Progression of nephropathy in type 2 diabetes: the glycation gap is a significant predictor after adjustment for glycohemoglobin (Hb [A.sub.1c]).
As an indication that nonglycemic determinants of Hb [A.sub.1c] maybe important for the clinical assessment of diabetic patients, Hempe et al. (18) reported that the hemoglobin glycation index (HGI) (the difference between observed Hb [A.sub.1c] and the value calculated from its regression on MBG) differed among type 1 diabetic patients, was stable over time for a given patient, and was not related to erythrocyte turnover. In fact, when type 1 diabetic patients in the DCCT were ranked by mean HGI during the study, the risks of retinopathy and nephropathy among patients in the top third were at all MBG levels several times those of patients in the bottom third (19). However, because HGI and Hb [A.sub.1c] (as the residual and dependent variable of a regression) are mutually correlated, it has been questioned (20) whether there is any value in attempting to partition observed Hb [A.sub.1c] into a "standard" value for a given MBG level plus a correction determined by greater or lesser propensity for glycation (HGI). For example, reanalysis of the DCCT data shows that HGI is not a significant risk factor for microvascular complications if Hb [A.sub.1c] itself is included in the analysis (20).
Fructosamine is a measure of average glycemia over a shorter period than Hb [A.sub.1c], but its concentration in plasma is much more stable than that of glucose itself and is much easier to measure than true MBG, which requires continuous blood glucose monitoring or sufficiently frequent 7-point profiles. In a study of 40 type 1 diabetic patients, a 1% increase in a "glycation gap" computed in the same way as the HGI, but using the regression of Hb [A.sub.1c] on fructosamine, was found to increase the risk of more advanced nephropathy 2.9 fold, although there was no significant correlation between more advanced nephropathy and either Hb [A.sub.1c] or fructosamine (21). In the study described here, data on 2314 type 2 diabetic patients followed up for an average of 6.5 years were used to investigate whether the fructosamine-based glycation gap is stable within individuals and is a significant predictor of progression of nephropathy when adjusted for Hb [A.sub.1c].
Materials and Methods
Ours is a tertiary center serving a broad cross-section of an almost exclusively white population of about 450 000 individuals including a majority of local diabetic patients requiring insulin or oral antidiabetic medications. In the present study, we included relevant data for all patients who between March 1992 and March 2007 were prescribed insulin or oral antidiabetic drugs for type 2 diabetes and who satisfied the following additional inclusion criteria: (a) the patient had been followed up for at least 1 year; (b) throughout follow-up, nephropathy had been evaluated at least once each year on the basis of albumin excretion rate (AER); (c) each such AER measurement was accompanied by determination of serum fructosamine and Hb A1c; and (d) the patient had no known hemoglobinopathy or erythrocyte disorder.
Like the DCCT (2), we distinguished between a primary cohort of patients with a 1- to 5-year history of diabetes, no ophthalmoscopic retinopathy, and AER <40 mg/24 h at "entry" (i.e., at the date of the earliest determinations included in the study) (1225 patients) and a secondary cohort of patients with a 1- to 15-year history of diabetes, slight or moderate ophthalmoscopic retinopathy, and AER <200 mg/24 h at entry (1089 patients). Other variables recorded at entry included age, sex, duration of diabetes, and type of therapy. Serum fructosamine and Hb [A.sub.1c] levels were included in the study not only at each AER evaluation, but for each occasion on which both had been determined (hereinafter "visit"), except that for patients showing progression of nephropathy (see below), no data were included after the visit on which progression was observed.
As in the DCCT, patients were deemed to have undergone progression of nephropathy if AER was [greater than or equal to] 100 mg/24 h and had been <40 mg/24 h at entry, or if AER was [greater than or equal to] 300 mg/24 h and had been <200 mg/24 h at entry (19). Following previous authors (19, 20), but with fructosamine as a predictor (21), the (gg) glycation gap was calculated as follows. First, Hb [A.sub.1c] was regressed on fructosamine by using a repeated-measures longitudinal regression model (22), and data for all patients and all visits included in the study, and a "raw" glycation gap (gg), were calculated for each visit as observed Hb [A.sub.1c] minus the Hb [A.sub.1c] value predicted from the regression equation obtained. After verification of the stability of gg (see the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol57/issue2), a characteristic mean glycation gap (GG) was then calculated for each patient by averaging his or her raw gg values (19,20). Finally, the patients were ranked by GG and classified by tertiles as high-, medium-, or low-GG subjects.
Hb [A.sub.1c] was determined by HPLC by using Menarini HA-8121 and HA-8140 analyzers (Menarini Diagnostics); the interassay CV was 1.6% at an Hb [A.sub.1c] level of 5.9% and 0.9% at an Hb [A.sub.1c] level of 11%. For this study, all Hb [A.sub.1c] values were converted from Japanese Diabetes Society/Japanese Society for Clinical Chemistry-referenced values to DCCT-aligned units by using the equation Hb [A.sub.1c NGSP] = 0.985 Hb [A.sub.1c JDS/JSCC] + 0.46% (where NGSP stands for National Glycohemoglobin Standardization Program) (23). Until January 2001, fructosamine was determined by the nitroblue tetrazolium (NBT) method (Roche Diagnostics), with an interassay CV of 2.9% at 296 [micro]mol/L fructosamine and 1.7% at 521 [micro]mol/L fructosamine and, after that date, by the GlyPro enzymatic method (Genzyme) with an interassay CV of 1.8% at 175 [micro]mol/L fructosamine and 0.91% at 640 jumol/L fructosamine (in both cases, a Cobas Mira analyzer was used). For use in Cox analyses, etc., NBT-derived values were transformed by using the equation [Fructosamine.sub.GlyPro] = 1.33 [Fructosamine.sub.NBT] - 127.5, which was obtained by a regression analysis of the results for 188 randomly selected samples for which both methods were used (r = 0.991). Albumin in 24-h urine samples was measured by using a Dade Behring BN nephelometer from Siemens Healthcare Diagnostics with an interassay CV of 2.2% at 79 mg/L albumin and 0.9% at 230 mg/L albumin. statistical analysis
Raw Hb [A.sub.1c] was regressed on raw fructosamine by using a longitudinal linear model following both (a) appropriate data weighting to take between-patient variation in number of visits into account and (b) confirmation of linearity by means of a spline-fitting algorithm that made no prior assumptions regarding slope. A random intercept model was used on the basis ofthe Akaike information criterion, which showed it to provide a better fit than either the random slope model or the random slope, random intercept model.
In evaluating the relative values of GG and Hb [A.sub.1c] for assessment of risk of progression of nephropathy by means of Cox regression analyses, Hb [A.sub.1c] and fructosamine were included as updated mean (u.m.) values, where the u.m. value of a variable (x) at each visit of a patient was obtained by estimation of the mean of all values of each variable in all visits up to and including that visit [the measured value of each variable (x)is termed "raw x" to prevent confusion with u.m. x].
Multivariate Cox regressions adjusted for tied event times were performed with progression of nephropathy as the dependent variable and age at baseline, sex, duration of diabetes at baseline, therapy, and cohort as "basic covariates," and u.m. fructosamine, u.m. Hb[A.sub.1c], and GG as individually optional additional covariates; u.m. Hb [A.sub.1c] and u.m. fructosamine were treated as time dependent. In additional Cox analyses in which the study group was trichotomized by GG tertiles, baseline Hb [A.sub.1c] was also an optional covariate. Continuous covariates were entered as linear terms (nonlinearity was ruled out by examining alternative models by using penalized smoothing splines), and the assumption of proportional hazards was validated before these analyses by examination of Schoenfeld residuals. All statistical calculations were performed by using either SPSS v.15 or Stata v.10.
Table 1 summarizes baseline characteristics of the 2314 type 2 diabetic patients studied. Mean follow-up time was 6.5 years (SD 4.2 years). All patients made at least 1 visit per year during follow-up, and 43.1% of the low-GG group, 47.9% of the medium-GG group, and 47.1% of the high-GG group made an average of 2 visits per year. In all, data collected in a total of 21 960 visits were included in the study (mean 9.5 visits per patient). AER was measured on 17 540 of these occasions (mean 8 measurements per patient).
The stability of the gg was shown by close correlation (r > 0.84) between gg values determined for the same patient on different occasions. Accordingly, the characteristic GG of each patient was calculated by averaging his or her gg values. The distribution function of GG was normal (Gaussian) (P > 0.05), with mean 0.01% and SD 1.01%; for 336 patients (14.5%) GG was > 1%, and for 363 (15.7%), < -1%.
Fig. 1 shows the regression of raw Hb [A.sub.1c] on raw fructosamine in the 3 GG groups (r = 0.800, 0.922, and 0.806 for the low-, medium-, and high-GG group, respectively) and in the whole study sample. As was to be expected on statistical grounds, in the whole sample, raw gg was not correlated with raw fructosamine but was strongly correlated with raw Hb [A.sub.1c] (r = 0.778; Fig. 2). Similarly, the overall mean Hb [A.sub.1c] level in the GG groups increased in the order low-GG < medium-GG < high-GG (P < 0.001), whereas mean fructosamine did not differ significantly among the GG groups (see the online Supplemental Table). Adjustment for the basic covariates reduced the correlation between raw Hb [A.sub.1c] and raw fructosamine in the whole sample to 0.621 and increased the correlation between rawHb [A.sub.1c] and gg to 0.819.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Of the 2314 patients studied, 487 (21%) underwent progression of nephropathy. In the primary cohort, progression affected 170 patients (13.9%), and in the secondary cohort, it affected 317 patients (29.1%). Cox analyses with just 1 covariate in addition to the basic set showed that the risk of progression of nephropathy was significantly increased by u.m. fructosamine, u.m. Hb [A.sub.1c], and GG, with hazard ratios of 1.28, 1.24, and 1.45, respectively, when the fructosamine-Hb [A.sub.1c] and Hb [A.sub.1c]-GG regressions are used to refer all hazard ratios to a 1% increase in Hb [A.sub.1c] (models 1-3, Table 2). When u.m. fructosamine and GG, which derive from uncorrelated raw variables, were both included in the analysis (model 5), both remained significant and their hazard ratios were the same as in models 1 and 3, respectively. By contrast, when redundant covariates derived from closely correlated raw variables were included (models 4 and 6), only the covariate with the greater predictive power emerged as significant: u.m. Hb [A.sub.1c] in model 4 and GG in model 6.
Fig. 3 shows the cumulative incidence of progression of nephropathy among patients in the high-, medium-, and low-GG groups after adjustment for the basic covariates, u.m. Hb [A.sub.1c], u.m. fructosamine, and baseline Hb [A.sub.1c]. The influence of GG was highly significant (P < 0.0001). Specifically, risk in the mediumand high-GG groups was respectively 1.6 and 2.5 times the risk in the low-GG group (P = 0.001 and P < 0.0001, respectively). The inclusion or exclusion of baseline Hb [A.sub.1c] and/or u.m. Hb [A.sub.1c] made virtually no difference to these results, and neither did the use of untransformed NBT measurements of fructosamine in the calculation of the corresponding visit-specific glycation gaps.
Fig. 4 shows mean Hb [A.sub.1c] and the incidence of progression of nephropathy in subgroups defined by the tertiles of both GG and mean fructosamine. Note that as GG rises, whereas mean fructosamine falls (i.e., along the left-to-right diagonal of each figure), mean Hb [A.sub.1c] remains constant (as is only logical), but the incidence of progression of nephropathy rises.
First, the results of this large study confirm that gg values obtained for the same patient at different times are closely correlated (r > 0.84). Thus, gg is relatively stable, even in this retrospective observational study in which it was subject to marked iatrogenic fluctuation due to the difference between the times taken byserum fructosamine and Hb [A.sub.1c] to reflect changes in antiglycemic medication. Second, we found that GG is a significant predictor of the risk of progression of nephropathy in type 2 diabetic patients and that this property persists when allowance is made for Hb [A.sub.1c]. Progression of nephropathy was significantly more frequent in the high- and medium-GG groups than in the low-GG group even after adjustment for u.m. and baseline Hb [A.sub.1c] and other covariates.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
A limitation of this study that parallels the cause of criticism (20, 24) of previous similar studies (18, 19) is that gg, as the residual of the regression of Hb [A.sub.1c] on fructosamine, contains within-patient as well as between-patient contributions to its variability among all samples. Nevertheless, its observed degree of stability for a given individual in this 15-year study of 2314 type 2 diabetics followed up for an average of 6.5 years supports the thesis that some individuals have an intrinsically greater propensity than others toward glycation of hemoglobin and seems sufficient to allow this propensity to be evaluated by some gg-based measure. The best available measure may well be the value of gg in the untreated patient, averaged over enough measurements to smooth out variations in serum fructosamine. In the present study, following Cohen et al. (21), we made up for the absence of pretreatment measurements the best we could by using GG, the average of the patient's gg determinations throughout the study.
A potential confounder for this study was between-patient differences in the mean age of circulating erythrocytes; in hematologically normal individuals, a large proportion of variation in Hb [A.sub.1c] is due to this cause (8). However, no significant differences in erythrocyte creatinine level have been observed between glycation groups, and hence there is no indication that the increase in the proportion of individuals with high HGI (7) or high GG (this study) that accompanies increasing mean Hb [A.sub.1c] is due to reduced erythrocyte turnover. In the present study, the low- and high-GG groups included similar proportions of patients with pathological creatinine concentrations (3.5% and 3.9%, respectively), so the prevalence of kidney disease cannot have influenced our results.
The report by McCarter et al. that HGI is a strong predictor of retinopathy and nephropathy among the type 1 diabetic patients of the DCCT (19) has been criticized on the grounds that the correlation between HGI and Hb [A.sub.1c], which is a mathematical necessity if the statistical model of the regression of Hb [A.sub.1c] on MGB is valid, might mean that HGI is merely a surrogate for Hb [A.sub.1c]. Reanalysis of the DCCT data shows no significant influence of HGI on the risk of nephropathy if Hb [A.sub.1c] is taken into account (20). McCarter et al. acknowledge this, but point out that what is of interest for researchers at this point is not whether Hb [A.sub.1c] is associated with risk of diabetic complications, which is well established, but whether this association is due wholly to blood glucose concentrations or is also contributed to by putative nonglucose determinants of Hb [A.sub.1c] (25). This is an open question, in part because the DCCT measurements of MBG (7-point profiles recorded on a single day every 3 months), though used successfully to predict the development of retinopathy (26), may have been insufficiently accurate measures of average blood glucose concentration for the purposes of McCarter et al. In other words, the DCCT data do not unequivocally clarify whether HGI reflects the effects of nonglucose determinants of Hb [A.sub.1c], or merely inaccuracy in average blood glucose measurement (24). There is little possibility that in the present study the glycation gap merely reflected inaccuracy in the measurement of fructosamine.
Unlike Cohen et al. (21) in their study of the glycation gap in type 1 diabetic patients, we found that not only higher GG, but also higher Hb [A.sub.1c], was associated with more advanced nephropathy; Cohen et al.'s negative result is probably attributable to their small sample size (40 patients). In both studies, however, ordinal logistic regression or Cox analyses showed GG to be closely associated with progression of nephropathy and that significant association persisted when Hb [A.sub.1c], fructosamine, and other factors were taken into account.
In the present study, u.m. fructosamine and GG were both significant when both were included in the Cox analysis (as is logical in view of the statistical independence of raw fructosamine and gg), but u.m. Hb [A.sub.1c] was not a significant predictor of progression of nephropathy when GG was included, nor was u.m. fructosamine if u.m. Hb [A.sub.1c] was included. This result suggests that the predictive value of Hb [A.sub.1c] maybe due less to its association with fructosamine (a measure of extracellular glycation, and hence of what maybe called glycation pressure) than to its relationship with the glycation gap, which may reflect nonglycemic determinants of hemoglobin glycation, or "glycability."
A substantial proportion of variation in Hb [A.sub.1c] between individuals is ofgenetic origin, yet is independent of genes influencing fasting glucose (9). Illustrative of this, glycemic control fails to account for consistent differences of up to about 0.4% between the Hb [A.sub.1c] levels of different racial groups (10). These genetic results reinforce the implications of the temporal stability of gg and HGI in the sense that Hb [A.sub.1c] levels must be determined by multiple factors, not all of which are related to glycemia. Candidate factors include erythrocyte 2,3-diphosphoglycerate levels, intra-erythrocyte pH, and total plasma amino acid levels (8); lipoperoxides and antagonistic antioxidants (12); glycolytic and deglycating enzymes (6, 7, 21); processes related to glucose transporter 1-mediated transport (11); and erythrocyte transmembrane glucose gradient (27). When factors such as these have been more definitely related to the glycability of hemoglobin and/or of other proteins glycated in diabetes, they will doubtlessly be used in evaluating risk of complications. Meanwhile, it maybe asked whether the glycation gap, as a gross measure of glycability, can usefully be employed for this purpose. Our results suggest that it can.
For prediction of diabetic complications in a temporal sense, it is of course necessary to be able to determine an individual's characteristic glycation gap in a much shorter time than the average 6.5-year follow-up of this study. As noted above, the best measure of the gap may be the value of gg in the untreated patient, averaged over enough measurements to smooth out variations in fructosamine. If pretreatment determinations are not available, a good alternative may be to average 3 or 4 determinations at 1- to 2-week intervals during a period of good, stable glycemic control. We have in fact observed that in the cases of a few patients whose first Hb [A.sub.1c] determinations may have taken place before treatment, the corresponding gg values were indeed very close to those observed during periods of good control.
Finally, it may not be too redundant to point out explicitly the possible implications of the present findings for clinical practice. If neither reliable MBG values nor fructosamine values are obtained, so that the only guiding biochemical parameter is Hb [A.sub.1c], and the glycation gap is not taken into account in interpreting Hb [A.sub.1c], then the MBG concentrations of a low-GG patient are probably higher than they appear to be. Although a low-GG patient intrinsically has below-average risk of complications, there might therefore be room to reduce risk further by tighter control of glycemia. On the contrary, the MBG concentrations of a high-GG patient are probably lower than they appear to be, and therapy that does not take this into account may introduce an unnecessarily large risk of hypoglycemia.
In conclusion, the results of the present study of a large sample of type 2 diabetic patients followed up for an average of 6.5 years show that the glycation gap is a stable characteristic of these patients that, when evaluated as GG, predicts the progression of nephropathy, even after adjustment for Hb [A.sub.1c] and fructosamine and independently of the latter. The joint use of the glycation gap and fructosamine as markers of hemoglobin glycability and glycation pressure, respectively, may improve the evaluation of diabetic patients, allowing glycemia control targets to be set in accordance with the extent of the individual patient's predisposition to protein glycation. Research into the most convenient means of evaluating the glycation gap of new and treated patients is in progress. Future studies should address the possible relationships of the glycation gap with complications of diabetes other than nephropathy and the glycation of proteins other than hemoglobin.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the Disclosures of Potential Conflict of Interest form. Potential conflicts of interest:
Employment or Leadership: None declared.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: S. Rodriguez-Segade and F. Camina, the Secretariat General for Research and Development of the Xunta de Galicia, Spain (Refs. PGIDIT02BTF20303PR, PGIDIT04BTF203016PR, and PGIDIT06BTF20302PR), the Spanish Ministry of Education and Science (Ref. SAF2004-07602). S. Rodriguez-Segade, Siemens Healthcare Diagnostics and Menarini Diagnostics. Expert Testimony: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
Acknowledgments: We also acknowledge the statistical guidance of Dr. Juan Manuel Paz and Dr. Francisco Gude of the Clinical Epidemiology Unit of our hospital.
(1.) Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-86.
(2.) Diabetes Control and Complications Trial Research Group. The association between glycemic exposure and long-term diabetic complications. Diabetes 1995;44:968-83.
(3.) Consensus Statement on the Worldwide Standardization of the Hemoglobin A1C Measurement: The American Diabetes Association, European Association for the Study of Diabetes, International Federation of Clinical Chemistry and Laboratory Medicine, and the International Diabetes Federation. Diabetes Care 2007;30:2399-400.
(4.) Nathan DM, Kuenen J, Borg R, Zheng H, Schoenfeld D, Heine RJ. Translating the A1C assay into estimated average glucose values. Diabetes Care 2008;31:1473-8.
(5.) Rohlfing CL, Wiedmeyer HM, Little RR, England JD, Tennill A, Goldstein DE. Defining the relationship between plasma glucose and HbA1c: analysis of glucose profiles and HbA1c in the Diabetes Control and Complications Trial. Diabetes Care 2002;25:275-8.
(6.) Madsen H, Kjaergaard JJ, Ditzgl J. Relationship between glycosylation of haemoglobin and the duration of diabetes: a study during the third trimester of pregnancy. Diabetologia 1982;22: 37-40.
(7.) Hudson PR, Child DF, Jones H, Williams CP. Differences in rates of glycation (glycation index) may significantly affect individual HbA1c results in type 1 diabetes. Ann Clin Biochem 1999;36: 454-9.
(8.) Gould BJ, Davie SJ, Yudkin JS. Investigation of the mechanism underlying the variability of glycated haemoglobin in nondiabetic subjects not related to glycaemia. Clin Chim Acta 1997;260:49-64.
(9.) Snieder H, Sawtell PA, Ross L, Walker J, Spector TD, Leslie RD. A1C levels are genetically determined even in type 1 diabetes: evidence from healthy and diabetic twins. Diabetes 2001;50: 2858-63.
(10.) Herman WH, Ma Y, Uwaifo G, Haffner S, Kahn SE, Horton ES, et al. Differences in A1C by race and ethnicity among patients with impaired glucose tolerance in the Diabetes Prevention Program. Diabetes Care 2007;30:2453-7.
(11.) Cohen RM, Snieder H, Lindsell CJ, Beyan H, Hawa MI, Blinko S, et al. Evidence for independent heritability of the glycation gap (glycosylation gap) fraction of A1C in nondiabetic twins. Diabetes Care 2006;29:1739-43.
(12.) Selvaraj N, Bobby Z, Sathiyapriya V Effect of lipid peroxides and antioxidants on glycation of haemoglobin: an in vitro study on human erythrocytes. Clin Chim Acta 2006;366:190-5.
(13.) Delpierre G, Collard F, Fortpied J, Van Schaftingen E. Fructosamine 3-kinase is involved in an intracellular deglycation pathway in human erythrocytes. Biochem J 2002;365:8018.
(14.) Genuth S, Wanjie S, Cleary P, Sell DR, Dahms W, Malone J, et al. Glycation and carboxymethyllysine levels in skin collagen predict the risk of future 10-year progression of diabetic retinopathy and nephropathy in the Diabetes Control and Complications Trial and Epidemiology of Diabetes Interventions and Complications participants with type 1 diabetes. Diabetes 2005;54:3103-11.
(15.) Yudkin JS, Forrest RD, Jackson Ca, Ryle AJ, Davies SJ, Gould BJ. Unexplained variability of glycated haemoglobin in non-diabetic subjects not related to glycaemia. Diabetologia 1990;33:208-15.
(16.) Kilpatrick ES, Maylor PW, Keevil BG. Biological variation of glycated hemoglobin: implications for diabetes screening and monitoring. Diabetes Care 1998;21:261-4.
(17.) Rohlfing C, Wiedmeyer HM, Little R, Grotz VL, Tennill A, England J, et al. Biological variation of glycohemoglobin. Clin Chem 2002;48:1116-8.
(18.) Hempe J, Gomez R, McCarter R, Chalew S. High and low haemoglobin glycation phenotypes in type 1 diabetes: a challenge for interpretation of glycaemic control. J Diabetes Complications 2002;16:313-20.
(19.) McCarter RJ, Hempe JM, Gomez R, Chalew SA. Biological variation in HbA1c predicts risk of retinopathy and nephropathy in type 1 diabetes. Diabetes Care 2004;27:1259-64.
(20.) Lachin JM, Genuth S, Nathan DM, Rutledge BN. The hemoglobin glycation index is not an independent predictor of the risk of microvascular complications in the Diabetes Control and Complications Trial. Diabetes 2007;56:1913-21.
(21.) Cohen RM, Holmes YR, Chenier TC, Joiner CH. Discordance between HbA1c and fructosamine. Diabetes Care 2003;26:163-7.
(22.) Diggle PJ, Liang KY, Zeger SL. Analysis of longitudinal data. New York: Oxford University Press; 1994.
(23.) Hoelzel W, Weykamp C, Jeppsson JO, Miedema K, Barr JR, Goodall I, et al. IFCC reference system for measurement of hemoglobin A1c in human blood and the National Standardization Schemes in the United States, Japan, and Sweden: a method-comparison study. Clin Chem 2004;50: 166-74.
(24.) Nathan DM, Genuth S, Rutledge B, Lachin J. Response to Comment on: Lachin et al. (2007). The hemoglobin glycation index is not an independent predictor of the risk of microvascular complications in the Diabetes Control and Complications Trial. Diabetes 2007;57:e5, DOI: 10.2337/db07-1638.
(25.) Chalew S, Hempe J, McCarter R. Comment on: Lachin et al. (2007). The hemoglobin glycation index is not an independent predictor of the risk of microvascular complications in the Diabetes Control and Complications Trial. Diabetes 2007; 57:e4, DOI: 10.2337/db07-1569.
(26.) Service FJ, O'Brien PC. The relation of glycaemia to the risk of development and progression of retinopathy in the Diabetes Control and Complications Trial. Diabetologia 2001;44:1215-20.
(27.) Khera PK, Joiner CH, Carruthers A, Lindsell CJ, Smith EP, Franco RS, et al. Evidence for interindividual heterogeneity in the glucose gradient across the human red blood cell membrane and its relationship to hemoglobin glycation. Diabetes 2008;57:2445-52.
Santiago Rodrfguez-Segade, [1,2] * Javier Rodriguez, [1,2] Jose M. Cabezas-Agricola,  Felipe F. Casanueva, [3,4] and Felix Camina 
 Department of Biochemistry and Molecular Biology, and the  University Hospital Clinical Biochemistry Laboratory and  Division of Endocrinology, University of Santiago de Compostela, Santiago de Compostela, Spain; and  Physiopathology of Obesity and Nutrition Biomedical Research Network Consortium (CIBERobn), Madrid, Spain.
 Nonstandard abbreviations: Hb [A.sub.1c], glycated hemoglobin; MBG, mean blood glucose; DCCT, Diabetes Control and Complications Trial; HGI, hemoglobin glycation index; AER, albumin excretion rate; gg, raw glycation gap; GG, mean glycation gap; NBT, nitroblue tetrazolium; u.m., updated mean.
* Address correspondence to this author at: Laboratorio de Bioquimica Clinica, Complejo Hospitalario Universitario de Santiago, Travesia de la Choupana s/n, 15706 Santiago de Compostela, Spain. Fax +34-981594912; e-mail email@example.com.
Received February 6, 2010; accepted October 18, 2010.
Previously published online at DOI: 10.1373/clinchem.2010.144949
Table 1. Baseline characteristics of the study group (n = 2314). Sex (a) Men 1118(48.3) Women 1196 (51.7) Age (years) (b) 59.9 (10.7) Cohort (a) Primary 1225 (52.9) Secondary 1089 (47.1) White race (%) 100 Known duration of diabetes (years) (b) 6.0 (5.6) Treatment (a) Insulin 687 (29.7) Oral antidiabetic drugs 1627 (70.3) Hb [A.sub.1c] (%) (b) 7.5 (2.0) Fructosamine ([micro]mol/L) (b) 322 (81) (a) n (%). (b) Mean (SD) Table 2. Results of various multivariate Cox models of progression of nephropathy. Model Variables (a) Hazard ratio P number (95% CI) (b) 1 u.m. fructosamine 1.283 (1.087-1.394) <0.0001 2 u.m. Hb [A.sub.1c] 1.238 (1.138-1.337) <0.0001 3 GG 1.447 (1.300-1.615) <0.0001 4 u.m. fructosamine 1.087 (0.920-1.181) 0.290 u.m. Hb [A.sub.1c] 1.201 (1.089-1.320) <0.0001 5 u.m. fructosamine 1.283 (1.087-1.394) <0.0001 GG 1.447 (1.299-1.614) <0.0001 6 u.m. Hb [A.sub.1c] 1.070 (0.951-1.201) 0.260 GG 1.358 (1.168-1.577) <0.0001 (a) Other than age at baseline, sex, duration of diabetes at baseline, therapy (insulin or oral antidiabetic drugs), and cohort (primary or secondary), which were adjusted for in all models. (b) Hazard ratios for GG and fructosamine are given for a 1% rise in Hb [A.sub.1c].
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Endocrinology and Metabolism|
|Author:||Rodriguez-Segade, Santiago; Rodriguez, Javier; Cabezas-Agricola, Jose M.; Casanueva, Felipe F.; Cami|
|Date:||Feb 1, 2011|
|Previous Article:||Action to control cardiovascular risk in diabetes (ACCORD) trial--clinical implications.|
|Next Article:||Urinary catalytic iron in obesity.|