点击显示 收起
1 From the Departments of Nutritional Sciences (RNC, DAS, and ACW) and Internal Medicine (EHN), University of Wisconsin-Madison, Madison, WI
2 Supported by Cognis Nutrition & Health (LaGrange, IL) and by NIH grant RR03186. 3 Reprints not available. Address correspondence to DA Schoeller, 1415 Linden Drive, Madison, WI 53706. E-mail: dschoell{at}nutrisci.wisc.edu.
ABSTRACT
Background: Conjugated linoleic acid (CLA) is a family of positional and geometric isomers with 2 conjugated double bonds formed from linoleic acid and linolenic acid. CLA has a wide range of biological effects, including body fat reduction.
Objective: The aim of our study was to determine CLA's effects on energy expenditure, macronutrient utilization, and dietary fat oxidation in overweight adults after 6 mo of supplementation.
Design: We recruited 23 subjects from our main CLA efficacy study who were receiving either 4 g/d of 78% active CLA isomers (3.2 g/d: 39.2% cis-9,trans-11 and 38.5% trans-10,cis-12) or 4 g/d of safflower oil. Energy expenditure and substrate utilization were measured before and after 6 mo of CLA supplementation by using whole-room indirect calorimetry. Dietary fat oxidation was measured by using stable isotope–labeled oleate and palmitate.
Results: Our substudy detected a difference in the change in fat utilization between the CLA (4 ± 8 g) and placebo (–7 ± 11 g) groups during sleep after 6 mo of supplementation. In addition, the percentage of energy from protein was reduced during sleep in the CLA group (CLA: –3.3 ± 2.6%; placebo: 0.3 ± 5.7%). We also detected a difference in the change in energy expenditure during sleep (CLA: 0 ± 38 kcal; placebo: –43 ± 90 kcal). We did not detect a change in labeled dietary fat oxidation after 6 mo of CLA supplementation given with a breakfast meal.
Conclusion: Mixed isomer CLA supplementation, but not placebo, positively altered fat oxidation and energy expenditure during sleep.
Key Words: Energy expenditure • fat oxidation • substrate utilization • overweight • obesity treatment
INTRODUCTION
Conjugated linoleic acid (CLA) is a family of positional and geometric isomers with 2 conjugated double bonds formed from linoleic acid, which displays a wide range of biological effects. CLA-related health benefits first received notice in 1985 when Pariza and Hargraves (1) isolated a compound in grilled ground beef that inhibited epidermal carcinogenesis in mice. In 1997, CLA's effects on reducing body fat were first reported (2). CLA is being further evaluated for its potential role in atherosclerosis (3-5), the immune response (6-9), body composition (2, 10, 11), lipid metabolism (12, 13), and glucose and insulin tolerance (14-16).
CLA's role in altering body composition is fairly well established in animal models, but studies in humans have shown mixed results. Some short-term human studies detected a decrease in body fat (17-21), whereas other studies did not detect an effect (11, 22). Gaullier et al (10, 23) conducted a 2-y study, the longest human study to date, and clearly showed a decrease of 2 kg in body weight and body fat mass in an overweight mixed-sex population with no serious adverse events.
The mechanism through which CLA attenuates adiposity is thought to involve either adipocyte apoptosis (24-26) or decreased adipocyte cell size secondary to a decrease in triacylglycerol uptake (27-29). Others proposed that CLA decreases percentage body fat by increasing energy expenditure (30), fat oxidation, or both (27); yet, when an increase in fat oxidation was detected, it was unknown whether it was from increased dietary fat oxidation or endogenous fat oxidation. Numerous studies investigated the effect of CLA on the fatty acid composition of agricultural production animals and found an increased ratio of saturated fatty acids to monounsaturated fatty acids (31-33); however, little is known about the ability of CLA to alter the type of fatty acids being oxidized in the body.
The primary aim of the present study was to investigate whether CLA increases total fat oxidation under eucaloric conditions. Secondary aims were to test for any alterations in dietary fat oxidation by using stable isotopes, endogenous fat, or potentially both and to test for any changes in energy expenditure by using a whole-room indirect calorimeter. Our study was a substudy performed as part of a larger study for which the primary aim was to determine whether 3.2 g/d of CLA for 6 mo would alter body fat mass in an overweight adult population (34).
SUBJECTS AND METHODS
Protocol
The present study was done as part of a double-blind, randomized, 6-mo intervention trial. The first participants began the trial in August 2004, and the last subjects ended the trial in March 2005. Subjects were screened and then underwent baseline evaluation. The substudy evaluation measured 24-h energy expenditure and substrate utilization by using a whole-room indirect calorimeter. Dietary fat oxidation was measured by mixing [1-13C]oleate and D31-palmitate into a breakfast meal and then collecting breath carbon dioxide and urine to measure the end products of oxidation. Subjects were then provided either 4 g/d of 78% active CLA isomers (3.2 g/d: 39.2% cis-9,trans-11 and 38.5% trans-10,cis-12) or 4 g/d of safflower oil placebo as 1-g gel capsule supplements. The baseline evaluations were repeated 6 mo later.
Subjects
The aim of the main study was to compare the effects of 4 g/d of 78% active CLA isomers (3.2 g/d: 39.2% cis-9,trans-11 and 38.5% trans-10,cis-12) and 4 g/d of safflower oil placebo supplement for 6 mo on body composition by using the 4-compartment model (34).
The main CLA efficacy study involved 48 overweight subjects, from which 23 subjects (4 men and 19 women) were selected at random for participation in the substudy. The substudy required 2 additional overnight visits to the University of Wisconsin-Madison Hospital's General Clinical Research Center (GCRC). During recruitment, subjects were informed about the substudy and were excluded from both studies if they chose to opt out of the substudy at the time of random assignment.
The main study and this substudy were approved by the Health Sciences Internal Review Board of the University of Wisconsin-Madison. The subjects provided written informed consent for both studies before the beginning of the study.
Inclusion criteria were the following: 18–44 y of age, overweight but not obese [body mass index (BMI; in kg/m2) 25 to <30], stable weight within 3 kg for at least the previous 6 mo, and no current or planned pregnancies. Exclusion criteria were the following: pregnancy or lactation, smoking, a physical limitation causing difficulty to participate in a prescribed walking program, recent use of a weight-loss program, metabolic disease history, psychiatric or eating disorders, abnormal electrocardiogram, fasting plasma cholesterol concentrations 300 mg/dL, and triacylglycerol concentrations 500 mg/dL.
Because of scheduling limitations, it was not possible to fully control for menstrual cycle. Of the 16 women, at baseline, 4 in the CLA group and 5 in the placebo group were in the first half of their menstrual cycle, whereas the remaining 5 and 2 women in the CLA and placebo groups, respectively, were in the second half of their menstrual cycle. For the final overnight visit, 2 women in the CLA group and 5 women in the placebo group were in the first half of their menstrual cycle, whereas the remaining 7 and 2 women in the CLA and placebo groups, respectively, were in the second half of their menstrual cycle.
Respiratory chamber
During the overnight visits in the GCRC, subjects stayed in a 12-m3 shell volume whole-room respiratory chamber for measurement of macronutrient utilization, energy expenditure based on carbon dioxide production (
Subjects arrived at the GCRC for admission at 2000; they had been asked to refrain from exercise for the 24 h before admission. Female subjects provided a urine sample at this time to test for pregnancy. Before entering the respiratory chamber at 2100, all subjects emptied their bladders. They were instructed to provide urine samples at 2300 on the night of admission and on the next day at the following time points: 0645, 0800, 1200, 1600, 2000, and 2200. Subjects were instructed to be in bed at 2300 and to lie quietly in the dark with minimal movements if unable to sleep. Subjects were awakened at 0645 and exited the chamber at 0700 for showering and other grooming. At 0745 they reentered the chamber, and at 0800 they provided a baseline breath sample. Once samples were collected, breakfast was given. This meal included a heated drink into which the tracer fats had been mixed. The 2 tracers were [1-13C]oleate (10 mg/kg body wt) and D31-palmitate (15 mg/kg body wt). Subjects then provided hourly breath samples and the remaining scheduled urine samples. If they had to void between urine collection time points, that urine was also collected. Meals were provided at 1200 and 1900, with a snack at 1600. Subjects were released from the respiratory chamber at 2100; however, they remained at the GCRC until their final urine sample was collected at 2200. This overnight stay was conducted before the start of CLA supplementation and then again at the end of the 6-mo supplementation period. Data from 3 time periods were analyzed: a 24-h time period; a waking time period, which was from 2100 to 2300 and from 0645 to 2100; and a sleeping time period, which was from 2300 to 0645.
Meals
A 3-d lead-in diet menu and the corresponding foods were provided to subjects before their admission dates. The lead-in diet consisted of frozen meals and shelf-stable items calculated to give 36% of energy as fat, 18% as protein, and 46% as carbohydrate. The diet was tailored to meet the energy needs of each subject on the basis of the Harris Benedict equation with 7% subtracted to account for the equation's tendency to overestimate caloric need for overweight and obese subjects. This was then multiplied by an assumed 1.65 physical activity level. If subjects did not eat all food listed on their menu, they recorded what they did not eat, and their lead-in diet records were adjusted accordingly. During hospital stays, subjects ate a diet of 36% fat, 18% protein, and 46% carbohydrate, which was formulated by the research dietitian. Breakfast, lunch, dinner, and the afternoon snack provided 25%, 30%, 35%, and 10% of their daily energy intake, respectively. The caloric need for each subject in the chamber was calculated in the same manner as the lead-in diet except that the physical activity level applied was only 1.4 because of the restrictive nature of the respiratory chamber. During the stay, subjects were instructed to eat all food provided. Any amount left on the plate was measured and subtracted from the inpatient diet records. The lead-in diet and inpatient diet did not allow for any caffeine or alcohol.
Sample collection and analysis
Nitrogen
Complete urine samples were collected and analyzed for total nitrogen, from which protein oxidation was calculated. While in the indirect calorimeter, the subjects' total volume of each void, scheduled and unscheduled, was recorded. From each void, 15-mL urine aliquots were immediately put into a 50-mL falcon tube containing citric acid (99% pure; Acros Organics) to prevent loss of ammonia. They were later pooled into a "waking" specimen and a "sleeping" specimen by taking an aliquot that was proportional to the volume of the void, thus creating a volume-weighted sampled. The 2 pooled samples for waking and sleeping were diluted 1:100 with nitrogen-free water. The diluted samples were then analyzed for total nitrogen in triplicate by using an Antek 9000N Chemiluminescent Nitrogen analyzer (Antek Instruments Inc, Houston, TX). This instrument was calibrated by using grammetric urea solutions, which were then used as standards during the sample runs. Pearson's regression coefficients for the calibration curve were >0.999.
Stable isotopes
The oxidation of stable isotope–labeled dietary fats was assessed on the basis of breath carbon dioxide and urinary water. The excess 13C recovery in the breath carbon dioxide was used to calculate dietary monounsaturated fat utilization during the chamber stay as previously described (35). Breath samples were collected by having subjects blow through a straw into a labeled 15-mL red top no-additive Vacutainer tube (Becton Dickinson, Franklin Lakes, NJ), which was capped immediately after being filled. The ratio of 13C:12C in the sample was analyzed by injecting a small amount into an isotope ratio gas chromatography mass spectrometer (Delta S; Finnigan MAT, Bremen, Germany). Each sample was injected in duplicate, and the mean 13C abundance (in ppm) was used in further calculations. The acceptable average SD 13C:12C analysis was 0.3 ppm. If the replicates were more variable, the analysis was repeated. 13CO2 recovery was calculated by subtracting baseline 13C in breath carbon dioxide taken at 0800, with correction for background 13C variations by using control subjects not dosed with labeled oleate, and then multiplying by the measured carbon dioxide production by using the trapezoid rule. An acetate correction, previously attained by testing sedentary subjects dosed with [13C]acetate in the respiratory chamber, was used to adjust for the loss of 13C in intermediary metabolism (35).
Deuterium recovery was calculated as previously described to determine dietary saturated fat utilization (35). At each scheduled time point for urine collection, 5-mL urine aliquots were put into a 5-mL Vacutainer with no additive and stored in a –70 °C freezer until analyzed for deuterium abundance. Each of the urine samples was decolorized by using carbon black and was then filtered through a 0.45-µm pore size acetate filter. The decolorized urine samples were then put into individual glass vials and capped, and the average 2H:1H ratio abundances (in ppm) were assessed in duplicate with the use of a Delta Plus isotope ratio mass spectrometer equipped with an H/D Device (Finnigan MAT). The acceptable average SD for 2H:1H analysis was 0.08 ppm. If results were more variable, the analysis was repeated. The deuterium abundances were corrected for each individual's baseline abundances in the urine specimen collected before the breakfast dose. The cumulative percentage dose recovery was calculated by multiplying the excess deuterium abundance at each time point by the subject's total body water, as measured by 18O dilution, then dividing by the amount of deuterium from the D31-palmitate provided in the breakfast shake. The resulting dietary fat oxidation value was assigned to the time of the midpoint between successive voids.
Statistical analysis
All statistical analyses were performed with SAS software (version 8.2; SAS Institute Inc, Cary, NC). Means ± SDs were calculated for the continuous outcomes. Between-group differences for changes from month 0 to month 6 were assessed by t tests and also by Wilcoxon's rank tests because normality could not be assured given the small sample size. The results were consistent between the 2 tests. Within-group differences for changes from month 0 to month 6 were assessed by paired t tests. All data were also analyzed with an adjustment for sex, which did not alter the findings. All values are presented as means ± SDs, and statistical significance was set at an level of 0.05 for all analyses.
RESULTS
The subjects were all overweight on the basis of BMI and were mostly female (Table 1). As expected with randomization, the participants of the substudy were representative of the main study at baseline. The 2 arms of the study, however, were not balanced, and this was due to 4 dropouts, all from the placebo group. One subject withdrew immediately after the initial inpatient testing because of time constraints, and another subject withdrew because of time constraints in scheduling the final visit. Two additional subjects were excluded because one started taking a medication known to affect body weight and the other began marathon training despite instructions not to alter exercise patterns. No data from these 4 individuals were used in the analysis. This left 12 subjects in the CLA treatment group and only 7 in the placebo group for data analysis. The decreased sample size may have limited our ability to attain the desired level of significance in the analysis of certain data, specifically the waking respiratory quotient. The only difference between the CLA and placebo groups at baseline was greater energy expenditure in the CLA group during sleep, which remained at month 6.
View this table:
TABLE 1. Study subjects' characteristics at baseline for the main study and the substudy1
Energy intake and energy balance
There were no differences between groups at month 0 or month 6 for energy intake for the inpatient diet or the lead-in diet. The macronutrient breakdown for the lead-in diet was 35 ± 0.6%, 18 ± 0.9%, and 48 ± 1.5% for fat, protein, and carbohydrate, respectively, for the CLA group and was 36 ± 1.5%, 18 ± 1.8%, and 48 ± 1.2% for fat, protein, and carbohydrate, respectively, for the placebo group. The inpatient dietary macronutrient breakdown for both groups was 18 ± 0.1% protein and 46 ± 0.1% carbohydrate, whereas fat was 36 ± 0.05% for the CLA group and 37 ± 0.05% for the placebo group.
24-h Time point
No differences were detected in 24-h energy expenditure or fat oxidation (Table 2). There was no between-group difference in protein utilization. Energy balance was also not different between groups at either month 0 or month 6.
View this table:
TABLE 2. 24-h Substrate utilization at baseline and at month 61
Waking hours
During waking hours (2100–2300 and 0645–2100), we did not detect a difference in energy expenditure or substrate oxidation either within or between groups (Table 3).
View this table:
TABLE 3. Waking substrate utilization at baseline and at month 61
Sleeping hours
During sleep (2300–0645), CLA treatment was associated with a number of differences in energy expenditure and substrate utilization (Table 4). The CLA group expended more energy than did the placebo group at baseline (P = 0.01) and at month 6 (P = 0.04). This appeared to be due to a sex mismatch between groups. The 3 male subjects had greater resting metabolic rates than did their female counterparts in the CLA group, whereas there were no male subjects in the placebo group; nonetheless, the change in sleeping energy expenditure in the placebo group (–43 ± 90 kcal/sleep) also differed from that in the CLA group (0 ± 38 kcal/sleep; P < 0.05). At baseline, the CLA group had a higher protein oxidative utilization, in terms of percentage of energy (P = 0.02) and grams per night (P = 0.02), than did the placebo group. After 6 mo, the CLA group utilized less protein for oxidation than at baseline in terms of percentage of energy (P = 0.001) and grams per night (P = 0.004). We also detected a between-group difference for change in protein utilization as a percentage of energy; the decrease with treatment was greater in the CLA group (P = 0.03).
View this table:
TABLE 4. Sleeping substrate utilization at baseline and at month 61
The change in fat oxidation during sleep (0–6 mo) was significantly different in the CLA group (4 ± 8 g/sleep) from that in the placebo group (–7 ± 11 g/sleep; P < 0.05). Compared with the placebo group, the CLA group was less dependent on carbohydrates (%) and utilized fewer carbohydrates after 6 mo of supplementation (P = 0.05). In turn, this shift in fat and carbohydrate oxidation in the CLA group resulted in a lower RQ after 6 mo than that observed in the placebo group (P = 0.05).
Dietary fatty acid oxidation
No difference in the percentage cumulative dose recovery of D31-palmitate (P = 0.89) or [1-13C]oleate (P = 0.99) was detected between the 2 groups 13 h after administration (Figures 1 and 2).
FIGURE 1.. Mean (±SEM) cumulative percentage dose recovery of dietary [2H]palmitate in the placebo (n = 7) and conjugated linoeic acid (CLA; n = 12) groups. No statistically significant differences were detected (t test).
FIGURE 2.. Mean (±SEM) cumulative percentage dose recovery of dietary [13C]oleate in the placebo (n = 7) and conjugated linoeic acid (CLA; n = 12) groups. No statistically significant differences were detected (t test).
DISCUSSION
We recently reported the results of the main efficacy study of CLA supplementation (34). CLA supplementation for 6 mo was associated with a decrease in weight (–0.6 ± 2.5 kg), whereas a slight weight gain was observed with placebo (1.1 ± 3.2 kg) (P = 0.04). These changes were largely due to changes in fat mass after 6 mo, which were also different between the CLA and placebo groups (–1.0 ± 2.2 and 0.7 ± 3 kg, respectively; P = 0.02). In our current study, we investigated the physiologic mechanism responsible for the CLA-related loss in body fat. We found that fat oxidation and energy expenditure increased during sleep in subjects who were randomly assigned to the CLA group; however, our primary outcome of 24-h fat oxidation only trended toward an increase after 6 mo of CLA supplementation.
The effects of CLA on fat and carbohydrate oxidation during sleep were consistent with the nonsignificant decrease in RQ during the 24-h period. The increased fat oxidation was also consistent with results from an animal study in which it was reported that CLA decreased the RQ in mice during sleep with no difference detected during waking hours (37). These results suggest that the mechanism for the efficacy of CLA supplementation for fat loss is at least partially due to increased fat oxidation during sleep.
Additional studies with a larger sample size are required to determine whether the lack of statistical significance for waking hours was a type II error. One possible reason for not finding a significant difference during waking hours was that the SD of fat oxidation in grams was smaller during sleep than during waking hours in our subjects (P = 0.005), which reduced the sensitivity of our study with regard to fat oxidation during waking or 24-h periods.
Our findings differ from those of previous studies of energy expenditure and substrate utilization (10, 22, 27). A human study conducted by Zambell et al (22) found no treatment effect at rest or during a walking protocol in terms of energy expenditure, fat oxidation, and respiratory exchange ratios for 20 min after 30 min of rest at baseline, during week 4 and week 8 of treatment. Their measurement time period was much shorter than our 24-h continuous monitoring; thus, they may not have detected a treatment effect because of short measurement times and the accompanying potential for a type II error. Azain et al (27) found no treatment effect on energy expenditure by indirect calorimetry in rats placed in a respiratory chamber for 48 h at weeks 1, 4, and 7 of treatment. A decrease in RQ was detected at week 4; however, there was no effect at week 7 (27). The irregularity of RQ data from week 4 to week 7 may be a result of the time period length analyzed. Because Azain et al collected data for an entire 48-h time period, the potential sleep time effects of CLA may have been obscured by the greater variation during waking just as they were in our study. Thus, future CLA studies analyzing energy expenditure and macronutrient oxidation would benefit from doing analysis for an entire 24-h period and then further dividing these periods into waking and sleeping hours.
A confounder of our study was that sleeping protein oxidative utilization differed between the CLA and placebo groups at baseline. To ensure that this did not affect our global RQ data, a nonprotein respiratory quotient (NPRQ) was calculated. After 6 mo of supplementation, NPRQ decreased in the CLA group and increased in the placebo group. In addition to the baseline difference, protein utilization also decreased in the CLA group with treatment. This decrease in protein utilization is consistent with the findings from animal model studies (38, 39) that CLA protected against urinary protein losses in an attempt to protect lean body mass stores.
The second aim of our study was to determine whether the alteration in fat oxidation was due to a change in dietary fat, endogenous fat, or both. We did not detect a change in the oxidation of dietary fat as monitored by labeled oleate and palmitate administered during a breakfast meal. This would suggest that the change in fat oxidation was due to increased endogenous fat. Considering, however, that increased fat oxidation was observed during sleep and that the dietary fat oxidation measures were made during the waking period, further studies of dietary fat oxidation during sleeping hours by using stable-isotope–labeled fats given at a dinner meal are warranted to fully assess CLA's effects on dietary fat oxidation.
We did not directly test any biochemical mechanisms that may have underlain our observation of increased fat oxidation during sleep. Current theories that may explain our results, however, include mechanisms associated with the inhibition of fat deposition within adipocytes, altered lipase activity, and apoptosis of adipocytes. The increase in fat oxidation was likely to have been associated with lipases involved in the storage and mobilization of fat, which are lipoprotein lipase (LPL), hormone-sensitive lipase (HSL) (40), and adipose triacylglycerol lipase (ATGL; 41). CLA supplementation studies have shown decreased adipocyte LPL activity and increased HSL activity (42, 43). No studies to date have researched CLA's effects on ATGL activity. Because rat studies have shown that adipocyte LPL and HSL are linked to circadian rhythms, which lead to a decrease in adipocyte LPL and an increase in HSL during sleeping hours, it is plausible that our sleeping compared with waking results are explained by CLA's effect being further enhanced during the diurnal cycle (44, 45). The increase in nighttime fat oxidation that we observed without an increase in dietary fat oxidation during the day is more consistent with proposed mechanisms involving greater fat mobilization than with those involving reduced storage.
As reported elsewhere, we found that body fat mass decreased by 1 ± 2.2 kg with CLA treatment, whereas it increased by 0.7 ± 3.0 kg with placebo after 6 mo (34). These changes corresponded to an average decrease of 5.5 g fat/d compared with an increase of 3.8 g fat/d in the placebo group. These changes are smaller than those that we observed in fat oxidation during the sleep period and are also smaller than the trend in 24-h fat oxidation. As such, these changes in fat oxidation were sufficient to account for the effect of CLA on body fat mass. In contrast, with calculation of the change in energy balance in each group, assuming an energy value of 9.5 kcal/g of endogenous fat, we found that the change in energy balance was twice the change in sleep period energy expenditure and was much larger than the trend for the change in 24-h energy expenditure. This suggests that the changes in fat mass and hence energy balance that we observed in the main study and the changes in fat oxidation that we observed in the present study were related to a decrease in energy intake rather than to an increase in energy expenditure. This, however, can only be suggested because the 95% CIs for energy expenditure are relatively large.
We did not collect data that would have allowed us to determine whether the suggested change in energy intake may have been due to changes in energy absorption on CLA supplementation. Tepstra et al (46) investigated alterations in energy absorption before and after 39 d of CLA supplementation by measuring energy content and composition in feces and excreta. They observed an increase in energy excretion in Balb-c mice without a corresponding increase in energy intake compared with controls. Thus, future CLA research using humans could investigate CLA's ability to alter fecal energy excretion.
In conclusion, CLA altered fat oxidation during sleeping hours and showed a similar trend during waking hours. On the basis of the absence of an alteration in the oxidation of dietary labeled oleate and palmitate in a breakfast meal, CLA does not appear to affect the type of dietary fat oxidized during waking hours. Further studies could be directed at determining whether CLA supplementation alters the type of endogenous fat oxidized on greater utilization of fat by the body, whether dietary fat oxidation is increased during sleeping hours, and whether energy excretion is altered.
ACKNOWLEDGMENTS
We thank the staff of the GCRC for their invaluable efforts and Zhumin Zhang for performing the statistical analysis.
The authors' responsibilities were as follows—RNC: collected and analyzed the data and drafted the manuscript; DAS: designed the study and contributed significantly to the interpretation of the data and to the manuscript revisions; ACW: conducted the associated study and assisted in data collection; and EHN: assisted in monitoring patient safety and health during the length of the study. None of the authors had any conflict of interest.
REFERENCES