Methods:
A convenience sampling will be done for patients diagnosed with diabetes or prediabetes
according to the American Diabetes Association (ADA) criteria, who assist to the Obesity
Clinic and meet the selection criteria. Data from patients that during the study protocol
are called to undergo bariatric surgery procedure will be used until the time of surgery
(intention-to-treat analysis). The prevalence of comorbidities at baseline and the type
and dose of drugs used for treatment will be recorded. Patients will be randomized by a
random numbers system generated with a computational software and will be assigned to a
group: metformin (1,700 mg/day) or metformin (1,700 mg/day) and dapagliflozin 10 mg. All
patients will receive dietary treatment and follow-up during the study by the Nutrition
Service. Once assigned to the corresponding group, patients will receive an
identification code that will be retained throughout the study. One of the researchers
not directly involved in patient care, will assign tablets needed for daily intake for a
month and then the number of tablets required for 3 months in a sealed envelope. A run-in
period will be used to assess tolerance to treatments. This period will be for a month.
At this time patients may notice an increase of uresis and will be instructed to increase
fluid intake. Anthropometric and biochemical variables will be recorded baseline and at
1, 3, 6 and 12 months. The determination of glucagon, ghrelin, adiponectin, resistin,
interleukin 6 (IL-6) and interleukin 10 (IL-10) will be held on initial appointment and
before surgery (depending on the response of each patient). For the determination of
these cytokines and peptides, an ELISA kit (Enzyme Linked Immunosorbent Assay) will be
used. Adherence to treatment will be evaluated and will consist in consumption of 90
percent of pills granted. The patient will be required to return the drug blister. The
registration of adverse events will take place from the start of treatment and throughout
the study. Each event will be evaluated by researchers and classified according to the
Common Terminology Criteria for Adverse Events (CTCAE) version 4.03 list. CTCAE considers
that an adverse event grade 4 or 5 indicates discontinuation of treatment; grade 3 must
be submitted to review by the medical team, and grades 1 or 2 require registration and
intervention but not discontinuation of treatment. Adverse events will be recorded in the
report sheet. An intention-to-treat analysis will be used if patients require treatment
with insulin or sulfonylureas (grade 3 adverse event). They will provide information
until the time of its inception. Patients who do not achieve weight loss at one year
follow-up, will be discarded to their primary care hospital for continue with dietary
recommendations.
Sample Size:
A convenience sample will be performed of patients with obesity class III and prediabetes
or diabetes according ADA criteria at Obesity Clinic who meet selection criteria. The
sample size was calculated using a mean difference formula with data from Zhang et al.
who evaluated weight loss in patients using metformin/dapagliflozin and compared with
placebo. Sample size required is 90 patients: 45 patients in metformin group and 45
patients in dapagliflozin/metformin group. Considering a loss of 20 percent of population
during study, the final sample size required is 108 patients: 54 patients in metformin
and 54 patients in dapagliflozin group.
Statistical analysis Quantitative variables were described using measures of central
tendency and dispersion according to the data distributions. Qualitative variables were
described using frequencies or percentages. The Shapiro-Wilk test was used to establish
normality in the distribution of quantitative variables. Because information on the
different quantitative variables was available up to different time periods, a mixed
model analysis was performed using the restricted maximum likelihood (REML) method, as
well as a mixed model considering the fixed effects of assigned treatment, sex, and
prediabetes or diabetes status, and the random effects of variability in weight loss over
time. Comparisons between groups were analyzed using Sidak's post-hoc test. Pearson's
chi-square test or Fisher's exact test were used to evaluate the association between
qualitative variables. Statistical significance was defined as a p < 0.05. Data analysis
was performed using SPSS version 17.0 and STATA version 11.0.
