis, worsening renal function, high catabolic rate, or volume contraction can all be reflected by a high predialysis BUN.

The dose of dialysis delivered to the patient can be defined alternatively by the amount of clearance provided by the hemodialyzer during the dialysis session. Using the measured (instantaneous) clearance of the dialyzer for urea (K_d, mL/min) and the dialysis session length (T_d, minutes), the dose of dialysis can be defined as K_d × T_d or the volume of the patient cleared of urea (depurated volume) during the treatment (mL). This value can be indexed further to the total reservoir or distribution volume of urea in the patient (V, mL) to compare treatment efficacy among patients of different body sizes as V is equal to the patient’s total body water. This expression is analogous to conventional dosing of drugs as milligrams per kilogram of body weight. The value obtained with this kinetic expression, Kt/V, (Appendix, Equation 11) is unitless and represents the fractional clearance of the urea distribution volume.^{48,50,64,157,172} Kt/V has become the international reference for dialysis dosing and delivery.⁷⁶

This assessment of dialysis dose and intensity advances our understanding of the delivery of dialysis during individual treatments but requires the additional measurement of K_d (Appendix, Equation 5) and the imprecise estimation of V from the patients weight and hydration status. These predictions of dialysis dose are limited by simplifying assumptions regarding urea generation, fluid removal, and solute transference during the session, which require more extensive evaluation. A more fundamental understanding and precise description of solute dynamics during dialysis can be derived from kinetic modeling of the intradialytic and interdialytic changes in BUN similar to pharmacokinetic profiles used to describe drug metabolism.^48,64,141 Urea kinetic modeling (UKM) is fundamental to understanding the prescription, monitoring, and quality assurance of hemodialysis procedures and must be familiar to all practitioners of this therapeutic modality. It dissects the mutually independent influences of dialysis, residual renal function, nutrition, catabolism, and distribution volume on the intermittent perturbations in urea concentration during and between the dialysis sessions. This kinetic approach to urea metabolism also yields the fractional clearance of urea (Kt/V) as a measure of dose in addition to urea generation rate (G), protein catabolic rate (PCR), and the distribution volume of urea (V) that are ionic dialysance otherwise beyond clinical assessment.

The simplest kinetic assessment of urea during intermittent hemodialysis is represented by a single-pool, fixed-volume model, in which the entire volume of distribution of urea (i.e., total body water) is presumed to behave as a single pool with no change in volume or urea input during the treatments (see Figure 29-1). ^{47,48,141-143} In this simplified model, the only kinetic variable is total urea clearance (K), which represents the sum of residual renal clearance (K_r) and the clearance of the dialyzer (Kd) (see Figure 29-1; Appendix, Equations 5 and 6).¹⁸⁶ The absolute removal of urea in this system will be reflected by the change in urea concentration at any time during dialysis such that:

(1)

where C_t is the urea concentration at time = t; C₀ is the predialysis urea concentration at t = 0; K is the total urea clearance; and V is the volume of urea distribution. Rearrangement of Equation 1 provides Equation 2 for single-pool (sp) conditions,

(2)

Equation 2 is the fundamental kinetic expression for the fractional clearance of urea (dialysis dose) during a single dialysis session. In the simplified single-pool model, the kinetic prediction of dialysis dose can be derived very simply from the measured predialysis and postdialysis BUN concentrations. It must be emphasized, that this expression represents a gross oversimplification of the events and kinetic variables during therapeutic hemodialysis and should be used only to provide a rough estimate of the dialysis dose.

During a therapeutic dialysis session, the relationships between G, V, and K (illustrated in Figure 29-1) are more complex, highly interdependent, and cannot be described mathematically by a single simple relationship. Mathematical description of each variable, however, can be defined in terms of the other two with formal urea kinetic modeling (Appendix, Equations 8 through 10). When one of the variables (G, V, or K) is known, the others can be resolved by simultaneous iterative solution of the equations to yield a unique solution for the unknowns when residual renal clearance (K_r), instantaneous dialyzer clearance (K_d), ultrafiltration volume, and the measured changes in BUN during and after the treatment are known.^{47,48,141-143} These computations are performed easily with commercially available software or can be programmed into routine spreadsheet applications.

The simplified single-pool, fixed-volume model presumes conditions not generally valid in therapeutic dialysis sessions and loses accuracy if total body water (TBW) changes during or between treatments. The model also loses accuracy during high-efficiency treatments of short duration, when the urea distribution does not behave as a single homogenous compartment. Delayed diffusion from the intracellular compartment or variations in diffusion among discrete fluid compartments (e.g., skin, muscle, gut) with different perfusion and transference characteristics creates a solute disequilibrium between compartments that promotes a postdialysis rebound of urea that is not predicted by immediate postdialysis blood sampling.^{47,55,126,144} Deviations in the assumptions for single-pool, fixed-volume kinetics can be minimized by measurement of the postdialysis urea at 45 to 60 minutes after the end of the dialysis treatment rather than immediately postdialysis. By this time, intercompartmental shifts (or rebound) have reestablished solute equilibrium, and the plasma concentration reflects the equilibrated concentration of urea across all body compartments.^47,151,163As stated previously, therapeutic hemodialysis deviates considerably from the single-compartment model illustrated in Figure 29-1. Retained solutes, including urea, can be distributed in multiple compartments, which are partially secluded from the dialyzer by delayed transfer or differences in regional perfusion. Most dialysis treatments also require ultrafiltration, and urea generation proceeds throughout the session which further deviate the serum urea concentration from single-pool predictions. These deviations from single-pool, fixed volume assumptions can be improved to provide greater accuracy to urea kinetic analysis by using more mathematically complex double-pool¹⁴² or noncompartmental kinetic modeling methods (Figure 29-5). The double-pool variable volume kinetic model accounts for intercompartmental solute diffusion during and after completion of hemodialysis, and _dpKt/V is regarded as the standard for dialysis dose. Optionally, correction algorithms that account for these compartmental deviations have been applied to single-pool assessments using additional blood sampling and appropriate software in human patients.^41,48,69 These correction formulas minimize many of the limitations of single-pool estimates but have not been validated in animals. More accurate predictions of dialysis dose also can be obtained using single-pool kinetic calculations by incorporating an equilibrated BUN obtained 45 to 60 minutes after cessation of the treatment as the end-dialysis value. Use of the equilibrated BUN in the single-pool calculations yields _eKt/V as a measured dialysis dose that closely approximates the _dpKt/V and better reflects whole patient clearance. Both the _eKt/V and the _dpKt/V assessments of dialysis dose will be lower than dose predicted as the _spKt/V.

Figure 29-5 Graphic illustration of the double-pool variable-volume kinetic model of the urea metabolism during high efficiency hemodialysis. In this model, the urea generation rate (G), the renal clearance (Kr), and the dialyzer clearance (Kd) are the major determinants of urea content in the central compartment (volume V1). An additional peripheral compartment (volume V2) continuously exchanges solutes and water with the central pool. The bidirectional rate constant for urea transference between the two pools is indicated by Kc. When Kc = 1, urea diffuses freely between the compartments and the system reverts to a single-pool model. A lower Kc implies a slower diffusional component into and out of the peripheral compartment. If the peripheral compartment remains unaccounted for, single-pool kinetic modeling results in a lower apparent V, a more rapid decrease of the urea concentration in the central pool, a greater postdialysis rebound, and overestimation of the dose of dialysis, Kt/V. Anatomically, the two compartments can represent the extracellular and intracellular spaces, respectively, or body areas with different perfusion characteristics.

Online measurement of these kinetic determinants of dialyzer performance and dialysis dose can be computed in real-time with ionic dialysance techniques that advance the monitoring of individual dialysis sessions and ensure adequate dialysis delivery (Figure 29-6). Automated, bloodless kinetic modeling systems using ionic clearance are available on many modern delivery systems and provide kinetic assessments for each dialysis treatment as an alternative to blood-based modeling techniques.^{24,67,108,114,118} Dialysance of a dialyzer is a measure of solute mass transfer across the dialysis membrane when the solute is present in both the blood and dialysate. Ionic dialysance is a kinetic assessment of the transfer characteristics of the ionic solutes in the blood and dialysate. The collective concentration ionic solutes in solution can be measured by the conductivity of the solution, which is proportional to the electric current conducted through the solution. The conductivity of both plasma and the dialysate is influenced primarily by the concentration of sodium and chloride and will change with perturbations of these solutes.^67,114 The clearance of a solute by the dialyzer is equal to its dialysance when the solute is present only in the blood and is absent in the dialysate. The collective dialysance of small-molecular-weight ions (such as sodium) is considered equivalent to the dialysance of urea, and consequently ionic dialysance can be used as a reasonable surrogate for the dialysance of urea. In conventional single-pass hemodialysis, circuits in which the dialysate contains no urea, urea dialysance becomes equal to urea clearance, and ionic dialysance becomes an acceptable predictor of the urea clearance of the dialyzer, K_d-urea. Analogous to measurement of blood-based dialyzer clearance, ionic dialysance is computed from measurements of dialysate conductivity (concentration of ionic solutes) at the inlet and outlet ports of the dialyzer in response to transient changes in inlet dialysate conductivity and the instantaneous dialysate and blood flow rates.*

Figure 29-6 Screen shots of the ionic dialysance display of the Gambro Phoenix illustrating the ionic dialysance (solid line, left axis) and blood flow (dashed line, right axis) throughout a dialysis session. A, Demonstrates constant dialyzer performance and extraction ratio during the treatment with a K_d-ionic of approximately 195 mL/min at a Q_b of 300 mL/min (extraction ratio, 0.65). B, Illustrates a marked and progressive decrease in K_d-ionic after 1.5 hours of treatment associated with extensive clotting of the dialyzer necessitating termination of the treatment.

When ionic dialysance is programmed sequentially during the dialysis treatment, serial updates of the instantaneous clearance (K_d-ionic) of the dialyzer can be monitored, and the depurated volume for treatment (K_d-ionic × t) is predicted at the end of the session. The _ionicKt/V, as a surrogate for _spKt/V, provided when the ionic dialysance is indexed to urea distribution volume, V. The availability and simplicity of ionic dialysance to predict dialysis delivery at every treatment should promote a better understanding of the kinetics of dialytic therapy and the efficacy of dialysis prescriptions.

Sudden or progressive decreases of K_d-ionic during the treatment can alert possible clotting in the dialyzer or development of access recirculation that may compromise the adequacy of the treatment. It is also possible to make interim projections of the _ionicKt/V for the session to ensure the treatment targets will be met by the end of the scheduled session time. If therapeutic targets will not be met under current circumstances, adjustments to treatment time, blood flow, and dialysate flow, access repositioning, or dialyzer exchange can be initiated to modify the forecast treatment to ensure adequacy.²⁹

Routine animal hemodialysis is provided intermittently three times weekly based on human convention. As for humans, this schedule represents a compromise between clinical benefits, time constraints, and financial burden. However, recent experience in human patients with daily dialysis schedules has demonstrated marked theoretical and clinical benefits to the increased dialysis frequency.^† Because diffusion is a first order process, dialysis becomes more efficient as the frequency of dialysis increases.^28,46,66,99 Critical analysis of varying dialysis schedules has shown the total weekly dose calculated as the sum of the individual treatments is not equivalent among dialysis schedules with differing frequencies. Daily treatment schedules have equivalent clinical outcomes to traditional three times a week hemodialysis schedules even when delivered at a lower total weekly dose. For example, six treatments per week at a _spKt/V of 1.0 per treatment are more efficient than three conventional treatments per week with a _spKt/V of 2.0 per treatment. To reconcile these differences, the concept of standard Kt/V (_stdKt/V) has been proposed to compensate for the differences in efficiency when comparing schedules with different intermittence.^50,64-6699 Standard Kt/V is an hypothetical continuous urea clearance that would achieve a constant blood urea concentration identical to the average predialysis urea concentration for all intermittent treatments provided during the week. This theoretical concept allows comparisons among dialysis schedules with differing dialysis times and intervals, including the extreme case of continuous therapy.

A dialysis schedule with three 4-hour treatments per week with a _spKt/V of 2.0 per treatment is equivalent to a _stdKt/V of 2.7. Increasing the schedule to six 2-hour treatments per week (_spKt/V, 1.0 per treatment) with the same total 12 hours of weekly dialysis substantially increases the amount of dialysis delivered to the equivalent _stdKt/V of 3.9 (Appendix, Equation 12). Stated differently, a three times a week, 240-minute treatment schedule (_stdKt/V, 2.7) requiring 12 hours of treatment could be provided with equivalent efficacy by considerably shorter treatments of 70 minutes per session if provided six times weekly for a total weekly dialysis time of 7 hours. Although reduction of the individual treatment time is possible according to this analogy and for illustrative purposes, this recommendation would not be clinically prudent.^52,55,66,113 Conversely, decreasing the frequency of dialysis to two treatments per week would require extension of each treatment to almost 24 hours to achieve an equivalent _stdKt/V. These quantitative predictions illustrate the marked benefits to increased frequency of therapy and are in accordance with recent clinical observations, suggesting it is difficult to compensate for decreased frequency of therapy with longer treatment times.^47,50,172

As an alternative to _sdtKt/V for comparing the equivalency of intermittent and continuous therapies, including residual renal function, the intermittent kinetics of hemodialysis can be converted to a continuous equivalent clearance (EKR).^{25,28,50,183,186} This concept is more intuitive for most clinicians as the relative contribution of dialysis can be compared directly with residual renal function and with other intermittent or continuous dialytic therapies (Appendix, Equation 7). Total patient clearance (renal clearance, Kr, and dialyzer clearance, EKR) is expressed in the familiar term (milliliters per minute) of clearance, similar to the glomerular filtration rate, and the resulting total clearance can be used to predict the expected uremic morbidity, similar to patients with earlier stages of chronic kidney disease.

A prerequisite for the validity of most urea kinetic modeling algorithms is the presumption of steady-state urea metabolism (i.e., constant food intake (quality and quantity), constant endogenous nitrogen metabolism and catabolism, stable body weight, and a regular dialysis schedule).⁴⁸ These conditions rarely exist for most veterinary applications of hemodialysis that are prescribed for acute kidney failure; however, classic double-pool, equilibrated, and EKR analyses appear valid under these conditions in human patients if careful attention is paid to the accuracy of all input variables.^26,44,87

The rationale to scale dialysis dose to the nebulous index (V) that cannot be readily measured has kinetic justification and historical acceptance. The first order kinetics of urea removal by dialysis proceeds with an elimination constant equal to K_d/V and is a measure of the intensity of the treatment. Even though V is not measured directly, it is derived mathematically to yield the expression, Kt/V, with kinetic modeling. Recently, however, the universality of scaling dialysis dose to the urea distribution volume has been questioned in human patients as the relative distribution volume varies independently of body size, between genders, and in patients of differing body composition.⁴³ Consequently, scaling dialysis dose to V may promote under treatment in some individuals and relative overtreatment in others. The comparative significance of this issue has not been addressed in animals, but it is likely the diversity of size, species, and breed, in addition to gender, in animal patients that could impose even greater variance in the relative urea distribution volume than seen in humans.

The effect of dose of dialysis on outcome has been demonstrated in humans with end-stage chronic kidney disease in several large-scale clinical studies.* The dose of dialysis that is adequate to manage dogs and cats with either acute or chronic kidney failure needs to be established using appropriate tools for treatment quantification. However, until these parameters are established, routine application of UKM extends therapeutic insights of dialysis delivery far beyond reliance on routine chemistry tests and clearly benefits the assessment and clinical management of uremic animals. Kinetic parameters and quantitation of dialysis delivery are important tools for quality assurance of dialytic therapy in animals; however, they are not therapeutic goals per se.¹⁸⁶ The provision of a yet-to-be-defined minimal dose of dialysis is only one of the requirements of therapeutic adequacy, and management of uremia necessitates an individually tailored global approach to the animal.

Hemodialysis prescription for acute kidney injury (AKI)

The rapid accumulation of retained solutes in acute uremia intensifies expression of the clinical signs and metabolic disturbances compared with the uremia of chronic kidney disease. Hemodialysis prescriptions are prioritized to resolve hyperkalemia, profound azotemia, fluid imbalance, metabolic acidosis, and persistent nephrotoxins and to accommodate ongoing therapies (e.g., parenteral feeding). The therapeutic efficiency of hemodialysis must be applied judiciously to prevent overtreatment when the risks of dialysis disequilibrium syndrome, hypovolemia, hypotension, and bleeding are high. Consequently, dialysis goals for initial treatments in animals with AKI differ considerably from the goals and prescription for later dialysis treatments.

Dialysate Composition

Dialysate composition and its temperature and flow rate are active components of the dialysis prescription. Dialysate is formulated to maximize elimination of uremia toxins, prevent depletion of normal blood solutes, replenish depleted solutes, and minimize physiologic and metabolic perturbations during and after the dialysis sessions. Conventional dialysate formulations for dogs and cats include sodium, approximately 145 mmol/L (dogs), 150 mmol/L (cats); potassium, 0.0 to 3.0 mmol/L, bicarbonate, 25 to 40 mmol/L; chloride, approximately 113 mmol/L (dogs), approximately 117 mmol/L (cats); calcium, 1.5 mmol/L; magnesium, 1.0 mmol/L; and dextrose, 200 mg/dL, which are produced online from standard dialysate concentrates. Dialysate flow conventionally is 500 mL/min but can be decreased to reduce solute clearance during initial treatments or increased to maximize the intensity of maintenance treatments. For practical purposes, however, there is little additional solute clearance until dialysate flow exceeds twice the counter current blood flow rate.^78,158 Urea extraction across the dialyzer is nearly complete at the blood flow rates used during initial treatments, so it is not practical (or possible on most delivery systems) to reduce dialysate flow sufficiently to alter dialysis efficiency.

Rapid solute removal exposes the patient to nonphysiologic osmotic shifts that can cause osmotic disequilibrium between the vasculature, the interstitium, and cells. The accompanying shifts of fluid out of the vasculature and interstitium can cause signs of hypovolemia, hypotension, cramping, nausea, vomiting, and neurologic manifestations of dialysis disequilibrium syndrome. The patient may experience additional hypovolemia, hypotension, and poor catheter performance when ultrafiltration is superimposed on these effects. These signs are especially likely to develop early in the treatment when solute removal is the greatest. To offset these physiologic trends, the sodium composition of the dialysate can be modeled (or profiled) so that dialysate sodium is adjusted systematically during the treatment to counteract solute disequilibrium, promote vascular refilling, and lessen or prevent these adverse signs.^{22,60,129,169} Dialysate sodium can be programmed to change in stepped or linear adjustments from hypernatremic (155 to 160 mmol/L) during the initial stages of the dialysis treatment to isonatremic or hyponatremic (150 to 140 mmol/L) at the termination of the treatment. During the hypernatremic phase of the profile, the sodium gradient from dialysate to plasma causes sodium loading and expansion of intravascular volume during this critical time when the extracorporeal circuit has filled, ultrafiltration has started, and solute removal is greatest.

The efficacy of sodium profiling has not been validated in animals but appears beneficial in human patients predisposed to hypotension or intradialytic discomfort.* A modeled dialysate with a sodium concentration of 155 mmol/L for the initial 20% to 25% of the treatment, 150 mmol/L for the next 40% of the treatment, and 140 to 145 mmol/L for the remainder of the treatment has been used for small dogs that are not hypertensive and predisposed to hypovolemia.^34,59 For cats, sodium modeling using the respective sodium concentrations of 160 mmol/L, 155 mmol/L, and 145 to 150 mmol/L appears to prevent hypotension in the face of the large extracorporeal volume required for hemodialysis. The effects of sodium modeling on intravascular volume are illustrated in Figure 29-7 in which expansion (refilling) of blood volume coincides with the application of a high-to-low dialysate profile in a dog receiving concurrent ultrafiltration.

Modeling dialysate sodium from isonatremic or hyponatremic to hypernatremic (dogs: 145 mmol/L for the initial 20% to 25% of the treatment, 150 mmol/L for the next 40% of the treatment, and 155 mmol/L for the remainder of the treatment; cats: 150 mmol/L, 155 mmol/L, and 160 mmol/L, respectively) has been used prophylactically to forestall the neurologic manifestations of dialysis disequilibrium in severely azotemic animals. This sodium profile promotes osmotic (sodium) loading of the ECF in the later stages of treatment when urea disequilibrium can cause osmotic fluid shifts into the intracellular compartment, exacerbating cerebral edema and increased intracranial pressure. This profile has been derived empirically but appears to offer a margin of protection in animals with BUN concentrations greater than 200 mg/dL. Conceptually, this low-to-high dialysate profile could increase the osmolality of the ECF by 20 mOsm/kg (approximately equivalent to the osmotic effects of 60 mg/dL of urea disequilibrium), promoting an osmotic buffer to lessen fluid shifts into cells (Figure 29-8).

Figure 29-8 Hypothetical plot of the changes in serum urea nitrogen and dialysate sodium concentration during a dialysis treatment employing low-to-high sodium profiling from 150 to 160 mmol/L. The 20 mOsm/kg (NaCl) change in serum osmolality resulting from the sodium modeling could help offset, in part, the 89 mOsm/kg change in serum osmolality resulting from the dialytic change in urea of 250 mg/dL during the treatment. The osmotic buffer provided by dialysate sodium profiling is equivalent to the approximately 60 mg/dL change in blood urea nitrogen.

Sodium profiling will alter the patient’s sodium balance if the cumulative sodium transfer is other than neutral. A positive sodium balance is expected with the low-to-high profile and is accepted for initial treatments in patients at risk for dialysis disequilibrium. Patients may develop untoward complications, including postdialysis thirst, interdialysis weight gain, hyperkalemia, and hypertension if the profile consistently promotes sodium accumulation. This can be significant during maintenance hemodialysis in animals as documented in human patients.⁵¹ Routine high sodium dialysate profiles have been shown to exaggerate potassium rebound and increase interdialytic serum potassium concentrations in human dialysis patients.⁴⁵ In dogs undergoing maintenance hemodialysis, sodium profiling from low to high promoted a 50% reduction in the prevalence of hyperkalemia compared with dialysis with standard dialysate.¹²² Severe hypertension can been seen in animals in association with prolonged use of high sodium dialysate profiling, and the profile must be adjusted to produce neutral sodium balance, or an isonatric dialysate should be used if these signs are recognized.

A standard dialysate potassium concentration of 3 mmol/L can be used for most animals with acute or chronic renal failure. Essentially all potassium is sequestered in the intracellular compartment, and the excessive potassium load must transfer from this compartment. Consequently, serum potassium concentration may not be corrected adequately during short dialysis sessions if a standard dialysate is used in animals with severe hyperkalemia or during treatments using slow blood flow rates. Similarly, the extracellular potassium load in animals treated medically for severe hyperkalemia before dialysis may become sequestered in cells and not accessible for dialytic removal during short dialysis treatments.

Consistent evidence in human patients suggests large dialysis potassium gradients or rapid changes in serum potassium concentration during sessions employing a dialysate potassium <1.0 mmol/L can alter the intracellular/extracellular potassium ratio and resting cell membrane potential to increase the risk for ventricular arrhythmias and sudden cardiovascular death.^{88,103,132,134,135} Sudden intradialytic cardiovascular death is uncommon in animals perhaps due to the acute versus chronic nature of the patient populations and the relative differences in cardiovascular comorbidities between animals and humans. Nevertheless, these identified risks should prompt reconsideration of current recommendations. At a minimum, the appearance of ventricular arrhythmias during the treatment warrants changing to a dialysate containing 2 or 3 mmol/L of potassium. Until additional evidence becomes available, the future use of dialysate solutions with 0 mmol/L of potassium should be prescribed with caution.

Anticoagulation

The interaction of blood with the materials and irregularities of the dialysis membrane and extracorporeal circuit activate the coagulation cascade, promote thrombosis in the extracorporeal circuit, and necessitate routine anticoagulation of patients during the dialysis session.¹⁷⁰ In fact, it was the discovery of the anticoagulant, hirudin, that enabled the initial development of hemodialysis.¹ All triggers and components of the coagulation cascade, and activation and aggregation of platelets participate variably to induce clotting during dialysis. Active strategies to balance anticoagulation and coagulation must be employed to prevent these events for dialysis to succeed yet remain safe.

Inadequate anticoagulation promotes thrombosis of the dialyzer, causing inefficient treatments, blood loss in the extracorporeal circuit, and potential for an abrupt cessation of the treatment. Excessive anticoagulation can cause serious bleeding, although this is infrequent. Unfractionated heparin has been used as the standard anticoagulant for intermittent hemodialysis for 40 years.³⁶ Despite this experience, coagulation remains variable from animal-to-animal and treatment-to-treatment and remains problematic to control. The large extracorporeal circuit and slow blood flow rate required in severely uremic animals contributes to the anticoagulation difficulties and makes monitoring essential throughout the dialysis session. Automated activated clotting time (ACT) is used most commonly to prescribe and monitor safe heparin requirements, but other coagulation measures can be used with equal reliability. Automated ACT has proven reliable and predictive with point-of-care convenience and cost-effectiveness. Low-molecular-weight heparins are used with increased frequency in human dialysis, but there appears to be little difference in their respective efficacy during the dialysis session. The transition to low-molecular-weight heparins is directed to minimize heparin-induced complications, especially heparin-induced thrombocytopenia.¹⁷⁰ To date, there is little experience with the use of low-molecular-weight heparins in veterinary dialysis, but the relative safety of unfractionated heparin and cost have obviated a need for change.

The predisposition for clotting the dialysis circuit varies with individual characteristics of the animal in addition to its underlying disease, the choice of hemodialyzer membrane, predialysis hematocrit, extracorporeal blood flow rate, volume of the extracorporeal circuit, predialysis ACT, and rate of ultrafiltration. A standard protocol for anticoagulation during hemodialysis includes a loading dose of heparin from 10 to 25 units/kg IV (cats) and from 25 to 50 units/kg IV (dogs). The loading dose is administered 5 to 10 minutes before starting dialysis to establish an ACT in the target range of 1.5 to 1.8 times the reference ACT or approximately 150 to 180 seconds. After starting dialysis, a continuous infusion of heparin at 20 to 50 U/hr (cats) or 50 to 100 U/kg/hr (dogs) is provided to maintain the ACT in the target range. The hourly heparin dose is adjusted or intermittent boluses of heparin are administered based on sequential ACT measurements performed every 30 to 60 minutes to maintain the ACT target. The target ACT can be increased to 200 to 250 seconds if the animal demonstrates a propensity to clot the extracorporeal circuit. The loading and hourly dose is set to an ACT target of 125 to 150 seconds if there is moderate risk of bleeding.

Under some clinical circumstances, the risks of bleeding from heparin administration are too great despite the necessity to provide dialysis. These equally compelling circumstances mandate alternative anticoagulation strategies that preclude the use of heparin or systemic anticoagulation of the patient. The decision to avoid systemic anticoagulation or perform a “no heparin” treatment is determined by the animal’s relative risks for consequential bleeding. Active bleeding, recent or impending major surgery, percutaneous biopsy (within 24 to 48 hours), severe trauma, hyphema, gastric ulceration, uremic lung, and a predisposition for CNS hemorrhage represent contraindications for systemic heparinization and candidate conditions for “no heparin” hemodialysis.

There are four common alternatives for “no heparin” hemodialysis. In many circumstances, a “no heparin” protocol really means drastically reduced heparin or regional heparin rather than a complete absence of heparin use. In cases where there is a severe, life-threatening potential for bleeding (i.e., CNS hemorrhage, uremic lung), the risks for any degree of anticoagulation may be too extreme for any use of heparin.

Reduced Heparin

Most treatments are performed with some heparin delivery to the extracorporeal circuit. If the predialysis ACT is already increased or within the standard target range, the heparin prime usually can be eliminated. In most animals (even those with increased ACT measurements) some heparin is delivered to the extracorporeal circuit to prevent overt clotting (especially in treatments employing slow blood flow rates). This may be on the order of 10 U/hr (cats) or 10 U/kg/hr (dogs) up to 100 or 200 U/hr in small and large dogs, respectively. Clotting occurs preferentially in the venous header of the dialyzer followed, in order of frequency, in the venous drip chamber, the fiber bundle, the arterial header of the dialyzer, and in the arterial drip chamber. The drip chambers are particularly prone to clotting if the blood flow rate is <20 mL/min as there is little stirring at these flows, and the chamber volume remains static. Under these conditions, the heparin infusion can be split and delivered directly into both the arterial and venous chambers. It also is helpful during slow treatments to pause the treatment every 20 to 30 minutes by placing the system in bypass and increasing the blood flow to dislodge accumulating thrombin aggregates and clots.

No Heparin

Truly no heparin treatments can be performed, but they demand considerable attention. The extracorporeal circuit usually is recirculated with heparin during the setup to promote binding of heparin to the plastic surfaces. This procedure had merit for some membranes (i.e., Hemophan), which could bind heparin covalently. It is uncertain if this procedure is rational with newer synthetic membranes, which may not bind heparin. Before starting the treatment, the excessive heparin is removed from the circuit by a saline rinse (refresh) that is at least three times the volume of the circuit. No heparin prime or maintenance dose of heparin is provided during the treatment. The risk of clotting can be minimized by maintaining the blood flow rate as high as possible, keeping the treatment time to less than 2.5 hours, and flushing the extracorporeal circuit with 25 to 50 cc of saline every 15 to 30 minutes. Saline flushing dislodges accumulating thrombin aggregates and clots, and permits visual inspection of the dialyzer for clotting. The excessive volume associated with flushing can be removed by ultrafiltration if necessary.

Regional Anticoagulation with Heparin and Protamine

This procedure introduces heparin into the arterial blood line to anticoagulate the extracorporeal circuit and reverses (antagonizes) the actions of heparin with protamine in the venous outflow before returning the blood to the patient. Regional anticoagulation avoids anticoagulating the patient while permitting effective anticoagulation of the dialysis circuit. Although conceptually attractive, it is rarely performed because of the difficulty of precisely regulating the balance of anticoagulation and antagonism.

Regional Citrate Anticoagulation

In another regional approach to anticoagulate, only the extracorporeal circuit uses the sequential administration of citrate and calcium.⁷ Trisodium citrate is infused into the arterial blood path to chelate calcium, decrease ionized calcium, and prevent activation of the coagulation cascade while blood is circulated in the extracorporeal circuit. On the venous side, calcium is reinfused to normalize ionized calcium and reestablish normal coagulation before blood is returned to the patient. Regional citrate anticoagulation is used routinely in a variety of extracorporeal therapies, including CRRT and apheresis, but is not used commonly in intermittent hemodialysis. Although this represents an attractive approach, there are a variety of predictable complications that require careful monitoring and procedural fine tuning. The balancing of citrate and calcium infusions is critical and often problematic. If the citrate infusion is inadequate, the system is predisposed to clotting. If it is excessive, the patient is predisposed to hypocalcemia and metabolic acidosis. If the calcium supplementation is inadequate or excessive, the patient develops hypocalcemia or hypercalcemia, respectively. Other possible complications include hypernatremia from the trisodium citrate infusion and metabolic alkalosis from metabolism of excessive citrate or returned calcium-citrate complexes. This may become a more standardized procedure for hemodialysis in animals as greater experience is gained with the variety of extracorporeal techniques in which it is used more routinely.

Real-time monitoring of “no heparin” treatments is critical to their success and to prevent overt clotting complications. The goals of monitoring are to adjust intradialytic procedures to minimize progressive clotting in the system and to abort the treatment before catastrophic clotting causes loss of the entire extracorporeal blood volume (Figure 29-9). The presence of clotting often can be predicted by visual inspection of the extracorporeal circuit and measurement of dialyzer performance throughout the treatment. The most subtle evidence of potential clotting is often seen in the arterial and/or venous drip chambers as sticky fibrin tags or a film on the surface of the chambers. Visible clots in the headers of the dialyzer can be recognized as dark shapes within the blood pool. Clotting in the fiber bundle also can be recognized by darkened streaks in the bundle, but this appearance will vary with different membranes. Evidence of fibrin deposition and clotting can be identified more readily by flushing the extracorporeal circuit periodically with saline to displace the blood in the circuit. Serial flushing helps clear developing clots to prevent their extension and to document worsening of the clotting.

Figure 29-9 Mean ± SD changes in fiber bundle volume of high-flux dialyzers during hemodialysis treatments in a dog using adequate heparin administration (squares) and treatments using inadequate heparin (approximately half) dosing (triangles). Progressive clotting of the dialyzers develops after 180 minutes of treatment, and can be detected and monitored by sequential changes in fiber bundle volume.

At a constant blood flow and ultrafiltration rates, progressive clotting in the fiber bundle can be identified by a progressive increase in transmembrane pressure (TMP). As the surface area of the dialyzer declines with clotting of the fiber bundle volume, a greater TMP is required to achieve the ultrafiltration goal set in the ultrafiltration controller. Some delivery systems permit adjustment of the “maximum TMP” setting to a value just above the TPM established for the treatment. If clotting occurs and the TPM increases above this preset value, a “maximum TMP alarm” will sound to alert the operator to the increasing TPM and likely clotting. This process works very well for the early detection of clotting in the fiber bundle, but the maximum TMP limit must be readjusted if the blood flow or the ultrafiltration rates are changed during the treatment.

Evidence of clotting in the fiber bundle can also be detected by decreasing dialyzer performance throughout the treatment. Dialyzer performance can be monitored by sequential bedside measurements of the urea clearance of the dialyzer (at constant blood flow) or by real-time measurement of ionic dialysance available and displayed on some dialysis delivery systems (See previous Dialysis Adequacy section). Progressive or sudden decreases in clearance suggest active clotting in the dialyzer. These monitoring techniques provide a quantitative prediction of the degree of clotting (loss of membrane surface area), which could be used to trigger contingencies to modify the anticoagulation protocol or to stop the treatment. For example, a 25% to 30% decline in clearance predicts a similar magnitude of clotting and could prompt a contingency to replacement of the hemodialyzer, increase the blood flow rate, initiate saline flushing of the extracorporeal circuit, or discontinue the treatment for fear of greater blood loss. It should be recognized that these techniques only detect clotting in the fiber bundle and could fail to detect severe clotting in the arterial or venous header of the dialyzer or clotting in the venous drip chamber, which could cause a sudden and catastrophic interruption of the treatment and loss of the entire extracorporeal blood volume. In some cases obstruction of the arterial or venous headers with cessation of all flow of blood in the extracorporeal circuit can develop without activation of the arterial or venous pressure alarms and continued rotation of the blood pump.

It is possible to accurately and sequentially measure the fiber bundle volume of the dialyzer in vivo in real-time during the course of the dialysis session using indicator dilution techniques with ultrasonic detectors.⁹⁰ The fiber bundle volume is computed from the ultrasound-detected transit time of an injected saline bolus through the dialyzer. The volume of blood in the bundle is determined from the relationship, FBV = Q_b × T, where FBV is the volume in the blood compartment of the dialyzer, Q_b is the extracorporeal blood flow rate, and T is the mean transit time through the dialyzer. A decrease in mean transit time reflects the increased velocity of blood flowing through a smaller volume. A decrease in bundle volume reflects the loss of flowing blood channels due to clotting.

Hemodialysis prescription for chronic kidney disease

Experience with long-term intermittent hemodialysis for animals with chronic kidney disease is less than for acute uremia, yet hemodialysis is clearly indicated, effective, and affords a good quality of life for animals with CKD. Many of the considerations used to prescribe acute hemodialysis are equally valid for chronic dialytic therapy. Adequacy standards for animals with CKD await future definition, but intensive hemodialysis provided every 2 to 3 days can augment its medical management. As animals are supported beyond their fated life expectancy with dialysis, the spectrum and severity of uremic signs increase. Collectively, chronic malnutrition, fluid overload, hyperkalemia, hyperparathyroidism, metabolic bone disease, refractory hypertension, progressive anemia, infection, and drug interactions and toxicities replace concerns of hypothermia, hypovolemia, and dialysis disequilibrium syndrome so prevalent in animals with acute kidney injury.

The dialysis prescription for chronic kidney disease is targeted to reduce the azotemia maximally during each session. Animals starting hemodialysis with severe uremia should be approached similarly to those with acute uremia until the predialysis BUN is less than 100 mg/dL. Thereafter, high-efficiency dialysis schedules are well tolerated. Chronic dialysis prescriptions have been derived empirically but should promote a predialysis BUN less than 70 mg/dL, a postdialysis BUN less than 10 mg/dL, and a time-averaged BUN less than 50 mg/dL. The targeted spKt/V should be greater than 2.0 per session to provide an equivalent renal clearance (EKR) of at least 10% of normal renal function. The choice of dialyzer and dialysate composition generally are similar to those for maintenance treatments in animals with acute uremia. Blood flow rate can be increased cautiously to 15 to 25 mL/kg/min or the performance limits of the vascular access, and dialysis time lengthened to 300 minutes or longer. The temptation to reduce dialysis time with opportunities to use higher efficiency dialyzers and faster blood and dialysate flow rates should be avoided. Longer treatment times may appear to have limited additional efficiency for urea removal, but many solutes, including creatinine, phosphate, potassium, and middle-molecular-weight solutes, have different kinetic profiles and are slower to dialyze or have delayed transference from cellular or sequestered compartments.^{47,50,55,63,101} Effective clearance of these solutes requires longer treatments than would be adequate for urea removal.

Three treatments per week is the traditional schedule for human patients with end-stage CKD and is used for animal patients with serum creatinine concentrations greater than 8 mg/dL. A twice-weekly dialysis schedule has been used for animals with serum creatinine concentrations between 5 mg/dL and 8 mg/dL before starting dialysis therapy, but a twice-weekly schedule likely represents the minimum recommendation that will be beneficial. Even highly efficient individual treatments performed twice weekly provide only small contributions to the weekly solute clearance required for therapeutic adequacy.^{47,55,63,64,101} There are finite limits to the efficacy of individual dialysis treatments to improve the time-averaged solute concentrations of a patient. Solute generation and rebound proceed unopposed by dialysis during the interdialytic period. These processes contribute substantially to the cumulative solute retention throughout the week and become more significant as the interdialysis interval lengthens.* The limitations of hemodialysis can only be improved with more frequent and longer dialysis schedules that impart greater efficiency to this intermittent clearance technique rather than more intensive dialysis.^† A twice-weekly dialysis schedule only will be effective if the patient has sufficient residual renal function (i.e., a continuous clearance) to offset the effects of solute accumulation in the interdialysis interval to maintain predialysis azotemia and TAC within therapeutic guidelines (see Figure 29-2).

Chronic maintenance hemodialysis is an indefinite therapeutic commitment, and efforts must be taken to prevent long-term complications that are not as evident during shorter-term treatments. Maintenance of the vascular access is paramount, and rigorous attention must be paid to ensure that minor infections are resolved, and the catheter is protected from physical damage or movement within the subcutaneous tunnel. Animals supported with chronic hemodialysis still must be given standard medical therapy to manage the nutritional deficiencies, anemia, mineral disturbances, acidosis, and hypertension associated with end-stage CKD.^59,131 Prolonged survival unmasks features of CKD rarely identified in animals managed only with medical therapy. Malnutrition, hyperkalemia, fluid retention, renal osteodystrophy, hypercalcemia, and refractory hypertension become consistent clinical features and therapeutic challenges.

Support for renal transplantation

Renal transplantation is a management option for both dogs and cats with renal failure when other options for treatment are exhausted and there is no likelihood for recovery of renal function.^2,4,17 Hemodialysis frequently is used as a bridge to renal transplantation to resolve the uremia and metabolic disturbances contributing to the risks of anesthesia and surgery. Hemodialysis expands the pool of animals acceptable for renal transplantation that otherwise would be considered unsuitable and unlikely to survive because of the severity of their uremia.^34,35 Finite periods of dialytic support may be used for animals with acute kidney injury in which transplantation provides the most favorable long-term or most cost-effective outcome. The hemodialysis prescription for animals awaiting renal transplantation is predicated on the severity of the uremia and attendant signs as described for acute and chronic kidney disease, but the course of dialysis should be as short as possible to minimize development of complications that would jeopardize the success or opportunity for transplantation. Any dialysis-associated infection could delay indefinitely or preclude transplantation and must be avoided. Repeated administration of blood products may sensitize the recipient, making it incompatible with a potential donor. After transplantation, hemodialysis frequently is used to manage acute uremia precipitated by delayed graft function, surgical complications, acute rejection, or pyelonephritis.

Use of hemodialysis to correct disorders of fluid balance

Animals with oliguric or anuric AKI have too little excretory function to eliminate administered fluids and become subject to life-threatening fluid accumulation.³⁵ Similarly, polyuric animals with severe CKD accumulate orally administered fluids associated with tube feeding and parenteral fluids used to supplement hydration or to manage episodes of decompensation. Hypervolemia and circulatory overload develop under both circumstances as expressed by chemosis, pleural effusion, peripheral or pulmonary edema, congestive heart failure, and hypertension. Once established, overhydration may not resolve with cessation of fluid delivery or diuretic administration, leaving no medical therapies to manage these disorders. Restoration of fluid balance is an important indication for hemodialysis and a consistent component of the dialysis prescription.

During hemodialysis, fluid can be extracted from the patient by the process of ultrafiltration. The volume and rate of fluid removal must be prescribed for each dialysis session based on the estimated volume excess and deviation from the animal’s ideal dry body weight. Ideal dry body weight is a progressively derived value determined as the body weight at which additional fluid removal would produce hypotension or signs of hypovolemia.^83,84 Ideal dry weight usually is predicted from recent historical weight measurements before the onset of illness, or it is estimated from the postdialysis weight when blood pressure was controlled or there was no demonstrated fluid accumulation. Ideal dry weight should not be considered a static parameter but should be redefined regularly to compensate for ongoing changes in the animal’s lean body mass and body fat. Failure to update the targeted ideal dry weight can trigger a prescription for excessive or inadequate ultrafiltration, leading to hypovolemia or progressive overhydration, respectively, as the patient gains or loses nonfluid mass.⁸³ Progressive deviation from dry weight also can be recognized by routine assessment of body condition.^{33,34,112,117} The determination of dry weight can be elusive on the basis of clinical parameters alone and often is facilitated by more objective techniques including blood volume assessment and bioimpedance spectroscopy.^187,188

The rate and volume of ultrafiltration achieved is contingent on the hemodynamic stability of the animal. All available hemodialyzers have sufficient ultrafiltration performance to remove fluid from the vascular space faster than its rate of redistribution (refill) from the interstitium and intracellular compartments. This imbalance can promote hypovolemia, hypotension, and circulatory collapse if ultrafiltration is not prescribed and monitored carefully. The process of ultrafiltration is precisely regulated by the dialysis machine, but small errors or deviations in the tolerance of these systems can cause unscheduled volume losses in small animals during the course of a dialysis session. Slow rates of ultrafiltration between 5 and 10 mL/kg/hr generally are tolerated by dogs and cats, but faster rates must be prescribed cautiously and adjusted according to the animal’s vital signs and blood pressure or by use of fluid monitoring equipment (e.g., in-line blood volume monitor, venous oxygen saturation, continuous weight, bioimpedance spectroscopy).* In-line blood volume monitors are especially useful to assess the efficacy and the safety of ultrafiltration (Figure 29-10).^167,168

A lack of change in blood volume during ultrafiltration indicates the rate of fluid removal from the vasculature is precisely matched by a fluid transfer from the extravascular fluid load. If blood volume does not decrease after starting ultrafiltration, a faster rate of fluid removal could be attempted to increase the efficiency of fluid removal. As the vascular refill rate lags behind the ultrafiltration rate, the relative change in blood volume becomes negative in proportion to the deficit in vascular refilling. The change in blood volume stabilizes when the forces for vascular refilling match ultrafiltration. Moderate fluid loads can be removed at a steady 5% to 8% decrease in relative blood volume without overt clinical consequences. More intensive ultrafiltration at a stable 10% to 12% decrease in blood volume is tolerated by some animals with readily transferable fluid loads, but greater decreases in blood volume are likely to lead to clinically evident hypovolemia. The rate of change in blood volume during ultrafiltration helps predict the animal’s ability to surrender the fluid burden and attain dry weight.⁸⁴ Steep changes in relative blood volume at greater than 10% per hour (especially at the initiation of the treatment) forecast an excessive ultrafiltration rate that is unlikely to plateau at a safe level (see Figure 29-10). If ultrafiltration is stopped transiently, a rapid positively directed change in blood volume indicates the fluid load has not been corrected completely, whereas no change suggests the animal is at dry weight. A positive change in blood volume may be seen when the dialysate sodium is greater than the animal’s serum sodium concentration causing a shift of fluid into the animal or after administration of intravenous or oral fluids or mannitol (see Figure 29-7).

Figure 29-10 Change in hematocrit (HCT, A), relative blood volume (ΔBV%, B), and venous oxygen saturation (Sat%, C) assessed by an in-line monitor in a dog with acute uremia during hemodialysis and continuous ultrafiltration. The figure illustrates the decreases in relative blood volume and venous oxygen saturation associated with hypovolemia induced by ultrafiltration. The late increase in oxygen saturation reflects the supplemental administration of oxygen (arrow).

Figure 29-7 Relative percent changes in blood volume (ΔBV%) assessed by an in-line blood volume monitor in response to stepped sodium profiling and simultaneous ultrafiltration during hemodialysis in an uremic dog. The concentration and duration of each dialysate sodium step are indicated by the bars at the top of the figure. The sodium profiling supports a positive blood volume during the simultaneous ultrafiltration.

Animals often tolerate ultrafiltration better at the beginning of the treatment than at the end, and the rate of fluid removal can be profiled to achieve greater fluid losses at the beginning and scaled back later in the session to achieve the same treatment goal. Sodium profiling can be used to offset the hypovolemic and hypotensive effects of aggressive ultrafiltration to maximize fluid removal. Sodium loading during the hypernatremic stages of the modeling profile expands intravascular volume and facilitates redistribution of fluid from the interstitium and intracellular compartments (see Figure 29-7). The administration of small doses of 6% hydroxyethyl starch (hetastarch; at 1 to 2 mL/kg) can facilitate achievement of the ultrafiltration target by maintaining intravascular volume, supporting vascular refilling, and preventing hypotension. The net volume of fluid subsequently removed will far exceed the volume administered and improve the efficiency of the ultrafiltration prescription. Progressive hypovolemia from excessive ultrafiltration is detectable with in-line blood volume monitors well before development of hemodynamic signs, permitting adjustment of the ultrafiltration rate to avert hemodynamic complications. Changes in blood pressure and heart rate are rarely sensitive or early predictors of hypovolemia under these conditions.

Venous oxygen saturation also is a sensitive indicator of hemodynamic stability. Sudden or progressive decreases reflect directional decreases in cardiac output secondary to hypovolemia and can foreshadow impending hypotensive events. Venous oxygen saturation can be measured continuously with an in-line hematocrit monitor or observed visibly as darkening (desaturation) of blood in the extracorporeal circuit (Figure 29-10).¹⁶⁷ Any decrease in venous oxygen saturation should prompt immediate assessment of the patient and possible adjustment to the ultrafiltration goals.

Ultrafiltration and diffusive solute removal are independent processes controlled by separate functions of the delivery system. Animals with life-threatening fluid overload and severe azotemia are at risk for excessive solute removal and dialysis disequilibrium syndrome if the treatment is protracted to resolve the overhydration. Conversely, they remain at increased cardiopulmonary risk if the overhydration is not corrected during low intensity treatments. Both of these contrasting dialysis requirements and risks can be managed safely by prescribing periods of ultrafiltration without hemodialysis throughout the treatment or by scheduling independent periods of ultrafiltration before or after the azotemia has been treated to an appropriate URR. During ultrafiltration without dialysis, the machine is placed in bypass mode to stop dialysate flow to the dialyzer (and diffusive solute removal), while blood flow and transmembrane pressure gradients are maintained to continue ultrafiltration. This technique permits slower and more complete fluid removal without producing unsafe rates of diffusive hemodialysis. Isolated ultrafiltration can be used in nonuremic patients to treat fluid congestion associated with heart failure and pulmonary edema refractory to diuretics.* Resolution of the fluid burden from patients with congestive heart failure may improve hemodynamic function, clinical well-being, pulmonary function, drug dependency, and exercise capacity.^5,111,154 Similar indications exist in animals, and this aspect of extracorporeal therapy should be evaluated further. Ultrafiltration requirements for individual treatments can be increased to offset administered loads of blood products, drugs, and alimentation solutions. Ultrafiltration becomes especially important in oliguric animals with no excretory capacity and no tolerance for additional volume. The volume of essential fluid-containing therapies should be balanced by equivalent or proportional fluid removal during the dialysis session to balance the anticipated fluid input. Net fluid balance at the end of the dialysis treatment is the difference between the delivered ultrafiltered volume and the volume of the priming solution administered at the beginning of the treatment and the amount of rinse-back fluid used to return blood to the animal. Air can be used as a rinse-back medium to displace the extracorporeal blood rather than fluid to maximize net fluid removal.

Ultrafiltration contributes marginally to total solute removal during the treatment by convective transfer. Convective solute removal does not change the plasma concentration of solutes as occurs with dialysis because the transfer occurs with plasma water at the existing concentration. Dialysis dose predicted by URR, simple urea kinetic models, and measurement of postdialysis serum urea concentrations will underestimate true dialysis dose because of a failure to account for the convective contributions.⁴⁸

Use of hemodialysis to correct electrolyte imbalances

Uremic animals experience a wide spectrum of electrolyte imbalances because the kidneys are responsible for homeostatic regulation of body electrolytes. Hyperkalemia is the most common and life-threatening electrolyte imbalance encountered in animals with either acute or chronic uremia and can cause severe cardiovascular instability and death. The toxicity of potassium is intensified by acidosis, hypocalcemia, and hyponatremia that may coexist with uremia. Hyperkalemia is a consistent complication of acute uremia intensifying with the severity of the azotemia and presence of oligoanuria.^35,36,95

Predialysis hyperkalemia has been recognized with increased frequency in dogs maintained on hemodialysis for greater than 2 weeks.¹²² Preliminary findings demonstrated 20 of 27 dogs (74%) undergoing dialyzed for longer than 2 weeks had episodes of predialysis hyperkalemia associated with approximately 50% of 544 hemodialysis sessions. The hyperkalemia ranged in severity between 6 and 10 mmol/L and may be associated with varying degrees of hyponatremia, hypercalcemia, and metabolic acidosis. Its prevalence is associated with the duration of dialytic support, degree of azotemia, ultrafiltration requirements, and the intensity of dialysis.¹²² Chronic hyperkalemia often is difficult to manage and poses a persistent and life-threatening risk. The causes remain unknown but likely involve dialysis-induced disruptions of cell potassium or cell volume regulation, excesses in dietary potassium load, or altered potassium regulation associated with severe chronic uremia. The use of a dialysate containing 0 mmol/L of potassium decreased the prevalence of hyperkalemia at future dialysis sessions by 50% compared with a standard dialysate containing 3.0 mmol/L potassium. Use of a standard dialysate may actually increase the prevalence for severe hyperkalemia (serum potassium >7.0 mmol/L). More recently, evidence in dogs with chronic kidney disease suggests hyperkalemia can be induced directly by the use of therapeutic renal diets commonly fed to dogs with CKD and dogs undergoing maintenance dialysis.¹⁴⁹ These observations support a role for an excessive dietary potassium load rather than causal effects of the dialysis prescription.¹⁴⁹

Life-threatening electrocardiographic abnormalities resulting from hyperkalemia may be reversed completely within minutes of initiating hemodialysis using a dialysate containing 0 mmol/L of potassium (Figure 29-11). The mechanism for the immediate effect is not known but is disassociated from improvements in serum potassium concentration, which remains unchanged at this stage of the treatment. Consequently, for dialysis sessions in which the predialysis serum potassium is greater than 6.0 mmol/L, a dialysate containing 0 mmol/L of potassium has been recommended.^34,38,59 Transfer of potassium from secluded intracellular pools may lag behind its rate of removal from the extracellular compartment by the dialyzer, causing transient hypokalemia at the end of dialysis sessions. ¹³⁶ A rebound hyperkalemia occurs following the delayed transfer from intracellular pools within hours of ending dialysis that extends to the next dialysis treatment. Daily dialysis may be required until the bulk of the potassium burden is corrected.

Figure 29-11 A, Predialysis electrocardiogram (ECG) from a uremic dog with a serum potassium concentration of 9.6 mmol/L and evident cardiotoxicity. B, ECG from the same animal within 15 minutes of starting hemodialysis with a 0 mmol/L potassium dialysate. The improvement in the ECG was independent of changes in the peripheral potassium concentration.

In contrast to medical treatments for hyperkalemia, which merely shift extracellular potassium to intracellular pools or antagonize its neuromuscular toxicity, hemodialysis eliminates excessive potassium loads from both extracellular and intracellular pools.¹³⁴ Additional guidelines for the dialytic management of hyperkalemia were discussed previously under the Hemodialysis Prescription for Acute Kidney Injury section.

The dialysate sodium concentration can be proportioned to concentrations ranging from 125 to 160 mmol/L. It also can be programmed (or profiled) to change in user-defined patterns throughout the dialysis session to achieve specific treatment goals or to correct predialysis dysnatremias. Hyponatremia caused by sodium losses from excessive vomiting, diarrhea, diuretic administration, parenteral sodium-free fluid administration, or oral water can be corrected by programming the dialysate sodium concentration to increase in stepped increments or continuous gradients to the desired postdialysis concentration. Hypernatremia caused by excessive bicarbonate or hypertonic saline administration may be difficult or inappropriate to correct with additional fluid administration but can be resolved easily by adjusting the dialysate sodium concentration in progressive or incremental steps until the desired sodium concentration is reached. The rate of correction can be regulated precisely without overcorrection. Excessive isonatremic loads of sodium can be eliminated by ultrafiltration alone without simultaneous changes in serum sodium concentration. With the exception of minor Gibbs-Donnan effects, the ultrafiltrate is formed with the same sodium concentration present in plasma water. Consequently, large sodium loads can be eliminated without perturbations in sodium concentration or the risk of inducing sodium disequilibrium, which may trigger redistribution of fluid and electrolytes from intracellular stores.^45,135

Use of hemodialysis in acute intoxications

Elimination of toxins and support for the consequences of the intoxication are important but overshadowed applications of hemodialysis.^20,79,175 This use of hemodialysis is especially important if there has been a delay in medical management, there is limited endogenous clearance of the toxin or its metabolites, or there is no specific antidote for the toxicant. Hemodialysis can be used to eliminate toxins from the body before they promote cellular damage or before they are converted to more toxic metabolites. The dialytic removal of exogenous toxins is governed by the same molecular characteristics that define dialytic clearance of endogenous toxins. Molecular size, concentration in plasma water, distribution volume, degree of protein binding, and lipid solubility significantly influence the potential for a toxin’s elimination.^14,161,185 Toxins or drugs with low-molecular-weights (<1500 Da), small volumes of distribution, and minimal protein binding are excellent candidates for diffusive and convective clearance. A small volume of distribution (<1.0 L/kg) predicts the toxin is protein-bound or restricted to the extracellular space and accessible for extracorporeal clearance. A toxin with a large distribution volume (>1.0 L/kg) is likely to be concentrated in tissues and will have minimal transference or availability in plasma water for removal. Only the free fraction of protein-bound toxins can be dialyzed readily, and toxins or drugs that are highly protein bound may not be good candidates for dialytic removal.

Ethylene glycol has a molecular weight of 62 Da, negligible protein binding, and a volume of distribution equivalent to total body water (0.5 to 0.8 L/kg) and consequently is an excellent candidate for dialytic removal. With timely dialysis, ethylene glycol can be removed from the body before its enzymatic oxidation to more toxic metabolites, including glycoaldehyde, glycolate, glyoxylate, and oxalate.^{14,20,34,137,161} Toxins that are highly bound to serum proteins, including diazepam, salicylates, nonsteroidal antiinflammatory drugs (NSAIDs), and tricyclic antidepressants, are dialyzed less effectively, but dialysis may still be a therapeutic option. Redistribution (rebound) of a toxin or drug from peripheral tissues or cellular compartments to plasma may limit the efficacy of dialysis to resolve the poisoning. If redistribution of the toxin from extravascular pools is much slower than its dialytic removal, the animal may become reintoxicated within hours after completing dialysis. For these sequestered toxins, the length and frequency of dialysis may need to be increased to facilitate their whole-body elimination.

Hemoperfusion is an adsorptive extracorporeal therapy used to manage endogenous and exogenous intoxications that are not cleared efficiently by hemodialysis. Adsorption is the principle of molecular attachment of a solute to a material surface. In contrast to the physical separation between blood and dialysate that occurs during hemodialysis, during hemoperfusion blood is exposed directly to an adsorbent with the capacity to selectively or nonselectively bind toxins of defined chemical composition within the blood path. Hemoperfusion is a small but defined niche in medical therapeutics, which should be incorporated more broadly into extracorporeal therapies in veterinary medicine. ^20,79,152 Hemoperfusion is effective at eliminating high-molecular-weight, protein-bound, or lipid-soluble toxins or drugs that are cleared poorly, if at all, by hemodialysis (i.e., diffusion and convection). Toxic indications include mushroom poisoning (amanitin toxins and phalloidin), herbicides, insecticides, overmedication, hepatic failure, and sepsis.^79,152,161 Candidate toxins include barbiturates, salicylates, antimicrobials, antidepressants, chemotherapeutics, and NSAIDs that historically have been regarded as poorly removed by either hemodialysis or hemoperfusion. Hemoperfusion represents an important extension of the extracorporeal therapies that can be provided at regional hemodialysis programs for the management of intoxications for which there are no effective or efficient therapeutic alternatives.

Typical adsorbents used for hemoperfusion contain a vast and complex network of interstices and pores of varied shape and size. Toxic solutes interact by electrostatic and hydrogen bonds within the pores in the sorbent becoming entrapped and thus cleared from the blood. Selection of the adsorbent is critical for effective and safe hemoperfusion and must meet the following general criteria: (1) high adsorptive capacity for the drug(s) or toxin(s) to be removed; (2) nontoxic and hemocompatible; (3) minimal adsorptive selectivity for normal blood constituents; (4) sterile, free of endotoxins, and noncarcinogenic; and (5) compositional stability when exposed to blood (free of leachables).

Activated charcoal has been the adsorbent used most commonly to eliminate endogenous and exogenous toxins in vivo.^27,152,161 Toxic substances are cleared according to their molecular size and affinity for the charcoal, concentration in extracellular fluid, distribution volume, degree and affinity of protein binding, and lipid solubility. Activated carbons can remove solutes with a molecular mass ranging from 60 to greater than 21,000 Da.^27,185

Activated charcoal has a robust adsorptive capacity approaching 1000 m²/g but generally is nonselective in its solute binding. It can be manufactured with pore sizes ranging from <10 Å to greater than 100 Å to control the selectivity of solute removal. The larger the pore size, the larger the molecules that can be removed. With small pores (<10 Å), there is less concern about depleting molecules such as albumin, which would jeopardize safety. To prevent the release of fine residuals that could embolize in the kidneys, spleen, or lungs, most activated charcoals used for hemoperfusion are enveloped in an ultra-thin surface coatings (albumin-cellulose, cellulose, dextran) to prevent the release of residuals and improve biocompatibility and hemocompatibility without undue compromise to their adsorbent efficiency. The surface membrane has little influence on solutes with low molecular mass such as creatinine, uric acid, hippuran, indoles, and vitamin B₁₂. For solutes with higher molecular mass (>3,500 Da), however, the surface membrane limits diffusion to the pores in the interior of the carbon. The combination of hemodialysis for small solute removal and hemoperfusion for removal of larger, protein-bound, or lipid-soluble molecules provides a broad spectrum of blood purification in animal poisonings⁶¹ (Figure 29-12).

Figure 29-12 Combined hemoperfusion (HP) and hemodialysis (HD) for the treatment of enrofloxacin overdose in a uremic cat. A neonatal extracorporeal circuit was modified to include a 50-mL Clark biocompatible HP system (Clark Research and Development Inc., Folsom, La.) activated charcoal cartridge upstream to a Cobe 100HG hemodialyzer (Gambro Renal Products, Lakewood, Colo.). This combined blood purification technique resulted in a marked decrease in the blood enrofloxacin concentration (numbers in parentheses) through the extracorporeal circuit from 17.8 to 6.7 µg/mL (Δ62%) across the HP cartridge and from 6.7 to 1.3 µg/mL (Δ81%) across the hemodialyzer and 93% reduction across both devices at 10 minutes of treatment. Combined HD/HP provided a safe and effective additional route of clearance for enrofloxacin in this cat with renal compromise.

Despite the theoretical benefits, decisions to initiate extracorporeal therapies for patients with acute intoxications remains problematic. The current availability of experienced programs is limited, and the established benefits of extracorporeal therapies for known toxins are poorly defined. Extracorporeal therapy is generally indicated if the clinical signs of intoxication are progressive or deteriorating and if the toxin can be cleared faster with the intervention than by endogenous clearance. For an intoxicant such as ethylene glycol, experience with hemodialysis is extensive, documented, and effective; treatment decisions are easily justified. Hemodialysis is the most efficient and cost-effective means to clear this toxin (and its metabolites) from the animal and to prevent the renal and extrarenal consequences associated with the intoxication. It can be recommended and justified above all other treatments. For other toxins, documentation of efficacy and outcome is limited, but the window and opportunity for possible benefit is finite and decreases hourly following exposure.

The goals for extracorporeal therapies (hemodialysis and hemoperfusion) are to eliminate the toxin and its metabolites entirely from the animal as quickly as possible and to correct the accompanying fluid, electrolyte, and acid-base disturbances, and attending uremia. For suspected poisonings amenable to extracorporeal elimination hemodialysis or hemodialysis/hemoperfusion should be initiated immediately upon diagnosis to ensure rapid elimination of the toxin regardless of previous antidotal therapy or the absence of clinical signs. If the animal needs to be transported, appropriate antidotal therapy should be administered in addition to general supportive therapies.

For some toxins (such as ethylene glycol), it generally is possible to eliminate 90% to 95% or more of the toxin with a single intensive extracorporeal treatment. For intoxications other than ethylene glycol, experience is more anecdotal and recommendations for extracorporeal blood purification must be made with less evidence. Therapeutic decisions must be balance against the historical consequences of the intoxication, the efficacy of alternative therapies, and the potential for adverse consequences of the procedure. Hemoperfusion should be considered for lipid soluble, highly protein-bound toxins with molecular mass larger than the cutoff limits of the dialysis membrane. Therapy is more efficacious when the volume of distribution is small (plasma volume or ECF volume [i.e., <0.5 L/kg]). The treatment goal for detoxification is 100% elimination of the toxin and toxic metabolites. This is often difficult to achieve when experience with a specific toxins is limited and detection assays are unavailable during the procedure. For toxins and drugs including amatoxins, fluoroquinolones, and NSAIDs where morbidity is certain and molecular characteristics are favorable, urgent decisions to initiate combined hemodialysis/hemoperfusion must be ad hoc but are generally justified because of the low morbidity of these procedures and the possibility to alter the clinical course and outcome. As evidence for outcome benefits is acquired, more definitive recommendations can be justified. A precise definition of the window of opportunity for these therapeutic interventions is necessary for realistic therapeutic recommendations. Many toxins, such as amatoxins, will have sharply delimited therapeutic window in which the toxin can be eliminated before the clinical course is solidified and intervention is unjustified.

Hemodialysis is indicated for the treatment of poisoning or drug overdose with ethylene glycol, methanol, ethanol, salicylate, lithium, phenobarbital, acetaminophen, theophylline, aminoglycosides, tricyclic antidepressants, and possibly metaldehyde. ^{14,20,79,161,185} Hemodialysis secondarily corrects the acid-base and electrolyte abnormalities and the azotemia that accompany some intoxications (e.g., ethylene glycol, salicylate). Hemodialysis should be initiated once conventional treatments are deemed ineffective and continued until the concentration of the toxin has decreased to an acceptable level and the clinical toxicity has disappeared. Dialysis treatments should be continued for prolonged periods for toxins with delayed toxicity (i.e., paraquat), low blood concentrations, or significant redistribution following treatment.

Ethylene glycol (antifreeze poisoning) is a common intoxication in companion animal practice.^{34-36,59,95,96,173} Clinical signs develop within minutes and progress variably from lethargy, nausea, vomiting, dehydration, agitation, and depression to convulsions, coma, and death. Severe metabolic acidosis and hypocalcemia are seen with significant exposure, and in later stages of the intoxication (12 to 24 hours), hypertension, cardiopulmonary failure, and acute oliguric renal failure dominate the clinical presentation. Ethylene glycol concentrations are highly variable and significantly higher in nonazotemic compared with azotemic dogs presented for antifreeze poisoning (Figure 29-13).¹³⁷ Serum ethylene glycol and glycolic acid concentrations may persist for days at toxic concentrations in azotemic or anuric animals despite therapy with alcohol or 4-methylpyrazole. These inhibitors of alcohol dehydrogenase merely delay the enzymatic conversion of ethylene glycol, and their efficacy relies on the potential for renal elimination of both the toxin and its metabolites.

Figure 29-13 A, Box and whisker plots of the serum concentrations for ethylene glycol (left) and glycolic acid (right) in azotemic (light boxes; n = 20) and nonazotemic (dark boxes; n = 6) dogs presenting for hemodialysis. B, Box and whisker plots of the change in serum ethylene glycol (left) and glycolic acid (right) concentrations before and following hemodialysis in 26 azotemic and nonazotemic dogs poisoned with antifreeze.¹²³

The goals for hemodialysis are to eliminate the ethylene glycol and its metabolites from the animal as quickly as possible and to correct the accompanying fluid, electrolyte, and acid-base disturbances and attending uremia. For suspected poisonings, hemodialysis should be initiated immediately to ensure rapid elimination of the toxin regardless of previous administration of antidotal therapy or the absence of clinical signs. If the animal needs to be transported, an initial dose of ethanol or 4-methylpyrazole should be administered, and existing dehydration and metabolic acidosis should be corrected.^38,173 It generally is possible to eliminate 90% to 95% or more of the toxin with a single intensive dialysis treatment (Figure 29-14).^34,35,137 However, the necessary amount of dialysis to deliver when toxicologic results are unavailable to confirm toxin removal during the treatment is problematic. Urea (MW, 60 Da) is similar in molecular size and distribution volume to ethylene glycol (MW, 62 Da) and can serve as an index for changes in ethylene glycol clearance similar to its surrogate role for removal of small-molecular-weight uremic toxins. The URR can be used to predict ethylene glycol reduction and the dialyzed blood volume required to achieve the removal goal (Figure 29-14).^34,137 To achieve a 90% ethylene glycol reduction during the course of treatment, it is necessary to select treatment parameters that would promote the same URR for that patient.

Figure 29-14 Box and whisker plots demonstrating the reduction ratios for ethylene glycol (light boxes), urea (stippled boxes), and glycolic acid (dark boxes) in azotemic (n = 20) and nonazotemic (n = 6) dogs. These observations demonstrate that both ethylene glycol and glycolic acid have removal kinetics similar to those for urea, and urea reduction ratio can serve as a convenient surrogate to predict removal of these toxins with hemodialysis.

For nonazotemic animals, 90% to 100% of the toxin should be removed during the first dialysis treatment. A second treatment is provided if delivery is incomplete during the first session or if there is a rebound in ethylene glycol after treatment. Vascular access with a temporary dialysis catheter generally provides adequate blood flow. The highest efficiency hemodialyzer compatible with the extracorporeal volume requirement of the animal should be used to maximize diffusive removal of the toxins. Blood flow rates between 15 and 25 mL/kg/min or faster are tolerated. A standard dialysate flow between 500 and 600 mL/min is used but can be increased if the blood flow rate is greater than 300 mL/min. A dialysate formulated with 3 or 4 mmol/L potassium, 30 to 35 mmol/L bicarbonate, and a physiologic sodium concentration is appropriate unless specific electrolyte, acid-base, or hemodynamic disorders are present. A neutral sodium phosphate additive should be added to the dialysate for nonuremic animals to prevent hypophosphatemia (see previous Dialysate Additives section). Ethanol should be added to the dialysate concentrate to achieve a dialysate ethanol concentration of approximately 0.1% in an effort to inhibit ongoing metabolism of ethylene glycol to its toxic metabolites during the extended hours of dialysis (see previous Dialysate Additives section). Ultrafiltration can be used to correct pulmonary edema or congestive heart failure secondary to the toxin or fluid administration. However, ultrafiltration is minimally effective for pulmonary effusions arising from respiratory distress syndrome or uremic pneumonitis associated with antifreeze poisoning.

In uremic animals, the goals for aggressive toxin removal are constrained by requirements to prevent dialysis disequilibrium syndrome, and dialysis must be delivered carefully to accommodate all of the patient’s needs. A temporary hemodialysis catheter generally is placed to expedite the initial treatment, but it is replaced with a permanent tunneled catheter after 2 weeks if additional dialysis is required. If the BUN concentration is less than 125 mg/dL, an intensive treatment as used in nonuremic animals is suitable. For animals with BUN concentrations greater than 150 mg/dL, the dialysis prescription should target a 90% to 100% ethylene glycol reduction, but it must be delivered with a slow, extended treatment tailored to the hourly URR targets appropriate for the degree of azotemia (see Table 29-2). For severely uremic animals, safe urea reduction and greater toxin removal is achieved when dialysis is provided over 6 to 10 hours. The remainder of the dialysis prescription should be formulated to specific complications accompanying the uremia, fluid volume status, acid-base and electrolyte disturbances, and hemodynamic stability. Ethanol can be added to the dialysate concentrate as described previously for nonazotemic animals. Mannitol (Mannitol Injection USP, Abbott Laboratories, North Chicago, Ill.) can be administered at 0.5 to 1.0 g/kg intravenously 45 to 60 minutes after starting dialysis in both mild and severely azotemic animals to prevent manifestations of dialysis disequilibrium syndrome.

Application of extracorporeal therapies should not be limited to single modalities but should be sequenced and combined to best match the clinical course and kinetics of the toxicant. Continuous versus intermittent therapies should not be considered mutually exclusive but rather complimentary. There is little justification not to include a dialytic device with a hemoperfusion cartridge when contemplating hemoperfusion. For many toxins, hemodialysis has the potential to improve toxin clearance in concert with hemoperfusion despite theoretical predictions to the contrary (Figure 29-12). Dose, blood concentration, changes in protein binding of the toxin, concurrent drugs/toxins, acid-base status, membrane type, and other variables may influence the diffusive potential of a toxin under different clinical conditions. The presence of the hemodialyzer in the extracorporeal circuit provides the potential for better thermal regulation and opportunity to correct coexisting fluid volume, electrolyte, acid-base, or uremic complications. Placement of the hemodialyzer after the hemoperfusion cartridge also helps to prevent depletion of calcium and glucose by charcoal sorbents and isolates the hemoperfusion cartridge from the ultrafiltration control system that may promote excessive fluid removal and hemoconcentration in the dialyzer should pressure increase in the sorbent bed.

Hemoperfusion with activated charcoal is generally safe but poses potential disadvantages or complications not generally experienced with hemodialysis. One of the principal concerns is the innate hemocompatibility of the adsorbent. Hemoperfusion with activated charcoal (as well as other sorbent materials) can cause thrombocytopenia and leukopenia because platelets and leukocytes become adhered to the sorbent or entrapped in fibrin films or clots formed on the charcoal. Platelets are destroyed also by surface irregularities of the charcoal bed. These effects are not unique to activated charcoal and are potential complications of polymer-based sorbents. Thrombocytopenia can be especially problematic if daily treatments are required that preclude adequate regeneration of platelets between treatments. If hemoperfusion is not combined with hemodialysis, the patient may experience significant cooling because of the duration the extracorporeal blood is exposed to room temperature. The sorbent bed also may become saturated at unpredictable times during the treatment resulting in incomplete removal of the toxin. Saturation of the sorbent is easily demonstrated by measuring the extraction ratio [(A_tox – V_tox)/A_tox] across the device or its whole blood clearance [Q_b × extraction ratio], where A_tox and V_tox are the concentrations of the solute or toxin at the inlet and outlet of the hemoperfusion cartridge, respectively, and Q_b is the blood flow rate.

Outcome and prognosis

Dogs have been supported on chronic intermittent hemodialysis as long as 1.5 years, but complications related to vascular access and anemia management often curtail dialytic support beyond 6 months. Regional availability and financial and time constraints further limit use of hemodialysis for management of either acute or chronic uremia for extended periods of time. For acute kidney injury, renal recovery is not predicated on dialysis, but rather the cause, extent of damage, comorbid diseases, multiple organ involvement, and availability of diagnostic and therapeutic services. Overall survival rates of dogs and cats treated with hemodialysis for AKI is 41% to 52%, but survival time is highly dependent on the cause.^3,59,96,149 The survival rate for AKI of infectious causes varies between 58% to 100%.^{3,59,62,123,149} Hemodynamic and metabolic causes have a reported 40% to 72% survival rate.^62,123,149 Only 20% to 40% of patients dialyzed with AKI from toxic causes (especially ethylene glycol ingestion) survive.* This average of 50% survival has been reported in most studies over the past 15 years and is a marked improvement from the 15% expected survival rate from the earliest years of dialysis or for survival of animal patients with comparable stages of AKI treated conventionally without hemodialysis.^36,39 Of the nonsurviving patients, about half die or are euthanized because of extrarenal conditions (e.g., pancreatitis, respiratory complications, DIC, and financial limitations). About a third of nonsurviving animals are euthanized due to failure of recovery of renal function within a narrow window of time that is dictated usually by economic constraints. Ongoing uremic signs, dialysis complications, and unknown causes account for the remaining patient deaths. Of surviving patients, approximately half regain normal renal function (defined by normal serum creatinine concentration) and half have persistent chronic kidney disease. A clinical scoring system for severity and outcome prediction of dogs with AKI receiving hemodialysis has been developed that may facilitate better characterization of dialysis outcomes in the future.¹⁴⁹

Studies assessing survival in animals treated with conventional therapy cannot be compared with those managed with dialysis.³⁶ There are often large disparities in the severity of the renal injury for animals managed with these respective approaches, and the window for survival is finite for animals treated conventionally and indefinite for animals treated with dialysis. Severity scoring and staging of AKI likely will provide more precise understanding of outcomes in the future.

Overall, these observations illustrate hemodialysis has a vital role in the therapeutic stratification of dogs and cats with uremia that remain nonresponsive to conventional medical therapy. Hemodialysis improves survival for animals with AKI beyond what would be expected with conventional management of the same animals. Clinical evidence and experience in human patients suggest a role for earlier intervention with renal replacement to avoid the morbidity of uremia and to promote better metabolic stability and recovery.