Analysis of Repeated Measurements of Serum Carotenoid Levels and All-Cause and Cause-Specific Mortality in Japan | Cardiology | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure.  Study Design Using Repeated Measurements vs Baseline Measurements
Study Design Using Repeated Measurements vs Baseline Measurements
Table 1.  Baseline Characteristics of Participants by All-Cause Mortality
Baseline Characteristics of Participants by All-Cause Mortality
Table 2.  All-Cause Mortality by Repeated Measurement of Serum Carotenoid Levels
All-Cause Mortality by Repeated Measurement of Serum Carotenoid Levels
Table 3.  Cancer Mortality by Repeated Measurements of Serum Carotenoid Levels
Cancer Mortality by Repeated Measurements of Serum Carotenoid Levels
Table 4.  Cardiovascular Disease Mortality by Repeated Measurements of Serum Carotenoid Levels
Cardiovascular Disease Mortality by Repeated Measurements of Serum Carotenoid Levels
1.
Boffetta  P, Couto  E, Wichmann  J,  et al.  Fruit and vegetable intake and overall cancer risk in the European Prospective Investigation Into Cancer and Nutrition (EPIC).   J Natl Cancer Inst. 2010;102(8):529-537. doi:10.1093/jnci/djq072 PubMedGoogle ScholarCrossref
2.
Leenders  M, Sluijs  I, Ros  MM,  et al.  Fruit and vegetable consumption and mortality: European Prospective Investigation Into Cancer and Nutrition.   Am J Epidemiol. 2013;178(4):590-602. doi:10.1093/aje/kwt006 PubMedGoogle ScholarCrossref
3.
Aune  D, Giovannucci  E, Boffetta  P,  et al.  Fruit and vegetable intake and the risk of cardiovascular disease, total cancer and all-cause mortality—a systematic review and dose-response meta-analysis of prospective studies.   Int J Epidemiol. 2017;46(3):1029-1056. doi:10.1093/ije/dyw319 PubMedGoogle ScholarCrossref
4.
Paiva  SA, Russell  RM.  Beta-carotene and other carotenoids as antioxidants.   J Am Coll Nutr. 1999;18(5):426-433. doi:10.1080/07315724.1999.10718880 PubMedGoogle ScholarCrossref
5.
Omenn  GS, Goodman  GE, Thornquist  MD,  et al.  Risk factors for lung cancer and for intervention effects in CARET, the Beta-Carotene and Retinol Efficacy Trial.   J Natl Cancer Inst. 1996;88(21):1550-1559. doi:10.1093/jnci/88.21.1550PubMedGoogle ScholarCrossref
6.
Omenn  GS, Goodman  G, Thornquist  M,  et al.  The beta-carotene and retinol efficacy trial (CARET) for chemoprevention of lung cancer in high risk populations: smokers and asbestos-exposed workers.   Cancer Res. 1994;54(7)(suppl):2038s-2043s.PubMedGoogle Scholar
7.
Bates  CJ, Hamer  M, Mishra  GD.  Redox-modulatory vitamins and minerals that prospectively predict mortality in older British people: the National Diet and Nutrition Survey of people aged 65 years and over.   Br J Nutr. 2011;105(1):123-132. doi:10.1017/S0007114510003053 PubMedGoogle ScholarCrossref
8.
Greenberg  ER, Baron  JA, Karagas  MR,  et al.  Mortality associated with low plasma concentration of beta carotene and the effect of oral supplementation.   JAMA. 1996;275(9):699-703. doi:10.1001/jama.1996.03530330043027 PubMedGoogle ScholarCrossref
9.
Ito  Y, Wakai  K, Suzuki  K,  et al; JACC Study Group.  Serum carotenoids and mortality from lung cancer: a case-control study nested in the Japan Collaborative Cohort (JACC) study.   Cancer Sci. 2003;94(1):57-63. doi:10.1111/j.1349-7006.2003.tb01352.x PubMedGoogle ScholarCrossref
10.
Ito  Y, Kurata  M, Suzuki  K, Hamajima  N, Hishida  H, Aoki  K.  Cardiovascular disease mortality and serum carotenoid levels: a Japanese population-based follow-up study.   J Epidemiol. 2006;16(4):154-160. doi:10.2188/jea.16.154 PubMedGoogle ScholarCrossref
11.
Ito  Y, Suzuki  S, Yagyu  K, Sasaki  R, Suzuki  K, Aoki  K.  Relationship between serum carotenoid levels and cancer death rates in the residents, living in a rural area of Hokkaido, Japan.   J Epidemiol. 1997;7(1):1-8. doi:10.2188/jea.7.1 PubMedGoogle ScholarCrossref
12.
Lee  JE, Männistö  S, Spiegelman  D,  et al.  Intakes of fruit, vegetables, and carotenoids and renal cell cancer risk: a pooled analysis of 13 prospective studies.   Cancer Epidemiol Biomarkers Prev. 2009;18(6):1730-1739. doi:10.1158/1055-9965.EPI-09-0045 PubMedGoogle ScholarCrossref
13.
Osganian  SK, Stampfer  MJ, Rimm  E, Spiegelman  D, Manson  JE, Willett  WC.  Dietary carotenoids and risk of coronary artery disease in women.   Am J Clin Nutr. 2003;77(6):1390-1399. doi:10.1093/ajcn/77.6.1390 PubMedGoogle ScholarCrossref
14.
Hirvonen  T, Virtamo  J, Korhonen  P, Albanes  D, Pietinen  P.  Intake of flavonoids, carotenoids, vitamins C and E, and risk of stroke in male smokers.   Stroke. 2000;31(10):2301-2306. doi:10.1161/01.STR.31.10.2301 PubMedGoogle ScholarCrossref
15.
Suzuki  K, Honjo  H, Ichino  N,  et al.  Association of serum carotenoid levels with urinary albumin excretion in a general Japanese population: the Yakumo study.   J Epidemiol. 2013;23(6):451-456. doi:10.2188/jea.JE20130058 PubMedGoogle ScholarCrossref
16.
Suzuki  K, Ito  Y, Inoue  T, Hamajima  N.  Inverse association of serum carotenoids with prevalence of metabolic syndrome among Japanese.   Clin Nutr. 2011;30(3):369-375. doi:10.1016/j.clnu.2010.12.006 PubMedGoogle ScholarCrossref
17.
Suzuki  K, Ito  Y, Nakamura  S, Ochiai  J, Aoki  K.  Relationship between serum carotenoids and hyperglycemia: a population-based cross-sectional study.   J Epidemiol. 2002;12(5):357-366. doi:10.2188/jea.12.357 PubMedGoogle ScholarCrossref
18.
Wakimoto  P, Block  G.  Dietary intake, dietary patterns, and changes with age: an epidemiological perspective.   J Gerontol A Biol Sci Med Sci. 2001;56(Spec No 2):65-80. doi:10.1093/gerona/56.suppl_2.65 PubMedGoogle ScholarCrossref
19.
Kabat  GC, Kim  M, Adams-Campbell  LL,  et al; WHI Investigators.  Longitudinal study of serum carotenoid, retinol, and tocopherol concentrations in relation to breast cancer risk among postmenopausal women.   Am J Clin Nutr. 2009;90(1):162-169. doi:10.3945/ajcn.2009.27568 PubMedGoogle ScholarCrossref
20.
Kabat  GC, Heo  M, Ochs-Balcom  HM,  et al.  Longitudinal association of measures of adiposity with serum antioxidant concentrations in postmenopausal women.   Eur J Clin Nutr. 2016;70(1):47-53. doi:10.1038/ejcn.2015.74 PubMedGoogle ScholarCrossref
21.
World Medical Association.  World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects.   JAMA. 2013;310(20):2191-2194. doi:10.1001/jama.2013.281053PubMedGoogle ScholarCrossref
22.
Watarai  R, Suzuki  K, Ichino  N,  et al.  Association between serum levels of carotenoids and serum asymmetric dimethylarginine levels in Japanese subjects.   J Epidemiol. 2014;24(3):250-257. doi:10.2188/jea.JE20130137 PubMedGoogle ScholarCrossref
23.
D’Agostino  RB, Lee  ML, Belanger  AJ, Cupples  LA, Anderson  K, Kannel  WB.  Relation of pooled logistic regression to time dependent Cox regression analysis: the Framingham Heart Study.   Stat Med. 1990;9(12):1501-1515. doi:10.1002/sim.4780091214 PubMedGoogle ScholarCrossref
24.
Shardell  MD, Alley  DE, Hicks  GE,  et al.  Low-serum carotenoid concentrations and carotenoid interactions predict mortality in US adults: the Third National Health and Nutrition Examination Survey.   Nutr Res. 2011;31(3):178-189. doi:10.1016/j.nutres.2011.03.003 PubMedGoogle ScholarCrossref
25.
Li  C, Ford  ES, Zhao  G, Balluz  LS, Giles  WH, Liu  S.  Serum α-carotene concentrations and risk of death among US adults: the Third National Health and Nutrition Examination Survey Follow-up Study.   Arch Intern Med. 2011;171(6):507-515. doi:10.1001/archinternmed.2010.440 PubMedGoogle ScholarCrossref
26.
Sahyoun  NR, Jacques  PF, Russell  RM.  Carotenoids, vitamins C and E, and mortality in an elderly population.   Am J Epidemiol. 1996;144(5):501-511. doi:10.1093/oxfordjournals.aje.a008957 PubMedGoogle ScholarCrossref
27.
Kilander  L, Berglund  L, Boberg  M, Vessby  B, Lithell  H.  Education, lifestyle factors and mortality from cardiovascular disease and cancer: a 25-year follow-up of Swedish 50-year-old men.   Int J Epidemiol. 2001;30(5):1119-1126. doi:10.1093/ije/30.5.1119 PubMedGoogle ScholarCrossref
28.
Akbaraly  TN, Favier  A, Berr  C.  Total plasma carotenoids and mortality in the elderly: results of the Epidemiology of Vascular Ageing (EVA) study.   Br J Nutr. 2009;101(1):86-92. doi:10.1017/S0007114508998445 PubMedGoogle ScholarCrossref
29.
Walston  J, Xue  Q, Semba  RD,  et al.  Serum antioxidants, inflammation, and total mortality in older women.   Am J Epidemiol. 2006;163(1):18-26. doi:10.1093/aje/kwj007 PubMedGoogle ScholarCrossref
30.
Lauretani  F, Semba  RD, Dayhoff-Brannigan  M,  et al.  Low total plasma carotenoids are independent predictors of mortality among older persons: the InCHIANTI study.   Eur J Nutr. 2008;47(6):335-340. doi:10.1007/s00394-008-0732-9 PubMedGoogle ScholarCrossref
31.
Huang  J, Weinstein  SJ, Yu  K, Männistö  S, Albanes  D.  Serum beta carotene and overall and cause-specific mortality.   Circ Res. 2018;123(12):1339-1349. doi:10.1161/CIRCRESAHA.118.313409 PubMedGoogle ScholarCrossref
32.
De Waart  FG, Schouten  EG, Stalenhoef  AF, Kok  FJ.  Serum carotenoids, alpha-tocopherol and mortality risk in a prospective study among Dutch elderly.   Int J Epidemiol. 2001;30(1):136-143. doi:10.1093/ije/30.1.136 PubMedGoogle ScholarCrossref
33.
Buijsse  B, Feskens  EJ, Schlettwein-Gsell  D,  et al.  Plasma carotene and alpha-tocopherol in relation to 10-y all-cause and cause-specific mortality in European elderly: the Survey in Europe on Nutrition and the Elderly, a Concerted Action (SENECA).   Am J Clin Nutr. 2005;82(4):879-886. doi:10.1093/ajcn/82.4.879 PubMedGoogle ScholarCrossref
34.
Virtamo  J, Pietinen  P, Huttunen  JK,  et al; ATBC Study Group.  Incidence of cancer and mortality following alpha-tocopherol and beta-carotene supplementation: a postintervention follow-up.   JAMA. 2003;290(4):476-485. doi:10.1001/jama.290.4.476 PubMedGoogle ScholarCrossref
35.
Hashim  D, Gaughan  D, Boffetta  P, Lucchini  RG.  Baseline serum β-carotene concentration and mortality among long-term asbestos-exposed insulators.   Cancer Epidemiol Biomarkers Prev. 2015;24(3):555-560. doi:10.1158/1055-9965.EPI-14-0952 PubMedGoogle ScholarCrossref
36.
Holick  CN, Michaud  DS, Stolzenberg-Solomon  R,  et al.  Dietary carotenoids, serum beta-carotene, and retinol and risk of lung cancer in the alpha-tocopherol, beta-carotene cohort study.   Am J Epidemiol. 2002;156(6):536-547. doi:10.1093/aje/kwf072 PubMedGoogle ScholarCrossref
37.
Krinsky  NI.  Carotenoids as antioxidants.   Nutrition. 2001;17(10):815-817. doi:10.1016/S0899-9007(01)00651-7 PubMedGoogle ScholarCrossref
38.
Stahl  W, Sies  H.  Antioxidant activity of carotenoids.   Mol Aspects Med. 2003;24(6):345-351. doi:10.1016/S0098-2997(03)00030-X PubMedGoogle ScholarCrossref
39.
Clinton  SK.  Lycopene: chemistry, biology, and implications for human health and disease.   Nutr Rev. 1998;56(2 Pt 1):35-51. doi:10.1111/j.1753-4887.1998.tb01691.x PubMedGoogle Scholar
40.
Aune  D, Keum  N, Giovannucci  E,  et al.  Dietary intake and blood concentrations of antioxidants and the risk of cardiovascular disease, total cancer, and all-cause mortality: a systematic review and dose-response meta-analysis of prospective studies.   Am J Clin Nutr. 2018;108(5):1069-1091. doi:10.1093/ajcn/nqy097 PubMedGoogle ScholarCrossref
41.
Druesne-Pecollo  N, Latino-Martel  P, Norat  T,  et al.  Beta-carotene supplementation and cancer risk: a systematic review and metaanalysis of randomized controlled trials.   Int J Cancer. 2010;127(1):172-184. doi:10.1002/ijc.25008 PubMedGoogle ScholarCrossref
42.
Al-Delaimy  WK, Ferrari  P, Slimani  N,  et al.  Plasma carotenoids as biomarkers of intake of fruits and vegetables: individual-level correlations in the European Prospective Investigation Into Cancer and Nutrition (EPIC).   Eur J Clin Nutr. 2005;59(12):1387-1396. doi:10.1038/sj.ejcn.1602252 PubMedGoogle ScholarCrossref
43.
Jenab  M, Riboli  E, Ferrari  P,  et al.  Plasma and dietary carotenoid, retinol and tocopherol levels and the risk of gastric adenocarcinomas in the European Prospective Investigation Into Cancer and Nutrition.   Br J Cancer. 2006;95(3):406-415. doi:10.1038/sj.bjc.6603266 PubMedGoogle ScholarCrossref
44.
Burri  BJ, La Frano  MR, Zhu  C.  Absorption, metabolism, and functions of β-cryptoxanthin.   Nutr Rev. 2016;74(2):69-82. doi:10.1093/nutrit/nuv064 PubMedGoogle ScholarCrossref
45.
Sugiura  M, Kato  M, Matsumoto  H,  et al  Serum concentration of β-cryptoxanthin in Japan reflects the frequency of satsuma mandarin (Citrus unshiu Marc.) consumption.   Journal of Health Science. 2002;48(4):350–353. doi:10.1248/jhs.48.350 Google ScholarCrossref
46.
Ito  Y, Kurata  M, Hioki  R, Suzuki  K, Ochiai  J, Aoki  K.  Cancer mortality and serum levels of carotenoids, retinol, and tocopherol: a population-based follow-up study of inhabitants of a rural area of Japan.   Asian Pac J Cancer Prev. 2005;6(1):10-15.PubMedGoogle Scholar
47.
Iribarren  C, Folsom  AR, Jacobs  DR  Jr, Gross  MD, Belcher  JD, Eckfeldt  JH.  Association of serum vitamin levels, LDL susceptibility to oxidation, and autoantibodies against MDA-LDL with carotid atherosclerosis: a case-control study: the ARIC study investigators: Atherosclerosis Risk in Communities.   Arterioscler Thromb Vasc Biol. 1997;17(6):1171-1177. doi:10.1161/01.ATV.17.6.1171 PubMedGoogle ScholarCrossref
48.
Suzuki  K, Inoue  T, Hioki  R,  et al.  Association of abdominal obesity with decreased serum levels of carotenoids in a healthy Japanese population.   Clin Nutr. 2006;25(5):780-789. doi:10.1016/j.clnu.2006.01.025 PubMedGoogle ScholarCrossref
49.
Suzuki  K, Ishii  J, Kitagawa  F,  et al.  Association of serum carotenoid levels with N-terminal pro-brain-type natriuretic peptide: a cross-sectional study in Japan.   J Epidemiol. 2013;23(3):163-168. doi:10.2188/jea.JE20120087 PubMedGoogle ScholarCrossref
50.
Jurek  AM, Greenland  S, Maldonado  G, Church  TR.  Proper interpretation of non-differential misclassification effects: expectations vs observations.   Int J Epidemiol. 2005;34(3):680-687. doi:10.1093/ije/dyi060 PubMedGoogle ScholarCrossref
51.
Bjelakovic  G, Nikolova  D, Gluud  LL, Simonetti  RG, Gluud  C.  Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases.   Cochrane Database Syst Rev. 2012;(3):CD007176. doi:10.1002/14651858.CD007176.pub2 PubMedGoogle Scholar
52.
Gaziano  JM, Sesso  HD, Christen  WG,  et al.  Multivitamins in the prevention of cancer in men: the Physicians’ Health Study II randomized controlled trial.   JAMA. 2012;308(18):1871-1880. doi:10.1001/jama.2012.14641 PubMedGoogle ScholarCrossref
53.
Kristman  V, Manno  M, Côté  P.  Loss to follow-up in cohort studies: how much is too much?   Eur J Epidemiol. 2004;19(8):751-760. doi:10.1023/B:EJEP.0000036568.02655.f8 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Nutrition, Obesity, and Exercise
    June 11, 2021

    Analysis of Repeated Measurements of Serum Carotenoid Levels and All-Cause and Cause-Specific Mortality in Japan

    Author Affiliations
    • 1Department of Preventive Medical Sciences, Fujita Health University School of Medical Sciences, Toyoake, Japan
    JAMA Netw Open. 2021;4(6):e2113369. doi:10.1001/jamanetworkopen.2021.13369
    Key Points

    Question  Are repeated measurements of serum carotenoid levels associated with all-cause and cause-specific mortality?

    Findings  In this cohort study of 3116 individuals who received physical examinations, higher levels of serum carotenoid ascertained with multiple measurements were associated with lower risks of all-cause, cancer, and cardiovascular disease mortality. The decreases in risk associated with higher serum carotenoid levels were greater for analysis using repeated measurements of serum carotenoid compared with analysis using single measurements.

    Meaning  These findings suggest that higher carotenoid levels are associated with lower risk of mortality.

    Abstract

    Importance  The associations of levels of diverse serum carotenoids ascertained via repeated measurements with all-cause, cancer, and cardiovascular disease (CVD) mortality risk have not been considered in previous prospective studies.

    Objective  To investigate the association between repeated measurement of serum carotenoid levels and all-cause and cause-specific mortality risk.

    Design, Setting, and Participants  This cohort study’s baseline data were collected using information from physical examinations from 1990 to 1999. Eligible participants were followed up until December 2017, with a median (interquartile range) follow-up period of 22.3 (15.5-25.3) years. Included individuals were age 40 years or older at the baseline data collection, were residents of the study site in the town of Yakumo, Hokkaido, Japan, and participated in a physical examination at least once from 1990 to 1999. Among eligible participants, after excluding 332 individuals, 3116 individuals were included in the analysis. Data analysis was conducted in April 2020.

    Exposures  Repeated measurements of 6 serum carotenoid levels and 4 associated indices.

    Main Outcomes and Measures  All-cause, cancer, and CVD mortality, categorized by International Classification of Diseases, Ninth Revision and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes, were recorded. A time-dependent Cox regression model was performed to examine associations between time-varying serum carotenoid levels and mortality.

    Results  Among 3116 individuals who received physical examinations, the mean (SD) age was 54.7 (10.6) years and 1883 (60.4%) were women. During the follow-up period, 762 deaths from all causes, 253 deaths from cancer, and 210 deaths from CVD were ascertained. In a time-dependent Cox regression analysis, for every 25% higher serum levels of total carotenoids, risks were statistically significantly lower for all-cause mortality (hazard ratio [HR], 0.85; 95% CI, 0.82-0.87; P < .001), cancer mortality (HR, 0.82; 95% CI, 0.78-0.87; P < .001), and CVD mortality (HR, 0.86; 95% CI, 0.81-0.91; P < .001). Using only baseline measures, for every 25% higher serum levels of total carotenoids, risks were also statistically significantly lower for all-cause mortality (HR, 0.92; 95% CI, 0.89-0.95; P < .001), cancer mortality (HR, 0.87; 95% CI, 0.83-0.93; P < .001), and CVD mortality (HR, 0.93; 95% CI, 0.88-0.99; P = .03) but with larger HRs than those associated with repeated measurements.

    Conclusions and Relevance  This study found that higher levels of serum carotenoids in analysis using repeated measurements were associated with significantly lower all-cause and cause-specific mortality over a follow-up period of 25 years.

    Introduction

    Dietary intake of vegetables and fruits has been found to contribute to the reduced risks of overall and cause-specific mortality.1-3 The protective effects associated with vegetables and fruits may be attributed to the carotenoids they contain, which are mainly found in green and yellow vegetables and fruits.4 Serum carotenoids act as antioxidants in the human body, and their association with human health has received considerable attention for several decades. Although some intervention studies have found adverse cancer outcomes associated with serum carotenoids among individuals who smoke and workers exposed to asbestos,5,6 numerous prospective and interventional studies have reported the association of carotenoids with prevention of cancer, cardiovascular disease (CVD), and other pathophysiological conditions.7-17

    A considerable number of prospective studies10-13 examined the association between baseline serum carotenoid levels and all-cause, cancer, and CVD mortality. However, most studies did not consider changes in serum carotenoid levels over a long follow-up period. Given that serum carotenoid level is a better indicator associated with dietary intake of vegetables and fruits in a time series, previous studies failed to reflect changes in nutritional status during follow-up for risk estimation. Nutrient intake can be affected by multiple internal and external factors, such as physiological status with aging, food supply, and public perception.18 Therefore, a prospective study with repeated measurements of serum carotenoid levels may be able to capture changes in dietary vegetable and fruit intake and is essential to estimate more precise associations with mortality. Previous longitudinal studies assessed the association of repeated measurements of serum carotenoid levels with the incidence of breast cancer19 and colorectal cancer.20 In these studies, the authors applied time-dependent analyses and found evidence that several carotenoids were inversely associated with cancer incidence. Inspired by these findings for cancer incidence, we hypothesized that higher levels of serum carotenoid using repeated measurements would be associated with a greater decrease in risks of mortality compared with the results using only baseline measurements. However, to the best of our knowledge, few studies have been conducted to examine the associations between repeated measurements of serum carotenoid levels and all-cause and cause-specific (ie, cancer and CVD) mortality risk. Therefore, we aimed to examine the association between time-varying serum carotenoid levels and all-cause and cause-specific mortality risk using a time-dependent Cox regression model.

    Methods
    Study Population

    This study was conducted based on an annual population-based physical examination in the town of Yakumo, Hokkaido, Japan. The graphical timeline is shown in the Figure. The baseline data were collected using information from the physical examination from 1990 to 1999. The eligibility criteria for this cohort study were being age 40 years or older at the baseline data collection, residing at the study site, and participating in the physical examination at least once from 1990 to 1999. A total of 3448 individuals were eligible, among whom 332 individuals were excluded from the study; 63 individuals were excluded owing to a lack of informed consent, 261 individuals owing to missing values for questionnaire-based clinical history (including 84 individuals missing values for stroke, 84 individuals missing values for diabetes, 195 individuals missing values for cancer, and 96 individuals missing values for angina, with overlap among these groups), and 8 individuals owing to invalid information about end points, such as date and cause of death. Finally, we included 3116 individuals (1233 [39.6%] men and 1883 women [60.4%]) in our statistical analysis. The flow of selecting eligible participants is shown in eFigure 1 in the Supplement. The eligible participants were followed up until December 2017 using mortality records with permission from the Japanese Ministry of Internal Affairs and Communications and Ministry of Health, Labour and Welfare. During this period, 408 participants (13.1%) were lost to follow-up owing to their relocation. We censored them at the end of their relocation months. Serum carotenoid levels of participants in physical examinations were measured until the physical examination in 2011 (mean [SD] times for repeated measurements was 4.52 [4.71]).

    This cohort study was conducted in accordance with the Declaration of Helsinki,21 and the study protocol was approved by the ethics review committee of Fujita Health University. Written informed consent was obtained from all participants. This study is reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

    Exposure Assessment

    We repeatedly measured serum carotenoid levels of eligible participants during the follow-up period. The samples were collected at the annual physical examinations and stored at −80 °C until carotenoid levels were measured in our laboratory. We assessed the serum levels annually of zeaxanthin and lutein, canthaxanthin, β-cryptoxanthin, lycopene, α-carotene, and β-carotene using high-performance liquid chromatography. The details of the measurement are described elsewhere.22 Based on high-performance liquid chromatography results, we calculated total carotene levels by summing α-carotene, β-carotene, and lycopene levels; total xanthophyll levels by summing zeaxanthin and lutein, canthaxanthin, and β-cryptoxanthin levels; provitamin A levels by summing β-cryptoxanthin, α-carotene, and β-carotene levels; and total carotenoid levels by summing the levels of all measured carotenoids.

    Outcome Assessment

    After the baseline data collection from 1990 to 1999, municipal officials in Yakumo reviewed death certificates to assess the direct and indirect causes of death until December 2017. Based on this information, we assigned appropriate International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10) codes for each death. The underlying causes of death were classified according to ICD-9 for death certificates from 1990 to 1994 and ICD-10 for death certificates from 1995 to December 2017 (ie, the end of the follow-up period). We transformed the codes from ICD-9 into codes from ICD-10 before our analysis. In this study, total mortality includes all ICD-10 codes and cancer and CVD mortality were identified using the following ICD-10 codes: C00-C96 (ie, malignant neoplasms) for cancer mortality and I00-I99 (ie, diseases of the circulatory system) for CVD mortality.

    Covariates

    Eligible residents answered and completed the self-reported questionnaire, which included items regarding general health, lifestyle habits, medical history, clinical symptoms, and dietary habits before the health examination. In this study, lifestyle-related variables, such as tobacco smoking history (ie, current, ever, or never smoked) and alcohol consumption history (ie, current, ever, or never consumed alcohol), were collected. For each clinical condition, such as stroke, angina, diabetes, and cancer, participants were classified into 2 groups based on their answer in the questionnaire of never or yes (including ongoing treatment, past treatment, and no treatment for the condition). Fasting blood samples were collected during the physical examination and were centrifuged within 1 hour of sampling. Biochemical analyses for serum triglyceride and alanine aminotransferase (ALT) were conducted in Yakumo General Hospital laboratory. Systolic blood pressure (SBP) was measured using automatic upper-arm sphygmomanometers at the physical examination site after the participant had gotten at least 5 minutes of rest. Anthropometric measurements (ie, height and weight) were performed with help of trained public health nurses at the physical examination site. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared.

    Statistical Analysis

    Serum levels of triglyceride, ALT, and carotenoids had nongaussian distributions and therefore were expressed as medians and interquartile ranges (IQRs). These variables were used as log-transformed values in our statistical analysis. Other continuous variables with normal distributions were presented as means with SDs. The crude death rate was calculated according to quartiles and was presented as deaths per 1000 person-years. To estimate hazard ratios (HRs) and 95% CIs for the association of serum carotenoid levels with mortality using repeated measurements during follow-up, we performed a time-dependent Cox regression model.23 To ensure proportionality in this analysis, we performed the Schoenfeld method (eFigure 2 in the Supplement). The HRs in this analysis indicated a mortality risk in individuals for every 25% higher values in each carotenoid level. This meant, for example, mortality risk for every 14 μg/dL higher β-carotene levels (to convert to micromoles per liter, multiply by 0.01863) among the population and for every 5μg/dL higher β-cryptoxanthin levels among the population. For a detailed rationale for using the 25% difference in serum carotenoid levels, see the eAppendix in the Supplement. We included sex and age at baseline as potential confounders in the model adjusted for sex and age. Additionally, the fully adjusted model included sex, age, smoking habits, alcohol intake, SBP, triglyceride level, ALT level, and BMI. These covariates are also included as time-varying covariates in time-dependent Cox regression analyses. The rationale for inclusion as covariates in models was that these variables are associated with exposures (ie, carotenoids in this study) and outcomes (ie, mortality in this study).

    As an additional analysis, we first compared the HR estimated in time-dependent Cox regression analysis with the HR using only the baseline levels of serum carotenoids. For analysis using the baseline data set, we applied Cox regression analysis, the common method in survival analysis. To examine different levels of increase in serum carotenoid levels, we performed statistical analyses with 4 different settings of serum carotenoid increase (ie, for every 15%, 20%, 25%, and 30% higher levels). To reduce potential bias from recent measurements, we also performed sensitivity analyses after the exclusion of individuals who measured serum carotenoids within the year before mortality and after stratification by sex.

    Statistical significance was defined as a P value less than .05, and all tests were 2-sided. All statistical analyses were performed using R statistical software version 3.6.2 (R Project for Statistical Computing).

    Results
    Basic Characteristics and Recorded Death

    Among 3116 individuals in the analysis, the mean (SD) age was 54.7 (10.6) years and 1883 (60.4%) were women; 2354 individuals survived (75.5%), and 762 individuals died (24.5%). The mean (SD) age was 63.2 (8.6) years for individuals who died and 52.0 (9.7) years for those who survived (Table 1). Among individuals who died, 432 (56.7%) were men, compared with 801 (34.0%) men among those who survived. During a median (IQR) follow-up of 22.3 (15.5-25.3) years, 762 deaths from all causes, 253 deaths from cancer, and 210 deaths from CVD were recorded.

    All-Cause Mortality

    In time-varying analyses adjusted for sex and age, individuals who had higher serum levels of most carotenoids in repeated measurements had statistically significantly lower risk for all-cause mortality (zeaxanthin and lutein: HR, 0.90; 95% CI, 0.87-0.93; P < .001; β-cryptoxanthin: HR, 0.90; 95% CI, 0.88-0.92; P < .001; lycopene: HR, 0.90; 95% CI, 0.88-0.92; P < .001; α-carotene: HR, 0.90; 95% CI, 0.88-0.92; P < .001; β-carotene: HR, 0.90; 95% CI, 0.89-0.92; P < .001; total carotene: HR, 0.88; 95% CI, 0.86-0.90; P < .001; total xanthophyll: HR, 0.87; 95% CI, 0.84-0.90; P < .001; provitamin A: HR, 0.88; 95% CI, 0.86-0.90; P < .001; and total carotenoid: HR, 0.84; 95% CI, 0.81-0.87; P < .001); however, individuals who had higher levels of canthaxanthin did not have lower risk of all-cause mortality (HR, 1.00; 95% CI, 0.98-1.03; P = 0.93) (Table 2). After further adjustment for time-varying smoking habits, alcohol intake, SBP, ALT levels, triglyceride levels, and BMI, higher serum canthaxanthin levels in repeated measurements were still not associated with a significantly lower risk for all-cause mortality (HR, 0.99; 95% CI, 0.97-1.02; P = .47). However, after adjustments, higher serum levels of other carotenoids in repeated measurements were associated with statistically significantly lower risk for all-cause mortality (zeaxanthin and lutein: HR, 0.90; 95% CI, 0.87-0.93; P < .001; β-cryptoxanthin: HR, 0.91; 95% CI, 0.88-0.93; P < .001; lycopene: HR, 0.91; 95% CI, 0.89-0.92; P < .001; α-carotene: HR, 0.90; 95% CI, 0.88-0.92; P < .001; β-carotene: HR, 0.91; 95% CI, 0.89-0.93; P < .001; total carotene: HR, 0.89; 95% CI, 0.87-0.91; P < .001; total xanthophyll: HR, 0.87; 95% CI, 0.84-0.90; P < .001; provitamin A: HR, 0.89; 95% CI, 0.87-0.91; P < .001; and total carotenoid: HR, 0.85; 95% CI, 0.82-0.87; P < .001).

    Cancer Mortality

    Similarly, higher levels of most serum carotenoids in repeated measurements were associated with statistically significantly lower cancer mortality risk in analyses adjusted for sex and age (zeaxanthin and lutein: HR, 0.89; 95% CI, 0.85-0.94; P < .001; β-cryptoxanthin: HR, 0.90; 95% CI, 0.86-0.93; P < .001; lycopene: HR, 0.90; 95% CI, 0.87-0.93; P < .001; α-carotene: HR, 0.87; 95% CI, 0.84-0.91; P < .001; β-carotene: HR, 0.89; 95% CI, 0.86-0.92; P < .001; total carotene: HR, 0.87; 95% CI, 0.83-0.90; P < .001; total xanthophyll: HR, 0.86; 95% CI, 0.81-0.91; P < .001; provitamin A: HR, 0.87; 95% CI, 0.84-0.91; P < .001; and total carotenoids: HR, 0.82; 95% CI, 0.78-0.86; P < .001) (Table 3). After multivariable adjustment for covariates, higher levels of serum carotenoids remained associated with statistically significantly lower risks of cancer mortality (zeaxanthin and lutein: HR, 0.89; 95% CI, 0.85-0.95; P < .001; β-cryptoxanthin: HR, 0.90; 95% CI, 0.86-0.94; P < .001; lycopene: HR, 0.90; 95% CI, 0.87-0.93; P < .001; α-carotene: HR, 0.87; 95% CI, 0.84-0.91; P < .001; β-carotene: HR, 0.90; 95% CI, 0.87-0.93; P < .001; total carotene: HR, 0.87; 95% CI, 0.84-0.90; P < .001; total xanthophyll: HR, 0.86; 95% CI, 0.81–-0.91; P < .001; provitamin A: HR, 0.88; 95% CI, 0.84-0.91; P < .001; and total carotenoid: HR, 0.82; 95% CI, 0.78-0.87; P < .001). However, higher levels of serum canthaxanthin in repeated measurements were not associated with statistically significantly lower cancer mortality risk in either model.

    CVD Mortality

    In analyses adjusted for sex and age, higher levels of serum carotenoids, aside from canthaxanthin, in repeated measurements were associated with statistically significantly lower risk of CVD mortality (zeaxanthin and lutein: HR, 0.91; 95% CI, 0.86-0.97; P = 0.002; β-cryptoxanthin: HR, 0.91; 95% CI, 0.86-0.95; P < .001; lycopene: HR, 0.90; 95% CI, 0.87-0.93; P < .001; α-carotene: HR, 0.89; 95% CI, 0.85-0.93; P < .001; β-carotene: HR, 0.90; 95% CI, 0.87-0.93; P < .001; total carotene: HR, 0.88; 95% CI, 0.84-0.91; P < .001; total xanthophyll: HR, 0.88; 95% CI, 0.83-0.94; P < .001; provitamin A: HR, 0.88; 95% CI, 0.85-0.92; P < .001; and total carotenoid: HR, 0.84; 95% CI, 0.80-0.90; P < .001) (Table 4). In the fully adjusted analysis, higher levels of most serum carotenoids in repeated measurement were associated with statistically significantly lower risk for CVD mortality but with higher HRs (zeaxanthin and lutein: HR, 0.91; 95% CI, 0.86-0.97; P = .004; β-cryptoxanthin: HR, 0.92; 95% CI, 0.88-0.97; P < .001; lycopene: HR, 0.91; 95% CI, 0.87–0.94; P < .001; α-carotene: HR, 0.90; 95% CI, 0.86-0.94; P < .001; β-carotene: HR, 0.92; 95% CI, 0.88-0.96; P < .001; total carotene: HR, 0.89; 95% CI, 0.85-0.93; P < .001; total xanthophyll: HR, 0.89; 95% CI, 0.83-0.95; P < .001; provitamin A: HR, 0.90; 95% CI, 0.86-0.94; P < .001; and total carotenoid: HR, 0.86; 95% CI, 0.81-0.91; P < .001).

    Comparison With Analysis Using Only Baseline Measurements

    To evaluate the associations of carotenoid levels in analysis using repeated measurements with mortality risk, we compared the results from a time-dependent Cox regression model with the results using baseline serum carotenoid levels (eTable 1 in the Supplement). Compared with the analysis using baseline serum carotenoid levels, the decrease in risk for all-cause, cancer, and CVD mortality using repeated measurements was greater for most serum carotenoids; the HRs for all-cause, cancer, and CVD mortality were lower by 3% to 7%, 0% to 5%, and 4% to 10%, respectively. For example, the HR for higher levels of zeaxanthin and lutein using baseline measurements, compared with repeated measurements, was greater for all-cause mortality (0.95; 95% CI, 0.93-0.98; P = .003 vs 0.90, for a difference of 5.3%), cancer mortality (0.92; 95% CI, 0.87-0.97; P = .002 vs 0.89, for a difference of 3.3%), and CVD mortality (0.98; 95% CI, 0.93-1.04; P = .51 vs 0.91, for a difference of 7.1%). Overall, the HR was greater for higher total serum carotenoid levels using baseline measures, compared with repeated measurements, for all-cause mortality (0.92; 95% CI, 0.89-0.95; P < .001 vs 0.85, for a difference of 7.6%), cancer mortality (0.87; 95% CI, 0.83-0.93; P < .001 vs 0.82, for a difference of 5.7%), and CVD mortality (0.93; 95% CI, 0.88-0.99; P = .03 vs 0.86, for a difference of 7.5%).

    Sensitivity Analysis

    As a sensitivity analysis, we examined the associations between serum carotenoid levels found via repeated measurements and all-cause, cancer, and CVD mortality with different ranges of increase in serum carotenoid levels (eTable 2, eTable 3, and eTable 4 in the Supplement). The results of the significance level tests remained identical, but the magnitude of the estimates changed according to differences in serum carotenoid levels. After excluding participants who had serum carotenoid levels measured within 1 year from death (eTable 5 in the Supplement), we found that the HRs for mortality of serum carotenoid levels found via repeated measurements were higher compared with the results in Table 2, Table 3, and Table 4. In a sex-stratified analysis, we found no difference between men and women in the associations between carotenoid levels and mortality risk, although serum levels in all carotenoids were significantly different by sex.

    Discussion

    In this cohort study, we examined the association between carotenoid levels found via repeated measurements and mortality risk in a Japanese population. Over a follow-up period of 25 years, individuals with higher serum levels of most types of carotenoids had lower all-cause, cancer, and CVD mortality risk. We found that for results using repeated measurements, compared with those using baseline data only, most types of carotenoids had lower HRs for all-cause, cancer, and CVD mortality risk.

    We examined the associations between serum carotenoid levels and all-cause and cause-specific mortality using a repeated measurement data set. An important limitation of previous studies was the use of single baseline measurements of carotenoid levels, which could not reflect changes in blood carotenoid levels during the follow-up period.24,25 We found that for results using repeated measurements, most carotenoids were associated with greater decreases in all-cause, cancer, and CVD mortality risk compared with results obtained using only baseline data. Despite a lack of a previous similar study, we speculated that our results may be associated with accumulation of antioxidant properties by continued high serum carotenoid levels during the follow-up period. Given that more recent measurements of an exposure variable are associated with risk of all-cause and cause-specific mortality, these decreased HRs in the time-dependent Cox regression model may be associated with the more recent measurements, which is referred to as the late-stage effect. Therefore, we addressed this issue and performed sensitivity analyses after excluding participants who had serum carotenoid levels measured within the past year before their deaths. However, we found no significant bias from the late-stage effect.

    Previous cohort studies7-11,24-36 have examined the association between serum carotenoid levels and risk of all-cause, cancer, and CVD mortality. Most of these studies found inverse associations between serum carotenoid levels and mortality risk. Of these carotenoids, α-carotene, β-carotene, and lycopene are major antioxidants and play multiple bioactive roles in the human body.37-39 Moreover, a meta-analysis40 of prospective studies summarized the associations between the blood concentrations of different types of carotenoids and all-cause mortality, reporting that α-carotene, β-carotene, lycopene, and total carotene levels were inversely associated with mortality risk. As an example, for β-carotene, the authors reported that relative risk (RR) for each 25 μg/dL increase in serum level was 0.81 (95% CI, 0.72-0.90). We found a comparable association between serum β-carotene and all-cause mortality (HR for each 14 μg/dL increase in serum level, 0.91; 95% CI, 0.89-0.93). Although some interventional studies34,41 reported the null outcome or harmful cancer mortality outcomes associated with high-dose β-carotene supplementation, higher serum carotenoid levels did not have an association with increased risk of mortality in our study.

    Serum carotenoid levels reflect an individual’s dietary fruit and vegetable intake.42,43 Based on the results of older and more recent studies, higher serum carotenoid levels may be attributed to the sufficient intake of fruits and vegetables, rather than high-dose supplementation of carotenoids, and these foods may include various micronutrients and be associated with beneficial outcomes. For β-cryptoxanthin, 3 prospective studies7,24,32 found an association between serum levels and mortality risk,7,24,32 and 1 study40 found an inverse association with mortality, with an HR of 0.85. In the current study, we also found a significant association between serum β-cryptoxanthin levels at baseline and all-cause mortality risk. Compared with the reported risk reduction of serum β-cryptoxanthin in mortality in Aune et al40 (RR for each 15 μg/dL increase in serum level, 0.84; 95% CI, 0.76-0.94), our study using repeated measurements found an association between β-cryptoxanthin levels and decreased all-cause mortality risk with an HR for each 5 μg/dL increase in serum level of 0.91 (95% CI, 0.88-0.93). The dietary intake of β-cryptoxanthin varies across countries worldwide, and Japan has one of the populations that commonly consumes this type of carotenoid.44 Therefore, blood β-cryptoxanthin levels have been found to be higher in Japanese populations than in European populations,45 which is likely to contribute to the associations between β-cryptoxanthin and CVD mortality as found in our study. Although 3 prospective studies24,32,46 examined the associations between zeaxanthin and lutein and mortality risk, the results were inconsistent. The study by Iribarren et al47 found an inverse association between zeaxanthin and lutein and atherosclerosis. Therefore, we hypothesized that serum zeaxanthin and lutein levels may be associated with decreased CVD mortality risk. Additionally, in our study, serum zeaxanthin and lutein levels at baseline were associated with lower all-cause and cancer-specific mortality risk. For canthaxanthin, there are no prospective studies on the association between blood levels and mortality. Our study found no association between serum canthaxanthin levels at baseline and all-cause or cause-specific mortality, whereas our study found that serum canthaxanthin levels are associated with some beneficial biological activities.48,49 The associations between zeaxanthin and lutein, β-cryptoxanthin, and canthaxanthin levels and mortality risk are still not fully elucidated, and further studies with long-term follow-up should be performed to validate these associations.

    Limitations

    This study has several limitations. First, it was conducted among a homogenous population in Japan. Serum carotenoid levels found in this Japanese population were different from those of other ethnic populations. For example, the mean value of serum β-carotene in Japanese men from this study was 34.2 μg/dL, while the previously reported value for US men was 21.2 μg/dL.31 Therefore, these results may not be generalizable to other populations. Thus, further studies should be conducted to examine these associations in other, larger populations with repeated measurements of serum carotenoid levels. Second, unmeasured confounders may be associated with potential bias even in prospective studies. We tried to reduce the risk of bias by performing a sensitivity analysis and by adjusting for multiple covariates. However, owing to the limited items in our questionnaire in earlier periods, we could not adjust for important factors, such as physical activity and dietary quality. Third, as can happen in most studies using death certificates, nondifferential misclassifications for the cause of death could bias our findings toward the null (ie, bias the HR toward 1.00).50 Fourth, we did not incorporate the associations of multivitamin supplements with observed outcomes. Previous studies reported controversial results on the associations of multivitamin supplements with health outcomes.51,52 Therefore, although we cannot determine the direction of associations from dietary supplements in observed outcomes, it is highly possible that there was some kind of association. Fifth, in this study, 408 participants (13.1%) were lost to follow-up and censored. Although the percentage of loss to follow-up was relatively low, we cannot rule out biased estimation by loss to follow-up.53

    Conclusions

    This study found that higher levels of serum carotenoids in an analysis using repeated measurements were associated with statistically significantly lower all-cause and cause-specific mortality risk over a follow-up period of 25 years, as reported in previous studies. In the time-dependent Cox regression analysis, we found that the decreases in mortality risk associated with lower serum levels of carotenoids were greater using repeated measurements compared with using baseline data alone. These results suggest that higher levels of serum carotenoids are associated with lower risk of all-cause, cancer, and CVD mortality.

    Back to top
    Article Information

    Accepted for Publication: April 14, 2021.

    Published: June 11, 2021. doi:10.1001/jamanetworkopen.2021.13369

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Fujii R et al. JAMA Network Open.

    Corresponding Author: Koji Suzuki, PhD, Department of Preventive Medical Sciences, Fujita Health University School of Medical Sciences, 1-98 Dengakugakubo, Kutsukake-cho, Toyoake 470-1192, Japan (ksuzuki@fujita-hu.ac.jp).

    Author Contributions: Dr Suzuki and Mr Tsuboi had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Suzuki.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Fujii, Tsuboi, Ishihara.

    Critical revision of the manuscript for important intellectual content: Maeda, Suzuki.

    Statistical analysis: Fujii, Ishihara.

    Obtained funding: Suzuki.

    Administrative, technical, or material support: Maeda, Suzuki.

    Supervision: Suzuki.

    Conflict of Interest Disclosures: None reported.

    Funding/Support: This study was supported by Fujita Health University.

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Additional Contributions: Dr. Yoshinori Ito (P.D, Nagoya University) contributed to sample collection and carotenoid measurement and was not compensated for this work.

    References
    1.
    Boffetta  P, Couto  E, Wichmann  J,  et al.  Fruit and vegetable intake and overall cancer risk in the European Prospective Investigation Into Cancer and Nutrition (EPIC).   J Natl Cancer Inst. 2010;102(8):529-537. doi:10.1093/jnci/djq072 PubMedGoogle ScholarCrossref
    2.
    Leenders  M, Sluijs  I, Ros  MM,  et al.  Fruit and vegetable consumption and mortality: European Prospective Investigation Into Cancer and Nutrition.   Am J Epidemiol. 2013;178(4):590-602. doi:10.1093/aje/kwt006 PubMedGoogle ScholarCrossref
    3.
    Aune  D, Giovannucci  E, Boffetta  P,  et al.  Fruit and vegetable intake and the risk of cardiovascular disease, total cancer and all-cause mortality—a systematic review and dose-response meta-analysis of prospective studies.   Int J Epidemiol. 2017;46(3):1029-1056. doi:10.1093/ije/dyw319 PubMedGoogle ScholarCrossref
    4.
    Paiva  SA, Russell  RM.  Beta-carotene and other carotenoids as antioxidants.   J Am Coll Nutr. 1999;18(5):426-433. doi:10.1080/07315724.1999.10718880 PubMedGoogle ScholarCrossref
    5.
    Omenn  GS, Goodman  GE, Thornquist  MD,  et al.  Risk factors for lung cancer and for intervention effects in CARET, the Beta-Carotene and Retinol Efficacy Trial.   J Natl Cancer Inst. 1996;88(21):1550-1559. doi:10.1093/jnci/88.21.1550PubMedGoogle ScholarCrossref
    6.
    Omenn  GS, Goodman  G, Thornquist  M,  et al.  The beta-carotene and retinol efficacy trial (CARET) for chemoprevention of lung cancer in high risk populations: smokers and asbestos-exposed workers.   Cancer Res. 1994;54(7)(suppl):2038s-2043s.PubMedGoogle Scholar
    7.
    Bates  CJ, Hamer  M, Mishra  GD.  Redox-modulatory vitamins and minerals that prospectively predict mortality in older British people: the National Diet and Nutrition Survey of people aged 65 years and over.   Br J Nutr. 2011;105(1):123-132. doi:10.1017/S0007114510003053 PubMedGoogle ScholarCrossref
    8.
    Greenberg  ER, Baron  JA, Karagas  MR,  et al.  Mortality associated with low plasma concentration of beta carotene and the effect of oral supplementation.   JAMA. 1996;275(9):699-703. doi:10.1001/jama.1996.03530330043027 PubMedGoogle ScholarCrossref
    9.
    Ito  Y, Wakai  K, Suzuki  K,  et al; JACC Study Group.  Serum carotenoids and mortality from lung cancer: a case-control study nested in the Japan Collaborative Cohort (JACC) study.   Cancer Sci. 2003;94(1):57-63. doi:10.1111/j.1349-7006.2003.tb01352.x PubMedGoogle ScholarCrossref
    10.
    Ito  Y, Kurata  M, Suzuki  K, Hamajima  N, Hishida  H, Aoki  K.  Cardiovascular disease mortality and serum carotenoid levels: a Japanese population-based follow-up study.   J Epidemiol. 2006;16(4):154-160. doi:10.2188/jea.16.154 PubMedGoogle ScholarCrossref
    11.
    Ito  Y, Suzuki  S, Yagyu  K, Sasaki  R, Suzuki  K, Aoki  K.  Relationship between serum carotenoid levels and cancer death rates in the residents, living in a rural area of Hokkaido, Japan.   J Epidemiol. 1997;7(1):1-8. doi:10.2188/jea.7.1 PubMedGoogle ScholarCrossref
    12.
    Lee  JE, Männistö  S, Spiegelman  D,  et al.  Intakes of fruit, vegetables, and carotenoids and renal cell cancer risk: a pooled analysis of 13 prospective studies.   Cancer Epidemiol Biomarkers Prev. 2009;18(6):1730-1739. doi:10.1158/1055-9965.EPI-09-0045 PubMedGoogle ScholarCrossref
    13.
    Osganian  SK, Stampfer  MJ, Rimm  E, Spiegelman  D, Manson  JE, Willett  WC.  Dietary carotenoids and risk of coronary artery disease in women.   Am J Clin Nutr. 2003;77(6):1390-1399. doi:10.1093/ajcn/77.6.1390 PubMedGoogle ScholarCrossref
    14.
    Hirvonen  T, Virtamo  J, Korhonen  P, Albanes  D, Pietinen  P.  Intake of flavonoids, carotenoids, vitamins C and E, and risk of stroke in male smokers.   Stroke. 2000;31(10):2301-2306. doi:10.1161/01.STR.31.10.2301 PubMedGoogle ScholarCrossref
    15.
    Suzuki  K, Honjo  H, Ichino  N,  et al.  Association of serum carotenoid levels with urinary albumin excretion in a general Japanese population: the Yakumo study.   J Epidemiol. 2013;23(6):451-456. doi:10.2188/jea.JE20130058 PubMedGoogle ScholarCrossref
    16.
    Suzuki  K, Ito  Y, Inoue  T, Hamajima  N.  Inverse association of serum carotenoids with prevalence of metabolic syndrome among Japanese.   Clin Nutr. 2011;30(3):369-375. doi:10.1016/j.clnu.2010.12.006 PubMedGoogle ScholarCrossref
    17.
    Suzuki  K, Ito  Y, Nakamura  S, Ochiai  J, Aoki  K.  Relationship between serum carotenoids and hyperglycemia: a population-based cross-sectional study.   J Epidemiol. 2002;12(5):357-366. doi:10.2188/jea.12.357 PubMedGoogle ScholarCrossref
    18.
    Wakimoto  P, Block  G.  Dietary intake, dietary patterns, and changes with age: an epidemiological perspective.   J Gerontol A Biol Sci Med Sci. 2001;56(Spec No 2):65-80. doi:10.1093/gerona/56.suppl_2.65 PubMedGoogle ScholarCrossref
    19.
    Kabat  GC, Kim  M, Adams-Campbell  LL,  et al; WHI Investigators.  Longitudinal study of serum carotenoid, retinol, and tocopherol concentrations in relation to breast cancer risk among postmenopausal women.   Am J Clin Nutr. 2009;90(1):162-169. doi:10.3945/ajcn.2009.27568 PubMedGoogle ScholarCrossref
    20.
    Kabat  GC, Heo  M, Ochs-Balcom  HM,  et al.  Longitudinal association of measures of adiposity with serum antioxidant concentrations in postmenopausal women.   Eur J Clin Nutr. 2016;70(1):47-53. doi:10.1038/ejcn.2015.74 PubMedGoogle ScholarCrossref
    21.
    World Medical Association.  World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects.   JAMA. 2013;310(20):2191-2194. doi:10.1001/jama.2013.281053PubMedGoogle ScholarCrossref
    22.
    Watarai  R, Suzuki  K, Ichino  N,  et al.  Association between serum levels of carotenoids and serum asymmetric dimethylarginine levels in Japanese subjects.   J Epidemiol. 2014;24(3):250-257. doi:10.2188/jea.JE20130137 PubMedGoogle ScholarCrossref
    23.
    D’Agostino  RB, Lee  ML, Belanger  AJ, Cupples  LA, Anderson  K, Kannel  WB.  Relation of pooled logistic regression to time dependent Cox regression analysis: the Framingham Heart Study.   Stat Med. 1990;9(12):1501-1515. doi:10.1002/sim.4780091214 PubMedGoogle ScholarCrossref
    24.
    Shardell  MD, Alley  DE, Hicks  GE,  et al.  Low-serum carotenoid concentrations and carotenoid interactions predict mortality in US adults: the Third National Health and Nutrition Examination Survey.   Nutr Res. 2011;31(3):178-189. doi:10.1016/j.nutres.2011.03.003 PubMedGoogle ScholarCrossref
    25.
    Li  C, Ford  ES, Zhao  G, Balluz  LS, Giles  WH, Liu  S.  Serum α-carotene concentrations and risk of death among US adults: the Third National Health and Nutrition Examination Survey Follow-up Study.   Arch Intern Med. 2011;171(6):507-515. doi:10.1001/archinternmed.2010.440 PubMedGoogle ScholarCrossref
    26.
    Sahyoun  NR, Jacques  PF, Russell  RM.  Carotenoids, vitamins C and E, and mortality in an elderly population.   Am J Epidemiol. 1996;144(5):501-511. doi:10.1093/oxfordjournals.aje.a008957 PubMedGoogle ScholarCrossref
    27.
    Kilander  L, Berglund  L, Boberg  M, Vessby  B, Lithell  H.  Education, lifestyle factors and mortality from cardiovascular disease and cancer: a 25-year follow-up of Swedish 50-year-old men.   Int J Epidemiol. 2001;30(5):1119-1126. doi:10.1093/ije/30.5.1119 PubMedGoogle ScholarCrossref
    28.
    Akbaraly  TN, Favier  A, Berr  C.  Total plasma carotenoids and mortality in the elderly: results of the Epidemiology of Vascular Ageing (EVA) study.   Br J Nutr. 2009;101(1):86-92. doi:10.1017/S0007114508998445 PubMedGoogle ScholarCrossref
    29.
    Walston  J, Xue  Q, Semba  RD,  et al.  Serum antioxidants, inflammation, and total mortality in older women.   Am J Epidemiol. 2006;163(1):18-26. doi:10.1093/aje/kwj007 PubMedGoogle ScholarCrossref
    30.
    Lauretani  F, Semba  RD, Dayhoff-Brannigan  M,  et al.  Low total plasma carotenoids are independent predictors of mortality among older persons: the InCHIANTI study.   Eur J Nutr. 2008;47(6):335-340. doi:10.1007/s00394-008-0732-9 PubMedGoogle ScholarCrossref
    31.
    Huang  J, Weinstein  SJ, Yu  K, Männistö  S, Albanes  D.  Serum beta carotene and overall and cause-specific mortality.   Circ Res. 2018;123(12):1339-1349. doi:10.1161/CIRCRESAHA.118.313409 PubMedGoogle ScholarCrossref
    32.
    De Waart  FG, Schouten  EG, Stalenhoef  AF, Kok  FJ.  Serum carotenoids, alpha-tocopherol and mortality risk in a prospective study among Dutch elderly.   Int J Epidemiol. 2001;30(1):136-143. doi:10.1093/ije/30.1.136 PubMedGoogle ScholarCrossref
    33.
    Buijsse  B, Feskens  EJ, Schlettwein-Gsell  D,  et al.  Plasma carotene and alpha-tocopherol in relation to 10-y all-cause and cause-specific mortality in European elderly: the Survey in Europe on Nutrition and the Elderly, a Concerted Action (SENECA).   Am J Clin Nutr. 2005;82(4):879-886. doi:10.1093/ajcn/82.4.879 PubMedGoogle ScholarCrossref
    34.
    Virtamo  J, Pietinen  P, Huttunen  JK,  et al; ATBC Study Group.  Incidence of cancer and mortality following alpha-tocopherol and beta-carotene supplementation: a postintervention follow-up.   JAMA. 2003;290(4):476-485. doi:10.1001/jama.290.4.476 PubMedGoogle ScholarCrossref
    35.
    Hashim  D, Gaughan  D, Boffetta  P, Lucchini  RG.  Baseline serum β-carotene concentration and mortality among long-term asbestos-exposed insulators.   Cancer Epidemiol Biomarkers Prev. 2015;24(3):555-560. doi:10.1158/1055-9965.EPI-14-0952 PubMedGoogle ScholarCrossref
    36.
    Holick  CN, Michaud  DS, Stolzenberg-Solomon  R,  et al.  Dietary carotenoids, serum beta-carotene, and retinol and risk of lung cancer in the alpha-tocopherol, beta-carotene cohort study.   Am J Epidemiol. 2002;156(6):536-547. doi:10.1093/aje/kwf072 PubMedGoogle ScholarCrossref
    37.
    Krinsky  NI.  Carotenoids as antioxidants.   Nutrition. 2001;17(10):815-817. doi:10.1016/S0899-9007(01)00651-7 PubMedGoogle ScholarCrossref
    38.
    Stahl  W, Sies  H.  Antioxidant activity of carotenoids.   Mol Aspects Med. 2003;24(6):345-351. doi:10.1016/S0098-2997(03)00030-X PubMedGoogle ScholarCrossref
    39.
    Clinton  SK.  Lycopene: chemistry, biology, and implications for human health and disease.   Nutr Rev. 1998;56(2 Pt 1):35-51. doi:10.1111/j.1753-4887.1998.tb01691.x PubMedGoogle Scholar
    40.
    Aune  D, Keum  N, Giovannucci  E,  et al.  Dietary intake and blood concentrations of antioxidants and the risk of cardiovascular disease, total cancer, and all-cause mortality: a systematic review and dose-response meta-analysis of prospective studies.   Am J Clin Nutr. 2018;108(5):1069-1091. doi:10.1093/ajcn/nqy097 PubMedGoogle ScholarCrossref
    41.
    Druesne-Pecollo  N, Latino-Martel  P, Norat  T,  et al.  Beta-carotene supplementation and cancer risk: a systematic review and metaanalysis of randomized controlled trials.   Int J Cancer. 2010;127(1):172-184. doi:10.1002/ijc.25008 PubMedGoogle ScholarCrossref
    42.
    Al-Delaimy  WK, Ferrari  P, Slimani  N,  et al.  Plasma carotenoids as biomarkers of intake of fruits and vegetables: individual-level correlations in the European Prospective Investigation Into Cancer and Nutrition (EPIC).   Eur J Clin Nutr. 2005;59(12):1387-1396. doi:10.1038/sj.ejcn.1602252 PubMedGoogle ScholarCrossref
    43.
    Jenab  M, Riboli  E, Ferrari  P,  et al.  Plasma and dietary carotenoid, retinol and tocopherol levels and the risk of gastric adenocarcinomas in the European Prospective Investigation Into Cancer and Nutrition.   Br J Cancer. 2006;95(3):406-415. doi:10.1038/sj.bjc.6603266 PubMedGoogle ScholarCrossref
    44.
    Burri  BJ, La Frano  MR, Zhu  C.  Absorption, metabolism, and functions of β-cryptoxanthin.   Nutr Rev. 2016;74(2):69-82. doi:10.1093/nutrit/nuv064 PubMedGoogle ScholarCrossref
    45.
    Sugiura  M, Kato  M, Matsumoto  H,  et al  Serum concentration of β-cryptoxanthin in Japan reflects the frequency of satsuma mandarin (Citrus unshiu Marc.) consumption.   Journal of Health Science. 2002;48(4):350–353. doi:10.1248/jhs.48.350 Google ScholarCrossref
    46.
    Ito  Y, Kurata  M, Hioki  R, Suzuki  K, Ochiai  J, Aoki  K.  Cancer mortality and serum levels of carotenoids, retinol, and tocopherol: a population-based follow-up study of inhabitants of a rural area of Japan.   Asian Pac J Cancer Prev. 2005;6(1):10-15.PubMedGoogle Scholar
    47.
    Iribarren  C, Folsom  AR, Jacobs  DR  Jr, Gross  MD, Belcher  JD, Eckfeldt  JH.  Association of serum vitamin levels, LDL susceptibility to oxidation, and autoantibodies against MDA-LDL with carotid atherosclerosis: a case-control study: the ARIC study investigators: Atherosclerosis Risk in Communities.   Arterioscler Thromb Vasc Biol. 1997;17(6):1171-1177. doi:10.1161/01.ATV.17.6.1171 PubMedGoogle ScholarCrossref
    48.
    Suzuki  K, Inoue  T, Hioki  R,  et al.  Association of abdominal obesity with decreased serum levels of carotenoids in a healthy Japanese population.   Clin Nutr. 2006;25(5):780-789. doi:10.1016/j.clnu.2006.01.025 PubMedGoogle ScholarCrossref
    49.
    Suzuki  K, Ishii  J, Kitagawa  F,  et al.  Association of serum carotenoid levels with N-terminal pro-brain-type natriuretic peptide: a cross-sectional study in Japan.   J Epidemiol. 2013;23(3):163-168. doi:10.2188/jea.JE20120087 PubMedGoogle ScholarCrossref
    50.
    Jurek  AM, Greenland  S, Maldonado  G, Church  TR.  Proper interpretation of non-differential misclassification effects: expectations vs observations.   Int J Epidemiol. 2005;34(3):680-687. doi:10.1093/ije/dyi060 PubMedGoogle ScholarCrossref
    51.
    Bjelakovic  G, Nikolova  D, Gluud  LL, Simonetti  RG, Gluud  C.  Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases.   Cochrane Database Syst Rev. 2012;(3):CD007176. doi:10.1002/14651858.CD007176.pub2 PubMedGoogle Scholar
    52.
    Gaziano  JM, Sesso  HD, Christen  WG,  et al.  Multivitamins in the prevention of cancer in men: the Physicians’ Health Study II randomized controlled trial.   JAMA. 2012;308(18):1871-1880. doi:10.1001/jama.2012.14641 PubMedGoogle ScholarCrossref
    53.
    Kristman  V, Manno  M, Côté  P.  Loss to follow-up in cohort studies: how much is too much?   Eur J Epidemiol. 2004;19(8):751-760. doi:10.1023/B:EJEP.0000036568.02655.f8 PubMedGoogle ScholarCrossref
    ×