Printer Friendly

COMPARISON OF DIAGNOSTIC ACCURACY OF GLYCATED HEMOGLOBIN (HBA1C) AND 2-HOUR POST GLUCOSE FOR DIABETES USING AMERICAN DIABETIC ASSOCIATION (ADA) CUT OFFS IN PAKISTANI POPULATION.

Byline: Sheharbano Imran, Saira Furqan, Hafsa Majid, Lena Jafri, Najmul Islam and Aysha Habib Khan

ABSTRACT

Objective: To determine median HbA1c level in patients screened for DM and compare the diagnostic accuracy of HbA1c with the currently recommended ADA cut off, using 2 hours (h) plasma glucose (2-h PG) post 75 grams oral glucose tolerance test (OGTT) as gold standard.

Study Design: Cross sectional study.

Place and Duration of Study: Aga Khan University Hospital Karachi, from Jul 2014 to Nov 2015.

Patients and Methods: Consecutive subjects screened by OGTT for diabetes mellitus (DM) were included. Blood sample for plasma glucose and HbA1c were analyzed on ADVIA 1800.

Results: Total 146 subjects were included with median age 45 (IQR 54.2-35) years; 53% (n=77) being females. Median HbA1c of the study subjects was 6% (IQR 6.6-5.6). Positive correlation was observed between age and HbA1c [5.7% (IQR 6.2-5.3) 40 y subjects; r 0.34, p=0.000). Males had higher HbA1c than females [6.1% (IQR 6.8-5.7) years. 5.9% (IQR 6.4-5.4); r 0.17, p=0.036). HbA1c levels were significantly different amongst the different ethnic groups residing in Pakistan (p-value<0.03). HbA1c was positively correlated with FPG (r=0.59, p-value<0.001) and 2-h PG (r=0.56, p-value<0.001). Sensitivity, specificity, PPV, NPV for HbA1c at cutoff of 6.5% in diagnosing diabetes mellitus (DM) were 70%, 89%, 72% and 80% respectively.

Conclusion: High median HbA1c levels is noted in our subjects. At cut off level of HbA1c [greater than or equal to]6.5%, 70% of subjects with DM were diagnosed using ADA criteria of 2-h PG for diabetes as gold standard. Ethnic differences were also observed in levels of HbA1c. There is a need to define cut off for our population.

Keywords: Diabetes Mellitus, Ethnic Groups, Glycosylated Hemoglobin (HbA1c).

INTRODUCTION

Incidence of Diabetes Mellitus (DM) is on rise worldwide. World Health Organization (WHO) has predicted a 170% rise in its incidence in developing countries, which accounts for 228 million patients with DM; consisting 75% of the total world diabetic population1. In Pakistan prevalence of DM in three provinces, i.e., Sindh, Baluchistan and Khyber Pakhtunkhwa is 13.9%, 8.6% and 11.7% respectively2,3. In Pakistan, prevalence of DM is high ranging from 7.6 to 11%4. Diagnosis of DM is a challenge, as patient remains asymptomatic for several years and would not be tested5,6. Commonly used diagnostic tests are fasting plasma glucose (FPG) and 75-gm oral glucose tolerance test (OGTT). However, these tests require not only overnight fasting, but results are also affected by calorie restriction, dietary in take of last 72 hours and exercise prior to testing.

Also glucose level shows diurnal variation with higher level in morning and lower in afternoon; missing cases of u ndiagnosed DM if glucose estimation is performed in afternoon7. In addition, temperature affects glucose measurement by decreasing level by 3-8mg/dl per hour of glucose, when sample is kept at room temperature8,9. Most factors that affect FPG have minimal or no effect on HbA1c level. Short term lifestyle changes, recent food in take or little exercise do not affect HbA1c level. Samples are also stable at room temperature for up to 8 hours and analytical variability is <2%. Added advantage of HbA1c is that it predicts a subject's glycemic control over a period of 2-3 months9. Previously only FPG and OGTT were included in diagnostic criteria for DM.

In 2009 an International Expert Committee comprised of representatives of the American Diabetes Association (ADA), International Diabetes Federation (IDF) and European Association for Study of Diabetes (EASD) recommended that HbA1c can be utilized for labeling DM at a cutoff of [greater than or equal to]6.5% and using assay techniques which are certified by National Glycohemoglobin Standardization Program (NGSP) and traceable to Diabetes Control and Complications Trial (DCCT) reference method. In 2011 WHO also endorsed the recommendations of the International Expert Committee. HbA1c is now commonly used for diagnosing DM world-wide, but very few reports have shown its utility in diagnosing DM in our population. As the cutoffs defined for HbA1c are developed in Caucasian population. So, we wanted to evaluate the diagnostic accuracy of the ADA defined HbA1c cutoff in our population.

Literature has reported that ethnic differences also affect HbA1c level, and no such study observing differences in HbA1c level due to ethnicities has been done in our population. This study was designed to calculate the median HbA1c level in patients suspected of DM, and compare the diagnostic accuracy of HbA1c with the currently recom-mended ADA cut off, using OGTT as gold standard and to evaluate HbA1c level in different ethnicities of Pakistan.

PATIENTS AND METHODS

A cross sectional study was conducted from July 2014 to November 2015 at the Section of Chemical Pathology, Department of Pathology and Laboratory Medicine and Department of Medicine, Aga Khan University Karachi Pakistan. Study was commenced after taking approval from institution's ethical review committee (ERC no: 2306-Med-ERC-12). All consecutive subjects aged 18 to 65 years, referred for screening of DM by OGTT, at the Clinical Laboratory of AKUH were approached and included in the study after informed consent. Subjects on medications affecting glucose or HbA1c level e.g. corticosteroids, pregnancy, history of blood transfusion in last 3 months, chronic kidney disease, known hemoglobinopathies, anemia (hemoglobin of <11.0 g/dl in females, and <13.7g/dl in males), and those taking iron or vitamin B12 supplementation were excluded from the study.

After informed consent demographic details including weight, height, waist circumference, ethnicity (based on ancestral language) along with smoking history and family history of DM were recorded on a preformed questionnaire at the time of OGTT. Blood samples were collected in vacutainer containing EDTA for HbA1c and sodium fluoride for glucose. Samples for FPG were collected after a minimum of 8 hours overnight fasting and samples for 2-h PG were taken after oral administration of a standard dose of 75 g anhydrous glucose dissolved in water. Samples for FPG and 2-h PG were analyzed immediately, while whole blood samples for HbA1c were stored at -80AdegC until analysis. Whole blood HbA1c was analyzed by immunoassay method on ADVIA 1800 chemistry analyzer (Siemens Diagnostics, NY, US) using manufacturer provided recommendations. HbA1c assay is NGSP certified and traceable to the DCCT reference method.

For assessing system performance, three levels of quality control materials (low, medium and high) were run before analysis of each batch. Cutoffs used for HbA1c were normal a$?5.7%, prediabetes 5.7-6.4%, and diabetic [greater than or equal to]6.5%10,11. Plasma glucose was measured by hexokinase method on ADVIA 1800 chemistry analyzer (Siemens Diagnostics, NY, US). System performance was assessed using three levels of quality control i.e. low, medium and high. Clinical laboratory of AKUH is participating in external proficiency testing program of College of American Pathologists (CAP) during this study period and performance was acceptable. Cutoffs for FPG are, for normal <100mg/dl, prediabetes 100 to 125mg/dl and diabetic [greater than or equal to]126mg/dl and cutoffs for 2-h PG are normal <140mg/dl, prediabetes 140-199mg/dl and diabetic [greater than or equal to]200mg/d.10,11. Data was analyzed using the Statistical Package for the Social Sciences (SPSS version 19.0) for testing normality Kolmogorov-Smirnov test was applied.

Median value and IQR were computed for skewed quantitative variable and mean +- SD for others; whereas frequency and percentage were calculated for categorical variables. Subjects were labeled normal weight, overweight and obese using criteria proposed for Asian populations by International Association for the Study of Obesity and International Obesity Task Force of World Health Organization 200112,13. Median levels of HbA1c, FPG and 2-h PG were compared amongst ethnic groups by KruskalWallis test taking pa$?0.05 as significant. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) at 6.5% cutoff level of HbA1c for diabetes were calculated using 2x2 table.

Table-I: Demographic and biochemical parameters including ethnicity in study subjects.

Parameters###Overall###Urdu speaking###Punjabi###Sindhi###Others###p-value

n(%)###146###85(58)###28(19)###19(13)###14(10)###-

Age(years)###45(54.2-35)###45(53-35)###43.5(53.7-35.5)###40(56-35)###27.9(31.1-25)###0.524

Median with IQR

BMI(kg/m2)###27(30.5-24.3)###26.59(30.2-23.6)###27.8(32-24.7)###27.4(30.3-23.2)###27.9(31.1-25)###0.669

Median with IQR

Waist cicumference###35.5(40-33)###35(40-32)###38(40-34)###36(40-34)###36.5(40.2-31.5)###0.579

Median with IQR

HbA1c(%)###6(6.6-5.6)###6(6.6-5.5)###5.9(6.2-5.45)###6(7.4-5.6)###6.2(7-5.85)###0.216

Median with IQR

FPG(mg/dl)###94.5(110.5-86.7)###94(107-84)###94(105.7-87)###102(125-91)###104.5(129.5-87.2)###0.18

Median with IQR

2-h PG(mg/dl)###135.5(206-110.7)###134(196-105)###128(172-114)###156(269-117)###176(240.7-123.5)###0.157

Median with IQR

Table-II: Distribution of fasting plasma glucose (FPG) and 2-hour PG levels in study subjects according to ADA criteria and comparison with HbA1c levels (n=146).

###Plasma glucose(mg/dl)###HbA1c(%)

OGTT###Distributions###n(%)###Median(IQR)###Overall

###based on cutoff###Median(IQR)###HbA1c <5.7###HbA1c 5.7-6.4###HbA1c [greater than or equal to]6.5

###n(%)###Median(IQR)###n(%)###Median(IQR)###n(%)###Median(IQR)

###Normal <100###87(60)###88(94-81)###5.7(6-5.37)###35(24)###5.3(5.4-5)###2(1.4)###5.4###1(0.7)###5.1

FPG(mg/dl)###Prediabetes###37(25)###109(115-104)###6.3(6.8-6)###44(30)###5.9(6.1-5.72)###21(14.4)###6.2(6.3-5.95)###2(1.4)###6.1

###[greater than or equal to]100-125

###Diabetic###22(15)###143(184-133.5)###7.3(8.3-6.6)###7(5)###6.8(6.8-6.6)###15(10.3)###6.8(7.1-6.7)###19(13)###7.4(8.6-7)

###[greater than or equal to]126

###Normal###76(52)###114(125.7102)###5.7(6-5.3)###35###5.3(5.4-5)###44###5.9(6.1-5.7)###7###6.8(6.8-6.6)

###<140

2-hPG (mg/dl)###Pre-diabetes###28(19)###163(177.7-150.7)###6.1(6.3-5.9)###2###5.4###21###6.2(6.3-5.95)###15###6.8(7.1-6.7)

###[greater than or equal to] 140-199

###Diabetic###42(29)###246(269-222)###6.8(7.5-6.35)###1###5.1###2###6.1###19###7.4(8.6-7)

###[greater than or equal to]200

RESULTS

Total of 146 subjects who underwent OGTT for diagnosis of DM, over a period of 16 months, were included in this study. Test of normality showed that data for BMI, waist circumference, FPG, 2-h PG and HbA1c was skewed. Median age of study subjects was 45 (IQR 54.2-35) years, with 38% (n=55) subjects being 40 years. Majority were females 53% (n=77). Subjects with family history of diabetes were 65% (n=95) while 17.8% (n=26) were smokers. Majority of subjects were Sindhis, Punjabis and Urdu speaking or muhajirs (immigrants from India), while Pakhtuns and Balochis were categorized as others, as their number was small as shown in table-I. Median BMI of the subjects was 27kg/m2 (IQR 30.5-24.3) with 17.8% (n=26) were normal with median BMI 21.9 (IQR 22.3-19.9), 32.2% (n=47) being over weight with BMI 25.1 (IQR 26.3-24.5) and 50% (n=73) subjects were obese having BMI 30.5 (IQR 35-28.7).

Median waist circumference was 35.5 cm (IQR 40-33) with 22% (n=32) female [median waist circumference 40cm (IQR 40-38)], while 15% (n=22) males [median waist circumference 42cm (IQR 44-40)] had high waist circumference. Median HbA1c of the study subjects was in prediabetic range; 6% (IQR 6.6-5.6). Positive correlation was observed between age and HbA1c [5.7% (IQR 6.2-5.3) 40y subjects; r 0.34, p=0.000). While amongst genders; males had higher HbA1c level [6.1% (IQR 6.8-5.7) as compared to female with level of 5.9% (IQR 6.4-5.4); r 0.17, p=0.036). No significant difference was observed for HbA1c, FPG, 2-h PG levels, age, BMI and waist circumference amongst ethnicities; as shown in table-I. While table-II shows the frequency of subjects diagnosed as normal, pre-diabetic and diabetic according to the current ADA criteria of FPG and 2-h PG levels in OGTT and corresponding median HbA1c levels in each group according to ADA cut-off.

On basis of HbA1c 46% [n=67; median HbA1c levels 6% (IQR 6.2-5.8)] were labeled as pre-diabetic and 28% [n=41; median HbA1c 7% (IQR 7.55-6.75)] were labeled diabetic. Out of 22 subjects identified as diabetic on FPG, mean HbA1c was impaired in 2 subjects and normal in 1 subject (table-II). One subject out of 22 diabetic on FPG was identified as normal on 2-hr PG value. Among 29% (or 42/146) of the subjects identified as diabetic by 2-h PG; only 71% (or 30/42) patients had HbA1c [greater than or equal to]6.5% while 19% (or 8/42) had impaired and 9.5% (or 4/42) had normal HbA1c. Using 2-h PG as gold standard, sensitivity, specificity, PPV, NPV for HbA1c at cutoff of 6.5% were 70%, 89%, 72% and 80% respectively. While sensitivity, specificity, PPV, NPV of FPGat cutoff of 126mg/dl for diagnosing diabetes mellitus was 67%, 84%, 79% and 73% respectively.

DISCUSSION

Prevalence of DM keeps increasing in this part of the world and screening our population remains a challenge. Additionally, glycemic control is fundamental to the management of diabetes, as optimal control prevents the risk of complications related to uncontrolled DM. Both plasma glucose and HbA1c have been used as a tool for monitoring and predicting risk of developing complications. 2010 ADA guidelines, updated in 2015 recommends that HbA1c can be used for of DM10,11. Since then HbA1c is being increasingly used for screening DM. Overall mean HbA1c is high in subjects screened by OGTT included in this study. Subjects with normal 2-h PG had mean HbA1c of 5.6 +- 0.5%. Compared to these, subjects with impaired glucose tolerance and diabetic had higher HbA1c levels; 6.2% +- 0.6 and 7.1% +- 1 respectively. Higher mean HbA1c levels are also observed in other studies in newly diagnosed diabetic patients as compared to normal subjects14,15.

Mean HbA1c also shows positive correlation with age, suggesting that elderly subjects are at higher risk of developing diabetes16,17. Similar findings are reported by different studies from our part of the world18,19. After stratification based on ethnicities, our results show no significant difference in HbA1c levels. Although literature have reported that HbA1c levels varies in different ethnicities, the exact cause for which is not known. However, no such variation can be observed in this study. Variation in HbA1c in Indian population has been observed, where different HbA1c cut off values for southern and northern Indian had been reported; [greater than or equal to]6.0% vs. [greater than or equal to]6.4% respectively19,20. These findings suggest that larger prospective outcome based studies are required to establish separate HbA1c cut-off for diagnosing diabetes in our population.

There is a considerable debate regarding role of HbA1c as diagnostic marker for screening of DM19,20 Present study also demonstrates HbA1c misdiagnosed 13 cases as compared to 2-h PG. On assessing the HbA1c cutoff level of 6.5% (for labeling DM) using 2-h PG as the reference, showed sensitivity of 70% and specificity of 89%, which is almost similar to FPG. Findings are similar to a study conducted in India which showed sensitivity and specificity of 65% and 88% at cut off level of 6.5% of HbA1c15. Good screening tests should have sensitivity of at least 80% or above so that the suspected cases are not missed on screening. This advocates that using HbA1c cut off >6.5 for our population can miss few cases of diabetes compared to 2-h PG and newer cut offs for HbA1c should be derived for our population.

Similar findings wereobserved by Likhari et al21 who showed that South Asians had higher mean HbA1c level of 6.5% +- 0.7 than Caucasian 6.1 +- 0.6%. In another study by Razak et al22 reported higher mean HbA1c in non-diabetic population among South Asians (5.9%) as compared to Chinese (5.7%) and Europeans (5.4%). These findings advocate that probably different cut offs should be used for HbA1c for our population and it should be used in combination with clinical history and other biochemical tests when diagnosing diabetes. Limitation of this study includes small sample size of different ethnic groups. A larger prospective study is required to confirm these findings. Also in this study HbA1c was performed only once and ideally it should be repeated in subjects with abnormal results to confirm findings.

CONCLUSIONS

Higher median HbA1c levels are seen in patients suspected of DM. Sensitivity of HBA1c is 70% using ADA cut-off of 6.5% using 200mg/dl 2-h PG levels in our pilot study.

CONFLICT OF INTEREST

This study has no conflict of interest to declare by any author.

REFERENCES

1. Vyas Shital I. Risk factors and effect of type 2 diabetes mellitus in North Gujarat community. PhD Thesis published on Shodhgange: a reservoir of Indian theses @ INFLIBNET. http://hdI.handle.net/10603/44302.

2. Khan IA, King H. Pakistan National Diabetes Survey Prevalence of Glucose Intolerance and Associated Factors in North West Frontier Province (NVVFP) of Pakistan. JPMA-Journal of the Pakistan Medical Association, OPMA 1999; 49: 206.

3. Shera AS, Jawad F, Maqsood A. Prevalence of diabetes in Pakistan. Diabetes research and clinical practice 2007; 76(2): 219-22.

4. Bajaj S, Jawad F, Islam N, Mahtab H, Bhattarai J, Shrestha D, et al. South Asian women with diabetes: psychosocial challenges and management: consensus statement. Indian journal of endocrinology and metabolism 2013; 17(4): 548.

5. Zalecenia PT. Guidelines on the management of diabetic patients. A position of Diabetes Poland. Clinical Diabetology 2017; 6(A): 1-80.

6. American Diabetes Association. Standards of medical care in diabetes. Diabetes care 2013; 36(Supplement 1): S11-66.

7. Kalsbeek A, la Fleur S, Fliers E. Circadian control of glucose metabolism. Molecular Metabolism 2014; 3(4): 372-83.

8. Chan H, Lunt H, Thompson H, Heenan HF, Frampton C, Florkowski CM. Plasma glucose measurement in diabetes: impact and implications of variations in sample collection procedures with a focus on the first hour after sample collection. Clinical Chemistry and Laboratory Medicine (CCLM) 2014; 52(7): 1061-8.

9. Berard LD, Blumer I, Houlden R, Miller D, Woo V, Canadian Diabetes Association Clinical Practice Guidelines Expert Committee. Monitoring glycemic control. Can J Diabetes 2013; 37(Suppl 1): S35-9.

10. A Alonso GT, Pyle L, Frohnert B. Change in hemoglobin A1c one year following the 2014 American Diabetes Association guideline update. Diabetes Res Clin Pract 2017; 129: 169-72.

11. World Health Organization. Use of Glycated Haemoglobin (HbA1c) in the Diagnosis of Diabetes Mellitus: Abbreviated Report of a WHO Consultation 2011. Geneva: World Health Organization 2013.

12. Choo V. WHO reassesses appropriate body-mass index for Asian populations. The Lancet 2002; 360: 235.

13. Gill T. Epidemiology and health impact of obesity: an Asia Pacific perspective. Asia Pac J Clin Nutr 2006; 15: 3.

14. Cagliero E, Levina EV, Nathan DM. Immediate feedback of HbA1c levels improves glycemic control in type 1 and insulin-treated type 2 diabetic patients. Diabetes care 1999; 22: 1785-9.

15. Kumar PR, Bhansali A, Ravikiran M, Bhansali S, Dutta P, Thakur JS, et al. Utility of glycated hemoglobin in diagnosing type 2 diabetes mellitus: A community-based study. J Clin Endocrinol Metab 2010; 95: 2832-5.

16. Yang L, Shen X, Yan S, Xu F, Wu P. The effectiveness of age on HbA1c as a criterion for the diagnosis of diabetes in Chinese different age subjects. Clinical endocrinology 2015; 82(2): 205-12.

17. Pani LN, Korenda L, Meigs JB, Driver C, Chamany S, Fox CS, Sullivan L, D'Agostino RB, Nathan DM. Effect of aging on A1C levels in individuals without diabetes: evidence from the Framingham Offspring Study and the National Health and Nutrition Examination Survey 2001-2004. Diabetes Care 2008; 31(10): 1991-6.

18. Ali SK, Khan DA, Khan FA. Determination of glycosylated hemoglobin reference range in adult population attending a military care set up in Rawalpindi. Pak Armed Forces Med J 2011; 61(4): 165-67.

19. Sequeira IR, Poppitt SD. HbA1c as a marker of prediabetes: A reliable screening tool or not. Insights in Nutrition and Metabolism 2017; 1(1): 11-20.

20. Sherwani SI, Khan HA, Ekhzaimy A, Masood A, Sakharkar MK. Significance of hba1c test in diagnosis and prognosis of diabetic patients. Biomarker insights 2016; 11: 95.

21. Likhari T, Gama R. Glycaemia-independent ethnic differences in HbA1c in subjects with impaired glucose tolerance. Diabetic Medicine 2009; 26: 1068-9.

22. Razak F, Anand S, Vuksan V, Davis B, Jacobs R, Teo KK, et al. Ethnic differences in the relationships between obesity and glucose-metabolic abnormalities: A cross-sectional population-based study. Int J Obes 2005; 29: 656-67.
COPYRIGHT 2018 Asianet-Pakistan
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Pakistan Armed Forces Medical Journal
Geographic Code:9PAKI
Date:Oct 31, 2018
Words:3754
Previous Article:FREQUENCY OF ACUTE COMPLICATIONS OF HEMODIALYSIS IN ADULT AGE GROUP AT A TERTIARY CARE CENTER.
Next Article:ROLE OF PANCREATIC STENT PLACEMENT IN PREVENTION OF POST ENDOSCOPIC RETROGRADE CHOLANGIOPANCREATOGRAPHY PANCREATITIS: EXPERIENCE AT A TERTIARY CARE...
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |