Corrected creatinine reference intervals

Pål Rustad

(Updated 13/3-2006)

According to "LETTER TO THE EDITOR. Creatininium reference intervals for corrected methods. Mårtensson A, Rustad P, Lund H, Ossowicki H. Scand J Clin Lab Invest 2004; 64, p 439-442 it is proposed new reference intervals for creatinine:

• Female: 45-90 µmol/L
• Male: 60-105 µmol/L

(i.e. the lower reference limit for women is decreased with 5 and the upper limit for men is increased with 5 µmol/L releative to the original NORIP suggestion)

SFKK has officially accepted the new suggestion.

These reference intervals applies to:

• Enzymatic methods
• Vitros measurement values corrected according to Ortho: New value= Old value x 1.02 - 8.1 µmol/L
• Roche compensated Jaffé method
• Other Jaffé methods corrected by using creatinine free serum (CFS) (may be purchased from EQUALIS, see certificate here).

There may be several ways to correct measurement values for the last group after having measured CFS (mean value Ko) and X (mean value Vold):

• Use -Ko as the calibrator value for the "0"-calibrator of your measurement system. If your high calibrator has a target value traceable to a reference method (and the non-creatinine chromogens in the calbrator is about the same as in CFS), there should be little bias when measuring X (see method recommended in certificate).
• If the correction have to be made using slope and intercept according to Vnew = Vold * Slope + Intercept where Vnew and Vold is values before and after correction for an arbitrary sample, these can be calculated as (see "Calculations" below):
  • Slope = Tc/(Tc - f*Ko) where Tc is target value for local high calibrator and f is the fraction of non-creatinine chromogens of the calibrator relative to the creatinine-free serum. If the calibrator is of aquous type f is probably 0 (and slope = 1 and intercept is -Ko), but if the matrix is serum-like, one can expect a value near 1.
  • Intercept = -Tc*Ko/(Tc - f*Ko)

If the bias for X after this correction is still considerable, the reasons might be wrong value of f or Tc.

A split sample comparison with an enzymatic method covering the whole measurement range will help you to confirm your corrections to the Jaffé method.

Calculations (equations are numbered within [])

Vold= Ap*Tc/Ac [1] where Ap is total absorbance for a sample and Ac is total absorbance for calibrator.

Vnew = (Ap- Io)*Tc/(Ac-Is) [2] where Io is absorbance of creatinine free serum and Is is absorbance of non-creatinine chromogens in calibrater.

When creatinine free serum is measured with uncorrected method, the relation between the absorbance Io and the measured concentration Ko is:

Ko=Io*Tc/Ac [3]

With [1] and [3] in [2] to eliminate the unknown Ap and Ac, we get:

Vnew = (Vold-Ko)*Tc/(Tc - Ko*Is/Io). With the abbreviation f = Is/Io inserted, we get:

Slope = Tc/(Tc - Ko*f)

Intercept = -Ko*Slope

The fraction f = Is/Io is a measure of non-creatinine chromogens in the calibrator relative to that in the creatinine-free serum. If the calibrator do not contain non-creatinine chromogens (i.e. aquous matrix), f=0. If the calibrator has a serum matrix, f is expected to have a value near 1.