Measurement of renal function during drug development

A large number of drugs are eliminated to a clinically important degree by the kidney. For many drugs, this means that in patients with decreased renal function, doses need to be decreased to avoid exposure of the patient to excessive drug concentrations that may cause toxicity. In general, if 30% or more of a drug's total clearance is dependent on the kidney, dose reduction should be considered in patients with decreased renal function. Thus, it is important early in drug development to quantify the renal contribution to overall drug elimination. If a drug is eliminated by the kidney, it is necessary to determine the pathways of elimination; for example, glomerular filtration, as opposed to active secretion. If the latter process operates, in addition to the need to decrease doses in patients with diminished renal function, there is also a likelihood of drug interactions that must be considered in labelling and clinical use. During development of a drug, there occasionally will be a change in renal function raising the question of nephrotoxicity. Sometimes such effects are indeed adverse renal effects; sometimes it is a functional effect such as inhibition of the secretion of creatinine 1, 2 or it may cause a false positive renal screening test. In the former case, development may cease; the latter simply dictates a need to inform clinicians of the risk of a false positive renal test. A host of methods, both in vivo and in vitro, have been employed to determine renal function accurately and to assess mechanisms of renal elimination of drugs. A variety of tests are used clinically to measure renal function. In drug development these tests are best viewed as screening in nature. Net renal function is assessed clinically by measurement of serum creatinine or blood urea nitrogen (BUN) concentrations and, on occasion, measurement of creatinine clearance. The inaccuracy of these markers derives from the fact that their renal elimination is not purely by filtration at the glomerulus 1–6. Creatinine has a secretory component of elimination that is about 15% of total excretion in subjects with normal renal function, and which is even greater in patients with decreased renal function. Thus, an increase in serum creatinine can occur from either a decrease in overall renal function (GFR) or because a drug has blocked its secretion. The implications of these two scenarios differ markedly. Urea is not only filtered but is also reabsorbed by the proximal nephron 7. In conditions of increased proximal tubular reabsorption of solute including a variety of oedematous disorders, BUN rises, but this increase is not due to a decrease in global renal function per se. Because of the limitations of these clinical indices of renal function, considerable research has been devoted to the identification and validation of better markers of renal function. Some of these are listed in Table 18–15. However, many limitations remain. Even with these markers, a variety of methods can be employed when using them. For example, and as will be discussed below, with any of the GFR markers one can use a single injection technique and measure marker disappearance rate. Alternatively, a continuous infusion of the marker can be given with collection of serial blood and urine samples to calculate renal clearance directly. When designing studies to assess renal function precisely, it is vitally important that not only is the best marker selected, but the best method for using the marker is used. Experimentally, CAV is often assumed to be approximated by the plasma concentration at the midpoint of the urine collection interval. This assumption can be erroneous if the plasma concentration is changing in a nonlinear fashion, as typically occurs after intravenous bolus administration of a marker or drug or during the absorptive phase of a drug given by mouth. The schematic in panel a of Figure 1 illustrates this potential problem wherein the midpoint plasma concentration is clearly not the average concentration during the interval specified. In contrast, panel b of Figure 1 depicts an example where the change in plasma drug concentration over the duration of the collection interval is linear wherein the midpoint plasma sample is representative of the average concentration. These same considerations apply when markers are used to quantify renal function using a single injection technique followed by measuring only plasma disappearance of the marker. On the one hand, if the marker is only filtered, if renal function is stable, and if the disappearance curve is linear, then this method is both simple and yields accurate measures of GFR. On the other hand, if any of these assumptions is violated, then the single injection method is flawed. As noted previously, care must be taken to select the best method for the marker or drug and the condition in which it is studied. Lastly, panel c depicts a method in which any problems related to nonlinearity are avoided by administering the substance as a continuous intravenous infusion to achieve steady-state plasma concentrations. Unfortunately, not all markers, particularly those that are radiolabelled, can be given over the lengths of time needed for this method and many drugs do not have an intravenous formulation. Schematic of different drug administration methods for determining the renal clearance of a marker or drug. Panel a shows plasma drug concentration-time curves after a single i.v. (descending curve) and oral (ascending curve) dose. Panel b shows a linear plasma drug disappearance curve. Panel c shows a constant plasma drug concentration at steady state during continuous i.v. infusion (from reference 16, with permission). Another potential source of error in using measurements of the renal clearance of a marker or drug is in the urine collection itself 18. If the subject/patient has a bladder catheter, the accuracy of the urine collection is less of an issue. However, urine is usually collected by spontaneous voiding. It goes without saying that the times of the collection must be accurate. In addition, the collection interval and urine volume must be sufficient to obtain an accurate collection such that urine flow can be started and the bladder can be as completely emptied as possible. Small urine volumes collected over short intervals are likely to be subject to considerable error. If short (e.g. 20–30 min) collection intervals are to be used, patients must receive sufficient fluids to assure urine volumes of several hundred millilitres. In this manner, even if residual urine volume is as high as 5 ml, any error would be small and unimportant. Note, that renal clearance in all of the preceding equations represents all components of renal clearance. For a substance like creatinine, it represents filtration clearance and secretory clearance. For some drugs it might represent filtration clearance plus secretory clearance less reabsorptive clearance. Thus, this value per se offers no mechanistic information as to the pathway(s) of renal excretion. On the other hand, if a marker is used that is only filtered, then renal clearance in these equations represents GFR. If, for example, a drug's renal clearance is quantified along with a valid marker of GFR, and the drug's renal clearance exceeds GFR, then one can reliably conclude that the drug of interest has a secretory component of elimination. Such data do not exclude filtration or reabsorption and do not quantify the secretory component, but they do allow the conclusion that secretion occurs. Once the renal clearance of a drug is measured, inferences can be made concerning its mechanism of excretion. Since only unbound drug can be filtered, if renal clearance of unbound drug is equal to GFR, it can be assumed that all renal elimination is by filtration (unless secretion is exactly balanced by reabsorption). If renal clearance of unbound drug is less than GFR, then it can be assumed that the difference is accounted for by reabsorption. Such an observation may prompt additional studies to assess the influence of urinary flow rate and pH on the reabsorptive component. Lastly, if renal clearance of the unbound drug is greater than GFR, active secretion must be involved. Further assessment might entail additional studies with inhibitors of secretion to identify which secretory pathway is operative (thereby allowing prediction of potentially interacting drugs) and to quantify this component of renal elimination. For many drugs, the studies mentioned above can only provide qualitative and semiquantitative data concerning the various components of renal excretion. It is often impossible in man to determine the precise quantitative contribution of the different pathways. For example, a drug might be filtered and secreted and then reabsorbed. Precise quantification of the magnitude of each component of elimination would require the ability to completely block one or more excretion pathways with assessment of the effect of such a manoeuvre on overall renal elimination. Such blockade is usually impossible. This fact notwithstanding, knowing qualitatively the mechanism of a drug's renal elimination can allow appropriate cautions and predictions as to the influence of renal disease and the potential for drug interactions, as will be discussed subsequently. A general rule of thumb is that for renal elimination to contribute meaningfully to overall drug elimination, the amount of drug excreted unchanged in the urine should be 30% or more of the dose. There are two aspects of renal metabolism that invalidate the 30% rule. The proximal tubule has peptidases that are capable of digesting peptides and proteins. If a drug or polypeptide therapeutic agent is freely filtered at the glomerulus, the agent can be metabolized at the proximal tubule so that no or negligible amounts of agent appear unchanged in the urine. In fact, all excretion may be dependent upon the kidney, yet no unchanged drug is detected in urine at all. A common example of this phenomenon refers to insulin. Renal metabolism accounts for approximately 30% of overall insulin elimination. Accordingly, in patients with severe renal insufficiency, insulin requirements are often decreased 19. Another compound that undergoes renal metabolism is imipenem which is digested by dipeptidases after being filtered at the glomerulus 19. This means that imipenem alone is not effective for treating urinary tract infections. Effectiveness is maintained clinically by administering imipenem in combination with cilastatin, a compound that inhibits proximal tubular dipeptidases, thereby allowing imipenem to escape proximal tubular metabolism and be effective as an antibiotic in the lower urinary tract. A third example is superoxide dismutase (SOD) 20. Superoxide dismutase is a sufficiently small protein that it is freely filtered at the glomerulus, yet at low doses no SOD can be detected in the urine. This is because it is completely metabolized at the proximal tubule. As doses are increased, the metabolic capacity of the proximal tubule is exceeded, and SOD can be detected in the urine. Only at these higher doses can the contribution of renal clearance to overall clearance (Figure 2) be discerned. The clinical importance of this observation is that, based on the inability to detect unchanged drug in the urine, it might be concluded that the kidney plays no role in SOD disposition and that the protein is completely metabolized by the liver. To the contrary, a substantial component of SOD elimination depends upon the kidney, such that patients with decreased renal function will be at risk of accumulation of the protein and that dosage adjustment may be necessary. If such information had not been generated prior to efficacy trials, no provision for dosage adjustment would have been made for patients with decreased renal function. Such patients might then be exposed to higher concentrations than expected and suffer adverse events. Use of the drug might then appear to have risks much greater than would have been the case if the true pathways of elimination had been determined in early studies followed by appropriate protocol modification and dose adjustment in later phase trials. Clearance of superoxide dismutase (SOD) as a function of dose. Note that total clearance remains relatively constant while renal clearance as a percent of total clearance increases with dose. This occurs because at higher doses proximal tubular metabolism of SOD is overwhelmed allowing recovery of unchanged SOD in the urine and a more accurate assessment of the contribution of renal clearance. Estimates of renal clearance with low doses are inaccurate because the SOD cleared by the kidney never reaches the urine owing to proximal metabolism. The phenomenon described with SOD likely applies to many if not most peptides, peptidomimetics and small proteins that are used or being developed. In fact, it should be presumed that any peptide or protein that is sufficiently small to be freely filtered at the glomerulus would have the same characteristics described for SOD and that specific studies should be performed to determine whether or not renal excretion is a significant component of overall elimination. Another aspect of drug metabolism that can be misleading with respect to the importance of renal function in drug elimination is the phenomenon called ‘futile cycling’21–24. This occurs with drugs that are metabolized to acyl-glucuronides. Acyl-glucuronides are those where a carboxylic acid group in the parent drug molecule is conjugated through an ester linkage. Unlike ether glucuronides, ester glucuronides are chemically unstable, being readily hydrolysed, such that there is an equilibrium between parent drug and the acyl-glucuronide. Most glucuronides are excreted by the kidney. Thus in a subject with normal renal function an acyl-glucuronide would be formed in the liver, and this metabolite would be readily excreted into the urine. However, in patients with decreased renal function, excretion of the acyl-glucuronide is impaired, and it can accumulate in the plasma, where it spontaneously hydrolyses back to parent drug (Figure 3). The net result is an increase in circulating parent drug concentration and risk of adverse effects. The likelihood of such a ‘futile cycle’ should be expected with any drug where formation of an acyl-glucuronide metabolite is possible. Examples of such drugs include clofibrate, ciprofibrate, and a variety of arylpropionic nonsteroidal anti-inflammatory drugs such as ketoprofen 21–24. Schematic of a ‘futile cycle’ of drug metabolism where acyl-glucuronide metabolites are retained in patients with renal insufficiency. They are then able to spontaneously hydrolyse back to parent drug in the circulation resulting in its accumulation even though the parent drug per se undergoes no direct renal excretion. The kidney performs the physiologic functions of filtration, secretion and reabsorption. It is important to understand the implications of these different processes for drug handling. For example, filtration only refers to the fraction of circulating drug that is not bound to plasma protein. Secondly, filtration has size limitations owing to the dimensions of the glomerular pores and their electrical charge (negative). Size limitations are not relevant to traditional xenobiotics, but are relevant to substances such as high molecular weight dextrans and proteins. In general, molecules less than 30 Angstroms are freely filtered at the glomerulus. For drugs that are filtered, it is predictable that decreases in overall renal function, as reflected by a decreased GFR, will result in drug accumulation and may require dose adjustment in patients with renal disease. For drugs that are actively secreted, glomerular filtration rate is less important than renal plasma flow. For some drugs, their avidity for the transport pump is sufficiently greater than the binding affinity to circulating proteins that the drug is effectively ‘stripped’ from the protein and secreted into the urine. As such, they can achieve renal clearance rates that approach renal blood flow, i.e. in the limit all the drug presented to the kidney can be removed from the perfusing blood and secreted into the urine. Drugs that are actively secreted will also accumulate in patients with decreased renal function, because renal blood flow and GFR usually decrease in parallel. However, in addition, these drugs are subject to competition for transport. Thus, there is concern with clinical situations in which concomitant drugs might be administered that could inhibit transport, thereby exposing the patient to increased circulating drug concentrations. The tubular reabsorption of drugs follows the principles of passive nonionic diffusion and is dependent upon urinary pH and the pKa of the drug. A urine pH that favours the nonionized state of the drug enhances its lipophilicity and increases its passive reabsorption across the tubular membrane. For drugs that are subject to reabsorption, the extent of this process is also dependent on urinary volume. The clinical implications of kidney tubular reabsortion relate mostly to the effect of disease on the process. Normally, the urinary pH is acidic. Thus the considerations for drug reabsorption occur mainly where there are alterations in urinary pH such as in syndromes of renal tubular acidosis or during administration of acids or bases for other clinical indications. As noted previously, the overall elimination of many drugs has an important component that is dependent upon renal function 25. Among the different pathways of renal elimination, glomerular filtration is the most important in terms of the numbers of drugs that are affected. Compendia are published periodically that enumerate these drugs and offer recommendations as to the need for adjusting doses in patients with renal insufficiency (e.g 25). From a drug development perspective, it is critically important in the earliest phase of development to determine whether or not renal function is an important component of elimination. As noted previously, excretion of 30% or more of a parent drug or active metabolite in the urine of dose adjustment in patients with renal insufficiency using the drug in such Even if a drug this of excretion in the urine active the implications are the same in terms of the potential need for dose adjustment in patients with renal insufficiency because renal blood flow and GFR change in parallel. If early of drug development that renal elimination is studies need to be performed to whether there is an active secretory component to elimination. Such studies require the use of markers of GFR, true glomerular filtration rate to renal clearance of the unbound drug. If the latter exceeds GFR, then secretion is a component of elimination in addition to additional studies need to be to the secretory pathway and the potential for drug If secretion is not a component of renal elimination, studies are to assess the between total clearance of the drug of interest and GFR. These will allow recommendations for dose adjustment in clinical use. In addition, it is to identify of patients in use of the drug should be avoided secretion for the active secretion of drugs from blood to urine have been for acids bases and is also of these transport have been as have for them. in a variety of have been but considerable research remains to their of these is capable of a variety of and there is potential for competition for transport. Thus, acids can with other bases with A example refers to a of the acid This compound decreases the renal secretion of other such as and The are less include and competition for transport between is possible. of is even less for this pathway could with one but this is in interest has on over the This in to of and related is for this by of the is also in the the and the kidney. studies have that is able to a variety of into the urine The for in the kidney is include and As noted if the unbound renal clearance of a drug exceeds that of a marker for GFR (e.g. then active secretion is From the of the it may be to which might be involved. Such predictions can in and in vivo studies at the role of such and other that inhibit renal excretion. for example, can be in a and this can be used in to assess whether or not a drug is a can also be in such upon the in appropriate in vivo studies should be performed to interactions and their Such information can then be used for labelling and for or during efficacy trials. nonionic tubular reabsorption Drugs that are excreted into the kidney tubular are also subject to reabsorption at the This process follows the principles of passive nonionic such that the nonionized of a drug is more and can be reabsorbed The degree of reabsorption may with urinary For drugs, an urine in an increased concentration of nonionized drug, thereby the extent of reabsorption. In contrast, an urine increases the degree of thereby renal excretion. This is the for the use of an in treating For drugs such as the is true such that an urine favours the nonionized reabsorption and overall drug elimination (Figure 20. of urinary pH on the elimination of Since is a as urine pH more the concentration of nonionized drug in tubular increases and is subject to reabsorption. This decreases renal elimination and the (from reference with permission). If, during drug it is that the compound undergoes substantial tubular reabsorption, studies can be to determine whether or not urinary pH has an important on its overall elimination. However, in whether to such it is necessary to that urinary pH is acidic. There are only a conditions where urine pH is These include syndromes of renal tubular acidosis and where patients are administered high doses of If the drug development might be used in such then it is important to studies of urine pH For drugs that urine pH dependent elimination, urine volume can also influence the total amount of drug that is excreted Since of urinary volume usually only occur in patients with such as this is of clinical There are many in which an of renal function on drug Thus, it is important to be with methods for the accurate of renal function and of the in which the kidney to drug A of these at the earliest of drug development will allow better of doses for clinical trials, better of and and more accurate labelling a drug is