3.1 Baseline Characteristics
Baseline characteristics of participants divided by quartiles of CDAI are shown in Table 1. A total of 18046 eligible patients were enrolled as the primary cohort, including 8395 males (47.62%) and 9453 females (52.38%). Most were under 65 years old. Higher levels of CDAI were more common in males under 65 years old, with significantly higher proportions of high CDAI in males than females, possibly related to differences in dietary habits. CDAI levels were higher in non-Hispanic Asian race, higher education and income levels, and those taking dietary supplements. Never smokers and former smokers had higher proportions of high CDAI than current smokers, and in current smokers, the CDAI group 1 percentage (31.58%) was highest, indicating this group had lower antioxidant intakes but higher oxidative stress from smoking, conferring greater health risks. The proportion of participants with diabetes and hypertension decreased with increasing CDAI, showing a negative correlation. Since these conditions increase all-cause mortality, this suggests a protective effect of antioxidant-rich diets. BMI decreased with increasing CDAI, especially between groups 3 and 4, indicating antioxidant-rich diets help lower BMI in obese (BMI ≥ 28).
Table 1
Baseline characteristic map after stratification according to CDAI quartile.
Factor | Total | CDAI group 1 | CDAI group 2 | CDAI group 3 | CDAI group 4 | P |
Age(%) | | | | | | < 0.001 |
18–65 | 13954(77.32) | 3329(23.86) | 3426(24.55) | 3523(25.25) | 3676(26.34) | |
≥ 65 | 4092(22.68) | 1183(28.91) | 1086(26.54) | 987(24.12) | 836(20.43) | |
Gender(%) | | | | | | < 0.001 |
Female | 9453(52.38) | 3051(32.28) | 2651(28.04) | 2262(23.93) | 1498(15.75) | |
Male | 8593(47.62) | 1461(17.00) | 1861(21.66) | 2248(26.16) | 3023(35.18) | |
Race(%) | | | | | | < 0.001 |
Mexican American | 2400(13.30) | 514(21.42) | 597(24.88) | 622(25.92) | 667(27.79) | |
Non-Hispanic Asian | 2001(11.09) | 372(18.59) | 471(23.54) | 568(28.39) | 590(29.49) | |
Non-Hispanic Black | 4208(23.32) | 1339(31.82) | 1060(25.19) | 936(22.24) | 873(20.75) | |
Non-Hispanic White | 6940(38.46) | 1617(23.30) | 1772(25.53) | 1774(25.56) | 1777(25.61) | |
Other Hispanic | 1813(10.05) | 496(27.36) | 433(23.88) | 462(25.48) | 422(23.28) | |
Other Race - Including Multi-Racial | 684(3.79) | 174(25.44) | 179(26.17) | 148(21.64) | 183(26.75) | |
Education level(%) | | | | | | < 0.001 |
< 11th grade | 3398(19.85) | 1118(32.90) | 869(25..57) | 765(22.51) | 646(19.01) | |
High school graduate/GED or equivalent | 3829(22.37) | 1120(29.25) | 967(25.25) | 926(24.18) | 816(21.31) | |
Some college or AA degree | 5384(31.45) | 1336(24.81) | 1361(25.28) | 1347(25.02) | 1340(24.89) | |
College graduate or above | 4507(26.33) | 618(13.71) | 1079(23.94) | 1284(28.49) | 1526(33.86) | |
Income(%) | | | | | | < 0.001 |
Under $20,000 | 6658(39.02) | 1673(25.13) | 1591(23.90) | 1679(25.22) | 1715(25.76) | |
$20,000–$75,000 | 8788(51.51) | 2273(25.86) | 2281(25.96) | 2133(24.27) | 2101(23.91) | |
Over $75,000 | 1616(9.47) | 286(17.70) | 405(25.06) | 463(28.65) | 462(28.59) | |
Multivitamin supplement use(%) | | | | | | < 0.001 |
Yes | 8650(47.93) | 1805(20.87) | 2147(24.82) | 2300(26.59) | 2398(27.72) | |
No | 9393(52.05) | 2707(28.82) | 2364(25.17) | 2209(23.52) | 2113(22.50) | |
Smoking status(%) | | | | | | < 0.001 |
Never | 10387(58.39) | 2511(24.17) | 2647(25.48) | 2629(25.31) | 2600(25.03) | |
Former | 3953(22.22) | 830(21.00) | 962(24.34) | 1083(27.40) | 1078(27.27) | |
Now | 3448(19.38) | 1089(31.58) | 838(24.30) | 744(21.58) | 777(22.53) | |
Alcohol use(%) | | | | | | 0.048 |
Yes | 12153(87.19) | 2849(23.44) | 3026(24.90) | 3117(25.65) | 3161(26.01) | |
No | 1789(12.83) | 462(25.82) | 453(25.32) | 413(23.09) | 461(25.77) | |
Hypertension(%) | | | | | | < 0.001 |
Yes | 6501(36.02) | 1783(27.43) | 1706(26.24) | 1534(23.60) | 1478(22.73) | |
No | 11526(63.90) | 2722(23.62) | 2797(24.27) | 2974(25.80) | 3033(26.31) | |
DM(%) | | | | | | < 0.001 |
Yes | 2433(13.48) | 714(29.35) | 648(26.63) | 590(24.25) | 481(19.77) | |
No | 15137(83.88) | 3696(24.42) | 3744(24.73) | 3799(25.10) | 3898(25.75) | |
BMI | 17856(29.39) | 4450(29.96)cd | 4467(29.56)d | 4468(29.38)ad | 4471(28.65)abc | < 0.001 |
3.2 CDAI and All-Cause Mortality
Cox regression models were constructed using univariate and multivariate analysis of gender, age, race, BMI, use of dietary supplements, smoking status, history of diabetes, and history of hypertension. Univariate Cox regression showed all factors except BMI were significantly associated with all-cause mortality. HRs showed higher CDAI was associated with lower all-cause mortality (vs group 1, group 2 = 0.835, group 3 = 0.721, group 4 = 0.611), indicating higher CDAI is more beneficial. Older age (≥ 65 years), former and current smokers, hypertension and diabetes markedly increased HRs for all-cause mortality, suggesting deleterious health effects. Males had a higher HR than females (vs females, HR = 1.502), likely related to more unhealthy behaviors like smoking. Multivariate Cox regression showed BMI was associated with all-cause mortality (HR = 0.982) but the impact was relatively small. Other variables showed similar trends as the univariate model. The forest plot (Fig. 2) depicts HRs and 95% CIs for each factor. Despite adjustment for other factors, all-cause mortality HR decreased with increasing CDAI (vs group 1, group 2 = 0.852, group 3 = 0.771, group 4 = 0.699), indicating antioxidant-rich diets are beneficial. Restricted cubic spline (Fig. 3, p < 0.05, p-Nonlinear = 0.037 < 0.05) showed a nonlinear relationship between CDAI and all-cause mortality, suggesting CDAI of -0.71 ~ 25.52 lowered all-cause mortality risk, consistent with the Cox regression results.
Subsequently, models were adjusted for gender, age, race, BMI, use of dietary supplements, smoking, diabetes, and hypertension to explain the role of covariates. Model 1: adjusted for gender, age, and race; Model 2: adjusted for gender, age, race, use of dietary supplements, smoking status, and BMI; Model 3: adjusted for gender, age, race, use of dietary supplements, smoking status, BMI, diabetes and hypertension. All three models showed an inverse association between all-cause mortality and CDAI (Table 2), with substantial lowering of HRs, by 36.9%, 36.8%, and 30.1% for group 4 in models 1, 2 and 3, respectively. This demonstrates markedly lowered all-cause mortality with increasing CDAI. Additionally, the benefits of CDAI were diminished with increasing unhealthy factors (smoking, diabetes, hypertension). This illustrates the considerable harm of smoking, diabetes, and hypertension, thus maintaining a healthy lifestyle in addition to a high antioxidant diet yields greater health benefits.
Table 2
The relationship between CDAI and all-cause mortality after model adjustment.
Delamination | Model 1 | Model 2 | Model 3 |
| HR | 95.0% CI | P | HR | 95.0% CI | P | HR | 95.0% CI | P |
CDAI group 1 | Reference | | | Reference | | | Reference | | |
CDAI group 2 | 0.827 | 0.704–0.972 | 0.021 | 0.840 | 0.711–0.992 | 0.040 | 0.852 | 0.720–1.007 | 0.060 |
CDAI group 3 | 0.715 | 0.604–0.847 | < 0.001 | 0.721 | 0.605–0.859 | < 0.001 | 0.771 | 0.647–0.919 | 0.004 |
CDAI group 4 | 0.631 | 0.527–0.756 | < 0.001 | 0.632 | 0.524–0.763 | < 0.001 | 0.699 | 0.579–0.844 | < 0.001 |
3.3 CDAI and Cancer Mortality
Univariate Cox regression showed all factors except BMI were associated with cancer mortality, similar to all-cause mortality. Other factors had similar impacts as for all-cause mortality. Multivariate Cox regression showed use of dietary supplements and diabetes history were not significantly associated with cancer mortality when adjusted for other factors. We believe this is because these two factors have smaller impacts on cancer mortality compared to other factors. The forest plot (Fig. 4) showed cancer mortality HR decreased with increasing CDAI (vs group 1, group 2 = 0.705, group 3 = 0.690, group 4 = 0.727). However, group 4 showed less reduction compared to group 3, also seen in univariate analysis. This suggests that while higher dietary antioxidant intake lowers cancer mortality, excessive intake does not confer additional benefits. Restricted cubic spline (Fig. 5, p > 0.05, p-Nonlinear = 0.44 > 0.05) showed no evidence of a nonlinear relationship between CDAI and cancer mortality.
For cancer mortality, all three adjusted models showed an inverse association between CDAI and risk, with HR reduction of 29.0%, 31.0%, and 27.3% for group 4 in models 1, 2, and 3 (Table 3). This demonstrates high dietary antioxidant intake significantly lowers cancer mortality risk. Also, group 4 had lower benefits than group 3, consistent with multivariate regression, further indicating excessive antioxidant intake does not provide additional benefits for cancer mortality.
Table 3
The relationship between CDAI and cancer mortality after model adjustment.
Delamination | Model 1 | Model 2 | Model 3 |
| HR | 95.0% CI | P | HR | 95.0% CI | P | HR | 95.0% CI | P |
CDAI group 1 | Reference | | | Reference | | | Reference | | |
CDAI group 2 | 0.688 | 0.490–0.966 | 0.031 | 0.704 | 0.499–0.992 | 0.045 | 0.705 | 0.500-0.995 | 0.047 |
CDAI group 3 | 0.655 | 0.463–0.925 | 0.016 | 0.667 | 0.469–0.949 | 0.024 | 0.690 | 0.484–0.982 | 0.039 |
CDAI group 4 | 0.710 | 0.500-1.006 | 0.054 | 0.690 | 0.481–0.990 | 0.044 | 0.727 | 0.507–1.044 | 0.084 |
3.4 Stratified Analysis
Given differences in CDAI distribution by gender, analyses were stratified by males and females and models were adjusted in each group to better explore the CDAI-mortality relationships. Models 1, 2, and 3 showed inverse associations between CDAI and all-cause mortality for both genders. In model 3, a greater benefit was seen for group 3 (HR = 0.779) vs group 4 (HR = 0.784) in females, while in males group 4 (HR = 0.650) had greater benefits than group 3 (HR = 0.758). Also, group 4 had a 35.0% lower all-cause mortality risk compared to group 1 in males, markedly higher than the 21.6% for females (Table 4). This indicates CDAI has greater protective effects in males under multivariate adjustment. For cancer mortality, an inverse association was still seen in males, but in model 3 the risk reduction for group 4 (HR = 0.678) vs group 3 (HR = 0.605) was smaller (Table 5), consistent with multivariate regression and model adjustment results, further confirming excessive CDAI does not provide additional benefits.
Table 4
The relationship between CDAI and all-cause mortality stratified by gender according to model adjustment.
Stratification by gender | Model 1 | Model 2 | Model 3 |
HR | 95.0% CI | P | HR | 95.0% CI | P | HR | 95.0% CI | P |
Female | |
CDAI group 1 | Reference | | | Reference | | | Reference | | |
CDAI group 2 | 0.850 | 0.678–1.065 | 0.157 | 0.877 | 0.695–1.107 | 0.270 | 0.912 | 0.721–1.152 | 0.438 |
CDAI group 3 | 0.665 | 0.513–0.861 | 0.002 | 0.695 | 0.533–0.906 | 0.007 | 0.779 | 0.596–1.018 | 0.068 |
CDAI group 4 | 0.688 | 0.505–0.936 | 0.017 | 0.705 | 0.513–0.971 | 0.032 | 0.784 | 0.569–1.081 | 0.138 |
Male | |
CDAI group 1 | Reference | | | Reference | | | Reference | | |
CDAI group 2 | 0.805 | 0.638–1.014 | 0.066 | 0.804 | 0.633–1.020 | 0.073 | 0.797 | 0.627–1.014 | 0.064 |
CDAI group 3 | 0.742 | 0.591–0.932 | 0.010 | 0.730 | 0.576–0.924 | 0.009 | 0.758 | 0.598–0.961 | 0.022 |
CDAI group 4 | 0.610 | 0.485–0.768 | < 0.001 | 0.598 | 0.471–0.759 | < 0.001 | 0.650 | 0.511–0.827 | < 0.001 |
Table 5
The relationship between CDAI and cancer mortality stratified by gender according to model adjustment.
Stratification by gender | Model 1 | Model 2 | Model 3 |
HR | 95.0% CI | P | HR | 95.0% CI | P | HR | 95.0% CI | P |
Female | |
CDAI group 1 | Reference | | | Reference | | | Reference | | |
CDAI group 2 | 0.740 | 0.459–1.192 | 0.215 | 0.760 | 0.469–1.233 | 0.267 | 0.785 | 0.482–1.278 | 0.331 |
CDAI group 3 | 0.736 | 0.445–1.215 | 0.231 | 0.746 | 0.447–1.246 | 0.263 | 0.811 | 0.484–1.359 | 0.426 |
CDAI group 4 | 0.763 | 0.423–1.378 | 0.371 | 0.766 | 0.420–1.398 | 0.385 | 0.810 | 0.441–1.486 | 0.496 |
Male | |
CDAI group 1 | Reference | | | Reference | | | Reference | | |
CDAI group 2 | 0.630 | 0.389–1.020 | 0.060 | 0.638 | 0.393–1.037 | 0.070 | 0.640 | 0.393–1.042 | 0.073 |
CDAI group 3 | 0.585 | 0.364–0.940 | 0.027 | 0.595 | 0.367–0.963 | 0.034 | 0.605 | 0.373–0.981 | 0.041 |
CDAI group 4 | 0.670 | 0.431–1.042 | 0.076 | 0.642 | 0.407–1.013 | 0.057 | 0.678 | 0.428–1.072 | 0.096 |