Kidney function

Introduction

Kidney function in oncologic patients is an important parameter for several reasons. In some specific cases, like urologic cancers, the tumor itself can cause acute or chronic kidney dysfunction. However, the renal function parameter is more frequently followed in oncologic patients because, on one hand, they are likely to get nephrotoxic drugs, and on the other hand, some oncologic patients have chronic kidney disease (CKD) that would necessitate a dosage adaptation for potentially toxic chemotherapies. ^, Indeed, excretory renal function plays a fundamental role both in the pharmacokinetics and pharmacodynamics of several drugs. This is particularly the case for water soluble compounds and/or their active metabolites. Even for non-renally-excreted drugs, severe CKD can modify the pharmacokinetics by several mechanisms. For this reason, it is now recommended that both pharmacokinetics and pharmacodynamics of every new drug be studied in the context of CKD. Dosage-adjustment according to excretory renal function is required for many medications. However, there is a debate in the literature regarding the best way to estimate excretory function or glomerular filtration rate (GFR) for the purpose of pharmacotherapy.

Serum creatinine

The word “creatinine” was probably used for the first time by Justus von Liebig in 1847. This German chemist was thus describing the product obtained from heating creatine with mineral salts. Nowadays, serum creatinine is one of the most frequently prescribed analysis in Clinical Chemistry. Serum creatinine is the only renal plasma biomarker used in daily clinical practice to estimate GFR. ^, However, a good interpretation of the creatinine result remains sometimes problematic, or at least not so simple. To explain these difficulties, there are both physiologic (serum creatinine is not an “ideal” renal marker) and analytical reasons. First, serum creatinine can be measured by two main methods: methods derived from the classical Jaffe reaction on one part, and enzymatic methods on the other part. The Jaffe method is based on a reaction between picrate and creatinine in alkaline milieu that gives a red-orange product. Some components (so-called pseudochromogens) can however also interact with picrate: acetoacetate, pyruvate, ketonic acids, proteins, glucose, and ascorbic acid. These pseudochromogens take part in 15% to 20% of the Jaffe reaction if the serum creatinine is in the normal range. This limitation of Jaffe methods remains even after different technological innovations. The second method is known as the enzymatic and is based on successive enzymatic steps. Different types of reactions have been described, but they all share a higher specificity to measure serum creatinine, compared to Jaffe assays. Enzymatic methods are thus considered as more accurate and precise than Jaffe methods. These methods are recommended even if they are more expensive and not fully free from some interferences. There are two methods to measure creatinine but for each method, there are also several different assays (according to the manufacturers). Until recently, a great heterogeneity was observed between the assays, because of differences in calibration. ^, Nowadays, several improvements have been done in a quest to standardization, and it is recommended to measure serum creatinine with a standardized, calibrated, and so-called isotope dilution mass spectrometry ( IDMS )-traceable method. This traceability is of the highest importance in the context of creatinine-based equations. ^,

Beyond analytical issues, there are also physiologic limitations to serum creatinine. The molecular weight of creatinine is 113 Daltons. Creatinine is the anhydric catabolite of creatine and phosphocreatine. Creatinine is a catabolite final product and has no physiologic role. The vast majority of creatine (98%) will be found in muscles where creatine is phosphorylated in phosphocreatine by creatine kinase. Each day, 1% to 2% of the muscle creatine is converted into creatinine. ^, It is thus obvious that serum creatinine concentration is highly dependent on muscle mass. If the global creatine concentration is constant in healthy subjects, this concentration will strongly vary notably in muscular pathologies or, from a larger point of view, in all diseases with anorexia and muscle mass decreasing, as it is frequently observed in oncologic diseases. In these situations, serum creatinine will decrease, or will not increase when GFR is decreasing. ^, ^, The second important limitation is the tubular secretion of creatinine, which explains that creatinine clearance overestimated measured GFR (mGFR). Moreover, even if creatinine clearance, calculated on a 24-hour urine collection, is a relatively simple way to assess GFR, such a clearance also lacks precision, especially because of large errors in urine collection. Therefore creatinine clearance is not recommended for GFR estimation. A last point needs to be underlined: the relationship between serum creatinine and GFR is not linear but hyperbolic ( Fig. 1.1 ). ^, To keep this point in mind is fundamental for a good interpretation of a creatinine result. Actually, this hyperbolic relation implies that a little creatinine change will have great consequences in terms of GFR in low creatinine levels, although the same creatinine variation in higher ranges will be negligible in terms of GFR. ^, ^,

The creatinine-based equations in oncology

Serum creatinine concentration is dependent on muscular and thus on gender, age, and ethnicity. Moreover, the relationship between serum creatinine and GFR is hyperbolic (see Fig. 1.1 ). These observations lead authors to propose creatinine-based equations that include simple variables like age, gender, ethnicity and, for some of them, weight. Because of the hyperbolic association, a negative exponent is applied to serum creatinine value in these equations, making the association between estimated GFR (eGFR) and mGFR by a reference method, simple and linear. Several equations have been proposed since the 1950s. Before 2000, the Cockcroft and Gault (CG) equation was the most popular and has been used by generations of doctors to estimate GFR. This equation has, however, several limitations, notably a lack of precision caused by the variable weight included in the formula. Today, new equations based on serum creatinine and including only age and gender are recommended and their results can be automatically given by laboratories. These equations are applicable only with IDMS-traceable assays. Among these equations, we can cite the most used, that is, the modification of diet in renal disease (MDRD) equation (most valuable in CKD patients) ^, and chronic kidney disease epidemiology (CKD-EPI) equations (also valuable in healthy populations). Recently, some authors have challenged the superiority of these equations both in the general and CKD populations, and proposed new algorithms, known as the revised Lund-Malmö equation or the full age spectrum equations ( Table 1.1 ). Globally, both in the general or CKD population, one can expect an accuracy around 85% or 90% for creatinine-based equations. This means that results of eGFR equations will be within ±30% of the results of mGFR by a reference method in 85% to 90% of subjects/patients. ^, Beyond the experts’ discussion to know which equation performs the best, we have to keep in mind that the most important variable in all these equations remains the serum creatinine. Therefore there is no reason to believe that eGFR, by any estimating equations, would be accurate in specific patients or specific populations for who the serum creatinine, in itself, is particularly inadequate. In the context of oncology, it will be particularly relevant in patients with anorexia, loss of weight, and cachexia. ^, In these patients, serum creatinine concentration is inaccurate, and can decrease in patients with stable function or remain stable in patients with worsening GFR. Because serum creatinine is inaccurate, eGFR based on serum creatinine will be unable to correctly assess GFR in these patients. In 21 young patients suffering from anorexia nervosa, GFR was measured by a reference method (namely, the Cr-Ethylenediaminetetra-acetic acid (EDTA) plasma clearance). The performance of both the CG and MDRD equations was dramatically poor, with an accuracy within 30% of only 63% and 30%, respectively. Both the MDRD and CKD-EPI equations strongly overestimate mGFR in these specific patients. Redal-Baigorri et al. studied the performances of creatinine-based equations in 185 patients with cancer. GFR was measured by Cr-EDTA plasma clearance before chemotherapy and serum creatinine was IDMS traceable. Majority of patients were suffering from pulmonary cancer (77%). The performance of MDRD and CKD-EPI was acceptable around 89% for both equations, but most of the patients were healthy from a renal point of view, with only 17% of patients with mGFR below 60 mL/min/1.73m². Lauritsen et al. included patients with disseminated germ cell cancer and treated by bleomycin, etoposide, and cisplatin. The GFR was measured by Cr-EDTA single sample plasma clearance in 390 patients, but more than 1600 measurements were obtained with data before and just after chemotherapy, and then after 1, 3 and 5 years. Patients were young with few comorbidities and thus with normal mGFR values. CG, MDRD, and CKD-EPI equations have relatively good performance in estimating GFR before and in the years following chemotherapy (around 85% and 90%). However, values after three cycles of chemotherapy were more disappointing with an accuracy within 30% at 76%, 80%, and 50%, for CG, MDRD, and CKD-EPI, respectively. Actually, mGFR significantly declined during chemotherapy (−9 mL/min/1.73m²) but serum creatinine also decreased (and so eGFR equations increased). Funakoshi et al. also studied the performance of the CKD-EPI equation (both original and Japanese version and de-indexed for body surface area [BSA]) in 50 patients with cancer scheduled for cisplatin therapy (majority had neck and head cancer) before and after treatment. Only subjects with mGFR over 50 mL/min were considered. The accuracy of both CKD-EPI equations was acceptable with an accuracy within 30% of 92% and better than the CG equation (accuracy of 78%) before the therapy. However, a significant decline in accuracy was observed for all equations after the cisplatin cycle: 60%, 68%, and 56% for the original CKD-EPI, the Japanese version, and the CG, respectively. All equations overestimated mGFR, especially in the lower GFR ranges. Moreover, in the postchemotherapy period, one-quarter of patients with CKD-EPI values over 60 mL/min had actually a measured GFR below 50 mL/min. Lindberg et al. also included 94 patients with advanced head and neck cancer treated by radio-chemotherapy including cisplatin. The GFR was measured by Cr-EDTA single sample plasma clearance before the first cycle ( n =94), after the third cycle of chemotherapy ( n =78), and after the planned five cycles ( n =35). The nonindexed BSA result was considered for mGFR and eGFR (CG and CKD-EPI). At baseline, five patients were eventually not treated by cisplatin because mGFR was below 50 mL/min. These five patients had eGFR with CKD-EPI above 50 mL/min (and three with CG). At the end of the treatment, three patients had mGFR below 50 mL/min, but this decline in mGFR was detected by both equations in only one patient. The Bland and Altman analyses showed a relatively acceptable systematic bias between mGFR and CG and a systematic overestimation of mGFR by CKD-EPI, but the precision of both equations was actually poor. Hingorani et al. measured GFR by iohexol plasma clearance in 50 patients who benefit from a hematopoietic cell transplant at baseline and after 100 days. The authors compared the performance of mGFR with CG (nonindexed for BSA), MDRD, and CKD-EPI (both indexed for BSA). At baseline, all patients were also treated by trimethoprim, which is known to block the tubular secretion of creatinine, leading to an increase in creatinine concentration independently of any change in mGFR. At baseline, CKD-EPI and MDRD underestimated mGFR and CG overestimated it. The accuracies were low for patients with mean normal GFR values. Indeed, accuracy within 30% at baseline was 79%, 70%, and 57% for CKD-EPI, MDRD, and CG equations, respectively. After 100 days, the accuracy observed was similar for CKD-EPI and MDRD and slightly better for CG. All these data in cancer patients compared mGFR by a reference method with eGFR, but the studies share methodologic limitations, notably the samples being relatively limited. However, it appears from these analyses that the accuracy of the equations is, at best, suboptimal in cancer patients. This observation is not fully unexpected as these patients are frequently frail and have decreased muscle mass because of their pathology and/or chemotherapy. In the same vein, the inaccuracy of these equations seems especially important during or after the chemotherapy cycles, which is, once again, not fully unexpected.

Table 1.1

Creatinine- and Cystatin C-Based Equations to Estimate Glomerular Filtration Rate

Cockcroft and Gault		((140-age) × weight)/(72 × SCr)
MDRD		175 × SCr ^-1.154× age ^-0.203× [0.742 if female] × [1.212 if black]
CKD-EPI SCr
women	SCr ≤ 0.7 mg/dL	144 × (Scr/0.7) ^-0.329× 0.993 ^age× [1.159 if black]
	SCr > 0.7 mg/dL	144 × (Scr/0.7) ^-1.209× 0.993 ^age× [1.159 if black]
men	SCr ≤ 0.9 mg/dL	141× (Scr/0.9) ^-0.411× 0.993 ^age× [1.159 if black]
	SCr > 0.9 mg/dL	141× (Scr/0.9) ^-1.209× 0.993 ^age× [1.159 if black]
CKD-EPI SCys	SCys ≤ 0.8 mg/L	133 × (Scyst/0,8) ^-0.499× 0.996 ^age[× 0.932 if female]
	SCys > 0.8 mg/L	133 × (Scys/0,8) ^-1.328× 0.996 ^age[× 0.932 if female]
CKD-EPI SCrCys
women	SCr ≤ 0.7 mg/dL and SCys ≤ 0.8 mg/dL	130 × (SCr/0.7) ^-0.248× (SCyst/0.8) ^-0.375× 0.995 ^age× [1.08 if black]
	SCr ≤ 0.7 mg/dL and SCys > 0.8 mg/dL	130 × (SCr/0.7) ^-0.248× (SCyst/0.8) ^-0.711× 0.995 ^age× [1.08 if black]
	SCr > 0.7 mg/dL and SCys ≤ 0.8 mg/dL	130 × (SCr/0.7) ^-0.601× (SCyst/0.8) ^-0.375× 0.995 ^age× [1.08 if black]
	SCr > 0.7 mg/dL and SCys ≥ 0.8 mg/dL	130 × (SCr/0.7) ^-0.601× (SCyst/0.8) ^-0.711× 0.995 ^age× [1.08 if black]
Men	SCr ≤ 0.9 mg/dL and SCys ≤ 0.8 mg/dL	135 × (SCr/0.9) ^-0.207× (SCyst/0.8) ^-0.375× 0.995 ^age× [1.08 if black]
	SCr ≤ 0.9 mg/dL and SCys > 0.8 mg/dL	135 × (SCr/0.9) ^-0.207× (SCyst/0.8) ^-0.711× 0.995 ^age× [1.08 if black]
	SCr > 0.9 mg/dL and SCys ≤ 0.8 mg/dL	135 × (SCr/0.9) ^-0.601× (SCyst/0.8) ^-0.375× 0.995 ^age× [1.08 if black]
	SCr > 0.9 mg/dL and SCys > 0.8 mg/dL	135 × (SCr/0.7) ^-0.601× (SCyst/0.8) ^-0.711× 0.995 ^age× [1.08 if black]
FAS creatinine		(107.3/SCr/Q _creat) × (0.988(Age-40)) if ^age> 40 y
FAS cystatin C		(107.3/Scyst/Q _cyst) × (0.988(Age-40)) if ^age> 40 y
FAS combined		[107.3/(0.5 × Scyst/Q _cyst+0.5 × SCr/Q _creat] × (0.988 ^(Age-40)) if ^age> 40 y
LM	women SCr < 150 µmol/L	e ^{[2.5+0.0121 × (150-SCr)]-0.0158 × age+0.438 × Ln(age)}
	women SCr ≥ 150 µmol/L	e ^{[2.5-0.926 × Ln(SCr/150)]-0.0158 × age+0.438 × Ln(age)}
	men SCr < 180 µmol/L	e ^{[2.56+0.00968 × (180-SCr)]-0.0158 × age+0.438 × Ln(age)}
	men SCr ≥ 180 µmol/L	e ^{[2.56-0.926 × Ln(SCr/180)]-0.0158 × age+0.438 × Ln(age)}
CAPA		130 × SCyst ^-1.069× age ^-0.117-7