Relationship between Serum Carotenoids and Hyperglycemia: a Population-based Cross-sectional Study

The present study investigated the relationship between hyperglycemia and both serum carotenoids and intake of vegetables and fruits. Subjects with a history of diabetes mellitus (DM group, n=133) or with hyperglycemia diagnosed using a 5.6% cutoff value for hemoglobin A1c (High HbA1C group, n=151) were recruited from among inhabitants of a rural area in Hokkaido, Japan. Intake frequencies of vegetables and fruits were assessed using a questionnaire administered by public health nurses. Serum levels of carotenoids and retinol were measured using high-performance liquid chromatography. The relationships between high HbA1C or DM and both serum carotenoids and intake frequencies of vegetables and fruits were analyzed using logistic regression modeling for a case-control study; each case (High HbA1C or DM) was matched to two controls (healthy subjects without any history of disease) matched for sex and age (within 3 years). The odds ratio (OR) for high HbA1C was 0.49 (95% confidence interval: 0.29-0.85) on high intake frequency of carrot and pumpkin and the OR for DM was 1.21 (95% CI: 0.79-1.84). No significant relationships were observed between high HbA1c and intake frequencies of other vegetables and fruits. The ORs on high serum levels of α- and β-carotenes, lycopene, β-cryptoxanthin and zeaxanthin and lutein were 0.38 (0.22-0.65), 0.35 (0.21-0.59), 0.57 (0.35-0.93), 0.35 (0.20-0.59), and 0.88 (0.54-1.46) for high HbA1c, respectively. In conclusion, intake of vegetables and fruits rich in carotenoids might be a protective factor against hyperglycemia.


RESULTS
shows the characteristics of subjects in this study. The percentage of current alcohol drinkers was significantly higher in the High HbA1c group than in the Healthy 1 group. No significant differences in smoking, alcohol consumption, or exercise were observed between the DM and Healthy 2 groups. Table 2 shows the levels of HbA1c, serum components, blood pressure, and BMI among the study groups. The levels of HbA1c, serum fasting glucose, triglyceride, TBARS, -GTP, systolic blood pressure, diastolic blood pressure and BMI were significantly higher in the High HbAi~ group than in the Healthy 1 group. HDL-cholesterol was significantly lower in the High HbAlc group than in the Healthy 1 group . HbA1c, serum fasting glucose, triglyceride, -GTP, systolic blood pressure and BMI were significantly higher in the DM group than in the Healthy 2 group. HDL-cholesterol was significantly lower in the DM group than in the Healthy 2 group. HbA1c was significantly higher and serum total cholesterol, systolic blood pressure and diastolic blood pressure were significantly lower in the DM group than in the High HbA1c group. Table 3 shows intake frequency of fruit and vegetables among the study groups. The percentage of subjects with high intake frequency of carrot and pumpkin was significantly lower, and that of tomato tended to be lower in the High HbA1c group than in the Healthy 1 group. The percentage of subjects with intake frequency of oranges over 3-4 times/week tended to be lower in the DM group than in the Healthy 2 group. Table 4 shows the odds ratios of intake frequency of fruit and vegetables for high HbA1c and DM . A significantly lower odds ratio for high HbA1c was observed in the high intake group of carrot and pumpkin compared with the low intake group (0 .49, 95%CI: 0.29-0.85). Significantly lower odds ratio for DM was not observed between the high intake group of vegetables and fruits and the low intake group. Table 5 shows the levels of serum carotenoids, -tocopherol The results from the analyses conducted to compare the intake frequencies of fruit and vegetables and the serum carotenoids levels between DM and Healthy 1 group were also similar to those using Healthy 2 group as controls. retinol and -tocopherol were not significantly different between the high and low groups. Even after adjusting for confounding factors, these associations did not change. A significantly lower odds ratio for DM was observed in the high group of -cryptoxanthin than in the low group. No significant differences in serum other carotenoids, -tocopherol or retinol were observed between the high and low groups (Table 7).
In addition, the results obtained using the logistic regression between DM and Healthyl group were also similar to those using Healthy 2 group as controls.

DISCUSSION
Few studies have reported the relationships between serum carotenoids and DM or chronic hyperglycemia. The NHANES III study revealed that serum -carotene levels in subjects with impaired glucose tolerance and newly diagnosed with diabetes were lower compared to individuals with normal glucose tolerance. 13 In addition, serum lycopene levels demonstrated an inverse Antioxidant enzymes such as superoxide dismutase and glutathione peroxidase play protective roles against the reactive oxygen species.1 Chronic hyperglycemia reportedly causes glycation of those enzymes and reduces these functions,26.27 a factor for increased oxidative stress in vivo.
Carotenoids such as -carotene are known to protect cells from oxidative stress by quenching free radicals.9.10 In addition , there was significantly negative correlation between serumcarotene and HbA1c levels in this study subjects (r=0 .119, p<0.05). Carotenoids are particularly abundant in green-yellow vegetables and fruits. 28 Our previous report29 revealed that serum levels of carotenoids such as -and -carotene were correlated Table 3. Comparison of intake frequency of vegetables or fruits between the study groups .
Low: 1-2 times/week or less, High: 3-4 times/week or more High HbA1c: the subjects with no history of diabetes mellitus with HbA1c values of 5.6% or more. DM: the subjects with history of diabetes mellitus. Healthy 1: the healthy control group for High HbA1c group, Healthy 2: the healthy control group for DM group.
p : chi-square test (#: p<O.1, *: p<0.05) with intake frequencies of green-yellow vegetables and fruits. Low serum levels of carotenoids such as a -and 1 -carotene may reflect to low intake of fruit and vegetables in carotenoids. Spearman's rank correlation coefficient between intake frequency of carrot or pumpkin, however, and serum (3 -carotene was 0.20-0.25 and those between intake frequency of other vegetables or fruits and serum other carotenoids levels were approximately 0.20.29 There was insufficient validity of questionnaire of dietary food intake in this study, therefore, and clearly results might not be obtained. We suggested that individuals with high HbAlc values display lower serum carotenoids such as -and -carotene, due to both low intake frequencies of fruit and vegetables and increased production of the reactive oxygen species by chronic hyperglycemia. ' Another report showed that serum levels of vitamin C and E, which play as antioxidants, were lower in people with DM and impaired glucose tolerance." Furthermore, oral administration of large doses of -tocopherol to subjects with diabetes and of ascorbic acid to non-diabetic subjects decreased glycosylated hemoglobin concentrations. 7.30 Our results that the difference of serum -carotene levels between in High HbA1c group and in Healthy 1 group were approximately 30%. In the NHANES III study,13 the difference of serum /3 -carotene levels in subjects with impaired glucose tolerance or in subjects newly diagnosed with diabetes were 13% and 20% lower, respectively, compared to subjects with normal glucose tolerance. This difference might be affected by the difference of case and control subjects. Our case subjects were high HbA1c group that was HbA1c values of 5.6% or more, but subjects in the NHANES III study were classified by an oral glucose tolerance Table 4. Odds ratios (ORs) and 95% confidence intervals (CIs) of vegetables for high HbA1c and diabetes mellitus Low: 1-2 times/week or less, High: 3-4 times/week or more 95% confidence intervals in parentheses. Adjusted OR : Odds ratio adjusted for smoking habit and alcohol consumption. *: p<0 .05 test based on the World Health Organization criteria. Our control group was healthy subjects with no history of illness, but that of NHANES III study was subjects with normal glucose tolerance including those with illness, except DM. Moreover, serumcarotene levels in our study subjects were higher than in the NHANES III study subjects (geometric mean of the subjects with normal glucose tolerance: 0.425, mol/1).
In the DM group, intake frequency of oranges rich in -cryptoxanthin and serum levels of -cryptoxanthin were lower compared with the Healthy 2 group. Serum levels of other carotenoids such as -carotene were not significantly different between DM and Healthy 2 groups.
No differences in serum levels of total cholesterol and TBARS were observed between DM and Healthy 2 groups. The percentage of ex-smokers was higher in the DM group than the Healthy 2 group. It is possible that these results reflect the effect of medical treatment, including health education, in medical institutions.
Serum levels of lipid peroxidation in DM patients with good control are not elevated. 26 We suggest that the DM group in this study included many patients with good control.
There were two limitations in this study subjects. First , the subjects in the High HbA1c group were individuals with newly untreated DM or abnormal carbohydrate tolerance . It is impossible to exclude newly untreated DM from High HbA1c group , because we cannot diagnose diabetes mellitus from only results of health examination. According the Japan Diabetes Society criteria," the diagnosis of diabetes can be made by fasting blood glucose levels of 126mg/dl or higher and HbAic values of 6 .5% or higher. There were 12 subjects (7 .9%) who fulfilled those criteria (newly DM group) in High HbA1c group of this study. The same results were obtained when those subjects were excluded in analysis. Newly DM group had significantly lower levels of serum -carotene than Healthy 1 group . Serum other carotenoids levels in newly DM group were lower than Healthy 1 group , but p: t test (log-transformed ) *: p<0 .05, ***: p<0.001 not significant (data not shown). Second, many subjects had attended these health examinations two or more times in the present study. We therefore suggest that many subjects paid close attention to their health. Subjects in the DM group complied with health guidance and many good control patients were included.
Results from this study suggest that intake of fruit and vegetables rich in carotenoids might be a protective factor against hyperglycemia. Table 6. Odds ratios (ORs) and 95% confidence intervals (CIs) of serum levels of carotenoids, -tocopherol and retinol for high HbA1c 95% confidence intervals in parenthesis.