Rapid changes in chromatographically determined haemoglobin A1c induced by short-term changes in glucose concentration

Summary Chromatographically determined haemoglobin A_1c concentration was measured during short-term (1–24 h) changes in glucose concentration in vitro and in vivo. In vitro at 37 °C the HbA_1c concentration increased with glucose concentration and time both in normal and diabetic erythrocytes. In normal erythrocytes incubated in 20–100 mmol/l glucose, the increases in the HbA_1c concentration were maximal after 4–6 h and then stable for the next 18–20 h. During the first hour, increases in the HbA_1c concentration were linear with time and on average 0.034% HbA_1c × h^–1 × mmol/l glucose^–1. In erythrocytes, after a rapidly produced increase (2h), HbA_1c decreased to preincubation concentrations during a further incubation of the erythrocytes in a glucose-free medium at 37 °C for 4–6 h. The mean rate of linear decrease was 0.017% × h^–1 × mmol/l glucose^–1. After incubation of erythrocytes in 100 mmol/l glucose for 24 h, 1.3% HbA_1c remained stable for 6 h in saline. The rapid increase in HbA_1c concentration, as determined by chromatography, was not due to stable HbA_1c (ketoamine linked glucose) as no increase was found in the HbA_1c concentrations determined by the thiobarbiturate method. In juvenile diabetics controlled by an artificial beta-cell, rapid changes of blood glucose concentration (up to 20 mmol/l) resulted in increases in HbA_1c concentration of as much as 1.9% within 12 h (mean 1.1%). Rapid in vivo increases in HbA_1c concentration were reversible by normalization of the blood glucose concentration. That rapid changes in HbA_1c may occur in daily diabetic life was evidenced by differences in HbA_1c concentration between blood samples from out-patient diabetics incubated in saline for 16 hours at 4 °C and 37 °C (range of differences 0.2–1.4% HbA_1c). The differences correlated to the blood glucose concentration at the time of sampling blood for HbA_1c determination. Thus, incubation of blood at a low glucose concentration prior to determination of the glycosylated haemoglobin concentration may overcome interference from rapidly produced HbA_1c.     

The colorimetric determination of HbAic in normal and diabetic subjects

The parameters which influence the determination of the major glycosylated haemoglobin fraction (HbA_1c) with the thiobarbituric acid (TBA) method are described. Conditions for an optimal determination method are given. The correlation with the fast haemoglobin determination by column chromatography is > 0.9. The influence of storage on blood, washed erythrocytes and haemolysate in respect to HbA_1c values obtained with the method described, was also investigated. Normal and pathological values from 78 patients correlate well with the fasting blood sugar levels (r>0.9).     

Synthesis of glycosylated haemoglobin in vivo

Summary The synthesis of glycosylated haemoglobins in vivo was measured during 24 h of controlled hyperglycaemia in seven insulin dependent diabetics. The mean blood glucose concentration was 22 mmol/l, while electrolytes and other metabolites were kept normal by infusion of 4–23 IU of insulin during hyperglycaemia. The study confirmed the velocity and magnitude of unstable HbA_1c formation previously found in vitro. The stable HbA_1c formed in 24 h was on average 0.006% of total haemoglobin/ mmol glucose. This compares well with the rate of HbA_1c synthesis reported in normal subjects using ⁵⁹Fe-kinetic measurements, and is in accordance with the concept of slow changes in stable HbA_1c with time and glucose concentration. To investigate the possibility that the rate of HbA_1c synthesis varies with erythrocyte age, glycosylated haemoglobins were measured in erythrocyte fractions after density separation on Percoll-Albumin gradients. We found both in normal subjects and in insulin treated diabetics that the 5% least dense cells contained 70%–80% of whole blood HbA_1c. Assuming the least dense cells to be the youngest erythrocytes, this observation is inconsistent with a slow linear increase in HbA_1c. Similar results were obtained in six newly diagnosed insulin dependent diabetic patients both before and after the first 30 days of insulin treatment, even though a marked decrease in young cell HbA_1c would be expected with the improved glucose control observed. We therefore conclude that density separation of erythrocytes is an inadequate technique to study age related HbA_1c synthesis.     

CLD (chronic liver diseases)-HbA1C as a suitable indicator for estimation of mean plasma glucose in patients with chronic liver diseases

In patients with chronic liver diseases (CLD), turnover of erythrocytes is increased whereas that of serum albumin is decreased. Thus, glycated hemoglobin (HbA_1C) and glycated albumin (GA) cannot be used as adequate indicators for chronic plasma glucose control in diabetic patients with CLD. In this investigation, we have proposed CLD-HbA_1C, a novel long-term glycemic control marker by using measured HbA_1C and GA. We studied 82 patients with CLD in whom glycemic control was regarded as to be stable. Daily plasma glucose profiles were monitored and estimated levels of HbA_1C were calculated on the conversion formula established by Rohlfing et al. [C.L. Rohlfing, J.D. England, H.M. Wiedmeyer, A. Tennill, R.R. Little, D.E. Goldstein, Defining the relationship between plasma glucose and HbA1c, Diabetes Care 25 (2002) 275–278]. Cholinesterase (ChE) as an indicator for hepatic function was determined at the same time when HbA_1C and GA levels were measured. CLD-HbA_1C was defined as the average of measured HbA_1C and GA/3, based upon the results that among healthy individuals, GA levels were roughly estimated at approximately threefold higher than HbA_1C levels. While measured HbA_1C levels in patients with CLD were generally lower than estimated HbA_1C levels, GA/3 values were generally higher than estimated HbA_1C levels. Such discrepancies lineally increased in accordance with a decrease in ChE levels. On the other hand, CLD-HbA_1C levels were highly correlated with estimated HbA_1C levels (R = 0.883), while no significant correlation between CLD-HbA_1C and ChE was noted.In conclusion, CLD-HbA_1C has been found a superior chronic glycemic control marker than HbA_1C or GA in diabetic patients with chronic liver diseases.     

The measurement of hemoglobin A1. Evaluation of an ‘aldimine eliminator’

Summary The efficiency of a newly introduced ‘hemoglobin A₁ aldimine eliminator’ in minimizing the effect of recent fluxes of glucose on the determination of total glycosylated hemoglobin (HbA₁) was evaluated by comparing HbA₁ values measured during morning fast and again 6 h postprandially in 26 insulin-dependent diabetic subjects by Isolab’s Fast Hemoglobin Test System employing dialyzed and aldimine eliminator-added non-dialyzed hemolysate. The HbA₁ values determined by this microcolumn procedure were also compared with those of the conventional macrocolumn method of Trivelli. HbA₁ measured by the microcolumn procedure and aldimine eliminator using non-dialyzed hemolysates did not differ from the HbA₁ values based on dialyzed hemolysates, and a good correlation was found between the macrocolumn method and the Isolab’s Fast Hemoglobin Test System employing both aldimine eliminator added non-dialyzed hemolysate (r=0.88, p<0.001) and dialyzed hemolysate (r=0.97, p<0.001). When 6 h changes were assessed, the mean blood glucose had increased from 11.7 to 15.5 mmol/l (p<0.001), and no significant increase in HbA₁ occurred when HbA₁ was assayed in aldimine eliminator-added non-dialyzed hemolysates (mean fast HbA₁: 11.7% and mean postprandial HbA₁: 11.8%). Therefore, the use of the HbA₁ aldimine eliminator appears to be valuable, practical and time-saving. This study has been supported by a grant from the City of Aalborg Medical Fund.     

Changes in glycosylated hemoglobin after initiation of hydroxychloroquine or methotrexate treatment in diabetes patients with rheumatic diseases

Objective

Prior research demonstrates that hydroxychloroquine (HCQ) lowers glycosylated hemoglobin (HbA_1c) in diabetes patients without rheumatic disease. We examined medical records of patients with diabetes mellitus (DM) and concomitant rheumatic illness to measure changes in HbA_1c after starting HCQ or methotrexate (MTX).

Methods

We used electronic medical records to identify patients beginning treatment with either HCQ or MTX who had a diagnosis of DM (or a pretreatment HbA_1c value of ≥7%) and at least 1 HbA_1c measurement both before and within 12 months after initiation of treatment. Using a structured medical record abstraction, we examined rheumatic disease diagnosis, cumulative steroid use, duration (months) between drug initiation and lowest HbA_1c value, a change in DM medication, body mass index (BMI), age, and sex. Adjusted linear regression models determined changes in HbA_1c from pretreatment values to the lowest posttreatment values within 12 months.

Results

We identified 45 patients taking HCQ and 37 patients taking MTX who met the inclusion criteria. Rheumatoid arthritis had been diagnosed in approximately half of the patients in each group. Age, sex, and mean pretreatment HbA_1c levels were similar across groups. The mean BMI of those taking HCQ (35.4 kg/m²) was slightly higher than that of those taking MTX (32.2 kg/m²) (P = 0.13). Glucocorticoid use appeared more common in those taking MTX (46%) than in those taking HCQ (29%) (P = 0.17). The mean reduction in HbA_1c from pretreatment values to the lowest posttreatment values was 0.66% (95% confidence interval [95% CI] 0.26, 1.05) in those taking HCQ compared with 0.11% (95% CI −0.18, 0.40) in those taking MTX. In fully adjusted analyses, the reduction in HbA_1c among those taking HCQ was 0.54% greater than the reduction among those taking MTX (P = 0.041).

Conclusion

HCQ initiation was associated with a significantly greater reduction in HbA_1c as compared with MTX initiation among diabetes patients with rheumatic disease.

     

Plasma lipoproteins and apolipoproteins in insulin-dependent and young non-insulin-dependent Arab women

Summary Plasma lipids, lipoproteins and apolipoproteins (apo) were analysed in 30 young Arab IDDM and 50 young insulin-requiring NIDDM women. The mean age of IDDM and NIDDM groups was 20.2 and 34.5 years, and mean duration of diabetes was 5.7 and 4.6 years, respectively. Two groups of 40 and 60 healthy women (matched for age and BMI) provided corresponding control groups. In comparison with control subjects, diabetics showed marked increases in the following parameters: total cholesterol (TC), low density lipoprotein (LDL) cholesterol, total triglycerides (TG), very low density lipoprotein (VLDL) triglycerides, phospholipids, apoB, LDL apoB, glucose and glycosylated hemoglobin (HbA_1c) as well as the ratios of total cholesterol/high density lipoprotein (HDL) cholesterol, LDL-cholesterol/HDL-cholesterol, LDL cholesterol/high density lipoprotein 2 (HDL₂) cholesterol and apoB/apoAI. Plasma LCAT activity, concentrations of HDL₃ apoAI and apoAII in plasma and lipoprotein fractions were normal in both the diabetic groups. Levels of C-peptide, HDL, HDL₂ and HDL₃ cholesterol, plasma apoAI, HDL apoAI and HDL₂ apoAI were markedly decreased in the diabetic groups as compared to their corresponding controls. There was no significant correlation between fasting glucose or HbA_1c and any of the above parameters. Despite insulin therapy in both the diabetic groups studied, abnormalities in lipids, apoB and apoAI still persisted. Our data suggest a possible higher risk of atherosclerosis in these patients.     

The relationship between blood glycosylated haemoglobin and home capillary blood glucose levels in diabetics

Summary Serial capillary blood glucose levels from insulin treated patients were recorded over 24 hour periods at fortnightly intervals for three months. Total glycosylated haemoglobin as % of HbA was measured at the end of this period by the Flückiger method, and % HbA₁ by column chromatography. There were highly significant correlations between mean blood glucose levels over the three months and % HbA₁ (r=0.93, 95% confidence limits 0.84–0.98), and with total glycosylated haemoglobin (r=0.88, 95% confidence limits 0.75–0.94). There was also a good correlation between results obtained by the two methods (r=0.81, p<0.0001). There were less strong correlations between % HbA₁ and blood glucose levels during each of the three months before the estimation, with percentage of glucose levels greater than 10 mmol/l and with mean fasting blood glucose. These data support the hypothesis that % HbA₁ and total glycosylated haemoglobin are satisfactory measurements of short term diabetic control.     

Glycosylated hemoglobin in patients with newly discovered juvenile-onset diabetes mellitus in the days immediately following the beginning of insulin treatment

Summary We studied the behavior of fast hemoglobin fractions in newly discovered diabetic patients, before and in the 10 days immediately following the beginning of insulin therapy, in order to verify whether or not the rapid improvement of glycemic control involved a rapid reduction of total HbA₁ and of its fractions. We observed a rapid and highly significant fall of HbA_1(a+b+c) and HbA_1c levels after only 1 or 2 days of insulin therapy, followed by a slower decrement in the other 3–10 days. HbA_1(a+b) showed a slower decrement trend, reaching levels significantly below baseline values only after 7–10 days. These results suggest that rapid changes occurring in glycosylated hemoglobin levels after the beginning of insulin treatment in newly discovered diabetic patients involve mainly HbA_1c. The kinetics of glycosylated hemoglobin reduction, with a first rapid decrement followed by a slower one, may suggest the hypothesis that rapid changes are due to reversible Schiff base de-glycosylation, the ketoamine linkage being the true index of long term glycemic control.     

Altered adrenal and thyroid function in children with insulin-dependent diabetes mellitus

In 129 children, aged 12.6±3.8 years, affected by type 1 diabetes mellitus, the levels of dehydroepiandrosterone sulfate (DHEAS), cortisol, T₃, fT₃, T4, fT4, rT₃, TSH, cholesterol, and triglycerides were evaluated and compared with those of a control group of 458 healthy age-matched children. The results were also correlated with hemoglobin HbA_1C. The DHEAS-standard deviation score (DHEAS-SDS; –0.36±0.77) was significantly different from zero in diabetic children, while the cortisol serum level was higher than in control subjects (485±94 vs 359±132 nmol/l). Moreover, the DHEAS-SDS and DHEAS-SDS/cortisol ratio correlated negatively with HbA_1c. Diabetic patients also showed lower T₃ values (2.22 ± 0.4 vs 2.32±0.3 nmol/l) and a higher rT₃/T₃ ratio (0.17±0.09 vs 0.15±0.05) than controls. There was a negative correlation between T₃ and HbA_1C. Cholesterol (4.77±1.08 vs 4.51±0.76 mmol/l) and triglycerides (0.82 ±0.53 vs 0.63±0.37 g/L) levels were higher in diabetic children and positively correlated with HbA_1c, but not with DHEAS-SDS. We can therefore conclude that diabetes, particularly if poorly controlled, tends to induce a dissociation of cortisol and DHEAS secretion and a low T₃ syndrome, similar to that seen in other illnesses.     

Comparison of effects of quinapril and metoprolol on glycaemic control, serum lipids, blood pressure, albuminuria and quality of life in non-insulin-dependent diabetes mellitus patients with hypertension. Swedish Quinapril Group

Objective

To compare the long-term effects of the angiotensin-converting enzyme (ACE)-inhibitor quinapril and the cardioselective beta-adrenergic blocking agent metoprolol on glycaemic control, with glycosylated haemoglobin (HbA_1c) as the principal variable, in non-insulin-dependent diabetes mellitus (NIDDM) patients with hypertension.

Design

A randomized, double-blind, double-dummy, multicentre study during 6 months preceded by a 4 week wash-out and a 3 week run-in placebo period. Quinapril (20 mg) and metoprolol (100 mg, conventional tablets) were given once daily. No change was made in the treatment of diabetes (diet and hypoglycaemic agents).

Subjects

Seventy-two patients fulfilling the criteria were randomized and entered the double-blind period. Twelve patients did not complete the study. Sixty patients, 26 on quinapril and 34 on metoprolol, were available for the final analysis.

Main outcome measures

The effect was assessed by changes in HbA_1c, the fasting serum glucose and the post-load serum glucose, C-peptide and insulin levels during the oral glucose tolerance test.

Results

In the quinapril group, the fasting serum glucose, oral glucose tolerance and the C-peptide and insulin responses, determined as the incremental area under the curves (AUC), showed no change, but the mean HbA_1c level increased from 6.2 ± 1.1% to 6.5 ± 1.3% (P < 0.05). In the metoprolol group, the rise in the mean level of HbA_1c, from 6.3 ± 1.0% to 6.8 ± 1.3% (P < 0.01), tended to be more marked than after quinapril, although there was no significant difference between the increments. The mean fasting serum glucose showed an increase from 9.1 ± 1.9 mm to 10.1 ± 2.8 mm (P < 0.01) which correlated significantly with the duration of diabetes (P < 0.01) and the increase in fasting serum triglycerides (P < 0.001). Moreover, in the metoprolol group we found significant decreases in the oral glucose tolerance as well as C-peptide and insulin responses to the glucose load.

Conclusions

Treatment with quinapril for 6 months appears to have advantages over metoprolol in NIDDM patients with hypertension. Although treatment with quinapril or metoprolol over 6 months was concomitant with a rise in the HbA_1c, increased fasting blood glucose, decreased oral glucose tolerance and decreased C-peptide and insulin responses to a glucose challenge were observed only in patients treated with metoprolol.

     

Glycosylated hemoglobin as predictor of adverse fetal outcome in type 1 diabetic pregnancies

The study aimed to determine whether a consistent dose-response association can be demonstrated, after adjustment for maternal age and White classification, between glycosylated hemoglobin (HbA_1c) values before conception and in the first trimester of pregnancy of insulin-dependent diabetic mothers and adverse fetal outcome (abortions and major malformations). This is a historical follow-up study based on medical records in a geographically defined catchment area. The study comprised 60 pregnancies with HbA_1c determinations before pregnancy and 161 with HbA_1c in the first trimester in women with type 1 diabetes admitted between 1980 and 1992. Relative risk calculations indicated a highly significant and consistent correlation between HbA_1c values above 6.6% and adverse fetal outcome after adjustment for differences in maternal age and White classification. Our data support a clinically significant and consistent relationship between adverse fetal outcome and HbA_1c in the first trimester of pregnancy of type 1 mothers, without any indication of a cut-off level below which further improvement in HbA_1c was of minor importance. Received: 25 October 1996 / Accepted in revised form: 18 April 1997     

Carotid intima-media thickness and stiffness in relation to type 2 diabetes in Chinese

We investigated carotid intima-media thickness (IMT) and quantitative carotid stiffness (QCS) index in relation to plasma glycosylated hemoglobin A_1C (HbA_1C) and duration of diabetes mellitus in 337 Chinese diabetic patients. In categorical analyses, carotid IMT was 710 μm in subjects with a duration of diabetes mellitus ≤2 years, 760 μm in subjects with a duration of diabetes mellitus more than two years and with plasma HbA_1C < 6.5% (P < 0.05), and 790 μm in subjects with a duration of diabetes mellitus more than two years but with plasma HbA_1C ≥ 6.5% (P < 0.01). The corresponding values for QCS values were 4.5, 4.6 and 5.1 (P < 0.05), respectively. In multiple stepwise regression analyses carotid IMT was significantly associated with the duration of diabetes mellitus, systolic blood pressure and serum concentration of total cholesterol, whereas QCS was significantly associated with age, HbA_1C, systolic and diastolic blood pressure (P < 0.05). In conclusion, carotid IMT as a structural measure of arterial wall is increased in patients with a longer history of diabetes mellitus, whereas QCS as functional index is mainly influenced by the quality of blood glucose control.     

Impact of postprandial and fasting glucose concentrations on HbA1c in patients with type 2 diabetes

AimThis study aimed to assess the relative contributions of postprandial and fasting glucose concentrations to overall hyperglycaemia.MethodsPatients with type 2 diabetes (n = 973) carried out self-monitored blood glucose (SMBG) profiles on entry into the European Exenatide (EUREXA) trial. Glucose area under the curve was calculated for postprandial excursions (AUC_ppg) and total daytime concentrations > 6.1 mmol/L (AUC_total), as well as for the percentage of glycaemia due to postprandial excursions (%_ppg). In addition, OGTT scores were assessed for each patient. Results were evaluated according to defined HbA_1c categories.ResultsThere was a significant linear relationship between HbA_1c and the derived variables of AUC_ppg, AUC_total and %_ppg (P < 0.001 for each), with explained variance greatest for AUC_total (r² = 37.4%). AUC_ppg increased only slightly up to an HbA_1c of 7.0%, but showed a steeper increase in higher HbA_1c categories. Also, the increase in AUC_total with increasing HbA_1c was much more pronounced. As a result, the postprandial glucose excursion as a proportion of total glucose (%_ppg) decreased across HbA_1c categories from 61.0% at HbA_1c < 6.5% to 22.0% at HbA_1c ≥ 9.0%. HOMA-IR remained virtually unchanged through all HbA_1c categories, while HOMA-B showed no large changes up to HbA_1c 7.0%, but then decreased at higher HbA_1c values. The ΔI30/ΔG30 ratio decreased in the HbA_1c 7.0–7.9% category, but did not change greatly at higher HbA_1c categories.ConclusionWith increasing HbA_1c, there was a decrease in the contribution of postprandial hyperglycaemia to total glycaemia, and fasting hyperglycaemia became more important. This is consistent with impaired insulin release, particularly first-phase release, at higher HbA_1c levels.     

Haemoglobin glycation (Hb1Ac) increases during red blood cell storage: a MALDI‐TOF mass‐spectrometry‐based investigation

Haemoglobin A_1c (HbA_1c) represents a key biomarker in diabetes diagnosis and management, as it is indicative of recent blood glucose concentrations. Glycation of haemoglobin is a non‐enzymatic irreversible process that is promoted by the prolonged exposure of erythrocytes to high glucose concentrations, a condition that is known to occur under blood banking conditions. However, controversial data indicate no clear hint as to whether and to which extent HbA_1c accumulates during red blood cell storage. Hereby, we propose the application of a validated MALDI‐TOF mass‐spectrometry‐based method to this issue and report the observation about HbA_1c levels apparently increasing over storage progression.     

The Effect of a Hypertension Self-Management Intervention on Diabetes and Cholesterol Control

Background

Most patient chronic disease self-management interventions target single-disease outcomes. We evaluated the effect of a tailored hypertension self-management intervention on the unintended targets of glycosylated hemoglobin (HbA_1c) and low-density lipoprotein cholesterol (LDL-C).

Methods

We evaluated patients from the Veterans Study to Improve the Control of Hypertension, a 2-year randomized controlled trial. Patients received either a hypertension self-management intervention delivered by a nurse over the telephone or usual care. Although the study focused on hypertension self-management, we compared changes in HbA_1c among a subgroup of 216 patients with diabetes and LDL-C among 528 patients with measurements during the study period. Changes in these laboratory values over time were compared between the 2 treatment groups using linear mixed-effects models.

Results

For the patients with diabetes, the hypertension self-management intervention resulted in a 0.46% reduction in HbA_1c over 2 years compared with usual care (95% confidence interval, 0.04%-0.89%; P = .03). For LDL-C, there was a minimal 0.9 mg/dL between-group difference that was not statistically significant (95% confidence interval, −7.3-5.6 mg/dL; P = .79).

Conclusions

There was a significant effect of the self-management intervention on the unintended target of HbA_1c,but not LDL-C. Chronic disease self-management interventions might have “spill-over” effects on patients' comorbid chronic conditions.     

Risks of Diabetic Nephropathy with Variation in Hemoglobin A1c and Fasting Plasma Glucose

Background

This study examined whether annual variation in glycosylated hemoglobin A_1c (HbA_1c) and fasting plasma glucose (FPG), as represented by the coefficient of variation (CV), can predict diabetic nephropathy independently of mean FPG, mean HbA_1c, and other risk factors in patients with type 2 diabetes.

Methods

A computerized database of patients with type 2 diabetes aged ≥30 years and free of diabetic nephropathy (n = 3220) who were enrolled in the Diabetes Care Management Program of China Medical University Hospital before 2007 was used in a time-dependent Cox proportional hazards regression model.

Results

The incidence rates of diabetic nephropathy were 16.11, 22.95, and 28.86 per 1000 person-years in the first, second, and third tertiles of baseline HbA_1c-CV, respectively; the corresponding incidence rates for FPG-CV were 9.46, 21.23, and 37.51 per 1000 person-years, respectively. After multivariate adjustment, the corresponding hazard ratios for the second and third tertiles versus the first tertile of annual HbA_1c-CV were 1.18 (95% confidence interval [CI], 0.88-1.58) and 1.58 (95% CI, 1.19-2.11), respectively, and 1.55 (95% CI, 0.99-2.41) and 4.75 (95% CI, 3.22-7.01) for FPG-CV, respectively. The risks of diabetic nephropathy for HbA_1c-CV and FPG-CV stratified according to age, gender, renal function, and hypertension status were provided.

Conclusions

Annual FPG and HbA_1c variations have a strong association with diabetic nephropathy in patients with type 2 diabetes. Whether intervention for reducing glucose variation should be administered needs to be examined in a future study.     

A rapid micro-scale method for the measurement of haemoglobin A1(a+b+c)

Summary A rapid method is described for the measurement of total glycosylated haemoglobins (HbA_1(a+b+c). The procedure utilizes 0.05 ml of blood and takes forty minutes to complete manually. Eighty blood samples can be analysed without automation by one person in a day. Each analysis uses less than 2 mg of potassium cyanide, resulting in a method that is both safe and rapid for routine hospital laboratories. The inter-assay coefficient of variation was 4% and that for intra-assay measurements 3%, over the range 5–20% HbA_1(a+b+c). The method confirmed that the level of HbA_1(a+b+c) is elevated in imperfectly controlled diabetics. Amongst patients with blood glucose levels of less than 10 mmol/l the mean level of HbA_1(a+b+c) was found to be 8.5%; samples from 14 known diabetics gave a mean value of 10.9%, whereas 17 known non-diabetic samples gave a mean value of 8.3%. In the group of samples from 27 diabetic individuals with blood glucose levels above 10 mmol/l the mean level of HbA_1(a+b+c) was found to be 13.5%.     

Changes in glycaemic control and risk of coronary artery disease in type 1 diabetes mellitus: findings from the Pittsburgh Epidemiology of Diabetes Complications Study (EDC)

Aims/hypothesis To complete a comparative analysis of studies that have examined the relationship between glycaemia and cardiovascular disease (CVD)/coronary artery disease (CAD) and perform a prospective analysis of the effect of change in glycosylated Hb level on CAD risk in the Pittsburgh Epidemiology of Diabetes Complications Study (EDC) of childhood-onset type 1 diabetes mellitus (n = 469) over 16 years of two yearly follow-up. Methods Measured values for HbA₁ and HbA_1c from the EDC were converted to the DCCT-standard HbA_1c for change analyses and the change in HbA_1c was calculated (final HbA_1c minus baseline HbA_1c). CAD was defined as EDC-diagnosed angina, myocardial infarction, ischaemia, revascularisation or fatal CAD after medical record review. Results The comparative analysis suggested that glycaemia may have a stronger effect on CAD in patients without, than in those with, albuminuria. In EDC, the change in HbA_1c differed significantly between CAD cases (+0.62 ± 1.8%) and non-cases (−0.09 ± 1.9%) and was an independent predictor of CAD. Conclusions/interpretation Discrepant study results regarding the relationship of glycaemia with CVD/CAD may, in part, be related to the prevalence of renal disease. Measures of HbA_1c change over time show a stronger association with CAD than baseline values.