Drugs given IV enter the central circulation directly and generally have a rapid onset of action. The absolute bioavailability is determined by comparing the area under the serum/plasma concentration–time curve (AUC) after oral administration with that observed after IV administration. Drugs given by other routes must first pass through important organs of elimination (e.g., liver) before entering the systemic circulation; thus a smaller proportion of the drug becomes available at the site of drug action. Even drugs given IV and by inhalation must pass though the lungs before reaching arterial blood. Similar to other organs, the lungs remove substantial amounts of some agents (e.g., IV administered adenosine). For drugs administered orally, the rate and extent of gastrointestinal (GI) absorption are important considerations. Absorption can be characterized by determining the maximum attained serum or plasma concentration (C _max ), as well as the time after ingestion when the C _max was observed (T _max ). Differences in these two parameters (Cmax and Tmax) among patient groups were historically considered evidence of altered GI absorption, but in many cases the absolute bioavailability was unchanged. The bioavailability of a drug depends on the extent of metabolism during its first pass through the intestinal wall and liver before reaching the systemic circulation. When this AUC-derived measure of bioavailability was assessed, there were few drugs shown to be affected by the presence of CKD or AKI. ^,

First-pass biotransformation may also occur in the gut; bioflavonoids in grapefruit juice can inhibit cytochrome P450 (CYP) 3A4 and noncompetitively inhibit the metabolism of drugs metabolized by this enzyme. The grapefruit juice–CYP3A4 interaction was first noted with the calcium channel blocker felodipine. This interaction also increases the bioavailability of cyclosporine by as much as 20%. A wide variety of other drugs are similarly affected, including several medications used for depression and anxiety (e.g., selective serotonin reuptake inhibitors and serotonin–norepinephrine reuptake inhibitors) and statins. Herbal medicine (e.g., hypericin, one of the constituents of St. John’s wort) can activate the adenosine triphosphate–binding cassette transporter or P-glycoprotein (multidrug resistance) transporter (e.g., cyclosporine) in gut mucosa, leading to reduced drug absorption.

Although GI symptoms are common in patients with ESKF, little specific information about gastrointestinal function is available. The salivary concentration of urea increases when urea accumulates in plasma. Ammonia forms from urea in the presence of gastric urease and buffers gastric acid, increasing gastric pH. The ammonia is absorbed and converted to urea again by the liver. The gastric alkalinizing effect of this internal urea–ammonia cycle decreases the absorption of drugs that are best absorbed in an acidic environment. Drug malabsorption may be further aggravated by the increased use of various therapies to reduce gastric acidity and/or reduce phosphate absorption, especially in patients who are dialysis dependent. ^, The resultant chelation and formation of nonabsorbable complexes reduce the bioavailability of some drugs, including several antibiotics and digoxin.

The processes of GI drug absorption are complex, may be saturable and dose dependent, and are more variable in patients with ESKF than in those with normal kidney function. Gastroparesis, commonly observed in patients with diabetes mellitus, many of whom also have CKD, prolongs gastric emptying and delays drug absorption (i.e., T _max is observed to be delayed). Conversely, diarrhea decreases gut transit time (T _max is shortened and diminishes drug absorption by the small bowel). Gut mucosal integrity becomes impaired across the spectrum of CKD, as evidenced by increasing levels of circulating translocated endotoxins.

The volume of distribution (V _D ) of a drug does not necessarily correspond to a specific anatomic space. Rather, the V _D is a mathematic construct based on the plasma concentration achieved following the IV administration of a given dose of a drug. Agents that are highly protein bound and those that are water soluble tend to be restricted to the vascular compartment and extracellular fluid (ECF) space, respectively, and thus have volumes of distribution less than 0.20 L/kg. Highly lipid-soluble drugs and those extensively bound to tissues often exhibit volumes of distribution in excess of 1 L/kg. The drug distribution volume of highly water-soluble or protein-bound drugs may be increased in patients with AKI or CKD if edema and/or ascites are present ( Table 56.1 ). ^{, , ,} Drug distribution is one of the most important and complicated factors to quantify in patients with AKI. There is a fine balance between detrimental fluid overload and adequate hydration to preserve and optimize perfusion and function. Critically ill patients should be managed in a slightly negative fluid balance after initial adequate fluid resuscitation has been achieved. If patients are volume expanded, the administration of the usual doses of many drugs will result in inadequately low plasma concentrations.

The distribution volume of drugs may be altered by fluid removal during dialysis. Changes in body cell mass (nonfat, nonwater, and nonbone mineral mass) commonly occur over time in patients on dialysis, often accompanied by sarcopenia. Failure to detect a reduction in body cell mass may lead to inappropriate maintenance of the same dry weight target and drug dosage regimen, despite a real increase in total body water, and thus the distribution volume of many drugs. Finally, the method used to calculate the V _D may be influenced by impaired kidney function. There are three commonly used V _D terms: V _D of the central compartment (V _c ), V _D of the terminal phase (V _ß ), and V _D at steady state (V _ss ). The V _c for many drugs approximates ECF volume and thus may be increased or decreased by acute changes. Oliguric acute kidney failure is often accompanied by fluid overload and a resultant increased V _c for many drugs. The V _β represents the proportionality constant between plasma concentrations in the terminal elimination phase and the amount of drug remaining in the body. V _β is affected by distribution characteristics of the drug and by the terminal elimination rate constant. V _β and V _ss will often be similar in magnitude, with V _β being slightly larger.

Alterations of plasma protein binding in patients with CKD can also affect drug distribution and elimination. The V _D of a drug, the quantity of unbound drug available for action, and the degree to which the agent is eliminated by hepatic or kidney excretion are all influenced by protein binding. Drugs that are protein bound attach reversibly to albumin or α ₁ -glycoprotein in plasma ( Fig. 56.2 ). Whereas organic acids bind to a single binding site, organic bases probably have multiple sites of attachment.

Decreased protein binding in uremia.

Displacement of the drug from its binding site by an accumulation of undefined uremic toxins or a uremia-induced conformational change in the binding-site geometry results in more free drugs in the plasma.

A combination of decreased serum albumin concentration and reduction in albumin affinity for the drug reduces protein binding in dialysis-dependent patients. The latter reflects, in part, the effect of protein-bound organic acids such as hippuric acid, indoxyl sulfate, and 3-carboxy-4-methyl-5-propyl-2-furanpropionic acid, which accumulate in advanced CKD, thereby decreasing the protein binding of many acidic drugs. ^, This impaired plasma protein binding increases the unbound fraction of several acidic drugs in CKD, the magnitude of which correlates directly with the level of azotemia and may be corrected with dialysis. ^{, ,} Binding affinity is influenced by changes in the structural orientation of the albumin molecule, as well as the accumulation of endogenous inhibitors of protein binding that compete with drugs for their binding sites.

Toxicity can occur if the total plasma concentration of these drugs is pushed beyond the therapeutic range. These effects are profound for some commonly used drugs, such as seen with ceftriaxone, furosemide, metolazone, warfarin, diazepam, valproic acid, and phenytoin. Therefore when feasible, unbound plasma concentrations should be measured to guide therapy with drugs such as valproic acid and phenytoin. Predicting the clinical consequences of altered protein binding is difficult. Although decreased binding results in more unbound drugs being available at the site of drug action or toxicity, the distribution volume is increased, resulting in lower plasma concentrations after a given dose. More unbound drugs are available for metabolism and excretion, which increases the clearance and decreases the half-life of the drug in the body. Drugs with decreased protein binding in patients on dialysis are listed in Table 56.2 .

The disposition of drugs metabolized by the liver may be altered by changes in plasma protein binding. The systemic clearance of drugs with a high hepatic extraction ratio is not thought to be impacted by the effects of CKD in contrast to that of drugs with a low extraction ratio. The systemic clearance of a highly protein-bound drug with a low hepatic extraction ratio depends on the simultaneous effects of kidney function on protein binding and intrinsic metabolic drug clearance. Due to uremic toxins, protein binding of drugs will decrease in AKI and CKD, and hepatic metabolism will be inhibited. The effects of advanced CKD on protein binding with increased clearance and the effects on intrinsic metabolism with decreased drug clearance both might offset each other in terms of total systemic clearance.

Many active or toxic metabolites depend on the kidneys for their removal from the body. The accumulation of these partially active metabolites in patients with impaired kidney function (AKI and CKD) can explain, in part, the high incidence of adverse drug reactions in this patient population. For example, although the liver usually rapidly metabolizes morphine, it is categorized as excreted mainly in the urine because of the excretion of its active metabolites, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G). These metabolites readily cross the blood-brain barrier and bind to opiate receptors, exerting strong analgesic effects and potential adverse effects. In patients with CKD, morphine’s active metabolites accumulate, making prolonged narcosis and respiratory depression more likely. ^, Similarly, the biotransformation of meperidine results in the production of normeperidine. In patients with impaired kidney function, repeated doses of meperidine may result in the accumulation of this potentially toxic metabolite, with resultant seizures. Another useful example is the case of mycophenolic acid. The metabolite mycophenolic acid glucuronide accumulates in AKI and CKD patients, and though it lacks pharmacologic activity, it may be associated with gastrointestinal (GI) side effects. Table 56.3 highlights important drugs that form active or toxic metabolites in CKD and have been associated with adverse events.

The estimated eGFR ( Table 56.5 ) is used in clinical practice based on the serum creatinine and/or cystatin C (CysC) concentrations given the convenience. The measured creatinine clearance (mClCr) is the only endogenous method to determine glomerular hyperfiltration in pregnancy as cystatin C is not reliable in pregnancy.

Estimated GFR based on different current creatinine formulas may yield different CKD stage categorizations, which would impact drug dose recommendations based on original studies that used different formulas. It is not practical to repeat all the pharmacokinetics studies with standardized creatinine-determined eGFR, and therefore it is still reasonable to use drug dosing adjustments that appear in FDA- and European Medicines Agency (EMA)-approved product labeling, but clinicians should be aware of these concerns.

Traditionally, drug dosing was based on not the measured endogenous creatinine clearance (mCrCl) but on estimated creatinine clearance (eClCr) using the Cockcroft and Gault (CG) formula. ^{, , ,} The CG equation is not suitable for clinical reporting because body weight may not be available in the electronic health record and it does not apply across the spectrum of GFR. The Modification of Diet in Renal Disease (MDRD) equations ^, were initially used by clinical laboratories, although they were only validated for patients with a GFR <60 mL/min. Therefore the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation was developed to allow estimation of GFR throughout the full range of the CKD. In 2021, the CKD-EPI research group developed a new CKD-EPI eGFRcr equation without the inclusion of race, which is the currently preferred formula, although it does perform variably in diverse populations. ^{, , ,} Serum cystatin C has been proposed as an alternative marker to estimate GFR. The combined use of both serum markers, cystatin C and creatinine, permits more accurate estimation of kidney function than either of them alone.

Furthermore, the new cystatin C–based European Kidney Function Consortium EKFC equation estimates the eGFR independent from sex and race has been shown to improve GFR assessment.

Both, the 2021-CKD-EPI (either creatinine or cystatin C alone or in combination) and the EKFC equations estimate the GFR for a standard 1.73 m ² body surface area (BSA); thus for an individual patient, the BSA must be determined separately so that the GFR can be expressed in milliliters per minute (mL/min).

Adjusting drug doses based on the combined measurement of creatinine and cystatin C appears to be an effective and valid tool in the limited number of applications (mainly relating to chemotherapy and antibiotic dosing, e.g., vancomycin) for which it has been studied.

Some specialists argue that use of the CKD-EPI equations for drug dosing is not appropriate given that the pharmacokinetics studies were performed using eCrCl via the CG equation. ^, Furthermore, many studies have highlighted discordance between drug dosing recommendations based on these equations. For now it is recommended that estimation of GFR using the most appropriate formula should be an initial step in decision making rather than a conclusive step in decision making. Careful consideration of the context and clinical status is necessary.

The original equation to estimate GFR, as described by Schwartz and colleagues, is dependent on the child’s age and length:

where k is defined by age group: infant (1–52 weeks) = 0.45; child (1–13 years) = 0.55; adolescent male = 0.7; and adolescent female = 0.55. The serum creatinine level in μmol/L can be converted to mg/dL by multiplication using 0.0113 as the conversion factor. A newer version of the Schwartz equation was developed from a population of 349 children (age 1–19 years) with mild to moderate CKD enrolled in the Chronic Kidney Disease in Children (CKiD) study:

Lee and associates have reported that this new Schwartz equation performed better than the original Schwartz equation for patients with moderate CKD but was less accurate in patients with mild CKD. In pediatric patients, methods incorporating cystatin C have several advantages for evaluating kidney function. The most recent eGFR equation evaluated in pediatrics includes use of cystatin C, blood urea nitrogen (BUN), serum creatinine level (in mg/dL), and demographic data derived from over 600 pediatric patients enrolled in the CKiD study :

This equation had the highest R ² value (0.863) and the highest frequency of values within 30% of iohexol-measured GFR (91.3%) when compared with seven other GFR estimating equations.

During AKI, the GFR is a moving target so a kinetic GFR has been proposed (see later), although changes in serum creatinine are most routinely used. The mCrCl may also permit assessment of the augmented kidney clearance in the initial phase of sepsis, especially in the intensive care unit (ICU), where urine output is routinely monitored. At present, the staging of AKI is based on sequential measurement of the serum creatinine level and urine output. Because the GFR is inferred from the serum creatinine or cystatin C, all estimates of kidney function are delayed and lag the real-time GFR. Although several methods have been proposed to estimate GFR in this patient population, none have been rigorously evaluated. The latest proposed method to estimate GFR in patients with AKI is the kinetic GFR (kinetGFR), which is based on age (years), weight (kg), and serum creatinine (μmol/L) and holds true for increasing and decreasing kidney function.

This approach is based on an estimate of the creatinine production similar to the CG equation. The kinetic eGFR incorporates changing creatinine values over specified time intervals, as well as the actually measured serum creatinine values. It relates the increase in serum creatinine within a specified time interval to the maximum increase in creatinine level in one day. The kinetic eGFR tries to tackle the problem that there is always a delay between rapidly changing kidney function and measurable variables—namely, serum creatinine or urine output. The calculation of a patient’s kinetic eGFR may permit use of the eGFR-based dose-adjustment recommendations derived from patients with CKD for patients with AKI. Rigorous independent studies will be needed to confirm its validity and utility in clinical practice.

Some ESKF patients on dialysis and those with AKI on CKRT have residual kidney function that substantially contributes to better patient outcomes and the elimination of drugs and their metabolites. Measuring the elimination of iohexol after an IV dose has been reported to be an accurate and safe measure of residual kidney function in patients on dialysis and can inform drug dosing. ^, The eGFR derived from serum creatinine during continuous KRT represents the combined effect on drug clearance by residual kidney function and by KRT.

Critically ill patients frequently develop AKI ranging from 5% on admission to 50% in the ICU, as non–same-day hospitalization care is frequently complicated by AKI. In most cases, drug dosing is based on drug disposition information derived from studies in stable patients with CKD. There are large gaps in knowledge about drug metabolism and disposition in patients with AKI; thus patients may be at significant risk for underdosing and overdosing. Before the KDIGO consensus on AKI, more than 30 definitions of AKI were published in the literature. ^, AKI may occur as part of multiorgan dysfunction in critically ill patients or as isolated AKI. AKI-related, in-hospital mortality rates vary from up to 70% in ICU patients to 35% in other hospitalized patients.

The potential effects of AKI on drug dosing are of major consequence because AKI patients are often critically ill and require multiple drug therapies, some of which may be nephrotoxic or require dose modification in the setting of AKI. The pharmacokinetic changes in absorption, distribution, metabolism, and excretion are foundational to optimal patient care. ^{, , , ,} The clinician must appreciate these factors and realize that they may worsen and improve over the period of evolution or recovery of the AKI episode. Critically ill patients with AKI typically have minimal oral intake of food and liquids and commonly require parenteral administration of drugs otherwise given orally (e.g., antihypertensives and immunosuppressives).

There is a paucity of dosing algorithms to guide pharmacotherapy, derived from investigations of the pharmacokinetics and pharmacodynamics of multiple-dose studies in patients with AKI, and much variability exists. Most of the critical care literature and almost all FDA or EMA product labeling contain drug dosage recommendations derived from observations of patients with CKD and ESKF. The limited data available in the setting of AKI have predominantly been developed by clinicians; rarely is this information incorporated into official product labeling. The principles of drug dosage regimen modification described earlier for use in CKD thus remain the foundation for therapy optimization in patients with AKI. For patients on CRRT, dosing recommendations can be found in the literature. Generally, the dose in CRRT corresponds to around an eGFR of 30 mL/min, which is higher than that in intermittent HD.

The optimization of pharmacotherapy for patients receiving maintenance HD and emergent HD are both critically dependent on the availability of reliable information from well-designed pharmacokinetic studies. ^{, , ,} The impact of HD on drug dosing requirements is dependent on the drug characteristics and dialysis prescription. Drug-related factors include molecular weight (MW) or size, degree of protein binding, and distribution volume. ^, The vast majority of HD filters used before the mid-1990s were generally impermeable to drugs with an MW > 1 kDa. Dialysis membranes are now predominantly composed of semisynthetic or synthetic materials, which have larger pore sizes, and this allows the ready passage of drugs that have an MW up to 20 kDa (i.e., digoxin). For more on clearance of poisons with dialysis, see Chapter 66 .

Drug clearance during dialysis can occur by three different processes. ^, Drug removal by conventional HD occurs primarily by diffusion down a concentration gradient from the plasma to the dialysate. Removal of low-MW drugs is enhanced by increasing blood and dialysate flow rates and by using large-surface-area dialyzers. Larger molecules require more porous membranes for increased removal. The clearance of a drug by conventional HD can be estimated from the unbound fraction (f _u ) and the following relationship:

where Cl _HD is the drug’s clearance by HD, Cl _urea is the dialyzer clearance of urea, and MW _drug is the MW of the drug. The urea clearance for most conventional dialyzers varies between 150 and 200 mL/min and is comparable with values reported with high-flux hemodialyzers. With high-flux HD, the V _D and degree of protein binding of the drug become more important determinants of dialyzer clearance. The convective transport and removal of drugs during high-flux HD depends primarily on filtration pressure gradient, treatment time, blood, and dialysate flow rates. Despite the widespread adoption of high-flux HD in certain parts of the world, there are sparse quantitative data on drug clearance.

There has been little investigation about the effects of sustained low-efficiency hemodialysis (SLED) regimens on drug disposition or comparison among modalities. Only a few agents have been evaluated during the delivery of one of these dialytic variants. Slow nocturnal dialysis required a significant increase in gentamicin dosage to achieve therapeutic levels compared with conventional thrice-weekly dialysis. ^, The variability in drug clearance was high and did not correlate with small solute clearance. Similar findings were reported with cefazolin. The cefazolin clearance during nocturnal HD was slightly lower (Cl = 1.65 L/hr) than during high-flux intermittent HD (Cl = 1.85 L/hr). However, a greater percentage of cefazolin was removed in 8 hours of nocturnal HD (80%) than conventional 4-hour high-flux HD (60%). The investigators concluded that a dosing regimen of a 2-g D _start followed by 1-g IV dose after each HD was sufficient to achieve concentrations 6 × MIC for Staphylococcus species for at least 70% of the dosing interval. As a result, though drug dosing in patients receiving one of these dialytic variants may be empirically increased, drug dosing regimens should be guided by drug-level monitoring when feasible. Drugs with a molecular size of 500 to 5000 Da appear to be particularly likely to have an increased clearance with nocturnal dialysis. Studies of modeled clearance have suggested that frequent HD regimens would be associated with enhanced clearance (and thereby the potential of under dosing) of daptomycin, for example. ^, This enhanced clearance was confirmed in the setting of AKI when the pharmacokinetics associated with extended daily dialysis (EDD) was investigated. These findings should be transferable to maintenance HD, with a degree of caution about the effects on distribution volumes that might arise in the setting of septic shock. One of the other effects of prolonged HD appears to be a reduction in rebound of drug concentrations after the termination of dialysis. ^, The rebound is explained by the fact that the drug transfer from the peripheral to central compartment needs time rendering rapid intermittent HD is less efficient in clearing drugs than prolonged or continuous HD.

There are >100 different dialysis or hemofilters available, and at least four distinct variants of HD are currently being used. The effect of HD or hemofiltration on the disposition of a drug may vary markedly and, because dialyzer or hemofilter clearance is rarely evaluated more than once, clinicians have to extrapolate data from one procedure to another. The enhanced efficiency of 21st-century dialyzers mean that most of the literature for medications developed before 2000 probably reflects an underestimation of the impact of HD. Consequently, the dosages of some medications may need to be empirically increased by 25% to 50%. Therapeutic drug monitoring should be used for drugs with narrow therapeutic indices to optimize safety and efficacy.

Assessment of The Impact of Hemodialysis

The most common means for assessing the effect of HD is to calculate the dialyzer clearance of a drug (Cl ^{pharmacodynamic} ) from plasma, as follows:

C l D p = Q p · ( [ A p − V p ] / A p ) ,

where Q _p is plasma flow through the dialyzer, A _p is the concentration of drug in plasma going into the dialyzer, and V _p is the plasma concentration of drug leaving the dialyzer. ^, This equation tends to underestimate HD clearance for drugs that readily partition into and out of erythrocytes. In addition, venous plasma concentrations may be artificially high if extensive ultrafiltration is performed, so Cl ^{pharmacodynamic} will be lower than it really is. Because of these limitations, the recovery clearance approach remains the benchmark for the determination of dialyzer clearance and can be calculated as follows ^, :

C l D r = R / AU C 0 − t ,

where R is the total amount of drug recovered unchanged in the dialysate and AUC _0–t is the area under the predialyzer plasma concentration–time curve during the period that the dialysate was collected. The HD clearance values reported in the literature may vary significantly, depending on which of these methods were used. The least number of assumptions is needed when the fraction (FR) of the amount in the body is derived from half-life T1/2 during time t on dialysis.

F R = 1 − 0.693 T 1 / 2 · t

It is common practice in most HD units to administer drugs after dialysis to minimize the loss of drugs that would result from the additional clearance during HD. However, performing HD immediately after dosing has been proposed as an option for removal of toxic antibiotics ^{, ,} and chemotherapeutic agents. For anticancer drugs, the administration of a normal dose up to HD 2 to 12 hours before dialysis makes sense. This strategy delivers the desired maximum plasma concentration effect while minimizing the toxic drug or metabolic effects (see Table 56.7 ). Emerging pharmacokinetic and pharmacodynamic considerations suggest that administration after HD may not be the optimal approach for some antibacterial agents, such as aminoglycosides, colistin, and vancomycin. ^{, ,} High-bolus dosing immediately before or during the last hour of dialysis has been proposed for some antibiotics, but there have been few clinical studies.

If the drug is given after dialysis, the postdialysis dose (D _HD ) should first replace the amount eliminated during the interval between dialysis sessions (D _fail ) that is the result of clearance by the patient’s residual kidney function and nonkidney clearance, in addition, the fraction of drug removed by HD (FR) should be estimated to calculate the necessary supplementary dose (D _suppl ). The dose the patient should receive after HD would thus be the sum of these two doses (see Fig. 56.5 ):

D HD = D fail + D suppl

D HD = D fail + FR · ( D start − D fail )

Supplementary dose after hemodialysis (HD) .

To maintain therapeutic target concentrations, a supplementary dose must be given after hemodialysis to replace the removed fraction of the dose. The dose after dialysis (D _HD ) combines both the adjusted maintenance dose (D _fail ) and supplementary dose (D _suppl ) .

In practice, the dose after dialysis corresponds to the initial/starting dose D _start or slightly less (D _HD ≤ D _start ).

CKRT and hybrid KRTs are commonly used to manage patients with AKI and ESKF in ICUs. CKRT provides less of a challenge for drug dosing than intermittent HD because its continuous nature is analogous with drug removal by native kidneys and potentially amenable to the use of standard, first-order drug clearance equations to calculate dosing. However, in practice, CKRT rarely proves as continuous as planned in unstable ICU patients. The CKRT modality and the delivered dialysis dose can also have significant effects on drug clearance. MW, membrane characteristics (highly variable between systems), blood flow rate, ultrafiltration rate, and dialysate flow rate determine the rate and extent of drug removal. Because most drugs have MWs < 1.5 kDa, drug removal by CRRT does not depend greatly on MW. The use of higher hemofiltration volumes, especially if infused prefilter, can also affect clearance. The removal of urea, creatinine, and vancomycin was increased by 15% to 25% by the predilution compared with postdilution modality. ^,

CKRT clearances have been noted to decline over time because of accumulation of protein on the dialysis membrane over time. Clotting within the hemofilter’s hollow fibers also reduces the overall surface area for clearance. Although these factors have received little direct investigation, it appears that they do affect drug clearance. ^,

Drug protein binding also affects how much is removed during CKRT because only unbound drugs are available for elimination by CKRT. Protein binding of >80% provides a substantial barrier to drug removal by convection or diffusion. During continuous venovenous hemofiltration, drug clearance generally approximates the ultrafiltration rate. The addition of diffusion by continuous venovenous hemodiafiltration increases drug clearance and is dependent on the ultrafiltration and dialysate flow rates. As is the case during high-flux dialysis, drug removal often parallels the removal of creatinine. Thus the simplest method for estimating drug removal during CKRT is to estimate the total CrCl. ^{, ,}

Hybrid KRTs, including sustained or slow low-efficiency dialysis (SLED), EDD, continuous SLED, slow low-efficiency daily dialysis, and slow low-efficiency daily hemodiafiltration, which use higher dialysate flow rates and shorter treatment periods (6–12 hours in duration), are frequently used. As of 2020, hybrid KRT and slow low-efficiency dialysis pharmacokinetic data have been published for few drugs. ^{, ,} The improvement of KRT machines and filters has rendered old dosing guidelines for drugs, especially antibiotics, obsolete and potentially hazardous because they are underdosed. Although there are only a few FDA or EMA official drug dosing recommendations for patients receiving CKRT, several published dosing guidelines are widely used. ^{, , , ,} Unfortunately, these recommendations have generally not been prospectively evaluated and their influence on patient outcomes is largely unknown.

In the absence of FDA or EMA recommendations, tertiary reference sources, or any published studies relating to the handling of a drug by CKRT (common with agents that are new to the market), it may be necessary for the clinician to formulate a dosing regimen using the pharmacokinetics principles presented in this chapter. The lack of pharmacokinetics data for antibiotics during prolonged KRT and intermittent HD in critically ill patients has made the attainment and maintenance of effective concentrations a major challenge. ^, If the V _D is large (>1 L/kg), there is a low likelihood that CRRT will substantially remove the drug. The use of high-flux dialyzers or hemofilters allows for drugs with an MW < 20 kDa to be readily removed. If the clearance of the drug by CKRT or hybrid KRT is <25% of the patient’s estimated total body clearance, a further dosing adjustment is probably unnecessary. By contrast, if CKRT or hybrid KRT results in an augmentation of drug clearance by 25% to 50%, an D _start based on the patient’s estimated volume status should be given, and maintenance doses similar to those given to a patient with an eGFR of 30 to 50 mL/min can be used. Such estimates obviously must be supplemented by regular drug concentration measurements, where technically feasible.

Peritoneal dialysis, as practiced today, is unlikely to enhance total body clearance of any drug by >10 mL/min because most typical peritoneal dialysis prescriptions can achieve a urea clearance of about ≤10 mL/min. As most drugs are larger than urea, their clearance is even less; thus it is likely to be from 5 to 7.5 mL/min or less. Many studies performed in the 1970s and 1980s showed that drug clearances by peritoneal dialysis were in this low range, so one can conclude that peritoneal dialysis does not enhance drug removal to a degree that would require a special dosage regimen modification. Thus oral or IV drug therapy recommendations for patients with an eGFR <15 mL/min are likely clinically useful.

Intraperitoneal drug administration is well accepted for the treatment of peritoneal dialysis–associated peritonitis and other infections. ^, Administration intervals depend on the half-life of the drug, which is mainly determined by residual kidney and nonrenal metabolic clearance. Long-standing experience with intermittent antibiotic administration in continuous ambulatory peritoneal dialysis (CAPD) exists for the glycopeptides vancomycin and teicoplanin, which can be administered at 5- to 7-day intervals, as well as for aminoglycosides and cephalosporins, which are suitable for once-daily dosing. ^,

Patients treated by automated peritoneal dialysis (APD), with frequent short-dialysis cycles, may achieve higher plasma concentrations as compared with antibiotic loading in a single extended dwell period in patients on CAPD. Conversely, the higher dialysate flow and small-molecule clearance achieved with APD regimens may lead to a greater peritoneal clearance of antibiotics with the need for more frequent dosing. One common regimen involves adding antibiotics to each bag of dialysate solution used during APD exchanges. The choice of antibiotic, dosage, and frequency may vary depending on the patient’s clinical condition, local guidelines, and presence of any allergies or drug interactions. The International Society of Peritoneal Dialysis (ISPD) guidelines emphasize the importance of individualizing antibiotic treatment based on factors such as the patient’s medical history, previous infectious complications, and local microbiologic data.

For intermittent maintenance dosing, a long nighttime dwell time should be used in patents on CAPD and a long daytime dwell time in patients on APD. In clinical practice, intraperitoneal antibiotic dosing has not been unequivocally successful in eradicating bacterial growth, partially questioning the concept of antibiotic back diffusion into the peritoneal cavity. Thus antibiotics may need to be administered IV to enhance the antimicrobial effect.