Nan
Yao
a,
Shoumeng
Yan
a,
Yinpei
Guo
a,
Han
Wang
a,
Xiaotong
Li
a,
Ling
Wang
a,
Wenyu
Hu
a,
Bo
Li
*a and
Weiwei
Cui
*b
aDepartment of Epidemiology and Biostatistics, School of Public Health, Jilin University, Changchun, 130021, P. R. China. E-mail: li_bo@jlu.edu.cn; Tel: +86 431 85619451
bDepartment of Nutrition and Food Hygiene, School of Public Health, Jilin University, Changchun, 130021, P. R. China. E-mail: cuiweiwei@jlu.edu.cn; Tel: +86 431 85619455
First published on 12th May 2021
Background: Excess body weight, including overweight and obesity, is one of the major factors influencing human health, and plays an important role in the global burden of disease. Carotenoids serve as precursors of vitamin A-related retinoids, and are considered to have potential effects on many diseases. However, the influence of carotenoids on people with excess body weight is unclear. Methods: This meta-analysis was conducted to assess the effects of carotenoids on overweight or obese subjects utilizing the available evidence. We searched PubMed, Medline, Cochrane Library, Web of Science and EMBASE databases up to September 2020. Random effects models were used to calculate the standard mean differences (SMDs) and odds ratios (ORs) with their 95% confidence intervals (95% CIs). Results: A total of seven randomized controlled trials and eight observational studies met the inclusion criteria and contained 28944 subjects and data on multiple carotenoid subgroups, including lycopene, astaxanthin, cryptoxanthin, α-carotene, and β-carotene. In all included Randomized Controlled Trial (RCT), the intervention duration was 20 days at the shortest and 16 weeks at the longest, and the range of intervention doses was 1.2–60 mg d−1. Our study found that the insufficiency of serum carotenoids was a risk factor for overweight and obesity (OR = 1.73, 95% CI [1.57, 1.91], p < 0.001). Moreover, carotenoid supplementation was significantly associated with body weight reductions (SMD = −2.34 kg, 95% CI [−3.80, −0.87] kg, p < 0.001), body mass index decrease (BMI, SMD = −0.95 kg cm−2, 95% CI [−1.88, −0.01] kg cm−2, p < 0.001) and waist circumference losses (WC, SMD = −1.84 cm, 95% CI [−3.14, −0.54]cm, p < 0.001). Conclusion: In summary, the carotenoids show promising effects in overweight or obese subjects. Additional data from large clinical trials are needed.
Carotenoids are fat-soluble pigments found in plants, fungi, bacteria, algae and human foods.6 For example, green vegetables contain large amounts of lutein, lycopene is present in mature tomatoes, chili peppers include capsaicin, and crustaceans have high levels of astaxanthin. Moreover, multiple carotenoids, such as β-carotene, are commonly found in human serum. In addition, α-carotene, β-cryptoxanthin, lycopene, lutein and zeaxanthin have been observed.7 Notably, carotenoids are considered to have potential effects on many diseases. A positive connection between higher concentrations of carotenoids and a lower risk of chronic diseases has been illustrated by epidemiological studies, while β-carotene and lycopene are negatively related to the risk of cardiovascular disease.8 Furthermore, fat-soluble carotenoids were shown to be present in lipid droplets within adipocytes and have effects on lipid absorption and transport,9 indicating a correlation between carotenoids and excess body weight.10
However, the relationship between carotenoid concentrations and subjects with overweight or obesity has not been demonstrated in meta-analyses. Carotenoid interventions were reported to cause reductions in weight, BMI and other anthropometric measures in overweight and obese people.11,12 Nevertheless, some studies have shown that carotenoid interventions have no significant effect on weight change in patients with excess body weight.13,14 The effect of carotenoids on people with excess body weight is unclear. Consequently, we conducted a meta-analysis of all related observational studies and randomized controlled trials (RCTs) to evaluate the association between carotenoids and overweight or obese subjects.
Exclusion criteria were as follows: (1) single-arm studies or without a placebo or a mixture of carotenoid and antioxidant intervention or unquantifiable doses of carotenoid intervention; (2) patients with cancer, pregnancy and any other medication that could influence the carotenoid concentrations; and (3) nonhuman studies, reviews and conference literature.
Cochran's Q statistic and the I2 statistic were used to evaluate the statistical heterogeneity. p < 0.05 was defined as significant for heterogeneity. Egger's test was used to calculate the publication bias. The potential publication bias was evaluated via Egger's test, where the trim-and-fill method (sensitivity analysis) was used to correct outcomes and evaluate the impact of bias on the outcomes.
For the observational studies, the subgroup analyses were conducted based on the age (<18 years and ≥18 years), sex of the subjects (male, female and both of male and female), region (Asia, South America, North America and Oceania), population type (overweight and obesity) and carotenoid type (lycopene, astaxanthin, cryptoxanthin, zeaxanthin/lutein, α-carotene, β-carotene, total carotenoids and lutein/zeaxanthin).
Additionally, we used subgroup analyses based on the intervention time (≤12 weeks and >12 weeks), region (Asia and Europe), population type (overweight: 25 kg cm−2 ≤ BMI < 30 kg cm−2, obesity: BMI > 30 kg cm−2, and both overweight and obesity: BMI > 30 kg cm−2) and population sex (male, female and both of male and female) to evaluate the source of heterogeneity for the RCTs.
Author | Region | Year | N | Gender (M/F) | Gender | BMI (kg cm−2) | OR (95%CI) | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Case | Control | Case | Control | Male | Female | Case | Control | Case | Control | |||
Data were shown as the means ± S.D.; α car, α-carotene; ASTA, astaxanthin; β car, β-carotene; CRY, cryptoxanthin; car, total carotenoids; LYC, lycopene; lut/zea, lutein/zeaxanthin; zea/lut, zeaxanthin/lutein; F, female; M, male; pre, premenopausal; post, postmenopausal; o, old subjects (≥65 years); y, young subjects (19 ≤ 65 years); 1, obese; 2, overweight.a Some basic data of the literature are missing.b Weight/height z-score (Obesity ZWH ≥2).c Men.d Women.e Data shown as: % (SE). | ||||||||||||
Rebecca et al. [β car-M] | Brazil | 2019 | 40 | 49 | 40/0 | 49/0 | 89 | 0 | — | — | 1.46 (1.2, 1.8) | 1 |
Rebecca et al. [β car-F] | Brazil | 2019 | 44 | 105 | 0/44 | 0/105 | 0 | 149 | — | — | 1.19 (1, 1.4) | 1 |
Inong R. et al. [α car-2] | U.S.A. | 2014 | 413a | 587a | 237/230a | 283/364a | 537 | 617 | — | — | 1.2 (0.68, 2.08) | 1 |
Inong R. et al. [β-car-2] | U.S.A. | 2014 | 413a | 587a | 237/230a | 283/364a | 537 | 617 | — | — | 1.75 (1.12, 2.7) | 1 |
Inong R. et al. [α car-1] | U.S.A. | 2014 | 413a | 587a | 237/230a | 237/230a | 537 | 617 | — | — | 2.17 (1.39, 3.45) | 1 |
Inong R.et al. [β car-1] | U.S.A. | 2014 | 413a | 587a | 237/230a | 237/230a | 537 | 617 | — | — | 2.86 (1.89, 4.35) | 1 |
Inong R. et al. [β car-2] | U.S.A. | 2014 | 413a | 587a | 237/230a | 237/230a | 537 | 617 | — | — | 1.59 (0.93, 2.78) | 1 |
Inong R. et al. [β car-1] | U.S.A. | 2014 | 413a | 587a | 237/230a | 237/230a | 537 | 617 | — | — | 4.17 (2.63, 6.25) | 1 |
Luciane et al. [car] | Brazil | 2007 | 72 | 399 | 34/38 | 217/182 | 251 | 220 | 17.24 ± 2.93 | 17.92 ± 3.83 | 2.51 (1.43, 4.39) | 1 |
Roseli et al. [car] | Brazil | 2005 | 23 | 23 | — | — | 24 | 22 | 3.36 ± 2.16b | 0.58 ± 0.74b | 0.08 (0.01, 0.71) | 1 |
Koji et al. [LYC-M] | Japan | 2006 | 55 | 137 | 55/0 | 137/0 | 192 | 0 | 26.97 ± 1.6 | 22.27 ± 1.9 | 0.64 (0.29, 1.37) | 1 |
Koji et al. [ASTA-M] | Japan | 2006 | 55 | 137 | 55/0 | 137/0 | 192 | 0 | 26.97 ± 1.6 | 22.27 ± 1.9 | 1.03 (0.5, 2.1) | 1 |
Koji et al. [CRY-M] | Japan | 2006 | 55 | 137 | 55/0 | 137/0 | 192 | 0 | 26.97 ± 1.6 | 22.27 ± 1.9 | 0.89 (0.41, 1.9) | 1 |
Koji et al. [zea/lut-M] | Japan | 2006 | 55 | 137 | 55/0 | 137/0 | 192 | 0 | 26.97 ± 1.6 | 22.27 ± 1.9 | 1.55 (0.74, 3.24) | 1 |
Koji et al. [α car-M] | Japan | 2006 | 55 | 137 | 55/0 | 137/0 | 192 | 0 | 26.97 ± 1.6 | 22.27 ± 1.9 | 1.13 (0.55, 2.3) | 1 |
Koji et al. [β car-M] | Japan | 2006 | 55 | 137 | 55/0 | 137/0 | 192 | 0 | 26.97 ± 1.6 | 22.27 ± 1.9 | 0.83 (0.36, 1.84) | 1 |
Koji et al. [LYC-F] | Japan | 2006 | 119 | 279 | 0/119 | 0/279 | 0 | 398 | 27.07 ± 1.8 | 21.97 ± 2.0 | 1.15 (0.71, 1.88) | 1 |
Koji et al. [ASTA-F] | Japan | 2006 | 119 | 279 | 0/119 | 0/279 | 0 | 398 | 27.07 ± 1.8 | 21.97 ± 2.0 | 1.52 (0.94, 2.44) | 1 |
Koji et al. [CRY-F] | Japan | 2006 | 119 | 279 | 0/119 | 0/279 | 0 | 398 | 27.07 ± 1.8 | 21.97 ± 2.0 | 1.74 (1.07, 2.84) | 1 |
Koji et al. [zea/lut-F] | Japan | 2006 | 119 | 279 | 0/119 | 0/279 | 0 | 398 | 27.07 ± 1.8 | 21.97 ± 2.0 | 1.31 (0.79, 2.16) | 1 |
Koji et al. [αcar-F] | Japan | 2006 | 119 | 279 | 0/119 | 0/279 | 0 | 398 | 27.07 ± 1.8 | 21.97 ± 2.0 | 1.38 (0.84, 2.25) | 1 |
Koji et al. [β car-F] | Japan | 2006 | 119 | 279 | 0/119 | 0/279 | 0 | 398 | 27.07 ± 1.8 | 21.97 ± 2.0 | 1.4 (0.85, 2.31) | 1 |
Allison et al. [car-2] | Australia | 2011 | — | — | — | — | 280 | 617 | 27.0 (23.5–30.9)c | 27.8 (23.5–32.1)d | 0.76 (0.32, 1.79) | 1 |
Allison et al. [car-1] | Australia | 2011 | — | — | — | — | 280 | 617 | 27.0 (23.5–30.9)c | 27.8 (23.5–32.1)d | 1.43 (0.55, 3.7) | 1 |
Koji et al. [CRY-M] | Japan | 2003 | 50 | 108 | 50/0 | 108/0 | 158 | 0 | 27.7 ± 2.4 | 22.0 ± 2.0 | 2.13 (0.81, 5.56) | 1 |
Koji et al. [zea/lut-M] | Japan | 2003 | 50 | 108 | 50/0 | 108/0 | 158 | 0 | 27.7 ± 2.4 | 22.0 ± 2.0 | 1.69 (0.67, 4.35) | 1 |
Koji et al. [α car-M] | Japan | 2003 | 50 | 108 | 50/0 | 108/0 | 158 | 0 | 27.7 ± 2.4 | 22.0 ± 2.0 | 4.35 (1.54, 12.5) | 1 |
Koji et al. [β car-M] | Japan | 2003 | 50 | 108 | 50/0 | 108/0 | 158 | 0 | 27.7 ± 2.4 | 22.0 ± 2.0 | 3.03 (1.09, 8.33) | 1 |
Koji et al. [LYC-F] | Japan | 2003 | 52 | 106 | 0/52 | 0/106 | 0 | 158 | 22.1 ± 2.0 | 27.6 ± 1.9 | 2.33 (0.93, 5.88) | 1 |
Koji et al. [CRY-F] | Japan | 2003 | 52 | 106 | 0/52 | 0/106 | 0 | 158 | 22.1 ± 2.0 | 27.6 ± 1.9 | 1.92 (0.76, 4.76) | 1 |
Koji et al. [zea/lut-F] | Japan | 2003 | 52 | 106 | 0/52 | 0/106 | 0 | 158 | 22.1 ± 2.0 | 27.6 ± 1.9 | 1.64 (0.65, 4) | 1 |
Koji et al. [α car-F] | Japan | 2003 | 52 | 106 | 0/52 | 0/106 | 0 | 158 | 22.1±.0 | 27.6 ± 1.9 | 2.86 (1.14, 7.14) | 1 |
Koji et al. [β car-F] | Japan | 2003 | 52 | 106 | 0/52 | 0/106 | 0 | 158 | 22.1±.0 | 27.6 ± 1.9 | 2.56 (1.01, 6.67) | 1 |
Joel E. et al. [α car-F-pre-1] | U.S.A. | 2006 | 1320 | 1980 | 0/1320 | 0/1980 | 0 | 3300 | 22.71 (1.22)e | 55.18 (1.40)e | 4.44 (3.37, 5.84) | 1 |
Joel E. et al. [α car-F-pr6-2] | U.S.A. | 2006 | 1212 | 1980 | 0/1212 | 0/1980 | 0 | 3192 | 22.10 (1.10)e | 55.18 (1.40)e | 1.82 (1.39, 2.38) | 1 |
Joel E. et al. [α car-F-post-1] | U.S.A. | 2006 | 1267 | 1239 | 0/1267 | 0/1239 | 0 | 2506 | 30.13 (1.05)e | 36.88 (1.31)e | 2.75 (2.03, 3.72) | 1 |
Joel E. et al. [α car-F-post-2] | U.S.A. | 2006 | 1365 | 1239 | 0/1365 | 0/1239 | 0 | 2604 | 32.99 (1.04)e | 36.88 (1.31)e | 1.47 (1.47, 1.91) | 1 |
Joel E. et al. [α car-M-y-2] | U.S.A. | 2006 | 1244 | 2346 | 1244/0 | 2346/0 | 3590 | 0 | 19.37 (0.71)e | 41.57 (1.04)e | 2.67 (2.01, 3.56) | 1 |
Joel E. et al. [α car-M-y-1] | U.S.A. | 2006 | 2285 | 2346 | 2285/0 | 2346/0 | 4631 | 0 | 39.06 (0.90)e | 41.57 (1.04)e | 1.25 (1.01, 1.55) | 1 |
Joel E. et al. [α car-M-o-2] | U.S.A. | 2006 | 363 | 716 | 363/0 | 716/0 | 1079 | 0 | 21.51 (1.55)e | 33.47 (1.74)e | 1.52 (0.98, 2.35) | 1 |
Joel E. et al. [α car-M-o-1] | U.S.A. | 2006 | 854 | 716 | 854/0 | 716/0 | 1570 | 0 | 45.02 (1.82)e | 33.47 (1.74)e | 1.39 (1.02, 1.89) | 1 |
Joel E. et al. [β car-F-pre-1] | U.S.A. | 2006 | 1320 | 1980 | 0/1320 | 0/1980 | 0 | 3300 | 22.71 (1.22)e | 55.18 (1.40)e | 6.16 (4.35, 8.74) | 1 |
Joel E. [β car-F-pre-2] | U.S.A. | 2006 | 1212 | 1980 | 0/1212 | 0/1980 | 0 | 3192 | 22.10 (1.10)e | 55.18 (1.40)e | 2.05 (1.56, 2.69) | 1 |
Joel E. et al. [β car-F-post-1] | U.S.A. | 2006 | 1267 | 1239 | 0/1267 | 0/1239 | 0 | 2506 | 30.13 (1.05)e | 36.88 (1.31)e | 2.93 (1.99, 4.32) | 1 |
Joel E. et al. [β car-F-post-2] | U.S.A. | 2006 | 1365 | 1239 | 0/1365 | 0/1239 | 0 | 2604 | 32.99 (1.04)e | 36.88 (1.31)e | 1.72 (1.2, 2.48) | 1 |
Joel E. et al. [β car-M-y-2] | U.S.A. | 2006 | 1244 | 2346 | 1244/0 | 2346/0 | 3590 | 0 | 19.37 (0.71)e | 41.57 (1.04)e | 2.71 (2, 3.69) | 1 |
Joel E. et al. [β car-M-y-1] | U.S.A. | 2006 | 2285 | 2346 | 2285/0 | 2346/0 | 4631 | 0 | 39.06 (0.90)e | 41.57 (1.04)e | 1.4 (1.14, 1.72) | 1 |
Joel E. et al. [β car-M-o-2] | U.S.A. | 2006 | 363 | 716 | 363/0 | 716/0 | 1079 | 0 | 21.51 (1.55)e | 33.47 (1.74)e | 2.06 (1.22, 3.48) | 1 |
Joel E. et al. [β car-M-o-1] | U.S.A. | 2006 | 854 | 716 | 854/0 | 716/0 | 1570 | 0 | 45.02 (1.82)e | 33.47 (1.74)e | 1.68 (1.15, 2.46) | 1 |
Joel E. et al. [CRY-F-pre-1] | U.S.A. | 2006 | 1320 | 1980 | 0/1320 | 0/1980 | 0 | 3300 | 22.71 (1.22)e | 55.18 (1.40)e | 4.21 (3, 5.92) | 1 |
Joel E. et al. [CRY-F-post-1] | U.S.A. | 2006 | 1267 | 1239 | 0/1267 | 0/1239 | 0 | 2506 | 30.13 (1.05)e | 36.88 (1.31)e | 2.47 (1.86, 3.27) | 1 |
Joel E. et al. [CRY-F-post-2] | U.S.A. | 2006 | 1365 | 1239 | 0/1365 | 0/1239 | 0 | 2604 | 32.99 (1.04)e | 36.88 (1.31)e | 1.4 (1.09, 1.79) | 1 |
Joel E. et al. [CRY-M-y-2] | U.S.A. | 2006 | 1244 | 2346 | 1244/0 | 2346/0 | 3590 | 0 | 19.37 (0.71)e | 41.57 (1.04)e | 2.4 (1.69, 3.41) | 1 |
Joel E. et al. [CRY-M-y-1] | U.S.A. | 2006 | 2285 | 2346 | 2285/0 | 2346/0 | 4631 | 0 | 39.06 (0.90)e | 41.57 (1.04)e | 1.18 (0.96, 1.46) | 1 |
Joel E. et al. [CRY-M-o-2] | U.S.A. | 2006 | 363 | 716 | 363/0 | 716/0 | 1079 | 0 | 21.51 (1.55)e | 33.47 (1.74)e | 1.43 (0.83, 2.44) | 1 |
Joel E. et al. [CRY-M-o-1] | U.S.A. | 2006 | 854 | 716 | 854/0 | 716/0 | 1570 | 0 | 45.02 (1.82)e | 33.47 (1.74)e | 1.19 (0.77, 1.83) | 1 |
Joel E. et al. [LYC-F-pre-1] | U.S.A. | 2006 | 1320 | 1980 | 0/1320 | 0/1980 | 0 | 3300 | 22.71 (1.22)e | 55.18 (1.40)e | 2.07 (1.48, 2.88) | 1 |
Joel E. et al. [LYC-F-pre-2] | U.S.A. | 2006 | 1212 | 1980 | 0/1212 | 0/1980 | 0 | 3192 | 22.10 (1.10)e | 55.18 (1.40)e | 1.39 (1, 1.93) | 1 |
Joel E. et al. [LYC-F-post-1] | U.S.A. | 2006 | 1267 | 1239 | 0/1267 | 0/1239 | 0 | 2506 | 30.13 (1.05)e | 36.88 (1.31)e | 1.83 (1.27, 2.62) | 1 |
Joel E. et al. [LYC-F-post-2] | U.S.A. | 2006 | 1365 | 1239 | 0/1365 | 0/1239 | 0 | 2604 | 32.99 (1.04)e | 36.88 (1.31)e | 1.37 (1.08, 1.74) | 1 |
Joel E. et al. [LYC-M-y-2] | U.S.A. | 2006 | 1244 | 2346 | 1244/0 | 2346/0 | 3590 | 0 | 19.37 (0.71)e | 41.57 (1.04)e | 1.27 (0.99, 1.63) | 1 |
Joel E. et al. [LYC-M-y-1] | U.S.A. | 2006 | 2285 | 2346 | 2285/0 | 2346/0 | 4631 | 0 | 39.06 (0.90)e | 41.57 (1.04)e | 0.98 (0.77, 1.24) | 1 |
Joel E. et al. [LYC-M-o-2] | U.S.A. | 2006 | 363 | 716 | 363/0 | 716/0 | 1079 | 0 | 21.51 (1.55)e | 33.47 (1.74)e | 0.77 (0.49, 1.2) | 1 |
Joel E. et al. [LYC-M-o-1] | U.S.A. | 2006 | 854 | 716 | 854/0 | 716/0 | 1570 | 0 | 45.02 (1.82)e | 33.47 (1.74)e | 1.06 (0.71, 1.58) | 1 |
Joel E. et al. [lut/zea-F-pre-1] | U.S.A. | 2006 | 1320 | 1980 | 0/1320 | 0/1980 | 0 | 3300 | 22.71 (1.22)e | 55.18 (1.40)e | 3.7 (2.66, 5.16) | 1 |
Joel E.et al. [lut/zea-F-pre-2] | U.S.A. | 2006 | 1212 | 1980 | 0/1212 | 0/1980 | 0 | 3192 | 22.10 (1.10)e | 55.18 (1.40)e | 1.74 (1.29, 2.35) | 1 |
Joel E. et al. [lut/zea-F-post-1] | U.S.A. | 2006 | 1267 | 1239 | 0/1267 | 0/1239 | 0 | 2506 | 30.13 (1.05)e | 36.88 (1.31)e | 2.49 (1.85, 3.37) | 1 |
Joel E. et al. [lut/zea-F-post-2] | U.S.A. | 2006 | 1365 | 1239 | 0/1365 | 0/1239 | 0 | 2604 | 32.99 (1.04)e | 36.88 (1.31)e | 1.46 (1.06, 2.02) | 1 |
Joel E. et al. [lut/zea-M-y-2] | U.S.A. | 2006 | 1244 | 2346 | 1244/0 | 2346/0 | 3590 | 0 | 19.37 (0.71)e | 41.57 (1.04)e | 1.81 (1.37, 2.4) | 1 |
Joel E. et al. [lut/zea-M-y-1] | U.S.A. | 2006 | 2285 | 2346 | 2285/0 | 2346/0 | 4631 | 0 | 39.06 (0.90)e | 41.57 (1.04)e | 1 (0.75, 1.32) | 1 |
Joel E. et al. [lut/zea-M-o-2] | U.S.A. | 2006 | 363 | 716 | 363/0 | 716/0 | 1079 | 0 | 21.51 (1.55)e | 33.47 (1.74)e | 1.85 (1.2, 2.86) | 1 |
Joel E. et al. [lut/zea-M-o-1] | U.S.A. | 2006 | 854 | 716 | 854/0 | 716/0 | 1570 | 0 | 45.02 (1.82)e | 33.47 (1.74)e | 1.68 (1.17, 2.42) | 1 |
Author | Region | Year | Intervention | N | Age (years) | Gender (M/F) | BMI (kg cm−2) | Weight (kg) | WC (cm) | TG (mg dl−1) | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Dose (mg) | Time (week) | Study | Control | SG | Control | SG | Control | SG | Control | SG | Control | SG | Control | SG | Control | |||
Data were shown as mean ± S.D.a Mean ± S.E.; SG, supplementation groups; BMI, body mass index; WC, waist circumference; TG, triglyceride; NAFLD, non-alcoholic fatty liver disease; NLF, normal liver fat. | ||||||||||||||||||
Ryo et al. | Japan | 2018 | 9 | 12 | 41 | 39 | 48.9 ± 1.38a | 50.8 ± 1.39a | 35/6 | 32/7 | 27.19 ± 0.24 | 27.09 ± 0.22 | 77.81 ± 1.00a | 75.48 ± 1.14a | 89.02 ± 0.91a | 87.89 ± 0.83a | 131.8 ± 12.6a | 141.1 ± 13.9a |
Zohre et al. | Iran | 2015 | 60 | 20/7 | 40 | 35 | 20–30 | 20–30 | 0/40 | 0/35 | 28.22 ± 0.35 | 28.28 ± 0.29 | 71.82 ± 1.31a | 72.39 ± 1.19a | — | — | — | — |
Hye et al. | South Korea | 2011 | 20 | 12 | 14 | 13 | 31.1 ± 9.4 | 30.1 ± 9.5 | 12/2 | 11/2 | 28.1 ± 2.4 | 26.3 ± 1.3 | 83.6 ± 9.4 | 77.1 ± 10.8 | 97.1 ± 6.3 | 92.1 ± 6.2 | 110.6 ± 51.5 | 113.4 ± 40.5 |
Fatemeh et al. | Iran | 2020 | 6 | 12 | 23 | 23 | 38.1 ± 7.6 | 35.6 ± 9.1 | 11/12 | 11/12 | 33.2 ± 3.3 | 33.1 ± 3.9 | 93.8 ± 13.8 | 95.5 ± 17.0 | 103.6 ± 8.1 | 104.4 ± 11.9 | — | — |
Akira et al.-trial1 | Japan | 2018 | 1.2 | 12 | 13 | 10 | 41 ± 9 | 44 ± 9 | 13/0 | 10/0 | 28.2 ± 2.9 | 28.2 ± 2.7 | 83.0 ± 10.1 | 83.5 ± 9.7 | — | — | — | — |
Akira et al.-trail2 | Japan | 2018 | 2 | 12 | 46 | 45 | 44 ± 11 | 43 ± 10 | 46/0 | 45/0 | 26.9 ± 1.9 | 27.4 ± 1.8 | 80.5 ± 9.6 | 81.0 ± 7.8 | 94.4 ± 5.6 | 94.5 ± 5.7 | — | — |
M. Abidov et al.-NAFLD | Russia | 2010 | 2.4 | 16 | 36 | 36 | 36.1 ± 2.1a | 37.4 ± 2.8a | 0/36 | 0/36 | >30 | >30 | 94.1 ± 2.1 | 93.5 ± 2.4 | 110.6 ± 1.6 | 109.2 ± 1.4 | 195 ± 19 | 191 ± 15 |
M. Abidov et al.-NLF | Russia | 2010 | 2.4 | 16 | 19 | 19 | 34.7 ± 3.5a | 34.7 ± 3.2a | 0/19 | 0/19 | >30 | >30 | 94.5 ± 2.1 | 93.9 ± 1.4 | 103.1 ± 1.7 | 102.2 ± 1.4 | 177 ± 12 | 174 ± 12 |
Maria et al. | Denmark | 2019 | 7 | 12 | 6 | 6 | 56.2 ± 5.9 | 56.1 ± 5.8 | 2/4 | 4/2 | 32.7 ± 3.3 | 33.8 ± 3.5 | — | — | — | — | 136 ± 13.8 | 136 ± 13.8 |
We performed a meta-analysis of the carotenoid concentration to evaluate whether low carotenoid concentrations were a risk factor in obese or overweight people. Low serum carotenoid levels were a risk factor for overweight or obese subjects compared to control subjects (OR = 1.73, 95% CI [1.57, 1.91], p < 0.001) (Fig. 2). Moreover, publication bias was not observed in the serum carotenoid concentration, as shown by Egger's test (coefficient = −0.5664128, t = −0.95, p = 0.346).
Fig. 2 Meta-analysis results of the serum carotenoid concentration in the subjects with obesity or overweight vs. the control subjects. |
In addition, we performed subgroup analysis based on the age, sex, region, population type, and carotenoid type. Studies were subdivided into two groups: the adults (age ≥18 years) and the minors (age <18 years). For both groups, the risk of insufficient serum carotenoid concentration was higher in the subjects with excess weight than the normal. Moreover, the subgroup analysis was conducted to determine if there were differences in the effects of the region: Asia, South America, North America and Oceania. The risk of insufficient serum carotenoid concentration for the subjects with excess weight was higher than that for the normal weight subjects in the Asia group and the North America group. Based on the gender distribution of the subjects included, all studies were divided into three categories: male, female, and both male and female. The risk of low serum carotenoid concentrations was higher in all of the groups than in the controls. Moreover, studies were classified into two groups (overweight and obese) based on the BMI of the subjects. For the two groups, we observed a statistically significant difference between the two groups compared to the control group. In addition, the types of carotenoids in the included studies were lycopene, astaxanthin, cryptoxanthin, zeaxanthin/lutein, α-carotene, β-carotene, total carotenoids and lutein/zeaxanthin. The risk of low serum carotenoid concentrations was higher than that of the controls in all of the groups, except the astaxanthin group and total carotenoid group. In summary, subgroup analyses indicated that region and carotenoid type may contribute to the heterogeneity of the results (Table 3).
Grouped by | No. of studies | OR (95% CI), P | I 2 (%), P |
---|---|---|---|
BMI, body mass index; LYC, lycopene; ASTA, astaxanthin; CRY, cryptoxanthin; zea/lut, zeaxanthin/lutein; α car, α-carotene; β car, β-carotene; car, total carotenoids; lut/zea, lutein/zeaxanthin. | |||
Age | |||
≥18 | 62 | 1.733 (1.553, 1.932), <0.001 | 81.4, <0.001 |
<18 | 10 | 1.821 (1.359, 2.442), <0.001 | 82.2, <0.001 |
Gender | |||
Male | 32 | 1.449 (1.280, 1.640), <0.001 | 68.5, <0.001 |
Female | 32 | 1.995 (1.711, 2.327), <0.001 | 85.2, <0.001 |
Male and female | 10 | 1.829 (1.295, 2.583), 0.001 | 71.2, <0.001 |
Region | |||
Asia | 22 | 1.457 (1.250, 1.698), <0.001 | 6.9, 0.368 |
South America | 4 | 1.413 (0.988, 2.021), 0.058 | 77.1, 0.004 |
North America | 46 | 1.850 (1.636, 2.091), <0.001 | 86.2, <0.001 |
Oceania | 2 | 1.005 (0.531, 1.905), 0.987 | 0, 0.334 |
BMI | |||
Overweight | 27 | 1.621 (1.465, 1.795), <0.001 | 84.7, <0.001 |
Obesity | 47 | 1.800 (1.532, 2.114), <0.001 | 64.6, <0.001 |
Type | |||
LYC | 12 | 1.278 (1.070, 1.525), 0.007 | 60.7, 0.003 |
ASTA | 2 | 1.349 (0.907, 2.007), 0.139 | 0, 0.376 |
CRY | 12 | 1.766 (1.365, 2.284), <0.001 | 81, <0.001 |
zea/lut | 4 | 1.459 (1.027, 2.071), 0.035 | 0, 0.948 |
α car | 14 | 1.902 (1.504, 2.404), <0.001 | 85.4, <0.001 |
β car | 18 | 2.050 (1.639, 2.564), <0.001 | 85.9, <0.001 |
car | 4 | 1.024 (0.393, 2.669), 0.962 | 74.6, 0.008 |
lut/zea | 8 | 1.835 (1.394, 2.415), <0.001 | 83.0, <0.001 |
Fig. 3 Meta-analysis results of carotenoid supplementation for body weight in overweight or obese subjects. |
Fig. 5 Meta-analysis results of carotenoid supplementation for WC in subjects with overweight or obesity. |
Carotenoid intervention compared to no carotenoid intervention for subjects with overweight or obesity |
---|
Population: Subjects with overweight or obese |
Settings: Two studies were conducted in Europe, five studies were conducted in Asia |
Intervention: Carotenoid intervention |
Comparison: No carotenoid intervention |
Outcomes | SMD (95%CI)a | No. of participants (studies) | Quality of the evidence comments (GRADE) |
---|---|---|---|
Body weight (kg) | −2.340 (−3.800, −0.870) | 486 (6RCTs) | ⊕⊕⊕⊖Moderateb |
WC (cm) | −1.840 (−3.140, −0.540) | 388 (5RCTs) | ⊕⊕⊖⊖Lowb,c |
BMI (kg m−2) | −0.950 (−1.880, −0.010) | 342 (5RCTs) | ⊕⊕⊕⊖Moderateb |
Fat ratio (%) | −0.754 (−1.762, 0.254) | 171 (2RCTs) | ⊕⊕⊕⊖Moderateb |
TG (mg dL−1) | −2.095 (−3.201, −0.989) | 263 (4RCTs) | ⊕⊕⊕⊖Moderateb |
TC (mg dL−1) | −2.095 (−3.201, −0.989) | 119 (3RCTs) | ⊕⊕⊕⊖Moderateb |
LDL (mg dL−1) | −1.300 (−3.225, 0.625) | 119 (3RCTs) | ⊕⊕⊕⊖Moderateb |
HDL (mg dL−1) | 0.757 (0.101, 1.413) | 199 (3RCTs) | ⊕⊕⊕⊖Moderateb |
GRADE working group grades of evidence |
---|
SMD: standard mean deviation; CI: confidence interval; RCT: randomized controlled trial; WC: waist circumference; BMI: body mass index; TG: triglycerides; TC: total cholesterol; LDL: low density lipoprotein; HDL: high density lipoprotein.a Results for variations of treatments compared with controls.b Bias risk: downgraded by one level, as most of the included literature did not perform the Blind method allocation scheme hiding.c Inconsistency: downgraded by one level, as a high heterogeneity existed and its source was not completely clear. |
High quality: We are very confident that the true effect lies close to that of the estimate of the effect |
Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different |
Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect |
Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect |
Factors | Numbers of studies | SMD (95%CI), P | I 2 (%), P |
---|---|---|---|
SMD: standard mean difference; LDL, low density lipoprotein; HDL, high density lipoprotein; TC, total cholesterol; TG, triglyceride. | |||
Fat ratio (%) | 2 | −0.754 (−1.762, 0.254), 0.143 | 90.1, 0.001 |
HDL (mg dL−1) | 2 | 0.757 (0.101, 1.413), 0.024 | 0, 0.465 |
LDL (mg dL−1) | 3 | −1.300 (−3.225, 0.625), 0.186 | 90.7, <0.001 |
TC (mg dL−1) | 2 | −2.095 (−3.201, −0.989), <0.001 | 72, 0.059 |
TG (mg dL−1) | 4 | −1.875 (−4.382, 0.632), 0.143 | 96.9, <0.001 |
The publication bias outcomes of the included studies are presented in Table S5.† Publication biases were observed in body weight and LDL (P < 0.05). However, there was no significant difference in SMD before and after trimming and filling. Therefore, the influence of publication bias was considered slight and the results were stable (Table S5†).
Available animal and human studies have identified a positive role for carotenoids in weight loss.11,42 Such findings are analogous to the conclusions of our study. We concluded that the overweight and obese population with carotenoid intervention had a significant decrease in anthropometric measures (weight, WC and BMI), compared to the control population. A critical mechanism contributing to these beneficial effects appears to be the “abnormal immune response” in obesity. Calder et al. proposed that the obesity is highly correlated with low-grade inflammation,43 in which adipose tissue releases many inflammatory mediators.44,45 When carotenoids are sufficient due to the intervention, the circulating concentrations of inflammatory markers declined, which improves the state of obesity. Additionally, Thomas-Valdés and Bohn revealed a very powerful connection between oxidative stress and obesity. Oxidative stress triggers obesity by several mechanisms, such as stimulating the deposition of WAT, enhancing the proliferation and differentiation of preadipocytes and increasing the size of mature adipocytes.46 Carotenoids are essential in regulating oxidative metabolism and reducing cellular differentiation to treat obesity.47 Moreover, the association between the nuclear receptor superfamily and carotenoids may provide a mechanism for the antiobesity effect of carotenoids. Carotenoids, such as β-carotenoids, can be converted to retinoic acids48 in their all-trans or 9-cis configuration, which are highly-potent activators of the retinoic acid receptors (RARs) and the retinoid-X receptors (RXRs).49 Among them, all-trans retinoic acid (atRA) is known to have an inhibitory effect on adipogenesis.48 Several mechanisms could explain the inhibition of adipogenesis by atRA. One mechanism involves atRA inhibition of the function of the early adipogenic transcription factor CCAAT/enhancer-binding protein-β (C/EBPβ) in the adipogenic program through RAR-dependent induction of the C/EBPβ inhibitory protein Smad3.50 In addition, atRA uses the RAR pathway in preadipocytes to generate specific proteins that inhibit adipogenesis.51 Moreover, atRA could be an effective signal for the transcription of the prolipolytic substance UCP1 gene, inducing the expression of UCP1 through the mediation of RXRs.52,53 For these reasons, carotenoids have the potential to influence the regulation of obesity and related metabolic parameters.
Lipid metabolic disorders are intimately connected to the development of cardiovascular disease. High-density lipoprotein (HDL) is considered a protective indicator of cardiovascular events,54,55 while high levels of serum total cholesterol (TC) are regarded as a risk factor.56 Our meta-analysis identified a significant increase in HDL and a decrease in TC parameters in the carotenoid intervention group compared to the control group, indicating that carotenoids can prevent cardiovascular disease. Under low-grade inflammation in patients with hyperlipidaemia, inflammatory factors (Serum Amyloid A Protein, SAA) change HDL into a nonfunctional type by replacing apolipoprotein AI in HDL.57 Nonfunctional HDL is considered to be incapable of catabolizing excess lipids, struggling to maintain the normal serum TC levels and increasing the risk of cardiovascular disease. However, carotenoids have the ability to reverse this phenomenon by reducing HDL3-associated SAA, transforming HDL2&3-associated enzymes PON-1 (paraoxonase-1) and LCAT (lecithin: cholesterol acyltransferase), and ultimately releasing HDL from SAA.58 Additionally, carotenoids can regulate insulin resistance or insufficient secretion, which is one of the reasons why obesity contributes to abnormal lipid metabolism.59 This deficiency probably results from activation of the NF-κB transcription factor by TNF-α (tumour necrosis factor-α), and carotenoids could decrease the reactivation. In conclusion, carotenoids play an important role in multiple stages of the lipid metabolic process.60 Considering the correlation between dyslipidaemia and coronary heart disease, the effective intervention of carotenoids in dyslipidaemia discovered in this study is estimable to some extent.
In accordance with the outcomes of the subgroup meta-analysis, the anthropometric parameters in the female group, including body weight and TGs, were significantly reduced in terms of SMD values compared to those of the female and male group. The mechanism might be related to oestrogen, a potent steroid hormone with higher concentrations in females from adolescence to menopause than in males.61 An animal experiment conducted by Clegg exploring the association between sex hormones and leptin sensitivity has demonstrated that oestrogen increases the expression of UCP1 mRNA, ultimately leading to an increase in brown adipocyte breakdown and a reduction in body weight.62 We conjectured that the significant decrease in the TG concentrations in the female population after carotenoid intervention was linked to leptin. Leptin, a protein hormone secreted primarily by white adipose tissue, ameliorates elevated blood lipids produced by insulin resistance in obese patients. Carotenoids can increase leptin expression by acting on WAT mRNA, while oestrogens modulate the expression of leptin-specific receptors to increase leptin sensitivity.62–64 This finding may offer an explanation for the better intervention effect of carotenoids in women who have more oestrogen. Moreover, the studies with ≤12 weeks did not show statistically significant changes between the intervention and control groups based on the outcomes of the subgroup analysis. The reason might be the lack of guidelines for carotenoid supplementation during the intervention time. We could only refer to the most common intervention duration of the supplemental trial included in this meta-analysis, which was 12 weeks. Simultaneously, developing guidelines for carotenoid supplementation in obese patients is urgently recommended.
This meta-analysis has some limitations. Some of the papers did not offer the sex distribution of the study population, so we could not include them in the subgroup analysis based on sex. Furthermore, most of the results were highly heterogeneous. This finding might be attributed to differences in subjects and geography, as well as the type of results. Most importantly, RCTs as reliable evidence of effectiveness, have not been widely conducted in studies of carotenoid supplementation in obese patients. The limited sample size of the included RCTs may bias the true effect of carotenoid supplementation on the obese population from the findings of this study.
Footnote |
† Electronic supplementary information (ESI) available: Tables S1–5, Fig. S1 and PRISMA checklist. See DOI: 10.1039/d1fo00004g |
This journal is © The Royal Society of Chemistry 2021 |