Article in press - uncorrected proof

Clin Chem Lab Med 2010;48(8):11131119 2010 by Walter de Gruyter Berlin New York. DOI 10.1515/CCLM.2010.238

Assessment of the Nova StatSensor whole blood point-ofcare creatinine analyzer for the measurement of kidney function in screening for chronic kidney disease

Mark Shephard1,*, Michael Peake2, Olivia Corso3, Anne Shephard1, Beryl Mazzachi1, Brooke Spaeth1, Jeffrey Barbara3 and Timothy Mathew4
1

Community Point-of-Care Services, Flinders University, Adelaide, South Australia, Australia 2 Express Laboratory, SA Pathology, Flinders Medical Centre, Adelaide, South Australia, Australia 3 Renal Unit, Flinders Medical Centre, Adelaide, South Australia, Australia 4 Kidney Health Australia, Adelaide, South Australia, Australia

glomerular filtration rate -60 mL/min, but there will be many falsely low estimated glomerular filtration rate results that require laboratory validation. Creatinine results need improvement. Clin Chem Lab Med 2010;48:11139. Keywords: chronic kidney disease; creatinine; estimated glomerular filtration rate; POCT; renal function.

Introduction
Chronic kidney disease (CKD) has a prevalence of approximately 16% and 13% in Australian and American environments, respectively (1, 2). The disease is usually silent and progressive, and end-stage renal disease is placing increasing burdens on health care budgets, with increasing numbers of patients requiring dialysis (3). Early signs of CKD include proteinuria, increased blood pressure and reduced glomerular filtration rate (GFR) (4). GFR can be estimated (eGFR) using laboratory measurements of serum creatinine, and considerable international efforts have been made to align creatinine results to isotope dilution mass spectrometry (IDMS) equivalent standards (5). In addition, simplified equations to convert serum creatinine results to eGFR based on age, gender, and ethnic background have now been adopted (57). Although efforts to align the calibration of laboratory creatinine estimations are well-advanced (8), there can be valid reasons for measuring creatinine in non-laboratory situations. These include screening programs for CKD to facilitate the early detection and follow-up of at-risk patients. Recently, Kidney Health Australia conducted a targeted communitybased program for CKD risk called Kidney Evaluation for You (KEY). This pilot program was the first program for CKD risk assessment undertaken in the primary health care setting in Australia (9). The KEY study used the i-STAT point-of-care testing (POCT) analyzer (Abbott Point-of-Care Inc., Princeton, NJ, USA) for measuring creatinine, with subsequent calculation of eGFR. However, the i-STAT device required a venous whole blood sample of approximately 100 mL, which is less ideal than a fingerprick sample for a screening situation. In addition, i-STAT has been reported to produce higher creatinine results than the Roche enzymatic creatinine assay (Roche Diagnostics, Sydney, Australia) (10). A recently released point-of-care device from Nova Biomedical (Waltham, MA, USA) that measures creatinine using just 1.2 mL of whole blood and converts creatinine results

Abstract
Background: Point-of-care testing for creatinine using a fingerprick sample and resultant estimated glomerular filtration rate has potential for screening for chronic kidney disease in community settings. This study assessed the applicability of the Nova StatSensor creatinine analyzer for this purpose. Methods: Fingerprick samples from 100 patients (63 renal, 37 healthy volunteers; range 46962 mmol/L) were assayed using two StatSensor analyzers. Lithium heparin venous plasma samples collected simultaneously were assayed in duplicate using the isotope dilution mass spectrometryaligned Roche Creatinine Plus enzymatic assay on a Hitachi Modular P unit. Method comparison statistics and the ability of the StatSensor to correctly categorise estimated glomerular filtration rate above or below 60 mL/min were calculated pre- and post-alignment with the laboratory method. Results: StatSensor 1 creatinine results (y) were much lower than the laboratory (ys0.75xq10.2, average bias 47.3, 95% limits of agreement 208 to q113 mmol/L). For estimated glomerular filtration rates above or below 60 mL/min, 100% and 87% of results respectively agreed with the laboratory estimated glomerular filtration rate (79% and 96% post-alignment). StatSensor 2 statistics were similar. The 95% limits of agreement between StatSensor creatinine results were 35 to q34 mmol/L. Conclusions: Isotope dilution mass spectrometry alignment of the StatSensor will identify most patients with estimated
*Corresponding author: Mark Shephard, Associate Professor, Director and Senior Research Fellow, Community Point-of-Care Services, Flinders University Rural Clinical School, Flinders University, GPO Box 2100, Adelaide, South Australia 5001, Australia E-mail: Mark.Shephard@flinders.edu.au Received December 22, 2009; accepted March 1, 2010; previously published online May 19, 2010

2010/686

Article in press - uncorrected proof

1114 Shephard et al.: Assessment of Nova StatSensor creatinine analyzer

to eGFR has been actively promoted to fill a niche in the POCT market. We evaluated the performance of this device against the IDMS-aligned Roche enzymatic creatinine assay, and assessed the potential use of the Nova device for detecting silent kidney disease in the community. The results of this study are also relevant for radiology patients using potentially nephrotoxic contrast media (11) and in other POCT environments.

ular P unit. The performance of this assay has been validated vs. both the IDMS reference method and international reference materials (SRM 967) (1215).

Imprecision
Imprecision (coefficient of variation, CV%) for creatinine measurement on the Nova StatSensor device was assessed in three ways. Within-run and day-to-day imprecision were calculated using repeated analysis (ns10 and 20, respectively) of three levels of Nova StatSensor quality control (QC) material (Cat No 43921-3; QC lot numbers 5008340241, 5008100242 and 5008344243 for within-day and 5009037241, 5009037242 and 5009043243 for dayto-day). Between-device imprecision was calculated from the difference between results obtained on the same samples analyzed on the two Nova devices wusing the equation ss6(Sd2/2n), where ssstandard deviation, dsdifference between individual results on the two devices and nsnumber of duplicates (98 in this data set) and CV% ss/m=100 where ssstandard deviation and msmean creatinine concentrationx.

Materials and methods

Ethics approval
Ethics approval to conduct this study was obtained from the Flinders Clinical Research Ethics Committee (application number 222/08).

Patient samples
One hundred subjects (48 males and 52 females) participated in the study; 63 were patients attending either the renal clinic or dialysis clinic at the Renal Unit, Flinders Medical Centre (FMC), and 37 subjects were healthy volunteers. Capillary whole blood specimens were obtained from each subject and immediately analyzed in singlicate with two Nova StatSensor creatinine devices using the same reagent strip lot number. A venous whole blood specimen anticoagulated with lithium heparin (Greiner blood tube, Greiner Labortechnik GmbH, Cat No 456083; Kremsmuenster, Austria) was obtained from each subject at the same time and sent to the pathology laboratory at FMC, Adelaide, South Australia. In the laboratory, the venous whole blood sample was centrifuged (4500 g for 5 min) and a plasma sample aliquoted for duplicate laboratory analysis.

Linearity
Linearity of the Nova analyzers was assessed by increasing the creatinine concentration of a base pool of lithium heparin anticoagulated venous whole blood (58 mmol/L) by 1000 mmol/L using a concentrated solution of creatinine prepared from National Institute of Standards and Technology Standard Reference Material (NIST SRM) 914a (NIST, United States Department of Commerce). By mixing the base pool and the spiked sample in various ratios, whole blood samples were prepared in which the base pool was supplemented with 100, 250, 500, 750 and 1000 mmol/L of creatinine. All samples were assayed with both Nova 1 and Nova 2 analyzers, and linearity assessed using Clinical and Laboratory Standards Institute (CLSI) EP6-A guidelines (16) (see Statistical analyses).

Test method Accuracy

The Nova Biomedical StatSensor creatinine meter measured creatinine in 1.2 mL of whole blood in 30 s. The sample was added to a reagent strip which was inserted into the device prior to sample application. In the reagent strip, creatinine is converted to hydrogen peroxide in an enzymatic cascade involving creatininase, creatinase and sarcosine oxidase. The signal generated from H2O2 was detected amperometrically. Calibration was factory-encoded into the reagent strip. Fingerprick analyses were conducted according to manufacturer directions by a non-laboratory operator trained by Nova Biomedical. The accuracy of creatinine results obtained with each Nova StatSensor device was compared to the mean of duplicate creatinine results from the IDMS-aligned laboratory method using PassingBablok linear regression analysis (17). Differences between results were graphed against the Roche enzymatic assay using a modified Bland-Altman difference plot (18). eGFR on the Nova StatSensor, calculated automatically from the measured creatinine using the modification of diet in renal disease (MDRD) equation (factor 186), was also plotted against eGRF from the laboratory method, calculated by the laboratory information system (LIS) from the measured creatinine using the standardised MDRD equation (factor 175). The ability of the Nova StatSensor to correctly categorise eGFR above or below 60 mL/min was then assessed by calculation of sensitivity, specificity, positive and negative predictive values.

Comparison method
Creatinine was also measured in the laboratory by assaying plasma from each patient in duplicate using the IDMS-aligned Roche Creatinine Plus enzymatic assay (Cat No 1775685) with a Hitachi ModTable 1 Nova StatSensor day-to-day method imprecision (ns20). QC Target Acceptable range 44124 115230 398663 Mean Device 1 99.6 199.6 605.4 Device 2 100.8 195.5 601.0 SD

CV% Device 2 8.97 16.09 31.28 Device 1 8.9 8.7 5.4 Device 2 8.9 8.2 5.2

Range Device 1 83123 157237 543665 Device 2 84123 158232 532668

Device 1 8.89 17.31 32.65

Low QC Mid QC High QC

84 173 531

Creatinine units are mmol/L. Devices 1 and 2 are two separate Nova analyzers. Acceptable range is that stated by the manufacturer.

Table 2 Method correlation statistics for the Nova (y) vs. the IDMS-aligned Roche enzymatic creatinine method (x) before and after IDMS alignment. PB slope (95% Cl) PB intercept (95% Cl) r Mean (x) (range) Mean (y) (range) Mean bias (95% CI) 95% Limits of agreement n

Nova (y)

Creatinine concentration, mmol/L 0.83 0.84 0.97 0.97

Factory calibration Nova 1

-150

35.9 to 21.2 to 3.6) 34.8 to 21.3 to 3.1) 207.7 to 113.0 to 31.1) 213.7 to 120.8 to 29.3) 30.1 to 38.5 27.5 to 37.6 105.2 to 96.5 111.3 to 100.3

62 62 100 98

Nova 2

-150

Nova 1

All

Nova 2

All

0.96 (0.821.11) 0.92 (0.771.06) 0.75 (0.700.79) 0.74 (0.690.79) 0.83 0.84 0.97 0.97

3.5 (14.5 to 7.5) 2.4 (9.1 to 12.8) 10.2 (6.2 to 17.4) 14.0 (8.0 to 20.5)

82.6 (46144) 82.6 (46144) 217.0 (46962) 209.8 (46962)

75.2 (32131) 75.9 (31125) 169.6 (32566) 163.4 (31545)

7.3 (11.0 6.7 (10.3 47.3 (63.6 46.5 (63.6

After recalibration Nova 1

-150

62 62 100 98

Nova 2

-150

Article in press - uncorrected proof

Nova 1

All

Shephard et al.: Assessment of Nova StatSensor creatinine analyzer 1115

Nova 2

All

1.31 (1.121.51) 1.25 (1.081.44) 1.00 (0.941.06) 0.99 (0.931.05)

19.9 (36.0 to 5.7) 12.3 (27.4 to 0.3) 0.4 (5.7 to 9.1) 4.5 (3.7 to 12.7)

82.6 (46144) 82.6 (46144) 217.0 (46962) 209.8 (46962)

86.8 (29161) 87.6 (28153) 212.6 (29741) 204.3 (28713)

4.2 (0.2 to 8.7) 5.0 (0.8 to 9.3) 4.3 (14.5 to 5.9) 5.5 (16.4 to 5.3)

Article in press - uncorrected proof

1116 Shephard et al.: Assessment of Nova StatSensor creatinine analyzer

After alignment of the Nova StatSensor results to the laboratory creatinine method, this process was repeated, now using an eGFR factor of 175 for the Nova StatSensor.

Statistical analyses
Statistical analyses, including assessment of linearity, were performed using the statistical package Analyse-it (clinical laboratory version 2.21).

Results
Imprecision

Within-run and day-to-day imprecision averaged 3.3% and 8.9%, respectively at 100 mmol/L creatinine, and 2.8% and 5.3% at 600 mmol/L creatinine for the two StatSensor analyzers (Table 1). For the laboratory assay, year-long imprecision (approx. 1200 QC data points) was 1.9% and 1.4% at similar low and high concentrations of creatinine. The imprecision of the Roche enzymatic assay was consistent with data reported during the use of this method to develop the IDMSaligned 175 MDRD equation (14). Between-device imprecision was 7.8% (all concentrations, ns98), 7.8% for creatinine -150 mmol/L (ns62), and 6.2% for creatinine )150 mmol/L. The StatSensor creatinine day-to-day imprecision did not meet either the desirable or minimum analytical goal for imprecision derived from biological variation criteria (CV -2.2% and 3.2%, respectively) (19), or the criteria required to keep the analytical error in eGFR calculations below 10% (5). The laboratory assay met both these requirements.
Linearity

atinine results compared with the Roche enzymatic assay, as shown in Figure 1A. Similar findings have been reported at a recent conference (2022), with interference from creatine and urea described in one abstract (20), while the effect of hemtocrit was ruled out as a cause of discordant results in another study (21). Because of the underestimation of creatinine, eGFR results from StatSensor 1 were incorrectly categorised as )60 mL/min for 7/53 patients (false normal results). There were no false abnormal results (eGFR -60 mL/min) (Figure 2A and Table 3). StatSensor 2 had very similar summary statistics, indicating a consistent factory calibration and analytical performance for the two instruments (ns98 pairs, average bias between StatSensors 0.7 mmol/L, two samples had insufficient volume for assay on both devices). However, as illus-

Minor deviations from linearity were observed and are shown in parenthesis after the addition of 100 (3.9%), 250 (5.4%), 500 (5.5%), 750 (2.0%) and 1000 (3.3%) mmol/L creatinine to a base pool of blood containing 58 mmol/L creatinine. The CLSI EP6-A linearity protocol measures the degree to which a curve (polynomial line of best fit) approximates a straight line.
Initial method comparison

Creatinine concentrations in the samples tested ranged from 46 to 962 mmol/L by the laboratory method. Table 2 summarises the method correlation statistics for the Nova StatSensor device vs. the IDMS-aligned Roche enzymatic method, split by creatinine concentration. Using a factory-based calibration, StatSensor 1 (y) produced slightly lower results than the IDMS-aligned Roche enzymatic assay (x) for 62 samples with creatinine concentrations -150 mmol/L (ys0.96x3.5, average bias 7.3, 95% limits of agreement 36 to q21 mmol/L). For 100 samples spanning the full concentration range, the StatSensor results were much lower (ys0.75xq10.2, average bias 47.3, 95% limits of agreement 208 to q113 mmol/L). Patients on dialysis had significantly lower StatSensor cre-

Figure 1 Plot showing differences between Nova StatSensor 1 and IDMS-aligned laboratory Roche enzymatic creatinine results. (A) Using the StatSensor factory encoded calibration. (B) Following recalibration to align creatinine results to the Roche enzymatic assay.

Article in press - uncorrected proof

Shephard et al.: Assessment of Nova StatSensor creatinine analyzer 1117

Figure 2 Plot of eGFR results (mL/min) from Nova StatSensor 1 vs. the laboratory Roche enzymatic method. (A) Using the StatSensor factory encoded calibration. (B) Following recalibration to the Roche enzymatic assay.

trated in Figure 3, for individual samples, there was more variation in results than expected between the two POC analyzers (95% limits of agreement 35 to q34 mmol/L for all samples and 16 to q17 mmol/L for samples with laboratory creatinine -150 mmol/L).
Correction of observed method bias

However, both before and after recalibration, we were concerned by the number of StatSensor creatinine results showing poor agreement with the laboratory method (Figure 1). Predialysis results from one patient were omitted from graphs and statistical calculations because of very inconsistent results (lab 541,538: Nova, factory calibration 186,154 mmol/L).

Using the Passing-Bablok slope and intercept factors, the significant overall negative bias observed across the full creatinine concentration range with the factory-calibrated Nova 1 device was corrected using a reciprocal recalibration equation: Nova (recalibrated)swNova (factory calibration)= 1.3333x 13.53 mmol/L. Method comparison statistics postrecalibration are provided in Table 2. StatSensor 1 eGFR results were then recalculated using the 175 MDRD equation and replotted against the IDMSaligned laboratory eGFR (Figure 2B). Once Nova StatSensor 1 was recalibrated to the laboratory assay, eGFR G60 mL/min was correctly identified vs. the laboratory assay for 37/47 patients (79%). There were 10 false abnormal results. An eGFR -60 mL/min was identified correctly for 51/53 patients (96%). There were two false normal results (Table 3).

Discussion
Screening programs for early detection of CKD are increasing important because the burden of the disease continues to rise globally, and many risk factors, such as hypertension, smoking and obesity can be readily modified (3). In the US, the Kidney Early Evaluation Program (KEEP) provides such a national screening agenda (23). As part of KEEP, creatinine (and eGFR) is measured at a central laboratory and results are returned to patients at a later date. POCT for creatinine confers particular advantages for the CKD screening process as participants can be provided with immediate feedback on their kidney function during their community assessment. In Australia, the recent KEY study combined POCT for creat-

Article in press - uncorrected proof

1118 Shephard et al.: Assessment of Nova StatSensor creatinine analyzer

Table 3 Predictive values for Nova StatSensor vs. Roche Hitachi enzymatic assay using an eGFR cut-off value of 60 mL/min to detect reduced kidney function. Device Nova Nova Nova Nova 1 2 1 2 Calibration Factory calibration Factory calibration Post lab recalibration Post lab recalibration Sensitivity, % 86.8 82.4 96.2 92.2 Specificity, % 100.0 100.0 78.7 78.7 PV (qve test), % 100.0 100.0 83.6 82.6 PV (yve test), % 87.0 83.9 94.9 90.2

Positive test (reduced kidney function) eGFR -60 mL/min; negative test eGFR G60 mL/min.

inine (and calculation of eGFR), glucose, hemoglobin A1c, cholesterol, and urine albumin:creatinine ratio (ACR) with family history, blood pressure and measurement of height, weight and body mass index, as well as an exit interview with a renal nurse to provide an integrated on-site approach to targeted community-based CKD risk assessment (9). POCT was considered pivotal to the success of KEY, with )96% of participants stating POCT was convenient and helped them understand their results better (9). In the KEY study, creatinine was measured on an i-STAT device that proved useful and robust in this setting. However, for large scale risk assessment in settings such as pharmacies or workplaces, capillary whole blood is the sample of choice for creatinine measurement. The Nova StatSensor device, which is simple to use and can measure creatinine using a fingerprick sample, has considerable potential for use in this niche but its analytical performance is of paramount importance in deciding its suitability for screening. Based on the results of this evaluation, agreement with laboratory creatinine results did not meet expectations, especially using the factory-based calibration. The device exhibited a significant negative bias at high creatinine concentrations, with wide limits of agreement compared to the well-established Roche enzymatic assay. The reason for some unusual divergences from laboratory creat-

inine results and falsely low results for patients undergoing dialysis requires further investigation. Imprecision (8.9%) exceeded established criteria at creatinine concentrations -150 mmol/L. For detecting eGFR -60 mL/min, the Nova StatSensor recorded a 13% (7/53) false normal rate, meaning these patients with stage 3 CKD would be missed. Once Nova StatSensor 1 was recalibrated to the IDMSaligned Roche enzymatic Hitachi assay, the number of false normal results decreased to 4% (2/53). However, 21% (10/ 47) of results were now incorrectly classed as abnormal (i.e., eGFR -60 mL/min). For community-based programmes and hospital use, we consider that with recalibration this instrument will identify most patients with eGFR -60 mL/min, but there will be many falsely low eGFR results that will require laboratory validation and contribute to unnecessary stress among patients. Throughout the evaluation period, some further problems were experienced with the StatSensor method. These included poor reproducibility with certain batches of QC material (not used in the imprecision studies) and instability with different reagent strip lot numbers. In summary, the Nova StatSensor did not measure creatinine as well as expected, and we believe that the assay needs urgent improvement. Using the factory calibration, 18% of creatinine results below 150 mmol/L differed by more than 20 mmol/L from the Roche enzymatic assay, and 62% of results above 150 mmol/L differed by more than 20%. Many samples had large differences in creatinine results compared to the IDMS-aligned laboratory method (Figure 1); this is concerning given international efforts to standardise creatinine results (5). Despite this, it could still be useful as a screening test for CKD in community and other settings, as the risk of missing CKD stage 3 with recalibration of the instrument was -5% in this study.

Acknowledgements
We would like to acknowledge the support provided by Nova Biomedical and their Australian agents Regional HealthCare during this study, which included the provision of devices, reagent strips and quality control materials used for the evaluation. We would particularly like to thank Chuck Kirchner, Ron LaPierre, Patrick Chan and Jeffrey Dubois (Nova BioMedical) and Bob Griffen and Neil Spence (Regional HealthCare) for their cooperation during the evaluation period. We would also like to acknowledge the analytical and technical support provided by Brad Rumbelow (FMC) and Pauline Rudevics (CPS), and the secretarial assistance of Cheryl Marshall (CPS) in preparing this manuscript.

Figure 3 Difference between creatinine results (mmol/L) on individual patient samples, when measured on both Nova StatSensor devices.

Article in press - uncorrected proof

Shephard et al.: Assessment of Nova StatSensor creatinine analyzer 1119

Conflict of interest statement

Authors conflict of interest disclosure: The authors stated that there are no conflicts of interest regarding the publication of this article. Research support provided by Nova Biomedical played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication. Research funding: None declared. Employment or leadership: None declared. Honorarium: None declared. 12.

13.

14.

References
1. Chadban SJ, Briganti EM, Kerr PG, Dunstan DW, Welborn TA, Zimmet P, et al. Prevalence of kidney damage in Australian adults: the AusDiab kidney study. J Am Soc Nephrol 2003;14: 1318. 2. Coresh J, Selvin E, Stevens LA. Prevalence of chronic kidney disease in the United States. J Am Med Assoc 2007;298:2038 47. 3. El Kossi M, Bello AK, Hamer R, El Nahas AM. In: Goldsmith D, Jayawardene S, Ackland P, editors. ABC of kidney disease. Oxford: Blackwell Publishing, 2007:114. 4. Chronic kidney disease (CKD) management in general practice. Melbourne: Kidney Health Australia, 2007. 5. Myers GL, Miller WG, Coresh J, Fleming J, Greenberg N, Greene T, et al. Recommendations for improving serum creatinine measurement: a report from the Laboratory Working Group of the National Kidney Disease Education Program. Clin Chem 2006;52:518. 6. Mathew T, Johnson DW, Jones GR. Chronic kidney disease and automatic reporting of estimated glomerular filtration rate: revised recommendation. Med J Aust 2007;187:45963. 7. Jones GR, Mathew T, Johnson D, Peake M. Implementation of the routine reporting of eGFR in Australia and New Zealand. Scand J Clin Lab Invest 2008;241(Suppl):239. 8. Panteghini M. Enzymatic assays for creatinine: time for action. Scand J Clin Lab Invest 2008;241(Suppl):848. 9. Mathew TH, Corso O, Ludlow M, Boyle A, Cass A, Chadban SJ, et al. Screening for chronic kidney disease in Australia: a pilot study in the community and workplace. Kidney Int 2010; 77:S916. 10. Nichols JH, Bartholomew C, Bonzagi A, Garb JL, Jin L. Evaluation of the IRMA TRUpoint and i-STAT creatinine assays. Clin Chim Acta 2007;377:2015. 11. Thomsen HS, Morocos SK, Members of Contrast Media Safety Committee of European Society of Urogenital Radiology. In

15.

16.

17.

18.

19.

20.

21.

22.

23.

which patients should serum creatinine be measured before iodinated contrast medium administration? Eur Radiol 2005;15: 74954. Peake M, Whiting M. Measurement of serum creatinine current status and future goals. Clin Biochem Rev 2006;27:173 84. Available at http://www.aacb.asn.au/files/File/Measurement%20of%20Serum%20Creatinine.pdf (accessed 1 October 2009). Miller WG, Myers GL, Ashwood ER, Killeen AA, Wang E, Thienpont LM, et al. Creatinine measurement: state of the art in accuracy and interlaboratory harmonization. Arch Pathol Lab Med 2005;129:297304. Levey AS, Coresh J, Greene T, Marsh J, Stevens LA, Kusek JW, et al. Expressing the modification of diet in renal disease study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem 2007;53: 76672. Preiss DJ, Godber IM, Lamb EJ, Dalton RN, Gunn IR. The influence of a cooked-meat meal on estimated glomerular filtration rate. Ann Clin Biochem 2007;44:3542. Clinical and Laboratory Standards Institute (CSLI). Evaluation of the Linearity of Quantitative Measurement Procedures: A Statistical Approach; Approved Guideline (EP6-A). 2003. Wayne, PA 19087, USA. Bablok W, Passing H, Bender R, Schneider B. A general regression procedure for method transformation. J Clin Chem Clin Biochem 1988;26:78390. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurements. Lancet 1986;i:30710. Ricos C, Iglesias N, Garcia-Lario J-V, Simon M, Cava F, Hernandez A. Within-subject biological variation in disease: collated data and clinical consequences. Ann Clin Biochem 2007; 44:34352. Schnabl KL, Bagherpoor S, Dubois J, Yip P. Evaluation of the analytical performance of the Nova StatSensor whole blood creatinine meter and reagent strip for chronic kidney disease patients. Clin Chem 2009;55(Suppl):A978. Straseski J, Phelan L, Clarke W. Creatinine levels in patients with renal disease: discordance between POC meter and automated enzymatic methods. Clin Chem 2009;55(Suppl):A102. Korpi-Steiner NL, Wockenfus AM, Lakin JM, Koch CD, Karon BS. Clinical concordance of eGFR measurement between laboratory plasma and whole blood point-of-care creatinine methods. Clin Chem 2009;55(Suppl):A99100. Whalley-Connell AT, Sowers JR, Stevens LA, McFarlane SI, Shlipak MG, Norris KC, et al. CKD in the United States: Kidney Early Evaluation Program (KEEP) and National Health and Nutrition Examination Survey (NHANES) 19992004. Am J Kidney Dis 2008;51(Suppl 2):S1320.

Copyright of Clinical Chemistry & Laboratory Medicine is the property of De Gruyter and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.