Carbamylation of albumin is a cause for discrepancies between albumin assays

Clinica Chimica Acta 434 (2014) 6–10

Contents lists available at ScienceDirect

Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim

Carbamylation of albumin is a cause for discrepancies between albumin assays Maarten B. Kok a,⁎, Frans P.W. Tegelaers a, Bastiaan van Dam b, Jan L.M.L. van Rijn c, Johannes van Pelt a a b c

Medical Center Alkmaar, Laboratory of Clinical Chemistry, Hematology and Immunology, Alkmaar, The Netherlands Department of Internal Medicine, Alkmaar, The Netherlands Westfriesgasthuis, Laboratory of Clinical Chemistry, Hoorn, The Netherlands

a r t i c l e

i n f o

Article history: Received 28 January 2014 Received in revised form 31 March 2014 Accepted 31 March 2014 Available online 5 April 2014 Keywords: Bromocresol BCP BCG Hemodialysis Albumin Carbamylation

a b s t r a c t Background: Several investigators have reported discrepancies between the bromocresol-purple (BCP), bromocresol-green (BCG) and immunonephelometric (INP) assays in dialysis patients. This study compared the abovementioned assays and investigated whether hemodialysis itself or carbamylation of albumin is the cause for this discrepancy. Methods: Samples obtained from hemodialysis patients were analyzed by BCP, BCG and INP. Furthermore, albumin was carbamylated in vitro using isocyanate. Isocyanate converts lysine to homocitrulline. Results: No differences were observed between samples of the pre- and post-hemodialysis groups for BCP. In the control group, BCG averaged 6 g/L higher than INP. BCP did not statistically deviate from INP. In the dialysis group BCG averaged 5 g/L higher when compared to INP, whereas BCP averaged 2 g/L lower. BCP was affected by carbamylation of albumin. BCG and INP measurements were affected to a much lesser extent. Homocitrulline content of hydrolysates was increased in both the carbamylated albumin as well as in the dialysis population. Conclusion: In a hemodialysis population albumin concentrations are not consistently estimated by both BCG and BCP methods. Relative to INP measurements BCG overestimates the albumin concentration (4–10 g/L), whereas BCP leads to an underestimation (0–4 g/L). Carbamylation of albumin is the main attributor to the discrepancy found with BCP. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Analysis of plasma albumin is generally performed using either a bromocresol-green (BCG) or a bromocresol-purple (BCP) based colorimetric method. BCG and BCP are the most frequently utilized methods, since they are cost effective, quick and readily available on a variety of chemistry platforms. Other techniques, such as immunonephelometry (INP) or immunoturbidimetry are used less frequently, since these are more time-consuming and expensive. Main advantage of the use of immunonephelometry is the absence of cross-reactivity with nonalbumin proteins [1]. In patients with end-stage kidney disease, albumin is an important predictor of mortality and morbidity. Furthermore, it is required for the interpretation of total calcium levels in patients requiring dialysis in case of hypoalbuminemia [2,3]. Adjustment of total calcium for low

Abbreviations: BCP, bromocresol purple; BCG, bromocresol green; INP, immunonephelometry; CKD, chronic kidney disease; KDIGO, the Kidney Disease Improving Global Outcomes guidelines. ⁎ Corresponding author. E-mail address: [email protected] (M.B. Kok).

http://dx.doi.org/10.1016/j.cca.2014.03.035 0009-8981/© 2014 Elsevier B.V. All rights reserved.

albumin levels is advised by the Nephrology Guidelines (KDIGO; [4]) and several correction formulas have been suggested [5,6]. Several publications have reported discrepancies in albumin concentration between the BCP, BCG and INP assays [1,5,7–13]. In healthy subjects, BCP has been reported to closely agree with INP, whereas BCG reveals a positive bias when compared to BCP or INP, especially at low albumin levels [8]. Carfray et al. demonstrate that BCG has a positive bias up to 10 g/L when compared to either BCP or INP and also show that this discrepancy increases with hypoalbuminemia [14]. Ueno et al. describe that acute phase proteins such as haptoglobin may interfere in the BCG assay [13]. Xu et al. described that serum globulins are the main contributor to the discrepancy between BCG and BCP [7]. Pinell et al. report that immunoglobulins do not affect the BCP assay [15]. Interestingly, several investigators have also reported that differences found between BCP, BCG and INP are more pronounced in a population of patients with chronic renal failure requiring hemodialysis [9, 10,16]. A discrepancy of up to 9 g/L has been observed between BCP and BCG [9,10]. Previous studies have hypothesized that the reported increase of albumin in plasma of uremic patients could be attributed to either uremic toxins [17] or possibly to an unidentified chemical modification of albumin, that inhibits the binding of BCP to albumin, but not that of BCG or INP [8,10]. Jaisson et al. and Kraus et al. have

M.B. Kok et al. / Clinica Chimica Acta 434 (2014) 6–10

previously reported on the occurrence of modification of proteins coined carbamoylation or carbamylation. Carbamylation results from the covalent binding of isocyanic acid (or isocyanate) to the ε-amino group of lysine to form homocitrulline proteins. Isocyanate may be formed under normal physiological conditions but is amplified in a hyperuremic state, such is the case for patients suffering from (end stage) renal failure. Furthermore, isocyanate may also be formed by oxidation of thiocyanate by myeloperoxidase [18,19]. Recently, carbamylated proteins have shown to be associated with increased mortality and morbidity rates in patients suffering from end stage renal disease [20–22]. Furthermore Gillery et al. described that carbamylation products are considered potential toxins [23]. Ito et al. showed that approximately two binding sites for BCP are present on albumin. Interestingly, they also showed that both binding sites contain a lysine moiety, which may be available for modification by isocyanic acid [24]. We hypothesize that modification of this specific amino acid may prevent binding of BCP to albumin and leads to falsely lower results when measuring albumin. In this study, we evaluated whether hemodialysis itself influenced the BCG, BCP and INP by comparing samples obtained from 80 dialysis patients pre- and post-hemodialysis. Furthermore, we investigated whether in vivo carbamylation of ε-amino-end groups is a cause for the differences between BCP and BCG or INP. To this aim, purified human albumin was carbamylated by adding isocyanate. The carbamylated albumin as well as samples of patients were analyzed for homocitrulline content.

performed according to the scheme for the 'Combi Synchron' and the 'Combi General Clinical Chemistry'. Measurements of total protein and albumin by means of BCP and INP were performed by de Laboratory for Clinical Chemistry, Hematology and Immunology of the Medical Center Alkmaar, Alkmaar, The Netherlands. Both the BCP (Beckman, ALBm albumin, REF467858) and the rate biuret based total protein assay (Beckman, TPm, total protein, REF465986) were performed on a Beckman Synchron Dx860i using a modular chemistry cartridge based assay. Albumin INP measurements were performed on a Siemens Prospec 2 using N antiserum to Albumin (Siemens, NAS ALB). Albumin measured by means of BCG (Roche, Albumin gen. 2, ALB2) was performed by the Laboratory for Clinical Chemistry of the Westfriesgasthuis, Hoorn, The Netherlands on a Roche COBAS 6000. All measurements were performed on the same day. 2.3. KCNO experiment To verify whether carbamylation affected the BCP and BCG assay 2 mL of human albumin (200 g/L, Albuman, Sanquin Blood Supply, The Netherlands) was modified by adding 4 mL isocyanate (KCNO; Merck, Darmstadt, Germany) in demineralized water at a concentration of 225 mM, 225 μM and 225 nM, to obtain a final albumin concentration of approximately 70 g/L and a concentration of respectively 150 mM, 150 μM and 150 nM of KCNO. As a control, the KCNO solution was substituted by an equimolar solution containing KCl (Merck, Darmstadt, Germany) in demineralized water. After addition of the KCNO or KCl solution, both total protein and albumin were measured instantaneously. Subsequently, the samples were incubated at 37 °C. After 24 h of incubation, total protein and albumin were measured. Samples were stored at −30 °C and thawed before further analysis. Carbamylation of albumin was verified by electrophoresis as was previously described by Jaisson et al. [25]

2. Materials and methods 2.1. Patient samples To investigate the discrepancy in measurement of albumin in hemodialysis patients between BCP and BCG, both methods were compared with INP. To this aim, heparin plasma of 80 hemodialysis patients was collected pre- and post-dialysis. Furthermore, control-group heparin plasma samples were collected from 35 patients not diagnosed with kidney disease. In the control group, subjects creatinine and urea levels were within the reference interval.

2.4. Measurement of homocitrulline In total 15 patient samples (6 hemodialysis patients and 9 control patients) and 2 albumin samples (KCNO experiment and control) were analyzed. For every sample 500 μL of plasma was hydrolyzed for 24 h at 102 °C using 37% HCl (Merck, Darmstadt, Germany) to obtain a final concentration of 6 M. The amino acid mixture was filtered using a 0.2 μm low protein binding filter (AcroDisc, Pall Lifesciences, NY, USA). The filtrate was analyzed by UPLC–MS/MS (Acquity — Quattro Premier XE, Waters) using stable isotope dilution (SID) technique at the Laboratory Genetic Metabolic Diseases of the Academic Medical Center of the University of Amsterdam, The Netherlands according to Waterval et al. [26]. For homocitrulline, the limit of detection (LOD) was 0.6 μmol/L. The limit

2.2. Measurement of albumin/total protein Heparin plasma was used for analysis. All methods were validated according to the standard protocols provided by CLSI. Quality of the measurements was assessed by measuring both internal and external (Dutch Foundation for Quality Assessment in Medical Laboratories, SKML, www.skml.nl) quality control standards. Proficiency testing was 55

A

55

B

50

50

45

45

45

40 35 30

40 35 30

25 20 20

BCP/BCG g/L

50

BCP/BCG g/L

BCP/BCG g/L

55

7

BCG y = 0.89x + 9.9 R² = 0.84

25

30

35

40

INP g/L

45

50

55

20 20

40 35 30

25

BCP y 1.00x + 0.6 R² = 0.98

C

BCG

25

BCP

y = 0.78x + 12.7 y = 0.81x + 4.8 R² = 0.89 R² = 0.86

25

30

35

40

INP g/L

45

50

55

20 20

BCG y =0.92x + 8.8 R² =0.84

25

30

35

40

45

BCP y =0.80x + 6.3 R² =0.84

50

55

INP g/L

Fig. 1. Comparison between three methods for albumin. Bromocresol green (BCG) and Bromocresol purple (BCP) as a function of INP in plasma of patients with normal kidney function (A) and patients with end-stage kidney failure requiring hemodialysis (B/C). Data shown in B represents samples that were obtained pre-dialysis. Data shown in C represents samples that were obtained post-dialysis. The solid lines represent the equation for BCG/BCP as a function of INP determined by linear regression analysis. The dashed line is the identity line (x = y).

M.B. Kok et al. / Clinica Chimica Acta 434 (2014) 6–10

of quantification (LOQ) for homocitrulline was 1 μmol/L, which corresponds with 0.03 μmol/g albumin. 2.5. Statistics Data was analyzed using linear regression analysis to determine whether the different linear relationships were significantly different. To this aim the interaction term Z was introduced to obtain the following equation: y = b0 + b1X + b2Z + b3XZ + ε. In this equation b0 is the intercept and b1 is the slope of the first linear relationship. Z = 0 for the first linear relationship and Z = 1 for the second linear relationship. Using linear regression, a significant difference between the first relationship and the second relationship was defined as a value for p-value b 0.05 for either b2 and/or b3. b2 and b3 are the respective differences in the intercept and slope of the second relationship that are compared to the first relationship. 3. Results To investigate the discrepancy in measurement of albumin in hemodialysis patients between bromocresol purple (BCP) and bromocresol green (BCG) methods, both methods were compared to an immunonephelometric based method. Firstly, a comparison of the three methods was made for subjects with a normal kidney function. These results are presented in Fig. 1A. Linear regression analysis shows that the correlation between BCP and immunonephelometry (INP) is very good and does not significantly deviate from the identity line (x = y). Furthermore, it is shown that BCG has a significant bias throughout the entire range averaging 6 g/L in comparison with both BCP and INP (range: 4–11 g/L; p b 0.005 (1-sided paired student t-test), See Table 1). The differences between the assays are more pronounced at lower concentrations of albumin. As can be appreciated in Fig. 1B (pre-dialysis) and 1C (post-dialysis), BCP as a function of INP shows a significantly different linear relationship for the dialysis group when compared to the relationship found in subjects with normal kidney function. The linear relationship significantly deviates in both slope and intercept (p b 0.05). This is further illustrated in Table 1, which clearly demonstrates that BCP in comparison with INP averages 2.7 g/L lower (p b 0.005 (1-sided paired student t-test)) in the predialysis group when compared to the control group. Table 1 further shows that the average discrepancy between BCG and INP remains constant. To rule out whether hemodialysis itself influenced either the BCG, the BCP or the INP, 80 pre- and post-hemodialysis samples were compared. These results are shown in Fig. 1B (pre-dialysis) and Fig. 1C (post-dialysis) and Table 2. For BCP no significant difference between pre- and post-hemodialysis was observed in neither slope nor offset, thereby indicating that no interfering substance is present or introduced during dialysis (See Table 1). Table 1 further shows that the average discrepancy between BCG and INP remains constant. To investigate whether in vivo carbamylation of ε-amino-end groups is a cause for the differences observed between BCP and Table 1 Averages of albumin concentration. Differences (Δ) are calculated by subtracting albumin concentration measured by INP. Method Control (n = 35) Pre-dialysis (n = 80) Post-dialysis (n = 80)

INP BCP BCG INP BCP BCG INP BCP BCG

Average g/L 33.8 34.5 39.8 35.6 33.6 40.5 37.2 36.2 42.9

Δ g/L – 0.7 6.0 – −2.0 4.9 – −1.0 5.7

Range g/L 12–49 13–50 24–49 26–45 26–41 33–48 23–49 24–49 29–55

Table 2 Slope and offset as determined with linear regression of albumin measurements using both BCP and BCG as a function of albumin measured by INP. SD were calculated by means of linear regression analysis and represents the standard deviation on the slope or the offset.

Control Pre-dialysis Post-dialysis Control Pre-dialysis Post-dialysis

Method

slope

SD

Offset

SD

BCP BCP BCP BCG BCG BCG

1.00 0.81⁎ 0.80⁎ 0.89 0.78⁎#

0.03 0.04 0.04 0.03 0.03 0.05

0.6 4.8⁎ 6.3⁎ 9.9 12.7 8.8

1.0 1.5 1.5 0.8 0.8 1.7

0.92#

⁎ Statistical significant difference compared to the control group as determined by linear regression analysis. # Statistical significant difference between pre- and post-dialysis group as determined by linear regression analysis.

BCG/INP, purified human albumin was carbamylated by adding isocyanate (KCNO). Fig. 2 shows that after 24 h of incubation with 150 mM KCNO, albumin measured by BCP is approximately 40 g/L lower than is the case with non-carbamylated albumin. Total protein measurements, on the other hand, remained unaffected by carbamylation and gave identical values for all four concentrations used. BCP is also affected at lower concentration. BCG and INP were only affected to a lesser extent for non-physiological concentrations (150 mM and 150 μM). For the lowest concentrations of isocyanate no discrepancy could be demonstrated for BCG and INP. To prove whether carbamylation leads to the modification of lysine to homocitrulline, both modified albumin samples as well as dialysis and control samples were hydrolyzed and analyzed for amino acid content. The homocitrulline content was undetectable (i.e. lower than the LOQ) for the non-modified albumin, which was obtained by Cohn's fractionation by the manufacturer. After carbamylation, the homocitrulline content of albumin increased to 30 μmol/g albumin. Results of the patient samples are shown in Fig. 3 and reveal a significant increase (p b 0.005, 1-sided non-paired students t-test) in homocitrulline content, whereas lysine, alanine and total protein content of the dialysis group were comparable to the control group. These results demonstrate that albumin is affected by carbamylation both in vitro and in vivo. 4. Discussion In this study, we set out to investigate the consequence of hemodialysis and carbamylation on the measurement of albumin using BCG, BCP #

* * *

* * NS

* NS

80 70 60

Albumin g/L

8

Concentration KCNO

50

150 mM 40

150 µM

30

150 nM

20

Control

10 0

BCP

BCG

INP

Method Fig. 2. The effect of carbamylation of albumin as a function of isocyanate (KCNO) concentration for BCP, BCG and INP. Error bars represent the standard deviation. * Significant difference (p b 0.005), # significant difference (p b 0.05), and NS means not significant. Error bars represent the standard deviation.

M.B. Kok et al. / Clinica Chimica Acta 434 (2014) 6–10

p < 0.005 140 120 100

Dialysis NS

NS

Control NS

80 60 40 20 0

Fig. 3. Results of the amino acid analysis of hydrolysates for patients treated with hemodialysis and patients of the control group. Result is reported as plasma concentration. Error bars represent the standard deviation. NS means not significant.

or INP. Patient samples were obtained from patients with renal failure requiring hemodialysis (CKD stage 5 dialysis) and patients without renal failure (control group). In the control group, a good correlation was found between BCP and INP, whereas BCG deviated by an average of 6 g/L when compared to INP. These findings were in line with previous observations [9,10,27]. The discrepancy between BCG and BCP/INP was more pronounced at lower levels of albumin. This phenomenon has been previously reported [7,8,11]. Xu et al. have identified serum globulins to be the main contributor to this discrepancy [7]. The comparison between samples obtained pre- and post-hemodialysis showed no effect of hemodialysis on BCP and INP. As several authors have reported previously [1,5,7–10], subjects in the dialysis group did show a different trend for the BCP assay when compared to the control group. Our results show an underestimation up to 6 g/L and were larger for higher concentrations of albumin. We hypothesize that modification of lysine by isocyanate results in less effective binding of BCP and thus in a lower value for the albumin concentration. Parikh et al. [8] reported that the gap between BCG and BCP was even greater in patients undergoing peritoneal dialysis. We postulate that the decrease may be the result of higher average urea levels in this population. As a result more isocyanate will be formed which will subsequently lead to higher grades of carbamylation of albumin. Further experiments show that the discrepancy could be attributed to the carbamylation of albumin as a result of high levels of urea. We have shown that isocyanate concentrations at levels found in patients with end stage renal disease (150 nM) [28] are able to cause a decrease in albumin using BCP, whereas BCG and INP remained unaffected at this concentration of isocyanate. Higher, non-physiological, concentrations of cyanate did affect all three methods. These findings are in concordance with Calvo et al., who showed that the BCG method was affected by carbamylation at superphysiological concentrations of cyanate [29]. We postulate that a higher degree of carbamylation may affect the 3-dimensional structure of albumin, thereby affecting both the avidity of the antibodies used with INP and the binding of BCG. We hypothesize that the differences found between BCP, BCG and INP observed in Fig. 2 are the result of differences in the matrix of the samples. These samples only contained albumin diluted with cyanate and/or KCl solutions. No other plasma constituents were added. Quality of the measurements was assessed by measuring both internal and external (Dutch Foundation for Quality Assessment in Medical Laboratories, SKML) quality control samples. No discrepancies or

9

outliers were identified for the external control samples. Interestingly, two separate groups (i.e. BCP and BCG) were created by the SKML for the interpretation and processing of the results of the external quality assessment. Between these two groups discrepancies could be observed that concur with our data. One would expect these differences to be reflected in the reference intervals that are used for albumin, but those were equal in both hospitals (35–50 g/L). Using the concentration (78.4 μmol/L) for homocitrulline from Fig. 3 we were able to estimate the amount of lysines that are modified on each molecule of albumin. We assumed the average concentration of total protein to be 70 g/L of which 35 g/L is albumin. We expect isocyanate to react equally with lysine residues on either albumin or other proteins. Therefore only half of the homocitrulline (39.2 μmol/L) can be attributed to albumin. When assuming a molar weight of 66 kDa, 35 g/L of albumin is equivalent with 530 μmol/L. Calculations show that 1:14 molecules of albumin contains a homocitrulline. Conversion of lysine to homocitrulline may alter the binding efficiency of BCP resulting in a lower value for albumin. Based on a plasma albumin concentration of 35 g/L, a calculated ratio of 1:14 would lead to a decrease of 2.4 g/L, which closely resembles the actual difference observed in this study. It could be debated whether either BCG, BCP or INP is a suitable assay for measuring albumin in patients requiring dialysis. Not only is albumin an important predictor of mortality and morbidity, but also it is used in the estimation of nutritional status and is required for the interpretation of total calcium levels in patients requiring dialysis, especially in case of hypoalbuminemia [2,3]. In case of hypoalbuminemia, the BCG method leads to an underestimation of the adjusted total calcium level, whereas BCP leads to an overestimation as a consequence of the discrepancies found with these assays. Since hypercalcemic patients are generally treated with more expensive calcium free phosphate binders and calcimimetics, the choice of assay may also directly affect cost per patient. When comparing drug prescription to patients with end-stage renal disease between the Medical Center Alkmaar, Alkmaar, The Netherlands, where albumin is measured by means of BCP and the Westfriesgasthuis, Hoorn, The Netherlands, who measure albumin using the BCG assay, we observed a large difference in cost per patient as a result of differences albumin concentration. Interestingly, cost per patient for management of phosphate, calcium and PTH were 2.5-fold higher in Medical Center Alkmaar, mainly as a result of the prescription of more calcium free phosphate binders (unpublished data). Our conclusions are in concordance with Kato et al. [30] and Labriola et al. [1] who also affirm that the choice of method (BCP, BCG, INP) has a major impact on both the classification of patients according to the KDIGO guidelines as well as on the prescription of the abovementioned drugs within this population [1]. Lately, KDIGO have published a position statement in which it is acknowledged that measurement of ionized calcium is the preferred method for evaluating serum calcium levels [31]. 5. Conclusion This study has shown that BCG, BCP and INP show large discrepancies when measuring albumin, especially when determining the concentration of albumin in patients requiring dialysis. Values measured with BCG are consistently 4–10 g/L higher when compared to the albumin concentration as determined by BCP or INP. For plasma of dialysis patients, carbamylation of albumin is the main attributor to the discrepancy found for BCP and leads to an underestimation of 0–6 g/L. Both interferences lead to variations with a wide distribution that cannot be easily corrected for. In our opinion, an immunochemical based method such as immunonephelometry of immunoturbidimetry would be the preferred method for measuring albumin in patients requiring dialysis. Especially, since both techniques are seemingly unaffected by carbamylation and are not affected by interfering substances, such as immunoglobulins, as is the case with BCG. However, most studies that

10

M.B. Kok et al. / Clinica Chimica Acta 434 (2014) 6–10

show albumin to be a predictor of mortality and morbidity have used BCG as a method for measuring albumin [2,3,32–34]. Since our study shows discrepancies between the different methods, these observations cannot be generalized to other methods for measuring albumin. Unfortunately, no gold standard is available for measuring albumin in plasma. This goal can only be achieved by the development of a reference material or a reference method for measuring albumin as was stated previously by Lo et al. [35] and Lieske et al. [36] As a consequence we would like to advocate that future guidelines take into account that utilization of different methods for the same analyte may lead to differences in results and therefore reference intervals and will, subsequently, affect the interpretation of the guidelines and eventually treatment of patients. Acknowledgment We would like to thank Dr. Wim Kulik of the Laboratory Genetic Metabolic Diseases of the Academic Medical Center of the University of Amsterdam, The Netherlands for performing the measurement of homocitrulline. Furthermore we would like to thank Drs. Tjeerd van de Ploeg, Science Department of the Medical Center Alkmaar, Alkmaar, The Netherlands, for his contribution to the statistical analysis of the data. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.cca.2014.03.035. References [1] Labriola L, Wallemacq P, Gulbis B, Jadoul M. The impact of the assay for measuring albumin on corrected (‘adjusted’) calcium concentrations. Nephrol Dial Transplant 2009;24:1834–8. [2] Lowrie EG, Lew NL. Death risk in hemodialysis patients: the predictive value of commonly measured variables and an evaluation of death rate differences between facilities. Am J Kidney Dis 1990;15:458–82. [3] Owen Jr WF, Lew NL, Liu Y, Lowrie EG, Lazarus JM. The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis. N Engl J Med 1993;329:1001–6. [4] KDIGO clinical practice guideline for the diagnosis, evaluation, prevention, and treatment of Chronic Kidney Disease-Mineral and Bone Disorder(CKD-MBD). Kidney Int Suppl 2009:S1–S130. [5] Clase CM, St Pierre MW, Churchill DN. Conversion between bromcresol green- and bromcresol purple-measured albumin in renal disease. Nephrol Dial Transplant 2001;16:1925–9. [6] Jain A, Bhayana S, Vlasschaert M, House A. A formula to predict corrected calcium in haemodialysis patients. Nephrol Dial Transplant 2008;23:2884–8. [7] Xu Y, Wang L, Wang J, Liang H, Jiang X. Serum globulins contribute to the discrepancies observed between the bromocresol green and bromocresol purple assays of serum albumin concentration. Br J Biomed Sci 2011;68:120–5. [8] Parikh C, Yalavarthy R, Gurevich A, Robinson A, Teitelbaum I. Discrepancies in serum albumin measurements vary by dialysis modality. Ren Fail 2003;25:787–96. [9] Joseph R, Tria L, Mossey RT, et al. Comparison of methods for measuring albumin in peritoneal dialysis and hemodialysis patients. Am J Kidney Dis 1996;27:566–72. [10] Wells FE, Addison GM, Postlethwaite RJ. Albumin analysis in serum of haemodialysis patients: discrepancies between bromocresol purple, bromocresol green and electroimmunoassay. Ann Clin Biochem 1985;22:304–9.

[11] Speicher CE, Widish JR, Gaudot FJ, Hepler BR. An evaluation of the overestimation of serum albumin by bromcresol green. Am J Clin Pathol 1978;69:347–50. [12] Blagg CR, Liedtke RJ, Batjer JD, et al. Serum albumin concentration-related Health Care Financing Administration quality assurance criterion is method-dependent: revision is necessary. Am J Kidney Dis 1993;21:138–44. [13] Ueno T, Hirayama S, Ito M, et al. Albumin concentration determined by the modified bromocresol purple method is superior to that by the bromocresol green method for assessing nutritional status in malnourished patients with inflammation. Ann Clin Biochem 2013;50:576–84. [14] Carfray A, Patel K, Whitaker P, et al. Albumin as an outcome measure in haemodialysis in patients: the effect of variation in assay method. Nephrol Dial Transplant 2000;15:1819–22. [15] Pinnell AE, Northam BE. New automated dye-binding method for serum albumin determination with bromcresol purple. Clin Chem 1978;24:80–6. [16] Beyer C, Boekhout M, van Iperen H. Bromcresol purple dye-binding and immunoturbidimetry for albumin measurement in plasma or serum of patients with renal failure. Clin Chem 1994;40:844–5. [17] Mabuchi H, Nakahashi H. Underestimation of serum albumin by the bromcresol purple method and a major endogenous ligand in uremia. Clin Chim Acta 1987;167: 89–96. [18] Jaisson S, Pietrement C, Gillery P. Carbamylation-derived products: bioactive compounds and potential biomarkers in chronic renal failure and atherosclerosis. Clin Chem 2011;57:1499–505. [19] Kraus LM, Kraus Jr AP. Carbamoylation of amino acids and proteins in uremia. Kidney Int Suppl 2001;78:S102–7. [20] Koeth RA, Kalantar-Zadeh K, Wang Z, et al. Protein carbamylation predicts mortality in ESRD. J Am Soc Nephrol 2013;24:853–61. [21] Kalim S, Tamez H, Wenger J, et al. Carbamylation of serum albumin and erythropoietin resistance in end stage kidney disease. Clin J Am Soc Nephrol 2013;8:1927–34. [22] Berg AH, Drechsler C, Wenger J, et al. Carbamylation of serum albumin as a risk factor for mortality in patients with kidney failure. Sci Transl Med 2013;5:175ra29. [23] Gillery P, Jaisson S. Post-translational modification derived products(PTMDPs): toxins in chronic diseases? Clin Chem Lab Med 2014;52:33–8. [24] Ito S, Yamamoto D. Identification of two bromocresol purple binding sites on human serum albumin. Clin Chim Acta 2010;411:1536–8. [25] Jaisson S, Delevallee-Forte C, Toure F, et al. Carbamylated albumin is a potent inhibitor of polymorphonuclear neutrophil respiratory burst. FEBS Lett 2007;581:1509–13. [26] Waterval WA, Scheijen JL, Ortmans-Ploemen MM, Habets-van der Poel CD, Bierau J. Quantitative UPLC-MS/MS analysis of underivatised amino acids in body fluids is a reliable tool for the diagnosis and follow-up of patients with inborn errors of metabolism. Clin Chim Acta 2009;407:36–42. [27] Maguire GA, Price CP. Bromcresol purple method for serum albumin gives falsely low values in patients with renal insufficiency. Clin Chim Acta 1986;155:83–7. [28] Nilsson L, Lundquist P, Kagedal B, Larsson R. Plasma cyanate concentrations in chronic renal failure. Clin Chem 1996;42:482–3. [29] Calvo R, Carlos R, Erill S. Underestimation of albumin content by bromocresol green, induced by drug displacers and uremia. Int J Clin Pharmacol Ther Toxicol 1985;23: 76–8. [30] Kato A, Takita T, Furuhashi M, et al. Influence of the assay for measuring serum albumin on corrected total calcium in chronic hemodialysis patients. Ther Apher Dial 2011;15:540–6. [31] Moe S, Drueke T, Cunningham J, et al. Definition, evaluation, and classification of renal osteodystrophy: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int 2006;69:1945–53. [32] Herrmann FR, Safran C, Levkoff SE, Minaker KL. Serum albumin level on admission as a predictor of death, length of stay, and readmission. Arch Intern Med 1992;152: 125–30. [33] Guijarro C, Massy ZA, Wiederkehr MR, Ma JZ, Kasiske BL. Serum albumin and mortality after renal transplantation. Am J Kidney Dis 1996;27:117–23. [34] Corti MC, Guralnik JM, Salive ME, Sorkin JD. Serum albumin level and physical disability as predictors of mortality in older persons. JAMA 1994;272:1036–42. [35] Lo SF, Miller WG, Doumas BT. Laboratory performance in albumin and total protein measurement using a commutable specimen: results of a College of American Pathologists study. Arch Pathol Lab Med 2013;137:912–20. [36] Lieske John C, Bondar O, Miller WG, et al. A reference system for urinary albumin: current status. Clin Chem Lab Med 2013;51:981–9.