Correlation Between Values of HbA1c, Fasting and 2 Hour Post Glucose Blood Sugar in a Sri Lankan Adult Population and Their Effectiveness in T2DM Screening

Presentation Number: MON 626
Date of Presentation: April 3rd, 2017

Prasad Katulanda*1, John Chambers2, Ranil Jayawardena1, Suren Madawanarachchi1, Dishan Samarathunga1, Vodathi Bamunuarachchi1, Gayan Madushanka1, Reena Nishadi1, Pathmanesan Pirakash1, Rasangi Anuradhika1, Viduni Wickremasinghe1 and Thilini Gangodawila1
1Diabetes Research Unit, Faculty of Medicine,University of Colombo, Sri Lanka, 2Imperial College, London

Abstract

Fasting plasma glucose (FPG), HbA1c & 2 hour post glucose blood sugar (2 h RBS) are currenly accepted methods of diagnosing diabetes mellitus(DM). FPG has a sensitivity of 40-60% & specificity of >90%. HbA1c has a sensitivity of 83.4% & specificity of 84.4%1 .

A cross sectional study was carried out as a part of preventive study. Adults (40- 70 years) previously undiagnosed of DM, were recruited by random cluster sampling. HbA1c, FPG & 2 h RBS were tested. FPG ≤ 99 mg/dl, 2 h RBS ≤ 139 mg/dl & HbA1c ≤ 5.6% were considered normal. FPG 100 – 125mg/dl & 2 h RBS 140 -199mg/dl were considered as impaired fasting glucose(IFG), while HbA1c 5.7 – 6.4% was considered pre diabetes. FPG ≥ 126 mg/dl, 2 hour-RBS ≥ 200 mg/dl & HbA1c ≥ 6.5 % were considered diabetes.

From 926 participants, 69.3% were females. Mean age was 54 (SD ±9.35) years. From the sample 64.7% (n=599) had normal FPG, 27.9% (n=258) had FPG in IFG range and 7.5 % (n=69) had FPG in diabetes range. 2 h RBS was normal in 48.7% (n=451). It was in DM range in 20.8% (n=193). Percentage who had 2 h RBS within IFG range was 30.5% (n=282). Percentage with normal HbA1c was 44.6% (n=413), 15.6% (n=144) had HbA1c in DM range. Another 39.8% (n=369) had HbA1c in pre diabetes range.

Strong positive correlations were present between HbA1c & FBS (r= 0.829, p<0.01) , HbA1c & 2 h RBS (r= 0.820, p<0.01) & 2 h RBS & FPG. (r= 0.732, N= 904, p<0.01). FPG & 2 h RBS correlated better with HbA1c, when HbA1c was in diabetes range (r= 0.796, N= 144, p<0.01 & r= 0.770, N=144, p<0.01 respectively) than when in pre diabetes range ( r= 0.305, N=369, p<0.01 & r= 0.317, N= 369, p<0.01 respectively).

Relationship between HbA1c, FPG & 2 h RBS was estimated. FPG= (-20.252) + 20.448*HbA1C (R2 = 0.687), 2 h RBS = (-203.025) + 61.701* HbA1C, (R2 = 0.672), HbA1C= 2.550 + 0.034*FPG (R2=0.687). HbA1c values 5.6% & 6.5% corresponded with FPG of 94.25 mg/dl & 112.66 mg/dl & 2 h RBS of 142.50 mg/dl & 198.03 mg/dl respectively. FPG of 99mg/dl & 126mg/dl corresponded with HbA1c of 5.91% & 6.83% respectively.

ROC curve indicated that FPG value 111.5mg/dl had a 68.1% sensitivity & 95.4% specificity. (1-specificity= 4.6). (Area under the curve was 0.868)

2 hour-RBS value 193.5mg/dl had a sensitivity of 82.3% & a specificity of 88.9% (1-specificity =11.1). (Area under the curve was 0.931)

Significant correlations were found between HbA1c & FPG, HbA1c & 2 h RBS and FPG & 2 h RBS. These were more significant for DM range HbA1c. Corresponding 2 h RBS to HbA1c cut off values for diagnosis of DM & pre DM were nearly similar to international standards (2 h RBS 142mg/dl corresponded to HbA1c of 5.6% & 2 h RBS 198mg/dl corresponded to HbA1c of 6.5%). Corresponding FPG values for a given HbA1c were significantly lower (FPG 94mg/dl corresponded to HbA1c of 5.6% & FPG 112mg/dl corresponded to HbA1C of 6.5%). According to ROC analysis 2 h RBS had higher sensitivity compared to FPG in screening of undiagnosed DM.

 

Nothing to Disclose: PK, JC, RJ, SM, DS, VB, GM, RN, PP, RA, VW, TG