Renal function declines with age as a result of the anatomical and physiological changes that occur with aging. Because renal function deteriorates with age, even in the seemingly healthy patient, clearance of renally excreted and metabolized drugs may be markedly impaired in the elderly patient. Therefore, to avoid toxic effects a dosage reduction is often necessary.1,2
Because both the production and excretion of creatinine decreases with age, an elderly patient can have markedly decreased renal function without having an elevated serum creatinine level. Thus, an estimate of creatinine clearance is necessary for determining the appropriate dosage of renally excreted drugs. Creatinine clearance can either be measured by a timed urine collection or estimated from serum creatinine.1,2
Timed urine collections are associated with significant collection errors, due to improper timing and missed samples. Timed overnight collections or shorter timed daytime collections may reduce the inconvenience of a 24-hour collection, but are still associated with collection errors.
Creatinine clearance equations
An accurate, reliable and unbiased method for calculating creatinine clearance from serum creatinine is much sought after by clinicians. Dozens of methods for estimating creatinine clearance have been published, but no single method is ideal for all patients. Furthermore, much controversy exists as to which method is best for a particular patient group.
In adults, the Cockcroft and Gault equation has become the defacto standard despite many documented problems.5,6,7,8 C&G produces consistent results in patients of average size and build, with stable renal function and a SCr less than 3 mg%. However, it is problematic in others.
The C&G equation was derived from a group of lean males, bringing into question the validity of this method in obese patients. The Salazar and Corcoran equation was derived from an obese patient population, and may be more appropriate for this group.16
In patients with unstable renal function, Jelliffe's multi step method may be more accurate. This method corrects for rising SCr and for chronic renal failure.15
The MDRD method was derived from a study of a large diverse patient population having a wide range of renal function. The MDRD equations have also been validated in a separate, equally large and diverse group. Therefore, some feel that MDRD is the most accurate CLCR method overall.26, 27 Nevertheless, for dosing purposes, NKDEP does not recommend using the MDRD study equation at this time because the clinical impact on drug dose adjustment has not been compared to current practice. Pharmacists should continue to use their current drug dosing methods.28
Please see our online creatinine clearance calculator to compare the results from these various adult creatinine clearance equations:
In children 12 and older, the Cockcroft and Gault equation gives a reasonably accurate estimate of creatinine clearance. For younger children, infants and neonates, no method of estimating CLCR is reliable. The Swartz equation is the standard equation for young children, however, the results are not consistent enough to be used for pk modeling.20
Dosage adjustment guidelines which are based on creatinine clearance have been published.3,4 Manufacturers are now required by FDA to provide dosage guidelines for patients with decreased creatinine clearance.
A dosage regimen may be adjusted either by lowering the dose or prolonging the dosage interval. The dosage reduction method is recommended for those drugs for which a relatively constant blood level is desired, e.g., beta-lactam antibiotics. The interval extension method is recommended for those drugs whose efficacy is related to the peak level, e.g., fluoroquinolone antibiotics.
The following screening criteria may be used to identify patients who are at risk of impaired renal function:
Cockcroft & Gault derived their equation from a group of 296 males, all within 10% of LBW, leading many to question the validity of this method in obese patients. Studies have generally found that use of TBW (total body weight) tends to over-estimate CLCR in obese patients, while use of LBW (lean body weight) tends to under-estimate CLCR.17 Various weight correction factors have been proposed, each has their proponents and detractors:
An alternative for this group is the Salazar and Corcoran method which was derived from an obese patient population.16
The Cockcroft & Gault equation tends to over-estimate CLCR in the elderly.12 Therefore, an empiric "correction" commonly employed is to round up the serum creatinine to 1.0 mg% in elderly patients. However, most studies have found this to be an inappropriate practice which under-estimates true CLCR.13
Very low serum creatinine
Use of a very low serum creatinine (0.5 mg% or less) in the C&G equation leads to a falsely elevated CLCR. Therefore, many practitioners designate 0.7 mg% as the minimum SCr which should be used in the equation.
Rising serum creatinine
If the serum creatinine is rising, it is likely not at steady-state. SCr may require one week to stabilize following a decrease in renal function. Conversely, after renal function improves to normal, the shift of SCr to its new steady-state level occurs rapidly, since the new half life is now quite short. Thus, the probability that SCr may not be at steady-state is much greater when SCr is rising, than when it is falling. Jelliffe's multi step method, which corrects for rising SCr, is more accurate than C&G in patients with unstable renal function.15
As stated above, C&G derived their equation from a group of men, the 0.85 factor for women was added afterwards, to correct for the smaller muscle mass of females. One study found that a 0.9 factor for women may be more accurate.10
Serum creatinine will be affected by dietary extremes. Patients who are following an unusual vegetarian diet may have a lower SCr than expected. A diet excessively rich in red meat will lead to the reverse error.
After selecting the drug, click on the Prospective dosing tab to view the dosing table. The program asks for the current dosage. This is necessary if recommending a dosage adjustment, otherwise, if determining an initial dosage, just leave blank. Next type in your dosage recommendation.
For some drugs a one-compartment model may be selected. If a one compartment model is selected, the program calculates an ideal dose, based on the dosing weight and creatinine clearance. After the user enters a practical dose and interval, the program calculates estimated steady state peak and trough levels. If an MIC is entered, PK/PD parameters will be calculated. For more information on this subject, please see this page:
The drug models are not hard-coded into the program. The parameters are found in the drug model database and are fully user-edit able. You can tailor each drug model to fit your patient population, or you can create your own models.
ABW = LBW + [CF x (TBW - LBW)]
. . . where CF = correction factor (usually 20 to 40%)
. . . where WT = patient's total weight
Total Daily Dose
|Dose Regimen (mg)|
|>/=60||900-3600||300 TID||400 TID||600 TID||800 TID||1200 TID|
|>30-59||400-1400||200 BID||300 BID||400 BID||500 BID||700 BID|
|>15-29||200-700||200 QD||300 QD||400 QD||500 QD||700 QD|
|15a||100-300||100 QD||125 QD||150 QD||200 QD||300 QD|
|a For patients with creatinine clearance <15 mL/min, reduce daily dose in proportion to creatinine clearance (e.g., patients with a creatinine clearance of 7.5 mL/min should receive one-half the daily dose that patients with a creatinine clearance of 15 mL/min receive).|
|Drug||Method||CLCR > 50ml/min||CLCR 10 to 50ml/min||CLCR < 10ml/min|
|Acyclovir||I||5mg/kg Q 8 hr||5mg/kg Q 12 hr||5mg/kg Q 24 hr|
|Ampicillin||D, I||1-2 g Q 6 hr||0.5 g Q 6-8 hr||0.5-1g Q 12 hr|
|Aztreonam||D, I||1-2 g Q 6-8 hr||0.5-1g Q 8 hr||0.5-1g Q 12 hr|
|Cefazolin||D, I||1-2 g Q 6-8 hr||1-2 q Q 12-24 hr||1 g Q 48 hr|
|Cefotetan||D, I||1-2 g Q 12 hr||1-2 g Q 12-24 hr||0.5 - 1g Q 24 hr|
Option 2 - One compartment model
If parameters are available, a one compartment model may be employed. Model parameters for common antibiotics are listed in Table 3.25
The one feature that the RxKinetics family of pk programs have in common is the ability to edit the default drug models. You can edit any model to better fit your patient population, you can even add your own 1-compartment models for any drug and for multiple patient populations, a "Swiss army knife" for clinical pharmacokinetics if you will. Please see the following tutorial for a basic overview of how to create a one compartment model: