Comprehensive analysis of systemic, metabolic, and molecular changes following prospective change to low-carbohydrate diet in adults with type 2 diabetes mellitus in India

Purpose South Asians, especially Indians, face higher diabetes-related risks despite lower body mass index (BMI) compared with the White population. Limited research connects low-carbohydrate high-fat (LCHF)/ketogenic diets to metabolic changes in this group. Systematic studies are needed to assess the long-term effects of the diet, such as ocular health. Method In this prospective, observational study, 465 candidates aged 25–75 years with type 2 diabetes included with institutional ethics approval. A total of 119 subjects were included in the final study assessment based on the availability of pathophysiological reports, tears, and blood samples collected at baseline, 3rd, and 6th months. Serum and tear samples were analyzed by an enzyme-linked lectinsorbent assay, to examine secreted soluble protein biomarkers, such as IL-1β (interleukin 1 Beta), IL-6 (interleukin 6), IL-10 (interleukin 10), IL-17A (interleukin 17A), MMP-9 (matrix metalloproteinase 9), ICAM-1 (intercellular adhesion molecule 1), VEGF-A (vascular endothelial growth factor A), and TNF-α (tumor necrosis factor-alpha). A Wilcoxon test was performed for paired samples. Spearman’s correlation was applied to test the strength and direction of the association between tear biomarkers and HbA1c. p-value of < 0.05 was considered significant. Results After a 3- and 6-month LCHF intervention, fasting blood sugar decreased by 10% (Δ: −14 mg/dL; p < 0.0001) and 7% (Δ: −8 mg/dL; p < 0.0001), respectively. Glycated hemoglobin A1c levels decreased by 13% (Δ: −1%; p < 0.0001) and 9% (Δ: −0.6%; p < 0.0001). Triglycerides reduced by 22% (Δ: −27 mg/dL; p < 0.0001) and 14% (Δ: −19 mg/dL; p < 0.0001). Total cholesterol reduced by 5.4% (Δ: −10.5 mg/dL; p < 0.003) and 4% (Δ: −7 mg/dL; p < 0.03), while low-density lipoprotein decreased by 10% (Δ: −11.5 mg/dL; p < 0.003) and 9% (Δ: −11 mg/dL; p < 0.002). High-density lipoprotein increased by 11% (Δ: 5 mg/dL; p < 0.0001) and 17% (Δ: 8 mg/dL; p < 0.0001). At the first follow-up, tear proteins such as ICAM-1, IL-17A, and TNF-α decreased by 30% (Δ: −2,739 pg/mL; p < 0.01), 22% (Δ: −4.5 pg/mL; p < 0.02), and 34% (Δ: −0.9 pg/mL; p < 0.002), respectively. At the second follow-up, IL-1β and TNF-α reduced by 41% (Δ: −2.4 pg/mL; p < 0.05) and 34% (Δ: −0.67 pg/mL; p < 0.02). Spearman’s correlation between HbA1c and tear analytes was not statistically significant. Conclusion The LCHF diet reduces the risk of hyperglycemia and dyslipidemia. Changes in tear fluid protein profiles were observed, but identifying promising candidate biomarkers requires validation in a larger cohort.


Introduction
Type-2 diabetes mellitus (T2DM) is one of the most prevalent chronic metabolic diseases.Approximately 537 million adults (20-79 years) are diagnosed with diabetes, and it is estimated that it will increase to 643 million by 2030 and 783 million by 2045.For approximately 116 million diabetic patients, China has the highest contribution.India ranks second with 77 million people, followed by the United States of America with 31 million, indicating that India is one of the most diabetes-risk countries in the coming decades (1,2).Chronic diabetes-related complications are also a significant risk factor for cardiovascular diseases, chronic renal failure, diabetic retinopathy, and several other allied comorbidities (3).The economic burden of T2DM for treatment and management of its complications contributes to approximately 12% of global health expenditure (2).
Several ocular complications include T2DM, for instance, cataract, glaucoma, retinopathy, punctuate keratitis, and recurrent corneal lesions (4,5).Patients with diabetes are known to experience dry eye as a common ocular symptom.Dry eye disease (DED) is considerably more common in diabetics than in healthy individuals, and it is also more common in those with T2DM than in those with type 1 diabetes mellitus (T1DM) (6,7).T2DM is the primary cause of blindness in industrialized countries for individuals aged 25-74 years (8), which is the fourth reason of blindness in developing countries (9).Asia comprises both of the top two nations with the highest number of DM patients, namely, China (116 million) and India (77 million) (10), revealing the decade-long fast economic growth and urbanization in Asia, together with notable dietary and lifestyle changes (11,12).
Genetic and sedentary lifestyles are crucial factors, especially diet, which play a significant role in the pathogenesis of T2DM.Lifestyle factors are reversible and focused on the effort to lower the T2DM-associated risk (13).Several systematic reviews and meta-analyses (SRMA) of randomized controlled trials (RCT) (14)(15)(16) revealed that interventions such as modified diet and/or enhanced level of physical activity can reverse or prevent the onset of T2DM and significantly contribute to the effective management of the disease (7).In RCT and observational studies, low-carbohydrate, high-fat (LCHF) diet or ketogenic diet has revealed promising improvements in glycemic control and weight loss, along with decreases in the quantity and/or doses of antidiabetic medications (17)(18)(19).Low-carbohydrate approaches are supported mainly by the hypothesis that reducing insulin secretion, a critical hormone that creates an anabolic, fat-accumulation state, results in improved cardiometabolic function and induces weight loss (20).This approach has been recently called the carbohydrate-insulin model (21).
People of Indian origin or South Asian origin have a distinct pathophysiological feature with a higher risk of cardiometabolic disorders compared with the White population even though they have lower body mass index (BMI).This phenomenon is known as the "South Asian Phenotype" (21,22).Distinct features such as abdominal adiposity combined with glucose intolerance and dyslipidemia such as high levels of triglycerides (TGs) and low-density lipoprotein (LDL), low levels of high-density lipoprotein cholesterol (HDL), and high levels of TGs relative to HDL with normal BMI (21-23).Studies have also revealed additional features, such as excess body fat per unit BMI, truncal obesity, higher c-reactive protein, and lower adiponectin as part of the South Asian phenotype (23,24).Therefore, a myriad of different diabetes-control diets has gained significant popularity in India, including LCHF diets; however, biochemical or molecular data regarding the performance of such diets have not been adequately described.
Even with established ethnic variations in South Asian body composition and metabolic markers, there is a lack of studies connecting low-carbohydrate or ketogenic diets to metabolic alterations in the diabetic population in India and South Asia.Moreover, to understand and document, ocular changes after a switch over to ketogenic diets, though anecdotally described, require systematic long-term prospective trials.The ketogenic diet or low carbohydrate high fat (LCHF) diet-induced metabolic flexibility, adaptive response of an organism's metabolism to maintain energy homeostasis to utilizing available fuels in situations in the shortage of carbohydrates as fuel, allowing the body to switch from using glucose as the chief fuel to using ketones instead, which has been previously documented in starvation as well as ketogenic diets in cases of epilepsy (25-27), causes a reduction in blood insulin, a rise in glucagon and depletion of visceral as well as subcutaneous fat stores and subsequently, leads to recovery in glycaemic factors.We aim to correlate this proposed metabolic reprogramming with changes in body mass index (BMI), body fat measurements, and metabolic parameters in subjects with T2DM.We further hypothesize that the hyperglycemia-induced elevated inflammatory molecular factors will be reduced by the LCHF diet.Such a reduction in the soluble factors in tear and serum may help understand the role of diabetes-associated inflammation in the progression of peripheral organ-specific complications, such as diabetic retinopathy (DR).

Study design
This was an observational, prospective study performed after it was approved by the Narayana Nethralaya Institutional Ethics Committee (EC reference No: NNIEC/2022/02/01).A written consent form for the study was obtained from the subjects for their participation.For this study, samples from subjects with T2DM adopting the LCHF diet and those receiving standard care were collected.All clinical and disease progression histories were noted before (baseline) and after initiation at 3 and 6 months of home dietary intervention.Subjects with diabetes presenting to the retina clinic were also enrolled in the study (400-600).A total of 465 diabetic subjects were interviewed through counseling and guidance.The subjects with T2DM were treated for their condition as per the standard care with the addition of the LCHF diet.Participants were closely monitored, and their past and current treatment histories were noted in this study.A total of 119 participants and sex-matched diabetic subjects under standard care were finalized and recruited but not on the diet as controls.The inclusion criteria for participants were: person with T2DM, age 18-75 years, and HbA1c > 6.5%, while the exclusion criteria were person with T1DM and had serious diseases such as CKD, IHD, cancers, and DR, and wetting length of <8 mm.

Dietary instruction for the LCHF diet
On recruitment, subjects provided a food log or a list of regularly consumed foods, such as drinks and snacks, to the nutritionist who educated them on carbohydrate metabolism and the role of insulin in lipogenesis and weight gain in simplified terminology.Subjects were recommended to restrict net carbohydrate (total carbohydrates minus fiber) intake to ≤50 g/day or 10-20% of their total calories, which is lower than the guideline of the LCHF diet.The daily recommendation for protein was capped at 20-25% of total calories based on their sex, physical activity level, and ideal body weight.The recommended total fat intake was 65-70% of total calories.Permitted food and beverages include meats, poultry, fish, eggs, low-carbohydrate nuts, seeds, non-starchy (over ground) vegetables, high-fat dairy products, fats, and oils such as olive oil, butter, and coconut oil, and beverages such as water and unsweetened tea or coffee.Sample meals, snack options, and recipes available online were discussed.Subjects were advised to eat only when hungry and avoid eating late at night.No caloric restriction was imposed.All subjects were recommended to drink eight glasses of water per day and encouraged to keep a food log.Food logs were reviewed at subsequent visits to monitor diet adherence.

Clinical examinations
Data were collected longitudinally (0, 3, and 6 months) as part of the follow-up to determine the health status.A general exam was conducted at each visit to document the height, weight, blood pressure (BP), waist, and hip circumference.Following this, a detailed, whole blood serum analysis was performed at each visit.

Sample collections
Blood and tear samples were collected at each visit according to the protocol.Tears were collected and processed according to the standard protocols of the clinical follow-up regime of the institute based on their disease status.
In brief, tear fluid was collected with the help of sterile Schirmer's strips (5 × 35 mm 2 ; Contacare Ophthalmics and Diagnostics, Gujarat, India) on each visit.The strips were positioned one in each of the conjunctival fornices, simultaneously in both eyes.The strips were collected after a sufficient amount of tear fluid (wetting of strips until the 20 mm mark or more) was absorbed and stored in a sterile microcentrifuge tube at −80°C.For extraction of tear fluid, Schirmer's strips were chopped into small pieces; 300 μL of 1x sterile PBS was added and agitated for 2 h at 4°C at 300 rpm and centrifuged to elute the tear fluid.The elute was collected as tear fluid and used for further processing to examine secreted soluble protein factors in tears as described previously (28).

Blood investigations
Blood samples were collected according to the standard protocol at each visit to monitor the metabolic changes.The following tests were performed: fasting blood sugars (FBS), post-prandial blood sugars (PPBS), glycosylated hemoglobin (HbA1C), renal function tests (RFT), liver function tests (LFT), lipid profile, serum insulin, C peptide, erythrocyte sedimentation rate (ESR), C reactive proteins (CRP), and blood ketones.

Assessment of the diabetic medication effect score (MES)
The medication effect score (MES) was assessed by types (category by drug mechanism) and numbers (total number of oral hypoglycemic agents and injected insulins) at each visit (29).
The MES was the overall consumption of anti-glycemic agents.It was calculated by the sum of the median absolute decline of HbA1c times the percentage of the maximum daily dose for each medication, including insulin (30,31).The higher the MES shown, the greater the utilization of medication.The assessment was performed at the beginning and the end of the 3-month RCT study.

Tear evaluations
Tear samples were collected using Schirmer's strips according to the standard clinical protocols, which required 2-4 min.The wetting length was noted, and the strips were immediately taken for testing or stored at −80°C in Eppendorf tubes for further evaluation.

Multiplex ELISA
Using a variant of the multiplex ELISA assay (Biomarker Pathfinder, Novomol-Dx, India and Bio-Techne, USA), IL-1β (interleukin 1 beta), IL-6 (interleukin 6), IL-10 (interleukin 10), IL-17A (interleukin 17A), MMP-9 (matrix metalloproteinase 9), ICAM-1 (intercellular adhesion molecule 1), VEGF-A (vascular endothelial growth factor A), and TNF-α (tumor necrosis factoralpha) were measured.Schirmer's strips were collected in 1.5 mL tubes, in which 300 μL of phosphate buffer solution (extraction buffer) was added.The cartridge was loaded into an analyzer system, which indicated the measured value based on established internal references for each analyte.In the case of blood samples, serum was prepared for each respective sample.In total, 50 μL of serum was used in the same multiplex ELISA, to test changes in systemic levels of these analytes.

Follow-up
Data of anthropogenic factors and blood and tear samples were collected longitudinally (0, 3, and 6 months) as part of the follow-up, to determine the disease and the ocular status.

Statistical analyses
The Shapiro-Wilk test was applied to determine the distribution of the data.As the data were not normally distributed, Friedman test with Dunn's multiple comparison test was performed to determine the significant difference in the parameters between baseline and 3-and 6-month follow-up visits and differences or inter-eye correlation, if any, in the tear fluid levels of analytes between the eyes, at all visits.To study the systemic effect of diet on overall ocular conditions, the average value of the analytes of tear from two eyes of the same subject was used.Since the data were not normally distributed, the repeated measure analysis by mixed-effect model-based analysis of variance (ANOVA) was not applicable.Furthermore, Spearman's rank correlation coefficient was performed to determine the association between the various parameters studied.There was no statistical difference between the values of the two eyes in the cohort.All data are presented as median (interquartile range, IQR) distributed variables.A two-tailed p-value of ≤0.05 was considered statistically significant.Statistical analysis was performed using GraphPad Prism v.8 and MedCalc MedCalc ® Statistical Software version 20.218 (MedCalc Software Ltd., Ostend, Belgium).

Enrollment and participant characteristics
Out of the 465 enrolled participants, 7 dropped out as they did not meet inclusive criteria, and 339 candidates declined to participate in the study.Thus, 119 participants completed 3-and 6-month studies.The flow of the study is presented in Figure 1.In no instance was the primary reason for subject withdrawal attributed to the LCHF study.There were also no adverse events associated with the adoption or maintenance of the LCHF study (Figure 2).

Medication effect score (MES)
Diabetes medication use declined significantly during the 1st follow-up of the LCHF diet, and it remained unchanged at the 2nd follow-up from the baseline.MES decreased at 1st follow-up by 20% (Δ: −0.2; p < 0.02749).Nonetheless, at 2nd follow-up time-point, MES did not change from the baseline (Table 1).
The correlation between HbA1c and tear's secreted soluble protein is evaluated at the baseline and 1st and 2nd follow-up time-points, as shown in Table 3.While none of the tear's analytes showed a correlation with HbA1c at either time-points, 1st fold change (1st follow-up/baseline) revealed that only IL-6 showed a weak negative correlation with HbA1c.Nonetheless, the correlation analysis between tear's analytes and HbA1c showed no statistical significance at 2nd fold change (2nd follow-up/baseline).The levels of the various analytes measured were not significantly different between the right and left eyes of the subject at all the visits.

Concentration and correlation of serum's secreted soluble protein and HbA1c
At 1st and 2nd follow-ups, none of the serum analyte concentrations showed significant modulation from the baseline (Table 2).
The correlation between HbA1c and serum's secreted soluble protein is evaluated at baseline and 1st and 2nd follow-up time-points, as shown in Table 3.At the baseline, only ICAM-1 showed a weak positive correlation with HbA1c; however, at 1st follow-up, none of the serum analytes showed any significant correlation with HbA1c.While at 2nd follow-up, TNF-α has shown a significant and moderately positive correlation with HbA1c.At 1st and 2nd fold changes, serum analytes showed no statistically significant correlation with HbA1c.

Discussion
Hyperglycemia and T2DM have been extensively studied in the past, while a low-carbohydrate diet emerged as an attractive approach to prevent and reverse T2DM-associated health disorders.The results of the present study improve the understanding of the effects of the ketogenic diet/LCHF diet in subjects with T2DM for short-and longterm, more preciously in Indian subjects with T2DM.Consistent with the previous studies, in the South-Indian cohort, subjects with T2DM showed improvement in anthropogenic factors, such as weight, BMI, and WHR at the end of the study.Subsequently, there was an improvement in glycemic factors such as FBS, PPBS, and HbA1c, which corresponded to a reduction in dietary carbohydrate intake and elevation in fat intake.All these observations occurred without changing insulin and ketone concentration at the end of the study.
The possible reason for beneficial health delivery of the LCHF diet on subjects with T2DM could be restriction in the carbohydrate intake (Table 4), which leads to a decline in the absorption of sugar or monosaccharides lowering fasting and post-prandial blood glucose and regulating glucose and fatty acid metabolism (33,34).The current study analyzed subjects at three different time-points, and the results showed a decrease in fasting blood sugar from diabetic range to pre-diabetic range.In addition, PPBS showed a drastic reduction at the end of the study, where approximately PPBS declined by 30%, indicating the effectiveness of the LCHF diet in lowering blood sugar.HbA1c was analyzed to evaluate the long-term effect of the LCHF diet.HbA1c level indicates crucial clinical significance about average plasma glucose of the previous 2-3 months for assessing the status of blood glucose control, which acts as a diagnostic tool for patients with diabetes and a screening test for persons at a high risk of diabetes progression (35).It has been reported that cardiac infractions and microvascular complications decreased by 14 and 37%, respectively, when HbA1c decreased by 1%.Thus, the HbA1c level established important clinical significance in estimating the blood glucose control, uncovering the potential challenges in the treatment and managing the therapeutic schedule (36,37).In this study, a reduction in HbA1c occurred after 3 and 6 months of LCHF diet consumption; the changes appeared at −1% and −0.6% time-points; HbA1c reduced by 16 and 12.6%, respectively.The possible explanation for such small changes could be a wide range of ages; thus, the response of age to treatment cannot be neglected.The average cutoff of HbA1c was 1 in the current study, indicating that diabetes management may also be achieved by ketogenic diet/LCHF diet effects.The ketogenic diet not only improved glucose metabolism, but several studies reported that the ketogenic diet enhanced lipid metabolism.For instance, Hussain et al. (38) showed that a ketogenic diet decreased TG and TC while increasing the HDL-c level significantly, and as a result, it recovers dyslipidemia-associated conditions.In the current study, at 1st and 2nd follow-ups, the LCHF diet showed that TG was reduced by 27 and 19 mg/dL, TC was reduced by 10.5 and 7 mg/dL, LDL-c was reduced by 11.5 and 11 mg/ dL, while HDL-c was increased by 5 and 8 mg/dL, respectively, from the baseline.Similar results have been reported by Dashti et al. (39), where TG decreased by 3.67 mmol/L, TC decreased by 1.88 mmol/L, and LDL decreased by 1.78 mmol/L, while HDL increased by 0.14 mmol/L.Subsequently, the current study showed a reduction in systolic blood pressure by 7% (Δ: −10 mmHg; p < 0.0001) in subjects with T2DM.LCHF diet intake delivered significant positive effects in subjects with T2DM as elevated TG and free fatty acids (FFAs) are pathogenic factors for insulin resistance and oxidative stress and can lead to the escalation of risk for cardiovascular diseases (40,41).Hence, recovery from dyslipidaemia is a positive outcome of the current study, which is not only beneficial for regulating insulin sensitivity but also for managing the incident and progression of diabetes-associated heart diseases (42,43).The diabetes medication was not restricted in the current study.While diabetes medication schedules are often complicated, with multiple agents, modified dosages, and frequent administration, it does not reduce the dependency on medication (29).There have been numerous studies that have explored reasons associated with non-adherence to anti-diabetic medication such as financial conditions, forgetfulness, younger age, education, ongoing diabetes complications, and problems in taking the medications alone (44,45).Hence, diet could be a promising approach to maintaining health, reversing, and dependency on diabetes medication.In the current study, the LCHF diet shows a reduction in MES at 1st follow-up (p < 0.0001), which remained unchanged at 2nd follow-up from the baseline.However, this could be a possible lack of reporting by participants at 2nd follow-up time-point.
Diet has been proposed as one of the most commonly preferred modes to counter diabetes-associated disorders (46).Indian diet is rich in carbohydrates, ~60% (47), which is a high-carbohydrate diet, and plays a significant role in the development of insulin resistance and T2DM (48).Thus, a low-carbohydrate diet was focused on studying the effects of the LCHF diet on the reversibility of T2DM-associated clinical parameters, such as glycemic parameters, lipid profile, and medication score.To monitor the adherence of subjects to a low-carbohydrate diet, we examine macronutrients such as carbohydrates, fat, and protein at respective time-points.The results have shown that carbohydrate intake was reduced significantly from the baseline, while fatty acid intake elevated significantly (Table 4), and those changes were maintained at the end of the protocol.This revealed that subjects compiled the instruction of the diet throughout the study.To observe the effects of a low-carbohydrate diet at a systemic level, glycemic factors such as FBS, PPBS, HbA1c, and serum ketone were analyzed at respective timepoints.At the end of the study, a low-carbohydrate diet reduced glycemic factors at the end of protocol, which supported the significance of a low-carbohydrate diet, consistent with the previous studies (39, 49).Similarly, the lipid profile was monitored to check the dyslipidaemia status of the subject during the study.Several studies have mentioned that dyslipidaemia plays a crucial role in the development of metabolic syndrome and T2DM (50).At the end of the study, total cholesterol, Tear fluid has gained prominence as a promising source for biomarker detection regarding the DR status due to its unique components, ease of collection, vicinity to the disease site, and minimal cell contamination (52,53).Various systemic diseases have also been shown to reflect disease-associated alterations in the tear profiles.Several studies have demonstrated the presence of secreted soluble protein in tears, and alterations in tear quality and quantity have been reported in diabetic patients (54, 55).Furthermore, elevation in tear inflammatory factors has been revealed from various ocular diseases (4,56).For instance, during inflammation, ICAM-1 expression is upregulated in the vascular endothelium of patients with DED (57).Similarly, a study showed that an elevation of protein and messenger RNA expression of ICAM-1 in lacrimal and conjunctival epithelial cells has been reported in patients with DED (58).In contrast, a study mentioned that diabetic mice deficient of ICAM-1 revealed a decrease in the preliminary lesions linked with DR (such as the damage of pericytes, degeneration, and increased capillary permeability) and leukostasis (59).Moreover, the expression of TNF-α has been associated with the pathogenesis of various chronic inflammatory disorders, such as T2DM (60).Diabetes patients have shown a higher concentration of TNF-α in plasma than non-diabetic, where a strong correlation was found between TNF-α and severity of DR (61).Similarly, TNF-α was found higher in diabetic patient's tears while similar results were observed in serum and vitreous (62).A study revealed that the TNF-α concentration in tears escalates with the severity of DR, the concentration being less in non-diabetic participants than in diabetic, and correspondingly, the correlation of TNF-α was found highly significant with DR severity (63).Interestingly, the intravitreal injection of TNF-α inhibitor resulted in lowering of damage of pericytes and capillary degeneration in diabetic mice (64,65).Similarly, TNF-αdeficient mice demonstrated a reduction in vascular changes induced by diabetes (66).In summary, soluble secreted proteins of tears such as inhibition of ICAM-1 and TNF-α exert valuable effects on the prevention of early diabetic retinopathy (59,67).Overall, data propose that inflammation and ocular diabetic complications are interconnected, and that, the concentration of the circulating Despite the above-mentioned limitations, our study reveals information that can be used as potential biomarkers for monitoring studies aiming at personalizing therapies and offering alternative solutions for insulin resistance monitored in advanced obesity and T2DM.
The studies involving humans were approved by the Ethics Committee of Narayana Nethralaya (EC reference No: NNIEC/ 2022/02/01).The studies were conducted in accordance with the local legislation and institutional requirements.The participants provided their written informed consent to participate in this study.Written

FIGURE 1 CONSORT
FIGURE 1 CONSORT diagram.Flow diagram of the study through the two time-points (enrolment, intervention allocation, follow-up, and data analysis).Adapted from Schulz et al. (32) licensed under CC BY-NC 2.0.

FIGURE 2
FIGURE 2Experimental design.Schematic representation of the observational study design.

TABLE 1
Baseline and follow-up characteristics of the participants and changes in anthropometric and biochemistry parameters after LCHF diet intervention # .

TABLE 3
Correlation of tear fluid secreted protein factors level with HbA1c at baseline and follow-up after LCHF diet intervention.