Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2019 Dec 31;14(12):e0227065. doi: 10.1371/journal.pone.0227065

Glycated albumin as a diagnostic tool in diabetes: An alternative or an additional test?

Fernando Chimela Chume 1,2, Mayana Hernandez Kieling 1, Priscila Aparecida Correa Freitas 1,3, Gabriela Cavagnolli 4, Joíza Lins Camargo 1,5,*
Editor: Petter Bjornstad6
PMCID: PMC6938306  PMID: 31891628

Abstract

Introduction

Studies have revealed that glycated albumin (GA) is a useful alternative to HbA1c under conditions wherein the latter does not reflect glycaemic status accurately. Until now, there are few studies with non-Asians subjects that report on the validity of GA test in diagnosis of type 2 diabetes mellitus (DM). Thus, the aim of this study was to assess the clinical utility of GA in diagnosis of DM.

Materials and methods

This diagnostic test accuracy study was performed in 242 Brazilian individuals referred for OGTT in a tertiary university hospital. ROC curves were used to access the performance of GA and HbA1c in the diagnosis of DM by oral glucose tolerance test (OGTT).

Results

OGTT, HbA1c and GA were performed in all 242 participants (40.5% male, age 54.4 ± 13.0 years [mean ± SD], body mass index 28.9 ± 6.3 kg/m2). DM by OGTT was detected in 31.8% of individuals. The equilibrium threshold value of GA ≥14.8% showed sensitivity of 64.9% and specificity of 65.5% for the diagnosis of DM. The AUC for GA [0.703 (95% CI 0.631–0.775)] was lower than for HbA1c [0.802 (95% CI 0.740–0.864)], p = 0.028. A GA value of 16.8% had similar accuracy for detecting DM as defined by HbA1c ≥6.5% (48 mmol/mol) with sensitivity of 31.2% and specificity of 93.3% for both tests. However, GA detects different subjects from those detected by HbA1c and OGTT.

Conclusions

GA detected different individuals with DM from those detected by HbA1c, though it showed overall diagnostic accuracy similar to HbA1c in the diagnosis of DM. Therefore, GA should be used as an additional test rather than an alternative to HbA1c or OGTT and its use as the sole DM diagnostic test should be interpreted with caution.

Introduction

Despite being largely preventable, the worldwide increase in type 2 diabetes mellitus (DM) is becoming a major health concern. It has been estimated that globally as many as 212.4 million people or half (50.0%) of all people aged 20–79 years old with DM are unaware of their disease [1]. Any improvement in the identification of hyperglycaemia will be of significant impact because delays in diagnosis and treatment may increase the incidence of cardiovascular outcomes and all-cause mortality related to this disease [2]. At present DM may be diagnosed based on plasma glucose criteria, either by fasting plasma glucose (FPG) or 2-h plasma glucose (2hPG) after a 75-g oral GTT (OGTT) or HbA1C criterion, all tests are equally appropriate [3].

Although OGTT measurement is still a standard recommendation for DM diagnosis, this method is onerous, time-consuming and requires two blood samples. In contrast, the sole use of FPG measurement in DM screening will fail to diagnose those subjects presenting only with 2hPG ≥11.1 mmol/L (≥200 mg/dL). Moreover, HbA1c, which is considered the reference standard for monitoring long-term glycaemic control in subjects with DM, is also a primary diagnostic tool for DM. HbA1c has several advantages compared with the FPG and OGTT, including greater convenience (fasting is not required), higher pre-analytical stability, and less day-to-day variations during stress and illness [3]. However, HbA1c is not suitable for conditions with altered blood red cell turnover, such as some haemoglobinopathies, thalassemia, chronic kidney disease and haemolytic anaemia [4]. Furthermore, the presence of haemoglobin variants (e.g. HbS trait, HbC trait), elevated foetal haemoglobin (HbF) and chemically modified derivatives of haemoglobin (e.g. carbamylated Hb in patients with renal failure) can interfere either positively or negatively with the HbA1c measurement and consequently adversely affect the interpretation of HbA1c results [4–6]. Therefore, it is important to consider alternative procedures in the diagnosis of DM.

Glycated albumin (GA) is a ketamine produced by binding of albumin and glucose by a nonenzymatic glycation reaction [7]. It reflects short-term mean glycaemic values (2–3 weeks) due to the shorter half-life of serum albumin, rather than 2–3 months mean glycaemic values observed in HbA1c [8]. Similar to HbA1c, GA correlates with diabetic complications such as retinopathy, chronic kidney disease, peripheral neuropathy, cardiovascular disease, and even death [9–11]. Additionally, GA is haemoglobin/erythrocyte independent, consequently, measurement of GA is not influenced by anaemia or other conditions considered potential factors that can affect the interpretation of HbA1c results [7, 8]. Besides, evidences suggest that GA is a better glycaemic indicator than HbA1c in diabetic subjects on haemodialysis [7].

Although data about GA performance in diagnosis and screening of DM have been available in Asian populations [12–15], limited data exists in other populations [16–18]. We hypothesized that GA may be used in the diagnosis of DM and in clinical conditions where the HbA1c test does not accurately reflect blood glucose concentrations GA may be an alternative marker. Therefore, the current study was designed to assess the clinical utility of GA in screening and diagnosis of DM in Brazilian individuals.

Material and methods

Study design

We conducted a cross-sectional study of diagnostic accuracy and reported corresponding results according to Standard for Reporting Diagnostic Accuracy (STARD) statement [19]. The study flow diagram is shown in Fig 1.

Fig 1. Study flow diagram.

Fig 1

Open in a new tab

OGTT, oral glucose tolerance test.

Participant selection

Outpatients older than 18 years referred to the Hospital de Clinicas de Porto Alegre (HCPA) between August 2008 and August 2017 to perform OGTT were consecutively invited to participate in this study. Subjects who accepted the invitation completed a questionnaire, underwent a physical examination and received blood tests. Serum sample was stored at -80°C for GA measurement. The stability of the GA assay in long-term stored specimens has already been evaluated [20].

Study exclusion criteria were: albumin levels <3.0 g/dl; subjects with established diagnosis of DM or who were receiving anti-diabetic medication; pregnant women; presence of anaemia, hemoglobinopathy, recent transfusion, rheumatic disorder, hepatic cirrhosis, nephrotic syndrome, chronic kidney disease, untreated thyroid dysfunction, and/or Cushing syndrome, since these disorders are known to influence values of GA and/or HbA1c.

Each participant provided a written informed consent. This study was reviewed and approved by the Research Ethics Committee of the Hospital de Clinicas de Porto Alegre (GPPG 080321 and 160448).

Glycaemic status was defined according to American Diabetes Association criteria [3]. DM was defined by: (a) FPG ≥7.0 mmol/L (≥126 mg/dL) and/or (b) 2hPG ≥11.1 mmol/L (≥200 mg/dL) during an OGTT and/or (c) HbA1c ≥6.5% (≥48 mmol/mol) for descriptive purposes. With the intention of a diagnostic accuracy study, OGTT [FPG ≥7.0 mmol/L (≥126 mg/dL) and/or 2hPG ≥11.1 mmol/L (≥200 mg/dL)] was used as reference standard test; HbA1c and GA were used as index tests.

Laboratorial methods

All subjects underwent a standard 75g OGTT after an overnight fast of at least 8 hours. Blood samples for glucose determination were collected by venepuncture into tubes containing sodium fluoride at fasting and at 2-hour after 75g glucose oral intake. Plasma glucose concentrations were measured by colorimetric enzymatic method in the biochemistry automated analyser Cobas® c702 (Roche Diagnostics, Germany).

HbA1c were measured in K2EDTA-anticoagulated whole blood by high performance liquid chromatography (HPLC) using VARIANT II™ System (BioRad Laboratories, Hercules, CA, USA). This HbA1c assay is certified by the National Glycohemoglobin Standardization Programme, aligned to the DCCT assay and it is also standardized by International Federation of Clinical Chemistry [21]. Analytical inter-assay coefficient of variation in our lab was <3.0% [22].

Fasting serum samples were stored at -80°C until it was used for measurement of GA. GA were determined by an enzymatic method (GlycoGap®, Diazyme Laboratories, Poway, CA) in the automated analyser Cobas® c702 (Roche Diagnostics, Germany). This method was previously validated in our lab and the intra-assay repeatability was 3.5% [22]. Total albumin was measured with bromocresol green colorimetric method. GlycoGap® GA assay quantifies the total of glycated serum proteins (GSP, μmol/L), which are converted to percent of GA by the following conversion equation: GA (%) = {[GSP (μmol/L) x 0.182 + 1.97]/total albumin (g/dL)} + 2.9 [22]. Previous results showed that the Diazyme method correlates well with the Lucica GA-L assay, a specific GA assay used in Asian and Europe, with a small bias, supporting the equivalence between GSP and GA [23].

Serum creatinine was measured by Jaffé colorimetric method, triglycerides, total cholesterol by enzymatic assays, both using Cobas® c702 analyser (Roche Diagnostic, Mannheim, Germany). Haemoglobin and haematocrit were assayed by routine techniques.

Body mass index (BMI) was calculated by dividing body weight (kg) by the square of body height (m). The waist circumference was measured midway between the lowest rib and the iliac crest in a standing position. Systolic blood pressure and diastolic blood pressure were measured on the right arm, in the sitting position, with an automated sphygmomanometer (HEM-780, Omron Healthcare, Kyoto, Japan) after at least 5-minute rest. Smoking and drinking habit (current, past or never), and ethnicity was determined by self-report.

Statistical analysis

Unless otherwise stated, data are presented as mean ± standard deviation (SD) for continuous variables and as percentages for categorical variables. Group comparisons were analysed by Student's t-test, Fisher’s exact test and the Chi-square test as appropriate. For clinical and laboratory descriptive purposes, individuals with and without DM were divided using ADA OGTT criteria. Relationships among variables were explored using Spearman's correlation coefficients and regression models. Receiver operating characteristic (ROC) curves were used to access the performance of GA and HbA1c in the diagnosis of DM by OGTT as the reference test. Also ROC curve was created to evaluate the performance of GA using OGTT and/or HbA1c as DM diagnostic reference test. Areas under the curves (AUC) of GA and HbA1c were compared by DeLong’s test. The optimal cut-off for serum GA was derived from the ROC curve with the shortest distance to sensitivity and specificity with maximum value of the Youden index. Combining sensitivity and specificity, we calculated likelihood ratios (LR) for different cut-off points. The LR+ was calculated by dividing the sensitivity of the test by 1−specificity (Sensitivity/1 − specificity), while LR− of a test can be calculated by dividing 1 − sensitivity by specificity (1 − Sensitivity/Specificity) [24]. The first cut-off point of GA in the ROC curve with specificity over 90%DM was chosen as the criterion for diagnosis of DM. To demonstrate the clinical applicability of the test, we combined likelihood ratios with pre-test probability of the disease to estimate post-test probability using Fagan’s nomogram [25]. Venn diagram was used to present the number of individuals with DM diagnosed by each test and overlaps.

The IBM SPSS software for Windows, version 20.0 (Statistical Package for Social Sciences—Professional Statistics, IBM Corp, Armonk, USA) and MedCalc, version 19.1 (MedCalc software, Ostend, Belgium) were used for data analysis. P values 0.05 were considered significant.

Results

A total of 242 participants were enrolled in the present study, of those 144 (69.5%) were women. One hundred ninety-five (80.2%) subjects self-reported European ancestry (mainly of Portuguese, German, Italian and Spanish descent). Participants presented mean age of 54.4 years (± 13.0) and values for GA, FPG, 2hPG, and HbA1c of 14.9 ± 2.2%, 6.2 ± 1.2 mmol/l (112 ± 21 mg/dL), 9.2 ± 4.1 mmol/l (165 ± 73 mg/dL), 5.79 ± 0.79% (40 ± 8.6 mmol/mol), respectively. GA values were not normally distributed [median 14.5% (GA minimum 8.2%, GA maximum 26.9%)]. Based on glucose criteria for the OGTT, DM was detected in 31.8% (77/242). HbA1c ≥6.5% (48 mmol/mol) identified 33 individuals with DM (13.6%), of those subjects 24 were also diagnosed with DM by OGTT. Based on both tests, a total of 86 participants had diagnosis of DM (35.5%).

The clinical and laboratory characteristics of all individuals are shown in Table 1. Individuals with DM diagnosed by ADA OGTT criteria, compared to the group without DM, were older and had higher values of GA, FPG, 2hPG and HbA1c. There were no significant differences in BMI and HDL. On the other hand, subjects with DM had higher values of total cholesterol, triglyceride and LDL. Additionally, the ethnic difference between groups was not accessed due to small sample size.

Table 1. Clinical and laboratory characteristics of the study participants divided by subjects with and without DM using ADA OGTT criteria.

Total Without DM DM P
n 242 165 77
Age (years) 53.4 ± 13.4 56.8 ± 11.9 58.5 ± 11.5 0.056
Sex (male/female) 98/144 75/90 23/54 0.025
Ancestry 0.147
    European, n (%) 195 (80.5) 135 (80.5) 62 (80.5)
    African, n (%) 28 (11.6) 16 (9.8) 12 (15.6)
    Other ancestry, n (%) 19 (7.9) 16 (9.8) 3 (3.9)
BMI (kg/m2) 28.9 ± 6.3 28.6 ± 6.4 29.6 ±+ 6.3 0.271
WC (cm) 99.26 ± 13.38 98.8 ± 14.0 100.3 ± 11.9 0.429
SBP (mm Hg) 131.69 ± 16.74 130 ± 16 139 ± 17 0.016
DBP (mm Hg) 80.38 ± 12.72 79 ± 12 85 ± 14 0.051
Family history of DM, n (%) 122 (51.3) 73 (45.1) 49 (64.5) 0.005
Hypertension, n (%) 146 (60.8) 91 (55.8) 55 (71.4) 0.021
Hypertension treatment, n (%) 139 (57.4) 87 (53.4) 52 (67.5) 0.038
Total cholesterol (mmol/l) 4.9 ± 1.1 4.8 ± 1.0 5.1 ± 1.2 0.017
Triglycerides (mmol/l) 1.9 ± 1.2 1.7 ± 1.0 2.2 ± 1.4 0.003
HDL (mmol/l) 1.2 ± 0.4 1.2 ± 0.9 1.2 ± 0.3 0.934
LDL (mmol/l) 3.7 ± 1.1 3.5 ± 1.1 3.9 ± 1.2 0.021
Serum Creatinine (μmol/l) 73.2 ± 19.4 73.4 ± 17.7 76.0 ± 17.7 0.334
Serum albumin (g/l) 44.0 ± 4.0 44.0 ± 4.0 44.0 ± 4.0 0.773
Haemoglobin (g/l) 13.8 ± 1.4 13.8 ± 1.4 13.8 ± 1.4 0.942
FPG (mmol/l) 6.2 ± 1.2 5.8 ± 0.6 7.2 ± 1.4 < 0.001
2hPG (mmol/l) 9.2 ± 4.1 7.2 ± 2.0 13.4 ± 4.1 < 0.001
HbA1c (%)
     (mmol/mol)		 5.8 ± 0.8
40.0 ± 8.6		 5.5 ± 0.6
38.0 ± 6.6		 6.3 ± 0.9
43.0 ± 12.0		 < 0.001
GA (%) 14.91 ± 2.2 14.4 ± 1.8 15.9 ± 2.6 < 0.001
Open in a new tab

Mean ± SD and for continuous variables. ADA, American Diabetes Association; BMI, body mass index; WC, waist circumference (cm); SBP, systolic blood pressure; DBP diastolic blood pressure; HDL, serum high density lipoprotein cholesterol; LDL, serum low density lipoprotein cholesterol; FPG, fasting plasma glucose; 2hPG, plasma glucose 2 h after oral glucose; HbA1c, glycated haemoglobin; GA, glycated albumin; OGTT, oral glucose tolerance test; DM, type 2 diabetes mellitus.

The correlations between GA and factors potentially associated with the measurement of serum GA in all participants are presented in S1 Table. GA and age were positively correlated (r = 0.294, p <0.001). GA concentrations increased by 0.44% per decade (GA = 12.503 + 0.044 x age). GA was inversely correlated with triglyceride (r = - 0.197, p < 0.001). For every 10 mg/dL increase in serum triglyceride, GA decreased by 0.04% (GA = 15.623–0.004 x [triglyceride]). However, in those participants recently diagnosed with DM, these correlations were not significant [age (r = 0.101, p = 0.380) and triglyceride (r = -0.020, p = 0.429)]. Whereas HbA1c was positively correlated with BMI, waist circumference, total cholesterol, triglyceride and low-density lipoprotein cholesterol (LDL); GA was negatively correlated with BMI, WC, total cholesterol, triglyceride and LDL, though most of these correlations were not significant (S1 Table).

ROC curves comparing the performance of GA and HbA1c in the diagnosis of DM by OGTT as the reference test are presented in Fig 2. The AUC for GA in the diagnosis of DM by the OGTT was lower than for HbA1c (p = 0.028), with values of 0.703 (95% CI 0.631–0.775) and 0.802 (95% CI 0.740–0.864), for GA and HbA1c, respectively. The equilibrium cut-off value for GA was 14.8%; sensitivity and specificity for GA in this cut point were 64.9% and 65.5%, respectively. GA ≥14.8% yielded LR+ and LR- of 1.88 and 0.54, respectively (Table 2). Inferring in our population a pre-test probability of 9.0% for DM [1] and considering GA ≥14.8% as DM diagnostic criterion, after a positive test (GA ≥14.8%) the post-test probability for DM would increase to 16%, while a negative test (GA <14.8%) would decrease the post-test probability for DM to 5%. In this study, using the equilibrium point of GA as the criterion for diagnosis of DM (GA <14.8%), 50 subjects with DM would have a true positive diagnosis; however, 27 subjects with DM and 57 subjects without DM would be falsely diagnosed.

Fig 2. Receiver operating characteristic (ROC) curves to access the performance of GA, and HbA1c in the diagnosis of DM by OGTT.

Fig 2

Open in a new tab

The AUC value for GA was 0.703 (SE: 0.037, 95% CI: 0.631–0.775, p <0.001) and for HbA1c was 0.802 (SE: 0.032, 95% CI: 0.740–0.864, P <0.001); (n = 242). HbA1c, glycated haemoglobin; GA, glycated albumin; OGTT, oral glucose tolerance test; AUC, area under the ROC curve; SE, standard error; CI, confidence interval.

Table 2. Performance of different cut-offs of GA and HbA1c to diagnose DM by OGTT. (n = 242).

Threshold Sensitivity (%) Specificity (%) LR+ LR-
GA (%) 13.0 93.5 15.2 1.10 0.43
14.0 84.4 44.2 1.51 0.35
14.8 64.9 65.5 1.88 0.54
15.0 62.3 69.7 2.06 0.54
15.5 48.1 77.6 2.14 0.67
16.0 42.9 84.8 2.83 0.67
16.6 36.4 90.3 3.75 1.41
16.8 31.2 93.3 4.68 0.74
17.0 29.9 93.9 4.93 0.74
17.5 20.8 96.4 5.71 0.82
HbA1c (%) [mmol/mol] 5.5 (37.0) 87.0 58.2 2.08 0.22
5.7 (39.0) 81.8 68.5 2.59 0.27
5.8 (40.0) 76.6 72.7 2.81 0.32
6.00 (42.) 61.0 82.4 3.47 0.47
6.5 (48.0) 31.2 93.3 4.68 0.74
6.8 (51.0) 22.1 98.2 12.14 0.79
Open in a new tab

DM, Type 2 diabetes mellitus; HbA1c, glycated haemoglobin; GA, glycated albumin; LR, likelihood ratio; OGTT, oral glucose tolerance test.

We also evaluated the performance of GA using OGTT and/or HbA1c as DM diagnostic reference test. There was no relevant change in the AUC of GA compared to the one obtained when OGTT solely is considered as diagnostic reference test [0.708 (95% CI 0.639–0.776)] versus [0.703 (95% CI 0.631–0.775)], respectively. The optimal cut-off value for serum GA, when OGTT and/or HbA1c are reference was 14.7% (sensitivity 64.0% and specificity 64.1%) versus 14.8% (sensitivity 64.9% and specificity 65.5%) when OGTT alone is reference.

GA value of 16.6% was the first point in the ROC curve presenting specificity higher than 90%. However, the cut-off of 16.8% had similar performance for detecting DM as defined by HbA1c ≥6.5% (≥48 mmol/mol) with sensitivity of 31.2% and specificity of 93.3% and presented LR+ of 4.68 and LR- of 0.74 (Table 2). Therefore, considering a pre-test probability of 9.0% [1], after a positive test (GA ≥16.8%), the post-test probability for DM would increase to 32%, while a negative test (GA <16.8%) would decrease the post-test probability for DM to 7% (Fig 3). In our study group, considering this point (GA ≥16.8%), the number of truly negative subjects would increase to 154, the number of false positive results would be reduced to 11, however also would reduce the truly positive results to 24. However, it should be noted that GA, HbA1c and OGTT do not necessarily detect DM in the same individuals (Fig 4). Among 77 subjects diagnosed with DM by OGTT, only 11 were identified as DM subjects by both GA ≥16.8% and HbA1c ≥6.5%. Thirteen of the remaining 66 subjects were identified only by GA ≥16.8% and another 13/66 were identified only by HbA1c ≥6.5%. A GA ≥16.8% would also identify 14 subjects that would not be detected neither by HbA1c nor OGTT.

Fig 3. Fagan´s Nomogram for GA ≥16.8% cut-off inferring a subject’s pre- and post-test probability of having DM.

Fig 3

Open in a new tab

Pre-test probability according to International DM Federation–IDF data [1]; (n = 242).

Fig 4. Number of individuals diagnosed with DM by each test (OGTT, HbA1c, GA) and overlaps.

Fig 4

Open in a new tab

Glycaemic status for HbA1c and OGTT according to ADA criteria [3], and GA >16.8%; (n = 242). HbA1c, glycated haemoglobin; GA, glycated albumin; OGTT, oral glucose tolerance test; ǂ number of individuals with DM diagnosed by one test criteria without overlapping of other test criteria; * number of individuals with DM diagnosed by all tests (GA ≥16.8%, HbA1c and OGTT) criteria overlapped; £ number of individuals with DM diagnosed by both HbA1c and OGTT criteria overlapped; # number of individuals with DM diagnosed by both GA ≥16.8% and OGTT criteria overlapped; ¥ number of individuals with DM diagnosed by both HbA1c and GA ≥16.8% criteria overlapped.

Discussion

This study evaluated the performance of GA test in the diagnosis of DM in Brazilians subjects. According to ROC analysis GA ≥14.8% was the equilibrium cut-off. The LR+ and LR- indicate that GA ≥14.8% is more likely to occur in people with the disease than in people without the disease. However, this cut-off did not show enough sensitivity to correctly define the proportion of people with DM, nor had high enough specificity to correctly define the proportion of people without the disease. On the other hand, GA value of 16.8% presented lower sensitivity but specificity over to 90%, with performance similar to HbA1c ≥6.5% (>48mmol/mol) for detecting DM, and therefore it may be an adequate cut-off point for detecting DM in individuals with high-risk of developing the disease.

Although GA is not currently recommended for the screening or diagnosis of DM, there are several studies which advocate GA as a screening test for undiagnosed DM, still some studies have recommended the test as a secondary screening tool [12–18]. The cut-off of GA ≥16.8% suggested in this study is similar to those proposed by other previous studies [12, 15].

Though GA data have been accumulating in Asian population [12–15], limited data are available in other regions. There are few studies with non-Asians that report on the validity of GA test in screening and diagnosis of DM [16–18]. One study [16] evaluated the performance of GA in obese youth mainly Hispanic North Americans and suggested GA ≥12% as the cut point when using 2hPG as a reference test and GA ≥14% when HbA1c is a reference test. Although no details of sensitivity and specificity were reported, GA was good predictor of DM with AUC >0.90 in both scenarios. Another study evaluated the performance of GA in Caucasian subjects from Italy [17] using HbA1c only as a reference test and reported that the optimal threshold value (GA >14.0%) had sensitivity of 72.2% and specificity of 71.8% for diagnosis of DM. Lately, a study which examined African subjects [18] using OGTT as reference test referred the optimal cut-off value for GA as 14.9%, similar to the optimal threshold of GA of 14.8% reported in this present study. However, it should be noticed that, in this African study the suggested point in the ROC curve is not the point with the best equilibrium between sensitivity and specificity, as sensitivity and specificity for this GA threshold were 64.8% and 93.5%, respectively.

Nevertheless, in comparison with these studies our data showed different sensitivity and specificity for the same cut-off values. Some factors may be related to these differences. Firstly, ethnic differences are an important reason, since GA levels may vary with race/ethnicity independently of glycaemia [26]. Secondly, the inclusion criteria may have an effect in GA performance, our study included Brazilian subjects who had known risk factors for DM presenting in a tertiary hospital, while other studies [12, 13] included subjects from general population.

In this study, only 31.2% (24/86) of diabetic individuals overlapped in the diagnosis of DM by both HbA1c ≥6.5% (≥48mmol/mol) and OGTT criteria. This confirms that there is a large gap between HbA1c and OGTT criteria for diagnosing diabetes [27]. Our data showed that GA ≥16.8% shows performance similar to HbA1c and detected also one third of diabetic individuals detected by OGTT. Nevertheless, HbA1c and GA do not necessarily detect the same people.

However, using GA would have advantage over HbA1c, because GA can be measured accurately in plasma or serum samples [12]. Consequently, GA could be analysed together with common biological markers, including glucose, cholesterol, triglycerides and creatinine, without requiring a blood collection in a separate tube, by contrast, HbA1c can only be measured in whole blood samples. One should be aware that as for HbA1c, it is important to recognize that GA is an indirect measure of blood glucose levels and other factors may impact glycation of albumin independently of glycemia status. Therefore, in conditions with altered albumin metabolism as liver cirrhosis, thyroid dysfunction, nephrotic syndrome with massive proteinuria, or inflammatory conditions, the use of GA may be misleading [7, 8]. Other interfering situations on GA levels already described are age and obesity [28, 29].

In the present study, GA and HbA1c were found to be associated with age. A similar association was observed in other studies [12; 13]. However, in a previous analysis from our group, when participants of another study [22] were grouped according to quartiles of age or decade of life there was no difference in GA levels among groups (data not published).

GA was inversely correlated with triglycerides. GA negatively correlated with BMI, WC and LDL, although this association failed to reach statistical significance. Different of GA, HbA1c is more sensitive to BMI and WC, this may also explain why GA identifies a substantial number of non-obese individuals with prediabetes not detected by HbA1c [30]. Nonetheless, the overall similarity of major DM risk factor associations for elevated HbA1c and GA is reassuring and suggests that, in general, elevations in GA are largely being driven by the same pathophysiological processes that act to raise blood glucose concentrations over time [9–11, 31].

This study has several strengths. It is the first to evaluate the diagnostic utility of GA for DM in Brazilian population. At enrolment, we excluded pregnant women, as well as individuals with anaemia, renal failure, rheumatic disorder, hepatic cirrhosis, or thyroid disease, as these conditions may interfere with the interpretation of HbA1c and GA [7, 8]. Therefore, we were able to evaluate the diagnostic efficacy of GA and HbA1c in the absence of confounding factors. Moreover, the majority of population in this study has European ancestry which allows the applicability of our results in similar populations.

There were also some limitations to the present study that must be considered when interpreting the results. First, the study sample size is small; however, the sample size was calculated a priori and it is sufficient to obtain an AUC of 0.70 with a power of 80% and an estimated alfa error of 5%. Secondly, it comprises mainly individuals at risk of DM with a high pre-test probability attending a tertiary hospital rather than a general population. Third, OGTT, HbA1c and GA were performed only once, even when the results were positive.

Conclusions

In this study we were able to demonstrate that GA presents overall diagnostic accuracy similar to HbA1c in the diagnosis of DM. Although GA ≥16.8% has comparable performance for diagnosing DM as HbA1c ≥6.5% (>48mmol/mol), GA, HbA1c and OGTT tests do not necessarily detect DM in the same individuals. GA should be used as an additional test rather than an alternative to HbA1c or OGTT and its use as the sole DM diagnostic test should be interpreted with caution to assure the correct classification of diabetic individuals.

Supporting information

S1 Table. Correlations of GA, HbA1c and factors potentially associated with the measurement of serum GA in all participants.

a Correlation is significant at the 0.01 level (2-tailed). b Correlation is significant at the 0.05 level (2-tailed). GA, glycated albumin; HbA1c, glycated haemoglobin; FPG, fasting plasma glucose; 2hPG, plasma glucose 2 h after oral glucose; BMI, body mass index; WC, waist circumference (cm); Trigl., Triglyceride; HDL, serum high density lipoprotein cholesterol; LDL, serum low density lipoprotein cholesterol.

(DOCX)

Click here for additional data file. (35.9KB, docx)

Acknowledgments

The authors thank the staff of Clinical Biochemistry Unit and Laboratory Diagnosis Division of Hospital de Clinicas de Porto Alegre for their professionalism and dedication during blood collections and technical assistance.

Abbreviations

DM

type 2 diabetes mellitus

FPG

fasting plasma glucose

OGTT

oral glucose tolerance test

HbA1c

glycated haemoglobin

GA

glycated albumin

GSP

glycated serum proteins

HCPA

Hospital de Clinicas de Porto Alegre

LR

likelihood ratios

ROC

receiver operating characteristic

AUC

area under the ROC curve

WC

waist circumference

2hPG

2-h plasma glucose after a 75-g OGTT

Data Availability

Data cannot be shared publicly due to ethical restrictions. The use of the data for further research needs to be approved by the Research Ethics Committee of the Hospital de Clinicas de Porto Alegre. Requests to access the data may be submitted to corresponding author (contact via jcamargo@hcpa.edu.br) or contact Human Research Ethics Committee of the Hospital de Clinicas de Porto Alegre (HCPA) via cep@hcpa.edu.br.

Funding Statement

This work was supported by the Research Incentive Fund (FIPE) of the Hospital de Clínicas de Porto Alegre (HCPA). FCC received scholarship from Ministry of Science and Technology, Higher Education and Professional Technician (MCTESTP) of the Republic of Mozambique. MKH received a undergraduate scholarship from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.International Diabetes Federation. IDF Diabetes Atlas. Eighth edition. Brussels, Belgium: International Diabetes Federation; http://www.diabetesatlas.org. 2017. Accessed 17 Jan 2019 [Google Scholar]
  • 2.Herman WH, Ye W, Griffin SJ, Simmons RK, Davies MJ, Khunti K, et al. Early detection and treatment of type 2 diabetes reduce cardiovascular morbidity and mortality: a simulation of the results of the Anglo-DanishDutch study of intensive treatment in people with screen-detected diabetes in primary care (ADDITION-Europe). Diabetes Care. 2015; 38(8):1449–55. 10.2337/dc14-2459 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.American Diabetes Association (ADA) Diagnosis and classification of diabetes mellitus. Diabetes Care 41(Supplement 1): S13–S27. http://care.diabetesjournals.org/content/41/Supplement_1/S13. (2018). [DOI] [PubMed] [Google Scholar]
  • 4.National Glycohemoglobin Standardization Program (NGSP) Factors that Interfere with HbA1c Test Results [Updated 10/5/2018]. Available from http://www.ngsp.org/factors.asp. Accessed 12 Dez 2018.
  • 5.Cavagnolli G, Pimentel AL, Freitas PA, Gross JL, Camargo JL. Factors affecting A1c in non-diabetic individuals: review and meta-analysis. Clin Chim Acta. 2015; 445:107–14. 10.1016/j.cca.2015.03.024 [DOI] [PubMed] [Google Scholar]
  • 6.Rohlfing C, Hanson S, Weykamp C, Siebelder C, Higgins T, Molinaro R, et al. Effects of Hemoglobin C, D, E and S traits on Measurements of Hemoglobin A1c by Twelve Methods. Clin Chim Acta. 2016; 455:80–3. 10.1016/j.cca.2016.01.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Freitas PAC, Ehlert LR, Camargo JL. Glycated albumin: a potential biomarker in diabetes. Arch Endocrinol Metab 2017; 61(3):296–304. 10.1590/2359-3997000000272 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Furusyo N, Hayashi J. Glycated albumin and diabetes mellitus. Biochim Biophys Acta. 2013; 1830(12):5509–14. 10.1016/j.bbagen.2013.05.010 [DOI] [PubMed] [Google Scholar]
  • 9.Selvin E, Rawlings AM, Grams M, Klein R, Sharrett AR, Steffes M, et al. Fructosamine and glycated albumin for risk stratification and prediction of incident diabetes and microvascular complications: a prospective cohort analysis of the Atherosclerosis Risk in Communities (ARIC) study. Lancet Diabetes Endocrinol. 2014; 2(4):279–88. 10.1016/S2213-8587(13)70199-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Selvin E, Rawlings AM, Lutsey P, Maruthur N, Pankow JS, Steffes M, et al. Fructosamine and Glycated Albumin and the Risk of Cardiovascular Outcomes and Death. Circulation. 2015; 132(4):269–77. 10.1161/CIRCULATIONAHA.115.015415 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Nathan DM, McGee P, Steffes MW, Lachin JM DCCT/EDIC Research Group. Relationship of glycated albumin to blood glucose and HbA1c values and to retinopathy, nephropathy, and cardiovascular outcomes in the DCCT/EDIC study. Diabetes. 2014; 63(1):282–90. 10.2337/db13-0782 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Wu WC, Ma WY, Wei JN, Yu TY, Lin MS, Shih SR, et al. Serum Glycated Albumin to Guide the Diagnosis of Diabetes Mellitus. PLoS One. 2016; 11(1):e0146780 10.1371/journal.pone.0146780 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Furusyo N, Koga T, Ai M, Otokozawa S, Kohzuma T, Ikezaki H, Schaefer EJ, et al. Utility of glycated albumin for the diagnosis of diabetes mellitus in a Japanese population study: results from the Kyushu and Okinawa Population Study (KOPS). Diabetologia. 2011; 54(12):3028–36. 10.1007/s00125-011-2310-6 [DOI] [PubMed] [Google Scholar]
  • 14.Hsu P, Ai M, Kanda E, Yu NC, Chen HL, Chen HW, et al. A comparison of glycated albumin and glycosylated hemoglobin for the screening of diabetes mellitus in Taiwan. Atherosclerosis. 2015; 242(1):327–33. 10.1016/j.atherosclerosis.2015.07.037 [DOI] [PubMed] [Google Scholar]
  • 15.Ma XJ, Pan JM, Bao YQ, Zhou J, Tang JL, Li Q, et al. Combined assessment of glycated albumin and fasting plasma glucose improves the detection of diabetes in Chinese subjects. Clin Exp Pharmacol Physiol. 2010; 37(10):974–9. 10.1111/j.1440-1681.2010.05417.x [DOI] [PubMed] [Google Scholar]
  • 16.Chan CL, Pyle L, Kelsey M, Newnes L, Zeitler PS, Nadeau KJ. Screening for type 2 diabetes and prediabetes in obese youth: evaluating alternate markers of glycemia—1,5-anhydroglucitol, fructosamine, and glycated albumin. Pediatr Diabetes. 2016; 17(3):206–11. 10.1111/pedi.12258 [DOI] [PubMed] [Google Scholar]
  • 17.Bellia C, Zaninotto M, Cosma C, Agnello L, Bivona G, Marinova M, et al. Clinical usefulness of Glycated Albumin in the diagnosis of diabetes: Results from an Italian study. Clin Biochem. 2018; 54:68–72. 10.1016/j.clinbiochem.2018.02.017 [DOI] [PubMed] [Google Scholar]
  • 18.Zemlin AE, Barkhuizen M, Kengne AP, Erasmus RT, Matsha TE. Performance of glycated albumin for type 2 diabetes and prediabetes diagnosis in a South African population. Clin Chim Acta. 2019; 488:122–128. 10.1016/j.cca.2018.11.005 [DOI] [PubMed] [Google Scholar]
  • 19.Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig L, et al. STARD Group. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. Clin Chem. 2015; 61(12):1446–52. 10.1373/clinchem.2015.246280 [DOI] [PubMed] [Google Scholar]
  • 20.Nathan DM, Steffes MW, Sun W, Rynders GP, Lachin JM. Determining stability of stored samples retrospectively: the validation of glycated albumin. Clin Chem. 2011; 57(2):286–90. 10.1373/clinchem.2010.150250 [DOI] [PubMed] [Google Scholar]
  • 21.Little RR, Rohlfing C, Sacks DB. The National Glycohemoglobin Standardization Program: Over 20 Years of Improving Hemoglobin A1c Measurement. Clin Chem. 2018; 296962 10.1373/clinchem.2018.296962 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Freitas PAC, Ehlert LR, Camargo JL. Comparison between two enzymatic methods for glycated albumin. Anal Methods 2016; 8:8173–8. [Google Scholar]
  • 23.Rodriguez-Capote K, Tovell K, Holmes D, Dayton J, Higgins TN. Analytical evaluation of the Diazyme glycated serum protein assay on the siemens ADVIA 1800: comparison of results against HbA1c for diagnosis and management of diabetes. J Diabetes Sci Technol 2015; 9(2):192–9. 10.1177/1932296814567894 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Akobeng AK. Understanding diagnostic tests 2: likelihood ratios, pre- and post-test probabilities and their use in clinical practice. Acta Paediatr. 2007; 96(4):487–91. 10.1111/j.1651-2227.2006.00179.x [DOI] [PubMed] [Google Scholar]
  • 25.Kallner A. Bayes' theorem, the ROC diagram and reference values: Definition and use in clinical diagnosis. Biochemia medica, 2017; 28(1):010101 10.11613/BM.2018.010101 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Selvin E1, Steffes MW, Ballantyne CM, Hoogeveen RC, Coresh J, Brancati FL. Racial differences in glycemic markers: a cross-sectional analysis of community-based data. Ann Intern Med. 2011; 154(5):303–9. 10.7326/0003-4819-154-5-201103010-00004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Cavagnolli G, Comerlato J, Comerlato C, Renz PB, Gross JL, Camargo JL., et al. HbA1c measurement for the diagnosis of diabetes: is it enough?. Diabetic Medicine 2011; 28: 31–35. 10.1111/j.1464-5491.2010.03159.x [DOI] [PubMed] [Google Scholar]
  • 28.Koga M. Glycated albumin; clinical usefulness. Clin Chim Acta. 2014; 433:96–104. 10.1016/j.cca.2014.03.001 [DOI] [PubMed] [Google Scholar]
  • 29.Wang F, Ma X, Hao Y, Yang R, Ni J, Xiao Y, et al. Serum Glycated Albumin Is Inversely Influenced by Fat Mass and Visceral Adipose Tissue in Chinese with Normal Glucose Tolerance. PLoS One. 2012; 7(11):e51098 10.1371/journal.pone.0051098 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Sumner AE, Duong MT, Bingham BA. Glycated Albumin Identifies Prediabetes Not Detected by Hemoglobin A1c: The Africans in America Study. Clin Chem. 2016; 62 (11) 1524–1532. 10.1373/clinchem.2016.261255 [DOI] [PubMed] [Google Scholar]
  • 31.Cohen MP. Clinical, pathophysiological and structure/function consequences of modification of albumin by Amadori-glucose adducts. Biochim Biophys Acta. 2013; 1830(12):5480–5. 10.1016/j.bbagen.2013.04.024 [DOI] [PubMed] [Google Scholar]

Decision Letter 0

Petter Bjornstad

24 Sep 2019

PONE-D-19-21107

Glycated albumin as a diagnostic tool in diabetes: an alternative or an additional test?

PLOS ONE

Dear Dr. Camargo,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

==============================

ACADEMIC EDITOR: Minor revisions requested

==============================

We would appreciate receiving your revised manuscript by Nov 08 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Petter Bjornstad

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2.  We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

Additional Editor Comments (if provided):

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript entitled “Glycated albumin as a diagnostic tool in diabetes: an alternative or an additional test?” by Chume et al. presents an evaluation of the diagnostic use of glycated albumin (GA) as a diagnostic tool for type 2 diabetes in a cohort of individuals from a single tertiary center in Brazil. IRB approval was endorsed and a participant consent process was completed for participation in this study. Overall, the study was well-developed and detailed. The experiments were well-conducted and the analysis was appropriate to evaluate the stated main study question.

TITLE and ABSTRACT:

1. As with the remainder of the paper, would recommend changing all notations of “patients” to either “participants” or “subjects” to follow people first language.

2. It would be helpful to characterize your study population in the results section of the abstract, if possible (i.e. number of participants, average age, BMI, etc.) to allow the reader a frame of reference.

3. Your conclusion statement that GA should be used as an adjunctive test instead of an alternative test to HbA1c or OGTT is somewhat confusing as you state in your introduction that either fasting plasma BG, HbA1c, or 2h OGTT can be used to diagnose T2D – why is GA different than HbA1c if you have come to the conclusion that “GA showed overall diagnostic accuracy similar to HbA1c in the diagnosis of DM” – how did you come to that conclusion?

4. In your abbreviations, I would label DM as "type 2 diabetes mellitus" instead of simply "diabetes mellitus" as you do not refer to other forms of diabetes mellitus such as type 1 diabetes mellitus in this manuscript. You should also define OGTT as an "oral glucose tolerance test" in this section.

INTRODUCTION:

1. Would consider splitting the first paragraph into two paragraphs given the paragraph length.

2. Would recommend adding a hypothesis statement in addition to the aim statement at the end of the “Introduction” section.

METHODS:

1. Why was HbA1c used only for descriptive purposes and comparison with GA if ADA criteria state that T2D can be diagnosed with a FPG ≥126 mg/dL, 2hG on OGTT ≥200 mg/dL, HbA1c ≥6.5%, or random plasma glucose ≥200 mg/dL with symptoms of hyperglycemia? Should it also be used to define participants with T2D if going by true ADA T2D diagnosis criteria? It seems like you use it to diagnose T2D in the results section (i.e. table 2) so the methods section should reflect that.

RESULTS:

1. The average BMI in this paper was noted to 28.9 +/- 6.3 kg/m2 which falls in the overweight category with a large percentage of study participants also being obese, how do you think that impacts your results for GA utility as you have rightly previously stated that both age and obesity are factors that impact GA levels?

2. It’s unclear what separating out clinical and laboratory characteristics of the cohort by the upper tertile of GA values adds to the data given you only reference the equilibrium threshold of 14.8% and the value of 16.8% as the cut off that demonstrates a similar sensitivity/specificity as HbA1c. Why did you select 16.0%? Would it make more sense to select 14.8% or 16.8%?

DISCUSSION:

1. You mention briefly that the HbA1c, GA, and OGTT tests do not reflect the same participants when a diagnosis is made of T2D. I think this is a really important point and it would be good to explore that more as the use of OGTT and HbA1c are currently both accepted for a diagnosis of T2D even though in your population, they only overlapped in terms of a diagnosis of T2D by both measures in 24/86 of the participants. Why do you think that is? It seems like that number would only decrease if combining HbA1c, OGTT, AND GA so what comments do you have about why all three of these measures are detecting T2D in completely different individuals? And does that mean that we should accept a diagnosis of T2D if any one of these tests is positive or if all 3 are positive? Or is one test superior to all of the others? If adding GA as an adjunctive test to the diagnosis of T2D, how would we interpret positive vs. negative results in terms of our diagnosis and management?

CONCLUSIONS:

1. What comments do you have about the generalizability of these results as this study was completed at a single center in Brazil?

2. It would also be worth mentioning that because this study was completed in a population at high risk (i.e. they were referred for an OGTT due to some predisposing factor), results about GA can only really be interpreted if obtained in a similar high-risk population (i.e. one with a high pre-test probability) rather than as a general population screening tool.

Reviewer #2: The authors examined the utility of glycated albumin as a screening tool for diabetes mellitus. Similar work has been performed in other (mostly Asian) populations, and found GA to be useful in some conditions. This report verifies the potential utility of GA as a secondary screening measure in Brazilian patients with high risk of developing DM who were seen at a tertiary hospital. It is promising work in an important area.

My biggest concern is that this is a study cohort with a relatively broad age range, and GA is well known to increase with age. Therefore, it might be worth adjusting these analyses for age, or investigating different cut points for different age groups.

Below are my other comments:

1. In the abstract results (line 43), please include the sensitivity and specificity information for the HbA1c cutoff, so readers can easily compare it to the reported GA cutoff.

2. When comparing paired areas under the ROC curve (e.g. two diagnostic tests on the same group), it is better to use DeLong’s test rather than t-tests.

3. Post-test probability should be calculated directly rather than estimated graphically.

4. Please explain why in table 1 the group is divided by upper tertile of GA, rather than at the equilibrium cut point 14.8% or high specificity cut point 16.8%.

5. Ethnic difference between groups should be assessed with Fisher’s exact test rather than chi-squared, though this is unlikely to make much of a difference. Please also remove the “trended toward significance” language.

6. Line 242 – Please clarify that some of the other studies (e.g. Chan et al.) also suggest GA as a secondary screening tool, not as a primary diagnostic test.

7. Figure 2 – it looks to me as if the HbA1c curve has been smoothed but the GA has not, though this might not be correct.

8. Figure 3 – I find this figure confusing. Please either clarify the legend or remove the figure.

9. Figure 4 – I don’t think this figure is necessary, but it doesn’t detract from the paper either.

10. Please include all CIs in the text as well as figures.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Tim Vigers

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2019 Dec 31;14(12):e0227065. doi: 10.1371/journal.pone.0227065.r002

Author response to Decision Letter 0

22 Oct 2019

ANSWERS TO REVIEWERS:

Reviewer #1: The manuscript entitled “Glycated albumin as a diagnostic tool in diabetes: an alternative or an additional test?” by Chume et al. presents an evaluation of the diagnostic use of glycated albumin (GA) as a diagnostic tool for type 2 diabetes in a cohort of individuals from a single tertiary center in Brazil. IRB approval was endorsed and a participant consent process was completed for participation in this study. Overall, the study was well-developed and detailed. The experiments were well-conducted and the analysis was appropriate to evaluate the stated main study question.

Thank you for your careful evaluation and relevant suggestions/comments.

TITLE and ABSTRACT:

1. As with the remainder of the paper, would recommend changing all notations of “patients” to either “participants” or “subjects” to follow people first language.

Answer: We amended the text accordingly throughout the manuscript.

2. It would be helpful to characterize your study population in the results section of the abstract, if possible (i.e. number of participants, average age, BMI, etc.) to allow the reader a frame of reference.

Answer: The text was amended in Page 2, lines 39 and 40.

3. Your conclusion statement that GA should be used as an adjunctive test instead of an alternative test to HbA1c or OGTT is somewhat confusing as you state in your introduction that either fasting plasma BG, HbA1c, or 2h OGTT can be used to diagnose T2D – why is GA different than HbA1c if you have come to the conclusion that “GA showed overall diagnostic accuracy similar to HbA1c in the diagnosis of DM” – how did you come to that conclusion?

Answer: Thank you for this comment. The GA cut-off of 16.8% had similar performance for detecting DM as defined by HbA1c >6.5% (>48 mmol/mol) with sensitivity of 31.2% and specificity of 93.3%. As shown in Venn diagram (Figure 3), among 77 subjects diagnosed with DM by OGTT, only 11 were identified with DM by both GA ≥16.8% and HbA1c ≥6.5%. However, 13/77 individuals were identified only by GA ≥16.8% and another 13/77 were identified only by HbA1c ≥6.5%. We amended the text for clarity and change to “GA detected different subjects with DM from those detected by HbA1c, though GA it showed overall diagnostic accuracy similar to HbA1c in the diagnosis of DM” in the Conclusion on page 2, lines 48 and 49. We also elucidate in the results mentioning “Among 77 subjects diagnosed with DM by OGTT, only 11 were identified as DM subjects by both GA ≥16.8% and HbA1c ≥6.5%. Thirteen of 77 subjects were identified only by GA ≥16.8% and another 13/77 were identified only by HbA1c ≥6.5%” on page 17, lines 272 - 275.

4. In your abbreviations, I would label DM as "type 2 diabetes mellitus" instead of simply "diabetes mellitus" as you do not refer to other forms of diabetes mellitus such as type 1 diabetes mellitus in this manuscript. You should also define OGTT as an "oral glucose tolerance test" in this section.

Answer: We amended the text accordingly throughout the manuscript.

INTRODUCTION:

1. Would consider splitting the first paragraph into two paragraphs given the paragraph length.

Answer: We amended the text accordingly. Page 4, line 70.

2. Would recommend adding a hypothesis statement in addition to the aim statement at the end of the “Introduction” section.

Answer: We added the following sentence on page 5, in paragraph 2, lines 99 – 101: “We hypothesized that GA may be used in the diagnosis of DM and in clinical conditions where the HbA1c test does not accurately reflect blood glucose concentrations GA may be an alternative marker”.

METHODS:

1. Why was HbA1c used only for descriptive purposes and comparison with GA if ADA criteria state that T2D can be diagnosed with a FPG ≥126 mg/dL, 2hG on OGTT ≥200 mg/dL, HbA1c ≥6.5%, or random plasma glucose ≥200 mg/dL with symptoms of hyperglycemia? Should it also be used to define participants with T2D if going by true ADA T2D diagnosis criteria? It seems like you use it to diagnose T2D in the results section (i.e. table 2) so the methods section should reflect that.

Answer: Thank you for this comment. This study followed the STARD 2015 reporting guideline for diagnostic accuracy studies which recommend to choose as reference standard test the best available method for establishing the presence or absence of the target condition. OGTT is a direct test for reflection of blood glucose level and numerous studies have shown that it diagnoses most individuals with DM compared with FPG and HbA1c. Besides, HbA1c has low sensitivity at the designated diagnostic cut point. Moreover, the purpose of the present study was to evaluate GA in the diagnosis of DM assessing the ability to be one HbA1c alternative test. However after your comment, we performed an analysis to evaluate the performance of GA using OGTT and/or HbA1c as DM diagnostic reference test. There was no relevant change in the AUC of GA compared to the one obtained when OGTT solely is DM diagnostic reference test [0.708 (95% CI 0.639 – 0.776)] versus [0.703 (95% CI 0.631 – 0.775)], respectively. The optimal cut-off value for serum GA, when OGTT and/or HbA1c are reference was 14.7% (sensitivity 64.0% and specificity 64.1%) versus 14.8% (sensitivity 64.9% and specificity 65.5%) when OGTT alone is reference.

We have added to the text the results of ROC analysis evaluating the performance of GA using OGTT and/or HbA1c as DM diagnostic reference test (page 14, lines 250 - 256). Also, we amended the text in the Methods section accordingly in page 9, lines 173 - 177.

RESULTS:

1. The average BMI in this paper was noted to 28.9 +/- 6.3 kg/m2 which falls in the overweight category with a large percentage of study participants also being obese, how do you think that impacts your results for GA utility as you have rightly previously stated that both age and obesity are factors that impact GA levels?

Answer: This is a very interesting question. We have analyzed the correlations between GA and factors potentially associated with its measurement but due to manuscript legth we did not presented previously, now we included these data in S1 Table. Also, we amended the text in the Methods (on page 8, lines 171-172), Results (page 14, lines 224-235) and Discussion (page 20, lines 335-344) sections accordingly. GA was associated with age, albumin and triglycerides. However, correlations of GA with BMI, waist circumference (WC) and LDL were not significant. We believe that the size of the study population may be not large enough to evaluate these relationships. Although, different of GA, HbA1c was more sensitive to BMI and WC. This may also explain why in the study of Sumner et al., 2016, GA identified a substantial number of non-obese individuals with prediabetes not detected by HbA1c. The follow reference was added to the manuscript: “Sumner AE, Duong MT, Bingham BA. Glycated Albumin Identifies Prediabetes Not Detected by Hemoglobin A1c: The Africans in America Study. Clin Chem. 2016; 62 (11) 1524-1532. doi: 10.1373/clinchem.2016.261255” (page 28, lines 496 - 498).

2. It’s unclear what separating out clinical and laboratory characteristics of the cohort by the upper tertile of GA values adds to the data given you only reference the equilibrium threshold of 14.8% and the value of 16.8% as the cut off that demonstrates a similar sensitivity/specificity as HbA1c. Why did you select 16.0%? Would it make more sense to select 14.8% or 16.8%?

Answer: Thank you for this observation. We had hypothesized that GA values in the upper tertile had a high probability to diagnose DM, therefore we choose to divide the subjects using 16% of GA value. However, after your comment, we amended the table and described the clinical and laboratory characteristics of the participants divided by subjects with and without DM using ADA OGTT criteria (Table 1 in page 12 and 13). Also, we amended the text in the Methods (page 8, lines 169 and 170) and Results (page 10, lines 207-217) sections accordingly.

DISCUSSION:

1. You mention briefly that the HbA1c, GA, and OGTT tests do not reflect the same participants when a diagnosis is made of T2D. I think this is a really important point and it would be good to explore that more as the use of OGTT and HbA1c are currently both accepted for a diagnosis of T2D even though in your population, they only overlapped in terms of a diagnosis of T2D by both measures in 24/86 of the participants. Why do you think that is? It seems like that number would only decrease if combining HbA1c, OGTT, AND GA so what comments do you have about why all three of these measures are detecting T2D in completely different individuals? And does that mean that we should accept a diagnosis of T2D if any one of these tests is positive or if all 3 are positive? Or is one test superior to all of the others? If adding GA as an adjunctive test to the diagnosis of T2D, how would we interpret positive vs. negative results in terms of our diagnosis and management?

Answer: This a very interesting point, thank you for this remark. FPG, 2h PG during 75g OGTT, and HbA1c are equally appropriate for DM diagnostic testing. Therefore, these tests may be used to screen and diagnose DM. Nevertheless, it should be noted that the tests do not necessarily detect DM in the same individuals. The relationships among FPG, 2h PG and HbA1c are imperfect and they reflect glycemia by different mechanisms resulting in detection of hyperglycaemia in different stages. According to National Health and Nutrition Examination Survey (NHANES) data, HbA1c test, with a diagnostic threshold of 6.5%, diagnoses only 30% of the diabetes cases identified collectively using HbA1c, FPG, or 2h PG (Cowie CC, Rust KF, Byrd-Holt DD, et al. Prevalence of diabetes and high risk for diabetes using A1C criteria in the U.S. population in 1988–2006. Diabetes Care 2010;33:562–568). Compared with FPG and HbA1c tests, 2h PG diagnoses more people with DM [Meijnikman AS, De Block CEM, Dirinck E, et al. Not performing an OGTT results in significant underdiagnosis of (pre)diabetes in a high risk adult Caucasian population. Int J Obes 2017;41:1615–1620]. We previously showed that HbA1c at 6.5% cut-off is not enough to diagnose all cases of DM (reference #27 in this manuscript). In this present study, based on exclusively HbA1c ≥6.5% for DM diagnosis, only 31.2% of diabetic subjects were detected collectively with OGTT.

Considering the above-mentioned findings and the similarities between GA and HbA1c, GA should be considered as an equally appropriate test for DM diagnosis. Like FPG, 2h PG, and HbA1c, GA would not necessarily detect DM in the same individuals. However, as for HbA1c, it is important to recognize that GA is an indirect measure of blood glucose levels and other factors may impact glycation of albumin independently of glycemia. We highlighted this throughout the discussion.

For a better interpretation of GA values, we believe a better understanding of GA role in DM and prevention of diabetes-specific complications should be extensively studied with long-term follow-up.

CONCLUSIONS:

1. What comments do you have about the generalizability of these results as this study was completed at a single center in Brazil?

Answer: Hospital de Clínicas de Porto Alegre is a large tertiary hospital with multiple specialties located in Porto Alegre, Southern Brazil. Since it is a public hospital with priority for patients of Sistema Único de Saúde (SUS, Brazilian public health care system) has become a reference for the state of Rio Grande do Sul and southern Brazil. Furthermore, in our state the majority of population has European ancestry (https://biblioteca.ibge.gov.br/visualizacao/livros/liv63405.pdf). The proportion of Caucasian descendent people is higher than 80% in this region, similar to the proportion of Caucasian descendent people in this study (80.2%). Therefore, we believe our study population represents a sample of southern Brazil population and may correspond to the majority of European population allowing the applicability of our results in similar populations. These data were added to the text in Page 21, lines 353-355).

2. It would also be worth mentioning that because this study was completed in a population at high risk (i.e. they were referred for an OGTT due to some predisposing factor), results about GA can only really be interpreted if obtained in a similar high-risk population (i.e. one with a high pre-test probability) rather than as a general population screening tool.

Answer: We have added and amended the text accordingly in the study limitations in page 21, lines 359 – 361.

Reviewer #2: The authors examined the utility of glycated albumin as a screening tool for diabetes mellitus. Similar work has been performed in other (mostly Asian) populations, and found GA to be useful in some conditions. This report verifies the potential utility of GA as a secondary screening measure in Brazilian patients with high risk of developing DM who were seen at a tertiary hospital. It is promising work in an important area.

My biggest concern is that this is a study cohort with a relatively broad age range, and GA is well known to increase with age. Therefore, it might be worth adjusting these analyses for age, or investigating different cut points for different age groups.

Answer: We made correlations between GA and factors potentially associated with the measurement of serum GA, and presented in S1 Table. GA was found to be associated with age. GA concentrations increased by 0.44% per decade (GA = 12.503 + 0.044 x age). However similar correlation was found between HbA1c and age. We amended the text accordingly on Page 14, lines 224-235 and Page 20, lines 335-344.

1. In the abstract results (line 43), please include the sensitivity and specificity information for the HbA1c cutoff, so readers can easily compare it to the reported GA cutoff.

Answer: The text was amended on Page 2, line 46.

2. When comparing paired areas under the ROC curve (e.g. two diagnostic tests on the same group), it is better to use DeLong’s test rather than t-tests.

Answer: We apologize, we made a mistake by mentioning that the AUC of GA and HbA1c were compared by T-test, we actually used DeLong’s test in MedCalc. The text was amended accordingly on Page 9, line 178 and lines 190-192.

3. Post-test probability should be calculated directly rather than estimated graphically.

Answer: The Fagan nomogram was only used as a visual aid to help interpretation and applicability. All post-test probabilities were calculated by using the following formula: Post-test probability = Post-test odds /(Post-test odds + 1). In this equation, positive post-test probability was calculated using the likelihood ratio positive, and the negative post-test probability was calculated using the likelihood ratio negative. Likelihood ratio was calculated from sensitivity and specificity of GA test. Post-test odds = Pre-test odds x Likelihood ratio, being Pre-test odds = (Pre-test probability /(1 - Pre-test probability). Pre-test probability was according to International Diabetes Federation – IDF data. However, for illustration of the results and better understanding of clinical applicability, we used the Fagan nomogram (Figure 4; where we mentioned the above formulas).

4. Please explain why in table 1 the group is divided by upper tertile of GA, rather than at the equilibrium cut point 14.8% or high specificity cut point 16.8%.

Answer: As mentioned before in our response for Reviewer 1, we had hypothesized that GA values in upper tertile had a high probability to diagnose DM, therefore we choose to divide the subjects using 16% of GA value. However, after reviewers comments, we amended the table and described the clinical and laboratory characteristics of the participants divided by subjects with and without DM using ADA OGTT criteria (Table 1 in page 12 and 13). Also, we amended the text in the Methods (page 8, lines 169 and 170) and Results (page 10, lines 207-217) sections accordingly.

5. Ethnic difference between groups should be assessed with Fisher’s exact test rather than chi-squared, though this is unlikely to make much of a difference. Please also remove the “trended toward significance” language.

Answer: We re-analyzed these different by Fisher´s exact test and as you said, the results were similar (p = 0.147 vs. 0.152 for Fisher’s exact and chi-squared test, respectively). However, we amended accordingly (Table 1 in page 12 and 13). Also, we amended the text in the Methods section (page 8, line 168), and removed the “tended to be significant” on page 10, lines 210-211.

6. Line 242 – Please clarify that some of the other studies (e.g. Chan et al.) also suggest GA as a secondary screening tool, not as a primary diagnostic test.

Answer: We amended the text for clarity and change to “Although GA is not currently recommended for the screening or diagnosis of DM, there are several studies which advocate GA as a screening test for undiagnosed DM, still some studies have recommended the test as a secondary screening tool” on page 18, line 290.

7. Figure 2 – it looks to me as if the HbA1c curve has been smoothed but the GA has not, though this might not be correct.

Answer: That is interesting observation, but it is just impression. We believe since HbA1c has few data points (markers) the curve chart gives impression as if has been smoothed, while GA has numerous data points give unsmoothed shape.

8. Figure 3 – I find this figure confusing. Please either clarify the legend or remove the figure.

Answer: We used Venn diagram to present the number of individuals with DM diagnosed by each test and overlaps. We would like to keep the figure to illustrate the relationships between the tests, highlighting how the they are similar and different. We amended the legend for clarity accordingly (page 29, lines 510 – 518).

9. Figure 4 – I don’t think this figure is necessary, but it doesn’t detract from the paper either.

Answer: As we mentioned above, the Fagan nomogram was used as a visual aid to help interpretation and applicability of our findings. Therefore, we believe it is of certain utility for readers and, if possible, we would like to keep this figure for clarity.

10. Please include all CIs in the text as well as figures.

Answer: We revised and amended the manuscript accordingly.

Last, we would like to express our appreciation to you and the reviewers advices for suggesting how to improve our paper.

Thank you very much!

Yours sincerely

Attachment

Submitted filename: Response to Reviewers PlosOne GA and DM.docx

Click here for additional data file. (40.5KB, docx)

Decision Letter 1

Petter Bjornstad

12 Dec 2019

Glycated albumin as a diagnostic tool in diabetes: an alternative or an additional test?

PONE-D-19-21107R1

Dear Dr. Camargo,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements. Congratulations!

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Petter Bjornstad

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Fantastic work with your responses to questions from the reviewers. A few errors that were noted that should be corrected before the manuscript makes it into print are as follows:

INTRODUCTION:

1. Type 2 DM mellitus (DM) on page 4 line 61 should be changed to type 2 diabetes mellitus (DM) for clarity. This should also be reflected on page 4 line 67 where it should read DM instead of DM type 2 (due to previous definition).

TABLE 1:

1. All fields with mean +/- SD should include a +/- and not just a + in this table.

DISCUSSION:

1. Page 21 line 352 should state "enrollment" instead of "enrolment."

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Acceptance letter

Petter Bjornstad

18 Dec 2019

PONE-D-19-21107R1

Glycated albumin as a diagnostic tool in diabetes: an alternative or an additional test?

Dear Dr. Camargo:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr Petter Bjornstad

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Table. Correlations of GA, HbA1c and factors potentially associated with the measurement of serum GA in all participants.

    a Correlation is significant at the 0.01 level (2-tailed). b Correlation is significant at the 0.05 level (2-tailed). GA, glycated albumin; HbA1c, glycated haemoglobin; FPG, fasting plasma glucose; 2hPG, plasma glucose 2 h after oral glucose; BMI, body mass index; WC, waist circumference (cm); Trigl., Triglyceride; HDL, serum high density lipoprotein cholesterol; LDL, serum low density lipoprotein cholesterol.

    (DOCX)

    Click here for additional data file. (35.9KB, docx)
    Attachment

    Submitted filename: Response to Reviewers PlosOne GA and DM.docx

    Click here for additional data file. (40.5KB, docx)

    Data Availability Statement

    Data cannot be shared publicly due to ethical restrictions. The use of the data for further research needs to be approved by the Research Ethics Committee of the Hospital de Clinicas de Porto Alegre. Requests to access the data may be submitted to corresponding author (contact via jcamargo@hcpa.edu.br) or contact Human Research Ethics Committee of the Hospital de Clinicas de Porto Alegre (HCPA) via cep@hcpa.edu.br.

    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES