Performance of SARS-CoV-2 serology tests: Are they good enough?

In the emergency of the SARS-CoV-2 pandemic, great efforts were made to quickly provide serology testing to the medical community however, these methods have been introduced into clinical practice without the complete validation usually required by the regulatory organizations. SARS-CoV-2 patient samples (n = 43) were analyzed alongside pre-pandemic control specimen (n = 50), confirmed respiratory infections (n = 50), inflammatory polyarthritis (n = 22) and positive for thyroid stimulating immunoglobulin (n = 30). Imprecision, diagnostic sensitivity and specificity and concordance were evaluated on IgG serologic assays from EuroImmun, Epitope Diagnostics (EDI), Abbott Diagnostics and DiaSorin and a rapid IgG/IgM test from Healgen. EDI and EuroImmun imprecision was 0.02–14.0% CV. Abbott and DiaSorin imprecision (CV) ranged from 5.2%–8.1% and 8.2%–9.6% respectively. Diagnostic sensitivity of the assays was 100% (CI: 80–100%) for Abbott, EDI and EuroImmun and 95% (CI: 73–100%) for DiaSorin at ≥14 days post PCR. Only the Abbott assay had a diagnostic specificity of 100% (CI: 91–100%). EuroImmun cross-reacted in 3 non-SARS-CoV-2 respiratory infections and 2 controls. The DiaSorin displayed more false negative results and cross-reacted in six cases across all conditions tested. EDI had one cross-reactive sample. The Healgen rapid test showed excellent sensitivity and specificity. Overall, concordance of the assays ranged from 76.1% to 97.9%. Serological tests for SARS-CoV-2 showed good analytical performance. The head-to-head analysis of samples revealed differences in results that may be linked to the use of nucleocapsid or spike proteins. The point of care device tested demonstrated adequate performance for antibody detection.

Determinants of diagnostic discordance for non-diabetic hyperglycaemia and Type 2 diabetes using paired glycated haemoglobin measurements in a large English primary care population: cross-sectional study

AIM

To investigate factors influencing diagnostic discordance for non-diabetic hyperglycaemia and Type 2 diabetes.

METHODS

Some 10 000 adults at increased risk of diabetes were screened with HbA_1c and fasting plasma glucose (FPG). The 2208 participants with initial HbA_1c ≥ 42 mmol/mol (≥ 6.0%) or FPG ≥ 6.1 mmol/l were retested after a median 40 days. We compared the first and second HbA_1c results, and consequent diagnoses of non-diabetic hyperglycaemia and Type 2 diabetes, and investigated predictors of discordant diagnoses.

RESULTS

Of 1463 participants with non-diabetic hyperglycaemia and 394 with Type 2 diabetes on first testing, 28.4% and 21.1% respectively had discordant diagnoses on repeated testing. Initial diagnosis of non-diabetic hyperglycaemia and/or impaired fasting glucose according to both HbA_1c and FPG criteria, or to FPG only, made reclassification as Type 2 diabetes more likely than initial classification according to HbA_1c alone. Initial diagnosis of Type 2 diabetes according to both HbA_1c and FPG criteria made reclassification much less likely than initial classification according to HbA_1c alone. Age, and anthropometric and biological measurements independently but inconsistently predicted discordant diagnoses and changes in HbA_1c .

CONCLUSIONS

Diagnosis of non-diabetic hyperglycaemia or Type 2 diabetes with a single measurement of HbA_1c in a screening programme for entry to diabetes prevention trials is unreliable. Diagnosis of non-diabetic hyperglycaemia and Type 2 diabetes should be confirmed by repeat testing. FPG results could help prioritise retesting. These findings do not apply to people classified as normal on a single test, who were not retested.

Use of HbA1c in the diagnosis of diabetes mellitus in the UK. The implementation of World Health Organization guidance 2011

The WHO was very clear that an HbA(1c) of 48 mmol/mol (6.5%) and above is diagnostic of diabetes. They were less clear regarding results that fell below 48 mmol/mol. The WHO recognize that individuals with HbA(1c) values below the cut-off point may still have diabetes, but give no guidance on how to investigate further. It is important that these individuals, who may be at increased risk of developing diabetes, are monitored correctly; a recommended scheme is given in Fig. 2; following these recommendations will ensure at-risk people are not overlooked and will be monitored closely. Even although it is not recommended to combine HbA(1c) with glucose measurement for diagnosis, the WHO did not discount the value of a fasting glucose and an oral glucose tolerance test to diagnose diabetes in selected individuals; it is the responsibility of the investigating doctor to choose how best to diagnose on an individual basis. This new method of diagnosing diabetes will identify a different cohort as having diabetes than is currently being diagnosed; but the process of investigation that does not require a fasting sample makes investigation easier, allowing more people to be investigated.

Use of haemoglobin A1c to detect impaired fasting glucose or Type 2 diabetes in a United Kingdom community based population

AIMS

To evaluate the diagnostic accuracy of haemoglobin A1c (HbA1c) in screening for impaired fasting glucose and Type 2 diabetes (T2DM).

METHODS

We screened 3904 adults aged 45-70 (mean age 58.6 [standard deviation (SD) 6.9] years, mean body mass index (BMI) 29.9 [SD 4.7]kg/m(2)), with fasting plasma glucose (FPG) and HbA1c as part of a large diabetes prevention programme. We assessed the diagnostic accuracy of HbA1c for predicting impaired fasting glucose (IFG), (defined either as FPG 5.6-6.9 mmol/l, or 6.1-6.9 mmol/l), and T2DM (FPG ≥ 7.0 mmol/l).

RESULTS

The prevalences of IFG were 13.8% (FPG 5.6-6.9 mmol/l) and 4.5% (FPG 6.1-6.9 mmol/l) and of T2DM was 2.1%. Using FPG 5.6-6.9 mmol/l as the IFG reference standard, HbA1c of 39-47 mmol/mol (5.7-6.4%) was 63% sensitive and 81% specific, and HbA1c 43-47 mmol/mol (6.1-6.4%) was 21% sensitive and 98% specific, in diagnosing IFG. HbA1c ≥ 48 mmol/mol (6.5%) was 61% sensitive and 99% specific in diagnosing T2DM. Having HbA1c 39-47 mmol/mol (5.7-6.4%), male sex, and body mass index >29.5 together increased the odds of IFG 6.5-fold (95% confidence interval (CI) 5.5-7.8) compared to the pre-test odds.

CONCLUSION

Defining 'pre-diabetes' at a lower HbA1c threshold of 39 mmol/mol (5.7%) instead of 47 mmol/mol (6.1%) increases its sensitivity in diagnosing IFG, but current American Diabetes Association definitions of 'pre-diabetes' based on HbA1c would fail to detect almost 40% of people currently classified as IFG. This has implications for current and future diabetes prevention programmes, for vascular risk management, and for clinical advice given to people with 'pre-diabetes' based on fasting glucose data.

Introduction of IFCC reference method for calibration of HbA: implications for clinical care

Abstract HbA(1c) is recommended for monitoring glycaemic control and quantifying the risk of complications in patients with diabetes. National guidelines for treatment of patients with diabetes in UK specify that HbA(1c) measurements should be Diabetes Control and Complications Trial (DCCT)-aligned i.e. comparable to the DCCT and UK Prospective Diabetes Study (UKPDS). The IFCC reference method for HbA(1c) will be introduced in Europe in December 2003 for calibration of all laboratory and POCT (point of care testing) methods for HbA(1c) following the recent EC "In Vitro Diagnostic" (IVD) directive. This reference method involves measurement of HbA(1c) and HbA(0) by electron-spray ionisation-mass spectrometry or capillary electrophoresis with the reference range approximately 2% HbA(1c) lower than the corresponding range from the DCCT. However, this EC IVD directive will not change reporting of DCCT-aligned HbA(1c) in the UK. Professionals involved in the care of patients with diabetes in the UK met with Dr Sue Roberts in London in July 2003. It was decided that in the UK DCCT-aligned HbA(1c) will continue to be reported from December 2003 for patient care and that laboratories currently reporting non-aligned DCCT HbA(1c) should change to reporting DCCT-aligned results as soon as possible. It was considered important for diabetes care in the UK that the reporting of HbA(1c) should not fragment. The UK HbA(1c) Standardization Committee was set up to hold "a watching brief " on HbA(1c) especially with relation to reporting of HbA(1c) in other countries.

A reliable non-separation fluorescence quenching assay for total glycated serum protein: a simple alternative to nitroblue tetrazolium reduction

A simple non-separation assay for the measurement of total glycated serum protein is described. It was found that the fluorescence intensity of a solution of a fluorescein-boronic acid derivative was quenched in proportion to the amount of serum added. This led to the development of an assay in which 10 microL of serum is added to 4 mL of a solution of the fluorescein-boronic acid derivative and the fluorescence intensity is measured after 15 min. The results, as measured by drop in fluorescence intensity, calibrated by a single standard, were compared with the results for nitroblue tetrazolium (NBT) reduction of fructosamine and showed good correlation (r=0.936, n=114). The intra-assay precision (seven samples each measured 10 times) was less than 2.1% (concentration range 190-660 micromol/L); inter-assay precision for seven samples in 10 assays was less than 2.5% (over the same concentration range). Dilution of serum that had a high concentration of total glycated protein showed the assay to be linear. Serum samples (with low, medium and high total glycated protein concentrations) showed less than 2.1% difference from base results with added glucose (up to 60 mmol/L), less than 9.7% difference with added bilirubin (up to 250 micromol/L) and less than 6.9% with added triglycerides (up to 50 mmol/L). Addition of haemoglobin (up to 0.9 g/dL) with high glycation (11.7% HbA1c) to plasma (298 micromol/L total glycated protein) showed less than 10% difference from the base result. Assays performed over a range of temperatures (12-34 degrees C) showed no significant differences in the results. The assay gives similar results to the currently used NTB method but with significantly less susceptibility to interferences. As such the method should be a useful aid in the management of diabetes.

Alkaline phosphatase activity measurement in the UK by AMP-buffered methods: an appraisal of current practice

Most UK clinical laboratories use alkaline phosphatase (ALP) methods similar to that proposed by the International Federation of Clinical Chemistry (IFCC), based on the use of 2-amino-2-methyl-1-propanol (AMP) buffer. We present evidence of significant differences in results produced by apparently similar commercial ALP methods using an AMP buffer. We compared Bayer DAX, Dade Dimension and Boehringer Mannheim Hitachi 717 methods. Boehringer and Dade results were higher than Bayer results (Bland and Altman analysis, log transformed data): Boehringer (+23.0%, limits of agreement 1.16-1.31 times Bayer); Dade (+21.9%, limits of agreement 1.13-1.32 times Bayer). Biases were predominantly due to differences in reagents rather than analyser characteristics. Compared to a reagent system prepared exactly as described by the IFCC, Bayer was sub-optimal and Dade and Boehringer methods produced results higher than the IFCC method. Reference ranges and results on patients' samples by the various methods showed large differences but no clinically significant difference was observed in external quality assessment schemes either between Bayer and Boehringer or against method means. Apparently similar methods produce different results in patients' sera: external quality assessment schemes are not useful in highlighting these differences.

Evaluation of the Menarini-Arkray HA 8140 hemoglobin A1c analyzer

We describe a multinational evaluation of the Menarini-Arkray HA 8140 hemoglobin (Hb) A1c analyzer, which utilizes a high degree of automation, including bar code reading, cap piercing, and whole-blood sampling. With-in- and between-batch CVs were < 2%. Linearity was confirmed throughout the working range of the analyzer. Common Hb variants, including Hb S, Hb C, and Hb F, did not interfere with the Hb A1c separation, and the potentially interfering labile Schiff base was effectively removed during the chromatographic procedure. The HA 8140 analyzer displayed good correlation to the Bio-Rad Variant analyzer, Tinaquant immunoassay, affinity chromatography, and an optimized "in-house" HPLC Hb A1c method. The methods when compared by Altman and Bland plots showed bias (upper, lower 95% confidence limits) of: Variant minus HA 8140 = 0.99 (0.23, 1.74), Tinaquant minus HA 8140 = 0.14 (-0.71, 0.98); affinity minus HA 8140 (after log transformation) = 1.13 (0.90, 1.41), and "in house" HPLC minus HA 8140 (after log transformation) = 0.91 (0.82, 1.01).

Adaptation of latex-enhanced assay for percent glycohemoglobin to a Dade Dimension analyzer

At present no method for glycohemoglobin (%HbA1c) is automated on a main-line analyzer to allow joint measurement with other indicators of diabetic control such as glucose and cholesterol. We describe an adaptation of a latex-enhanced competitive immunoassay for quantifying %HbA1c to the Dade International Dimension analyzer. After a manual hemolysis step, HbA1c and total hemoglobin (Hb) are determined separately. The concentration of glycated beta-subunit is obtained from the immunoassay, whereas Hb is assessed colorimetrically from a derivatized form. Both reactions were fully optimized for accuracy, precision, and specificity on the Dimension; stabilities of reagents and calibration were established; and potential interferences were assessed. The analyzer gave reliable results over the required clinical range of 1-15% HbA1c. Within-run and total assay variation were within 5% of the target CV limits, as determined by ANOVA with three representative sample pools across 20 days. Close agreement with an established HPLC procedure and a commercially available enzyme immunoassay was observed for 140 samples from clinically defined patient groups. Additional samples from patients with hemoglobinopathies (n = 20) demonstrated a more complex relationship between methods. We conclude that adaptation of the method for use with the Dimension analyzer is a valid method for quantifying %HbA1c.

The effect of low dose recombinant human growth hormone replacement on regional fat distribution, insulin sensitivity, and cardiovascular risk factors in hypopituitary adults

GH deficiency is associated with increased cardiovascular morbidity, which may be determined by alterations in vascular risk factors. We report the effect of partially treated hypopituitarism and subsequent GH replacement (mean dose, 0.2 IU/kg.week) on putative cardiovascular risk factors in 22 nondiabetic hypopituitary subjects in a 6-month, double blind, controlled study (active/placebo ratio, 11:11). All patients were subsequently treated with GH for a further 6 months. Total fat, percent body fat, and central fat were measured by dual energy x-ray absorptiometry. The hypopituitary patients had increased percent fat (P = 0.03) and central fat (P < 0.01) compared with body mass index-matched controls. Before GH treatment, fasting (total) and specific insulin positively correlated with body mass index (P = 0.02 and P < 0.001, respectively), waist/hip ratio (P = 0.05 and P = 0.01), and central fat (P = 0.03 and P = 0.003). Specific insulin and insulin sensitivity (IS), calculated by homeostatic model of assessment, were related to total fat (P < 0.001 and P = 0.02). GH treatment for 6 months led to a reduction in total fat (P < 0.02), percent fat (P = 0.002), central fat (P = 0.012), waist/hip ratio (P < 0.05), total cholesterol (P = 0.03), and apolipoprotein-B (P = 00001), as well as a decrease in the IS from 36.9% (range, 12-100%) to 25% (range, 2.5-55%; P = 0.0002). This was paralleled by a rise in fasting (total) and specific insulin (P = 0.016 and P = 0.002). The degree of correlation among indices of IS, body composition, and fat distribution increased after GH treatment. Fasting plasma glucose rose significantly, but was within the reference range. During 12 months of GH therapy, a significant increase in serum lipoprotein-(a) was observed (P < 0.05). Although GH has beneficial effects on central adiposity and lipid fractions, it is also associated with a decrease in IS; these effects may vary between individuals.

Laboratory evaluation of the DCA 2000 clinic HbA1c immunoassay analyser

The DCA 2000 clinical analyser for the measurement of haemoglobin A1c was evaluated for analytical quality. The analyser, which utilises inhibition of latex agglutination immunoassay, demonstrated good within-batch (1.9-3.1% CV) and between-batch (2.2% CV) imprecision, and was not affected by haemoglobin concentration. The analyser was linear throughout the analytical range, and was found to correlate well with agar electroendosmosis (r = 0.93), affinity chromatography (r = 0.97), HPLC (r = 0.90) and EIA (r = 0.98). The analyser was found to give reliable analytical results, and with its ease of use, will provide the diabetologist with HbA1c results in the clinic; although an analysis time of 9 min will limit sample throughput.

Enzyme immunoassay--a new technique for estimating hemoglobin A1c

We describe a method for estimating hemoglobin A1c (HbA1c) with a commercially available enzyme immunoassay system. The method is based on microtiter plate technology, utilizing an antibody raised to hemoglobin, the epitope being the Amadori product of glucose plus the first eight amino acids on the N-terminal end of the beta chain of hemoglobin. The enzyme immunoassay displays good within-batch (CV 2.3-2.4%) and between-batch (CV 2.6-5.0%) precision, and the results were not affected by different types of anticoagulant. The method was linear within the expected range of results and showed good correlation (r = 0.88-0.98) with established methods for estimating glycohemoglobin. Using this method, we obtained a reference interval of 2.8-4.9% (central 95%) for HbA1c in a nondiabetic population. The percentages of hemoglobin that were HbA1c in diabetics (6.86% +/- 2.51%) were significantly greater (P &lt; 0.001) than in nondiabetics (3.46% +/- 0.52%).

Enzyme immunoassay--a new technique for estimating hemoglobin A1c

We describe a method for estimating hemoglobin A1c (HbA1c) with a commercially available enzyme immunoassay system. The method is based on microtiter plate technology, utilizing an antibody raised to hemoglobin, the epitope being the Amadori product of glucose plus the first eight amino acids on the N-terminal end of the beta chain of hemoglobin. The enzyme immunoassay displays good within-batch (CV 2.3-2.4%) and between-batch (CV 2.6-5.0%) precision, and the results were not affected by different types of anticoagulant. The method was linear within the expected range of results and showed good correlation (r = 0.88-0.98) with established methods for estimating glycohemoglobin. Using this method, we obtained a reference interval of 2.8-4.9% (central 95%) for HbA1c in a nondiabetic population. The percentages of hemoglobin that were HbA1c in diabetics (6.86% +/- 2.51%) were significantly greater (P < 0.001) than in nondiabetics (3.46% +/- 0.52%).

Glycated haemoglobin and fructosamine in non-diabetic subjects with chronic renal failure

Measurements of glycated haemoglobin by electroendosmotic and chromatographic methods, fructosamine, and fructosamine:albumin ratio were made in 91 non-diabetic subjects with chronic renal failure managed conservatively (n = 25), by continuous ambulatory peritoneal dialysis (n = 22), by haemodialysis (n = 22), or by renal transplantation (n = 22). Results were compared with those in a control group of 43 non-diabetic subjects with normal renal function. Mean glycated haemoglobin measured by electroendosmosis was significantly greater in all groups with chronic renal failure except the transplant group. Mean glycated haemoglobin measured by affinity chromatography was not significantly different from controls in any of the groups with chronic renal failure. No difference in mean fructosamine concentration was detected in the transplant or conservatively managed groups compared to controls, but values were significantly lower in the CAPD group, and greater in the haemodialysis group predialysis. Post-haemodialysis samples showed a significant reduction in mean fructosamine concentration when compared with prehaemodialysis samples. Fructosamine:albumin ratios were elevated in all groups of patients with renal failure, with the exception of the transplant group. Of the four indices of glycaemic control considered in this study, only glycated haemoglobin measured by affinity chromatography appears to be unaffected by chronic renal failure.

Glycated haemoglobin and glycated protein and glucose concentrations in necropsy blood samples

Glycated haemoglobin and glycated protein (fructosamine) and blood glucose concentrations were measured in blood samples collected from 75 patients at necropsy. Estimation of blood glucose was a poor indicator of glycaemia before death. Measurement of glycated haemoglobin by affinity chromatography distinguished non-diabetic patients from diabetic patients. The distinction was not as clear cut when HbA1 was estimated using electroendosmosis. Seven patients, who at necropsy had no known history of diabetes, had glycated haemoglobin concentrations in the diabetic range. Two of these patients were found to be diabetic, and diabetes had been suspected at some time in another three patients. It is concluded that measurement of glycated haemoglobin or HbA1, in necropsy specimens is a valuable tool for assessing glycaemic control in known diabetic patients, and may be useful in diagnosing previously unsuspected diabetes.

Glycated haemoglobin analyses--assessment of within- and between-laboratory performance in a large UK region

The performance of glycated haemoglobin methods has been assessed in 20 laboratories. The methods used in these laboratories were affinity chromatography, electroendosmosis and ion-exchange chromatography. Assessment was based on the results returned by the participating laboratories on distributed specimens. Each method displayed acceptable precision and good linearity. Overall there were substantial differences in results reported by different laboratories, though within each method group these differences were less marked.

Glycosylated haemoglobin levels in patients referred for oral glucose tolerance tests

Oral glucose tolerance tests (OGTTs) were performed on 127 patients, and the results interpreted according to the WHO criteria. Glycosylated haemoglobin (GHb), assayed by affinity chromatography, was measured on the fasting blood specimen. All patients classified as diabetic were found to have GHb results greater than 9.9%. There was no overlap with the normal group, all of whom had GHb results less than 8.8%. In the diabetic patients, GHb showed good correlations with both fasting and 2-hour post-glucose blood glucoses. By contrast the non-diabetic group showed only poor correlations between GHb and the fasting and the 2-hour blood glucose concentrations. Overall, the best correlation was found between GHb and the mean of the five blood glucose results obtained during the OGTT.

Glycosylated haemoglobin and glycosylated albumin in non-diabetic and diabetic mothers, and their babies

Glycosylated haemoglobin (GHb), glycosylated albumin (GAlb), and blood glucose were measured from 283 non-diabetic, 7 insulin-dependent diabetic, and 5 gestational diabetic mothers, and in the cord blood of their babies. Significant correlation was found between mother and baby for GHb (r = 0.31) and GAlb (r = 0.42). However, GHb, GAlb and glucose concentration were significantly higher in maternal blood compared to cord blood (p less than 0.001). Only GAlb showed any difference between a pregnant and a non-pregnant population (p less than 0.001). All three parameters were significantly higher in the diabetic mothers compared to the non-diabetic mothers (p less than 0.001). No difference was found between the levels of GHb and GAlb in babies from these two groups. There was no difference in the level of GHb between gestational and insulin-dependent mothers although the latter showed a significantly higher blood glucose (p less than 0.05).

Affinity chromatography: a precise method for glycosylated albumin estimation

A method is described for the estimation, in plasma, of glycosylated albumin, using the commercially available affinity chromatographic material Glycogel B (immobilised m-aminophenylboronic acid on an agarose support) to separate the glycosylated and non-glycosylated albumin, followed by the analysis of albumin concentration using rocket immunoelectrophoresis. Glycosylated albumin and glycosylated haemoglobin showed good correlation (r = 0.91) when both are estimated by affinity chromatography. The glycosylated albumin method displays good within-batch (2.8-4.4% CV) and between-batch (2.9-5.4% CV) precision, and the method is not affected by haemolysis. Using this method a reference range of 2.0-5.4% was found for glycosylated albumin. Levels of glycosylated albumin in diabetics (10.06 +/- 3.23%) were found to be significantly higher (P less than 0.001) than those in non-diabetics (3.72 +/- 0.85%). It was found that loading more than 3 mg of albumin on to a 1 mL affinity column must be avoided, as this appears to overload the column.

Proteins of guinea-pig bile: selective resorption in the gall bladder

We have examined and compared the proteins present in guinea-pig bile as collected either from the common hepatic duct or from the gall bladder. Guinea-pig bile, collected from the common bile duct, has a rather low concentration of protein. Detailed examination shows that the concentrations of actively transported proteins such as immunoglobulin A and haptoglobin:haemoglobin complexes are markedly lower than in rats although the concentrations of proteins which, like albumin, leak non-specifically into bile are similar in the two species. We also find that the protein composition of guinea-pig bile is extensively and selectively modified by resorption of protein in the gall bladder.

A prospective study of the effect of 12 months treatment on serum lipids and apolipoproteins A-I and B in Type 2 (non-insulin-dependent) diabetes

Serum high density lipoprotein cholesterol, total cholesterol, triglyceride and apolipoproteins A-I and B were studied at diagnosis and after 12 months conventional treatment in a group of Type 2 (non-insulin-dependent) diabetic patients. No significant deleterious effect on serum lipids and apolipoproteins was seen in any of the subgroups during the treatment period, although in the overall group there was a significant increase in serum total cholesterol in females. Serum high density lipoprotein cholesterol increased significantly in obese males treated with calorie restriction alone. There was a significant increase in serum apolipoprotein A-I in obese females treated with calorie restriction and metformin and in non-obese females treated with carbohydrate restriction and glibenclamide.