Scund. J . clin. Lab. Invest., Vol. 36, 1976.
Selection of Routine Method for Determination of Glomerular Filtration Rate in Adult Patients
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
J. BRUCHNER-MORTENSEN& P. RODBRO Dept. of Clinical Physiology and Medicine C (Nephrology) , Aalborg Sygehus, Aalborg, Denmark
Brochner-Mortensen, J. & Rodbro, P. Selection a. --Jutine Method --lr Determination of Glomerular Filtration Rate in Adult Patients. Scund. J. clin. Lab. Invest. 36, 3543, 1976. The precision and reproducibility of three different clearance methods as used in clinical routine assessment of glomerular filtration rate (GFR) were investigated in 51 patients : total [61Cr]EDTA plasma clearance (E); 24-hr endogenous creatinine clearance (C) ; and creatinine clearance estimated from the plasma creatinine concentration, weight, and sexand age-dependent mean creatinine excretion rate (c). The precision and reproducibility (coefficient of variation) for single determinations were, in patients with E> 30 ml/min, 5.5 and 4.1% (E); 26.9% (C); and 23.2 and 11.0% (c). The corresponding figures for E20 years; and stable renal function (i.e. patients with acute renal diseases, oedema, fever, changes in antihypertensive drug treatment, or marked unidirectional changes in plasma creatinine were not included). For the calculations the patients were divided into two groups according to values of E : a 3 0 ml/min (n=36; mean 79 ml/min, range 33-126 ml/min) and 30 ml/min are included. even as a mean of three determinations, is considerably lower than that of single determinations of E, whereas the precision is similar for the two methods (mean of three determinations of C and single determination of E) in patients with low renal function ( c j Table I). Exclusion qf single determinations. The effect of excluding one out of three determinations of C in patients with major variation in C (most deviating or highest or lowest value) on precision of mean value of the remaining two determinations is reported in Table 111. The figures given are mean differences and standard deviations of the difference between Cc and various types of C. The mean difference was nearly zero or attained the same value, negative or positiwe, for Ci, and c,. The standard deviation of difference for Ci,Ch,and c, were of the same magnitude and corresponded to that of cs,whose mean difference was also nearly zero. Exclusion of patients. The effect of excluding patients with major variation in C on precision of mean value of three determinations of C is given in Fig. 2 and is compared with the precision of a single determination of E (5.5%) for the
ch,
J. Brochner-Mortensen & P. Rodbro
40
Table IV. Precision components (mean coefficients of variation) for calculated creatinine clearance (single determinations) Total Degree of Reprodu[WrIEDTA inaccuracy cibility Precision plasma clearance vu v, I/vua+v2 (ml/min) (%I (%I (%I 30 Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
< 30
20.4 20.4
11.0 6.5
23.2 21.4
group of patients with E 2 30 ml/min. By excluding 25% of the patients, the precision of c, is about lo%, whereas nearly 75% of the patients must be excluded before the precision for mean value of three determinations of C reaches the precision of a single determination of E. Calculated creatinine clearance
The values for degree of inaccuracy, reproducibility, and resulting precision are given in Table IV. DISCUSSION The reference value for GFR is the renal plasma clearance of inulin. This method, however, is unsuitable for daily clinical practice and, furthermore, is less reproducible than newer methods without urine collection, using radioactive tracers and plasma sampling (13). The relation between inulin clearance and total [WrIEDTA plasma clearance (E) has been thoroughly investigated (1, 9, 24). E slightly underestimates GFR at high values and slightly overestimates GFR at low values, when GFR is measured simultaneously with the renal plasma clearance of inulin. Unpublished data on 39 subjects from our department, covering a wide range of renal function and in whom E and the renal plasma clearance of [WrIEDTA were performed simultaneously, showed that E on the average was 3.7 ml/min higher than renal plasma clearance in all ranges. Several investigations have shown that renal plasma clearance of [61Cr]EDTA multiplied by mean factors ranging from 1.04 to 1.12 equals a simultaneously determined renal plasma clearance of inulin (1, 8, 13, 24, 34, 35,
39). In our department the minor inaccuracy of E versus GFR is corrected for in adult patients by multiplying (E-3.7 ml/min) by 1.10. This correction holds true in nearly all conditions, the sole exception being patients with severe oedema, where E becomes increasingly higher than the renal plasma clearance of [WrIEDTA. The endogeneous creatinine clearance overestimates GFR at all levels, percentually more at low function levels (6). It has been discussed whether it is possible with reasonable certainty to correct for this inaccuracy; at any rate the degree of inaccuracy is considerably higher than that for E (3, 4, 17, 24, 28, 32), a feature that in itself makes C less reliable than E for clinical use. The precision of E has hitherto not received much attention. Chantler & Barrat (12) found a reproducibility of 3.9% in 25 determinations in 5 healthy adults, figures comparable with ours of 4.1%. However, our results concern the reproducibility in daily clinical routine, and similar studies have not to our knowledge been previously published. la the early days of clinical routine determination of 24-hr endogenous creatinine clearance the mean of three determinations was measured with a precision of 10% (lo), corresponding to a precision of 17% for single determinations. In well-instructed, selected, healthy subjects the precision of 24-hr endogenous creatinine clearance has been found to be 11-12% for single determinations (1 2, 18). However, as stressed by Holland & Whitehead (27), speaking generally of laboratory tests : ‘Too frequently precision is determined under “optimal” conditions and produces an unrealistically optimistic view of what is likely to be achieved in routine practice’. de Wardener (15) has stressed that an accurately timed collection of urine is ‘incredibly difficult to obtain’ and mentions several accidents that may vitiate the simple manoeuvre of collecting urine for a 24-hr endogenous creatinine clearance. In 119 healthy ambulatory subjects who had C determined twice after a written instruction for the accurate collection of urine (40), the mean coefficient of variation was 26.1%. This figure is similar to our results and was also heavily influenced by a few very bad urine collectors
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
Precision of GFR Measurements
(40). Another material (14) comprising 35 hospitalized patients with normal renal function who collected urine for 7 to 41 days after verbal instruction showed a mean day-to-day variation in urinary creatinine excretion rate of 19.1%. The similar figure for our material was 23.9%. Thus our data are by no means exceptional, and they stress that the determination of 24-hr endogenous creatinine clearance has a high degree of imprecision of the single determination, unless the patients are strictly supervised for several days, preferably confined to bed. The precision is of course improved by using mean values of three determinations (Table II), but the data in Table 111 show that nothing is gained by rejecting selected single determinations. Better results are obtained when patients with major variations are excluded (Fig. 2); the exclusion of 25% of subjects yields values comparable to those of Brun (10). If this precision of 10% is thought acceptable, more than three determinations must thus be done in a quarter of the patients. To avoid the problems inherent in urine collection, some authors have tried to estimate C from the plasma concentration of creatinine (11, 19, 20, 26, 37). These early simplified methods suffer from lack of regard for the parallel fall in creatinine clearance and creatinine production with age (31). They have been judged too unreliable for clinical use (18,21). Kampmann et al. (31) suggested the calculation of creatinine clearance from a patient’s actual plasma creatinine concentration and a mean creatinine excretion rate per kg body weight, regard being taken for sex and age. The present results include values for creatinine clearance calculated in this way (c). The results show that whereas the reproducibility is fairly good, the precision is no better than that of C, because a mean creatinine excretion rate for a group is used, in which the interindividual variation is large (20.4%, Table IV). Accordingly, in contrast to C, the precision can only be slightly ameliorated by repeated measurements. Furthermore, other sources of inaccuracy may hamper this approach: 1) The mean excretion rate, ii, is determined from a large number of subjects with normal values of plasma creatinine (31), thus excluding patients with chronic nephropathy
41
with low renal function. In such patients the urinary creatinine excretion rate may be lower than that found in healthy subjects of the same age (21, 23, 29, 37), and consequently c will overestimate C. 2) Under postoperative conditions (25, 33) the creatinine excretion rate is decreased on the first postoperative day, and c will overestimate C; hereafter the creatinine excretion rate is increased, leading to an underestimation of C determined as c. 3) In obese and oedematous patients, the ratio lean body mass/ body weight is decreased, and c will in principle overestimate C. 4) Finally, in conditions of rapid change in renal function, c will under- or overestimate C, when the product of clearance and plasma concentration does not equal the production rate of creatinine. The present data concerning precision indicate that E is markedly superior to the other methods at normal renal function and less so at low function, indicating the supremacy of E in assessment of renal functional state, prediction of postoperative renal function (from a preoperative renography and clearance determination), and dose calculation for potentially toxic drugs excreted via the kidneys. For the latter purpose, however, c has been found sufficiently reliable (30). Although C generally is too imprecise for routine clinical use, it should be used when E is unsuitable (severe oedema) and when c fails for dose calculations (chronic nephropathy, postoperative conditions, obese or oedematous patients, and rapid changes in renal function). Furthermore, C is the method of choice in pregnant women, in whom E should be avoided because of the radioactivity and in whom c presumably will give false values. The collection of urine, however, must be performed with meticulous care to yield meaningful data. The present data concerning reproducibility indicate that for assessing chunges in renal function E is still superior but c comes increasingly into its own for this purpose, especially at low functional levels. For daily measurements another method than E must be chosen; unless there is rapid change in renal function or the patient is investigated postoperatively - where C should be used - c is the method of choice. In conclusion, the data for precision and
42
J. Brochner-Mortensen & P. Rodbro
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
reproducibility here presented together with the cited reports on accuracy (1, 3, 4, 9, 17, 24, 28, 32) clearly indicate - with minor exceptions that the total [51Cr]EDTA plasma clearance is the best method for determination of glomerular filtration rate in clinical routine work.
REFERENCES 1. Aurell, M. & Ditzel, J. Renal clearance of W r EDTA-complex. pp. 405-413 in Proceedings of the 7th International Congress of Clinical Chemistry, Vol. 3. S . Karger, Basel & New York, 1970. 2. Bailey, R. R., Rogers, T. G. H. & Tait, ;T. J. Measurement of glomerular filtration rate using a single injection of Wr-edetic acid. Aust. Ann. Med. 3, 255, 1970. 3. Bennett, W. M. & Porter, G. A. Endogenous creatinine clearance as a clinical measure of glomerular filtration rate. Brit. med. J. 4, 84, 1971. 4. Berlyne, G. M., Nilwarangkur, S., Varley, H. & Hoerni, M. Endogenous-creatinine clearance and glomerular-filtration rate. Lancet 2, 874, 1964. 5. Bianchi, C. Measurement of the glomerular filtration rate. Progr. nucl. Med. 2, 21, 1972. 6. Brod, J. pp. 82-85 in The Kidney. Butterworth & Co., London, 1973. 7. Brochner-Mortensen, J. A simple method for the determination of glomerular filtration rate. Scand. J. clin. Lab. Invest. 30, 271, 1972. 8. Brochner-Mortensen, J. The glomerular filtration rate during moderate hyperglycemia in normal man. Acta med. scand. 194, 31, 1973. 9. Brochner-Mortensen, J., Giese, J. & Rossing, N. Renal inulin clearance versus total plasma clearance of 61Cr-EDTA. Scand. J. clin. Lab. Invest. 23, 301, 1969. 10. Brun, C. Dogncreatininclearance. Nord. Med. 48, 942, 1952. 11. Brun, C. & Munck, 0. Some clinical methods for measuring renal function. Prog. Surg. 4,2,1964. 12. Chantler, C. & Barratt, T. M. Estimation of glomerular filtration rate from plasma clearance of 51-chromium edetic acid. Arch. Dis. Childh. 47, 613, 1972. 13. Chantler, C., Garnett, E. S., Parsons, V. & Veall, N. Glomerular filtration rate measurement in man by the single injection method using WrEDTA. Clin. Sci.37, 169, 1969. 14. Cram&, K., CramCr, H. & Selander, S. A comparative analysis between variation in 24-hour urinary creatinine output and 24-hour urinary volume. Clin. chim. Acta 15, 331, 1967. 15. de Wardener, H. E. pp. 34-35 in The Kidney. Churchill Livingstone, Edinburgh & London, 1973. 16. Ditzel, J., Vestergaard, P. & Brinklov, M. Den
glomerulrere filtrationshastighed bestemt med Wr-EDTA. Nord. Med. 85, 650, 1971. 17. Dodge, W. F., Travis, L. B. & Daeschner, C. W. Comparison of endogenous creatinine clearance with inulin clearance. Amer. J. Dis. Child. 113, 683, 1967. 18. Doolan, P. D., Alpen, E. L. & Theill, G. B. A clinical appraisal of the plasma concentration and endogenous clearance of creatinine. Amer. J. Med. 32, 65, 1962. 19. Edwards, K. D. G. & Whyte, H. M. Plasma creatinine level and creatinine clearance as test of renal function. Aust. Ann. Men. 8, 218, 1959. 20. Effersoe, P. Relationship between endogenous 24-hour creatinine clearance and serum creatinine concentration in patients with chronic renal disease. Acta med. scand. 156, 429, 1957. 21. Enger, E. & Blegen, E. M. The relationship between endogenous creatinine clearance and serum creatinine in renal failure. Scand. J . clin. Lab. Invest. 16, 273, 1964. 22. Garnett, E. S., Parsons, V. & Veall, N. hleasurement of glomerular filtration rate in man using a 51Cr/edetic-acid complex. Lancet I , 818, 1967. 23. Goldman, R. Creatinine excretion in renal failure. Proc. Soc. exp. Biol. (N.Y.) 85, 446, 1954. 24.Hagstam, K. E., Nordenfelt, I., Svensson, L. & Svensson, S. E. Comparison of different methods for determination of glomerular filtration rate in renal disease. Scand. J. d i n . Lab. Invest. 34, 31, 1974. 25. Hasner, E., Carlsen, K., Engdahl, I., Klrerke, N., Paaby, H., Schultze, B., Sprechler, M., Sorensen, B. & Tobiassen, T. pp. 57-65 in Adrenals in Surgery. E. Munksgaard, Copenhagen, 1952. 26. Haugen, H. N. & Blegen, E. M. Plasma creatinine concentration and creatinine clearance in clinical work. Ann. intern. Med. 43, 731, 1955. 27. Holland, W. W. &Whitehead, T. P. Value of new laboratory tests in diagnosis and treatment. Lancet 2, 391, 1974. 28. Hood, B., Attman, P. O., AhlmCn, J. & Jagenburg, R. Renal hemodynamics and limitations of creatinine clearance in determining filtration rate in glomerular disease. Scand. J. Urol. Nkphrol. 5, 154, 1971. 29. Jones, J. D. & Burnett, P. C. Creatinine metabolism in humans with decreased renal function: creatinine deficit. Clin. Chem. 20, 1204, 1974. 30. Kampmann, J. P., Kristensen, M., SiersbrekNielsen, K., Christiansen, N. J. B., Lumholtz, B., Kolendorf, K. & Molholm Hansen, J. Doseringsforslag ti1 behandling med kanamycin, gentamycin og digoksin. Ugeskr. Lag, 137, 1, 1975. 31. Kampmann, J. P., Siersbrek-Nielsen, K., Kristensen, M. & Molholm Hansen, J. Aldersbetingede variationer i urinkreatinin og endogen kreatininclearance. Ugeskr. Lag. 133, 2369, 1971.
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
Precision of GFR Measurements 32. Kim, K. E., Onesti, G., Ramirez, O., Brest, A. N. & Swartz, C. Creatinine clearance in renal disease. A reappraisal. Brit. med. J. 4, 11, 1969. 33. Kragelund, E. Renal function after major surgery assessed on the basis of the 24-hour creatinine clearance. Acta chir. scand. 117, 416, 1959. 34. Lavender, S., Hilton, P. J. & Jones, N. F. The measurement of glomerular filtration-rate in renal disease. Lancet 2, 1216, 1969. 35. Lingirdh, G. Renal clearance investigations with Wr-EDTA and la6I-hippuran. Scand. J. Urol. Nephrol 6, 63, 1972. 36. Sederberg-Olsen, P., Siersbaek-Nielsen, K. & Ibsen, H. Den glomerulaere filtration bedomt ved Wr-EDTA clearance efter engangsinjektion. Ugeskr. Leg. 132, 2438, 1970. Received 4 April 1975 Accepted 6 October 1975
43
37. Steinitz, K. & Tiirkand, H. The determination of the glomerular filtration by the endogenous creatinine clearance. J. clin. Invest. 19, 285, 1940. 38. Truninger, B., Donath, A. & Kappeler, M. Simnlified clearance techniques. The single injection method and its modifications. Helv. med. Acta. 34, 116, 1968. 39. Vorburger, C., Riedwyl, H. & Reubi, F. Vergleichende studien zwischen den renalen clearances von Na-Cr,Cr51-athylendiamintetraacetat,inulin und nathriumthiosulfat beim Menschen. Klin. Wschr. 47,415, 1969. 40. Wibell, L. & Bjorsell-t)stling, E. Endogenous creatinine clearance in apparently healthy individuals as determined by 24 hour ambulatory urine collection. Upsala J. med. Sci. 78,43, 1973.
Selection of Routine Method for Determination of Glomerular Filtration Rate in Adult Patients
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
J. BRUCHNER-MORTENSEN& P. RODBRO Dept. of Clinical Physiology and Medicine C (Nephrology) , Aalborg Sygehus, Aalborg, Denmark
Brochner-Mortensen, J. & Rodbro, P. Selection a. --Jutine Method --lr Determination of Glomerular Filtration Rate in Adult Patients. Scund. J. clin. Lab. Invest. 36, 3543, 1976. The precision and reproducibility of three different clearance methods as used in clinical routine assessment of glomerular filtration rate (GFR) were investigated in 51 patients : total [61Cr]EDTA plasma clearance (E); 24-hr endogenous creatinine clearance (C) ; and creatinine clearance estimated from the plasma creatinine concentration, weight, and sexand age-dependent mean creatinine excretion rate (c). The precision and reproducibility (coefficient of variation) for single determinations were, in patients with E> 30 ml/min, 5.5 and 4.1% (E); 26.9% (C); and 23.2 and 11.0% (c). The corresponding figures for E20 years; and stable renal function (i.e. patients with acute renal diseases, oedema, fever, changes in antihypertensive drug treatment, or marked unidirectional changes in plasma creatinine were not included). For the calculations the patients were divided into two groups according to values of E : a 3 0 ml/min (n=36; mean 79 ml/min, range 33-126 ml/min) and 30 ml/min are included. even as a mean of three determinations, is considerably lower than that of single determinations of E, whereas the precision is similar for the two methods (mean of three determinations of C and single determination of E) in patients with low renal function ( c j Table I). Exclusion qf single determinations. The effect of excluding one out of three determinations of C in patients with major variation in C (most deviating or highest or lowest value) on precision of mean value of the remaining two determinations is reported in Table 111. The figures given are mean differences and standard deviations of the difference between Cc and various types of C. The mean difference was nearly zero or attained the same value, negative or positiwe, for Ci, and c,. The standard deviation of difference for Ci,Ch,and c, were of the same magnitude and corresponded to that of cs,whose mean difference was also nearly zero. Exclusion of patients. The effect of excluding patients with major variation in C on precision of mean value of three determinations of C is given in Fig. 2 and is compared with the precision of a single determination of E (5.5%) for the
ch,
J. Brochner-Mortensen & P. Rodbro
40
Table IV. Precision components (mean coefficients of variation) for calculated creatinine clearance (single determinations) Total Degree of Reprodu[WrIEDTA inaccuracy cibility Precision plasma clearance vu v, I/vua+v2 (ml/min) (%I (%I (%I 30 Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
< 30
20.4 20.4
11.0 6.5
23.2 21.4
group of patients with E 2 30 ml/min. By excluding 25% of the patients, the precision of c, is about lo%, whereas nearly 75% of the patients must be excluded before the precision for mean value of three determinations of C reaches the precision of a single determination of E. Calculated creatinine clearance
The values for degree of inaccuracy, reproducibility, and resulting precision are given in Table IV. DISCUSSION The reference value for GFR is the renal plasma clearance of inulin. This method, however, is unsuitable for daily clinical practice and, furthermore, is less reproducible than newer methods without urine collection, using radioactive tracers and plasma sampling (13). The relation between inulin clearance and total [WrIEDTA plasma clearance (E) has been thoroughly investigated (1, 9, 24). E slightly underestimates GFR at high values and slightly overestimates GFR at low values, when GFR is measured simultaneously with the renal plasma clearance of inulin. Unpublished data on 39 subjects from our department, covering a wide range of renal function and in whom E and the renal plasma clearance of [WrIEDTA were performed simultaneously, showed that E on the average was 3.7 ml/min higher than renal plasma clearance in all ranges. Several investigations have shown that renal plasma clearance of [61Cr]EDTA multiplied by mean factors ranging from 1.04 to 1.12 equals a simultaneously determined renal plasma clearance of inulin (1, 8, 13, 24, 34, 35,
39). In our department the minor inaccuracy of E versus GFR is corrected for in adult patients by multiplying (E-3.7 ml/min) by 1.10. This correction holds true in nearly all conditions, the sole exception being patients with severe oedema, where E becomes increasingly higher than the renal plasma clearance of [WrIEDTA. The endogeneous creatinine clearance overestimates GFR at all levels, percentually more at low function levels (6). It has been discussed whether it is possible with reasonable certainty to correct for this inaccuracy; at any rate the degree of inaccuracy is considerably higher than that for E (3, 4, 17, 24, 28, 32), a feature that in itself makes C less reliable than E for clinical use. The precision of E has hitherto not received much attention. Chantler & Barrat (12) found a reproducibility of 3.9% in 25 determinations in 5 healthy adults, figures comparable with ours of 4.1%. However, our results concern the reproducibility in daily clinical routine, and similar studies have not to our knowledge been previously published. la the early days of clinical routine determination of 24-hr endogenous creatinine clearance the mean of three determinations was measured with a precision of 10% (lo), corresponding to a precision of 17% for single determinations. In well-instructed, selected, healthy subjects the precision of 24-hr endogenous creatinine clearance has been found to be 11-12% for single determinations (1 2, 18). However, as stressed by Holland & Whitehead (27), speaking generally of laboratory tests : ‘Too frequently precision is determined under “optimal” conditions and produces an unrealistically optimistic view of what is likely to be achieved in routine practice’. de Wardener (15) has stressed that an accurately timed collection of urine is ‘incredibly difficult to obtain’ and mentions several accidents that may vitiate the simple manoeuvre of collecting urine for a 24-hr endogenous creatinine clearance. In 119 healthy ambulatory subjects who had C determined twice after a written instruction for the accurate collection of urine (40), the mean coefficient of variation was 26.1%. This figure is similar to our results and was also heavily influenced by a few very bad urine collectors
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
Precision of GFR Measurements
(40). Another material (14) comprising 35 hospitalized patients with normal renal function who collected urine for 7 to 41 days after verbal instruction showed a mean day-to-day variation in urinary creatinine excretion rate of 19.1%. The similar figure for our material was 23.9%. Thus our data are by no means exceptional, and they stress that the determination of 24-hr endogenous creatinine clearance has a high degree of imprecision of the single determination, unless the patients are strictly supervised for several days, preferably confined to bed. The precision is of course improved by using mean values of three determinations (Table II), but the data in Table 111 show that nothing is gained by rejecting selected single determinations. Better results are obtained when patients with major variations are excluded (Fig. 2); the exclusion of 25% of subjects yields values comparable to those of Brun (10). If this precision of 10% is thought acceptable, more than three determinations must thus be done in a quarter of the patients. To avoid the problems inherent in urine collection, some authors have tried to estimate C from the plasma concentration of creatinine (11, 19, 20, 26, 37). These early simplified methods suffer from lack of regard for the parallel fall in creatinine clearance and creatinine production with age (31). They have been judged too unreliable for clinical use (18,21). Kampmann et al. (31) suggested the calculation of creatinine clearance from a patient’s actual plasma creatinine concentration and a mean creatinine excretion rate per kg body weight, regard being taken for sex and age. The present results include values for creatinine clearance calculated in this way (c). The results show that whereas the reproducibility is fairly good, the precision is no better than that of C, because a mean creatinine excretion rate for a group is used, in which the interindividual variation is large (20.4%, Table IV). Accordingly, in contrast to C, the precision can only be slightly ameliorated by repeated measurements. Furthermore, other sources of inaccuracy may hamper this approach: 1) The mean excretion rate, ii, is determined from a large number of subjects with normal values of plasma creatinine (31), thus excluding patients with chronic nephropathy
41
with low renal function. In such patients the urinary creatinine excretion rate may be lower than that found in healthy subjects of the same age (21, 23, 29, 37), and consequently c will overestimate C. 2) Under postoperative conditions (25, 33) the creatinine excretion rate is decreased on the first postoperative day, and c will overestimate C; hereafter the creatinine excretion rate is increased, leading to an underestimation of C determined as c. 3) In obese and oedematous patients, the ratio lean body mass/ body weight is decreased, and c will in principle overestimate C. 4) Finally, in conditions of rapid change in renal function, c will under- or overestimate C, when the product of clearance and plasma concentration does not equal the production rate of creatinine. The present data concerning precision indicate that E is markedly superior to the other methods at normal renal function and less so at low function, indicating the supremacy of E in assessment of renal functional state, prediction of postoperative renal function (from a preoperative renography and clearance determination), and dose calculation for potentially toxic drugs excreted via the kidneys. For the latter purpose, however, c has been found sufficiently reliable (30). Although C generally is too imprecise for routine clinical use, it should be used when E is unsuitable (severe oedema) and when c fails for dose calculations (chronic nephropathy, postoperative conditions, obese or oedematous patients, and rapid changes in renal function). Furthermore, C is the method of choice in pregnant women, in whom E should be avoided because of the radioactivity and in whom c presumably will give false values. The collection of urine, however, must be performed with meticulous care to yield meaningful data. The present data concerning reproducibility indicate that for assessing chunges in renal function E is still superior but c comes increasingly into its own for this purpose, especially at low functional levels. For daily measurements another method than E must be chosen; unless there is rapid change in renal function or the patient is investigated postoperatively - where C should be used - c is the method of choice. In conclusion, the data for precision and
42
J. Brochner-Mortensen & P. Rodbro
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
reproducibility here presented together with the cited reports on accuracy (1, 3, 4, 9, 17, 24, 28, 32) clearly indicate - with minor exceptions that the total [51Cr]EDTA plasma clearance is the best method for determination of glomerular filtration rate in clinical routine work.
REFERENCES 1. Aurell, M. & Ditzel, J. Renal clearance of W r EDTA-complex. pp. 405-413 in Proceedings of the 7th International Congress of Clinical Chemistry, Vol. 3. S . Karger, Basel & New York, 1970. 2. Bailey, R. R., Rogers, T. G. H. & Tait, ;T. J. Measurement of glomerular filtration rate using a single injection of Wr-edetic acid. Aust. Ann. Med. 3, 255, 1970. 3. Bennett, W. M. & Porter, G. A. Endogenous creatinine clearance as a clinical measure of glomerular filtration rate. Brit. med. J. 4, 84, 1971. 4. Berlyne, G. M., Nilwarangkur, S., Varley, H. & Hoerni, M. Endogenous-creatinine clearance and glomerular-filtration rate. Lancet 2, 874, 1964. 5. Bianchi, C. Measurement of the glomerular filtration rate. Progr. nucl. Med. 2, 21, 1972. 6. Brod, J. pp. 82-85 in The Kidney. Butterworth & Co., London, 1973. 7. Brochner-Mortensen, J. A simple method for the determination of glomerular filtration rate. Scand. J. clin. Lab. Invest. 30, 271, 1972. 8. Brochner-Mortensen, J. The glomerular filtration rate during moderate hyperglycemia in normal man. Acta med. scand. 194, 31, 1973. 9. Brochner-Mortensen, J., Giese, J. & Rossing, N. Renal inulin clearance versus total plasma clearance of 61Cr-EDTA. Scand. J. clin. Lab. Invest. 23, 301, 1969. 10. Brun, C. Dogncreatininclearance. Nord. Med. 48, 942, 1952. 11. Brun, C. & Munck, 0. Some clinical methods for measuring renal function. Prog. Surg. 4,2,1964. 12. Chantler, C. & Barratt, T. M. Estimation of glomerular filtration rate from plasma clearance of 51-chromium edetic acid. Arch. Dis. Childh. 47, 613, 1972. 13. Chantler, C., Garnett, E. S., Parsons, V. & Veall, N. Glomerular filtration rate measurement in man by the single injection method using WrEDTA. Clin. Sci.37, 169, 1969. 14. Cram&, K., CramCr, H. & Selander, S. A comparative analysis between variation in 24-hour urinary creatinine output and 24-hour urinary volume. Clin. chim. Acta 15, 331, 1967. 15. de Wardener, H. E. pp. 34-35 in The Kidney. Churchill Livingstone, Edinburgh & London, 1973. 16. Ditzel, J., Vestergaard, P. & Brinklov, M. Den
glomerulrere filtrationshastighed bestemt med Wr-EDTA. Nord. Med. 85, 650, 1971. 17. Dodge, W. F., Travis, L. B. & Daeschner, C. W. Comparison of endogenous creatinine clearance with inulin clearance. Amer. J. Dis. Child. 113, 683, 1967. 18. Doolan, P. D., Alpen, E. L. & Theill, G. B. A clinical appraisal of the plasma concentration and endogenous clearance of creatinine. Amer. J. Med. 32, 65, 1962. 19. Edwards, K. D. G. & Whyte, H. M. Plasma creatinine level and creatinine clearance as test of renal function. Aust. Ann. Men. 8, 218, 1959. 20. Effersoe, P. Relationship between endogenous 24-hour creatinine clearance and serum creatinine concentration in patients with chronic renal disease. Acta med. scand. 156, 429, 1957. 21. Enger, E. & Blegen, E. M. The relationship between endogenous creatinine clearance and serum creatinine in renal failure. Scand. J . clin. Lab. Invest. 16, 273, 1964. 22. Garnett, E. S., Parsons, V. & Veall, N. hleasurement of glomerular filtration rate in man using a 51Cr/edetic-acid complex. Lancet I , 818, 1967. 23. Goldman, R. Creatinine excretion in renal failure. Proc. Soc. exp. Biol. (N.Y.) 85, 446, 1954. 24.Hagstam, K. E., Nordenfelt, I., Svensson, L. & Svensson, S. E. Comparison of different methods for determination of glomerular filtration rate in renal disease. Scand. J. d i n . Lab. Invest. 34, 31, 1974. 25. Hasner, E., Carlsen, K., Engdahl, I., Klrerke, N., Paaby, H., Schultze, B., Sprechler, M., Sorensen, B. & Tobiassen, T. pp. 57-65 in Adrenals in Surgery. E. Munksgaard, Copenhagen, 1952. 26. Haugen, H. N. & Blegen, E. M. Plasma creatinine concentration and creatinine clearance in clinical work. Ann. intern. Med. 43, 731, 1955. 27. Holland, W. W. &Whitehead, T. P. Value of new laboratory tests in diagnosis and treatment. Lancet 2, 391, 1974. 28. Hood, B., Attman, P. O., AhlmCn, J. & Jagenburg, R. Renal hemodynamics and limitations of creatinine clearance in determining filtration rate in glomerular disease. Scand. J. Urol. Nkphrol. 5, 154, 1971. 29. Jones, J. D. & Burnett, P. C. Creatinine metabolism in humans with decreased renal function: creatinine deficit. Clin. Chem. 20, 1204, 1974. 30. Kampmann, J. P., Kristensen, M., SiersbrekNielsen, K., Christiansen, N. J. B., Lumholtz, B., Kolendorf, K. & Molholm Hansen, J. Doseringsforslag ti1 behandling med kanamycin, gentamycin og digoksin. Ugeskr. Lag, 137, 1, 1975. 31. Kampmann, J. P., Siersbrek-Nielsen, K., Kristensen, M. & Molholm Hansen, J. Aldersbetingede variationer i urinkreatinin og endogen kreatininclearance. Ugeskr. Lag. 133, 2369, 1971.
Scand J Clin Lab Invest Downloaded from informahealthcare.com by Nyu Medical Center on 06/22/15 For personal use only.
Precision of GFR Measurements 32. Kim, K. E., Onesti, G., Ramirez, O., Brest, A. N. & Swartz, C. Creatinine clearance in renal disease. A reappraisal. Brit. med. J. 4, 11, 1969. 33. Kragelund, E. Renal function after major surgery assessed on the basis of the 24-hour creatinine clearance. Acta chir. scand. 117, 416, 1959. 34. Lavender, S., Hilton, P. J. & Jones, N. F. The measurement of glomerular filtration-rate in renal disease. Lancet 2, 1216, 1969. 35. Lingirdh, G. Renal clearance investigations with Wr-EDTA and la6I-hippuran. Scand. J. Urol. Nephrol 6, 63, 1972. 36. Sederberg-Olsen, P., Siersbaek-Nielsen, K. & Ibsen, H. Den glomerulaere filtration bedomt ved Wr-EDTA clearance efter engangsinjektion. Ugeskr. Leg. 132, 2438, 1970. Received 4 April 1975 Accepted 6 October 1975
43
37. Steinitz, K. & Tiirkand, H. The determination of the glomerular filtration by the endogenous creatinine clearance. J. clin. Invest. 19, 285, 1940. 38. Truninger, B., Donath, A. & Kappeler, M. Simnlified clearance techniques. The single injection method and its modifications. Helv. med. Acta. 34, 116, 1968. 39. Vorburger, C., Riedwyl, H. & Reubi, F. Vergleichende studien zwischen den renalen clearances von Na-Cr,Cr51-athylendiamintetraacetat,inulin und nathriumthiosulfat beim Menschen. Klin. Wschr. 47,415, 1969. 40. Wibell, L. & Bjorsell-t)stling, E. Endogenous creatinine clearance in apparently healthy individuals as determined by 24 hour ambulatory urine collection. Upsala J. med. Sci. 78,43, 1973.
