Dialysis – Transplantation

Kidney International (1998) 53, 488–495; doi:10.1046/j.1523-1755.1998.00777.x

Carbamylated hemoglobin and carbamylated plasma protein in hemodialyzed patients

Cynthia M Balion, Thomas F Draisey and Roger J Thibert

Department of Chemistry and Biochemistry, University of Windsor, Windsor, Ontario, Canada

Correspondence: Dr Cynthia M. Balion, GEM Program, Department of Pathology, Children's & Women's Health Centre of British Columbia, 4480 Oak St., Room 2J17, Vancouver, British Columbia V6H 3V4, Canada

Received 13 September 1996; Revised 17 September 1997; Accepted 24 September 1997.

Top

Abstract

Carbamylated hemoglobin and carbamylated plasma protein in hemodialyzed patients. The carbamylation reaction in vivo involves the nonenzymatic, covalent attachment of isocyanic acid, the spontaneous dissociation product of urea, to proteins. Carbamylated proteins have been proposed as markers of uremia and indicators of uremic control. However, the utility of measuring carbamylated proteins has not been investigated adequately. Therefore, this study was done to determine the relationship between the carbamylation of long-lived protein (hemoglobin) with that of short-lived proteins (plasma proteins) in hemodialyzed patients. Significantly higher carbamylated hemoglobin (CHb; 157 plusminus 40 mug valine hydantoin/g Hb) and carbamylated protein (CTP; 0.117 plusminus 0.011 absorbance/mg protein) concentrations were found in hemodialyzed patients (N = 13) as compared to normal individuals (N = 9, 53 plusminus 20 mug valine hydantoin/g Hb and 0.08 plusminus 0.01 absorbance/mg protein, respectively). A high correlation was found between CHb and CTP concentrations (r = 0.87, P < 0.0001), demonstrating a strong relationship between these two different half-lived proteins. A six-month longitudinal study of seven hemodialyzed patients showed that the between subject correlations were significant for CHb versus CTP as well as CHb versus pre-dialysis urea. Correlations were not significant for CTP versus pre-dialysis urea or Kt/V, nor CHb versus Kt/V. Carbamylated hemoglobin fluctuated the most over this time period (30.1%plusminus 20.2%), pre-dialysis urea and CTP varied less (18.3%plusminus 13.4% and 14.9%plusminus 7.5%, respectively), and Kt/V varied the least (6.3%plusminus 3.3%). Within subject correlations were not significant between any two tests. It is unclear whether the lack of correlations found is real or a function of the small sample size. However, these data do show that CHb and CTP are positively associated and reflect the degree of urea exposure in the blood, but their usefulness for patients on maintenance hemodialysis is not clear.

Keywords:

carbamylated hemoglobin and plasma protein, hemodialysis, urea, proteins and urea, measuring carbamylated proteins

Abbreviations:

A/mg protein, absorbance/mg protein; CHb, carbamylated hemoglobin; CHSA, carbamylated human serum albumin; CNVal, N-carbamyl-DL-norvaline; CVal, N-carbamyl-DL-valine; EDTA, ethylenediaminetetraacetic acid; Hb, hemoglobin; Kt/V, dialysis dose; PBS, phosphate buffered saline; TAC, time averaged urea concentration; UKM, urea kinetic modeling; VH, valine hydantoin

Increases in blood urea levels have always been associated with a loss of kidney function. Recently, a new approach has been taken for its measurement that has broadened the scope of its use. Urea decomposes to give cyanate and ammonium ions1. Cyanate, in its unprotonated form, reacts with unpaired electrons through a nucleophilic reaction2,3. In vivo, the most stable reactions are with the electrons of the free amino groups (alpha and epsilon), and sulfhydryl groups. Carbamylation studies done in vitro4,5 have demonstrated that carbamylation causes changes in conformation, decreased or inhibition of enzyme activity, binding changes (to receptors and drugs), and physically causes an increase in anodicity on various types of proteins.

Elevated urea levels therefore not only indicate a loss of kidney function, but may also serve as a source for potential pathophysiological consequences. Several reports have inferred the association of carbamylation with symptom or loss of function such as cataracts6, neuropathies7, nausea, drowsiness, memory loss4,8, atherosclerosis9,10, and drug binding defects12,13,14. In fact, the presence of carbamylated amino acids15, membranes16,17, lipoproteins9, plasma proteins12,14,18,19,20, and hemoglobin11,21,22,23,24,25,26,27 have been found in patients with renal failure. However, no study has yet shown by direct measurement that symptoms seen in patients with uremia are a result of carbamylation.

Nonetheless, there has been ample progress in assessing carbamylated proteins as markers of uremia3,11,15,16,18,19,21,22,23,24,25,26,27,28,29,30,31. This concept is analogous to that of glycated proteins in diabetes32. When glucose (or cyanate or other potential compound) makes a stable attachment to a protein, it takes on the half-life of that protein. In the assessment of diabetics, it is these glucose-protein adducts that reveal the past dynamics of blood glucose levels, a time-integrated marker. Glucose or urea measurements alone are subject to day-to-day dietary and metabolic influences, thereby reflecting transient time periods. The most common marker proteins measured are hemoglobin, which is a long-term indicator (2 months), and albumin or total plasma protein, which are short-term indicators (2 to 3 weeks)33.

Carbamylated hemoglobin has been the major carbamylated protein studied owing to initial studies that sought to use cyanate as a therapy for sickle cell anemia34. Flückiger et al were the first to identify carbamylated hemoglobin in uremic patients and furthermore correlated HbA1 with time-averaged urea concentration28. Oimomi et al29 then investigated the possibility of using HbA1 as an indicator of control in patients with renal failure. A time-course relationship between urea and HbA1 demonstrated a correlation with HbA1 and urea concentrations measured one to two weeks earlier. In another study, CHb in renal transplant patients reflected the degree of uremia two to three weeks earlier27.

Evaluation of carbamylated hemoglobin in patients with various degrees of renal function has shown that its measurement has potential clinical value Table 111,21,22,23,24,25,26,27,31. There are significant differences between CHb in control groups22,24,31 compared to patients with chronic renal failure21,22,24,29,31, patients on continuous ambulatory dialysis22,24, hemodialysis22,24,28,31, and patients with failed transplants22. No differences are seen between CHb in control groups and patients with primary renal disease22, successful transplants22,24, or diabetic patients without renal failure21,24,29. However, CHb has been found to be capable of differentiating acute from chronic renal impairment25. The positive predictive value for this test was 80% with a sensitivity of 100% when the ratio of CHb/creatinine was set at 0.2.


Furthermore, there are significant correlations between carbamylated hemoglobin values compared to urea, creatinine, and time-averaged urea concentration (TAC) measurements11,21,22,23,24,25,26,27,28,29, but not with bicarbonate and phosphate values31.

Considerably less information is available on the relationship of other carbamylated compounds in renal patients. However, carbamylated plasma protein (CTP)12,14,18,19, leukocytes16, erythrocytes, and amino acids15 in uremic individuals have all been found to be higher compared to normal individuals. Table 2 lists the carbamylated protein values in normal individuals and patients with renal failure who have been reported in the literature.


Measurement of carbamylated protein could therefore provide the clinical information needed as an indicator of uremic control. It could be used to monitor the deterioration of renal function, to assess compliance to dietary therapy, and to evaluate the adequacy of dialysis treatment23. The benefit of using carbamylated proteins as markers is that they can be used as alternatives to TAC measurements, particularly in dialyzed patients. The currently accepted method to quantify dialysis therapy and define its adequacy is through a mathematical description of urea balance called urea kinetic modeling (UKM)39. The calculation of UKM parameters requires collection of numerous pieces of data, including accurately timed urea measurements, and a computer to solve the simultaneous equations. In comparison, measurement of carbamylated protein requires only a single measurement on one blood sample that could be performed easily in a clinical laboratory.

However, more clinical studies are needed to fully establish the usefulness of carbamylated protein measurement in dialyzed patients, particularly in assessing adequacy of dialysis. Therefore, a six-month longitudinal study was undertaken to evaluate the changes in carbamylated hemoglobin, carbamylated protein, urea, and dialysis dose (Kt/V, a UKM derived parameter) in seven thrice weekly hemodialyzed patients. Furthermore, the relationship between CTP and CHb concentrations in normal individuals and hemodialyzed patients were assessed.

Top

METHODS

Chemicals

N-Carbamyl-DL-valine (CVAL), N-carbamyl-DL-norvaline (CNVal), albumin (approx99%), and 4-aminodiphenylamine (semidine) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). High pressure liquid chromotomographic (HPLC) grade acetonitrile and acetic acid, diacetyl monoxime, and sulphuric acid were purchased from BDH Inc. (Toronto, ON, Canada). 2-Aminobenzoic acid was from Aldrich Chemical Co. Inc. (Milwaukee, WI, USA).

Patients

Seven stable outpatient hemodialyzed patients who were dialyzed thrice weekly in the Renal Dialysis Unit at the Hotel Dieu-Grace Hospitals (Grace site) were evaluated in the longitudinal study (selection based on availability of blood samples for six consecutive months). All patients had been on maintenance hemodialysis for more than one year and had not received any blood transfusions, nor were they on erythropoetin during the course of the study. Approval for the study was given by both the University of Windsor Ethics Committee and the Salvation Army Grace Hospital (now Hotel Dieu-Grace Hospitals), Windsor, Ontario, Canada.

For the cross-sectional study, a random selection of hemodialyzed patients (N = 13) and outpatients (N = 9) with normal urea and glucose, were chosen for the uremic group and normal group, respectively.

Blood samples

Ethylenediaminetetraacetic acid (EDTA) whole blood and serum were drawn during regular dialysis treatments and the routine tests performed on them. The whole blood samples were further processed for determination of CHb and CTP. These samples were stored overnight at 4°C routinely and up to two weeks in some cases. Although carbamylation does not cease at low temperatures40, no significant changes in CHb values were found during this time period. This is in agreement with Kwan et al21 and reflective of in vitro urea decomposition studies41. Only slight hemolysis (1 to 2%) was seen in a few samples. Red cells were separated from the plasma by centrifugation, washed three times with phosphate buffered saline (PBS) and resuspended in an equal volume of PBS. The plasma and the washed red blood cells were then used immediately or frozen at -20°C until required.

Measurements

Blood tests
 

Serum samples were analyzed for urea, creatinine, phosphate, total protein (TP), albumin, calcium (total and ionized), potassium, sodium, chloride, carbon dioxide, glucose, alkaline phosphatase, gamma-glutamyl transferase, and aspartate aminotransferase (Ektachem 700 or 700XR; Eastman Kodak Co., Rochester, NY, USA). Hemoglobin measurements were made using a total hemoglobin test kit (Sigma Diagnostics Canada, Mississauga, ON, Canada). Low concentrations of protein were measured manually using the Bio-Rad Protein Assay kit (Bio-Rad Laboratories Ltd., Mississauga, ON, Canada), with human serum albumin as standard.

Preparation of carbamylated albumin
 

Human serum albumin (99%) was carbamylated by incubation with 100 mM sodium cyanate in 0.1 M phosphate buffer, pH 7.4 at room temperature. The incubation time varied depending on the extent of carbamylation required. Unreacted cyanate was removed from the protein by gel filtration (Bio-Gel P-6G). The extent of carbamylation was estimated with the trinitrobenzene-sulfonic acid assay42. Cyanate removal was detected by reaction with 2-aminobenzoic acid43.

Measurement of protein carbamylation
 

The determination of protein carbamylation was based on the method of Hunninghake and Grisolia44. Plasma samples were first passed through a Sephadex G-25 column (1 times 8.5 cm) to remove urea and other potential interfering small molecules. Briefly, into a 10-ml borosilicate test tube the following was added in order: 1 ml sample or control (albumin or carbamylated albumin), 0.5 ml diacetylmonoxime (3% wt/vol), 0.5 ml H2SO4, and 0.5 ml semidine (0.2% wt/vol). The mixture was vortexed and heated for 10 minutes at 90°C. The mixture was then allowed to cool for 5 to 10 minutes (temperature 30° to 60°C). Then, 1 ml of a 0.001 M FeCl3-H2SO4 solution was added and mixed thoroughly. The mixture was cooled to room temperature and absorbances read within one hour at 525 nm. The extent of protein carbamylation was expressed in relative units (absorbance/mg protein) since no appropriate standard is available for this method. This assay was found to be linear with amount of carbamylated albumin (CHSA) tested. Carbamylated albumin (74%) was used as the positive control.

In vitro preparation of valine hydantoin
 

Valine hydantoin (VH) standard was prepared by incubating carbamyl valine (13.5 mg) with 11.6 M HCl (5 ml) for eight hours in a dri-bath set at 120°C and then left overnight at room temperature45. The solution was rotoevaporated until only a few drops of yellow liquid remained. The tube was then left under a vacuum overnight to remove the last traces of liquid. The resulting residue was slightly yellowish and had a melting point of 145°C. The residue was redissolved in approximately 3 ml of water and filtered through a 2-mum Acrodisc LC13 polyvinylidene fluoride (PVDF) syringe filter (Pall Gelman Sciences, Inc., Ann Arbor, MI, USA). The filtrate was then left to recrystallize overnight on a watch glass.

Preparation of valine hydantoin from hemoglobin
 

The preparation of valine hydantoin from hemoglobin was based on the method by Manning et al45. A 0.5 ml aliquot of red blood cells was placed into a heavy walled 15 ml screw top test tube. To this was added 5 ml of 2% (wt/vol) HCl in acetone (4°C). The mixture was vortexed, centrifuged for three minutes at 3000 times g and the supernatant discarded. The pellet was washed three times with 5 ml of acetone to remove all the heme. The off-white pellet was then washed with 5 ml of diethyl ether, centrifuged and heated in warm water (ca. 45°C) until dry. The pellet was then suspended in 0.5 ml of 50% glacial acetic acid and 0.5 ml of concentrated HCl. The tube was capped with a glass marble and the mixture heated for one hour at 100°C in a dri-bath. The globin is subsequently hydrolyzed. Carbamyl valine residues are released which then spontaneously cyclizes to form valine hydantoins. The hydrolyzate is cooled immediately in an ice bath. To this were added 0.8 ml of 10 N NaOH and 0.5 ml of saturated NaCl and the resulting suspension mixed well. The pH at this point should be between 4 to 5 for optimal extraction. Then 100 mul of the internal standard, 100 mg/ml carbamyl norvaline, was added. The valine hydantoin and carbamyl norvaline are then extracted by adding 5 ml of ethyl acetate and vortexing for one minute. The ethyl acetate supernatant (4 ml) was transferred to a clean 10-ml borosilicate tube and evaporated to dryness under a stream of air in a dri-bath (70°C). The residue was then placed over NaOH pellets in a vacuum dessicator overnight to remove the last traces of acid.

Measurement of valine hydantoin
 

The method of Kwan et al21 was adapted for the measurement of valine hydantoin. The mobile phase (6% acetonitrile) was prepared by first adjusting approximately 900 ml of Type I water (Milli-Q water system; Millipore Corp., Milford, MA, USA) to pH 4.0 with 0.5 M acetic acid, adding 60 ml acetonitrile and diluting to a final volume of 1 liter. The prepared extract was resuspended with 0.5 ml of the mobile phase, filtered through a 2-mum PVDF filter syringe, and 50 mul were autoinjected into a Shimadzu HPLC system (SCL-10A; Shimadzu Corporation, Kyoto, Japan) fitted with a heated (45°C) Apex II reversed-phase column (Jones Chromatography Ltd., Lakewood, CO, USA). The absorbance was read at 210 nm. Each run included two standards, 50 mul of valine hydantoin (10 mug/ml) and 50 mul of carbamyl norvaline (10 mug/ml).

Calculation of hemoglobin carbamylation
 

Carbamylated hemoglobin, expressed as mug valine hydantoin/g hemoglobin (mug VH/g Hb) was calculated according to the following formula:

Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

where A is the peak height of valine hydantoin (VH) in the sample, B is the peak height of the 50 mul (0.5 mug) CNVal (external standard), and C is the peak height of CNVal in sample (internal standard). The ratio 0.5 mug/B converts the peak height of the sample VH to mug VH, since the absorbance of VH is the same as CNVal at 210 nm. The ratio of B/C is a correction factor for loss during extraction and the value 20 adjusts for the volume injected (50 mul) compared to the total volume (1 ml) prepared from 0.5 ml of washed red blood cells. The amount of Hb in the sample is calculated by multiplying the concentration of Hb (g/liter) in the sample by the sample volume processed (0.5 ml).

The within-run coefficient of variation for the standards was 1.7% for both VH and CNVal. Between-run coefficient of variances for the standards VH and CNVal were 2.7% and 3.6%, respectively. Variation between samples was calculated from three patient samples prepared in duplicate and run in the same batch. The within-run difference was 10.2% for VH and 2.5% for CNVal.

Calculation of dialysis dose (Kt/V)
 

Dialysis dose (Kt/V), was calculated using a preprogramed calculator (Fresenius-Brent, Toronto, ON, Canada) incorporating the variable volume single-pool model developed by Gotch and Sargent46.

Statistical analysis

Simple correlation calculations were done by using Microcal Origin (Microcal Software Inc., Northhampton, MA, USA). Multiple regression analyses were performed using SAS (SAS Institute Inc., Cary, NC, USA), and weighted correlation coefficients were calculated using BMDP (BMDP Statistical Software Inc., Los Angelas, CA, USA).

Top

RESULTS

Carbamylated hemoglobin and carbamylated protein in normal individuals and hemodialyzed patients

Figure 1 shows the relationship between carbamylated hemoglobin and carbamylated total protein in normal (N = 9) and uremic patients (N = 13). The overall correlation of carbamylated hemoglobin with carbamylated total protein was highly significant (P < 0.0001) with a correlation of 0.87 (0.01). Individually, however, the normal group shows no association between CHb and CTP values (r = 0.07, P = 0.85), and the uremic group shows a fair association (r = 0.61, P = 0.026). Furthermore, a significant difference was found between the normal and uremic groups with carbamylated hemoglobin (P < 0.0001) and carbamylated total protein (P < 0.0001). The mean value for carbamylated hemoglobin in normal individuals was 53 (plusminus 20) mug VH/g Hb, and in uremic patients it was 157 (plusminus 40) mug VH/g Hb. The mean value for carbamylated total protein in normal individuals was 0.08 (plusminus 0.009) A/mg protein, and in uremic patients it was 0.117 (plusminus 0.011) A/mg protein.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Correlation between carbamylated plasma protein (CTP) and carbamylated hemoglobin (CHb) in normal and uremic subjects. Relationship between carbamylated hemoglobin and carbamylated total protein in 9 normal (filled circle) and 13 uremic (filled square) patients. The regression line was y = 2.9E-4x + 0.069 (r = 0.87, SD = 0.01, P < 0.0001). Duplicate determinations were done for each sample.

Full figure and legend (17K)

The relationship between carbamylated hemoglobin and carbamylated total protein also shows that the y-intercept crosses at 0.069 (absorbance/mg protein). This is a very high y-intercept considering that the highest y-value is only about twice that value. This may suggest that carbamylated total protein can be detected sooner than carbamylated hemoglobin.

Longitudinal study: Correlation with uremic markers

The results of the longitudinal study comparing urea (pre- and post-dialysis), Kt/V, carbamylated protein and carbamylated hemoglobin in seven patients are shown in Figure 2. There did not appear to be any clear relationships, or discernable patterns between any sets of curves that were consistent among all patients.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Serial measurements of uremic markers in hemodialyzed patients. A six-month longitudinal study of traditional dialysis markers, urea and Kt/V, with time-averaged urea markers, carbamylated hemoglobin (CHb) and carbamylated protein (CTP), in seven thrice weekly hemodialyzed patients. Complete data were unavailable for patients 5 and 7. Patients 3 and 6 were diabetic. Symbols are: (filled triangle) pre-urea; (triangle) post urea; (dotdotdotcircledotdotdot) CHb,mug VH/g Hb; (–filled square–) Kt/V, (dotdotdotfilled circledotdotdot) CTP, A/mg protein.

Full figure and legend (41K)

Because of the very different half-lives of carbamylated protein (3 weeks) and carbamylated hemoglobin (2 months), correlations done at single times are of little value in assessing the relative clinical utility of the two assays. Furthermore, it is difficult to compare parameters that have different measurement scales. To accommodate these obstacles another approach was taken to analyze the data. The data were recalculated in terms of relative percent change over time, that is, the initial time interval was set to 0% change and each subsequent value was expressed relative to the previous value in the time series. The mean and SD were calculated for each test using the mean relative percent value for each patient. Kt/V values varied the least (6.3%plusminus 3.3%) over this time period, whereas CHb varied nearly five times more (30.1%plusminus 20.2). Pre-dialysis urea and CTP values had similar relative fluctuations (18.3%plusminus 13.4% and 14.9%plusminus 7.5%, respectively), between those of Kt/V and carbamylated hemoglobin values. This indicates that CHb is a more sensitive in monitoring urea fluctuations than either pre-dialysis urea, CTP, or the dialysis dose, Kt/V.

Correlation calculations (using the relative % change data) did not seem to reveal much more in terms of relationships than did the visual observations of the initial data. No correlations between any two tests were seen in most cases. Furthermore, the correlations that did exist between any two tests were mixed, positive or negative, and in very few cases were these correlations significant (P < 0.05). The reason for these apparent ambiguities could be due to the nature of the tests, that is, they do not measure the same thing at the same time, or the lack of a larger sample size (frequency of sampling and number of patients).

Correlations between pre-dialysis urea and CHb, and pre-dialysis urea and CTP on the data sets (N = 7) were done for each time period (N = 6). There was a much higher association between pre-dialysis urea and CHb (r = 0.97, 0.86, 0.71, 0.52, 0.51, 0.40) than with pre-dialysis urea and CTP (r = 0.48, 0.77, -0.11, 0.66, 0.26, 0.08).

To achieve greater statistical power, multiple regression analysis was done to statistically answer the question as to whether there is a relationship between any two tests within a patient. The overall correlation coefficient was obtained by removing the variation due to individual subjects47. No significant correlation was found with any combination of tests, which is probably a reflection of the variation in the correlations of the individual patients, that is, regression to the mean.

Furthermore, the data were also looked at in terms of between subject correlation. That is, to answer the question whether on average do patients with high (or low) values for test A also tend to have high (or low) values for test B. The weighted correlation coefficients were determined between various tests using the subject means. The significance of the correlation was based on the number of subjects (N = 7) and not the number of data pairs (N = 40) because of the repeated observations per subject48. Significant correlations (at alpha= 0.05, r = 0.6786) were found between CHb versus CTP (P = 0.707) and CHb versus pre-dialysis urea (P = 0.671). The correlation between CTP versus pre-dialysis urea was almost significant (P = 0.534), and there was a very poor correlation between Kt/V versus pre-dialysis urea (P = 0.278). The other comparisons done were between Kt/V versus CTP or CHb (P=-0.327 and -0.387, respectively).

Laboratory tests for the seven hemodialyzed patients showed that all had low hemoglobin concentrations and low hematocrits. Albumin values were within the reference interval and pre-dialysis urea values were above the reference interval. Post-dialysis urea values were all lower than pre-dialysis urea values and in some patients within the reference range. Creatinine concentrations were all well above the reference interval before dialysis and decreased after dialysis, but remained above the reference interval. Glucose was elevated only in the two diabetic patients. Values for potassium, sodium, chloride, carbon dioxide, glucose, calcium (total and ionized), phosphate, alkaline phosphatase, gamma-glutamyl transferase, and aspartate aminotransferase were in the reference range for most patients. No relationships were evident in cases where there were elevated levels of an analyte when compared to CHb, CTP or Kt/V values.

Top

DISCUSSION

Carbamylated hemoglobin and carbamylated plasma protein in normal individuals and hemodialyzed patients

The correlation between carbamylated hemoglobin and carbamylated total protein concentrations in normal individuals and uremic subjects Figure 1 showed a high correlation (r = 0.87). Furthermore, the y-intercept showed a bias in favor of carbamylated total protein. Similarly, in glycation studies the bias favors fructosamine and corroborates other studies that show fructosamine as a shorter term indicator of glycemic control compared to glycated hemoglobin49.

Since assay methodology could also influence this bias, a general idea for how much carbamylation (number of carbamyl groups independent of site or rate of attachment) occurs on albumin versus hemoglobin was obtained by interpolating data from two representative studies14,21. The amount of carbamylation was assessed by comparing homocitrulline residues for albumin and carbamyl valine residues for hemoglobin. This comparison (molar ratio) showed that there was fivefold more carbamyl groups on albumin [0.08 (0.06 to 0.12)] as compared with hemoglobin (0.016 plusminus 0.005) in normal individuals. Similarly, there were also more carbamyl groups on albumin [0.027 (0.02 to 0.033)] than on hemoglobin (0.066 plusminus 0.035) in patients with renal failure (4-fold). Therefore, the potential for carbamylation is greatest for plasma proteins where the concentration of urea and cyanate are the highest.

Longitudinal study

This is the first longitudinal study that has looked at the changes in carbamylated hemoglobin and carbamylated protein in hemodialyzed patients. Parameters of dialysis dose (Kt/V), pre- and post-dialysis urea, CHb and CTP for all seven patients were measured and plotted with respect to the time interval (1 month). Visual observation of these parameters did not show any strong associations or trends Figure 2.

Although the correlations between pre-dialysis urea and CHb were high for most time periods, they did not agree and few were significant. The large range in correlation values (0.21, 0.51, 0.52, 0.71, 0.86, 0.98) suggests that there is no consistent relationship between pre-dialysis urea concentration and CHb. This discrepancy is probably due to the small sample size (N = 7) or could also be reflecting the large fluctuations in urea generated by dialysis. Studies with larger sample sizes have shown inconsistent relationships as well. Kwan et al21 showed a correlation of 0.69 (N = 20) in hemodialysis patients for pre-dialysis urea versus CHb. However, Smith et al22, using a similar method, showed a correlation of only 0.44 (N = 31) in hemodialysis patients. They also found higher correlations in all other groups studied (control r = 0.55, N = 20; chronic renal failure r = 0.75, N = 40; continuous ambulatory peritoneal dialysis r = 0.55, N = 24; transplant rejection r = 0.87, N = 14). Another study reported a correlation of 0.84 in a combined group of normal and uremic subjects19.

The correlations between pre-dialysis urea and CTP among the six time periods were very poor and widespread (-0.38, -0.11, 0.07, 0.48, 0.66, 0.77). This variation could also be due to the same factors as discussed above.

In general, the half-lives of carbamylated proteins (CTP) and carbamylated hemoglobin (CHb) can be assumed to be similar to the half-lives of plasma proteins and hemoglobin (3 weeks and 2 months, respectively). Therefore, detecting a relationship between the CTP and CHb values when measurements are taken only every 30 days is therefore difficult Figure 2. If time points had been available weekly, the possibility of detecting a trend may have increased. Also, since the patients studied were in good control already (dialysis regime and diet), very little variation or significant difference in any individual test was seen. Untreated patients or patients on continuous ambulatory peritoneal dialysis may have shown more suggestive results.

This longitudinal study showed no clear relationship between pre-dialysis urea or Kt/V and CTP or CHb measurements. Because hemodialysis patients are maintained on a fairly stable dialysis regime, the utility of measuring carbamylated proteins may be limited. However, they may be of value if concentrations are above a predetermined cutoff value or a pattern emerges over time. Kwan et al23 reported that CHb concentrations < 175 mug CVal/g Hb indicated adequate dialysis.

The interpretation of carbamylated hemoglobin concentration can be complicated by the anemia in uremia, the shortened life-span of erythrocytes, and transfusions that dilute the erythrocyte mass and the CHb concentration. Measurement of CTP, however, would not be accompanied by these same problems. The plasma protein content is relatively stable in patients with end-stage renal disease, although lower than normal (nutritional effect)50.

Nonetheless, measurement of carbamylated hemoglobin may reflect a therapeutic time interval that would be appropriate in the clinical management of the uremic outpatient. Carbamylated protein concentration, an index of integrated urea levels over the preceding two weeks, may reflect adequacy of dialysis over too short a time interval to be of much practical use in and outpatient setting. However, in cases of acute renal failure the estimation of carbamylated protein may be more appropriate than carbamylated hemoglobin25.

In summary, to our knowledge this study has for the first time shown the relationship between CHb and CTP in hemodialyzed patients. It has also found no correlation between these measurements and Kt/V, the currently used, but poorly evaluated39, indicator of morbidity and mortality in the hemodialyzed patient. Therefore, the data suggest that CHb and CTP are unique indicators of uremia, and their potential value necessitates their examination in a valid outcome study.

Top

References

REFERENCES

  1. DIRNHUBER, P & SCHÜTZ, F: The isomeric transformation of urea into ammonium cyanate in aqueous solutions. Biochem J 1948 42:628–632,  | ISI | ChemPort |
  2. STARK, GR: Reactions of cyanate with functional groups of proteins. III. Reactions with amino and carboxyl groups. Biochemistry 1965 4:1030–1036,  | Article | PubMed | ISI | ChemPort |
  3. BACHMANN, K, VALENTOVIC, M, SHAPIRO, R: A possible role for cyanate in the albumin binding defect in uremia. Biochem Pharmacol 1981 29:1598–1601,
  4. CARRERAS, J, CHABAS, A, DIEDERICH, D: Physiological and clinical implications of protein carbamylation, in The Urea Cycle, 1976, edited by Grisolia S, Baguena R, Major F, New York, John Wiley & Sons, Inc., pp 501–548
  5. BALION, CM: Development and evaluation of methods for the determination of carbamylated proteins in hemodialyzed patients. (Dissertation). University of Windsor, Windsor, ON, 1996
  6. MARTIN, S & HARDING, JJ: Site of carbamoylation of bovine gamma-II-crystallin by potassium [14C]cyanate. Biochem J 1989 262:909–915,  | PubMed |
  7. HERRERA-ORNELAS, L, PETRELLI, NJ, MADAJEWICZ, S, MITTELMAN, A, ALLFREY, VG: Phase-I clinical trial of sodium cyanate in patients with advanced colorectal carcinoma. Oncology 1985 42:236–241,  | PubMed |
  8. SCHÜTZ, F: Cyanate. Experientia 1949 5:133–172,  | Article |
  9. HÖRKKÖ, S, HUTTUNEN, K, KESÄNIEMI, YA: Decreased clearance of low-density lipoprotein in uremic patients under dialysis treatment. Kidney Int 1995 47:1732–1740,  | PubMed | ISI | ChemPort |
  10. STEINBRECHER, UP, FISHER, M, WITZTUM, JL, CURTISS, LK: Immunogenicity of homologous low density lipoprotein after methylation, ethylation, acetylation, or carbamylation: Generation of antibodies specific for derivatized lysine. J Lipid Res 1984 25:1109–1116,  | PubMed |
  11. KRAUS, LM, MIYAMURA, S, PECHA, BR, KRAUS, AP JR: Carbamoylation of hemoglobin in uremic patients determined by antibody specific for homocitrulline (carbamoylated set symbol-N-lysine). Mol Immunol 1991 28:459–463,  | PubMed |
  12. ERILL, S, CALVO, R, CARLOS, R: Plasma protein carbamylation and decreased acidic drug protein binding in uremia. Clin Pharmacol Ther 1980 27:612–618,  | PubMed |
  13. DENGLER, TJ, ROBERTZ-VAUPEL, GM, DENGLER, HJ: Albumin binding in uraemia: Quantitative assessment of inhibition by endogenous ligands and carbamylation of albumin. Eur J Clin Pharmacol 1992 43:491–499,  | Article | PubMed | ISI | ChemPort |
  14. BACHMANN, K, VALENTOVIC, M, SHAPIRO, R: Contribution of cyanate to the albumin binding defect of uremia. Biochem Pharmacol 1981 30:1059–1063,  | PubMed |
  15. KRAUS, LM & KRAUS, AP JR: Tyrosine and N-carbamyl-tyrosine in end-stage renal disease during continuous ambulatory peritoneal dialysis. J Lab Clin Med 1991 118:555–562,  | PubMed |
  16. KRAUS, LM, ELBERGER, AJ, HANDORF, CR, PABST, MJ, KRAUS, AP: Urea-derived cyanate forms set symbol-amino-carbamoyl-lysine (homocitrulline) in leukocyte proteins in patients with end-stage renal disease on peritoneal dialysis. J Lab Clin Med 1994 123:882–891,  | PubMed |
  17. TREPANIER, DJ, CAINES, PS, THIBERT, RJ, GOODWIN, M, DRAISEY, TF: Carbamylation of normal and uremic erythrocyte membrane protein. Clin Biochem 1995 28:332,
  18. OIMOMI, M, ISHIKAWA, K, KAWASAKI, T, KUBOTA, S, YOSHIMURA, Y, BABA, S: Plasma carbamylated protein in renal failure. N Engl J Med 1983 308:655–656,
  19. KOSHIISHI, I & IMANARI, T: State analysis of endogenous cyanate ion in human plasma. J Pharmacobio-Dyn 1990 13:254–258,  | PubMed |
  20. BALION, CM, CAINES, PS, DRAISEY, TF, THIBERT, RJ: Carbamylated hemoglobin and carbamylated plasma proteins in normal subjects and hemodialyzed patients. (abstract) Clin Chem 1995 41:S153–S154,
  21. KWAN, JTC, CARR, EC, BENDING, MR, BARRON, JL: Determination of carbamylated hemoglobin by high-performance liquid chromatography. Clin Chem 1990 36:607–610,  | PubMed |
  22. SMITH, WGJ, HOLDEN, M, BENTON, M, BROWN, CB: Carbamylated haemoglobin in chronic renal failure. Clin Chim Acta 1988 178:297–304,  | PubMed |
  23. KWAN, JTC, CARR, EC, NEAL, AD, BURDON, J, RAFTERY, MJ, MARSH, FP, BARRON, JL, BENDING, MR: Carbamylated haemoglobin, urea kinetic modelling and adequacy of dialysis in haemodialysis patients. Nephrol Dial Transplant 1991 6:38–43,  | PubMed |
  24. KWAN, JTC, CARR, EC, BARRON, JL, BENDING, MR: Carbamylated haemoglobin in normal, diabetic and uraemic patients. Ann Clin Biochem 1992 29:206–209,  | PubMed |
  25. DAVENPORT, A, JONES, SR, GOEL, S, ASTLEY, JP, HARTOG, M: Differentiation of acute from chronic renal impairment by detection of carbamylated haemoglobin. Lancet 1993 341:1614–1617,  | PubMed |
  26. KRAUS, AP JR, SONI, P, STEPHENS, MC, KRAUS, LM: Carbamylated hemoglobin in uremia: Continuous ambulatory peritoneal dialysis (CAPD) versus hemodialysis. (abstract) Blood 1983 62:47a,
  27. KWAN, JTC, CARR, EC, BARRON, JL, BENDING, MR: Carbamylated haemoglobin–A retrospective index of time-averaged urea concentration. Nephrol Dial Transplant 1993 8:565–567,  | PubMed |
  28. FLÜCKIGER, R, HARMON, W, MEIER, W, LOO, S, GABBAY, K: Hemoglobin carbamylation in uremia. N Engl J Med 1981 304:823–827,  | PubMed |
  29. OIMOMI, M, ISHIKAWA, K, KAWASAKI, T, KUBOTA, S, YOSHIMURA, Y, BABA, S: Carbamylation of hemoglobin in renal failure and clinical aspects. Metabolism 1984 33:999–1002,  | PubMed |
  30. ATIONU, A: Carbamylation and glycosylation of haemoglobin in vitro: Effects of cyanate and glucose. Med Lab Sci 1992 49:34–37,  | PubMed |
  31. STIM, J, SHAYKH, M, ANWAR, F, ANSARI, A, ARRUDA, JAL, DUNEA, G: Factors determining hemoglobin carbamylation in renal failure. Kidney Int 1995 48:1605–1610,  | PubMed | ISI | ChemPort |
  32. BERNSTEIN, RE: Nonenzymatically glycosylated proteins. Adv Clin Chem 1987 26:1–78,  | PubMed |
  33. ARMBRUSTER, DA: Fructosamine: Structure, analysis, and clinical usefulness. Clin Chem 1987 33:2153–2163,  | PubMed |
  34. CERAMI, A: Cyanate as an inhibitor of red-cell sickling. N Engl J Med 1972 287:807–812,  | PubMed |
  35. DAVENPORT, A, JONES, S, GOEL, S, ASTLEY, JP, FEEST, TG: Carbamylated hemoglobin: A potential marker for the adequacy of hemodialysis therapy in end-stage renal failure. Kidney Int 1996 50:1344–1351,  | PubMed |
  36. HAN, JS, KIM, YS, CHIN, HJ, JEON, US, AHN, C, KIM, S, LEE, JS, JANG, I-J, SHIN, S-G: Temporal changes and reversibility of carbamylated hemoglobin in renal failure. Am J Kidney Dis 1997 30:36–40,  | PubMed |
  37. FRAZAO, JM, BARTH, RH, BERLYNE, GM: Carbamylated hemoglobin in prerenal azotemia. Nephron 1995 71:153–155,  | PubMed |
  38. CALVO, R, CARLOS, R, ERILL, S: Underestimation of albumin content by bromocresol green, induced by drug displacers and uremia. Biochim Biophys Acta 1988 954:208–215,  | PubMed |
  39. BARTH, RH: Urea modelling and Kt/V: A critical appraisal. Kidney Int 1993 43: (Suppl 41) S252–S260,  | ISI |
  40. HAGEL, P, GERDING, JJT, FIEGGEN, W, BLOEMENDAL, H: Cyanate formation in solutions of urea: I. Calculation of cyanate concentration at different temperature and pH. Biochem Biophys Acta 1971 243:366–373,  | PubMed |
  41. MARIER, JR & ROSE, D: Determination of cyanate, and a study of its accumulation in aqueous solutions of urea. Anal Biochem 1964 7:304–314,  | Article | ISI | ChemPort |
  42. SNYDER, SL & SOBOCINSKI, PZ: An improved 2,4,6-trinitrobenzenesulfonic acid method for the determination of amines. Anal Biochem 1975 64:284–288,  | Article | PubMed | ISI | ChemPort |
  43. GUILLOTON, M & KARST, F: A spectrophotometric determination of cyanate using reaction with 2-aminobenzoic acid. Anal Biochem 1985 149:291–295,  | PubMed |
  44. HUNNINGHAKE, D & GRISOLIA, S: A sensitive and convenient micromethod for estimation of urea, citrulline, and carbamyl derivatives. Anal Biochem 1966 16:200–205,  | PubMed |
  45. MANNING, JM, LEE, CK, CERAMI, A, GILLETTE, PN: Gas chromatographic determination of the carbamylation of hemoglobin S by cyanate. J Lab Clin Med 1973 81:941–945,  | PubMed |
  46. GOTCH, FA & SARGENT, JA: A mechanistic analysis of the National Cooperative Dialysis Study (NCDS). Kidney Int 1985 28:526–534,  | PubMed | ISI | ChemPort |
  47. BLAND, JM & ALTMAN, DG: Calculating correlation coefficients with repeated observations: Part 1 – correlation within subjects. Br Med J 1995 310:466,
  48. BLAND, JM & ALTMAN, DG: Calculating correlation coefficients with repeated observations: Part 2 – correlation between subjects. Br Med J 1995 310:633,
  49. NEGORO, H, MORLEY, JE, ROSENTHAL, MJ: Utility of serum fructosamine as a measure of glycemia in young and old diabetic and non-diabetic subjects. Am J Med 1988 85:360–364,  | PubMed |
  50. LINDSAY, RM, HEIDENHEIM, AP, SPANNER, E, KORTAS, C, BLAKE, PG: Adequacy of hemodialysis and nutrition: Important determinants of morbidity and mortality. Kidney Int 1994 45: (Suppl 44) S85–S91,
Top

Acknowledgments

This work was supported by a research grant to Dr. R.J. Thilbert from the Natural Sciences and Engineering Research Council of Canada and from the University of Windsor Research Board. The authors thank Dr. B.S. Reen (Consultant Nephrologist, Etobicoke General Hospital, Etobicoke, Toronto, ON, Canada), and Dr. P.S. Raina (Consultant Epidemiologist, McMaster University, Hamilton, ON, Canada) for helpful discussions during the preparation of this manuscript.

Extra navigation

.
ADVERTISEMENT