Serum fructosamine versus glycosylated hemoglobin as an index of glycemic control, hospitalization, and infection in diabetic hemodialysis patients

Transcription

1 & 2010 International Society of Nephrology Serum fructosamine versus glycosylated hemoglobin as an index of glycemic control, hospitalization, and infection in diabetic hemodialysis patients Neal Mittman 1, Brinda Desiraju 1, Irfan Fazil 1, Hiteshkumar Kapupara 1, Jyoti Chattopadhyay 1, Chinu M. Jani 2 and Morrell M. Avram 1 1 Avram Division of Nephrology, Long Island College Hospital, Brooklyn, New York, USA and 2 Spectra Laboratories, Rockleigh, New Jersey, USA Diabetes is the most common cause of end-stage renal disease and an important risk factor for morbidity and mortality in dialysis patients. Glycemic control, utilizing serial measurement of glycosylated hemoglobin (HbA1c), is generally recommended to limit end-organ damage, including cardiovascular morbidity and mortality. We, along with others, have previously suggested that HbA1c may not be a reliable measure of glycemic control in dialysis patients, and have therefore explored the use of serum fructosamine (SF) as an alternative marker. The objective of this study was to compare HbA1c levels with SF in monitoring glycemic control and associated morbidity (infection and hospitalization) in diabetic patients in a large urban hemodialysis (HD) center. We enrolled 100 diabetic HD patients and followed them up prospectively for 3 years. Data on demographics, as well as biochemical and clinical data, including hospitalizations and infections, were recorded. The mean age was 63 years. In all 54% were women and the majority were African Americans (72%). As expected, HbA1c and albumin-corrected fructosamine (AlbF) levels were highly correlated and both were significantly associated with serum glucose. AlbF, however, was more highly correlated with mean glucose values when less than 150 mg/dl and was a more useful predictor of morbidity. By univariate logistic regression and by Poisson regression analysis, AlbF, but not HbA1c, was a significant predictor of hospitalization. Additionally, in patients dialyzed by arteriovenous (AV) access (that is, excluding those dialyzed via vascular catheters), AlbF, but not HbA1c, was a significant predictor of infection. In conclusion, AlbF is as reliable a marker as HbA1c for glycemic control in diabetic patients on HD, and may be advantageous for patients with serum glucose in a desirable therapeutic range (o150 mg/dl). In addition, AlbF, but not HbA1c, is associated with morbidity (hospitalizations and infections) in diabetic patients on HD. ; doi: /ki KEYWORDS: diabetes; fructosamine; HbA1c; hemodialysis; hospitalization; infection Correspondence: Neal Mittman, Avram Division of Nephrology, Long Island College Hospital, 339 Hicks Street, Brooklyn, New York 11201, USA. nmittman@chpnet.org TO CITE THIS ARTICLE: Mittman N, Desiraju B, Fazil I, Kapupara H, Chattopadhyay J, Jani CM, Avram MM. Serum fructosamine versus glycosylated hemoglobin as an index of glycemic control, hospitalization, and infection in diabetic hemodialysis patients. Kidney Int 2010; 78 (Suppl 117): S41 S45. The prevalence of diabetes is increasing worldwide. According to a 2009 Centers for Disease Control report, more than 23.6 million Americans have diabetes, of whom 5.7 million do not know that they have the disease. Diabetes is the leading cause of end-stage renal disease in the United States, accounting for approximately 44% of new patients requiring dialysis. 1 The number of patients with diabetes as the primary cause of end-stage renal disease increased by 6.6% between 2001 and Diabetes is a known mortality risk in hemodialysis (HD) patients. 3 Glycemic control can limit end organ damage, cardiovascular morbidity, and mortality in these patients. 4 Traditionally, serial measurements of blood glucose and glycosylated hemoglobin (HbA1c), which reflect the previous 2 3 months of glycemic control, have been used to monitor glycemic control in diabetic HD patients, as for diabetics in the general population. Although graded glycohemoglobin levels within the normal range have been associated with coronary artery disease in nondiabetics without renal disease, 5 and blood glucose level is a risk factor for cardiovascular disease in healthy individuals without diabetes, 6 little information is available in the dialysis population. Kalantar-Zadeh et al., 7 using a large dialysis provider s database, noted paradoxically better survival with a poorer glycemic index (that is, lower mortality hazard ratios with higher A1c values). However, HbA1c may not be an ideal measure of glycemic control in diabetic end-stage renal disease patients, as it may be lowered by shortened red cell survival or by increased red cell turnover due to repetitive proerythropoietic stimulation (for example, recombinant human erythropoietin), or affected by assay interference from uremia. Several alternative indices of glycemic control have been reported in literature, including serum fructosamine (SF), which has a shorter half-life than HbA1c SF, produced by the spontaneous, nonenzymatic S41

2 N Mittman et al.: Fructosamine and HbA1c in HD glycation of proteins, reflects very recent (1 3 weeks) glycemic control, potentially lessening the confounding effect of shortened red cell survival or of high red cell turnover seen with A1c (HbA1c) levels. Few reports in literature have assessed the validity of alternative indices of glycemic control in dialysis patients. Those published have involved small numbers of patients, and results have been contradictory. 8,11,12 Elevated SF levels have been associated with increased cardiovascular mortality in nonuremic nondiabetics, 13 but little information is available with regard to the association of SF with morbidity in HD patients. RESULTS The demographics of the patient population are shown in Table 1. The mean age was 63 years. Women comprised 54%, and the majority were African Americans (72%). Mean dialysis vintage was 36 months. Mean values for SF, albumincorrected fructosamine (AlbF), and HbA1c (A1c) were 373 mmol/l (range: mmol/l), 973 mmol/g (range: mmol/g), and 7.2% (range: 5 14%), respectively. As expected, HbA1c and fructosamine levels were highly correlated. Enrollment and follow-up A1c values strongly and positively correlated with the corresponding SF values (r ¼ 0.66, Po at enrollment, and r ¼ 0.75, Po at follow-up) and AlbF values (r ¼ 0.67, Po at enrollment, and r ¼ 0.74, Po at follow-up). Both measures of glycemic index, A1c and SF, were highly Table 1 Characteristics of hemodialysis patients Age (years, mean±s.d.) 63±11 (range 35 85) Gender, female (%) 54 Race African Americans (%) 72 White (%) 5 Hispanic (%) 21 Other (%) 2 Months on dialysis at enrollment (mean±s.d.) 36±44 (range ) correlated with glucose measurements, with similar correlation coefficients and levels of significance (r ¼ 0.70, Po versus r ¼ 0.63, P ¼ o for enrollment AlbF and A1c, respectively; and r ¼ 0.75, Po versus r ¼ 0.71, Po for follow-up AlbF and A1c, respectively). AlbF was much more highly correlated with mean glucose values when glucose values were less than 150 mg/dl (r ¼ 0.57, P ¼ versus r ¼ 0.31, P ¼ 0.19 for enrollment AlbF and A1c, respectively; and r ¼ 0.55, P ¼ versus r ¼ 0.45, P ¼ 0.01 for follow-up AlbF and A1c, respectively) (Figures 1 and 2). Table 2 compares the levels of AlbF and HbA1c by hospitalizations and by infection. Patients who were hospitalized during the study period had higher AlbF levels compared with those who were not hospitalized (974 versus 863 mmol/g, P ¼ 0.012), whereas there was no significant difference in A1c between the groups (P ¼ 0.96). Similarly, patients dialyzed using AV access (that is, excluding those using catheters) and having an infection during the study period had significantly higher AlbF (998 versus 903 mmol/g, P ¼ 0.03). By univariate logistic regression analysis, AlbF was a significant positive predictor of hospitalization (odds ratio ¼ 1.005, P ¼ 0.016) and of infection (odds ratio ¼ 1.003, P ¼ 0.039) in diabetic HD patients, whereas A1c did not predict hospitalization (odds ratio ¼ 1.01, P ¼ 0.96) or infection (odds ratio ¼ 1.34, P ¼ 0.17). By Poisson regression analysis, AlbF was a significant predictor of the number of hospitalizations per patient year (relative risk ¼ 1.12, P ¼ 0.007), the duration of hospitalization per patient year (relative risk ¼ 1.19, Po0.0001), and the number of episodes of infection per patient year (relative risk ¼ 1.29, P ¼ 0.001), whereas A1c did not predict frequency or duration of hospitalization or of infection in patients receiving dialysis by AV access (Table 3). DISCUSSION This study has shown that two indices of glycemic control, HbA1c and SF, correlate with each other and with mean blood glucose levels in our diabetic HD population. Whereas Enrollment albumincorrected fructosamine (µmol/g) r=0.57, P =0.001 Enrollment HbA1c (%) 10.0 r =0.31, P = Glucose < 150 mg/dl Glucose <150 mg/dl Figure 1 Correlation of fructosamine and glycosylated hemoglobin with serum glucose. Relationship between glucose and (a) enrollment albumin-corrected fructosamine (correlation coefficient r ¼ 0.57, P ¼ 0.001) and (b) enrollment glycosylated hemoglobin (HbA1c, correlation coefficient r ¼ 0.31, P ¼ 0.19) in diabetic hemodialysis patients with glucose level less than 150 mg/dl S42

3 N Mittman et al.: Fructosamine and HbA1c in HD a Follow-up albumincorrected fructosamine (µmol/g) r =0.55, P = r =0.45, P = Glucose < 150 mg/dl b Follow-up HbA1c (%) Glucose <150 mg/dl Figure 2 Correlation of fructosamine and glycosylated hemoglobin with serum glucose. Relationship between glucose and (a) follow-up albumin-corrected fructosamine (correlation coefficient r ¼ 0.55, P ¼ 0.001) and (b) follow-up glycosylated hemoglobin (HbA1c, correlation coefficient r ¼ 0.45, P ¼ 0.01) in diabetic hemodialysis patients with glucose level less than 150 mg/dl. Table 2 Fructosamine and HbA1c by hospitalization and infection in diabetic hemodialysis patients Hospitalization Infection Yes No P-value Yes No P-value HbA1c (%) 7.02± ± ± ± Albumin-corrected fructosamine (mmol/g) 974± ± ± ± Abbreviation: HbA1c, glycosylated hemoglobin. Table 3 Poisson regression analysis: predictors of hospitalization and infection in hemodialysis patients HbA1c (%) Albumin-corrected fructosamine (lmol/g) a Outcome RR (95% CI) P-value RR (95% CI) P-value Frequency of hospitalization 1.02 (0.36, 1.86) (1.03, 1.22) Duration of hospitalization 0.97 (0.93, 1.00) (1.15, 1.22) o0.001 Infections 1.22 (0.86, 1.72) (1.11, 1.51) Abbreviations: CI, confidence interval; HbA1c, glycosylated hemoglobin; RR, relative ratio. a Values of fructosamine were divided by 100 for scaling purposes. AlbF and A1c equally reflected mean blood glucose values or glycemic status in the entire population, we found that when serum glucose was stratified for acceptable glycemic control (mean serum glucose below 150 mg/dl), AlbF was more highly correlated with mean glucose than was A1c. This supports the hypothesis that A1c may not be the best measure of glycemic control in diabetic HD patients, as it may be affected by shortened red cell survival, by increased red cell turnover due to repetitive proerythropoietic stimulation (that is, recombinant human erythropoietin), or by assay interference from uremia. 11,14 Our findings are consistent with those of Shoji et al., 15 who reported that SF level corrected for protein concentration is an excellent glycemic index in diabetic patients with CKD, and is not affected by urea, which is known to influence the level of HbA1c. Lamb et al. 16 reported that SF is elevated in HD patients, but suggested that, when corrected for total protein concentration, it is probably a more reliable marker of glycemic index than glycated hemoglobin. Coronel et al. 8 studied glycemic control in diabetic continuous ambulatory peritoneal dialysis patients, reporting that there were good correlations between fructosamine and mean blood glucose over the preceding 3 weeks, and concluded that both assays can be used as glycemic indices. Some authors have suggested other glycemic indices as superior alternatives to A1c. Recently, Inaba et al. 9 reported that glycated albumin is a better glycemic indicator than are glycated hemoglobin values in HD patients with diabetes, whereas Peacock et al. 10 suggested that in diabetic HD patients, HbA1c levels significantly underestimate glycemic control, whereas those of glycated albumin more accurately reflect this control. On the other hand, Ichikawa et al. 11 evaluated multiple indices of glycemic control in diabetic HD patients, including HbA1c, fructosamine, glycated albumin, and 1,5-anhydroglucitol, concluding that HbA1c was the most reliable index. Similarly, Nunoi et al. 12 and Joy et al. 17 reported that, in HD patients, fructosamine correlated poorly with blood glucose, and that HbA1c was a superior marker of glycemic control. S43

4 N Mittman et al.: Fructosamine and HbA1c in HD We have examined the association of morbidity, that is, hospitalization and infection, with two glycemic markers, HbA1c and AlbF, in our diabetic HD patients. It should be noted that we used mean serum glucose values over a longer period than did other studies cited above (3 4 months as opposed to 2 4 weeks), a time period that should have more fairly leveled the comparison. As central venous catheter hemoaccess was so strongly associated with the incidence of infection (odds ratio ¼ 11, P ¼ 0.026, by logistic regression analysis) compared with AV access, we analyzed data in patients dialyzed with AV access only. In our diabetic HD patients, hospitalization and risk of infection were associated with increased levels of AlbF, but not with the current standard of diabetic care, namely, hemoglobin A1c. A reasonable inference can be made that higher levels of AlbF reflect poorer glycemic control, potentially resulting in more diabetic complications and, therefore, in more hospitalizations and infections. Increased nonenzymatic glycosylation of various proteins such as glycosylated hemoglobin, glycated albumin, and fructosamine has been reported to be associated with diabetic complications. In addition, increased levels of glycated proteins have been reported to diminish immune function, 18 increase oxidative stress, proinflammatory response, 19 and induce various other pathological events. In nonuremic diabetic patients, good preoperative glycemic control (HbA1c level o7.0%) is associated with a decrease in infectious complications across a variety of surgical procedures. 20 SF is increased in anemic Helicobacter pylori-infected nondiabetic nonuremic patients compared with healthy controls. 21 Published information regarding the relationship of A1c and AlbF with morbidity, hospitalization, and infection in diabetic HD patients is scarce. In a retrospective study, extremely high and low HbA1c values were associated with hospitalization risk in diabetic HD patients. 22 Glycated albumin, but not glycated hemoglobin, was associated with peripheral vascular calcification in diabetic HD patients. 23 As SF levels are directly affected by serum albumin concentration, 24 fructosamine values were corrected for serum albumin. Therefore, we examined whether serum albumin concentration itself was associated with hospitalization and infection. By logistic regression analysis, serum albumin concentration was associated with infection (odds ratio ¼ 0.19, P ¼ 0.04) but not with hospitalization (odds ratio ¼ 0.22, P ¼ 0.06); therefore, serum albumin itself may have been a contributory factor in predicting infection in our HD patients. In summary, SF, corrected for the level of serum albumin, seems to be a reliable index of glycemic control in diabetic patients on HD, and may be superior to HbA1c in some patients, especially those with acceptable diabetic control. This study also suggests that enrollment AlbF, an alternative index of glycemia control, was more strongly associated with hospitalization and infection than was HbA1c in diabetic HD patients followed up prospectively in a single center. Large, prospective trials are needed to confirm these findings, to define optimal diabetic control in this population at great risk for cardiovascular morbidity and mortality, and to define the contribution of serum albumin to morbidity associated with SF. MATERIALS AND METHODS One hundred diabetic HD patients treated at the Avram Center for Kidney Diseases in Long Island College Hospital were enrolled in this study, conducted from February 2005 through to August Patients were followed up to April On enrollment, demographics, and clinical and biochemical data were recorded. Morbidity data including number of hospitalizations, duration of hospitalization, and episodes of infections were recorded throughout the study period. Results of monthly evaluation of biochemical parameters, including serum albumin, total protein, serum creatinine, and serum glucose, were recorded. Serial blood glucose values measured using test strips (Precision Extra, Medisense, Abbott Laboratories, Alameda, CA, USA) and predialysis serum glucose measurements for the preceeding 4-monthly chemistry panels were collected. HbA1c levels were measured by an immunoturbidimetric method. SF levels were measured by a colorimetric method (COBAS INTEGRA, Roche Diagnostics, Indianapolis, IN, USA). Both HbA1c and fructosamine were measured at enrollment and then 6 months after enrollment. Fructosamine values were corrected for variations in the concentration of serum albumin according to the following formula by Lamb et al. 16 AlbF ¼ (fructosamine (mmol/l) 100/albumin (g/l)). The study protocol was approved by the Institutional Review Board of Long Island College Hospital, and informed consent was obtained from all patients. Statistical analysis Continuous variables were expressed as mean±s.d. Correlations were reported as either the Pearson correlation coefficient or the Spearman rank correlation coefficient. Poisson regression analysis was used to examine the association of HbA1c and AlbF with hospitalization and infection. Two-sided P-values of less than 0.05 were considered as statistically significant. Calculations were performed using SPSS for Windows (SPSS, Chicago, IL, USA) and SAS, version 8.2 software (SAS Institute, Carey, NC, USA). DISCLOSURE CMJ is an employee of Spectra Laboratories. The remaining authors declared no competing interests. ACKNOWLEDGMENTS This work was supported, in part, by the Nephrology Foundation of Brooklyn. We acknowledge with gratitude the assistance and counsel provided by Dr J Michael Lazarus, Chief Medical Officer of Fresenius of North America. REFERENCES 1. United States Renal Data System. Excerpts from the USRDS 2008 annual data report: Atlas of end-stage renal disease in the United States. Am J Kidney Dis 2009; 53(Suppl 1): S129 S United States Renal Data System. USRDS 2007 Annual Data Report, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney disease, Bethesda, MD, Oomichi T, Emoto M, Tabata T et al. Impact of glycemic control on survival of diabetic patients on chronic regular hemodialysis: a 7-year observational study. Diabetes Care 2006; 29: S44

5 N Mittman et al.: Fructosamine and HbA1c in HD 4. McMurray SD, Johnson G, Davis S et al. Diabetes education and care management significantly improve patient outcomes in the dialysis unit. Am J Kidney Dis 2002; 40: Selvin E, Coresh J, Golden SH et al. Glycemic control and coronary heart disease risk in persons with and without diabetes. The atherosclerosis risk in communities study. Arch Intern Med 2005; 165: Levitan EB, Song Y, Ford ES et al. Is nondiabetic hyperglycemia a risk factor for cardiovascular disease? Arch Intern Med 2004; 164: Kalantar-Zadeh K, Kopple JD, Regidor DL et al. A1C and survival in maintenance hemodialysis patients. Diabetes Care 2007; 30: Coronel F, Marcia M, Ciodoncha A et al. Fructosamine levels in CAPD: its value as glycemic index. Adv Perit Dial 1991; 7: Inaba M, Okuno S, Kumeda Y et al. Glycated albumin is a better glycemic indicator than glycated hemoglobin values in hemodialysis patients with diabetes: effect of anemia and erythropoietin injection. J Am Soc Nephrol 2007; 18: Peacock TP, Shihabi ZK, Bleyer AJ et al. Comparison of glycated albumin and hemoglobin A(1c) levels in diabetic subjects on hemodialysis. Kidney Int 2008; 7: Ichikawa H, Nagake Y, Takahashi M et al. What is the best index of glycemic control in patients with diabetes on hemodialysis? Nippon Jinzo Gakkai shi 1996; 38: 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: Browner WS, Pressman AR, Lui L-Y et al. The association between serum fructosamine and mortality in elderly women. The study of osteoporotic fractures. Am J Epidemiol 1999; 149: Nakao T, Matsumoto H, Okada T et al. Influence of erythropoietin treatment on hemoglobin A1c levels in patients with chronic renal failure on hemodialysis. Intern Med 1998; 37: Shoji T, Tabata T, Nishizawa Y et al. Clinical availability of serum fructosamine measurement in diabetic patients with uremia. Use as a glycemic index in uremic diabetes. Nephron 1989; 51: Lamb E, Venton TR, Cattell WR et al. Serum glycated albumin and fructosamine in renal dialysis patients. Nephron 1993; 64: Joy MS, Cefalu WT, Hogan SL et al. Long-term glycemic control measurements in diabetic patients receiving hemodialysis. Am J Kidney Dis 2002; 39: Cohen G, Rudnicki M, Walter F et al. Glucose-modified proteins modulate essential functions and apoptosis of polymorphonuclear leukocytes. J Am Soc Nephrol 2001; 12: Aronson D. Hyperglycemia and the pathobiology of diabetic complications. Adv Cardiol 2008; 45: Dronge AS, Perkal MF, Kancir S et al. Long-term glycemic control and postoperative infectious complications. Arch Surg 2006; 141: Vijayan G, Sundaram RC, Bobby Z et al. Increased plasma malondialdehyde and fructosamine in anemic H. pylori infected patients: effect of treatment. World J Gastroenterol 2007; 13: Williams ME, Lacson Jr E, Teng M et al. Extremes of glycemic control (HbA1c) increase hospitalization risk in diabetic hemodialysis patients in the USA. Am J Nephrol 2009; 29: Yamada S, Inaba M, Shidara K et al. Association of glycated albumin, but not glycated hemoglobin, with peripheral vascular calcification in hemodialysis patients with type 2 diabetes. Life Sci 2008; 83: Van Dieijen-Visser MP, Seynaeve C, Brombacher PJ. Influence of variations in albumin or total-protein concentration on serum fructosamine concentration. Clin Chem 1986; 32: S45