Effects of Colchicine in Adults with Metabolic Syndrome: A Pilot Randomized Controlled Trial

Abstract

Aims:

Low-grade inflammation is believed to contribute to the metabolic dysregulation of obesity. Colchicine can reduce inflammation by inhibiting microtubule propagation and inflammasome assembly. We evaluated efficacy and safety of colchicine for improving metabolic and inflammatory outcomes in people with obesity and metabolic syndrome (MetS).

Materials and Methods:

Adults with obesity and MetS, but who did not have diabetes, were randomized to colchicine 0.6mg or placebo capsules twice daily for three months. The primary outcome was change in insulin sensitivity (S_I) as estimated by insulin-modified frequently-sampled intravenous glucose tolerance tests. Secondary outcomes included changes in other metabolic parameters and inflammatory markers.

Results:

Of 40 participants randomized (21 colchicine, 19 placebo), 37 completed the trial. Compared with placebo, colchicine significantly reduced C-reactive protein (p<.005), erythrocyte sedimentation rate (p<.01), white blood cell count (p<.005), and absolute neutrophil count (p<.001). Change in S_I was not significantly different between colchicine and placebo arms (difference: +0.21 ×10⁻⁵; CI −1.70 to +2.13 ×10⁻⁵ min⁻¹ mU⁻¹ mL; p=.82). However, changes in some secondary outcomes, including HOMA-IR (p=.0499), fasting insulin (p=0.07) and glucose effectiveness (p=.08), suggested metabolic improvements in the colchicine vs. placebo group. Adverse events were generally mild and similar in both groups.

Conclusions:

This pilot study found colchicine significantly improved obesity-associated inflammatory variables and demonstrated a good safety profile among adults with obesity and MetS who did not have diabetes. These results suggest a larger, adequately powered study should be conducted to determine whether colchicine improves insulin resistance and other measures of metabolic health in at-risk individuals.

Introduction

Over one-third of adults in the US, 84 million people, have prediabetes.¹ 5–10% of individuals with prediabetes develop diabetes annually, and an estimated 70% of individuals with prediabetes will progress to diabetes during their lifetimes.² Although many new pharmaceutical options have been developed for the treatment of diabetes, limited proven medical therapies exist for prediabetes to prevent diabetes onset.³

Mouse models and human studies suggest that obesity-induced inflammation is an important mechanism promoting insulin resistance and the progression to diabetes.^4,5 Circulating inflammatory effectors, such as saturated fatty acids and uric acid, are found in higher concentrations in obesity and stimulate the innate immune system. The resultant chronic inflammatory state leads to progressive insulin resistance and decreased pancreatic beta-cell reserve.⁶ It has been proposed that suppression of this chronic low-level inflammatory state may slow the onset of diabetes and cardiovascular disease.^7,8

Nod-like Receptor Family Pyrin Domain Containing 3 (NLRP3), a member of the innate immune system, has recently been shown to play an integral role in promoting the inflammatory state in obesity.⁹ Upon stimulation by danger-associated molecular patterns such as saturated fatty acids or cholesterol esters, NLRP3 uses microtubules to synchronize with an adaptor protein and caspase-1 within the cytosol of macrophages to create a multimeric inflammasome.¹⁰ The activated NLRP3 inflammasome produces activated cytokines interleukin (IL)-1β and IL-18, thereby initiating and propagating the inflammatory cascade.^11,12

Increased Nlrp3 expression in adipose tissue is associated with increasing adiposity in mouse models,⁹ while knocking out components of the Nlrp3 inflammasome blocks the development of insulin resistance and reduces inflammatory levels in mice fed a high-fat diet.^12,13 Nlrp3^−/− mice with diet-induced obesity also develop less pancreatic inflammation, fibrosis, and beta-cell death as compared to wildtype littermates.¹⁴ Weight loss in people with obesity and diabetes is associated with improved insulin sensitivity and decreased mRNA expression of NLRP3 and IL1B in subcutaneous adipose tissue.⁹ Taken together, these studies suggest that suppression of the NLRP3 inflammasome in obesity has the potential to improve peripheral insulin resistance as well as beta-cell insulin production.

Colchicine, a microtubule inhibitor, is classically used to suppress or prevent inflammation in disorders such as gout, Familial Mediterranean Fever, and pericarditis.^15,16 One mechanism by which colchicine exerts its anti-inflammatory effects is by inhibiting NLRP3 inflammasome formation and activation.¹⁰ A recent retrospective study suggested that among patients with gout, long-term colchicine treatment may have glycemic benefit;¹⁷ however, to date no RCT has investigated colchicine’s long-term effects on glucose metabolism in adults with obesity and metabolic syndrome (MetS).¹⁸ We hypothesized that administration of colchicine to adults with MetS, but who had not yet developed type 2 diabetes, would improve their obesity-associated metabolic and inflammatory dysregulation.

Materials and Methods

Participants.

Adults (age ≥ 18 years) with obesity (BMI ≥ 30 kg/m²) were recruited through advertisements specifically targeted to this study (seeking individuals with “overweight” and “hypertension”), to attend visits at the National Institutes of Health Clinical Research Center (NIH CRC) in Bethesda, MD between 2014 and 2018. Participants were eligible if they demonstrated evidence of inflammation, defined as high-sensitivity C-reactive protein (hsCRP) ≥ 2.0 mg/L,¹⁹ insulin resistance, as defined by Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) ≥ 2.6,²⁰ and MetS, defined as at least 3 of the following characteristics: 1) Impaired fasting glucose (glucose ≥ 100 mg/dL but <126 mg/dL) or impaired glucose tolerance (glucose ≥ 140 mg/dL but <200 mg/dL at 2 hours of an oral glucose tolerance test); 2) Triglycerides ≥ 150 mg/dL or current treatment for hypertriglyceridemia; 3) Waist circumference ≥ 102 cm for men or ≥ 88 cm for women); 4) High-Density Lipoprotein-Cholesterol (HDL-C) < 40 mg/dL in men, < 50 mg/dL in women, or current treatment; 5) Hypertension: ≥130 mmHg systolic, ≥85 mmHg diastolic, or current treatment.²¹ Exclusion criteria included presence of significant medical illness (e.g. diabetes mellitus, uncontrolled hypertension, congestive heart failure); estimated glomerular filtration rate < 60 mL/min/1.73 m²; recent/current tobacco, nicotine, or illicit substance use; previous history of agranulocytosis, significant myositis, or gout; recent/current use of colchicine or medication known to affect colchicine metabolism/clearance; known allergy to colchicine; recent/current use of other anti-inflammatory medications (e.g. aspirin, other non-steroidal agents, corticosteroids); recent/current use of medication known to affect glucose or weight; change in body weight > 3% in the two months prior to enrollment; for women: irregular menses, pregnancy, breast feeding or planning pregnancy in the next six months; and for women of childbearing potential, being unwilling to use contraception during their participation in the study. A patient was considered as having diabetes mellitus if they met any of the following three criteria: (a) clear clinical diagnosis of diabetes, such as a patient in a hyperglycemic crisis or classic symptoms of hyperglycemia and a random plasma glucose ≥200 mg/dL, (b) two of the following: -b-i fasting plasma glucose ≥ 126 mg/dL, b-ii hemoglobin A1c ≥ 6.5%, or b-iii an oral glucose tolerance test glucose concentration of ≥ 200 mg/dL at 2 hours, or (c) one of the three criteria (b-i.-b-iii.) meeting the diabetes cutoff on two different days.²²

Study Design

The trial was a single-center double-blind, randomized, placebo-controlled parallel groups phase II study. The study protocol was approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Internal Review Board and was registered at ClinicalTrials.gov (NCT02153983). The study was conducted in compliance with the principles of the Declaration of Helsinki. The study was overseen by a Data and Safety Monitoring Board convened by NICHD. All participants provided written informed consent prior to study participation. The study consisted of a screening visit followed by a 3-month randomized double-blind treatment period. For premenopausal female subjects, study investigators aimed to schedule the baseline and final visits in the same phase of the menstrual cycle. Pre-specified stopping criteria for individual participants included erroneous randomization of an ineligible subject, pregnancy, or development of a serious unanticipated complication, including development of any medical problem listed in the exclusion criteria after starting on drug, or development of any verified abnormality in laboratory values or physical status that could be explained by a mild intercurrent illness (or by obesity and insulin resistance itself) as defined in the CTCAE v4.03²³ submitted to the IRB in conjunction with this protocol.

Screening Visit

Subjects were seen as outpatients at the NIH CRC and were asked to come in the fasted state. Measurements were performed including: weight using a calibrated digital scale and height using a stadiometer; heart rate; and blood pressure using an automated sphygmomanometer (Dinamap-Plus, Critikon, Tampa, FL) measured in the seated position after at least 5min rest. An electrocardiogram and laboratory testing (including a fasting plasma glucose, insulin, hemoglobin A1c, hsCRP, lipid panel, complete blood count, and comprehensive metabolic panel) were performed, followed by a standard 75-gram oral glucose tolerance test with glucose measured at the 2h timepoint.

Randomization and Interventions

Participants meeting inclusion/exclusion criteria were randomly assigned in a 1:1 randomization ratio to receive colchicine (Spectrum Chemical MFG Corp, New Brunswick, NJ) or placebo capsules, twice daily. Investigators assigned consecutive code numbers to participants from prespecified lists stratified by race and sex. The NIH CRC Pharmaceutical Development Section and Investigational Drug Management Section used permuted blocks with stratification to generate allocations that translated code numbers into study group assignments by using a pseudo-random number program. An Investigational New Drug application (#120722) to use colchicine USP was approved by the United States Food and Drug Administration. Identically-appearing placebo capsules and colchicine capsules (0.6mg/capsule) were prepared by the NIH CRC Pharmaceutical Development Section or Pine Pharmaceuticals (Tonawanda, NY). NIH Pharmacy personnel not involved with the conduct of the study dispensed study capsules in containers that differed only by participant code number. No participant, investigator, or other medical or nursing staff interacting with participants was aware of study group assignments during the trial.

Once baseline assessments were completed, subjects were prescribed one capsule of study medication by mouth twice daily. We decreased the dose by one capsule for 1 week when participants reported difficulty tolerating study medication and then attempted to increase it back to twice daily. Participants were contacted by telephone one week after study medication initiation and then monthly or as needed to adjust dose. Participants were seen for an interim visit at 6 weeks and for a final study visit after 12 weeks of study medication. No dietary advice was provided during the trial.

Initial Assessment

Subjects who met inclusion criteria were seen as outpatients at the NIH CRC for measurementsincluding: weight, height, heart rate, and blood pressure as described above. Whole body and visceral fat mass was performed by dual-energy x-ray absorptiometry (GE Lunar iDXA, GE Healthcare, Madison WI; software GE enCore 15 with CoreScan algorithm).²⁴ An insulin-modified three-hour frequently-sampled intravenous glucose tolerance test (FSIVGTT) was performed in the morning after an overnight fast as described previously.^25,26 Briefly, two intravenous catheters were placed in the antecubital veins, one for the administration of glucose and insulin and the other for blood sampling from the contralateral arm. A glucose load of 50% dextrose 0.3 g/kg given as a smooth bolus over 2 min was given at time 0 and a bolus of 0.05 U/kg insulin was given just before minute 20. Blood samples were collected at times −15, −10, −5, −1 (averaged as the baseline), then at 2, 3, 4, 5, 6, 8, 10, 14, 19, 22, 25, 30, 40, 50, 70, 100, 140, and 180 minutes after glucose injection for the measurement of plasma glucose and serum insulin concentrations. First phase insulin (AIR_G) was calculated as the insulin area under the curve (using the trapezoidal rule) obtained during the first 14 minutes. Insulin sensitivity (S_I) and glucose effectiveness (S_G) were estimated using minimal model analysis (SAAM II, The Epsilon Group, Charlotte, VA). Disposition Index (DI) was also calculated as AIR_G × S_I. At baseline and follow-up, samples obtained in the fasted state were collected for measurement of glucose, hsCRP, creatinine, creatinine kinase, ALT, AST, total- and HDL-cholesterol, and triglycerides by the NIH CRC clinical laboratory on a Roche (Indianapolis, IN) Cobas 6000 analyzer. Plasma for glucose was collected in tubes containing powdered sodium fluoride. Insulin was measured by an electrochemiluminescent immunoassay using a Roche Cobas e601analyzer. Fasting samples were used to estimate insulin resistance by the HOMA-IR index = insulin (μU/mL) × glucose (mg/dL)/405. Erythrocyte sedimentation rate (ESR) was calculated by the Modified Westergren method.

Subjects were interviewed by a clinician who used a structured questionnaire containing a list of symptoms designed to identify potential adverse drug reactions.²⁷ Adverse events were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) v4.03.²⁸

Outcomes and Follow-up Measures

The prespecified primary outcome measure was change in S_I from baseline to the 3-month timepoint. Secondary prespecified outcomes included changes in fasting insulin, HOMA-IR, the laboratory components of MetS, inflammatory markers, and anthropomorphic measurements.

Participants were seen at six weeks and exchanged their unused study medication for a new supply. We used the tally of returned capsules to assess adherence. Safety laboratories along with an interim history obtained using a structured list of queries was obtained. After three month’s treatment, subjects were reevaluated. Outcome measures obtained at baseline were repeated and efficacy of blinding was assessed via a questionnaire at the end of the treatment period. No interim analyses for efficacy were performed during the study.

Statistical Analysis

The protocol was designed as a pilot study, intended to determine the treatment effect size and standard deviation for colchicine’s effects on markers of metabolic health in subjects with MetS, to allow for determination of an adequate sample size for a larger study should results prove promising. Based on the limited data available from prior studies,^29,30 we estimated that a total sample size of 40 subjects (20 colchicine and 20 placebo) would likely have 80% power (β) to detect a 60% difference in change in insulin sensitivity between colchicine and placebo groups, at level of significance <0.05 (2-sided), allowing for up to 10% loss to follow-up.

All analyses were performed as intention-to-treat for all randomized subjects, using baseline values carried forward if there were missing data at the end-treatment visit. For the prespecified primary outcome and the secondary outcomes, an analysis of covariance (ANCOVA) was used to examine the differences between treatment arms, accounting for age, baseline body fat percentage, change in body fat percentage, and sex. For anthropometric variables, change in body fat percentage was not included in the model. Two-sided significance tests were performed for these analyses. Data were transformed as necessary to maintain assumptions of normality. We examined baseline characteristics by simple t tests or, in the case of categorical data, with exact tests. Chi-square analyses were also used to compare rates of adverse events between groups. SPSS v25.0 (IBM Corp, Armonk, NY) was used for all statistical analyses. As this was a pilot exploratory study, a p value <.05 was deemed significant for the primary and secondary outcomes.

Results

Study Participants

We randomized 40 adults to study drug (colchicine n=21, placebo n=19; Figure 1). There were no significant demographic differences between those who participated and the 9 individuals who declined. Baseline characteristics (Table 1) were not significantly different for participants in the treatment arms. As required for eligibility, participants had significant insulin resistance and elevated inflammatory markers (Table 1), and all met criteria for MetS. A total of 11 women reported being premenopausal (Table 1). At their baseline visit, 7 were studied in the follicular phase and 4 were studied in the luteal phase. At their final visit, 8 women (6 placebo and 2 colchicine) were studied in the same phase as for their baseline-study visit; 3 (2 colchicine and 1 placebo) were evaluated during a different menstrual phase. Among the 40 adults randomized, 37 (18 colchicine; 19 placebo) completed the 3-month study; no subjects left the study or were withdrawn due to adverse events (Figure 1). One subject was withdrawn due to an incidentally found renal mass, and two subjects were withdrawn when found to have exclusionary conditions that were not identified until after randomization. Adherence, as assessed by percentage of prescribed capsules used, was high and not significantly different between groups (colchicine 89% vs placebo 92%; p>.50). The efficacy of blinding appeared excellent; 61% of the colchicine group and 72% of the placebo group stated they believed they had received the active compound (p>.50).

Figure 1.

CONSORT flow diagram of participants during the trial.

Table 1 -.

Baseline Participant Characteristics

Variable†	Colchicine (n=21)	Placebo (n=19)
Age (y)	47.3±13.4	44.4±10.0
Race (n, %)
Black	6, 29%	5, 26%
Non-Black	15, 71%	14, 74%
Sex (Female; n, %)	16, 76%	15, 79%
Premenopausal females (n, %)	4, 19%	7, 37%
Height (cm)	168.4±9.1	167.2±8.2
Weight (kg)	116.3±26.5	119.7±31.2
Body Mass Index (kg/m2)	41.1±9.4	42.5±8.5
Body Fat (%)	48.9±4.2	49.0±5.9
SI (×10−5 min−1 mU−1 mL)	9.57±0.88	9.88±1.0
SG (×10−4 min−1)	147.9±44.0	161.7±36.2
AIRG (mU/mL min)	1209.4±655.1	1973.6±2985.1
DI (×10−2)	10.4±6.3	14.4±11.6
AUCG 0–19 minutes (mg/dL min)	4217.5±610.2	4483.8±508.4
AUCG 0–300 minutes (mg/dL min)	31387.9±3184.1	31476.4±2675.2
Fasting Glucose (mg/dL)	105.5±9.6	100.9±7.1
Fasting Insulin (μIU/mL)	26.3±11.3	24.2±10.7
HOMA-IR	6.9±0.7	6.0±0.6
Hemoglobin A1c (%)	5.6±0.1	5.5±0.1
hsCRP (mg/L)	8.1±7.4	6.7±4.2
ESR (mm/hr)	20.6±16.5	21.0±12.8
WBC (×1000/μL)	7.38±1.82	6.41±1.73
Neutrophils (×1000/μL)	4.41±1.47	3.52±1.15

^†Data are reported as mean ± SD except where otherwise indicated. S_I, insulin sensitivity; S_G, glucose disposal; AIR_G, acute insulin response to glucose; DI, Disposition Index; AUC_G, glucose area under the curve; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; hsCRP, high sensitivity C-reactive protein; ESR, erythrocyte sedimentation rate; WBC, white blood cell count. There were no significant demographic or anthropometric differences between groups.

Primary Endpoint

The effect of colchicine on S_I was not significantly different from that of placebo (mean±SD for change in S_I: colchicine 0.29 ×10⁻⁵±3.14 ×10⁻⁵ vs placebo 0.34 ×10⁻⁵±2.90×10⁻⁵ min⁻¹ mU⁻¹ mL; difference 0.21 ×10⁻⁵, 95%CI −1.70 ×10⁻⁵ to 2.13 ×10⁻⁵; p=.82; Figure 2A). S_I as estimated by the FSIVGTT did not significantly change in either group during the study.

Figure 2.

Metabolic and inflammatory changes after three months of study medication in participants randomized to colchicine (N=21) or placebo (N=19). A: Insulin sensitivity (S_I); B: fasting glucose; C: fasting insulin; D: Homeostasis Model Assessment of Insulin Resistance (HOMA-IR); E: high sensitivity C-reactive protein (hsCRP); F: erythrocyte sedimentation rate (ESR); G: white blood cell count (WBC); H: neutrophil count; and I: platelet count. Data are presented as mean ± SEM.

Secondary Endpoints

Colchicine significantly decreased inflammation compared to placebo. High-sensitivity C-reactive protein (hsCRP) decreased by 2.8±2.9 mg/L in the colchicine arm and increased by 0.4±2.8 mg/L in the placebo arm (difference −3.3, 95%CI −5.2 to −1.3 mg/L; p=.0015; Table 2; Figure 2E). Similarly, colchicine decreased ESR as compared to placebo (difference −5.9, 95%CI −10.1 to −1.7 mm/hr; p=.007). Significant decreases in WBC, absolute neutrophil count (ANC), absolute monocyte count, and platelets were also seen in the colchicine versus placebo group (all p’s<.005; Table 2; Figure 2). There were no significant between-group differences in other measured circulating leukocyte populations, including lymphocytes, basophils, or eosinophils.

Table 2 -.

Changes in Outcome Measures Among Treatment Groups

Change in Variable from Baseline†	Colchicine; N=21	Placebo; N=19	Difference	p-value
Metabolic
Δ SI (×10−5 min−1 mU−1 mL)	+0.41 (−0.90 to +1.72)	+0.20 (−0.12 to +1.58)	+0.21 (−1.70 to +2.13)	.82
Δ SG (×10−4 min−1)	−1.5 (−23.3 to +20.3)	−29.7 (−52.7 to −6.8)	+28.3 (−3.6 to +60.1)	.08
Δ AIRG (mU/mL min)	−34.0 (−251.5 to +183.6)	−138.0 (−366.9 to +90.9)	+104.0 (−214.3 to +422.3)	.51
Δ DI (×10−2)	+0.4 (−1.2 to +1.9)	−1.4 (−3.1 to +0.2)	+1.8 (−0.5 to +4.1)	.13
Δ AUCG 0–19 minutes (mg/dL min)	+6.5 (−175.5 to +188.4)	+106.1 (−96.2 to 308.4)	−99.6 (−338.6 to +139.4)	.40
Δ AUCG 0–300 minutes (mg/dL min)	+662.1 (−247.0 to 1571.1)	+1038.2 (+27.4 to +2049.0)	−376.1 (−1570.3 to +818.1)	.53
Δ Fasting Insulin (μIU/mL)‡	+0.0 (−3.7 to +3.6)	+3.7 (−0.1 to +7.6)	−3.8 (−9.2 to +1.6)	.07
Δ Fasting Glucose (mg/dL) ‡	−1.2 (−4.9 to +2.6)	2.6 (−1.4 to +6.5)	−3.7 (−9.2 to +1.7)	.17‡
Δ HOMA-IR‡	−0.3 (−1.1 to +1.1)	1.1 (−0.1 to +2.3)	−1.1 (−2.7 to +0.5)	.04998‡
Δ Hemoglobin A1c (%)‡	−0.02 (−0.16 to +0.11)	+0.05 (−0.10 to +0.19)	−0.07 (−0.27 to +0.13)	.41‡
Δ Triglycerides (mg/dL)	+11.2 (−3.8 to +26.1)	+2.0 (−13.7 to +17.7)	+9.1 (−12.7 to +31.0)	.40
Δ LDL-C (mg/dL)	−2.7 (−9.3 to +3.9)	−3.9 (−10.9 to +3.1)	+1.2 (−8.5 to +10.9)	.80
Δ HDL-C (mg/dL)	−2.0 (−3.9 to −0.2)	−0.6 (−2.5 to +1.3)	−1.4 (−4.1 to +1.2)	.28
Δ Total Cholesterol (mg/dL)	−2.6 (−10.7 to +5.5)	−2.1 (−10.6 to +6.3)	−0.5 (−12.3 to +11.3)	.94
Anthropometric
Δ Weight (kg)	−0.1 (−1.1 to +1.0)	+1.1 (+0.1 to +2.3)	−1.2 (−2.8 to +0.3)	.11
Δ Body Mass Index (kg/m2)	0.0 (−0.4 to +0.3)	+0.4 (0.0 to +0.8)	−0.4 (−1.0 to +0.1)	.10
Δ Fat mass (%)	+0.3 (−0.2 to +0.7)	+0.1 (−0.4 to +0.5)	+0.2 (−0.4 to +0.8)	.55
Inflammatory
Δ hsCRP (mg/L)	−2.8 (−4.1 to −1.5)	+0.5 (−0.9 to +1.8)	−3.3 (−5.2 to −1.3)	.0015
Δ ESR (mm/hr)	−5.4 (−8.3 to −2.5)	+0.5 (−2.6 to +3.5)	−5.9 (−10.1 to −1.7)	.007
Δ WBC (×1000/μL)	−1.09 (−1.7 to −0.5)	+0.3 (−0.29 to +0.89)	−1.39 (−2.21 to −0.56)	.002
Δ Neutrophils (×1000/μL)	−0.91 (−1.30 to −0.52)	+0.12 (−0.28 to +0.53)	−1.03 (−1.60 to −0.47)	.0007
Δ Monocytes (×1000/μL)	−0.06 (−0.11 to −0.01)	+0.06 (0.00 to +0.11)	−0.12 (−0.20 to −0.04)	.0045
Δ Lymphocytes (×1000/μL)	+0.07 (−0.07 to +0.21)	+0.11 (−0.04 to +0.26)	−0.04 (−0.25 to +0.17)	.70
Δ Platelets (×1000/μL)	−15.2 (−23.7 to −6.8)	+5.1 (−3.8 to +14.0)	−20.3 (−32.6 to −8.0)	.002

^†Data reported as estimated marginal means (95% CIs), adjusted for covariates as described in Methods.

^‡For log transformed variables, calculated as log₁₀ (end study value) - log₁₀ (baseline value), the untransformed data are shown for clarity even though the data required transformation for the statistical analysis. S_I, insulin sensitivity; S_G, glucose disposal; AIR_G, acute insulin response to glucose; DI, Disposition Index; AUC_G, glucose area under the curve; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; LDL-C, low density lipoprotein cholesterol; HDL-C, high density lipoprotein cholesterol; hsCRP, high sensitivity C-reactive protein; ESR, erythrocyte sedimentation rate; WBC, white blood cell count. P-values shown for ANCOVA analyses examining effect of group on change in variable as described in Methods section.

The colchicine group had an improvement of HOMA-IR (p=0.0499) and a trend towards greater improvement in fasting insulin (p=0.07) and S_G, the FSIVGTT-derived metric of non-insulin mediated glucose disposal (p=.08; Table 2; Figure 2). No significant group differences were seen for change in first-phase insulin response (AIR_G), hemoglobin A1c, or total, HDL- or LDL-cholesterol (p’s>.05). Changes in anthropometric measures were also not significantly different between treatment arms (Table 2).

Post-hoc Power Analyses

Using data from this trial, we calculated the necessary sample size to achieve 80% power to detect a difference in measures of metabolic health between colchicine and placebo groups, for a two-tailed α <0.05 (Supplemental Table 1). A sample size of <160 subjects was estimated as sufficient to evaluate fasting insulin, fasting glucose, HOMA-IR, S_G, or Disposition Index.

Safety Endpoints

Rates of adverse events were not significantly different between groups during the course of the study (Table 3). Most adverse events were mild (Grade 2 or less), and all spontaneously improved/resolved either during the study or after ceasing study drug. Anemia (Grade 1) developed in one subject taking placebo. One subject taking colchicine developed leukopenia (Grade 2) and neutropenia (Grade 2), while two colchicine subjects and one placebo subject developed thrombocytopenia (all Grade 1). Creatine kinase elevations developed in three colchicine (Grade 1: n=2, Grade 2: n=1) and five placebo subjects (Grade 1: n=3, Grade 2: n=2). Four colchicine (Grade 1: n=2, Grade 2: n=2) and two placebo (Grade 1: n=1, Grade 3: n=1) subjects developed an elevated ALT, while five colchicine (Grade 1: n=5) and three placebo (Grade 1: n=2, Grade 2: n=1) subjects developed an elevated AST. One subject given placebo developed a Grade 3 transaminitis (ALT elevation), which improved spontaneously. No subject’s estimated GFR fell below 60 ml/min/1.73 m² during the study.

Table 3 -.

Adverse Events

Variable†	Colchicine	Placebo	p-value
Subject Reported
Gastrointestinal	13 (14.0)	14 (15.9)	.72
Upper Respiratory Tract	6 (6.5)	11 (12.5)	.17
Fatigue	6 (6.5)	5 (5.7)	.82
Headache	3 (3.2)	4 (4.5)	.65
Neurologic	3 (3.2)	3 (3.4)	.94
Rash	2 (2.2)	2 (2.3)	.96
Laboratory Test Results
Anemia	0 (0)	1 (5.3)	.33
Thrombocytopenia	2 (11.1)	0 (0)	.16
Leukopenia	2 (11.1)	0 (0)	.16
Neutropenia	1 (5.6)	0 (0)	.33
Elevated CK	4 (10.8)	6 (16.2)	.50
Elevated ALT	7 (18.9)	2 (5.4)	.08
Elevated AST	6 (16.2)	4 (10.8)	.50

^†Data shown as n (% of visits) where adverse event was reported. There were no significant demographic or anthropometric differences between the groups. Most AE were minor (CTCAE v4.03 Grade 1 or 2); one subject taking placebo developed a Grade 3 ALT elevation which spontaneously improved. CK, Creatinine Kinase; ALT, alanine aminotransferase; AST, aspartate aminotransferase

Discussion

Results from this pilot randomized controlled trial demonstrate that colchicine was well tolerated by adults with obesity, MetS, and evidence for inflammation. Colchicine significantly reduced multiple markers of obesity-associated inflammation, including hsCRP and ESR. Colchicine is well-known to have anti-inflammatory properties, although its effect on obesity-associated inflammation has not previously been investigated. Classically it has been posited that colchicine blocks inflammation by impeding leukocyte locomotion, diapedesis, and, ultimately, recruitment to sites of inflammation.³¹ Colchicine may exert its effects through other mechanisms as well, including blocking the production of chemotactic and adhesion molecules.^32,33 Recently, it has been shown that colchicine also inhibits the formation of the NLRP3 inflammasome, an important component of the obesity-associated inflammatory cascade.^10,34 In the current study, participants in the colchicine arm had moderate but statistically significant reductions in WBC, monocytes, neutrophils, and platelets, without significant effects on lymphocyte count. Further work is needed to investigate specifically which inflammatory pathways are suppressed, and which remain unaffected, by colchicine in obesity.

The suppression of inflammation from colchicine treatment was not associated with significant improvement in our primary outcome, S_I estimated from the FSIVGTT. However, some of the secondary outcomes related to glucose homeostasis, including HOMA-IR and fasting insulin, had changes that suggest colchicine treatment may potentially improve hepatic³⁵ insulin sensitivity. Additionally, the trend towards improvement in DI in the colchicine group suggests that colchicine might potentially be able to delay the onset of diabetes mellitus in at-risk individuals. However, larger follow up studies will be needed to confirm and extend these findings.

Previous small, short-term prospective studies in healthy adults had suggested that colchicine might actually worsen metabolic parameters by inhibiting insulin secretion by unclear mechanisms.^36,37 However, more recent retrospective studies of individuals with chronic inflammatory conditions found that long-term colchicine use did not have negative effects on insulin secretion or glycemic control,³⁸ and might potentially have metabolic benefits.¹⁷ Similarly, our results confirmed that chronic colchicine use does not impair first-phase insulin response or insulin sensitivity. Other markers of metabolic health such as glycosylated hemoglobin and cholesterol were not significantly changed by colchicine versus placebo treatment. It is unclear whether these non-significant differences were due to lack of efficacy or inadequate power.

Previous studies of immunomodulatory agents for cardiometabolic diseases in humans have also produced mixed results. Anakinra, a recombinant human IL-1 receptor antagonist, has demonstrated an improvement in beta-cell secretory function and decrease in Hemoglobin A1c in adults with diabetes,³⁹ but this was not seen in those with impaired glucose tolerance.⁴⁰ Blockade of TNFα had no significant effects on metabolic parameters in participants with T2DM⁴¹ and only marginal impact in adults without diabetes but who had insulin resistance.⁴² Aspirin or the IL-1β antibody canakinumab also had unimpressive metabolic effects in at-risk individuals.^43–45 However, a recent large cardiovascular trial found that canakinumab did successfully decrease hsCRP and the primary composite end point of nonfatal myocardial infarction, nonfatal stroke, or cardiovascular death in individuals with prior myocardial infarction.⁴⁶ Similarly, the LoDoCo study found that subjects with known coronary artery disease randomized to colchicine 0.5mg daily had fewer composite cardiovascular events than the placebo arm.⁴⁷

As obesity-induced inflammation involves multiple complimentary and redundant inflammatory pathways, blocking a single cytokine (e.g. IL-1β or TNF-α) may not be sufficient to induce clinically significant metabolic improvements. However, colchicine affects multiple pro-inflammatory cell types, cytokines, and pathways activated in obesity.^18,48 Although we hypothesize that colchicine can improve metabolic dysregulation by its ability to impair NLRP3 inflammasome activation, its ability to block neutrophil diapedesis, promote M2 macrophage differentiation, reduce chemotactic and adhesion molecule production, and suppress superoxide production potentially all contribute to its anti-inflammatory and potentially pro-metabolic effects.^10,31,49

Strengths of this study include the prospective, randomized, double-blind, placebo-controlled trial design; participants’ excellent adherence to study regimen; and the small number of participants who did not complete the trial. The relatively small sample size limited the chance to see significant effects from colchicine on most metabolic measurements and precluded adjustment of significance tests for comparisons among the secondary study outcomes. However, as this investigation was designed as a pilot, it enabled us to calculate that a reasonable sample size (<160 participants) would allow an adequately-powered examination of relevant study endpoints. Another possible weakness is that we did not utilize the FDA-approved version of colchicine available at the time on the US market (Colcrys); rather we used colchicine USP powder in study capsules at the FDA-approved dose so that we could produce an identical-appearing placebo. Because we did not avoid enrollment of individuals with extremely high BMI, it is possible that results might have been different in a sample restricted, for example, to BMI < 40kg/m². Additionally, as the phase of the menstrual cycle can affect insulin sensitivity, we attempted to schedule baseline and final visits in the same self-reported phase for the 11 premenopausal female subjects. While we were largely successful in this effort, we did not have objective evidence (e.g. measured serum progesterone) to confirm that this was the case. Lastly, the subjects studied had only insulin resistance or prediabetes rather than diabetes. It is possible that a study of a more metabolically-unhealthy group might find greater difference between placebo and colchicine. Future studies are also needed to characterize more completely the metabolic and anti-inflammatory effects of colchicine administration in MetS, including measurements of IL-1 beta and IL-18 concentrations.

In conclusion, in this pilot study, colchicine successfully suppressed obesity-induced inflammation but did not significantly improve the study’s primary outcome. However, several metabolic measures did suggest improvement in glucose homeostasis in the colchicine arm. Larger trials are needed to investigate whether colchicine has efficacy in improving insulin resistance and/or preventing the onset of diabetes mellitus in at-risk individuals with obesity-associated inflammation.