The present case control study was conducted in Department of Biochemistry, Indira Gandhi Institute of Medical Sciences, Patna and Andaman and Nicobar Islands Institute of Medical Sciences, Port Blair on serum specimen collected for various diagnostic purposes including fasting and post-prandial plasma glucose from July, 2018 to January, 2019. The study was approved by Institutional Ethics Committee (registration no. ECR/640/Inst/BR/2014/RR-17) and informed consent was taken for both cases and control. Power of the study was 85% associated with 95% confidence interval. Fasting plasma glucose (FPG) was categorised as normoglycemic (80 to 120 mg/dL), mildly to moderately hyperglycemic (121 to 200 mg/dL) and severe hyperglycemic (>200 mg/dL) [7]. A total of 145 subjects were included in the study that were subdivided into three groups; normoglycemic controls (n=54), mild-moderate hyperglycemia (n=44) and severe hyperglycemia (n=47). Hyperglycemic samples were obtained from freshly diagnosed type 2 diabetes mellitus subjects. Individual that came to attend OPD for routine check-up were recruited as controls and were matched with respect to age and gender with hyperglycemic cases. For both controls and hyperglycemic cases, only those subjects were included that aged between 18-65 years with normal liver and renal function tests. Subjects with long term medication, active smokers, alcoholic individuals, any concurrent illness, pregnant females and abnormal liver and renal function tests were excluded from study. A 5 mL of overnight fasting blood was collected in plain vial and serum was separated by centrifugation and stored at -20°C till analysis on following day.

The sample was analysed for total carbohydrate in serum by modified phenol-sulfuric acid method [8]. C-2 hydroxyl group of free or protein linked carbohydrate (PLC) can be detected by phenol-sulfuric acid assay reliably [9,10]. Briefly, 25 μL of serum was precipitated with equal volume of 10% sodium tungstate and 2/3 N H2SO4 and 300 μL distilled water [11]. It was subjected to centrifugation at 5000 rpm for 20 minutes and then supernatant was decanted. Protein pellet was washed thrice with normal saline to remove traces of carbohydrate followed by resuspension in 500 μL of 6M urea/2M thiourea buffer [9,12]. A 100 μL of resuspended pellet was used for measuring PLC using modified phenol-sulfuric acid assay for 2.8 mL assay volume [8] and 25 μL resuspended pellet for analysing protein content using Bradford method at 595 nm for 2.5 mL assay volume [9] using spectrophotometer (Analytical Technologies Limited, Model No. UV-3366Plus). Carbohydrate content in resuspended pellet was calculated in mg per g of protein content in precipitate (designated as CPR) to offset any alteration in serum total protein or alterations in yield during extraction of protein.

Statistical analysis was done using Statistical Package for Social Sciences (SPSS) version V21.0) and Microsoft excel software. Variables in controls were compared with those of cases for significant mean with independent unpaired student t-test and pearson correlation coefficient. Mild-moderate and severe hyperglycaemic cases were categorised into quartiles according to distribution in controls. Risk was estimated using binary logistic regression while odds ratio (OR) was calculated for 4^th quartile of mild-moderate and severe hyperglycemia using interquartile range (IQR) as reference. In all tests of significance, two sided p value have been reported with p-value significant at <0.05.

Mean Fasting Plasma Glucose (FPG) levels were 86.72±9.29 mg/dL in normoglycemia, 164.81±15.59 mg/dL in Mild-moderate Hyperglycemia (MHG) and 256.71±52.67 mg/dL in Severe Hyperglycemia (SHG). FPG levels were significantly increased in MHG and SHG compared to controls (p<0.001 and p<0.001, respectively). Mean Serum Total Carbohydrate (STC) levels were significantly elevated in MHG and severe hyperglycemia SHG compared to controls (243±29.90 mg/dL in MHG Vs 179.42±21.79 mg/dL in controls; p=0.01 and 323.26±60.75 mg/dL in SHG Vs 179.42±21.79 mg/dL in controls; p<0.001, respectively). Protein Linked Carbohydrate (PLC) levels were not significantly increased in MHG compared to controls (146.38±24.67 mg/dL in MHG and 140.09±20.09 mg/dL in controls; p=0.17) while it was significantly higher in SHG than in controls (172.46±54.79 mg/dL and 140.09±20.09 mg/dL in controls; p=0.002). Alterations in mg of Carbohydrate Per Gram of Protein (CPR) were statistically significant both in MHG and SHG compared to controls {50.96±9.12 in MHG Vs 46.92±7.75 in controls; p=0.02 and 59.42±15.84 in SHG Vs 46.92±7.75 in controls; p<0.001). There were no alterations in mean levels of protein content in precipitate between either controls {3.01 (±0.37) g%} and MHG {2.89 (±0.34) g%; p=0.11} or between controls and SHG {2.92 (±0.45) g%; p=0.10) [Table/Fig-1].

A dose dependent relationship (χ² test for trend) was evident between risk of high CPR with SHG (p<0.001) while it was not statistically significant for MHG. When odds ratio (OR) was calculated for 4^th quartile with inter-quartile range (IQR) as reference, OR was statistically significant for CPR in SHG {OR (95% CI)=5.27 (2.09-13.25); p<0.001}, while OR was not statistically significant for MHG (p=0.067) [Table/Fig-2]. When mean levels of serum total carbohydrate, PLC and CPR were compared between MHG and SHG, all were significantly elevated in SHG (p<0.001, p=0.005 and p=0.003, respectively) while protein content in precipitate revealed no significant alterations (p=0.82). Both fasting plasma glucose (FPG) and serum total carbohydrates (STC) were positively correlated with CPR (r=0.36; p<0.001 and r=0.684; p<0.001, respectively) while FPG and STC were also positively correlated (r=0.815: p<0.001) [Table/Fig-3].

[Table/Fig-3]:

Correlation between TSC and CPR.

To the best of author’s knowledge, no study has correlated total carbohydrates linked to serum protein with grades of hyperglycemia using inexpensive assay described in this study. In the current study, levels of STC were significantly increased in MHG and SHG than controls (p=0.01 & p <0.001, respectively). Since MHG and SHG have higher free glucose in serum than controls, it is anticipated that STC levels would be higher in them, termed spurious by Kennedy AL et al. [13]. Consequently little research has been done to estimate STC in diabetes, nevertheless a study on oral squamous cell carcinoma found STC levels to be 97.25±15.16 mg% that are lower than in current study (179.42±21.79 mg%) in controls [14]. It could be due to lack of reference method and uncertain pre-analytical/analytical variables.

There was significant difference in mean levels of PLC between SHG and controls (p <0.002) while no significant correlation was obtained between MHG and controls (p=0.17). Significant alterations in mean value of PLC were also observed between MGH and SHG (p=0.005). Measurement of PLC in re-dissolved protein precipitate can be deluding without associated measurement of protein, as changes in protein levels could nullify or inflate any actual changes in PLC. We also found CPR to be significantly elevated in MHG than in controls (p=0.02) while elevation in SHG vs control was highly significant (p <0.001). Previous studies [15-18] corroborate the result of present study, as Nelson DM et al., measured Glycosylated Serum Proteins (GSP) using affinity chromatography and reported higher fraction of GSP in diabetic patients than controls [15]. Likewise, Kumar A et al., observed highly significant levels in diabetes than control subjects [16] while Lapolla et al., observed significantly higher levels in IGT, NIDD and IDD than in NGT [17]. Besides significant elevation in diabetic subject, GSP levels reported by Beisswenger et al., [18] are much closer to CPR levels in current study than other studies [15,16, 17]. The disagreement on GSP levels in studies [15-18] mentioned earlier may either be due to use of different analytical methodology or by virtue of estimating non-enzymatic protein glycation. The present study has estimated protein linked total carbohydrate i.e., a measure of both glycation and glycosylation leading to variation in CPR level.

It was also observed that an upward trend from 1^st quartile through 4^th quartile for CPR in MHG and SHG but it was statistically significant only for SHG (p<0.001). Likewise OR for 4th quartile compared with IQR as reference was statistically significant only for SHG (OR=5.27; p<0.001) and not MHG (p=0.067). CPR levels are also significantly elevated in SHG compared to MHG (p=0.003). None of the studies discussed so far [15-18] have grouped the subjects in three groups that underline an important observation that both PLC and CPR levels increases with levels of hyperglycemia. FPG, STC and CPR are positively correlated with each other and it implies a strong association between increased FPG or STC with CPR. Hyperglycemia induced alterations in protein structure/function range from non-enzymatic glycation of protein and lipids, receptor/non-receptor mediated effects of AGEs that can accelerate atherosclerotic process, PK-C activation that increases expression of TGF-β leading to thickening of capillary basement membrane and hyperglycemia induced oxidative stress [19]. Besides protein glycation, exemplified by HbA1c and glycated albumin, studies on enzymatically glycosylated proteins have revealed that patients with diabetic complications have higher serum or plasma glycoprotein levels (as haptoglobin, fibrinogen, caeruloplasmin) [20], fibrinogen [21], high globulin but not albumin (as α-1 globulin, α-2 globulin, β-globulin and γ-globulins) [22], Osteoprotegerin [23], glycosylated low density lipoprotein fraction,β2-GPI-LDL and ox-LDL [24]. In particular two state of art recent studies as Phuong et al., analysed serum by LC-MSMS and found that expression of five proteins (haptoglobin alpha 2 chain, alpha-2-HS-glycoprotein, clusterin percursor, Apolipoprotein E percursor, Anti-TNFa antibody light chain) increased significantly [25] while Liu M et al., analysed serum glycan profile by microarray and revealed that relative abundance of glycan pattern exhibited significant differences between diabetic nephropathy and non-diabetic renal disease group [26]. A plausible interpretation of these observations suggest that degree of hyperglycemia is positively associated with glycation/glycosylation of serum protein.

Most studies are focussed on role of glycated haemoglobin, glycated albumin and AGEs in diagnosis or monitoring of diabetes and in predicting diabetic complications while missing the general glycosylation process in tissues or circulation. The utility of HbA1c has been argued in select groups as age and racial groups while measurement of glycated albumin/fructosamine and AGEs lack standardisation, require expensive technique or influenced by variety of factors as altered albumin metabolism [3]. The current study has attempted to address overall glycation and glycosylation process in circulation by measurement of total carbohydrate linked to serum proteins inexpensively and rapidly. The results of present study show that measurement of PLC may be used as an adjunct to measurement of glycated haemoglobin and fasting plasma glucose as it could be conducted at a time between immediate control (self monitoring of blood glucose) and long term control (HbA1c estimation). Since the results may vary among laboratories, reference levels of PLC and CPR needs to be established by individual laboratories.

The present study had limited sample size, thus more elaborative case-controls studies are required for further evaluation of PLC in monitoring of diabetes.

Though glycemic status affects both glycation and glycosylation, studies have focussed mainly on glycated haemoglobin, glycated albumin and AGEs. HbA1c is widely used for monitoring long term glycemic control and self monitoring of blood glucose for immediate glycemic control. Thus there is a need of an assay or biomarker that could assess the overall glycation/glycosylation process in between immediate and long term control. The present study has addressed this issue using methods that are inexpensive and rapid. We found that protein linked carbohydrates (PLC) levels and carbohydrate protein ratio (CPR) were significantly elevated in SHG compared to controls (p<0.001 and p<0.01, respectively) and serum total carbohydrate (STC) was positively correlated with CPR (r=0.68, p<0.001). The alterations in levels of STC, PLC and CPR may be attributed to difference in glycemic status between these two groups. The methods used in current study were inexpensive, rapid and technically less demanding thus suitable for regions having limited resources.