Abstract
Background
Metabolic syndrome (MetS) is a multifactorial disorder and a predisposing factor for diabetes, heart diseases, and stroke. Glycated haemoglobin (HbA1c) has recently received considerable attention as a potential marker to identify subjects at risk of MetS. This study aimed at assessing the performance of fasting plasma glucose (FPG), the American Diabetes Association (ADA) HbA1c cut-off, and a population-derived HbA1c (pHbA1c) cut-off value as the glycaemic criterion for MetS in a non-diabetic population.
Methods
In this cross-sectional study, we recruited 728 non-diabetic Ghanaian adults. Venous blood sample was obtained and fasting plasma insulin and glucose, HbA1c, lipid profile, blood pressure and anthropometric measurements were performed for each respondent.
Results
The prevalence of MetS using the FPG, ADA HbA1c and pHbA1c criteria were 35.2%, 38.5% and 41.8%, respectively. The pHbA1c cut-off identified 6.6% and 3.3% more subjects with MetS when compared with FPG and the ADA HbA1c cut-offs, respectively while the ADA HbA1c cut-off identified 3.3% more subjects with MetS compared with the FPG criterion. The ADA HbA1c criterion showed a substantial agreement (ĸ = 0.79) with the FPG criterion while pHbA1c showed an almost perfect concordance (ĸ = 0.82) with the FPG criterion and an excellent sensitivity and specificity for identifying subjects with MetS in the study population.
Conclusion
Screening of MetS by introduction of the ADA HbA1c criterion in addition to the traditional FPG criterion enhances the detection of more people with MetS. However, the use of population-derived HbA1c cut-off value could potentially identify even greater number of high risk subjects in that specific population.
Keywords: Epidemiology, Cardiovascular system, Metabolism, Metabolic disorder, Nutrition, Fasting plasma glucose, Glycated haemoglobin, Metabolic syndrome
Epidemiology; Cardiovascular system; Metabolism; Metabolic disorder; Nutrition; Fasting plasma glucose; Glycated haemoglobin; Metabolic syndrome.
1. Introduction
Metabolic syndrome (MetS) is a multifactorial disorder. It is diagnosed based on clustering of closely related cardiovascular risk factors such as hypertension, obesity, hyperglycaemia, and dyslipidaemia. MetS is used as a clinical tool for the identification of patients with metabolic risk of cardiovascular diseases (CVD) [1, 2], and serves as a predictor of cardiovascular events and diabetes [3]. Insulin resistance is the primary pathophysiologic mechanism; however, due to the difficulties and the complex nature of the direct measurement of insulin sensitivity, some anthropometric, hemodynamic, and biochemical parameters have been employed in the diagnosis of MetS [4]. Various organizations have proposed different criteria for the diagnosis of MetS which have led to disparities in the identification of high risk subjects. This consequently resulted in the need for harmonization of the criteria for defining MetS (Joint Interim Statement-JIS) [5].
Despite the coherence in the harmonized definition, there are inter-population variabilities in the cut-off values for fasting plasma glucose (FPG). Moreover, the use of point glucose estimation may not be reliable. This heralded the proposal for the utilization of glycated hemoglobin (HbA1c) as one of the diagnostic criteria for diabetes and the categorization of subjects with increased risk for diabetes. As result, the American Diabetes Association (ADA) proposed that HbA1c value of 5.7–6.4% should be considered a criterion for increased risk for diabetes [6].
HbA1c is a predictor for diabetes [7] and used in the monitoring of long-term glycaemic control as it is the index of mean blood glucose over a period of three months. Advantages of its use over the fasting plasma glucose include the convenience of not requiring a fasting sample, requiring less time and the superior technical attributes [6]. Furthermore, HbA1c captures chronic hyperglycaemia, including postprandial glucose spikes in contrast to daily pre-prandial glucose snapshot offered by FPG [8]. Although it is debatable, some cross-sectional [9, 10, 11], case-control [3], follow-up [12], prospective observational [13], and large-scale longitudinal [8] studies have reported HbA1c to be a better predictor of cardiovascular risk compared to FPG even among non-diabetic population. Moreover, we have previously reported the usefulness of the ADA HbA1c cut-off over FPG for MetS diagnosis among non-diabetic population in Ghana [14]. Nonetheless, evidence suggests that the cut-off for HbA1c might differ by ethnicity due to significant discordance in the association between HbA1c and FPG in diverse populations [15, 16].
It is against this background that we assessed and compared the performance of fasting plasma glucose with the ADA HbA1c, and evaluated the plausibility of a population-derived HbA1c cut-off value as the dysglycaemic criterion for MetS in a non-diabetic population.
2. Materials and methods
2.1. Study design, population and area
This was a cross-sectional study conducted at St Francis Xavier Hospital, Assin Fosu, Central Region, Ghana. A total of 728 apparently healthy Ghanaian adults, living in Assin Foso were recruited for the study. The population of Assin North Municipality according to the 2010 Population and Housing Census is 161,341 representing 7.3% of the region's total population [17]. The protocol for this study was approved by the Committee on Human Research Publication and Ethics (CHRPE) of the School of Medical Sciences, Kwame Nkrumah University of Science and Technology. All participants gave their written informed consent after the aim and potential risks involved in participation had been explained to them.
2.2. Inclusion and exclusion criteria
Apparently healthy Ghanaian adults between the ages of 30 and 70 years old were included in this study. Subjects clinically diagnosed of hypertension, diabetes or family history of diabetes, and subjects suspected of malignancies, inflammatory disease, as well as subjects on lipid or glucose-lowering medication, or antihypertensive agents were excluded from the study. Subjects with anaemia, thalassemia, and sickle cell as well as subjects with FPG ≥ 7.0 mmol/L, classified as diabetics according to the American Diabetes Association [18], were also excluded from the study.
2.3. Blood pressure and anthropometric evaluation
Blood pressure (BP) was measured with an automated blood pressure apparatus (Omron MX3-Omron Matsusaka Co., Ltd. Japan). The average of the two readings taken five minutes apart was recorded as the blood pressure measurement. The weight was measured in light clothing without shoes, in an upright position using a calibrated analogue scale (Seca, Hamburg, Deutschland). Height was measured without shoes using a stadiometer (Seca, Hamburg, Deutschland). Body mass index (BMI) was calculated using the equation; [BMI (kg/m2) = weight/height2] [19]. Using a measuring tape, waist circumference (WC) was measured at the narrowest part of the waist between lower end of the twelfth rib and iliac crest, and hip circumference (HC) was measured at the widest part of the hips (below the iliac crest at the level of the greater trochanters). The WC, HC, height as well as other parameters were used to calculate waist to height ratio (WHtR) = WC (m)/height (m), waist to hip ratio (WHR) = WC (m)/HC (m), body adiposity index (BAI) = (HC (cm))/(height (m)1.5) –18 [20], and visceral adiposity index (VAI) = WC (m)/(39.68 +(1.88 x BMI)) x [triglyceride (TG)/(1.03)] x ((1.31)/high density lipoprotein (HDL)) for males and (WC (m))/(36.58 +(1.89 x BMI)) x (TG/(0.81))x ((1.52)/HDL) for females [21]. Blood pressure and anthropometric measurements were carried out on each participant by same trained personnel.
2.4. Blood sampling, processing and analysis
From each participant, about 5 ml of venous blood was obtained from the antecubital vein after an overnight fast. One milliliter (1 ml) was dispensed into a fluoride oxalate tube, 1ml into EDTA tube, and 3 ml into gel separator tubes. The tubes were placed in a centrifuge and spun at 3000 rpm for 10 min to obtain the plasma and serum. Plasma glucose was measured immediately and the serum for the measurement of other biochemical variables were stored at -20 °C until analysis. Insulin was assayed by sandwich ELISA method (Cat # EIA-2935; DRG International Inc., Springfield Township, USA). Fasting plasma glucose ((FPG): Cat # 20767131-322) and lipid profile (Total Cholesterol (TCHOL): Cat # 03039773-190; Triglycerides (TG): Cat # 20767107-322; High Density Lipoprotein (HDL): Cat # 07528566-190) were estimated enzymatically using Cobas Integra automated Chemistry analyzer (Cobas Integra 400 Plus, Roche Diagnostics, USA). Low density lipoprotein (LDL) concentration was determined using Friedewald's formula: LDL (mmol/L) = TCHOL (mmol/L) – HDL (mmol/L) – [triglyceride (mmol/L)/2.2] [22]. Homeostasis model assessment of insulin resistance (HOMA-IR) was calculated using the HOMA2 calculator. Whole blood was used for HbA1c estimation by turbidimetric inhibition immunoassay (Cat # 04528123-190) using the Cobas Integra automated Chemistry analyzer (Cobas Integra 400 Plus, Roche Diagnostics, USA) as previously described [14]. Briefly, well-mixed EDTA-anticoagulated whole blood was put into sample tubes. The tubes were immediately placed on a rack. Red blood cells were haemolysed by low osmotic pressure and the free haemoglobin subsequently degraded by pepsin to ensure availability of the N-terminal of the beta chain (β-N-terminal) of haemoglobin. Latex particles-bound monoclonal antibodies bind to the β-N-terminal of HbA1c while the remaining free antibodies are agglutinated using synthetic polymers with multiple copies of the β-N-terminal structure of HbA1c. The change in turbidity is measured at 552 nm and the final HbA1c value expressed as a percentage using the formula: HbA1c (%) = (HbA1c/Hb) × 87.6 + 2.27. The test was standardized with an intra-assay %CVs of 0.9%–1.5% and inter-assay %CVs of 1.1%–1.6%. The method for the HbA1c determination is among the 2019 National Glycohemoglobin Standardization Program (NSGP) list of certified methods [23].
Daily calibration and maintenance of the analyzer was performed according to the manufacturer's instructions. Quality control (QC) was assessed using quality control materials provided by the manufacturer [negative and positive controls (high and low HbA1c)] and calibration was performed using manufacturer-supplied calibrator (Cfas HbA1c).
2.5. Definition of metabolic syndrome
The MetS was defined according to the consensus criteria by the Joint Interim Statement (JIS) [5]. Subjects were classified as having MetS by the presence of three or more of the following: elevated WC (≥94 cm and ≥80 cm for African male and female, respectively); TG ≥ 1.7 mmol/l; HDL <1.0 mmol/l in men and <1.3 mmol/l in women; BP ≥ 130/85 mmHg; or FPG ≥ 5.6 mmol/l. Dysglycaemia was defined by FPG ≥ 5.6 mmol/l or HbA1c ≥ 5.7% according to the American Diabetes Association [18].
2.6. HbA1c cut-off value for the prediction MetS
In this study, we employed the receiver operating characteristic (ROC) curve analysis to obtain a study population specific HbA1c cut-off value (represented as pHbA1c) for the prediction of MetS as described by Park et al [9]. The cut-off was based on the HbA1c value which corresponds to the FPG value of 5.6 mmol/L (100 mg/dL) diagnostic of dysglycaemia as defined by the American Diabetes Association [18]. A pHbA1c value of 5.4% was derived which showed an almost perfect agreement (ĸ = 0.82) with the FPG-based criterion. Diagnosis of pHbA1c-based MetS was done by replacing FPG in the Joint Interim Statement (JIS) [5] with pHbA1c cutoff of 5.4%.
2.7. Data analysis
All categorical data were presented as frequencies (percentages) and continuous data as mean ± SD. Chi-squared/Fisher exact test used to compare categorical variables and independent t-test was used to compared continuous data. Pearson's correlation was used to evaluate the association between HbA1c and cardiovascular risk factors. The kappa (ĸ) statistic was used to evaluate the agreement between FPG- and HbA1c-based criteria for MetS. The sensitivity, specificity, positive predictive (PPV) and negative predictive (NPV) values of the HbA1c criteria were evaluated using the FPG-based criterion as the reference. Reliability was expressed as the J index [(TP×TN) - (FP×FN)]/[(TP + FN) (TN + FP)]. A p value <0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS version 25.0.
3. Results
The mean age of the study population was 50.4 (±10.1) years. A higher proportion of the participants were female (56.0%). Participants with MetS had significantly elevated haemodynamic, and anthropometric indices compared with subjects without MetS. Higher lipid profile parameters, fasting insulin, homeostasis model assessment of insulin resistance (HOMA-IR), and HbA1c as well as lower HDL were found among participants with MetS compared with participants without MetS across all three criteria (Table 1).
Table 1.
Baseline characteristics of the study population.
| Criteria | Total (n = 728) | FPG-based criterion |
ADA HbA1c-based criterion |
pHbA1c-based criterion |
|||
|---|---|---|---|---|---|---|---|
| MetS positive | MetS negative | MetS positive | MetS negative | MetS positive | MetS negative | ||
| Age (years) |
50.4 ± 10.1 |
50.13 ± 10.5 |
50.5 ± 10.0 |
51.1 ± 10.9 |
49.8 ± 9.5 |
51.1 ± 10.9 |
49.8 ± 9.5 |
| Sex∗ | |||||||
| Male | 320 (44.0) | 88 (27.5) | 232 (72.5)‡ | 88 (27.5) | 232 (72.5)‡ | 88 (27.5) | 232 (72.5)‡ |
| Female | 408 (56.0) | 168 (41.2) | 240 (58.8) | 168 (41.2) | 240 (58.8) | 216 (52.9) | 192 (47.1) |
| SBP | 117.8 ± 13.8 | 125.7 ± 11.1 | 113.5 ± 13.3† | 125.7 ± 9.0 | 112.1 ± 13.9† | 125.7 ± 9.0 | 112.1 ± 13.9† |
| DBP | 75.0 ± 9.4 | 77.4 ± 8.9 | 73.7 ± 9.4† | 77.3 ± 8.7 | 73.4 ± 9.5† | 77.3 ± 8.7 | 73.4 ± 9.5† |
| WC | 92.0 ± 11.9 | 101.7 ± 9.7 | 86.8 ± 8.4† | 99.5 ± 10.5 | 86.7 ± 8.8† | 99.5 ± 10.5 | 86.7 ± 8.8† |
| BMI | 25.8 ± 5.0 | 29.2 ± 5.3 | 23.9 ± 3.7† | 28.7 ± 5.0 | 23.7 ± 3.8† | 28.7 ± 5.0 | 23.7 ± 3.8† |
| WHR | 0.9 ± 0.1 | 0.9 ± 0.1 | 0.8 ± 0.1† | 0.9 ± 0.1 | 0.8 ± 0.1† | 0.9 ± 0.1 | 0.8 ± 0.1† |
| WHtR | 0.6 ± 0.1 | 0.6 ± 0.1 | 0.5 ± 0.1† | 0.6 ± 0.1 | 0.5 ± 0.1† | 0.6 ± 0.1 | 0.5 ± 0.1† |
| BAI | 31.4 ± 6.6 | 34.2 ± 7.0 | 29.8 ± 5.8† | 34.4 ± 6.6 | 29.2 ± 5.6† | 34.4 ± 6.6 | 29.2 ± 5.6† |
| VAI | 2.5 ± 2.1 | 4.1 ± 2.5 | 1.6 ± 1.1† | 3.8 ± 2.4 | 1.6 ± 1.1† | 3.8 ± 2.4 | 1.6 ± 1.1† |
| FPG | 4.9 ± 0.7 | 5.1 ± 0.7 | 4.7 ± 0.7† | 5.1 ± 0.7 | 4.7 ± 0.7‡ | 5.1 ± 0.7 | 4.7 ± 0.7† |
| TCHOL | 4.4 ± 1.1 | 4.5 ± 1.2 | 4.2 ± 1.0§ | 4.6 ± 1.1 | 4.2 ± 1.0‡ | 4.6 ± 1.1 | 4.2 ± 1.0‡ |
| TG | 1.4 ± 0.6 | 1.9 ± 0.6 | 1.1 ± 0.5† | 1.7 ± 0.6 | 1.1 ± 0.5† | 1.7 ± 0.6 | 1.1 ± 0.5† |
| HDL | 1.2 ± 0.4 | 1.0 ± 0.4 | 1.3 ± 0.3† | 1.0 ± 0.3 | 1.3 ± 0.3† | 1.0 ± 0.3 | 1.3 ± 0.3† |
| LDL | 2.9 ± 1.1 | 3.2 ± 1.1 | 2.7 ± 1.0‡ | 3.2 ± 1.1 | 2.7 ± 1.0† | 3.2 ± 1.1 | 2.7 ± 1.0† |
| Insulin | 15.4 ± 7.7 | 17.9 ± 8.5 | 14.1 ± 6.9† | 18.0 ± 8.4 | 13.6 ± 6.6† | 18.0 ± 8.4 | 13.6 ± 6.6† |
| HOMA-IR | 3.4 ± 1.9 | 4.2 ± 2.0 | 2.9 ± 1.6† | 4.1 ± 2.0 | 2.8 ± 1.6† | 4.1 ± 2.0 | 2.8 ± 1.6† |
| HbA1c | 5.2 ± 0.8 | 5.6 ± 0.7 | 5.0 ± 0.8† | 5.7 ± 0.7 | 4.9 ± 0.7† | 5.7 ± 0.7 | 4.9 ± 0.7† |
Unless otherwise indicated, data are presented as Mean ± SD. Categorical data were compared between MetS and non-MetS groups using Chi-squared/Fisher exact tests. Continuous data were compared using independent t-test was used to compared continuous data. SBP; Systolic blood pressure, DBP; Diastolic blood pressure, WC; Waist Circumference, BMI; Body Mass Index, WHR; Waist-to-Hip ratio, WHtR; Waist-to Height ratio, BAI; Body Adiposity Index. VAI; Visceral Adiposity Index. FPG; Fasting Plasma Glucose, TCHOL; Total Cholesterol, TG; Triglycerides, HDL; High Density Lipoprotein, LDL; Low Density Lipoprotein, HOMA-IR, Homeostatic Model Assessment for Insulin Resistance, HbA1c; Glycated haemoglobin, ADA; American Diabetes Association, pHbA1c; Population-specific HbA1c; SD; Standard deviation.
†; Significant at p < 0.0001.
‡; Significant at p < 0.01.
§; Significant at p < 0.05.
data is presented as n (%).
Systolic blood pressure, BMI, WHtR, BAI, VAI, total cholesterol, triglyceride, LDL, and FPG were found to be positively related with HbA1c, whereas HDL showed a significant negative correlation with HbA1c in both the unadjusted and adjusted models (Table 2).
Table 2.
Association between HbA1c and cardiovascular risk factors among the entire study population.
| Variables |
Unadjusted model |
Adjusted model |
||
|---|---|---|---|---|
| Haemodynamic | r | p-value | r | p-value |
| SBP | 0.15 | 0.004 | 0.15 | 0.005 |
| DBP |
0.06 |
0.268 |
0.06 |
0.269 |
| Anthropometric | ||||
| BMI | 0.33 | <0.0001 | 0.31 | <0.0001 |
| WHR | 0.10 | 0.058 | 0.09 | 0.097 |
| WHtR | 0.32 | <0.0001 | 0.29 | <0.0001 |
| BAI | 0.34 | <0.0001 | 0.33 | <0.0001 |
| VAI |
0.30 |
<0.0001 |
0.29 |
<0.0001 |
| Biochemical | ||||
| TCHOL | 0.13 | 0.016 | 0.12 | 0.028 |
| TG | 0.24 | <0.0001 | 0.24 | <0.0001 |
| HDL | -0.27 | <0.0001 | -0.31 | <0.0001 |
| LDL | 0.17 | 0.001 | 0.17 | 0.001 |
| FPG | 0.37 | <0.0001 | 0.35 | <0.0001 |
Pearson's correlation was used to evaluate the association between HbA1c and cardiovascular risk factors. The adjusted model was evaluated using partial correlation after controlling for age and sex. r; correlation coefficient, SBP; Systolic blood pressure, DBP; Diastolic blood pressure, BMI; Body Mass Index, WHR; Waist-to-Hip ratio, WHtR; Waist-to Height ratio, BAI; Body Adiposity Index. VAI; Visceral Adiposity Index. FPG; Fasting Plasma Glucose, TCHOL; Total Cholesterol, TG; Triglycerides, HDL; High Density Lipoprotein, LDL; Low Density Lipoprotein.
The prevalence of MetS using the FPG, ADA HbA1c and pHbA1c criteria were 35.2%, 38.5% and 41.8%, respectively. The prevalence of MetS was higher among females compared to males across all criteria. The pHbA1c cut-off identified 6.6% and 3.3% more subjects with MetS when compared with FPG and the ADA HbA1c cut-offs, respectively. Similarly, the ADA HbA1c cut-off identified 3.3% more subjects with MetS compared with the FPG criterion in this study population (Table 3).
Table 3.
Prevalence of MetS among the study population.
| Number of components | FPG-based criterion |
ADA HbA1c-based criterion |
pHbA1c-based criterion |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Total | Male | Female | Total | Male | Female | Total | Male | Female | |
| 0 | 72 (9.9) | 56 (7.7) | 16 (2.2) | 72 (9.9) | 56 (7.7) | 16 (2.2) | 72 (9.9) | 56 (7.7) | 16 (2.2) |
| 1 | 192 (26.4) | 120 (16.5) | 72 (9.9) | 176 (24.2) | 120 (16.5) | 56 (7.7) | 168 (23.1) | 120 (16.5) | 48 (6.6) |
| 2 | 208 (28.6) | 56 (7.7) | 152 (20.9) | 200 (27.5) | 56 (7.7) | 144 (19.8) | 184 (25.3) | 56 (7.7) | 128 (17.6) |
| 3 | 136 (18.7) | 48 (6.6) | 88 (12.1) | 144 (19.8) | 40 (5.5) | 104 (14.3) | 144 (19.8) | 32 (4.4) | 112 (15.4) |
| 4 | 96 (13.2) | 40 (5.5) | 56 (7.7) | 104 (14.3) | 40 (5.5) | 64 (8.8) | 104 (14.3) | 48 (6.6) | 56 (7.7) |
| 5 | 24 (3.3) | 0 (0.0) | 24 (3.3) | 32 (4.4) | 8 (1.1) | 24 (3.3) | 56 (7.7) | 8 (1.1) | 48 (6.6) |
| ≥3 | 256 (35.2) | 88 (12.1) | 168 (23.1) | 280 (38.5) | 88 (12.1) | 192 (26.4) | 304 (41.8) | 88 (12.1) | 216 (29.7) |
FPG; Fasting plasma glucose, HbA1c; Glycated haemoglobin, ADA; American Diabetes Association, pHbA1c; Population-specific HbA1c.
Among the subjects with MetS by either FPG or ADA HbA1c (n = 304), an overlap of 232 (31.9%) was observed showing a substantial agreement (ĸ = 0.79) between the FPG and ADA HbA1c criterion for MetS (Figure 1A). On the other hand, of the 312 subjects diagnosed with MetS by either FPG or the pHbA1c cut-off, an overlap of 248 (34.1%) and an almost perfect concordance (ĸ = 0.82) was observed (Figure 1B).
Figure 1.
Validation and concordance evaluation between FPG- and HbA1c-based criteria.
Using the FPG-based criterion as the reference, both ADA HbA1c and pHbA1c cut-offs presented with excellent sensitivities and specificities in identifying subjects with MetS among the entire study population. The pHbA1c cut-off, however, had higher sensitivities compared to the ADA HbA1c criterion though the ADA HbA1c criterion presented with marginally higher specificities. Furthermore, the area under the curve (AUC) and reliability of the pHbA1c criterion was consistently higher than the ADA HbA1c criterion (Table 4).
Table 4.
Diagnostic performance of the ADA HbA1c and pHbA1c cut-off for diagnosing MetS.
| Parameter | ADA HbA1c criterion |
pHbA1c criterion |
||||
|---|---|---|---|---|---|---|
| Total | Males | Females | Total | Males | Females | |
| Sensitivity (95% CI) | 90.6 (84.1–94.7) | 100.0 (90.2–100.0) | 85.7 (76.5–91.7) | 96.9 (91.9–99.0) | 100.0 (90.2–100.0) | 95.2 (87.9–98.5) |
| Specificity (95% CI) | 89.8 (85.0–93.1) | 100.0 (96.0–100.0) | 80.0 (71.9–86.2) | 88.1 (83.3–91.7) | 100.0 (96.0–100.0) | 76.7 (68.3–83.3) |
| PPV (95% CI) | 82.9 (76.6–89.1) | 100.0 (100.0–100.0) | 75.0 (66.3–83.7) | 81.6 (75.4–87.7) | 100.0 (100.0–100.0) | 74.1 (65.8–82.3) |
| NPV (95% CI) | 94.6 (91.7–97.6) | 100.0 (100.0–100.0) | 88.9 (82.9–94.8) | 98.1 (96.3–99.9) | 100.0 (100.0–100.0) | 95.8 (91.8–99.8) |
| TP | 232 | 88 | 144 | 248 | 88 | 160 |
| TN | 424 | 232 | 192 | 416 | 232 | 184 |
| FP | 48 | 0 | 48 | 56 | 0 | 56 |
| FN | 24 | 0 | 24 | 8 | 0 | 8 |
| LR+ (95% CI) | 8.9 (6.1–13.1) | - | 4.3 (3.0–6.2) | 8.1 (5.8–11.6) | - | 4.1 (2.9–5.7) |
| LR- (95% CI) | 0.1 (0.1–0.2) | - | 0.2 (0.1–0.3) | 0.04 (0.01–0.09) | - | 0.06 (0.02–0.16) |
| Accuracy (%) | 90.1 | 100.0 | 82.4 | 91.2 | 100.0 | 84.3 |
| AUC (%) | 90.2 | 100.0 | 82.9 | 92.5 | 100.0 | 86.0 |
| Reliability (J index) | 80.5 | 100.0 | 65.7 | 85.0 | 100.0 | 71.9 |
The receiver operator curve (ROC) analysis was used to evaluate the diagnostic performance of the ADA HbA1c and pHbA1c using the FPG-based criterion as the reference. Reliability was expressed as the J index [(TP×TN) - (FP×FN)]/[(TP + FN) (TN + FP)]. HbA1c; Glycated haemoglobin, ADA; American Diabetes Association, pHbA1c; Population-specific HbA1c, PPV; Positive predictive value, NPV; Negative predictive value, LR+; Positive likelihood ratio, LR-; Negative likelihood ratio, CI; Confidence interval.
4. Discussion
In this study, the prevalence of MetS using the fasting hyperglycaemia and HbA1c based on the ADA criterion was 35.2% and 38.5%, respectively. Hence, the ADA HbA1c cut-off identified 3.3% more subjects with MetS compared with the FPG criterion as consistent with a follow-up population-based cohort study by de Vegt et al. [24] among older non-diabetics. de Vegt et al. reported that elevated HbA1c was associated with a higher age- and sex-adjusted risk of cardiovascular mortality and suggested that HbA1c is a better predictor of CVD mortality than fasting hyperglycaemia among non-diabetic subjects. This finding is also in harmony with a cross-sectional study by Annani-Akollor et al. [14] among non-diabetic population in Ghana who reported that the use of the ADA HbA1c criterion improves MetS diagnosis compared to FPG. Additionally, Osei et al. [11], in a study to evaluate the significance of HbA1c in MetS among African-American patients with type 2 diabetes, reported that when HbA1c was divided into tertiles: 4.7% (3.3%–4.8%), 5.4% (4.9%–5.6%), and 5.8% (5.7%–6.4%), subjects having HbA1c value within 5.7%–6.4% were more predisposed to the metabolic abnormalities, buttressing the significance of HbA1c as a marker for MetS diagnosis.
Furthermore, HbA1c is associated with increasing prevalence of cardiovascular disease and metabolic syndrome [3, 8, 12, 13]. Eeg-Olofsson et al. [25] found that higher HbA1c levels is linked to increased risks of coronary heart disease, CVD and total mortality; however, low HbA1c levels is associated with no increase in risk. We also found that HbA1c was significantly associated with several CVD risk factors including haemodynamic (systolic blood pressure), anthropometric (BMI, WHtR, BAI, and VAI), and biochemical indices (total cholesterol, triglyceride, LDL, HDL and FPG), reinforcing our previous deposition that HbA1c could serve as a potential marker for CVD [14]. Nonetheless, racial discrepancies necessitate the development of population-specific HbA1c cut-offs.
Using the ROC curve, a population-specific HbA1c (pHbA1c) cut-off of 5.4%, corresponding to the FPG value of 5.6 mmol/L (100 mg/dL) for dysglycaemia as defined by the American Diabetes Association [18], was developed. The prevalence of MetS using pHbA1c as the diagnostic criterion for dysglycaemia was 41.8%, showing that the pHbA1c cut-off identified 6.6% and 3.3% more subjects with MetS when compared with FPG and the ADA HbA1c cut-offs, respectively. Thus, although diagnosis of MetS based exclusively on fasting hyperglycaemia may involve some risk of overlooking subjects at high risk for CVD which may be remedied by the ADA-HbA1c criterion, the development of population-specific cut-offs for HbA1c could identify a greater number of such high risk subjects in that specific population. It is worth noting however that with less than three components of MetS, pHbA1c presents with different efficiency. As expected, when the presence of MetS components is lower than three, pHbA1c identifies less number of participants compared to FPG-based and ADA-based HbA1c criteria. This implies that caution should be taken when using pHbA1c in assessing CVD, especially in population where <3 MetS components is prevalent. Importantly, the prevalence of MetS was consistently higher among females compared to males across all criteria. This finding is in keeping with previous studies in Ghana [2, 14, 26] and elsewhere [27, 28, 29], and suggests a trend where females are more prone to CVD than males [30]. Therefore, identification of sociodemographic and environmental correlates of MetS, particularly those that affect females, will be critical to the preventive efforts against the increasing prevalence of MetS.
Furthermore, validation and concordance evaluation showed an almost perfect agreement (ĸ = 0.82) between pHbA1c and FPG criterion compared with the ADA HbA1c and FPG criterion which presented with a substantial agreement (ĸ = 0.79). This implies that, despite the identification of higher number of subjects with MetS by the pHbA1c compared to the ADA HbA1c criterion, there is a lower risk of misdiagnosis when pHbA1c used in this study population.
In assessing and comparing the diagnostic performance of the ADA HbA1c and pHbA1c, we used the FPG-based criterion as the reference. We found that both ADA HbA1c and pHbA1c cut-offs had an excellent sensitivity and specificity for identifying subjects with MetS in the entire study population. The pHbA1c cut-off however presented with higher sensitivities compared to the ADA HbA1c criterion, although the ADA HbA1c criterion presented with marginally higher specificities. Furthermore, the AUCs and reliabilities of the pHbA1c criterion were consistently higher than the ADA HbA1c criterion.
Taken together, it is evident that the use of HbA1c as the criterion for dysglycaemia in defining MetS allows for detection of high risk subjects who may be missed by the traditional FPG-based criterion. Nonetheless, a population specific cut-offs for HbA1c will detect a greater number of high risk subjects in that population with substantial accuracy and reliability compared to the general HbA1c cut-off.
Notwithstanding, this study is limited by its cross-sectional nature as the cause-effect relationship between HbA1c and MetS could not be established. Furthermore, the estimation of blood pressure was based on the average of two measurements at a single visit and biochemical profiling was single point measurement which might have overestimated the prevalence rates, as in many epidemiological studies. Additionally, though subjects with diabetes as well as those with newly diagnosed diabetes were excluded by applying the criteria of fasting hyperglycaemia and HbA1c at baseline, there is a possibility that a number of subjects with postprandial hyperglycaemia were included. However, HbA1c represents chronic exposure to basal and postprandial hyperglycaemia; thus, minimizes the effect of this limitation on the study findings. Although representative sample size was used in this study, the number of participants recruited for the study remain limited. We recommend a larger scale study to extensively explore the usefulness of the population-specific HbA1c cut-off in MetS diagnosis. Also, future studies should consider establishing a population specific cut-off for WC.
5. Conclusion
Screening for MetS by introduction of the ADA HbA1c criterion in addition to the traditional FPG criterion enhances the detection of more people with MetS. However, the use of population-specific cut-off for HbA1c could identify a greater number of high risk subjects in that specific population.
Declaration
Author contribution statement
M. Annani-Akollor and E. Laing: Conceived and designed the experiments.
O. Addai-Mensah, L. Fondjo and E. Adu: Performed the experiments; Analyzed and interpreted the data.
E. Owiredu: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.
Funding statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Competing interest statement
The authors declare no conflict of interest.
Additional information
No additional information is available for this paper.
Acknowledgements
The authors are grateful to the Staff of the St Francis Xavier Hospital and all who participated in the study.
Contributor Information
Max Efui Annani-Akollor, Email: akollor2017@gmail.com.
Eddie-Williams Owiredu, Email: eddiewilliams.owiredu@gmail.com.
References
- 1.Simmons R., Alberti K., Gale E. The metabolic syndrome: useful concept or clinical tool? Report of a WHO Expert Consultation. Diabetologia. 2010;53(4):600–605. doi: 10.1007/s00125-009-1620-4. [DOI] [PubMed] [Google Scholar]
- 2.Agyemang-Yeboah F., Eghan B.A.J., Annani-Akollor M.E. Evaluation of metabolic syndrome and its associated risk factors in type 2 diabetes: a descriptive cross-sectional study at the komfo anokye teaching hospital, Kumasi, Ghana. BioMed Res. Int. 2019;2019:8. doi: 10.1155/2019/4562904. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Siu P., Yuen Q. Supplementary use of HbA1c as hyperglycemic criterion to detect metabolic syndrome. Diabetol. Metab. Syndrome. 2014;6 doi: 10.1186/1758-5996-6-119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Magliano D., Shaw J., Zimmet P. How to best define the metabolic syndrome. Ann. Med. 2006;38:34–41. doi: 10.1080/07853890500300311. [DOI] [PubMed] [Google Scholar]
- 5.Alberti K., Eckel R., Grundy S. Harmonizing the metabolic syndrome: a joint interim statement of the international diabetes federation task force on epidemiology and prevention; national heart, lung, and blood institute; American heart association; world heart federation; international atherosclerosis society; and international association for the study of obesity. Circulation. 2009;120(16):1640–1645. doi: 10.1161/CIRCULATIONAHA.109.192644. [DOI] [PubMed] [Google Scholar]
- 6.ADA . American Diabetes Association; 2010. Standards of Medical Care in Diabetes-2010.http://care.diabetesjournals.org/content/33/Supplement_1/S11 [26/04/2019]; S11-S61]. Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Inoue K., Matsumoto M., Akimoto K. Fasting plasma glucose and HbA1c as risk factors for type 2 diabetes. Diabet. Med. 2008;25:1157–1163. doi: 10.1111/j.1464-5491.2008.02572.x. [DOI] [PubMed] [Google Scholar]
- 8.Sun X., Du T., Huo R. Impact of HbA1c criterion on the definition of glycemic component of the metabolic syndrome: the China health and nutrition survey 2009. BMC Publ. Health. 2013;13 doi: 10.1186/1471-2458-13-1045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Park S., Yoon J., Won K. Usefulness of glycated hemoglobin as diagnostic criteria for metabolic syndrome. J. Kor. Med. Sci. 2012;27(9):1057–1061. doi: 10.3346/jkms.2012.27.9.1057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sung K., Rhee E. Glycated haemoglobin as a predictor for metabolic syndrome in non-diabetic Korean adults. Diabet. Med. 2007;24(8):848–854. doi: 10.1111/j.1464-5491.2007.02146.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Osei K., Rhinesmith S., Gaillard T. Is glycosylated hemoglobin A1c a surrogate for metabolic syndrome in nondiabetic, first-degree relatives of African-American patients with type 2 diabetes? J. Clin. Endocrinol. Metabol. 2003;88(10):4596–4601. doi: 10.1210/jc.2003-030686. [DOI] [PubMed] [Google Scholar]
- 12.Selvin E., Steffes M., Zhu H. Glycated hemoglobin, diabetes, and cardiovascular risk in nondiabetic adults. N. Engl. J. Med. 2010;362:800–811. doi: 10.1056/NEJMoa0908359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stratton I., Adler A., Neil A. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UK-PDS 35): prospective observational study. BMJ. 2000;321:405–412. doi: 10.1136/bmj.321.7258.405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Annani-Akollor M.E., Laing E.F., Osei H. Prevalence of metabolic syndrome and the comparison of fasting plasma glucose and HbA1c as the glycemic criterion for MetS definition in non-diabetic population in Ghana. Diabetol. Metab. Syndrome. 2019;11(1):26. doi: 10.1186/s13098-019-0423-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Herman W., Ma Y., Uwaifo G. Diabetes Prevention Program Research Group. Racial and ethnic differences in hemoglobin A1c among patients with impaired glucose tolerance in the Diabetes Prevention Program. Diabetes Care. 2007;42(1) doi: 10.2337/dc06-2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ziemer D., Kolm P., Weintraub W. Glucose-independent, black-white differences in hemoglobin A1c levels: a cross-sectional analysis of 2 studies. Ann. Intern. Med. 2010;152:770–777. doi: 10.7326/0003-4819-152-12-201006150-00004. [DOI] [PubMed] [Google Scholar]
- 17.Service G.S. 2014. Population and Housing Census: District Analytical Report. [Google Scholar]
- 18.Association A.D. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2014:S81–90. doi: 10.2337/dc14-S081. [DOI] [PubMed] [Google Scholar]
- 19.Owiredu E.-W., Dontoh E., Essuman S.E. Demographic and lifestyle predictors of prehypertension: a cross-sectional study among apparently healthy adults in Kumasi, Ghana. BioMed Res. Int. 2019;2019 doi: 10.1155/2019/1764079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Freedman D., Thornton J., Pi-Sunyer F. The body adiposity index (hip circumference÷ height1. 5) is not a more accurate measure of adiposity than is BMI, waist circumference, or hip circumference. Obesity. 2012;20(12) doi: 10.1038/oby.2012.81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fondjo L., Sakyi S., Owiredu W. Evaluating vitamin D status in pre-and postmenopausal type 2 diabetics and its association with glucose homeostasis. BioMed Res. Int. 2018;2018 doi: 10.1155/2018/9369282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Friedewald W.T., Levy R.I., Fredrickson D.S. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin. Chem. 1972;18(6):499–502. [PubMed] [Google Scholar]
- 23.NSGP . 2019. List of NGSP Certified Methods: National Glycohemoglobin Standardization Program.http://www.ngsp.org/docs/methods.pdf [26/08/2019]. Available from: [Google Scholar]
- 24.de Vegt F., Dekker J., Ruhé H. Hyperglycaemia is associated with all-cause and cardiovascular mortality in the Hoorn population: the Hoorn Study. Diabetologia. 1999;42:926–931. doi: 10.1007/s001250051249. [DOI] [PubMed] [Google Scholar]
- 25.Eeg-Olofsson K., Cederholm J., Nilsson P. New aspects of HbA1c as a risk factor for cardiovascular diseases in type 2 diabetes: an observational study from the Swedish National Diabetes Register (NDR) J. Intern. Med. 2010;268(5):471–482. doi: 10.1111/j.1365-2796.2010.02265.x. [DOI] [PubMed] [Google Scholar]
- 26.Ofori-Asenso R., Agyeman A.A., Laar A. Metabolic syndrome in apparently “healthy” Ghanaian adults: a systematic review and meta-analysis. Int. J. Chronic Dis. 2017;2017 doi: 10.1155/2017/2562374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Amirkalali B., Fakhrzadeh H., Sharifi F. Prevalence of metabolic syndrome and its components in the Iranian adult population: a systematic review and meta-analysis. Iran. Red Crescent Med. J. 2015;17(12) doi: 10.5812/ircmj.24723. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Wong-McClure R.A., Gregg E.W., Barceló A. Prevalence of metabolic syndrome in Central America: a cross-sectional population-based study. Rev. Panam. Salud Públic. 2015;38:202–208. [PubMed] [Google Scholar]
- 29.Aryal N., Wasti S.P. The prevalence of metabolic syndrome in South Asia: a systematic review. Int. J. Diabetes Dev. Ctries. 2016;36(3):255–262. [Google Scholar]
- 30.Woodward M. Cardiovascular disease and the female disadvantage. Int. J. Environ. Res. Publ. Health. 2019;16(7):1165. doi: 10.3390/ijerph16071165. [DOI] [PMC free article] [PubMed] [Google Scholar]