A community-based randomized controlled trial that is pragmatic in nature, following a parallel-group design will be conducted. Patients with T2D (N = 200) will be recruited from across 12 pharmacy locations in the province of British Columbia, Canada. The pharmacists will use a secure password-protected website maintained by the University of British Columbia (UBC) Centre for Health Evaluation and Outcome Sciences (CHEOS) to implement the random allocation sequence. The secure website provides allocation concealment until the interventions are assigned. The allocation sequence generated was by a researcher from CHEOS using PROC PLAN in SAS 6.4 software. Participants will be stratified by site (pharmacy) and glucose-lowering medications (two or fewer, three or more (or taking exogenous insulin)) and randomized in a 1:1 allocation ratio using permuted block sizes, to either the therapeutic nutrition group or the treatment-as-usual (TAU) group. Due to the nature of the interventions and the involvement required of the pharmacists, they cannot be blinded to final treatment allocations of the participants. The protocol was approved by the University of British Columbia Clinical Research Ethics Board (H16–01539) and the trial was registered on ClinicalTrials.gov (NCT03181165) on 8 June 2017. A flow diagram of the trial is presented in Fig. 1. Additional file 3 contains the SPIRIT 2013 Checklist.
Primary outcome
The primary outcome is the proportion of patients with T2D who are either using or no longer using glucose-lowering medications after the 12-week study period.
Secondary outcomes
The secondary outcomes are:
Hemoglobin A1C (%)
Percentage dose reduction in glucose-lowering medications (change in daily dose)
Body mass index (BMI) (kilograms/meters squared) and weight reduction (kilograms)
Waist circumference (centimeters)
Body fat percentage (%)
Health-related quality of life (SF-20)
Blood lipid profile (total, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol, triglycerides), liver function tests (alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transpeptidase (GGT)), fasting plasma glucose, pro-insulin, insulin, and C-peptide
Blood pressure (millimeters of mercury)
Percentage dose reduction in blood pressure-lowering medications (change in daily dose)
Participant recruitment
We aim to recruit 200 participants (equal probability of randomization to each arm) using a combination of newspaper advertising, online advertising, recruitment posters posted in pharmacies, and by word of mouth in each respective study site location. Potential participants who respond to the recruitment material will be directed to the nearest study site and invited to attend a screening appointment where the study will be fully explained and discussed. Participants will provide written informed consent and their primary care physician will be informed of their participation prior to any data collection. The pharmacists at each site will obtain consent.
Inclusion criteria
The inclusion criteria are:
Written informed consent
Men and women aged 30–75 years
Diagnosed with T2D by a physician
Using at least one glucose-lowering medication
BMI ≥ 30 kg·m2
Exclusion criteria
The exclusion criteria are:
History of heart attack or coronary artery disease within the previous 2 years
Any current unstable cardiovascular disorder
History of liver disease
History of kidney disease or impaired renal function
Currently pregnant or lactating, or planning on becoming pregnant within the next 12 months
Any diagnosed neurological disorder
History of bariatric surgery
History of cancer within the previous 5 years
Dietary restrictions that would inhibit adherence to the intervention diet
Trial procedures
Eligible participants who enroll in the study will undergo baseline assessment of medications and A1C, fasting glucose, triglycerides, liver ALT, AST, GGT, serum triglycerides, HDL and LDL cholesterol, and high sensitivity (hs)-C-reactive protein (hs-CRP). Weight, height, BMI, body fat percentage (measured by bioelectrical impedance), and waist circumference will be taken as anthropometric measurements, and quality of life will be assessed using the SF-20 questionnaire. Participants’ primary care physicians will be notified of all laboratory results as per typical blood requisition (by fax). Participants will fill out a 3-day diet record to assess habitual food intake and the Godin Leisure-Time Exercise Questionnaire to assess typical physical activity. Participants will also be asked to fill out a questionnaire regarding their ethnicity and socioeconomic status. Following the baseline assessments, participants are randomized to either the Pharm-TCR diet group or the TAU control group. A simplified cost-effectiveness analysis including the cost of medications and cost of the foods and pharmacists’ time during the intervention will be conducted as an exploratory outcome.
Pharm-TCR intervention (0–12 weeks)
Participants randomized to the Pharm-TCR group will meet with a pharmacist and a lifestyle coach at a community pharmacy. The pharmacist will review all medications with the participant and then develop a care plan, including medication recommendations (Additional file 1), to be shared and approved by the participant’s physician (by fax, using existing standard procedures for any medication adjustments recommended by pharmacists). Any adjustments to care made by the physician will be incorporated into the plan. The lifestyle coach or pharmacist will then educate the participant on the components of the intervention and review the Pharm-TCR food options, which are standardized by using a combination of whole foods and supplemental foods from a commercial weight loss diet (Ideal Protein). This commercial diet plan (provided to the participants free of charge) provides a variety of low-carbohydrate, energy-restricted, adequate-protein meals and snacks that are used in combination with a selected group of meats and a wide range of vegetables in order to create a personalized meal plan for each participant. Participants will not otherwise be remunerated for participating in the trial. The macronutrient composition of the meal plan equates to < 50 g of carbohydrates, ~ 35-45 g of fat, and ~ 110-120 g of protein for a total of approximately 850–1100 kcal per day. Pre-packaged foods are used for two meals and one snack each day, with the third meal being prepared from a select group of lower-fat protein sources (e.g., meat, eggs) and low-carbohydrate vegetables. The use of a commercial weight loss diet plan allows for tight control over the amount of carbohydrates consumed and the overall energy intake. This approach was chosen because the consistency makes medication adjustments more predictable and reproducible for the pharmacist. This is essential as the pharmacists will need to manage medication adjustments in anticipation of, and in response to, the rapid glucose-lowering effects of this specific low-carbohydrate low-calorie dietary intervention, even prior to substantial weight loss. For patients who are taking > 30 units of exogenous insulin per day at the start of the study, a modified TCR meal plan is used, which includes 80–100 g of carbohydrates per day. This is to reduce potential risk of hypoglycemia during the insulin de-escalation phase of the study (details below and in Additional file 1). Once a participant’s insulin dose is < 30 units per day they are transitioned to the regular TCR meal plan incorporating < 50 g of carbohydrate per day with remaining insulin dose being reduced by a further 50% on day 1.
Following the initial visit, participants will be required to report to the pharmacy each week to meet with the lifestyle coach and pharmacist in order to monitor progress, collect intervention foods, and receive counseling on dietary adherence for the week ahead (total of 12 visits). At this time, the pharmacist will review the current status of the care plan with the patient, make adjustments as necessary, and communicate as appropriate with their physician. The care plan is individualized to each participant based on their baseline medication usage and blood glucose measurements using a standardized medication deprescribing framework as outlined in Additional file 1. Participants’ medication usage, weight, height, waist circumference, body fat percentage, blood pressure, and capillary ketone levels are also recorded during these visits and a 3-day diet record is filled out again on week 6. On the final visit, participants are assessed for the same blood, anthropometric, quality of life, and habitual physical activity measures that were collected at baseline.
Treatment-as-usual control (0–12 weeks)
Participants randomized to the TAU group will undergo a medication review by the pharmacist, will be provided with standard medication advice by their pharmacist, and will be given information pamphlets on diet and lifestyle conforming with 2013 Diabetes Canada (formerly the Canadian Diabetes Association) Clinical Practice Guidelines. This information is evidence based and recommends participants to follow a low-fat, low glycemic-index diet and to engage in 150 min of moderate activity per week. Standard clinical practice recommends an A1C test and potential medication adjustments every 3 months (typically an increase in dose or number of glucose-lowering drugs if A1C does not reach below the clinical target of < 7.0%) and therefore the treatment period of 12 weeks was chosen.
Three-day food records will be collected at baseline, week 6, and week 12 to assess the energy and macronutrient content of the foods consumed. Following week 12, TAU participants will return to the pharmacy for a final visit where they are assessed again for all baseline blood, anthropometric, quality of life, and habitual physical activity measures.
In order to help retain TAU-group participants for follow-up testing they will be offered participation in the Pharm-TCR intervention after the TAU period has completed. If a TAU-group participant chooses to try the dietary intervention following completion of the TAU period, the participant is considered to have finished the trial and as such, no study data will be collected from them during this time period.
Patient safety
During the first 2–4 weeks of the program (when most medication adjustments are made), participants communicate frequently with their pharmacist and lifestyle coach via phone call and/or text messaging and take finger stick blood glucose measurements four times per day (a fasting measurement upon waking, and before lunch, dinner, and bedtime). This allows for feedback and assessment of the medication reductions and helps to reduce any potential risk of hypoglycemia. Participants are instructed on steps to take if blood glucose is < 3.9 mM and/or they have symptoms of hypoglycemia; the steps include consumption of an energy bar containing 15 g of carbohydrates and, if severe or if symptoms do not resolve, reporting to their doctor or emergency room. All observations/results that may pose a risk to health will be discussed with the study team and all adverse events will be recorded. Pharmacists at each site will be closely engaged with participants due to the nature of the intervention. Any safety concerns will be brought to the attention of the study physician, Dr. Jay Wortman and the lead pharmacist (Mr. Sean McKelvey). A safety committee of two independent physicians with experience in low-carbohydrate diets and T2D will review any adverse events and provide clinical insight into their etiology. These two independent physicians can request stoppage or changes to the ongoing study. The Principal Investigator (PI) (Dr. J. Little) will meet with Dr Wortman, Mr. McKelvey, and the two independent physicians every 6 months or as often as needed, to review any safety issues or data collection issues. If there are any serious adverse events, the independent physicians or study physician can recommend that any/all participants discontinue participation in the trial. Participants may also choose to discontinue at any time. Pharmacists at each site will have access to the data as they are collected. They will fax or email any data, with no personal identifiers, to the PI or study email address at the UBC. Only study investigators will have access to the faxed or emailed data. Any personally identifying data collected by the pharmacists will be de-identified prior to being sent to the research team.
Protocol delivery fidelity
A standard training protocol has been developed to minimize variability and maintain the intervention protocol across all pharmacies. Training videos followed by visits in person by the study coordinator will be used to explain the study design, aims, and protocol to the study personnel at each pharmacy. The study coordinator will remain easily accessible to the pharmacy personnel via email and phone throughout the study. The head study pharmacist teaches the medication adjustment framework to the lead pharmacist from each study site, by in-person workshops that include both a lecture portion and multiple case studies. The head study pharmacist will also be available to the site pharmacists via email and phone throughout the study.
Measurements
The measurements taken at each stage of the study are detailed in the Additional file 2. Height will be measured to the nearest millimeter, using a stadiometer (Seca, model 700, Germany). Body weight will be measured to the nearest 100 g (Tanita, model DC-430 U, IL, USA) with the participant wearing light clothing without shoes. Waist circumference is assessed by measuring the distance around the waist at the top of the iliac crest. Blood pressure will be measured using the PharmaSmart Model PS-2000C (BC, Canada) after the participant has been seated, at rest, with legs uncrossed for at least 5 minutes. Body fat percentage will be assessed by bioelectrical impedance analysis (Tanita, model DC-430 U, IL, USA). Briefly, participants stand on the scale with soles and heels of both of their feet in contact with the electrodes and the participants’ sex, height, and weight are used in combination with the measured resistance to calculate body fat percentage. Tanita has validated this method in people with T2D against dual x-ray absorptiometry (r = 0.89, p < 0.0001) [24].
Participants in the Pharm-TCR group will have their capillary ketone levels measured at each weekly visit using a Freestyle Precision Neo device (Abbott Laboratories, CHI, IL, USA). Clinical laboratory requisitions will be given to the participants for blood to be collected at a local clinical laboratory (LifeLabs or Valley Medical Labs) at baseline and post Pharm-TCR intervention or TAU control, to measure A1C, fasting glucose, triglycerides, ALT, AST, GGT, serum triglycerides, HDL and LDL cholesterol, and hs-CRP. A subsample of participants will have fasting C-peptide, insulin, and proinsulin measured to estimate homeostasis model assessment of insulin resistance (HOMA-IR) and beta-cell function based on HOMA-B and C-peptide–proinsulin ratio. This subsample analysis of the plasma samples collected by the clinical laboratory will be performed by the research team at a university laboratory.
Statistical analysis
Data will be analyzed on an intention-to-treat basis at the 12-week (post-testing) timepoint. The primary outcome measure is binary - using or not using glucose-lowering medication at 12 weeks - and data will be analyzed by logistic regression, with the stratified allocation factors included as fixed effects together with trial arm and sex. We will derive the odds ratio for use of glucose-lowering medication, and odds will be converted to predicted probabilities (risk differences). Participants who fail to complete the intervention or are lost to follow up from the Pharm-TCR group will be retained in the analyses and will be considered as not achieving the primary outcome. The secondary outcome measures of anthropometrics, blood pressure, medication dose changes, quality of life, and all blood measures will be analyzed using constrained baseline longitudinal data analysis via a linear mixed model [25]. This approach provides a principled method for addressing missing data, including missing baseline data as baseline values are part of the response vector in this model. In addition, study site will be included as a random effect in the linear mixed models to account for site variability in the outcome measures [26]. Model specification will be examined by visual inspection of residuals plots, with appropriate data transformations used if indicated. For the primary endpoint we shall report the odds ratios and predicted probabilities (risk differences) with 95% confidence intervals. Secondary continuous outcomes will be reported as mean differences with 95% confidence intervals. Whenever possible, data analysis will be performed blinded to the allocations.
It is not anticipated that any participants in the TAU group will discontinue glucose-lowering medications; however, to be conservative, the sample size was calculated based on an expected proportion of medication use of 80% in the TAU group and a 20% difference in the proportion of participants on glucose-lowering medications at 12 weeks in the Pharm-TCR group. A sample size of 100 per group provides 80% power to detect a 20% difference in the proportion of patients on glucose-lowering medications, accounting for 20% attrition with a two-sided P value of 0.05 (PASS 14 Power Analysis and Sample Size Software, 2015. NCSS, LLC. Kaysville, Utah, USA; ncss.com/software/pass).