Dyslipidemia, often found in type 2 diabetes mellitus (T2DM) patients, plays an important role in the progression of cardiometabolic syndrome. Two essential nutrients, chromium and biotin, may maintain optimal glycemic control. The authors report here a randomized, double-blind placebo-controlled trial (N=348; chromium picolinate and biotin combination [CPB]: 226, placebo: 122; T2DM participants with hemoglobin A1c [HbA1c] ≥7%) evaluating the effects of CPB on lipid and lipoprotein levels. Participants were randomly assigned (2:1 ratio) to receive either CPB (600 μg chromium as chromium picolinate and 2 mg biotin) or a matching placebo once daily for 90 days. Statistical analyses were conducted in all eligible participants. Subsequent supplemental analyses were performed in T2DM participants with hypercholesterolemia (HC) and in those using stable doses of statins. In the primary analysis, CPB lowered HbA1c (P<.05) and glucose (P<.02) significantly compared with the placebo group. No significant changes were observed in other lipid levels. In participants with HC and T2DM, significant changes in total cholesterol and low-density lipoprotein cholesterol (LDL-C) levels and atherogenic index were observed in the CPB group (P<.05). Significant decreases in LDL-C, total cholesterol, HbA1c , and very low-density cholesterol levels (P<.05) were observed in the CPB group taking statins. CPB treatment was well tolerated with no adverse effects, dissimilar from those associated with placebo. These data suggest that intervention with CPB improves cardiometabolic risk factors.
Cardiometabolic syndrome (CMS) is highly correlated with the risk of developing type 2 diabetes mellitus (T2DM) and cardiovascular disease (CVD).1 Epidemiologic studies show a correlation between CMS and an increased risk of coronary heart disease (CHD), ischemic stroke, cardiovascular mortality, and total mortality.2,3 Managing CMS involves both lifestyle and pharmacologic interventions designed to prevent, delay the onset of, or treat the individual components of the syndrome.
Insulin resistance is thought to be a major contributor to the metabolic dysregulation that leads to CVD.4 Lifestyle intervention involving control of dietary patterns, including the addition of dietary supplements, in conjunction with increased physical activity is highly effective in the management of the modifiable CMS risk factors associated with CVD and T2DM. The main goal of treating T2DM includes not only normalization of hyperglycemia but also reduction of hypertension and correction of dyslipidemia.5,6 Directly targeting metabolic risk factors for treatment is a relatively new approach for treating T2DM, however.
The National Cholesterol Education Program (NCEP) classifies high cholesterol as a total cholesterol (total-C) level >200 mg/dL and/or low-density lipoprotein cholesterol (LDL-C) >130 mg/dL. It is estimated that over 50 million Americans need treatment for high cholesterol; in other words, 1 in 5 people are at serious risk.7 Medical nutritional therapy is an alternative approach for managing CMS risk factors.
Chromium (Cr+3) is an essential mineral that appears to be an integral element in glucose homeostasis and insulin signaling.8 A recent double-blind placebo-controlled study reported that intervention with chromium picolinate (CP) resulted in significantly lowered hemoglobin A1c (HbA1c), fasting glucose, and free fatty acid levels and significantly reduced both central and visceral adiposity in patients with T2DM.9 Biotin has been reported to enhance the activity of glucokinase, a pancreatic enzyme involved in glucose uptake in the liver, and subsequently lower blood glucose levels and increased insulin release from pancreatic β cells.10 In vitro and in vivo studies suggest that chromium picolinate and biotin combination (CPB) enhances glucose uptake and improves glycemic control and insulin sensitivity. In an open label program, CPB reduced HbA1c levels in a high-risk group (HbA1c≥8%) of patients with T2DM when used as an adjuvant to standard of care therapy that included diabetes education and oral antidiabetic medications (OADs).11 In a recent double-blind placebo-controlled study, supplementation with CPB for 30 days significantly reduced plasma fructosamine, fasting glucose, and the area under the curve for glucose levels after an oral glucose tolerance test, when compared with placebo.12
In this well-controlled study, we prospectively assessed the effect of CPB on coronary risk factors in patients with T2DM and comorbid T2DM and hypercholesterolemia.
Participants and Methods
Study Design. Men and women aged between 18 and 70 years with a diagnosis of T2DM (≥1 year) that was poorly controlled (HbA1c≥7%) and who were on a stable dose of OADs (stable, ≥60 days) were eligible for participation in this study. During this study, insulin use was allowed only for rescue purposes. Table I provides a list of the major concurrent medications used by the participants in the trial. The study design is depicted in Figure.
Table I. Concurrent Medications
Placebo (n=122), %
Active (n=226), %
Rescue insulin (frequency ≤1/wk)
Individuals taking any supplement containing CP within the last 90 days or taking a supplement/multivitamin containing any other form of Cr+3≥120 μg/d within the last 30 days were ineligible for enrollment. Inclusion criteria included the following: (1) diagnosis of T2DM ≥1 year; (2) HbA1c level ≥7%; (3) body mass index (BMI) ≥25 kg/m2 and <35 kg/m2; and (4) fasting triglyceride (TG) level <400 mg/dL. Exclusion criteria included the following: (1) diagnosis of type 1 diabetes; (2) hypoglycemic event requiring intervention by emergency medical services ≤12 months; (3) diabetic ketoacidosis ≤12 months; (4) creatinine level ≥2× upper limit of normal (ULN), aspartate aminotransferase or alanine transaminase level ≥2× ULN, total bilirubin level ≥1.5× ULN; (5) any cardiovascular conditions requiring hospitalization within the previous 12 months; (6) history of cerebrovascular accident, pulmonary embolism, or an unresolved deep venous thrombosis; (7) uncontrolled high blood pressure; (8) history of any serious immunosuppressive disorder; (9) hepatic disease, impaired thyroid, impaired renal function, or other diseases known to affect glucose or lipid metabolism; (10) diagnosed or self-reported alcoholism or substance abuse problems; (11) any psychiatric or mental health issue; (12) any illness or complication factor that in the opinion of the investigator would jeopardize the participant's health; and (13) participation in any other clinical research trial for any product or device within 30 days before enrollment. Women who were pregnant or nursing were also ineligible for enrollment.
Participants. This randomized, double-blind placebo-controlled study was conducted at 17 geographically diverse sites in the United States. Two diabetes-related disease management groups (WellMed Medical Management and XLHealth) participated in the trial, allowing use of their physician-based facilities as recruitment centers for the study. Participants were recruited from the principal investigator's database, referrals from area physicians, or via institutional review board-approved advertisements. The study had 1 prescreening phone contact and 2 office visits (at baseline and at study completion), with 2 additional midstudy phone contacts between visits. Prior to enrollment, all participants were informed about the purpose and risks of the study and gave written consent to participate. Oversight for the study was provided by New England Institutional Review Board, Wellesley, MA. The trial was conducted in accordance and compliance with all federal, state, and local regulations. This trial is registered on clinicaltrials.gov (Identifier: NCT00289354).
Prospective study participants were screened either over the phone or in the study physician's office for eligibility before scheduling the baseline visit. The participant was instructed to arrive in a fasting state to provide blood and urine samples for the following purposes: to determine levels of HbA1c, blood glucose, and serum insulin; to assess serum lipid profile; and to perform blood chemistries and urinalysis (via dipstick). A urine dipstick pregnancy test was performed in all women. During the baseline visit (Day 0), participants received a brief physical examination that included measuring vital signs; taking a comprehensive medical history; and assessing concomitant medications, demographic information, and anthropomorphic characteristics. Participants who qualified for enrollment continued on their existing medications and were randomized blindly in a 2:1 ratio (with a randomization plan generated by www.randomization.com) to receive either CPB (600 μg Cr+3 as CP and 2 mg biotin) or a matching placebo once daily before the morning meal as an adjuvant to the current OAD standard of care. The intervention period was 90 days in duration. Participants were instructed not to change their diet, level of physical activity, or prescription medications. During the course of the study, a centralized call center contacted the participants twice via phone before the final visit. The purpose of the phone contact was to assess the patient's status, remind the participant to take the treatment, perform all diary-related tasks, and provide an opportunity for the participant to ask questions. The office visit procedures outlined above were repeated at the termination of the study during the final visit (Day 90).
At the Day-90 final visit, patient study compliance was assessed and the remaining amount of clinical study test product returned, along with other parameters as deemed appropriate by the investigator.
Sample Size. Determination of sample size was based on the primary efficacy variable of a reduction in HbA1c from baseline to the final visit. This determination was based on published reports from similar studies in the peer-reviewed literature. The sample size of patients per treatment group was sufficient to detect a 0.4% change in HbA1c versus placebo with a type 1 error of 0.05, a power of 80% (β-0.2), and allowance for an attrition rate of 20%.
Laboratory Methods. All laboratory analyses (HbA1c, glucose, and lipid levels; chemistry panel, etc) were performed by Physicians Reference Laboratory (www.prlnet.com). HbA1c was measured at screening using the A1cNow test (Metrika, Inc, Sunnyvale, CA) for the purposes of determining eligibility only. The reported coefficient of variation (CV) for this device is 5.4%, which is sufficient for the purpose of determining eligibility. HbA1c values used for efficacy analyses were provided by fasting blood samples collected at the baseline and final visits. These fasting blood samples were used for all other individual laboratory analyses.
Plasma lipoprotein panels included measurements of TG and cholesterol levels (total-C, LDL-C, and high-density lipoprotein cholesterol [HDL-C]; routine and Centers for Disease Control/National Heart, Lung, and Blood Institute Lipid Standardization Program). Additional lipid ratios such as total-C/HDL-C, LDL-C/HDL-C, TG/HDL-C, and atherogenic index were also calculated.
Safety and Tolerability. Standard safety assessments included adverse event reports, clinical laboratory data (hepatobiliary and renal biochemistries were included), hematology, urinalysis, vital signs, and physical examination conducted at the baseline and final study visits. Patients were monitored on a regular basis for self-reported and clinical adverse experiences.
Statistical Analyses. The population in the current analysis is defined as those participants who returned for at least 1 final efficacy visit and who had no significant protocol entry violations. The defined population encompassed 348 participants and included 226 and 122 in the treatment and placebo groups, respectively. Using this population, we further stratified participants into groups with total-C >200 mg/dL (hypercholesterolemia) and those receiving statin therapy (Figure).
All values are expressed as mean ± SD unless otherwise indicated. The normality of the data was confirmed by chi-square goodness-of-fit testing, and the homogeneity of variances between groups was confirmed by Bartlett's test. Statistical significance was accepted at P<.05. Data analyses were performed using SAS Software Release 8.2 (SAS, Inc, Cary, NC).
Participants were randomized in a 2:1 ratio to active treatment or placebo, respectively. The randomization scheme was developed using standardized software with block sizes of 6 participants per blocking unit. All medication bottles were prelabeled with the participant allocation number with space provided for participant initials and date of dispensation.
The supplemental analysis of lipid and lipoprotein levels included patients in both treatment groups who met the stratification criteria of having high cholesterol levels at the baseline visit (moderate to high cholesterol with serum total-C >200 mg/dL). The 2-group t test procedure was used in all analyses. Another supplemental analysis was conducted in the same hypercholesterolemic participants using further stratification criteria of receiving statins. Multiple regression analyses were conducted to illuminate the correlation coefficients of the outcome variables in the 2 treatments groups.
The comparison between treatments in the analysis of treatment-emergent adverse events was based on Fisher exact test.
Of the 600 patients screened for eligibility, 447 patients were allocated to randomization at the baseline Day 0 visit and 348 completed the trial (Figure). Baseline anthropometric characteristics and the duration of diabetes from diagnosis to study entry are presented in Table II; of the 348 patients, more than 50% of participants in the active group and more than 60% in the placebo control group were men. No significant difference was observed between the groups for the sex distribution (P=.061). Approximately 50% of the participants in both treatment groups were white, 30% were Hispanic, and 10% were black. Chi-square tests indicated that the racial-ethnic distribution of the participants was not significantly different between the 2 groups (P=.0975). Over 85% of the participants in each treatment group were younger than 45 years of age. There was a significant difference between the treatment groups in the distribution of ages (P=.0267), and there was a significant difference between the treatment groups in the number of cigarettes smoked daily (P=.0282). More participants in the active group were in the 1 to 5 cigarettes per day category (between 40% and 50%) and more participants in the placebo group were in the 11 to 20 cigarettes per day category (50%). There was no significant difference in the amount of alcohol consumed weekly between the treatment groups (P=.9834). Hypertension and T2DM were the 2 conditions most prominently seen in both treatments. All participants had a positive family history for at least 1 of the 8 conditions displayed.
Table II. Demographics and Characteristics of the Participants
All Participants (N=348)
Hypercholesterolemic Participants (n=137)
Body weight, kg/m2
Body mass index
Ethnicity, %, white/Hispanic/black/Asian/other
Smoking, % (yes/no)
Alcohol, % (yes/no)
Data are mean ± SD and were analyzed by an unpaired t test.
During the course of the study, more participants in the active group discontinued by being “lost to follow-up” than for any other reason (40.8% of discontinued participants), while more participants in the placebo control group discontinued because of “withdrawal of consent” than for any other reason (33.3% of discontinued participants).
Table III displays summary statistics by treatment group at baseline and the change between baseline and final visit results for efficacy analyses.
Table III. Baseline and Change in Risk Factors by Treatment
All Participants (N=348)
Hypercholesterolemic Participants (n=I37)
Change by Treatment
Change by Treatment
SBP, mm Hg
DBP, mm Hg
46.1 ± 11.1
5.21 ± 1.89
3.11 ± 1.11
aP<.05; bP<.01. Abbreviations: SBP, systolic blood pressure; DBP, diastolic blood pressure; HbAlc, hemoglobin Alc; HOMA, homeostasis model assessment; IR, insulin resistance; BCF, beta-cell function; Total-C, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TG, triglyceride; VLDL-C, very low-density lipoprotein cholesterol; AI, atherogenic index (non-HDL-C/HDL-C). Data are mean ± SD and were analyzed by an unpaired t test. There was no significance at the baseline for the risk factors by treatment.
Effect on Lipids, Lipoproteins, and Lipid Ratios. There was a significant difference between treatments on the change from baseline values for TG/HDL-C and atherogenic index (log [TG/HDL-C]; P<.05), but not for any other measures. In this study, however, Hispanic ethnicity was predictive of a greater reduction in the total-C/HDL-C values (P<.0303).
In the hypercholesterolemia/T2DM population, the results indicated a significantly larger reduction from baseline in the active versus the placebo group for total-C (−18.9 vs −5.44 mg/dL; P<.04), LDL-C (−21.9 vs −8.05 mg/dL; P<.01), and atherogenic index (−0.351 vs 0.055; P<.02) (Table III).
Effect on Glycemic Control. Mean change in HbA1c and fasting glucose levels from baseline was greater in the CPB group than the placebo group (P<.0339 and P<.0207, respectively). In the hypercholesterolemia/T2DM population, a significant difference in the effect of the 2 treatments on HbA1c level was observed (P<.03).
Effect of CPB Supplementation in Statin Users. In a supplemental analysis of statin users with hypercholesterolemia/T2DM, we evaluated the effects of CPB on lipid profiles, HbA1c levels, blood pressure, body weight, and body mass index compared with placebo. The active group (CPB) showed a significant decrease in blood pressure and VLDL-C (P<.05), with no significant changes in the placebo group. Greater effects were seen in the statin users with hypercholesterolemia/T2DM, where significant decreases (P<.05) in LDL-C, total-C/HDL-C, systolic blood pressure, body weight, body mass index, and HbA1c were observed in the active but not in the placebo group.
Safety. Safety analyses were conducted on the intent-to-treat population. This population was defined as all participants who were randomized to the study product, who received at least 1 dose, and who returned for an exit visit. Analyses of the intent-to-treat population revealed that the once-daily oral CPB intervention was well tolerated. The types and indices of treatment-emergent adverse events in the CPB group were generally similar to those in the placebo group, with no statistical differences observed. The clinical safety laboratory work profile for the active treatment group was not significantly different from that of the placebo group, including the absence of significant changes in vital signs, blood chemistries, or urinalysis. An analysis of adverse experiences revealed that 1.8%of CPB participants and 3.3% of placebo control participants discontinued due to an adverse event. A significantly lower proportion of participants in the active group experienced a serious adverse event than in the placebo group (0.9%vs 4.1%, respectively). No participants had alanine aminotransferase level >3× ULN or creatine phosphokinase >10× ULN. Data from other laboratory parameters and vital signs indicated no active treatment (CPB)-related trends or patterns.
The primary objective of the present study was to determine whether use of CPB improved markers of glycemic control such as fasting glucose and HbA1c levels in overweight to obese participants with T2DM who were taking OADs. The secondary objective, which was equally important, was to assess changes in lipid and lipoprotein levels in the same population. Recent data indicate that plasma, hair,13 and possibly tissue levels of chromium in individuals with CHD,14 T2DM,15 and comorbid T2DM and CVD16,17 are reduced and that chromium supplementation may help replete chromium stores to within the normal range.18
In recent studies,11,12 the use of CPB improved blood glucose, fructosamine, and HbA1c levels in T2DM. These results are consistent with those reported in previous studies evaluating the efficacy of CP8,9,19–21 and biotin22 in individuals with T2DM.23,24
Glycemic control remains the cornerstone of therapeutic strategies in patients with T2DM and its complications.25 Cr+3 appears to be an integral element in glucose homeostasis. Rohlfing and colleagues26 suggested that each 1% change in HbA1c represents a change of approximately 35 mg/dL mean plasma glucose or 2.0 mmol/L. Similarly, in the current study the active intervention participants demonstrated a highly significant reduction in HbA1c with a mean change of 0.53 (P<.05) and a significant reduction in blood glucose (P<.02). In the hypercholesterolemia/T2DM population, the active intervention group experienced a significant reduction in HbA1c with a mean between-group change of −0.31% (P<.03). According to the Diabetes Control and Complications Trial (DCCT)26 and United Kingdom Prospective Diabetes Study (UKPDS)27 trials, the risk of the progression of chronic disease complications associated with diabetes is closely related to the degree of glycemic control. These 2 trials showed that if people with diabetes can lower their HbA1c level by any amount, they would improve their quality of life.
In addition, it has been conclusively demonstrated that there is a strong association between low HDL-C levels and/or elevated TG levels and an increased risk of coronary heart disease in patients with diabetes.28–30 In this study, glycemic control, coronary risk lipids, lipoproteins, and atherogenic index were all significantly improved in the hypercholesterolemia/T2DM group during CPB treatment. These observations suggest that using CPB may be a promising therapeutic strategy for reducing the increased risk for diabetes and dyslipidemia in patients with CMS. Therefore, we hypothesize here that the improvement in glucose and lipid metabolism could be associated with an amelioration of insulin resistance.
Results of our study confirm that CPB represents a safe and effective medical nutritional therapy, adjuvant to standard of care, and provides additional options for improvements in patients with hypercholesterolemia/T2DM. The present study demonstrates that CPB is effective in lowering HbA1c and fasting glucose levels; however, the improvements observed in the lipid and lipoprotein levels were more pronounced in the subset of patients with hypercholesterolemia/T2DM in the CPB therapy group. These results are promising and indicate that further clinical investigations in patients with hypercholesterolemia/T2DM are warranted to illuminate the full benefits of CPB adjuvant therapy on risk factors associated with CMS.
Limitations of this study include that although lipids and lipid ratios were a secondary end point, the study was not directly designed to assess improvements in CVD risk in patients with T2DM. The study did not control for the type of anticholesterolemic agent the patients were receiving or their cholesterol level at baseline. In addition, 12-lead echocardiography was not performed and could have detected improvements in wave functions that would have indicated reduced CVD risk. The study was of limited duration at 90 days of intervention and to more accurately assess reductions in CVD risk, a longer trial should be conducted. Further studies to assess reductions in CVD risk should focus primarily on hypercholesterolemic and dyslipidemic T2DM patients.
CPB was associated with favorable changes in cardiometabolic risk factors and an improvement in glycemic control in patients with poorly controlled T2DM. Intervention with CPB appears to be generally well tolerated with no adverse effects dissimilar to those associated with placebo. CPB therefore has the potential to be a useful adjunct to lifestyle modification and pharmacotherapy in the treatment of CMS and T2DM.
Disclosures: Nutrition 21, Inc, provided funding and clinical supplies for the reported clinical investigation. This trial is registered on clinicaltrials.gov, Identifier: NCT00289354. A portion of the efficacy data was presented at the American Diabetes Association, 65th Scientific Sessions.31