If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Center for Research on Genomics and Global Health, National Human Genome Research Institute, National Institutes of Health, 12 South Drive, Room 4047, Bethesda, MD 20892, United States
Center for Research on Genomics and Global Health, National Human Genome Research Institute, National Institutes of Health, 12 South Drive, Room 4047, Bethesda, MD 20892, United States
Center for Research on Genomics and Global Health, National Human Genome Research Institute, National Institutes of Health, 12 South Drive, Room 4047, Bethesda, MD 20892, United States
Center for Research on Genomics and Global Health, National Human Genome Research Institute, National Institutes of Health, 12 South Drive, Room 4047, Bethesda, MD 20892, United States
Center for Research on Genomics and Global Health, National Human Genome Research Institute, National Institutes of Health, 12 South Drive, Room 4047, Bethesda, MD 20892, United States
Center for Research on Genomics and Global Health, National Human Genome Research Institute, National Institutes of Health, 12 South Drive, Room 4047, Bethesda, MD 20892, United States
The glycated hemoglobin (HbA1c) test can be unreliable in the presence of hemoglobinopathies. The co-existence of type 2 diabetes (T2D) with sickle cell anemia calls for alternative tests. Therefore, we established a reference interval for serum fructosamine and evaluated its utility as a potential glycemic biomarker that is not affected by abnormal hemoglobin.
Methods
The accuracies of serum fructosamine in monitoring and diagnosing T2D were evaluated using the Area under the Receiver Operating Characteristics and other measures in 618 Nigerians with or without sickle cell trait. The estimated diagnostic cut-off for serum fructosamine was then validated in an independent multi-ethnic cohort of 634 West Africans.
Results
Serum fructosamine was similar between individuals with or without sickle cell trait (median: 287 vs 275 umol/L, p = 0·11, respectively) despite statistically different HbA1c. Fructosamine was highly correlated with both HbA1c and fasting glucose independently of sickle cell trait. The areas under the curve (AUC) of serum fructosamine in identifying individuals with uncontrolled glycemia and individuals with T2D were similar and independent of sickle cell trait: 0·92 (95% confidence interval [95% CI ], 0·88-0·95 and 0.92 (95% CI, (0.89–0.95) respectively.
Conclusions
Serum fructosamine is a good alternative to HbA1c for monitoring and diagnosing T2D in the presence of sickle cell trait.
The global prevalence of type 2 diabetes (T2D) has been increasing steadily with low-and-middle- income countries (LMICs) seeing the largest increase in recent years. Approximately 79% of adults with diabetes live in LMICs which include most countries in sub-Saharan Africa. Notably, the number of individuals with diabetes in sub-Saharan Africa is expected to increase by 143% by 2045 [
]. Hemoglobinopathies including sickle cell disease, are common in sub-Saharan Africa. An estimated 10%-30% of the population in West and Central Africa carry the hemoglobin S (HbS) allele [
]. Sickle cell anemia, the most common form of sickle cell disease is an autosomal recessive hemoglobinopathy that is due to a mutation in the β-globin gene (HBB) [
]. Homozygotes for the HbS allele have a severe clinical course resulting in a shortened life expectancy, especially in LMICs. In contrast, heterozygotes (HbAS) referred to as sickle cell trait (SCT), have more subtle manifestations of the disease including extreme exertional injury [
]. The increased turnover of RBC and their abnormal glycosylation in carriers of the sickle gene have implications for tests of glycemic status using HbA1c that assay glycated hemoglobin [
]. HbA1c is indicative of the average blood glucose levels for the previous 2–3 months. However, it may not be optimal in populations with high prevalence of hemoglobinopathies [
].The presence of HbS can lead to inaccurate HbA1c measurements which consequently may lead to inacurrate estimation of past glycemia in individuals with SCT or carriers [
]. Despite the use of National Glycohemoglobin Standardization Program (NGSP)-certified devices whose assays are not affected by HbS interference, the relationship between HbA1c and SCT status is inconsistent across studies [
]. The accuracy of HbA1c may also be affected by other factors including co-existence of other hemoglobinophies (e.g. alpha-thalassemia, glucose 6-phosphate dehydrogenase deficiency (G6PD) [
]. Thus, there is a need for alternative tests that overcome these limitations. Additionally, the use of HbA1c in sub-saharan Africa is hampered by several factors including infrastructure, cost, and lack of population-specific optimum cut-off points [
A proposed alternative to HbA1c is serum fructosamine, a measure of glycated plasma proteins. Like HbA1c, fructosamine assesses peripheral glucose levels over time, but for about 2–3 weeks instead of 2–3 months for HbA1c. Its independence from hemoglobin levels makes it useful in conditions characterized by shortened RBC lifespan including hemoglobinopathies, pregnancy, and hemolytic anemia [
]. Therefore, fructosamine has glycemic monitoring and diagnostic potential although its utility as a reliable glycemic marker in diverse populations has not been well established [
]. To our knowledge, fructosamine has not been systematically explored in populations living with the double burden of hemoglobinopathies and T2D except in a couple of small studies [
In this study, we evaluated the potential utility of fructosamine as a measure of glycemic status in the presence of SCT. First, we established fructosamine reference intervals in an African population. Next, we evaluated the use of fructosamine for the diagnosis and monitoring of glycemic status in T2D. We also validated fructosamine diagnosis cut-off in an independent diverse African population.
2. Materials and methods
The study was conducted according to the Declaration of Helsinki and all relevant ethical regulations related to human subjects. The study protocol was approved by the institutional ethics review board (IRB) of the National Institutes of Health and the IRBs of each participating institution. Written informed consent was obtained from each research participant prior to enrollment.
2.1 Study participants
The participants included in this study were drawn from the Africa America Diabetes Mellitus (AADM) study, which has been previously described elsewhere [
]. Briefly, AADM is an ongoing genetic epidemiology study of T2D, that enrolled participants from multiple medical centers in Nigeria, Ghana, and Kenya. In the current phase of the study, we are conducting deep phenotyping and genotyping of T2D participants and controls at a single site in Ibadan, Nigeria. The first set of analyses (primary analyses) included 650 participants enrolled in Ibadan, Nigeria. To avoid heterogeneity in the sickle cell disease pool and unanalyzable groups (due to small number of individuals), we filtered out homozygous individuals for the S allele (n = 1), or heterozygous for the C allele (AC, SC) (n = 28). After excluding three individuals with missing genotypes, the dataset analyzed included 618 individuals (332 individuals without T2D and 286 with T2D) with normal hemoglobin (Hb[AA]) or with a sickle allele (Hb[AS]). Fasting glucose (FG), 2-hour oral glucose tolerance (2 h-OGTT), and fructosamine were all measured using the Roche Modular-E autoanalyzer (Roche Diagnostics, Indianapolis, IN). HbA1c levels were measured using the latex agglutination inhibition immunoassay on DCA Vantage (Siemens Medical Solutions USA, Inc, Malvern, PA). This HbA1c test is one of three point-of-care (POC) tests approved and certified by the National Glycohemoglobin Standardization Program (NGSP) and the International Federation of Clinical Chemistry (IFCC) [
For the validation study, 634 individuals enrolled from five research sites in Nigeria (Enugu, Lagos, Ibadan) and Ghana (Accra and Kumasi) as part of the parent AADM study and who were not included in the first set of analyses were included. This sample was used to evaluate the estimated diagnosis cut-off for fructosamine from the primary analyses for its sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and AUC in identifying T2D cases (Supplemental Fig. 1).
Fig. 1Overall population-based reference interval: histogram with the gaussian curve of fructosamine distribution. Central black line represents fructosamine median in the reference sample; the left and right black lines represent the lower and upper values of the 90% CI lower and upper limits of the reference interval [194–309] mol/l respectively. The grayed areas represent areas outside the 90% CI. Green horizontal represents the population-based RI. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
All participants underwent a clinical examination that included a medical history, anthropometry, blood pressure measurements, and blood collection. The definition of T2D was based on the American Diabetes Association (ADA) criteria: a fasting glucose concentration (FG) ≥ 126 mg/dl (7.0 mmol/l) or a 2-h postload value in the oral glucose tolerance test ≥ 200 mg/dl (11.1 mmol/l) on more than one occasion. Alternatively, a diagnosis of T2D was accepted if an individual was on pharmacological treatment for T2D and review of clinical records indicated adequate justification for that therapy. The detection of autoantibodies to glutamic acid decarboxylase (GAD) and/or a fasting C-peptide ≤ 0.03 nmol/l was used to exclude probable cases of type 1 diabetes. Given that most cases were anteriorly identified using ADA criteria and were confirmed to be using T2D medication as a result of a diagnosis of T2D, we used T2D medication as proxy in our analyses.
HbA1c levels were also used to identify T2D cases following ADA guidelines i.e. HbA1c ≥ 6·5% irrespective of SCT status . According to ADA criteria, glycemic control is achieved in individuals with T2D if FG is < 130 mg/dl or HbA1c is < 7%.
2.2 Statistical analysis
Data was analyzed using IBM SPSS Statistics for Windows Version 26·0 (Armonk, NY: IBM Corp) . Indeterminate results for any variable were set as missing. Any variables lower or higher than the limit of detection was set to the lowest or the highest limit of detection of that specific variable. This was only observed for HbA1c where any value > 14% was set to 14%. (n = 28).
For variables with skewed distributions, non-parametric tests were used to compare groups (individuals with T2D vs individuals without T2D or non-carriers (Hb[AA]) vs carriers (Hb[AS]). Spearman’s coefficients of correlation were used to test associations between fructosamine, HbA1c, FG, and 2 h-OGTT and between fructosamine, age, and gender. To determine fructosamine reference interval (RI), we used the International Federation of Clinical Chemistry (IFCC) recommendations [
]. Only apparently healthy individuals (reference group) were included for the RI calculation and were selected from the non-diabetic group using a stringent definition based on all 3 ADA criteria i.e. only individuals with FG < 100 mg/dl and 2 h-OGTT) < 140 mg/dl and HbA1c < 5·7 were included (n = 186). The resulting fructosamine data was transformed using a Cox-transformation followed by Tukey method -based on an interquartile range to identify and remove outliers (n = 7). A total of 179 individuals were included in the estimation of the RI. The detailed steps involved in calculating the RI are found in the supplemental method file.
The area under the receiver operating characteristic (AUROC) curve was used to evaluate fructosamine ability: 1) to monitor glycemic control in T2D, and 2) to diagnose T2D in the presence or absence of SCT. P-values ≤ 0·05 were considered significant. Fructosamine cut-offs were determined using Youden index (YI = sensitivity + specificity-1).
3. Results
The median age of the participants was 57 years and 46·3% of them had T2D (Table 1). As expected, individuals with T2D were slightly heavier with a significantly larger waist circumference (difference in median waist circumference ~2 cm) and had significantly higher fructosamine, HbA1c, and FG compared to individuals without T2D (Table 1). About 25% of the participants were carriers of the sickle cell mutation (HbAS), and the frequency was similar across individuals with or without T2D. There was no statistical difference in age, anthropometric and clinical biomarkers between carriers and non-carriers except for HbA1c that was higher in the carriers compared to the non-carriers (Table 1). The evaluation of the use of T2D medication and HbA1c as T2D classifiers using Cohen's kappa test showed that there was a moderate agreement between the two diagnosis variables (κ = 0·70). While there was no statistically significant association between T2D and SCT (p = 0·67), both the use of T2D medication and HbA1c estimated a higher frequency of T2D among carriers, Hb[AS] 47.7% and 38.7% respectively; the agreements between the two diagnosis variables remained about the same regardless of the presence of SCT (κHb[AA] = 0·68, κHb[AS] = 0·69).
Table 1Anthropometric and clinical characteristics of the study participants by type 2 diabetes status and sickle cell trait status.
All values in the table are median (IQR) except for sickle-cell trait (SCT) and sex which are frequencies. Medians among groups were compared by a nonparametric test (Median test). Frequencies among groups were compared by Chi-square test.
3.1 Determination of fructosamine reference interval (RI) in an African population.
The reference group consisted of a subset of the individuals without T2D that included only individuals without impaired glucose homeostasis. Overall, the individuals included had a healthier metabolic profile compared to the entire group of individuals without T2D (Supplemental Table 1).
Given that the usefulness of RI can be improved by stratification of a priori confounders, we evaluated the associations between fructosamine, age, BMI, and SCT and used the Harris & Boyd method to determine the utility of partitioned RI. There was no association between fructosamine and BMI (r = -0·14, p = 0·08), age (r = 0·13, p = 0·10), sex (r = -0·03, p = 0·69), and SCT (r = 0·08, p = 0·059). Additionally, all calculated standard deviation ratio between subgroups were < 1·5 (Supplemental Table 2). Thus, only an overall population-based RI was determined. The median fructosamine in this reference sample is 253 (37) and RI: [194–309] with 90% CI for the lower limit and upper limit being [189–197] and [299–312] respectively (Fig. 1).
Table 2Fructosamine diagnostic ability in identifying individuals with diabetes in the presence or absence of SCT.
Fructosamine (test variable)
T2D defined by the use of T2D medication
T2D defined by HbA1C
AUC (se)
p-value
95% CI
AUC (se)
p-value
95% CI
ALL
0·85(0·02)
<0·001
0·82-0·88
0·92 (0·01)
<0·001
0·89-0·95
Hb [AA]
0·84 (0·02)
<0·001
0·80-0·88
0·92 (0·02)
<0·001
0·89-0·95
Hb [AS]
0·87(0·03)
<0·001
0·80-0·93
0·91 (0·03)
<0·001
0·86-0·97
AUC: fructosamine’s area under the curve; se: standard error; CI: AUC confidence interval; T2D if use of T2D medication (left panel); T2D if HbA1C ≥ 6.5% (right panel)
3.2 Correlations between fructosamine and established glycemic markers
In individuals with T2D, fructosamine was highly associated with both HbA1c (rHb[AA] = 0·80, vs rHb[AS] = 0·82, p < 0·001) and FG (rHb[AA] = 0·72, rHb[AS] = 0·83, p < 0·001) whereas in individuals without T2D, the associations were weaker (0·16-0·27) for all 3 glycemic biomarkers in non-carriers; however, no association was found in carriers except for FG (rHb[AS] = 0·30, p = 0·010).The association strength seems to be stronger in carriers in both individuals with and without T2D (Supplemental Table 3).
3.3 Monitoring of glycemic control in T2D using fructosamine
Using glycemic control definition, set forth in the method section, we compared fructosamine ability to monitor glycemic control. In this study, 99% (283/286) of individuals with T2D were on pharmacological treatment for T2D. However, only about half achieved expected glycemic control as defined by FG < 130 mg/ml (49·3%) or HbA1c < 7% (40·6%). AUROC analyses demonstrated that fructosamine is a good biomarker for predicting individuals with T2D who did not achieve glycemic control in this population. When FG > 130 mg/dl and HbA1c > 7% were used to define “uncontrolled glycemic control”, fructosamine (AUC) were 0·88 and 0·92 respectively (Fig. 2 (A&D). Also, SCT had no effect on fructosamine discriminative power in predicting individuals with uncontrolled glucose levels (Fig. 2 (B-C;E-F). In clinical setting HbA1c is the reference test for monitoring glycemic control, so we only reported fructosamine monitoring accuracy using HbA1c as classifier . At cut-off point of 352·5 umol/L, fructosamine had 81·5% [CI 95%:0·76-0·87] sensitivity and 97·4% [CI 95%:0·95-1·00] specificity for identifying individuals with uncontrolled glucose levels (i.e. fructosamine > 352·5 umol/L).
Fig. 2Fructosamine ROCs to predict glycemic control in T2D cases.Fructosamine = test variable ROCs for predicting glycemic control defined as followed:Uncontrolled glycemia is defined as FG ≥ 130 mg/dl (Figures A, B, C); uncontrolled glycemia is defined as HbA1c > 7% (Figures D, E, F). #Given that HbA1C is the Gold standard for monitoring treatment effectiveness, it was used as reference to estimate fructosamine cut-off to identify individuals with uncontrolled glycemia.
We used AUROCs to test fructosamine ability to identify diabetes and found that fructosamine had a good discriminative ability in identifying T2D as shown by high AUCs (>0·80) and regardless of the presence or absence of SCT (Table 2, Supplemental Fig. 2, A-F). Given that there was no difference in fructosamine’s AUCs between carriers and non-carriers, the estimation of fructosamine cut-offs were conducted using the entire cohort (i.e. independently of SCT). Thus, when T2D is defined by HbA1c, the optimal sensitivity and specificity for fructosamine were 82·4% [95% CI:0·77-0·88] and 92·4%[95% CI: 0·90-0·95] respectively at the threshold of 309·5 umol/L. When T2D is defined by the use of T2D medication, the sensitivity and specificity were 69·7% [95% CI:0·64-0·75] and 91·1% [95% CI: 0·88-0·94] respectively at the threshold of 300·5 umol/L. Despite the difference (Δ = 9 umol/L) in the estimated fructosamine cut-offs by the two established biomarkers in identifying T2D cases, the overall global accuracy for fructosamine is approximatively the same (HbA1c: 1·75 and the use of T2D medication : 1·61). Additionally, we used each of the estimated fructosamine cut-off to classify the participants into diabetes or non-diabetes groups and evaluate agreement. For fructosamine ≥ 300·5 umol/L, the frequency of T2D was 37·8% and for fructosamine ≥ 309·5 umol/L, the frequency was 33·3%. There was almost perfect agreement between the two cut-offs in identifying T2D cases (κ = 0·90). Thus, we derived a single cut-off in order to simplify our analysis. Based on the calculated RI, any fructosamine levels > 309 umol/L is suggestive of abnormal levels, therefore, we estimated that fructosamine cut-off at 309·5 umol/L (i.e. fructosamine ≥ 309·5 umol/L) is suitable in this population to identify T2D cases. Using this fructosamine cut-off showed that fructosamine had a good concordance -as reflected by its sensitivities and specificities- and substantial agreement (κFructosmine/use of T2Dmedication) = 0·61, κFructosamine/HbA1c = 0·75) with the two established variables examined in this population. Fructosamine successfully identified T2D cases as shown by its predictive positive values (PPV) that were>85% (Fig. 3).
Fig. 3Fructosamine performance in screening for T2D cases. Figure displays the accuracy measures of fructosamine in the primary and validation analyses when T2D is defined by:HbA1C ≥ 6.5% (black bars),T2D medication use [Gray bars (primary analysis) and wave bars (in the validation analysis). FRA ≥ 309.5 umol/L was used to define T2D cases.NPV: negative predictive value; PPV: positive predictive value.
3.5 Evaluation of fructosamine cut-off in an independent cohort of African individuals with diabetes
Demographic and clinical characteristics of participants included in this validation stage are shown in Supplemental Table 4. Overall, participants had a median age of 51 years and the frequency of T2D was 52.7%. Individuals with diabetes were significantly older (Δage = 5·6), heavier (ΔBMI = 2·2), and had larger WC (ΔWC = 7·5) than controls. As expected, they also had significantly higher FG and fructosamine than controls. In comparison to the cohort used in the primary analyses, individuals in this cohort were 6 years younger, 5 BMI units lighter, had smaller WC and lower FG. However, the percentage of males, frequency of T2D, and fructosamine levels were higher (Table 1 and Supplemental Table 4). HbA1c was not available in this cohort, therefore the validation of fructosamine cut-off in identifying T2D was conducted using the use of T2D medication only as a reference classifier for T2D status. As done above, the estimated fructosamine cut-off of 309·5 umol/l was used to identify T2D cases. This resulted in T2D frequency of 40·7% which is in line with the trend observed in the primary cohort i.e. lower T2D frequency estimate by fructosamine. The difference in T2D frequency by fructosamine and use of T2D medication in both cohorts was very similar (12% in the validation cohort vs 13% in the primary cohort). The degree of agreement between the use of T2D medication and fructosamine as classifiers was also comparable (κ(validation) = 0.68 vs κ(primary) = 0.61). Similarly, fructosamine was able to discriminate T2D cases from individuals without T2D with a PPV of 94·6% (Fig. 3).
4. Discussion
In this analysis including more than a thousand Africans, we demonstrated that serum fructosamine is a good alternative to HbA1c for monitoring and diagnosing T2D in populations where SCT is prevalent. We established the first RI and suggested clinical decision limits for fructosamine in an African population to facilitate interpretation and comparisons between populations.
Fructosamine is an intermediate glycemic marker proposed as an alternate to HbA1c under certain conditions including hemoglobinopathies [
]. In this study,we did not observe a significant difference in fructosamine levels by SCT indicating that serum fructosamine is not affected by HbS mutation. However, in contrast to recently published study by Lacy, M et al. [
] in African Americans, we observed higher levels of HbA1c in carriers of SCT compared to non-carriers. While the reasons for this discrepancy are unknown, our study is one of several studies to report higher levels of HbA1c in carriers of SCT. In a cohort of Black patients, the genetic variant (rs334) responsible for SCT was associated with higher HbA1c despite adjusting for confounders such as BMI, waist circumference and principal component factors [
]. Additionally, a systematic review to assess how SCT affects the interpretation of HbA1c showed conflicting results across studies, with some studies reporting lower levels of HbA1c and others reporting higher or same levels in carriers of SCT compared to non-carriers [
In individuals with T2D, moderate to strong correlations have previously been observed between fructosamine, FG, and HbA1c as confirmed in this study [
Hammar N Fructosamine Is a Useful Indicator of Hyperglycaemia and Glucose Control in Clinical and Epidemiological Studies – Cross-Sectional and Longitudinal Experience from the AMORIS Cohort.
]. Overall, the strength of the associations between fructosamine and HbA1c was slightly stronger than that of fructosamine and FG, consistent with previous studies [
]. The stratified analysis by SCT status revealed the same trend in carriers and non-carriers. FG is a snapshot of glycemic state whereas fructosamine and HbA1c are measures of exposure, thus the expectation for fructosamine and HbA1c to be more correlated. Unlike in individuals with T2D and consistent with previous reports, the associations between fructosamine and three established glycemic markers were considerably weaker in individuals without diabetes regardless of SCT status [
]. It should be noted that other alternative markers including 1,5-Anhydroglucitol show the strongest correlations only in the highest glucose concentration range [
To our knowledge, this is the largest study to simultaneously evaluate fructosamine as a monitoring and diagnostic tool in a well-phenotyped African population using an analytic approach that did not rely on reference ranges established for fructosamine in other populations. As shown by previous studies, using thresholds from other populations or established by commercial companies could introduce misclassification and consequently misdiagnosis if fructosamine were to be adopted as viable diagnostic tool in LMICs [
Hammar N Fructosamine Is a Useful Indicator of Hyperglycaemia and Glucose Control in Clinical and Epidemiological Studies – Cross-Sectional and Longitudinal Experience from the AMORIS Cohort.
].We found that the percentage of individuals above the cut-off estimated (309·5 umol/l) in this study was comparable to the one obtained using HbA1c. The degree of agreement between fructosamine and HbA1c varies considerably across studies [
] but a systematic comparison of the agreements between studies is not feasible due the lack of standardization of methods used to measure concordancy, and/or factors including glycemic status over different time frames, glycation gap, and RBC-related pathologies [
]. Additionally, population differences have been reported for fructosamine; but no cut-off points specific for populations of African ancestry have been systematically evaluated [
]. The RI reported by Selvin et al. for fructosamine is wider [198·2-267·8 umol/l] in African Americans than in Caucasians [194·8-258 umol/l] from the same cohort but narrower than the RI estimated in this study [194–309 umol/l] [
HbA1c is the gold standard for monitoring glucose control. However, fructosamine can be an earlier indicator of poorly controlled glucose levels but is not often used in clinical settings and does not have published cut-offs for accessing glycemic control. To our knowledge, this is the first time fructosamine has been systematically evaluated as an index of glycemic control and we estimated an optimal threshold with high sensitivities and specificities for identifying both individuals with T2D and those with uncontrolled glucose levels. Although at the diagnosis threshold established in this study the sensitivity is lower when the use of T2D medication was used as T2D classifier than when HbA1c was used, both specificity and PPV increased. PPV depends on the prevalence of the disease and the increased observed PPV could be the reflection of the higher frequency of T2D estimated by the use of T2D medication compared to HbA1c. Despite differences in clinical and anthropometric markers between the primary and validation cohorts, the measures of accuracy for fructosamine were similar. This finding suggests that the estimated threshold can be used in other west-african populations given that the validation cohort represented multiple ethnic groups from Nigeria and Ghana.
5. Conclusions
The strengths of this study include establishing a reference interval for fructosamine; evaluating diagnosis, and monitoring thresholds for fructosamine in an under-studied population with the double burden of metabolic disorders and sickle cell anemia, and validation of the estimated threshold in an independent diverse west-african population. These types of studies are needed to address the challenges encountered in LMICs for both the diagnosis and monitoring of T2D but also to explore alternative glycemic biomarkers that are deemed more suitable for LMICs [
]. For example, HbA1c is more expensive and less accessible in LMICs and may not be the optimal biomarker in the presence of pathologies that are more prevalent in these countries [
]. This study did not rely on previously established fructosamine cut-offs in populations that are ancestrally and environmentally different from our target population but instead follow defined published procedures to produce population-specific reference interval and clinical decision limits (CDL) [
]. The main limitation of our investigation was our inability to validate in the secondary cohort the estimated fructosamine diagnosis threshold when T2D is defined by HbA1c. However, we note that HbA1c is currently not yet a routine procedure globally, especially in sub-Saharan Africa [
]. Factors such as malnutrition and disorders of albumin metabolism can affect fructosamine levels but were not evaluated in our investigation. For example, in disorders characterized by hypoalbuminemia including hypothyroidism, protein-losing enteropathy, nephrotic syndrome, or liver failure resulting in higher levels of fructosamine [
Lorena Alarcon-Casas Wright MaIBH, MD. The Challenge of the Use of Glycemic Biomarkers in Diabetes: Reflecting on Hemoglobin A1C, 1,5-Anhydroglucitol, and the Glycated Proteins Fructosamine and Glycated Albumin. Diabetes Spectrum 2012. 2012;25:141-8.
Lorena Alarcon-Casas Wright MaIBH, MD. The Challenge of the Use of Glycemic Biomarkers in Diabetes: Reflecting on Hemoglobin A1C, 1,5-Anhydroglucitol, and the Glycated Proteins Fructosamine and Glycated Albumin. Diabetes Spectrum 2012. 2012;25:141-8.
Lorena Alarcon-Casas Wright MaIBH, MD. The Challenge of the Use of Glycemic Biomarkers in Diabetes: Reflecting on Hemoglobin A1C, 1,5-Anhydroglucitol, and the Glycated Proteins Fructosamine and Glycated Albumin. Diabetes Spectrum 2012. 2012;25:141-8.
]. Though, this study did not specifically consider these limitations, future studies should address them, especially in under-studied populations.
The findings of this study address the gap in knowledge about systematic evaluation of alternate glycemic biomarkers in diverse populations, especially in the presence of population-specific confounders such as SCT.
Contributors
AD, HF, KE, AA, CR conceived the study. AD, JZ maintained and managed the data. They also run the initial quality controls of the data. AD, HF reviewed data, did the statistical analyses as well as literature search and drafted the manuscript. All authors collaborated in the interpretation of the results, revision and editing of the report.
Funding
Support for this study is provided by NIH Grant No. 3T37TW00041-03S2 from the Office of Research on Minority Health and the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant DK-54001.This research was supported in part by the Intramural Research Program of the Center for Research on Genomics and Global Health (CRGGH). The CRGGH is supported by the National Human Genome Research Institute, the NIDDK and the Office of the Director at the National Institutes of Health (Z01HG200362).
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We are grateful to the participants and staff of the AADM project.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Hammar N Fructosamine Is a Useful Indicator of Hyperglycaemia and Glucose Control in Clinical and Epidemiological Studies – Cross-Sectional and Longitudinal Experience from the AMORIS Cohort.
Lorena Alarcon-Casas Wright MaIBH, MD. The Challenge of the Use of Glycemic Biomarkers in Diabetes: Reflecting on Hemoglobin A1C, 1,5-Anhydroglucitol, and the Glycated Proteins Fructosamine and Glycated Albumin. Diabetes Spectrum 2012. 2012;25:141-8.
00278-3&id=gr1.jpg){kind=link}
00278-3&id=gr2.jpg){kind=link}
00278-3&id=gr3.jpg){kind=link}