Journal of Applied Physiology

Effects of resistance training and chromium picolinate on body composition and skeletal muscle in older men

Wayne W. Campbell, Lyndon J. O. Joseph, Stephanie L. Davey, Deanna Cyr-Campbell, Richard A. Anderson, William J. Evans


The effects of chromium picolinate (CrPic) supplementation and resistance training (RT) on skeletal muscle size, strength, and power and whole body composition were examined in 18 men (age range 56–69 yr). The men were randomly assigned (double-blind) to groups (n = 9) that consumed either 17.8 μmol Cr/day (924 μg Cr/day) as CrPic or a low-Cr placebo for 12 wk while participating twice weekly in a high-intensity RT program. CrPic increased urinary Cr excretion ∼50-fold (P < 0.001). RT-induced increases in muscle strength (P < 0.001) were not enhanced by CrPic. Arm-pull muscle power increased with RT at 20% (P = 0.016) but not at 40, 60, or 80% of the one repetition maximum, independent of CrPic. Knee-extension muscle power increased with RT at 20, 40, and 60% (P < 0.001) but not at 80% of one repetition maximum, and the placebo group gained more muscle power than did the CrPic group (RT by supplemental interaction,P < 0.05). Fat-free mass (P < 0.001), whole body muscle mass (P < 0.001), and vastus lateralis type II fiber area (P < 0.05) increased with RT in these body-weight-stable men, independent of CrPic. In conclusion, high-dose CrPic supplementation did not enhance muscle size, strength, or power development or lean body mass accretion in older men during a RT program, which had significant, independent effects on these measurements.

  • exercise
  • strength training
  • aging
  • basal metabolic rate
  • muscle power

the trace mineral trivalent chromium (Cr3+) is an essential nutrient involved in the regulation of carbohydrate, lipid, and protein metabolism via an enhancement of insulin action (2, 4, 5). Chromium supplementation, in the form of the organic compound chromium picolinate, was reported (24) to have ergogenic properties related to body composition changes in young men. The suggested benefits of chromium picolinate included decreased fat mass (FM) and percent body fat, increased fat-free mass (FFM), increased skeletal muscle size and strength, and enhanced senses of energy and vigor. These chromium picolinate-related benefits were subsequently promoted in the public (23) and research (36) literature but were not supported by other controlled research studies that used young men and women as subjects (20, 32, 34, 39, 46).

If chromium picolinate does have ergogenic properties, older people may especially benefit from the use of this supplement. Aging is associated with marked changes in body composition, skeletal muscle size and function, and energy metabolism. These changes include an increase in both FM and percent body fat, a decrease in FFM (primarily due to a loss of muscle mass), a decrease in muscle strength and power, and a decrease in resting metabolic rate (21, 25, 47). These body composition, functional, and metabolic changes contribute to the development of age-related disorders, including obesity, sarcopenia, impaired physical mobility and function, and physical frailty (25, 27). The promoted benefits of chromium picolinate are largely opposite to these age-related body composition, functional, and metabolic changes.

The ergogenic effects of chromium picolinate are thought to be manifest especially when supplementation is coupled with an exercise program that includes resistance training (RT) (24). High-intensity progressive RT has been shown to decrease FM, increase FFM, increase muscle strength and size, and increase resting metabolic rate and total energy requirements of older people (17, 30), including frail elderly people (28, 29). Thus chromium picolinate and RT may provide older people with safe and effective ways to slow the progression of, or to partially reverse, the age-related changes in body composition, muscle function, and energy metabolism and may prevent or treat the associated health disorders.

There are no known data on the effect of chromium picolinate supplementation on body composition, muscle strength and function, and energy metabolism in older people. Thus the purpose of the present study was to assess the effect of high-dose chromium picolinate supplementation on body composition, including body density, whole body muscle mass and muscle fiber area, muscle strength and power, and resting metabolic rate of older men during 12 wk of RT. We hypothesized that chromium picolinate supplementation would augment RT-induced changes in these parameters.



Twenty-three men volunteered to participate in this 13-wk study. Recruitment criteria included the following:1) men, age range 50–75 yr;2) body mass index range of 27–34 kg/m2;3) nondiabetic;4) physically able to safely engage in all aspects of the study protocol; and5) clinically normal cardiac function, blood pressure, liver function, and kidney function. Each man was cleared for the study on the basis of a screening evaluation that included a medical history, a physician-administered physical examination, a resting and resistance exercise electrocardiogram, routine blood and urine chemistries, and a 3-h, 75-g oral glucose tolerance test. For the resistance exercise stress test, heart rate, blood pressure, and electrocardiogram recordings were monitored while each prospective subject performed the bilateral knee-extension exercise for three sets of eight repetitions at 70–80% of his maximal strength. After receiving complete written and verbal explanations of the study, each man signed an informed consent agreement. The study protocol and informed consent agreement were reviewed and approved by the Institutional Review Board of The Pennsylvania State University (University Park). The protocol was also reviewed and approved by the General Clinical Research Center (GCRC) Advisory Committee of The Pennsylvania State University (University Park).

Eighteen of twenty-three men successfully completed the study protocol. The reasons for the five men withdrawing included the following:1) a request by a subject’s personal physician to avoid aggravation of a chronic hip injury;2) an irritation of chronic elbow tendonitis, unrelated to resistance exercise;3) a personal family commitment;4) a resistance exercise-induced aggravation of an existing knee condition; and5) a shoulder injury caused by slipping on ice.

Experimental design.

The 13-wk study was conducted at a GCRC located at the Noll Physiological Research Center on the University Park campus of The Pennsylvania State University. Throughout the study, each man maintained his customary lifestyle while living at home, except for when he participated in the planned dietary control periods, the scheduled testing, and the RT. Use of all nutritional supplements, other than the supplements used for the study, was discontinued 3 wk before and throughout the study.

Baseline testing and evaluations were completed during studyweek 1, at which time each man did not consume a supplement and remained sedentary. This was followed by a 12-wk period of supplementation and RT, with testing and evaluations repeated at study weeks 7 and13.

Each man was randomly assigned in a double-blind fashion to either a chromium picolinate group or a placebo group, on the basis of the date they were declared eligible to participate in the study. The starting dates of the subjects were staggered and ranged from June to November 1995. The chromium picolinate group consumed twice daily one commercially prepared capsule reported to contain 9.62 μmol chromium/capsule (500 μg chromium/capsule) as chromium picolinate (Nutrition 21, San Diego, CA), and the placebo group consumed one commercially prepared placebo capsule (Nutrition 21) twice daily. All of the men were to consume one capsule in the morning and one capsule in the evening. The capsules were distributed weekly in 7-day medication dispensers, and compliance was monitored by counting any returned capsules and by weekly interviews with the men. The chromium picolinate and placebo capsules were analyzed for chromium by using a graphite furnace atomic absorption spectrophotometer (model HGA 500, Perkin-Elmer, Norwalk, CT), as previously described (11, 32), and were found to contain 8.88 ± 0.21 μmol chromium/capsule (462 μg chromium/capsule) and <0.002 μmol chromium/capsule (<0.1 μg chromium/capsule), respectively. Thus each man in the chromium picolinate group consumed 17.8 μmol chromium/day (924 μg chromium/day) of supplemental chromium. The rationale for using this high dose of chromium was to provide an opportunity to clearly test our hypotheses with a chromium intake distinctly different from that which our subjects might normally consume in their habitual diets (i.e., the dose of chromium was not “nutritional”) and a dose recently reported to have beneficial effects on carbohydrate metabolism in middle-aged and older people with type 2 diabetes (10).

RT protocol.

All men participated twice weekly for 12 wk in progressive RT of the muscles involved with 1) unilateral knee extension, 2) unilateral knee flexion, 3) double leg press,4) seated chest press, and5) seated arm pull (Keiser pneumatic resistance equipment, Keiser Sports Health Equipment, Fresno, CA). Each man’s baseline maximal strength for each exercise was set as the greater of two one-repetition-maximum (1RM) values obtained during the first two resistance exercise sessions. On these days and on all 1RM assessment days, each man then performed two sets of eight repetitions at 80% of the 1RM for each exercise. 1RM assessments were repeated at study weeks 7 and13. On the non-1RM days, the RT consisted of the performance of each exercise for three sets at 80% of 1RM. Eight repetitions were completed for the first two sets, and the third set consisted of continuing repetitions until voluntary muscular fatigue or until 12 repetitions were completed. If 12 repetitions were completed for a given exercise, the resistance was increased by 5% for the next exercise session. All resistance exercise sessions were preceded by 10 min of easy cycling (heart rate <100 beats/min) and 10 min of stretching and were followed by a similar routine. Seventeen men completed 23 RT sessions (100% compliance), whereas one man completed 22 RT sessions.


During study weeks 2–6 and8–12, each man was asked to maintain his self-provided habitual diet. In an effort to control dietary chromium intake for the measurements of urinary chromium excretion and apparent chromium absorption, each man consumed, for 5 days during study weeks 1, 7, and13, the same quantity of foods and beverages (except water) prepared and provided by the Metabolic Research Kitchen at the GCRC. The controlled diet consisted of 2-day rotating menus: menu A andmenu B (Table1). Both menus were designed to provide 13, 57, and 30% of total energy as protein, carbohydrate, and fat, respectively. Each man’s total energy intake was estimated to be 1.5 times basal energy needs as predicted from the sex-specific Harris-Benedict equation (33). Total energy intake was calculated by using Nutritionist IV software (version 4.0, N-Squared Computing, First Data Bank, San Bruno, CA), assuming metabolizable energy values for protein, carbohydrate, and fat of 16.7, 16.7, and 37.7 kJ/g, respectively.

View this table:
Table 1.

Composition of a representative controlled 2-day cycle menu

One duplicate composite of menu A andmenu B for each man was collected into 3-liter polypropylene jugs and frozen at −20°C. Thawed composites were homogenized in the polypropylene jugs by using low-chromium titanium blender blades and analyzed for chromium concentration as previously described (11). Water samples were also collected from seven drinking fountains and faucets in and around the University Park campus and the State College, PA area and were analyzed for chromium concentration. A mixed diet standard reference material (SRM 8431; National Institute of Standards and Technology, Gaithersburg, MD) with a certified chromium concentration of 102 ± 6 ng/g was run with each analysis to verify the accuracy of individual assays. A mean chromium concentration of 103 ± 10 ng/g was determined for the standard reference material over all assays. We assumed that the chromium contents of the composite menus for each subject were the same as the chromium contents of the menus provided at study weeks 1, 7, and13.

Muscle power assessments.

Maximal arm and leg power were determined at studyweeks 1 and13 by using the Keiser seated arm-pull and leg-extension machines, respectively. Each man was asked to exert maximal pull using both arms or push using the dominant leg with the machine resistance set at 20, 40, 60, or 80% of the most recently determined 1RM. The test was repeated three times at each percentage of 1RM, with ∼30 s of rest in between efforts.

Urine collections.

At study weeks 1, 7, and13, 24-h urine collections were made during the last 3 days of controlled menu feedings. A RT session was performed on one of the urine collection days at weeks 7 and 13. All 24-h urine samples were collected into disposable 4-liter polyethylene containers (Fisher Scientific, Pittsburgh, PA). Aliquots of urine for chromium analyses were pipetted by a research technician, wearing powder-free gloves and using disposable borosilicate glass pipettes, into polypropylene test tubes (Sarstedt, Newton, NC) and were frozen at −20°C. Samples were analyzed for chromium by using a graphite furnace atomic absorption spectrophotometer (model HGA 500, Perkin-Elmer), as described (14). A pooled urine sample, with a gas chromatography-mass spectrometer-verified concentration of 0.32 ± 0.01 ng/ml, was run with each assay as a quality control. The mean chromium concentration of this control urine sample over all assays was 0.33 ± 0.02 ng/ml, as measured by atomic absorption spectrophotometry. Percent chromium absorption was estimated by using the following formula: %Cr absorption = [urinary Cr excretion/(food Cr intake + supplement Cr intake)] × 100.

Body composition measurements.

Fasting body weight was measured each weekday during studyweeks 1, 7, and13 and twice weekly during the other study weeks. Weights were taken to the nearest 0.1 kg with the subject wearing underwear, socks, T-shirt, and gym shorts. Nude body weight was calculated as total body weight − weight of socks, T-shirt, and gym shorts. Body height, without shoes, was measured one morning duringweek 1, to the nearest 0.1 cm, with a wall-mounted stadiometer and was assumed to remain constant throughout the study.

Body composition was assessed by using the techniques described below at study weeks 1, 7, and13. Total body water (TBW) was determined in the fasting state by using a 20.0-g dose of deuterium oxide (deuterium, 99.9%; Cambridge Isotope Laboratories, Woburn, MA) as described (17). Body density was determined in the fasting state by using hydrostatic weighing (1), with residual volume measured in the hydrostatic weighing tank via the nitrogen dilution technique (49).

Body composition was assessed by using both a two-compartment model (2c; FM and FFM) and a three-compartment model (3c; FM, TBW mass, and protein + mineral mass). Whole body FM and FFM were estimated from body density by using the 2c model equation of Siri (45). Whole body FFM, FM, and protein + mineral mass were estimated from TBW and body density by using the 3c model of Siri (45) as described (17). For the 3c model, FFM was calculated as body mass − FM, and protein + mineral mass was calculated as FFM − TBW mass. For the purpose of presentation, FM and FFM estimates derived by using the 2c model are designated FM2c and FFM2c, respectively, and FM and FFM estimates derived by using the 3c model are designated FM3c and FFM3c, respectively.

Total body muscle mass was estimated from the mean creatinine concentration of three consecutive 24-h urine samples, assuming an equivalence of 18.5 kg muscle/g urinary creatinine (35). The creatinine concentration of each urine sample was analyzed on a Technicon Autoanalyzer II (Technicon Instrument, Tarrytown, NY) by using the Jaffe reaction (15).

Skinfold thicknesses were measured to the nearest 0.5 mm at eight sites (biceps, triceps, chest, subscapula, mid-axial, abdomen, suprailiac, and thigh) on the right side of the body by using standard techniques (38) and Lange calipers (Cambridge Scientific Industries, Cambridge, MD). The sum of these eight skinfold thicknesses is reported. Body circumference measurements were taken at the chest (at the level of the fourth costosternal joints, in the horizontal plane, at the end of a normal expiration) and midthigh (midway between the inguinal crease and the distal border of the patella).

Muscle biopsy.

The needle biopsy technique (26) was used at studyweeks 1 and13 to obtain vastus lateralis samples from the dominant leg of each man (∼150 mg muscle/biopsy). A piece of each biopsied muscle sample was placed with the fibers arranged in a longitudinal fashion into a gelatin capsule filled with optimum cutting temperature embedding compound (Sakura Finetek, Torrance, CA). The capsule was covered with optimum cutting temperature compound and quick-frozen for 20 s in isopentane that was precooled to −160°C by using liquid nitrogen. The muscle sample was then placed into liquid nitrogen and transported to a −70°C freezer for storage. Serial 10-mm muscle sample sections were sliced onto microscope slides by using a cryostat (Cryocut 1800, Leica Instrument, Germany) maintained at −20°C and were stained for myofibrillar adenosine triphosphatase, on the basis of published methods (16, 42), by using a pH 4.3 incubation solution. Types I and II muscle fiber areas were assessed by using the National Institutes of Health Image Program, version 1.60a, as modified by Scion (Frederick, MD) with the use of an LG-3 Scientific Frame Grabber PCI card (Scion). Stained muscle sections were viewed by using the 10× objective of a Nikon Micro Photo-FXA microscope (Tokyo, Japan) that was calibrated with a micrometer mounted on the slide, and an Ikegami 370M model S camera unit (Tsushinki). The mean muscle fiber areas were calculated from a mean of 59 type I fibers (range 15–140 fibers) and 73 type II fibers (range 13–188 fibers).

Resting metabolic rate measurements.

Resting metabolic rate was measured in the fasting state one morning during study weeks 1 and13. Each man arrived at the GCRC by automobile, was escorted to a room with an indirect calorimeter, and rested in a semirecumbent position for ∼20 min with the lights low. A clear plastic box, part of a ventilated-hood system, was then placed over the subject’s head while he rested supine for another 15 min. Then the rates of oxygen consumption (l/min) and carbon dioxide production (l/min) were measured and averaged from 15 consecutive 1-min expired air samples. Resting metabolic rate was calculated by multiplying the oxygen consumption rate (l/min) by the kilojoules per liter of oxygen associated with the respiratory exchange ratio of the expired air (40).

Statistical methods.

Values are reported as means ± SD. The difference between the chromium picolinate group and placebo group for each of the independent variables reported was determined for study week 1 data (baseline) by using Student’s unpairedt-test. The main effects of RT and chromium picolinate supplementation and the interactions among these dependent variables on each of the independent variables were determined by using two-way repeated-measures analysis of variance. All calculations were performed by using PROC TTEST and PROC GLM of SAS version 6.11 (SAS Institute, Cary, NC). All data processing was performed by using Microsoft Excel 5.0 (Microsoft, Redmond, WA). Results were considered statistically significant ifP < 0.05 (2-sided).


Menus A andB used during the 5-day controlled diet periods at study weeks 1, 7, and13 provided a mean energy intake of 11.4 ± 1.1 MJ/day (2,722 ± 258 kcal/day) and 11.6 ± 1.2 MJ/day (2,772 ± 282 kcal/day), respectively (Table 1). The energy intake of menu A consisted of 13% protein, 57% carbohydrate, and 30% fat, and the energy intake ofmenu B consisted of 13% protein, 58% carbohydrate, and 29% fat. The chromium contents ofmenus A andB were 1.88 ± 0.27 μmol/day (98.0 ± 14.1 μg/day) and 1.09 ± 0.12 μmol/day (56.9 ± 6.0 μg/day), respectively.

The dietary intake, supplement intake, urinary excretion, and estimated absorption of chromium for menus A andB at study weeks 1, 7, and 13 are presented in Table 2. For the placebo group, the mean urinary chromium excretion at baseline was 5.77 ± 3.46 nmol/day (0.30 ± 0.18 μg/day) for menu A and 3.85 ± 2.88 nmol/day (0.20 ± 0.15 μg/day) for menu B and was similar at study weeks 7 and13. In contrast (time × diet interaction, P < 0.001), chromium picolinate supplementation increased urinary chromium excretion ∼50-fold, compared with baseline (42- and 48-fold formenu A and 66- and 57-fold formenu B at study weeks 7 and 13, respectively). At baseline, the mean absorption of chromium based on urinary chromium excretion ranged from 0.25 to 0.37% (for both groups consuming either menu). For the placebo group, the calculated absorption of chromium at study weeks 7 and 13 remained similar to baseline (0.24–0.59%). In contrast (time × diet interaction, P < 0.001), chromium picolinate supplementation increased the estimated absorption of chromium to a range of 0.93–1.15%.

View this table:
Table 2.

Chromium intake, urinary excretion, and estimated absorption of older men before, midway between, and after 12 wk of resistance training with or without chromium picolinate supplementation

Water samples collected from seven different drinking fountains and faucets contained 0.33 ± 0.09 ng chromium/ml (range 0.22–0.46 ng chromium/ml). Thus 4 l/day (≈1 gallon/day) of water intake would provide 0.9–1.8 μg chromium/day.

At baseline, the mean values for all of the independent variables related to body size, body composition, muscle size, strength and power, and resting energy expenditure were not different between the men in the chromium picolinate group and the men in the placebo group (see Tables 3 and 4 and Figs. 1 and 2).

For all measurements of body composition, statistical analyses established that no RT-by-chromium picolinate interaction or main effect of chromium picolinate existed (Table3). These analyses indicate that1) when RT resulted in changes in these measurements of body composition, these changes were not influenced by whether the men consumed chromium picolinate or the placebo, and 2) when data from all of the evaluation periods were considered together, there were no differences between groups. Because there were no RT-by-chromium picolinate interactions, the effect of RT was assessed by using data from men in both groups combined (n = 18).

View this table:
Table 3.

Physical characteristics, body composition, and resting metabolic rate of older men before and after 12 wk of resistance training with or without chromium picolinate supplementation

RT increased body density (P = 0.003; Table 3). By using the body density data in a 2c model, FM2c(P = 0.007) and percent body fat (P = 0.003) were decreased by 2.1 ± 2.9 kg and 2.1 ± 2.6%, respectively, and FFM2c was increased by 2.2 ± 2.3 kg (P < 0.001).

TBW was decreased by 3.0 ± 2.3 liters with RT (P < 0.001; Table 3). By using the TBW data and the body density data in a 3c model, FM3c(P = 0.047) and percent body fat (P = 0.026) were increased by 1.5 ± 2.9 kg and 1.6 ± 2.7%, respectively, and FFM3c decreased by 1.4 ± 2.1 kg (P = 0.016). The protein + mineral mass compartment (calculated as FFM3c − TBW) was increased by 1.6 ± 1.1 kg (P < 0.001).

Whole body muscle mass, assessed from 24-h urinary creatinine excretion, increased with RT by 5.1 ± 3.8 kg (P < 0.001; Table 3). Mean type I muscle fiber area was unchanged, and mean type II muscle fiber area was increased by 577 ± 1.62 μm2(P = 0.046) with RT, as assessed from vastus lateralis biopsy samples. The sum of eight skinfold thicknesses and midthigh circumference did not change with RT, and chest circumference was decreased by 1.37 ± 2.16 cm (P = 0.018).

The mean resting metabolic rate at study week 13 was not statistically different from baseline but did trend higher [mean change from baseline, 0.016 ± 0.034 l O2/min (P = 0.068) and 0.46 ± 1.07 MJ/day (P = 0.093) forn = 18 subjects combined].

1RM strength increased in all of the muscle groups trained (P < 0.0001; Table4). Quantitatively, the amount of strength gained was similar for men in the chromium picolinate and placebo groups for the right knee flexion, left knee flexion, double leg press, chest press, and arm pull exercises. However, the chromium picolinate group gained less strength in both the right and left knee-extension exercises than did the placebo group (RT-by-chromium picolinate interaction, P = 0.035 andP = 0.009, respectively).

View this table:
Table 4.

Maximal muscle strength of older men before and after 12 wk of resistance training with or without chromium picolinate supplementation

The data from the muscle power testing for the arm- pull and leg-extension exercises are presented in Figs.1 and 2, respectively. Chromium picolinate supplementation did not influence the changes in arm muscle power to RT. After 12 wk of RT, arm-pull muscle power at 20% of 1RM was increased and at 40, 60 and 80% of 1RM was not different, compared with baseline. Knee-extension power at 20, 40, and 60% of 1RM increased less in the chromium picolinate group than in the placebo group (RT-by-chromium picolinate interaction,P = 0.005,P = 0.024, andP = 0.069, respectively). Knee-extension power at 80% of 1RM was not different at studyweek 13 than at baseline and was not influenced by chromium picolinate supplementation.

Fig. 1.

Arm-pull muscle power before (baseline) and after 12 wk of resistance training (RT week 12) of older men who consumed either 17.8 μmol chromium/day (924 μg chromium/day) as chromium picolinate (A) or a low-chromium placebo (B) during RT. 1RM, 1 repetition maximum. Statistical analyses are as follows.

Embedded Image

Fig. 2.

Leg-extension muscle power before (baseline) and after 12 wk of RT of older men who consumed either 17.8 μmol chromium/day (924 μg chromium/day) as chromium picolinate (A) or a low-chromium placebo (B) during RT. Statistical analyses are as follows.

Embedded Image


In the present study, supplementation of older men with 17.8 μmol chromium/day as chromium picolinate for 12 wk did not influence any of the RT-induced changes in body composition. Chromium picolinate supplementation did not augment the accretion of FFM or muscle mass or enhance the loss of body fat. These findings are consistent with most (20, 32, 39, 46) but not all (24, 36) double-blind, placebo-controlled chromium picolinate supplementation trials conducted in young men and women (Table 5). The present study expands present knowledge by studying older men, by using a much higher dose of chromium picolinate, and by using a wider variety of methods to quantify body composition.

View this table:
Table 5.

Chromium picolinate supplementation effects on body composition

The original report of Evans (24) and the research by Kaats et al. (36) are the only two published studies that report chromium picolinate to have ergogenic effects on body composition. Several factors severely limit the ability to evaluate the quality of the data of Evans (24). First, there is no way to evaluate the accuracy or precision of these data because neither the within-group variability of the changes nor the absolute amount of weight, lean body mass, or FM were reported. Second, the use of a paired t-test to assign statistical significance to differences between mean values of group changes and differences in changes between groups may inappropriately enhance the chance of detecting a difference in response between groups. Third, the assessment of body composition was limited to indirect estimates of lean body mass and FM by using skinfold thicknesses and measurement of biceps and calf circumferences. Strengths of the Kaats et al. (36) data include the large number of subjects studied (n = 154) and the use of a double-blind study design. Weaknesses of the study include the following: 1) not controlling the amount of formula beverage consumed or quantifying that chromium picolinate was consumed via measurements of the chromium content of the beverages and urinary chromium excretion;2) the large subject dropout rate [65 of the 219 subjects who started the study (30%)], and3) artificially enhancing the actual between-group differences in body composition change through the use of undocumented calculations of a “body composition improvement score.”

Collectively, the present study and most of the published studies (20,32, 34, 39, 46) do not support the hypothesis (24) that chromium picolinate supplementation promotes the accretion of muscle mass and loss of body fat. The findings that chromium picolinate supplementation did not affect whole body density (assessed via hydrodensitometry), skinfold thicknesses, body circumferences, or whole body muscle mass (assessed via urinary creatinine excretion) support the findings of previous research studies that also used these techniques (20, 32, 34,39, 46). The present study expands present knowledge by showing that chromium picolinate supplementation also did not affect TBW, whole body protein + mineral mass, or type I and II fiber areas of the vastus lateralis. These studies assessed whether chromium picolinate augmented body composition changes in subjects during relatively short-term periods (6–16 wk) and in conjunction with other interventions that specifically attempted to alter body composition (e.g., RT or aerobic training). If chromium picolinate does influence body composition, the effects must be small. Also, the measurement techniques used not only must be able to precisely quantify small changes in body composition, but must be able to quantify differences in the amount of change in body composition (i.e., the interaction of chromium picolinate and the other intervention).

The findings that FFM3c decreased and FM3c increased when we used the 3c model, opposite to the increase in FFM2c and decrease in FM2c observed by using the 2c model, were surprising and difficult to explain (Table 3). These changes in FFM3c and FM3c were mostly related to the very large decrease in TBW having a major impact on the calculation. The mean 3.0-liter decrease in TBW is largely opposite to a 1.6-liter increase in TBW observed in older men and women who completed a similar RT program (17). The large decrease in TBW is consistent with a change in hydration status of the men, a possibility that was not tested in the present study. The men were encouraged to drink a great deal of fluids in conjunction with the exercise sessions. Also, the body weights of the men remained stable throughout the study, with no indication of abrupt declines in body weight, consistent with short-term shifts in TBW. Because the mean body density of these men was >1.000 g/ml at baseline, a decrease in TBW (density = 0.993 at 37°C) is consistent with the increase in body density measured by hydrostatic weighing.

The significant increase in protein + mineral mass is fully consistent with a RT-induced improvement in body composition, as was also shown independently by the increased whole body density, the increased whole body muscle mass, and the increased vastus lateralis type II fiber area. We had chosen to use the 3c model based on previous research (17,18) showing that the estimate of protein + mineral mass was sensitive to change and that a change in protein + mineral mass was reflective of changes in whole body and muscle protein breakdown.

Urinary creatinine excretion is considered a reliable, indirect indicator of whole body muscle mass in adult humans (35), including older people (48). It is also considered sensitive and reflective of RT-induced changes in muscle mass in older people (41). To use urinary creatinine excretion as an index of muscle mass, one must assume that the body creatine pool is almost exclusively found in skeletal muscle, that the conversion of creatine to creatinine occurs irreversibly and at a constant rate, and that the rate of renal excretion of creatinine is constant (35). These assumptions are not always precisely met. For example, urinary creatinine excretion occurs in direct proportion to the whole body creatine pool size and was documented to change to some extent with changes in dietary protein, creatine, and creatinine intakes, independent of changes in lean body mass (22, 35).

For the present study, the increase in urinary creatinine excretion indicates an increase in whole body muscle mass, a finding supported by the increase in vastus lateralis type II fiber area. However, the mean 5.1-kg increase in muscle mass estimated from this method exceeds the mean 2.2-kg increase in FFM estimated from body density. It is possible that the RT-induced gains in muscle mass were overestimated by the urinary creatinine excretion data, possibly because of changes in whole body creatine pool size (unmeasured in the present study) or dietary factors. The men were purposefully provided a meat-free diet for 2 days before and during the 3-day urine collection periods to minimize exogenous sources of creatine, consistent with standard protocol (35). Whereas urinary creatinine excretion is known to decrease by 5–8% within the first 2–3 wk of consuming a meat-free diet (Ref. 19 and unpublished observations), the magnitude of these dietary-induced changes would be <5% during the 5-day meat-free periods and within the daily within-subject variability of the measurements (8.9 and 7.7% at weeks 1 and13, respectively). Alternatively, it is possible that the RT-induced gains in FFM were underestimated by the hydrostatic weighing method. Indeed, Siri (45) indicated that, when the 2c model is used, the assumption is made that a change in body density reflects a change in body FM and that the apparent change in body density would be less if muscle mass were gained at the same time that FM was lost.

In contrast with the present findings, one previous study from our research group (17, 18) reported that RT by older people did not “improve” body composition. Both studies utilized the same equipment, protocol, and duration of RT. Differences between the previous (17, 18) and present studies, respectively, included studying both men and women vs. just men and using controlled diets intermittently vs. continuously. A strength of both studies was the use of multiple, independent measurements to assess body composition. Within each study, the different measurements of body composition were in general agreement. For example, in the previous research (17, 18), the conclusion that RT did not improve body composition in older people was supported by assessments of body density via hydrostatic weighing, protein + mineral mass from the 3c model by using body density and TBW, whole body muscle mass via 24-h urinary creatinine excretion, thigh muscle area via computed tomography scanning, body cell mass via40K-potassium scanning, and muscle fiber area via biopsies of the vastus lateralis. In the present study, the conclusion that RT did improve body composition in older men is supported by assessments of body density, protein + mineral mass, whole body muscle mass, and vastus lateralis type II fiber area. Collectively, these data emphasize the importance of using multiple techniques to assess baseline and intervention-induced changes in body composition and the inherent difficulties of precisely and accurately measuring relatively small changes in body composition associated with shorter-term RT protocols. There is little doubt that longer-term RT significantly and beneficially alters whole body composition and muscle mass in older people (41, 43). These studies also demonstrate that much more research is needed to better understand the factors (e.g., dietary, hormonal, exercise, gender) involved with maintaining and enhancing FFM and muscle mass in older people as preventive or therapeutic treatments for sarcopenia.

The increase in muscle strength with RT by the older men in the present study is consistent with previous research in older people (17, 30). Similarly, the finding that chromium picolinate did not augment the increase in strength is fully consistent with previous research in young men who resistance trained (20, 32, 34, 39).

To our knowledge, this is the first study to assess the effect on isotonic muscle power of RT by older people and to document increased power of the muscles involved with knee extension. Because 1RM increased with RT, the absolute force or torque at a given percentage of 1RM used for the muscle power testing was higher also. A lack of change in muscle power implies that, although the subjects were exerting more force or torque, their speed of movement was actually less. Alternatively, where power at a given percentage of 1RM increased with RT, it cannot be readily determined how this relates to the speed of movement. The functional significance of these power data was not assessed in the present study and requires further evaluation. Similarly, the finding that the men in the chromium picolinate group gained less strength in the right and left knee-extension exercises and gained less knee-extension power (at 20, 40, and 60% of 1RM) is of interest and requires further research. These data must be viewed as preliminary and should not be used to condemn the use of chromium picolinate supplements, because similar results were not obtained for the other muscle groups tested.

Urinary chromium excretion is suggested to be a “fairly accurate estimation of the amount of chromium absorbed” (11), because absorbed chromium is excreted primarily in the urine, with only minimal losses occurring via other routes. To this end, we chose to use the term “estimated absorption” to reflect the ratio of urinary chromium excretion to oral chromium intake from the diet and the supplement. Whereas a direct comparison between this method of estimating chromium absorption and the use of the stable isotope53Cr has not been published to date, separate assessments that use these two methods give relatively comparable results (11, 44). For the present study, the use of controlled menus of known chromium content during the 24-h urine collection periods enhanced the estimation of the absorption of chromium. The estimated absorption of chromium at baseline, ranging from 0.37 to 0.25% at dietary intakes of 1.07–1.92 μmol chromium/day (56–100 μg chromium/day; Table 2), is fully consistent with an estimated absorption of ≤0.5% at intakes of >0.77 μmol chromium/day (>40 μg chromium/day) (11).

The significant increases in urinary chromium excretion with chromium picolinate supplementation at RT weeks 6 and 12 for these older men are consistent with previous chromium picolinate supplementation studies in RT young men (32, 39). Hallmark et al. (32) reported an ∼10-fold increase in urinary chromium excretion (29.2 vs. 2.9 nmol chromium/day) with a chromium picolinate supplementation of 3.62 μmol chromium/day (188 μg chromium/day). In comparison, chromium picolinate supplementation of 17.8 μmol chromium/day (924 μg chromium/day) in the present study increased urinary chromium excretion ∼50-fold (199 vs. 3.9 nmol chromium/day; mean ofmenus A andB at RT week 6 vs. baseline for chromium picolinate group, Table 2). Provided that urinary chromium excretion is an index of the estimated absorption of chromium, ∼0.93–1.15% of the chromium contained in a chromium picolinate supplement was absorbed. This range of estimated percent absorption of supplemental chromium picolinate is greater than that reported for dietary chromium (Table 2) (11), a finding consistent with previous research in humans (31) and rats (8). Collectively, these data suggest that increased chromium picolinate supplementation proportionately increases urinary chromium excretion via a process that is without strict homeostatic control. It is important to note that this conclusion only relates to increasing doses of supplemental chromium picolinate and not to dietary chromium. At dietary chromium intakes <0.77 μmol chromium/day (<40 μg chromium/day), estimated absorption decreases with increasing chromium intake, i.e., estimated chromium absorption is inversely related to dietary intake at normal intakes (between 10 and 40 μg chromium/day) (11).

Aerobic exercise has been shown to increase urinary chromium excretion (12, 13) in relation to the acute stress of the exercise (9) and the training status of the subjects (7). In the present study, mean urinary chromium excretion of the placebo group was not significantly different from baseline at RT weeks 6 or12 (Table 2). These results suggest that, under these experimental conditions, RT did not alter the urinary excretion or the estimated absorption of chromium. These results are in contrast to the recent report that acute and chronic resistance exercise increased urinary chromium excretion in men aged 53–63 yr, as assessed by the measurement of urinary losses of a standard oral dose of the stable isotope 53Cr (44). Differences in results of the present study and those of Rubin et al. (44), respectively, may relate to1) measurement of the urinary excretion of total dietary chromium intake vs. a 5.77-μmol (300-μg) dose of 53Cr as chromium chloride and 2) the frequency (2 vs. 3 days/wk) and intensity (80% of 1RM for 5 exercises vs. 90% of 3RM for 14 exercises) of resistance exercise.

The chromium contents of controlled diet menus A and B were very high (1.88 and 1.09 μmol/day; 36 and 21 μg/1,000 kcal, respectively; Table 1) based on previous reports (3, 11). This was unexpected because the chromium content of many Western diets and institutional menus (obtained by using the duplicate-plate method) was analyzed and found to contain <0.96 μmol chromium/day (50 μg chromium/day). Published determinations of the mean chromium content of self-selected diets of women and men generally range from 0.48 to 0.63 μmol chromium/day or 14 to 18 μg chromium/1,000 kcal (3, 6, 11). For the present study, the menu items were chosen to fit within the energy and macronutrient guidelines of the study, without consideration of their chromium content. The individual food items were also selected, in part, because they were readily available in the local grocery stores. They were not selected to purposely achieve either a high or low dietary chromium content. Whereas the same food items were used throughout the study, the specific lot of a given food item was not controlled and may have varied in chromium content. To increase the convenience of storing the foods and beverages and distributing the meals to the subjects, many of the items used were processed and packaged in individual serving sizes. The processing and packaging of foods and beverages may increase the chromium content of the item due to exogenous chromium contamination. The low chromium concentration in water samples taken from selected drinking fountains and faucets indicated that ad libitum consumption of water from these sources would not have contributed significantly to total daily chromium intake. These results are in agreement with previous findings in men and women who participated in a long-term, defined-nutrient study, in which the chromium intake from ad libitum consumption of water, coffee, and tea was 0.6–1.1 μg/day (37).

In summary, these data demonstrate that chromium picolinate supplementation has no effect in the augmentation of changes in body composition or muscle size or function when used during a RT program by older men. The greater increase in muscle strength and power in the placebo-supplemented men is unexplained and should be investigated further.


Sincere thanks go to the dedicated volunteers who made this study possible. We thank the General Clinical Research Center dietary and nursing staffs for assistance with conducting this study. We also thank Noella Bryden, US Department of Agriculture Human Nutrition Research Center (Beltsville, MD), for performing the chromium analyses for this study. We are grateful to Keiser Sports Health Equipment (Fresno, CA) for the generous donation of the resistive training equipment used during the study.


  • Address for reprint requests: W. W. Campbell, VA Medical Center, NMEL/NLR, Rm. 3J106, 2200 Fort Roots Dr., North Little Rock, AR 72114–1706 (E-mail: campbellwaynew{at}

  • This study was supported by National Institute on Aging Grants T32 AG-0048, 1-R29-AG-13409 and RO1-AG-11811, by General Clinical Research Center Grant MO1-RR-10732, and by an independent monetary gift from Nutrition 21 (San Diego, CA).


View Abstract