Clinical study design and patient recruitment

We conducted a multi-arm parallel-group randomized study in SA-T2DM of Surinamese descent in The Hague area of the Netherlands between May 2018 and September 2020, recruited at general practitioners’ offices. In this study 3 parallel arms were selected: 1) receiving a diet regime of FMD boxes (Prolon®, L-Nutra Inc., Los Angeles, CA, USA); 2) receiving the dietary supplement EndocalyxTM (MicroVascular Health Solutions LLC, Alpine, UT, USA); or 3) receiving placebo capsules (Fig. 1A). For sample size estimation the primary outcome was determined as improvement of the Microvascular Health Index MVHSdynamic according to a pilot study with 13 healthy volunteers receiving the food supplement EndocalyxTM for 3 consecutive months (supplementary Fig. 1). Eligible patients, with inclusion criteria of age between 18 and 75 years old, self-identified as from South-Asian Surinamese descent, treatment with hypoglycemic drugs for type 2 diabetes and proven albuminuria with an albumin to creatinine ratio (ACR) between 0.3 and 30 mg/mmol in the last 12 months. Moreover, eGFR had to be above 45 mL/min/1.73 m2 (CKD-EPI formula [28]). The study (NCT03889236) was conducted in accordance with the Declaration of Helsinki (October 2013) and approved by the Ethics Committee of the Leiden University Medical Center (LUMC) in agreement with the Dutch law for medical research involving human subjects.

Fig. 1: Study design and patient inclusion.

Study design (A) and CONSORT Flow diagram (B).

Intervention, randomization and blinding

Eligible patients were randomized via randomization envelopes made by the Pharmacy department of the LUMC into the diet-, supplement-, or placebo-arm, after given informed consent (Fig. 1). In agreement with the CONSORT statement, randomisation was performed to provide blinding of the supplement and placebo arm to the participants, care providers and researchers. The Pharmacy department redistributed, labeled and blinded the capsules.

FMD (Prolon®) consisted of a 5-day low protein plant-based diet regime that contained energy bars, vegetable-based soups, kale chips, olives, energy drinks, a supplement and natural tea’s. Day 1 provided 1090 kcal (containing 34% carbohydrate, 56% fat and 10% protein), and days 2 to 5 were identical in formulation and provided 725 kcal (47% carbohydrate, 44% fat and 9% protein). As previous studies with this diet showed beneficial effects after 3 monthly cycles [16], our patients also followed the diet once a month, with a total of 3 cycles in 3 months. Trajectory patients’ satisfaction about FMD intervention was reviewed with the Diabetes Treatment Satisfaction Questionnaire (DTSQ) [29]. EndocalyxTM supplement was produced and provided by Microvascular Health Solutions (Alpine, UT, USA). One capsule contained fucoidan extracted from Laminaria japonica (106.25 mg), glucosamine sulphate (375.0 mg), hyaluronic acid (17.5 mg), a blend of superoxide dismutase and polyphenols (120.0 mg) and stabilizers/bulking agents. The placebo capsules were manufactured by the Pharmacy department of the LUMC and contained microcrystalline cellulose. Patients were instructed to take 4 capsules a day for 3 consecutive months. Patients received no dietary advice and maintained their normal diet during the study.

Data collection

Study visits were conducted at the general practitioner’s office of that specific patient and were executed by the researcher or research assistants. The web based relational database management system Castor Electronic Data Capture (EDC) (https://www.castoredc.com) was used for data storage. Patients were instructed to fast overnight and not to smoke before each study visit. In the first three months, patients had a study visit each month. After completing the first 3 months of the intervention, the interventions were discontinued, and patients had one follow-up study visit at month 6 (Fig. 1A).

At baseline, self-identified ethnicity, age and smoking status were collected. Medical history and medication use was extracted from the personal health records. Microvascular complications were defined as having retinopathy and/or neuropathy. Macrovascular complications were defined as having myocardial infarction, angina pectoris, cerebrovascular accident and/or peripheral artery disease.

Systolic and diastolic BP was measured twice with an automated blood pressure monitor (OMRON, Model M6, Omron Health Care Inc, IL, USA) after patients were sitting calmly for about 5 min. BMI was calculated by dividing the weight (measured with indoor clothing but without shoes) in kilograms by the self-reported height in meters squared. Waist circumference was measured with a measuring tape mid-way between the lower costal margin and the iliac crest. Fasting blood glucose levels were measured with a finger prick blood sample (Accu-chek Aviva, Roche, Basel, Swiss).

Blood samples were collected after overnight fast at the morning of the study visit at baseline, 3- and 6 months through vena puncture. Serum levels of C-peptide, Insulin growth factor 1 (IGF-1), creatinine, high sensitivity C-reactive protein (HsCRP), total cholesterol, high-density lipoprotein cholesterol, triglycerides levels and plasma levels of HbA1c were determined in the central clinical chemistry laboratory of the LUMC using standard assays. Low-density lipoprotein cholesterol was calculated using the Friedewald formula [30]. The CKD-EPI formula was used to estimate the glomerular filtration rate [28].

Plasma HPSE-1 (heparan sulfate, HS degradation) activity (Takara Bio Inc., Shiga, Japan), HYAL-1 [31] (chondroitin sulfate/hyaluronan, CS/HA) and HYAL-4 [32] (CS) activity in-house developed ELISA [33], optimized by use of recombinant active human HYAL-1 or HYAL-4 (7358-GH-020 and 6904-GH-020, Bio-techne, Abingdon, UK). HYAL-4 protein was measured (AMS Biotechnology, Abingdon, UK) according to the manufacturer’s instructions. Plasma levels of ANG2 (DANG20, R&D Systems, Abington, UK) and sTM (850.720.096, Diaclone, Besançon, France) were determined as described [32], and measured according to the protocol supplied by the manufacturer.

On the day of study visit, first morning urine was collected to determine albumin and creatinine (CCL, LUMC), albumin-creatinine ratio (ACR) was calculated and for albumin levels lower than 3.0 mg/mL (displayed as <3.0 by CCL), 2.9 was used for ACR calculation. Urinary HPSE-1 activity and MCP-1 concentration were measured (Takara Bio Inc., Shiga, Japan and R&D Systems Europe, Ltd., Abingdon, UK, resp.) according to manufacturer’s protocols and corrected for creatinine concentration.

Microvascular imaging

Sublingual microcirculation was assessed with SDF-imaging (CapiScope HVCS, KK Technology, Honiton, UK) coupled to the GlycoCheckTM software (Microvascular Health Solutions Inc., Salt Lake City, UT, USA). Image acquisition was automatically mediated through the Glycocheck™ software as described elsewhere [26, 27, 34]. The GlycoCheck™ software detects and extracts the following microvascular parameters: red blood cell velocity(VRBC), perfused capillary density, static and dynamic capillary blood volume (CBV), static and dynamic perfused boundary region (PBR), and the overall microvascular health score (MVHS), validated and described earlier [27, 35].

Glucose monitoring and diet compliance

To minimize the occurrence of hypoglycemia during the diet cycles, dosages of hypoglycemic medications were temporarily altered during the 5 day diet cycle. Sulfonylurea derivatives and short acting insulin were discontinued, long acting insulin 50% reduced, with fasting glucose monitoring on days 6, 7 and 8. Metformin, DPP4 inhibitors, SGLT-2 inhibitors or GLP-1 agonists were continued during the diet cycle. Compliance was checked on the morning of day 5 of the first FMD cycle by measuring fasting ketone body concentration in blood (CareSens Dual, Zkope Healthcare, Sittard, Netherlands) and with ketone sticks in a fresh morning urine sample (Ketostix, Bayer, Leverkusen, Germany). During the cycles patients were contacted by the investigator to check glucose monitoring and compliance with the diet.

Clinical data and resource availability

Most of the data generated or analyzed during this study are included in the published article (and its online supplementary files). The remainder clinical data generated during and/or analyzed during the current study are not publicly available due to hospital privacy restrictions but can be made available as anonymized data from the corresponding author upon reasonable request.

Statistical analysis

The primary endpoint was improvement of microvascular function within 3 months as determined with SDF-imaging (capillary density, CBV, PBR and MVH). Secondary endpoints were improvement in clinical parameters (BP, BMI, waist-to-hip ratio), laboratory markers (ACR, fasting glucose, HbA1c, C-peptide, IGF-1, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, hsCRP, HPSE-1, HYAL-1, HYAL-4, ANG2, sTM and MCP-1). Potential legacy effects were determined using the microvascular and clinical parameters, 3 months after discontinuation (at month 6).

Continuous variables with normal distribution were presented as mean with standard deviation (SD) and variables with skewed distribution as median with 25–75 percentile. Categorical data were expressed as proportions.

Treatment effects within and between groups (diet vs. placebo and supplement vs. placebo, respectively) were investigated with intention to treat analysis by linear mixed models for repeated measurements with Bonferroni post hoc test (values expressed as estimated marginal means (SE) or estimated mean differences with 95% CI). The models were adjusted for age, gender, microvascular and macrovascular history at baseline and hypertension at baseline as this can influence microvascular function. For the per protocol analysis, delta changes within 3 months intervention were compared between the intervention groups with an unpaired t test.

Results from the DTSQ questionnaire (see supplementary data) were compared between FMD and placebo groups with an unpaired t test.

Statistical analysis was performed using SPSS version 25 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism version 8 (Graphpad Inc., La Jolla, CA, USA). A significance level of 0.05 was considered statistically significant.