Interpreting Glycosylated . Hemoglobin in Diabetic Patients on Peritoneal Dialysis

Antonios H. Tzamaloukas

The limitations and usefulness of glycosylated hemoglobin (HbA) as a means of assessing glycemic control in peritoneal dialysis (PD) and the importance of good glycemic control in diabetics on P D were analyzed. HbA J levels may be elevated in renal failure because of either interference with the assay of HbA J by carbamylated hemoglobin, which is formed by dissociation of urea to cyanate and condensation of cyanate with hemoglobin, or increased rate of formation of HbA J secondary to acidosis or other uremic effects. Interference with the assay of HbA J is limited to methods separating HbA J from hemoglobin A by electrical charges and is not encountered when chemical or immunologic assays are used. The proposed effect of uremia on the rate offormation of HbA J was refuted by in vitro studies and in vivo multifactorial analysis, which showed that blood glucose level is the only important determinant of HbA J in patients with renal failure. HbAJ levels may be low in uremia because of shortened red cell survival and/or frequent transfusions of red cells, conditions rarely encountered in P D patients. In the great majority of the reported studies, HbA I' regardless of the method of assay used; correlated with direct glycemic indices in patients with renal failure treated conservatively, by hemodialysis or by PD. Poor glycemic control, indicated by high HbA J levels, is associated with adverse outcomes, including prolonged hospitalization and shortened survival, in diabetics on PD. HbAJ is a useful tool in assessing blood glucose control in PD. Good glycemic control is important for diabetics on PD.

Key words

Blood glucose concentration, carbamylated hemoglobin, glycosylated hemoglobin, diabetes mellitus, renal failure, uremia

From:

Veterans Affairs Medical Center and University of New Mexico, Albuquerque, New Mexico, U.S.A.

Introduction

Glycosylated hemoglobin (HbA.) is extensively used to assess glycemic control in peritoneal dialysis (PD) patients (1-5). Nevertheless, HbA. has potential pitfalls in PD (6). This review will address the following questions: (I) Is HbA1 appropriate for assessing long-term glycemic control in PD patients? (2) Are there any qualifiers to the answer to the fIrst question? (3) Is glycemic control important for diabetics on PD?

Synthesis of glycosylated hemoglobins

Glycosylated hemoglobins are separated chromatographically from hemoglobin A (HbA) because they migrate faster than HbA. There are four fast fractions: HbA.al' HbA1a2' HbA.b, and HbA.c. HbA.c' which is the largest fraction and the most consistent index of glycemia, is formed by post-translational nonenzymatic condensation of glucose with the N-terminal amino groups of the beta-chains of HbA (7). This condensation forms a stable ketoamine (Amadori reaction) in a process similar to the formation of the advanced glycosylation end products. Injections ofS9Fe-transferrin demonstrated in vivo continuous, slow formation ofHbAI throughout the life span of the red cells (7).

HbA J as a measure of glycemic control in diabetics without renal failure

The use of HbA1 as a measure of glycemic control in diabetics without renal failure was based on the demonstration of positive correlations between HbA. and different direct and indirect indicators of glycemia. High correlation coefficients were reported between HbAl and fasting blood glucose in long-term hospitalized diabetics (8), HbAl and physician rating of glycemic control, HbAl and fasting blood glucose in both hospitalized subjects and outpatients, and between HbA1 and random or peak blood glucose levels in diabetics with continuous metabolic monitoring (9). Significant, but lower, correlations were reported between HbAl and antecedent 24-hour excretion of glucose in urine, particularly when the urine collection preceded the HbAl measurement by two months (10). In the same study, HbA1 correlated with serum cholesterol (10).

A common finding of the studies of HbAl in diabetics without renal failure was a consistent decrease in HbAl levels as blood glucose levels were lowered. It was estimated that a decrease in HbAl by 1% corresponds approximately to a decrease in blood glucose by 35 mg/dl (11).

Prob/ems of HbAJ in PD

Table I shows conditions other than glycemic control that can cause changes in HbAl measured by chromatography (12). Independently of glycemic control, renal failure may cause either a rise or a fall in HbAl. Elevated levels of HbAl in renal failure may represent either interference with the assay of HbAl by compounds present in the uremic blood or altered (increased) rate of formation of HbAI. The identified compound interfering with the assay of HbAl is carbamylated hemoglobin, which is formed by post-translational nonenzymatic condensation of cyanate with the N-terminal amino groups of hemoglobin A (13,14). Spontaneous.dissociation ofurea leads to the formation of cyanate, the reactive form of which, isocyanic acid, reacts with the amino groups to form carbamylated hemoglobin. This stable compound is formed continuously during the life span of the red cells, representing thus an index of the integrated blood urea concentration in azotemic subjects (13). This is analogous to the use of HbAl as an index of integrated glycemia in diabetics.

Carbamylated hemoglobin coelutes with HbAl in a cation-exchange chromatographic column. The measurement ofHbAla and HbAlb is affected by the presence of carbamylated hemoglobin to a far greater extent than that of HbAlc (13,15). Carbamylated hemoglobin interferes with the measurement of HbAl only when assay methods are based on electrical charges. Chemical or other assay methods are not affected. Table II shows assay methods reported to be affected and those unaffected by the presence in the blood of carbamylated hemoglobin.

Other proposed effects of uremia on HbAl are through an increase in the rate offormation ofHbA1. A negative correlation between HbA1 and blood pH or bicarbonate was reported in patients with chronic renal failure treated conservatively or by hemodialysis (22,23). This finding was interpreted as indicating an increased rate of formation of HbAl in the presence of uremic acidosis (23). Finally, auremic defect in the process of protein glycosylation was postulated; this defect is partially corrected by dialysis (24).

An altered rate of glycosylation in uremia was not found by advanced study techniques. In vitro, the rate of hemoglobin glycosylation, measured by affinity chromatography, was the same in azotemic and nonazotemic environments (25). By multifactorial analysis, serum bicarbonate was not among the independent correlates ofHbAl in patients with renal failure treated conservatively, by hemodialysis or by PD (17,21). In the last study, serum glucose accounted for 0.54 (r2) of the variation in HbA1 measured by affmity chromatography, while a small improvement of r2 to 0.60 was found when serum urea and creatinine were added to the multiple regression model (21). Therefore, the rates ofhemoglobin glycosylation are either the same or only slightly different between uremia and normal renal function. It should also be noted that acidosis is infrequent in patients on continuous PD.

A decrease in HbA1 in uremia can be caused by decreased red cell life span or dilution from frequenttransfusions (26,27). However, shortened red cell survival has not been reported in PD patients, who even before the introduction of erythropoietin as a means to correct anemia in dialysis patients had satisfactory hematocrit values and rarely, if ever, required transfusions.

In summary, HbA1 metabolism does not appear to be altered in PD patients. The only influence on HbA1 levels that needs to be seriously considered in PD is the interference of carbamylated hemoglobin with the assays ofHBAl that are based on its separation from HbA by electrical charges. These methods are extensively used by clinicallaboratories. Even assays receiving interference by carbamylated hemoglobin are useful, because serum urea concentration varies little in stable PD patients (28). Consequently, changes in HbA1 levels in PD subjects usually indicate variations in glycemia, not changing blood urea levels.

Correlation of HhA, and glycemia in PD

Table III shows the correlations, found by bibliographic search, between HbA, and glycemia in dialysis patients. Lack of correlation was found in a small number of studies, which did not include diabetics among the study subjects. Significant correlations were reported in studies including diabetic subjects, regardless of the method of assay ofHbA1. However, the correlation coefficients tended to be higher when HbA1 was assayed by methods not receiving interference by carbamylated hemoglobin.

A different comparison used cut-off levels of either frequency of hyperglycemia plus hypoglycemia or HbA" measured by affinity chromatography, to characterize glycemic control as good or poor. In 226 diabetics on hemodialysis or PD, the classification by frequency of abnormal blood glucose measurements agreed closely with the classification by high HbAI, with a kappa ratio of 0. 810 (35). Finally, unlike the levels of fructosamine, which are affected, those of HbA1 are not affected by hemodialysis (33). In conclusion, HbA1 is clearly useful as a measure of glycemic control in diabetics on dialysis. Thitiarchakul et al. estimated that in diabetics on PD a 1% change in HbAl corresponds to an approximate change of 26 mg/dL in serum glucose concentration (34).

Importance of glycemic control, judged by HbA I' in diabetics on PD

This question was addressed in studies on diabetics with or without renal failure. In diabetics without renal failure, poor glycemic control, indicated by high HbA1 levels, correlates with abnormalities in erythrocyte, leukocyte, and platelet function, unfavorable changes in the coagulation cascade and in the plasma lipid profile, and increased tone of the blood vessels (36). In addition, poor glycemic control is associated with the acute morbidity and mortality of hypertonicity, hyperkalemia, hypoglycemia, and diabetic acidosis in diabetics with or without renal failure. An important sequela of hyperglycemia in diabetics on dialysis, not found in diabetics with preserved renal function, is the development of symptomatic fluid gain (37-39). Finally, in a study of 110 diabetics on PD with an average follow-up of 22.4 months, those with HbA1 > 10% had higher serum cholesterol concentration, longer annual hospitalization rate, and shorter patient survival than those with HbA1 < 10% (40).

Summary

This report can be summarized by answering the questions posed in the Introduction: (I) HbA1 is a useful tool in assessing integrated glycemic levels in diabetics on PD; (2) Although all assay methods provide meaningful results, HbA1 assay methods that are free from carbamylated hemoglobin interference should be preferable; and (3) good glycemic control is important for diabetics on PD.

References

  1. Amair P, Khanna R, Leibel B, et a/. Continuous . ambulatory peritoneal dialysis in diabetics with end stage renal disease. N Engl J Med 1982; 306:625-30.
  2. Madden MA, Zimmerman SW, Simpson DP. Continuous ambulatory peritoneal dialysis in diabetes mellitus. The risks and benefits of intraperitoneal insulin. Am JNephrol1982;2:133-9.
  3. Grefberg N, Danielson BG, Nilsson P. Continuous ambulatory peritoneal dialysis in the treatment of end-stage diabetic nephropathy. Acta Med Scand 1984; 215:427-34.
  4. Tzamaloukas AH, Rogers K, Ferguson BJ, et a/. Management of diabetics on CAPD with subcutaneous insulin. In: Khanna R, Nolph KD, Prowant BF, Twardowski ZJ, Oreopoulos DG, eds. Advances in continuous ambulatory peritoneal dialysis. Toronto: Peritoneal Dialysis Bulletin Inc., 1988; 4:126-32.
  5. Rottembourg J, Issad B, Allouache M, Baumelou A, Deray G, Jacobs C. Clinical aspects of continuous ambulatory and continuous cyclic peritoneal dialysis in diabetic patients. Perit Dial Int 1989; 9:289-94.
  6. Tzamaloukas AH, Oreopoulos DG. Subcutaneous versus intraperitoneal insulin in the management of diabetics on CAPD: a review. In: KhannaR,Nolph KD, Prowant BF, Twardowski ZJ, Oreopoulos DG, eds. Advances in peritoneal dialysis. Toronto: Peritoneal Dialysis Bulletin Inc., 1991; 7:81-5.
  7. Bunn HF, Haney DN, Kamin S, Gabbay KH, Gallop PM. The biosynthesis of human hemoglobin Alc. Slow glycosylation ofhemoglobin in viva. J Clin Invest 1976; 57: 1652-9.
  8. Koenig RJ, Peterson CM, Jones RL, Saudek C, Lehrman M, Cerami A. Correlation of glucose regulation and hemoglobin A1c in diabetes mellitus. N Engl J Med 1976; 295:417-20.
  9. Gonen B, Rubenstein AH, Tanega SP, Horwitz DL. Haemoglobin Al : an indicator of the metabolic control of diabetic patients. Lancet 1977; 2:734-7.
  10. Gabbay KH, Hasty K, Breslow JL, Ellison RC, Bunn HP, Gallop PM. Glycosylated hemoglobin and longterm blood glucose control in diabetes mellitus. J Clin Endocrinol Metab 1977; 44:859-(; 1.
  11. Goldstein DE. Is glycosylated hemoglobin clinically useful?NEnglJMed 1984;310:384-5. .
  12. Health and Public Policy Committee, American College ofPhysicians. Glycosylated hemoglobin assays in the management and diagnosis of diabetes mellitus. Ann IntemMed 1984; 101:710-13.
  13. Fluckiger R, Harmon W, Meier W, Loo S, Gabbay H. Hemoglobin carbamylation in uremia. N Engl J Med 1981; 304:823-7.
  14. Oimomi M, Ishikawa K, Kawasaki T, Kubota S, Yoshimura Y, Baba S. Carbamylation ofhemoglobin in renal failure and clinical aspects. Metabolism 1984; 33:999-1002.
  15. Smith WG, Holden M, Benton M, Brown CB. Glycosylated and carbamylated haemoglobin in uraemia. Nephrol Dial Transplant 1989; 4:96-100.
  16. Panzetta G, Bassetto MA, Feller P, et al. Microchromatographic measurement of hemoglobin Ai in uremia. Clin Nephro11983; 20:259-62.
  17. Paisey R, Banks R, Holton R, et al. Glycosylated haemoglobin in uraemia. Diabetic Med 1986; 3 :44H.
  18. Morgan LJ, Marenah CB, Morgan AG, Burden RP , John WG. Glycated haemoglobin and fructosamine in non-diabetic subjects with chronic renal failure. Nephrol Dial Transplant 1990;5:868-73.
  19. Tiran A, Pieber T, Tiran B, Halwachs-aumann G, Wilders Truschnig M. Photometric determination of HbA.c levels in clinical practice. Dtsch Med Wochenschr 1994; 119:833--6.
  20. Blumberg A, Marti HR. Hemoglobin Aic in patients under periodic hemodialysis or continuous ambulatory peritoneal dialysis. Schweiz Med Wochenschr 1986; 116:208-11.
  21. TzamaloUkas AH, Hsi KC, Quintana BI, Merlin TL, Avasthi PS. Glycosylated hemoglobin measured by atfmity chromatography in diabetic and nondiabetic patients on long-term dialysis therapy. West I Med 1989; 150:415-19.
  22. De Marchi S, Cecchin E, CamulTi C, et al. Origin of glycosylated hemoglobin Ai in chronic renal failure.lnt I ArtifOrgans 1983; 6:77-82.
  23. De Marchi S, Cecchin E, Basile A, et al. More on the increase in hemoglobin Ai in chronic renal failure: the role ofacidosis. Nephron 1983; 35:49-53.
  24. Sabater I, Quereda C, Herrera I, Pascual I, Villafruela II, Ortuno I. Nonenzymatic glycosylation ofhemoglobin and total plasmatic proteins in end-stage renal disease. AmINephro11991; 11:37-43.
  25. Wettre S, Lundberg M. Kinetics of glycosylated haemoglobin in uraemia determined on ion-exchange and affinity chromatography: no increase in the rate of glycosylation. Diabetes Res 1986; 3:107-10.
  26. Stewart ill. Haemolytic anaemia in acute and chronic renal failure. Q I Med 1967; 36:85-105.
  27. Freedman D, DandonaP, Fernando 0, MoorheadJF. Glycosylated haemoglobin in chronic renal failure and after renal transplantation. I Clin Patho11982; 35:737-9.
  28. Tzamaloukas AH, Murata GH, MalhotraD, Fox L, Goldman RS, A vasthi PS. The minimal dose of dialysis required for a target KT N in continuous peritoneal dialysis. Clin Nephrol1995; 44:316-21.
  29. Lindholm B, Bergstrom I, Karlander SG. Glucose metabolism in patients on continuous ambulatory peritoneal dialysis (CAPD).ASAIOTrans 1981;27:58-60.
  30. SalorantaC,GroopL, YlinenK, TeramoK, Tolppanen EM, TallgrenLG. Theusefulnessofmicro-and macrochromatographic determination of glycohemoglobin in diabetic patients with nephropathy. ClinNephroI1986;25:186-92.
  31. Bilo HJG, Struijk DG, Boeschoten EW, Oe PL, Donker AIM. A comparison of three methods to assess metabolic control in diabetic patients on CAPD treatment. Neth I Med 1988; 33:217-24.
  32. Coronel F, Macia M, Cidonha A, et al. Fructosamine levels in CAPD: its value as a glycemic index. In: Khanna R, Nolph KD, Prowant BF, Twardowski ZJ, Oreopoulos DG, eds. Advances in peritoneal dialysis. Toronto: Peritoneal Dialysis Bulletin Inc., 1991 ; 7:253-6.
  33. Nunoi K, Kodama T, Sato Y, et al. Comparison of reliability of plasma fructosamine and glycosylated hemoglobin assays for assessing glycemic control in diabetic patients on hemodialysis. Metabolism 1991 ; 40:986-9.
  34. Thitiarchakul S, Prowant BF, Moore HL, KhannaR, Nolph KD. Glycosylated hemoglobin (HbAic) in diabetic patients on continuous ambulatory peritoneal dialysis (CAPD) (Abstract). I Am Soc Nephrol1995; 6:514.
  35. Tzamaloukas AH, Murata GH, Zager PG, Eisenberg B, A vasthi PS. The relationship between glycemic control and morbidity and mortality in diabetics on dialysis. ASAIO I 1993; 39:880-5.
  36. Iovanovic L, Peterson CM. The clinical utility of glycosylated hemoglobin. Am I Med 1981; 70:331-8.
  37. TzamaloukasAH, AvasthiPS. Effectofhyperglycemiaon serum sodium concen1rntion and tonicity in outpatients on chronic dialysis. Am I Kidney Dis 1986; 7:477-82.
  38. Ifudu 0, Dulin AL, Friedman EA. Interdialytic weight gain COn'elates with glycosylated hemoglobin in diabetic hernodialysispatients. AmI Kidney Dis 1994; 23:686-91.
  39. Tzamaloukas AH, Saddler MC, Murata GH, et al. Symptomatic fluid retention in patients on continuous peritoneal dialysis. I Am Soc Nephro11995; 6: 198-206.
  40. Tzamaloukas AH, Yuan ZY, Murata GH, Balaskas EV, A vasthi PS, Oreopoulos 00. Clinical associations of glycemic control in diabetics on CAPD. In: KhannaR, Nolph KD, Prowant BF, Twardowski ZI, Oreopoulos DG, eds. Advances in peritoneal dialysis. Toronto: Peritoneal Dialysis Publications Inc., 1993; 9:291-4.
Corresponding author:
Antonios H. Tzamaloukas, MD, Renal Section ( 111 C), V A Medical Center, 2100 Ridgecrest Drive, SE, Albuquerque, New Mexico 87108 U.S.A.