Abstract
Objective
Plant-based diets have multiple health benefits for cancers; however, little is known about the association between plant-based dietary patterns and esophageal cancer (EC).This study presents an investigation of the prospective associations among three predefined indices of plant-based dietary patterns and the risk of EC.
Methods
We performed endoscopic screening for 15,709 participants aged 40−69 years from two high-risk areas of China from January 2005 to December 2009 and followed the cohort until December 31, 2022. The overall plant-based diet index (PDI), healthful plant-based diet index (hPDI), and unhealthful plant-based diet index (uPDI), were calculated using survey responses to assess dietary patterns. We applied Cox proportional hazard regression to estimate the multivariable hazard ratios (HRs) and 95% confidence intervals (95% CIs) of EC across 3 plant-based diet indices and further stratified the analysis by subgroups.
Results
The final study sample included 15,184 participants in the cohort. During a follow-up of 219,365 person-years, 176 patients with EC were identified. When the highest quartile was compared with the lowest quartile, the pooled multivariable-adjusted HR of EC was 0.50 (95% CI, 0.32−0.77) for hPDI. In addition, the HR per 10-point increase in the hPDI score was 0.42 (95% CI, 0.27−0.66) for ECs. Conversely, uPDI was positively associated with the risk of EC, and the HR was 1.80 (95% CI, 1.16−2.82). The HR per 10-point increase in the uPDI score was 1.90 (95% CI, 1.26−2.88) for ECs. The associations between these scores and the risk of EC were consistent in most subgroups. These results remained robust in sensitivity analyses.
Conclusions
A healthy plant-based dietary pattern was associated with a reduced risk of EC. Emphasizing the healthiness and quality of plant-based diets may be important for preventing the development of EC.
Keywords: Esophageal cancer, diet, dietary pattern, epidemiology, nutrition, prospective cohort study
Introduction
Esophageal cancer (EC) is the eighth most frequently diagnosed cancer worldwide and ranks as the sixth leading cause of cancer-related deaths, with approximately 604,100 new cases and 544,100 deaths (1). An overwhelming 53.7% of all EC cases globally are reported in China, with estimates of 324,422 new cases and 301,135 deaths in 2020 (2).
Diet is a major modifiable risk factor for EC. The esophageal mucosa is in direct contact with ingested food, and its physiology is closely influenced by hormonal and neural connections, which are affected by the quality, quantity, and timing of meals. Plant-based diets such as vegetarian or vegan diets are significantly associated with a decreased risk of total cancer (3). Nevertheless, epidemiological evidence on the association between plant-based diets and EC is scarce, particularly in Chinese populations, whose dietary patterns differ markedly from those of Western countries.
Plant-based diets represent dietary approaches that emphasize the consumption of foods derived from plants while restricting or excluding certain or all animal-based products (4). Epidemiological studies have provided evidence supporting the advantageous effects of specific plant-based foods, such as whole grains and fibre, in reducing EC risk (5,6). However, it is essential to note that not all plant foods are beneficial, as certain items, such as pickled vegetables, may be harmful to ECs (7). To better account for the healthiness and quality of plant-based foods in previous studies, plant-based diet indices (PDIs) were proposed. PDIs encompass the overall plant-based diet index (PDI), the healthful plant-based diet index (hPDI), and the unhealthful plant-based diet index (uPDI) (8,9). Greater hPDI scores have been inversely associated with cardiovascular diseases (10), diabetes (8) and certain cancers (11,12), while higher uPDI scores have shown adverse associations with the incidence of these diseases. The original PDIs were type 2 diabetes (T2D)-specific indices that used existing knowledge of food associations with T2D to differentiate between healthy and less healthy plant foods, rather than being tailored specifically for cancer or EC research. Thus, plant foods that did not exhibit a clear and unidirectional relationship with these health outcomes, such as alcoholic beverages, were deliberately excluded from the indices (8). However, strong evidence related to alcohol consumption indicates an increased risk of esophageal squamous cell carcinoma (ESCC) (13). In the study of the effects of PDIs on cognitive function in the Chinese population, the Chinese culinary attribute was introduced for the first time in traditional PDIs, providing an exploratory refinement. These adjustments involved restructuring the original PDIs framework, including garlic and pickled vegetables, and fine-tuning the scoring methodology (14,15). Currently, the association between plant-based diets and EC incidence has been examined in only one study within U.S. cohorts, which may not reflect the distribution of EC histological subtypes in China (16). Hence, we explored the associations of PDIs with the dietary habits of high-risk sites of EC in China by referring to the foods in the original PDIs.
The aim of the present study was to prospectively examine the associations between three predefined indices of plant-based dietary patterns and the risk of EC using a cohort from high-risk sites of the National Cancer Early Detection and Treatment Project (NCEDTP) (Linzhou and Feicheng) in China.
Materials and methods
Study population
This study was carried out at two high-risk sites in the NCEDTP (Linzhou and Feicheng) (17). All residents aged between 40−69 years were the target population and 31,826 participants were included. This age group was chosen based on early research, which indicated that the annual age-adjusted mortality rates for EC among individuals aged 40−69 years were 100−150 times greater than the rates adjusted for U.S. whites (18). Individuals who declined to undergo endoscopic screening were excluded. The participation rate in the target population was 49.36%, enrolling a total of 15,709 individuals from January 2005 to December 2009. Participants completed questionnaires about demographic information, height, weight, lifestyle factors, previous medical history of digestive system diseases, and family history of cancer. The study was approved by the Institutional Review Board of the National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (No. 20/386-2582).
Surveillance of EC
All patients completed baseline surveys during the period spanning from January 2005 to December 2009. Then, they received follow-up through monthly phone calls, home visits every three months, reviews of medical records, and annual assessments of cancer registries. We continued follow-up until the occurrence of EC, death, or the end of follow-up (31 December 2022), whichever occurred first. For this analysis, we excluded participants who were diagnosed with upper gastrointestinal cancer before or at baseline, had an unclear diagnosis, or whose information was missing from the questionnaire. We used the tenth edition of the International Classification of Diseases (ICD-10) for case identification: esophageal cancer (C15.0−C15.5, C15.8−C15.9). A total of 15,184 participants are included in the final analysis (Figure 1).
Figure 1.
Flowchart of study population selection.
Dietary assessment
The trained investigators performed a face-to-face survey using standard questionnaires, and participants answered the questions by themselves. The dietary section of the questionnaire included 15 commonly consumed food groups at high-risk sites for EC in China: fresh vegetables, fresh fruits, legumes, dried fruits and vegetables, staple foods, spring onions, garlic, tea, pickled vegetables, fried food, hot food, mildewed food, dairy products, eggs, meat, alcoholic beverages, and tea. The definitions of these food groups are reported in the (Supplementary Table S1). The questionnaire recorded the intake frequency and duration of time as “times/week”, “times/month”, “times/year” and “duration of time” for most food groups, except for the consumption of tea and alcoholic beverages. The questionnaire recorded the frequency of intake of tea and alcoholic beverages as “yes” or “no”.
Table S1. Definition of PDI, hPDI and uPDI.
| Variables | PDI | hPDI | uPDI |
| PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index. Positive, positive scores [quintile (Q) 1=1, Q2=2, Q3=3, Q4=4, Q5=5]; Reverse, reverse scores (Q5=1, Q4=2, Q3=3, Q2=4, Q1=5). a, mixed grains, mainly consisting of corn and wheat; b, mainly include fried cake, deep-fried dough sticks; c, mainly include hot soup, hot noodles; d, mainly include mildewed staple food; e, mainly include alcoholic beverages, beer, fruit wine; f, mainly include pork, beef, rabbit, chicken, or duck. | |||
| Plant food groups | |||
| Healthy | |||
| Fresh vegetable | Positive | Positive | Reverse |
| Fresh fruit | Positive | Positive | Reverse |
| Legumes | Positive | Positive | Reverse |
| Dried fruit and vegetable | Positive | Positive | Reverse |
| Staple fooda | Positive | Positive | Reverse |
| Spring onion, garlic | Positive | Positive | Reverse |
| Tea | Positive | Positive | Reverse |
| Less healthy | |||
| Pickled vegetable | Positive | Reverse | Positive |
| Fried foodb | Positive | Reverse | Positive |
| Hot foodc | Positive | Reverse | Positive |
| Mildewed foodd | Positive | Reverse | Positive |
| Alcoholic beveragee | Positive | Reverse | Positive |
| Animal food groups | |||
| Dairy products | Reverse | Reverse | Reverse |
| Egg | Reverse | Reverse | Reverse |
| Meatf | Reverse | Reverse | Reverse |
We assessed the plant-based dietary pattern by calculating three plant-based diet indices (PDI, hPDI, and uPDI), an adapted approach used in previous studies (8,14). The included food groups covered the most common food consumed in the daily diet at high-risk sites in China from 2005 to 2009. We incorporated a total of 15 food groups for evaluation; these groups were differentially associated with a set of health outcomes (19,20). The healthy plant foods consisted of fresh vegetables, fresh fruits, legumes, dried fruits and vegetables, staple foods, spring onions, garlic, and tea. The less healthy plant foods included pickled vegetables, fried food, hot food, mildewed food, and alcoholic beverages. The animal foods were represented by dairy products, eggs, and meat. The categorization of the dietary assessments is determined in advance of the study (Supplementary Table S1).
We scored three plant-based diet indices according to intake frequency. Although servings or quintiles of intake are commonly used, using a nonquantitative food frequency questionnaire to assess dietary patterns has been demonstrated to be reliable and valid in some studies (21-23). In addition, previous studies have shown that the frequency of intake is more important than the portion size for distinguishing between high and low consumption of fruits and vegetables (24). The average number of intakes was calculated according to the intake frequency and duration. The frequency of intake was converted to “almost every day”, “≥1 time/week”, “≥1 time/month”, “occasionally”, or “rarely or never”. The participants’ scores in the study followed a distinct pattern: the highest frequency of consumption for any plant food group received a score of 5, whereas the lowest frequency of consumption received a score of 1. In contrast, the scoring for the animal food group was the opposite: the highest frequency of consumption received a score of 1, while the lowest frequency of consumption received a score of 5. Tea and alcoholic beverages were coded as “yes” or “no” and were assigned scores of 1 and 5, depending on the specific indices used in the study. More details on constructing and scoring PDIs can be found in Supplementary Table S2. For the PDI, higher scores were assigned to any plant food group, whereas reverse scores were assigned to animal food groups. For the hPDI, higher scores were given to healthy plant food groups, while reverse scores were assigned to less healthy plant food groups and animal food groups. Conversely, for the uPDI, higher scores were assigned to less healthy plant food groups, and reverse scores were given to healthy plant food groups and animal food groups. The 15 food group scores for an individual were summed to obtain the indices, with a theoretical range of 15−75.
Table S2. Plant-based diet indices scoring.
| Food | N (times/week) | Frequency | PDI | hPDI | uPDI |
| PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index; N=(W*4+M+Y/12)*D/48; W, times/week; M, times/month; Y, times/year; D, duration of time. | |||||
| Plant food group | |||||
| Healthy | |||||
| Fresh vegetable | N≥7.00 | Almost everyday | 5 | 5 | 1 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 4 | 2 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 2 | 4 | |
| N=0 | Rarely or never | 1 | 1 | 5 | |
| Fresh fruit | N≥7.00 | Almost everyday | 5 | 5 | 1 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 4 | 2 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 2 | 4 | |
| N=0 | Rarely or never | 1 | 1 | 5 | |
| Legumes | N≥7.00 | Almost everyday | 5 | 5 | 1 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 4 | 2 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 2 | 4 | |
| N=0 | Rarely or never | 1 | 1 | 5 | |
| Dried fruits and vegetables | N≥7.00 | Almost everyday | 5 | 5 | 1 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 4 | 2 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 2 | 4 | |
| N=0 | Rarely or never | 1 | 1 | 5 | |
| Staple food | N≥7.00 | Almost everyday | 5 | 5 | 1 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 4 | 2 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 2 | 4 | |
| N=0 | Rarely or never | 1 | 1 | 5 | |
| Spring onion, garlic | N≥7.00 | Almost everyday | 5 | 5 | 1 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 4 | 2 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 2 | 4 | |
| N=0 | Rarely or never | 1 | 1 | 5 | |
| Drink tea | Yes | Yes | 5 | 5 | 1 |
| No | No | 1 | 1 | 5 | |
| Less healthy | |||||
| Pickled vegetable | N≥7.00 | Almost everyday | 5 | 1 | 5 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 2 | 4 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 4 | 2 | |
| N=0 | Rarely or never | 1 | 5 | 1 | |
| Fried food | N≥7.00 | Almost everyday | 5 | 1 | 5 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 2 | 4 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 4 | 2 | |
| N=0 | Rarely or never | 1 | 5 | 1 | |
| Hot food | N≥7.00 | Almost everyday | 5 | 1 | 5 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 2 | 4 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 4 | 2 | |
| N=0 | Rarely or never | 1 | 5 | 1 | |
| Mildewed food | N≥7.00 | Almost everyday | 5 | 1 | 5 |
| 1.00≤N<7.00 | ≥1 time/week | 4 | 2 | 4 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 2 | 4 | 2 | |
| N=0 | Rarely or never | 1 | 5 | 1 | |
| Alcoholic beverage | Yes | Yes | 5 | 1 | 5 |
| No | No | 1 | 5 | 1 | |
| Animal food groups | |||||
| Meat, egg, dairy products | N≥7.00 | Almost everyday | 1 | 1 | 1 |
| 1.00≤N<7.00 | ≥1 time/week | 2 | 2 | 2 | |
| 0.25≤N<1.00 | ≥1 time/month | 3 | 3 | 3 | |
| 0<N<0.21 | Occasionally | 4 | 4 | 4 | |
| N=0 | Rarely or never | 5 | 5 | 5 | |
Assessment of covariates
We assessed demographic characteristics, socioeconomic status, lifestyle factors, and medical status. The covariates included age (year), gender (male or female), marital status (married, unmarried, divorced, widowed), educational status (no education, elementary school, middle school and above), number of family members, annual income (CNY per person), smoking status (yes or no), prior history of digestive diseases (yes or no), and family history of cancer (yes or no). Additionally, we calculated body mass index (BMI) based on height and weight.
Statistical analysis
We summarized the baseline characteristics using descriptive statistics. Cox proportional hazard regression was used to estimate the hazard ratios (HRs) and 95% confidence intervals (95% CIs) for the associations between three plant-based diet indices and the risk of EC. The follow-up time in person-years was calculated from the baseline to the date of diagnosis of EC, the date of death, or the end of the follow-up period (31 December 2022), whichever came first. We categorized PDI, hPDI, and uPDI into quartiles (quartile 1 to quartile 4). In the basic model, PDI, hPDI, and uPDI were included separately. In the minimally adjusted model, we adjusted for age and gender. In the fully adjusted model, we further adjusted for marital status, educational status, number of family members, annual income, smoking status, prior history of digestive diseases, family history of cancer, and BMI categories (<23.9 kg/m2, ≥24.0 kg/m2) (25). We also estimated the risk of EC per 10-point increase in PDI, hPDI, and uPDI. We separately plotted three knots of cubic splines to explore nonlinearity between three plant-based diet indices and the risk of EC. The Wald test was used to assess whether the observed relationships were linear or nonlinear. We conducted correlation analysis to assess the relationships among various food groups. Additionally, we performed food component analysis to investigate the individual contributions of healthy plant food, less healthy plant food, and animal food groups.
We performed several sensitivity analyses to test the robustness of the results; these included the exclusion of individuals who developed EC within one year of follow-up. Additionally, we excluded individuals with diabetes, stroke, or myocardial infarction to account for the possibility that their usual diet may have changed following the diagnosis of these chronic diseases. Moreover, we explored the associations of PDI, hPDI, and uPDI with ESCC incidence. Given that alcohol was not originally included in the PDIs, we also included alcohol consumption as a covariate in our analysis rather than in the PDIs.
All the statistical analyses were carried out using R software (Version 4.1.1; R Foundation for Statistical Computing, Vienna, Austria). P<0.05 was considered to indicate statistical significance in two-sided tests.
Results
Baseline characteristics of participants
Supplementary Table S3 provides an overview of the baseline characteristics of the 15,184 participants included in the study, spanning from 2005 to 2009. The participants had a mean age of 51.84 years [standard deviation (SD): 7.50], and 54.01% were females. The majority (93.45%) were married, and 49.51% had completed nine years of compulsory education. Approximately 27.73% of the participants were current smokers, and 40.15% were obese. Approximately 11.75% had a history of digestive diseases, and 29.50% had a family history of cancer.
Table S3. Participants’ characteristics at baseline.
| Variables | n (%) | χ2 | P | ||
| Healthy population (N=15,008) | EC case (N=176) | Total (N=15,184) | |||
| EC, esophageal cancer; BMI, body mass index. | |||||
| Gender | |||||
| Male | 6,865 (45.74) | 118 (67.05) | 6,983 (45.99) | 31.784 | <0.001 |
| Female | 8,143 (54.26) | 58 (32.95) | 8,201 (54.01) | ||
| Age at baseline (year) | |||||
| 40−49 | 6,255 (41.68) | 36 (20.45) | 6,291 (41.43) | 32.988 | <0.001 |
| 50−59 | 6,132 (40.86) | 102 (57.95) | 6,234 (41.06) | ||
| 60−69 | 2,621 (17.46) | 38 (21.59) | 2,659 (17.51) | ||
| Marital status | |||||
| Married | 14,029 (93.48) | 160 (90.91) | 14,189 (93.45) | 1.873 | 0.171 |
| Unmarried/divorced/widowed | 979 (6.52) | 16 (9.09) | 995 (6.55) | ||
| Educational status | |||||
| No education | 2,201 (14.67) | 20 (11.36) | 2,221 (14.63) | 2.085 | 0.353 |
| Elementary school | 5,384 (35.87) | 61 (34.66) | 5,445 (35.86) | ||
| Middle school and above | 7,423 (49.46) | 95 (53.98) | 7,518 (49.51) | ||
| No. of family members | |||||
| 1−2 | 3,281 (21.86) | 58 (32.95) | 3,339 (21.99) | 15.412 | 0.004 |
| 3 | 2,856 (19.03) | 22 (12.50) | 2,878 (18.95) | ||
| 4 | 5,038 (33.57) | 51 (28.98) | 5,089 (33.52) | ||
| 5 | 2,502 (16.67) | 32 (18.18) | 2,534 (16.69) | ||
| ≥6 | 1,331 (8.87) | 13 (7.39) | 1,344 (8.85) | ||
| Annual income (CNY/person) | |||||
| ≤5,000 | 4,857 (32.36) | 69 (39.20) | 4,926 (32.44) | 3.715 | 0.054 |
| >5,000 | 10,151 (67.64) | 107 (60.80) | 10,258 (67.56) | ||
| BMI (kg/m2) | |||||
| <18.5 | 515 (3.43) | 8 (4.55) | 525 (3.46) | 6.150 | 0.407 |
| 18.5−23.9 | 8,453 (56.32) | 110 (62.50) | 8,562 (56.39) | ||
| 24.0−27.9 | 4,805 (32.02) | 51 (28.98) | 4,855 (31.97) | ||
| ≥28.0 | 1,235 (8.23) | 7 (3.98) | 1,242 (8.18) | ||
| Smoking status | |||||
| No | 10,868 (72.41) | 105 (59.66) | 10,973 (72.27) | 14.123 | <0.001 |
| Yes | 4,140 (27.59) | 71 (40.34) | 4,211 (27.73) | ||
| History of digestive diseases | |||||
| No | 13,246 (88.26) | 154 (87.50) | 13,400 (88.25) | 0.097 | 0.756 |
| Yes | 1,762 (11.74) | 22 (12.50) | 1,784 (11.75) | ||
| Family history of cancer | |||||
| No | 10,587 (70.54) | 118 (67.05) | 10,705 (70.50) | 1.023 | 0.312 |
| Yes | 4,421 (29.46) | 58 (32.95) | 4,479 (29.50) | ||
The baseline characteristics of 15,184 participants across the PDIs are displayed in Table 1. Compared with participants in the lowest quartile of the PDI and hPDI, those in the highest quartile were more likely to be female, be married, have a higher BMI, and have a higher income (Table 1, Supplementary Tables S4, S5). They were also more likely to be nonsmokers and not have a family history of cancer. Conversely, in the case of uPDI, participants in the highest quartile were more likely to be male and unmarried and to have a lower BMI and a lower income than were those in the lowest quartile (Table 1, Supplementary Table S6).
Table 1. Baseline characteristics of participants by lowest vs. highest quartile of three plant-based diet indices in the cohort.
| Variables | hPDI [n (%)] | uPDI [n (%)] | PDI [n (%)] | |||||
| Quartile 1 | Quartile 4 | Quartile 1 | Quartile 4 | Quartile 1 | Quartile 4 | |||
| PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index. | ||||||||
| Gender | ||||||||
| Male | 2,410 (53.72) | 1,452 (38.55) | 1,673 (40.49) | 1,645 (53.05) | 1,898 (50.00) | 1,216 (42.03) | ||
| Female | 2,076 (46.28) | 2,315 (61.45) | 2,459 (59.51) | 1,456 (46.95) | 1,898 (50.00) | 1,677 (57.97) | ||
| Age at baseline (year) | ||||||||
| 40−49 | 1,823 (40.64) | 1,615 (42.87) | 1,798 (43.51) | 1,261 (40.66) | 1,488 (39.20) | 1,281 (44.28) | ||
| 50−59 | 1,851 (41.26) | 1,522 (40.40) | 1,660 (40.17) | 1,271 (40.99) | 1,626 (42.83) | 1,141 (39.44) | ||
| 60−69 | 812 (18.10) | 630 (16.72) | 674 (16.31) | 569 (18.35) | 682 (17.97) | 471 (16.28) | ||
| Marital status | ||||||||
| Married | 4,153 (92.58) | 3,535 (93.84) | 3,899 (94.36) | 2,850 (91.91) | 3,501 (92.23) | 2,737 (94.61) | ||
| Unmarried/divorced/widowed | 333 (7.42) | 232 (6.16) | 233 (5.64) | 251 (8.09) | 295 (7.77) | 156 (5.39) | ||
| Educational status | ||||||||
| No education | 596 (13.29) | 675 (17.92) | 747 (18.08) | 392 (12.64) | 446 (11.75) | 570 (19.70) | ||
| Elementary school | 1,551 (34.57) | 1,400 (37.16) | 1,484 (35.91) | 1,113 (35.89) | 1,439 (37.91) | 937 (32.39) | ||
| Middle school and above | 2,339 (52.14) | 1,692 (44.92) | 1,901 (46.01) | 1,596 (51.47) | 1,911 (50.34) | 1,386 (47.91) | ||
| No. of family members | ||||||||
| 1−2 | 1,033 (23.03) | 765 (20.31) | 832 (20.14) | 701 (22.61) | 885 (23.31) | 574 (19.84) | ||
| 3 | 821 (18.30) | 750 (19.91) | 832 (20.14) | 531 (17.12) | 698 (18.39) | 584 (20.19) | ||
| 4 | 1,456 (32.46) | 1,269 (33.69) | 1,395 (33.76) | 1,051 (33.89) | 1,253 (33.01) | 993 (34.32) | ||
| 5 | 749 (16.70) | 671 (17.81) | 717 (17.35) | 520 (16.77) | 638 (16.81) | 486 (16.80) | ||
| ≥6 | 427 (9.52) | 312 (8.28) | 356 (8.62) | 298 (9.61) | 322 (8.48) | 256 (8.85) | ||
| Annual income (CNY/person) | ||||||||
| ≤5,000 | 1,538 (34.28) | 1,091 (28.96) | 1,133 (27.42) | 1,114 (35.92) | 1,366 (35.99) | 848 (29.31) | ||
| >5,000 | 2,948 (65.72) | 2,676 (71.04) | 2,999 (72.58) | 1,987 (64.08) | 2,430 (64.01) | 2,045 (70.69) | ||
| BMI (kg/m2) | ||||||||
| <18.5 | 149 (3.32) | 125 (3.32) | 144 (3.48) | 108 (3.48) | 161 (4.24) | 112 (3.87) | ||
| 18.5−23.9 | 2,649 (59.05) | 2,022 (53.68) | 2,234 (54.07) | 1,797 (57.95) | 2,136 (56.27) | 1,616 (55.86) | ||
| 24.0−27.9 | 1,355 (30.21) | 1,266 (33.61) | 1,372 (33.20) | 956 (30.83) | 1,197 (31.53) | 941 (32.53) | ||
| ≥28.0 | 333 (7.42) | 354 (9.40) | 382 (9.24) | 240 (7.74) | 302 (7.96) | 224 (7.74) | ||
| Smoking status | ||||||||
| No | 2,793 (62.26) | 3,075 (81.63) | 3,343 (80.91) | 1,932 (62.30) | 2,556 (67.33) | 2,298 (79.43) | ||
| Yes | 1,693 (37.74) | 692 (18.37) | 789 (19.09) | 1,169 (37.70) | 1,240 (32.67) | 595 (20.57) | ||
| History of digestive diseases | ||||||||
| No | 3,854 (85.91) | 3,347 (88.85) | 3,633 (87.92) | 2,679 (86.39) | 3,313 (87.28) | 2,514 (86.90) | ||
| Yes | 632 (14.09) | 420 (11.15) | 499 (12.08) | 422 (13.61) | 483 (12.72) | 379 (13.10) | ||
| Family history of cancer | ||||||||
| No | 3,104 (69.19) | 2,622 (69.60) | 2,867 (69.39) | 2,179 (70.27) | 2,700 (71.13) | 2,084 (72.04) | ||
| Yes | 1,382 (30.81) | 1,145 (30.40) | 1,265 (30.61) | 922 (29.73) | 1,096 (28.87) | 809 (27.96) | ||
Table S4. Demographics for participants according to PDI.
| Variables | n (%) | χ2 | P | |||
| Quartile 1 | Quartile 2 | Quartile 3 | Quartile 4 | |||
| PDI, overall plant-based diet index; BMI, body mass index. | ||||||
| Gender | ||||||
| Male | 1,898 (50.00) | 1,966 (45.50) | 1,903 (45.59) | 1,216 (42.03) | 43.502 | <0.001 |
| Female | 1,898 (50.00) | 2,355 (54.50) | 2,271 (54.41) | 1,677 (57.97) | ||
| Age at baseline (year) | ||||||
| 40−49 | 1,488 (39.20) | 1,764 (40.82) | 1,758 (42.12) | 1,281 (44.28) | 23.043 | <0.001 |
| 50−59 | 1,626 (42.83) | 1,812 (41.93) | 1,655 (39.65) | 1,141 (39.44) | ||
| 60−69 | 682 (17.97) | 745 (17.24) | 761 (18.23) | 471 (16.28) | ||
| Marital status | ||||||
| Married | 3,501 (92.23) | 4,052 (93.77) | 3,899 (93.41) | 2,737 (94.61) | 16.332 | <0.001 |
| Unmarried/divorced/widowed | 295 (7.77) | 269 (6.23) | 275 (6.59) | 156 (5.39) | ||
| Educational status | ||||||
| No education | 446 (11.75) | 581 (13.45) | 624 (14.95) | 570 (19.70) | 94.141 | <0.001 |
| Elementary school | 1,439 (37.91) | 1,563 (36.17) | 1,506 (36.08) | 937 (32.39) | ||
| Middle school and above | 1,911 (50.34) | 2,177 (50.38) | 2,044 (48.97) | 1,386 (47.91) | ||
| No. of family members | ||||||
| 1−2 | 885 (23.31) | 955 (22.10) | 925 (22.16) | 574 (19.84) | 15.592 | 0.211 |
| 3 | 698 (18.39) | 803 (18.58) | 793 (19.00) | 584 (20.19) | ||
| 4 | 1,253 (33.01) | 1,471 (34.04) | 1,372 (32.87) | 993 (34.32) | ||
| 5 | 638 (16.81) | 705 (16.32) | 705 (16.89) | 486 (16.80) | ||
| ≥6 | 322 (8.48) | 387 (8.96) | 379 (9.08) | 256 (8.85) | ||
| Annual income (CNY/person) | ||||||
| ≤5,000 | 1,366 (35.99) | 1,376 (31.84) | 1,336 (32.01) | 848 (29.31) | 35.738 | <0.001 |
| >5,000 | 2,430 (64.01) | 2,945 (68.16) | 2,838 (67.99) | 2,045 (70.69) | ||
| BMI (kg/m2) | ||||||
| <18.5 | 161 (4.24) | 134 (3.10) | 116 (2.78) | 112 (3.87) | 17.982 | 0.035 |
| 18.5−23.9 | 2,136 (56.27) | 2,452 (56.75) | 2,359 (56.52) | 1,616 (55.86) | ||
| 24.0−27.9 | 1,197 (31.53) | 1,369 (31.68) | 1,349 (32.32) | 941 (32.53) | ||
| ≥28.0 | 302 (7.96) | 366 (8.47) | 350 (8.39) | 224 (7.74) | ||
| Smoking status | ||||||
| No | 2,556 (67.33) | 3,037 (70.28) | 3,082 (73.84) | 2,298 (79.43) | 133.830 | <0.001 |
| Yes | 1,240 (32.67) | 1,284 (29.72) | 1,092 (26.16) | 595 (20.57) | ||
| History of digestive diseases | ||||||
| No | 3,313 (87.28) | 3,851 (89.12) | 3,722 (89.17) | 2,514 (86.90) | 15.152 | 0.002 |
| Yes | 483 (12.72) | 470 (10.88) | 452 (10.83) | 379 (13.10) | ||
| Family history of cancer | ||||||
| No | 2,700 (71.13) | 2,973 (68.80) | 2,948 (70.63) | 2,084 (72.04) | 10.013 | 0.018 |
| Yes | 1,096 (28.87) | 1,348 (31.20) | 1,226 (29.37) | 809 (27.96) | ||
Table S5. Demographics for participants according to hPDI.
| Variables | n (%) | χ2 | P | |||
| Quartile 1 | Quartile 2 | Quartile 3 | Quartile 4 | |||
| hPDI, healthful plant-based diet index; BMI, body mass index. | ||||||
| Gender | 211.685 | <0.001 | ||||
| Male | 2,410 (53.72) | 1,735 (47.37) | 1,386 (42.41) | 1,452 (38.55) | ||
| Female | 2,076 (46.28) | 1,928 (52.63) | 1,882 (57.59) | 2,315 (61.45) | ||
| Age at baseline (year) | ||||||
| 40−49 | 1,823 (40.64) | 1,523 (41.58) | 1,330 (40.70) | 1,615 (42.87) | 9.273 | 0.159 |
| 50−59 | 1,851 (41.26) | 1,475 (40.27) | 1,386 (42.41) | 1,522 (40.40) | ||
| 60−69 | 812 (18.10) | 665 (18.15) | 552 (16.89) | 630 (16.72) | ||
| Marital status | ||||||
| Married | 4,153 (92.58) | 3,409 (93.07) | 3,092 (94.61) | 3,535 (93.84) | 14.645 | 0.002 |
| Unmarried/divorced/widowed | 333 (7.42) | 254 (6.93) | 176 (5.39) | 232 (6.16) | ||
| Educational status | ||||||
| No education | 596 (13.29) | 482 (13.16) | 468 (14.32) | 675 (17.92) | 67.578 | <0.001 |
| Elementary school | 1,551 (34.57) | 1,306 (35.65) | 1,188 (36.35) | 1,400 (37.16) | ||
| Middle school and above | 2,339 (52.14) | 1,875 (51.19) | 1,612 (49.33) | 1,692 (44.92) | ||
| No. of family members | ||||||
| 1−2 | 1,033 (23.03) | 848 (23.15) | 693 (21.21) | 765 (20.31) | 26.469 | 0.009 |
| 3 | 821 (18.30) | 690 (18.84) | 617 (18.88) | 750 (19.91) | ||
| 4 | 1,456 (32.46) | 1,234 (33.69) | 1,130 (34.58) | 1,269 (33.69) | ||
| 5 | 749 (16.70) | 591 (16.13) | 523 (16.00) | 671 (17.81) | ||
| ≥6 | 427 (9.52) | 300 (8.19) | 305 (9.33) | 312 (8.28) | ||
| Annual income (CNY/person) | ||||||
| ≤5,000 | 1,538 (34.28) | 1,264 (34.51) | 1,033 (31.61) | 1,091 (28.96) | 35.924 | <0.001 |
| >5,000 | 2,948 (65.72) | 2,399 (65.49) | 2,235 (68.39) | 2,676 (71.04) | ||
| BMI (kg/m2) | ||||||
| <18.5 | 149 (3.32) | 125 (3.41) | 124 (3.79) | 125 (3.32) | 31.863 | <0.001 |
| 18.5−23.9 | 2,649 (59.05) | 2,085 (56.92) | 1,807 (55.29) | 2,022 (53.68) | ||
| 24.0−27.9 | 1,355 (30.21) | 1,170 (31.94) | 1,065 (32.59) | 1,266 (33.61) | ||
| ≥28.0 | 333 (7.42) | 283 (7.73) | 272 (8.32) | 354 (9.40) | ||
| Smoking status | ||||||
| No | 2,793 (62.26) | 2,622 (71.58) | 2,483 (75.98) | 3,075 (81.63) | 412.231 | <0.001 |
| Yes | 1,693 (37.74) | 1,041 (28.42) | 785 (24.02) | 692 (18.37) | ||
| History of digestive diseases | ||||||
| No | 3,854 (85.91) | 3,279 (89.52) | 2,920 (89.35) | 3,347 (88.85) | 34.457 | <0.001 |
| Yes | 632 (14.09) | 384 (10.48) | 348 (10.65) | 420 (11.15) | ||
| Family history of cancer | ||||||
| No | 3,104 (69.19) | 2,647 (72.26) | 2,332 (71.36) | 2,622 (69.60) | 11.771 | 0.008 |
| Yes | 1,382 (30.81) | 1,016 (27.74) | 936 (28.64) | 1,145 (30.40) | ||
Table S6. Demographics for participants according to uPDI.
| Variables | n (%) | χ2 | P | |||
| Quartile 1 | Quartile 2 | Quartile 3 | Quartile 4 | |||
| uPDI, unhealthful plant-based diet index; BMI, body mass index. | ||||||
| Gender | ||||||
| Male | 1,673 (40.49) | 2,193 (45.19) | 1,472 (47.51) | 1,645 (53.05) | 116.676 | <0.001 |
| Female | 2,459 (59.51) | 2,660 (54.81) | 1,626 (52.49) | 1,456 (46.95) | ||
| Age at baseline (year) | ||||||
| 40−49 | 1,798 (43.51) | 1,988 (40.96) | 1,244 (40.15) | 1,261 (40.66) | 12.758 | 0.047 |
| 50−59 | 1,660 (40.17) | 2,003 (41.27) | 1,300 (41.96) | 1,271 (40.99) | ||
| 60−69 | 674 (16.31) | 862 (17.76) | 554 (17.88) | 569 (18.35) | ||
| Marital status | ||||||
| Married | 3,899 (94.36) | 4,567 (94.11) | 2,873 (92.74) | 2,850 (91.91) | 23.664 | <0.001 |
| Unmarried/divorced/widowed | 233 (5.64) | 286 (5.89) | 225 (7.26) | 251 (8.09) | ||
| Educational status | ||||||
| No education | 747 (18.08) | 689 (14.20) | 393 (12.69) | 392 (12.64) | 66.462 | <0.001 |
| Elementary school | 1,484 (35.91) | 1,745 (35.96) | 1,103 (35.60) | 1,113 (35.89) | ||
| Middle school and above | 1,901 (46.01) | 2,419 (49.85) | 1,602 (51.71) | 1,596 (51.47) | ||
| No. of family members | ||||||
| 1−2 | 832 (20.14) | 1,064 (21.92) | 742 (23.95) | 701 (22.61) | 27.804 | 0.006 |
| 3 | 832 (20.14) | 931 (19.18) | 584 (18.85) | 531 (17.12) | ||
| 4 | 1,395 (33.76) | 1,647 (33.94) | 996 (32.15) | 1,051 (33.89) | ||
| 5 | 717 (17.35) | 795 (16.38) | 502 (16.20) | 520 (16.77) | ||
| ≥6 | 356 (8.62) | 416 (8.57) | 274 (8.84) | 298 (9.61) | ||
| Annual income (CNY/person) | ||||||
| ≤5,000 | 1,133 (27.42) | 1,569 (32.33) | 1,110 (35.83) | 1,114 (35.92) | 80.947 | <0.001 |
| >5,000 | 2,999 (72.58) | 3,284 (67.67) | 1,988 (64.17) | 1,987 (64.08) | ||
| BMI (kg/m2) | ||||||
| <18.5 | 144 (3.48) | 154 (3.17) | 117 (3.78) | 108 (3.48) | 20.280 | 0.016 |
| 18.5−23.9 | 2,234 (54.07) | 2,753 (56.73) | 1,779 (57.42) | 1,797 (57.95) | ||
| 24.0−27.9 | 1,372 (33.20) | 1,557 (32.08) | 971 (31.34) | 956 (30.83) | ||
| ≥28.0 | 382 (9.24) | 389 (8.02) | 231 (7.46) | 240 (7.74) | ||
| Smoking status | ||||||
| No | 3,343 (80.91) | 3,519 (72.51) | 2,179 (70.34) | 1,932 (62.30) | 313.378 | <0.001 |
| Yes | 789 (19.09) | 1,334 (27.49) | 919 (29.66) | 1,169 (37.70) | ||
| History of digestive diseases | ||||||
| No | 3,633 (87.92) | 4,330 (89.22) | 2,758 (89.03) | 2,679 (86.39) | 16.983 | <0.001 |
| Yes | 499 (12.08) | 523 (10.78) | 340 (10.97) | 422 (13.61) | ||
| Family history of cancer | ||||||
| No | 2,867 (69.39) | 3,467 (71.44) | 2,192 (70.76) | 2,179 (70.27) | 4.710 | 0.194 |
| Yes | 1,265 (30.61) | 1,386 (28.56) | 906 (29.24) | 922 (29.73) | ||
PDIs and risk of EC
During 219,365 person-years of follow-up, a total of 176 patients with EC were identified within the cohort. According to the fully adjusted model (Table 2), the PDI was modestly inversely associated with EC incidence, but the difference was not statistically significant (HR comparing extreme quartiles: 0.80, 95% CI: 0.50−1.30; HR per 10-unit increase: 0.84, 95% CI: 0.57−1.22; Ptrend=0.358). When we analyzed hPDI and uPDI separately, we found a stronger inverse association between hPDI and EC incidence (HR comparing extreme quartiles: 0.50, 95% CI: 0.32−0.77; HR per 10-unit increase: 0.42, 95% CI: 0.27−0.66; Ptrend<0.001) and a positive association for uPDI (HR comparing extreme quartiles: 1.80, 95% CI: 1.16−2.82; HR per 10-unit increase: 1.90, 95% CI: 1.26−2.88; Ptrend=0.002). No evidence of nonlinearity was found for hPDI or uPDI (Figure 2).
Table 2. Association between three plant-based diet indices and risks of EC.
| Varialbles | Sample (cases) | Not adjusted model | Minimally adjusted model | Fully adjusted model | |||||
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | ||||
| EC, esophageal cancer; PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index; HR, hazard ratio; 95% CI, 95% confidence interval. Not adjusted model, PDI, hPDI, and uPDI are included separately; Minimally adjusted model, adjusted age and gender; Fully adjusted model, adjusted for age, gender, marital status, educational status, number of family members, annual income, smoking status, history of digestive diseases, family history of cancer, and BMI categories. | |||||||||
| PDI by quartiles | |||||||||
| Quartile 1 | 3,796 (50) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,321 (51) | 0.91 (0.61−1.34) | 0.620 | 0.95 (0.64−1.41) | 0.807 | 0.96 (0.65−1.42) | 0.826 | ||
| Quartile 3 | 4,174 (49) | 0.90 (0.61−1.34) | 0.602 | 0.96 (0.64−1.42) | 0.819 | 0.97 (0.66−1.45) | 0.892 | ||
| Quartile 4 | 2,893 (26) | 0.69 (0.43−1.11) | 0.131 | 0.77 (0.48−1.24) | 0.286 | 0.80 (0.50−1.30) | 0.371 | ||
| Per 10 score increase | 15,184 (176) | 0.77 (0.55−1.08) | 0.123 | 0.80 (0.57−1.13) | 0.206 | 0.84 (0.57−1.22) | 0.358 | ||
| hPDI by quartiles | |||||||||
| Quartile 1 | 4,486 (78) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 3,663 (40) | 0.63 (0.43−0.92) | 0.016 | 0.67 (0.46−0.98) | 0.038 | 0.67 (0.46−0.99) | 0.042 | ||
| Quartile 3 | 3,268 (31) | 0.54 (0.36−0.83) | 0.004 | 0.60 (0.39−0.91) | 0.015 | 0.61 (0.40−0.92) | 0.020 | ||
| Quartile 4 | 3,767 (27) | 0.41 (0.27−0.64) | <0.001 | 0.48 (0.31−0.74) | 0.001 | 0.50 (0.32−0.77) | 0.002 | ||
| Per 10 score increase | 15,184 (176) | 0.38 (0.23−0.54) | <0.001 | 0.41 (0.26−0.63) | <0.001 | 0.42 (0.27−0.66) | <0.001 | ||
| uPDI by quartiles | |||||||||
| Quartile 1 | 4,132 (32) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,853 (46) | 1.21 (0.77−1.91) | 0.398 | 1.15 (0.73−1.80) | 0.551 | 1.12 (0.71−1.76) | 0.625 | ||
| Quartile 3 | 3,098 (46) | 1.90 (1.21−2.99) | 0.005 | 1.74 (1.10−2.73) | 0.017 | 1.67 (1.06−2.62) | 0.028 | ||
| Quartile 4 | 3,101 (52) | 2.13 (1.37−3.31) | <0.001 | 1.87 (1.20−2.92) | 0.005 | 1.80 (1.16−2.82) | 0.009 | ||
| Per 10 score increase | 15,184 (176) | 1.24 (1.50−3.35) | <0.001 | 1.97 (1.31−2.95) | 0.001 | 1.90 (1.26−2.88) | 0.002 | ||
Figure 2.
Cubic splines for three plant-based diet indices and risks of EC. (A) Cubic splines for PDI (Poverall=0.218, Pnonlinear=0.382); (B) Cubic splines for hPDI (Poverall<0.001, Pnonlinear=0.539); (C) Cubic splines for uPDI (Poverall<0.001, Pnonlinear=0.712). EC, esophageal cancer; PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index.
Subgroup analyses
Associations of PDI, hPDI, and uPDI with the risk of EC remained consistent across the subgroups (Supplementary Table S7). The results indicated that the directions of these associations were not significantly different among subgroups stratified by age at baseline, gender, marital status, educational status, number of family members, annual income, smoking status, BMI, history of digestive diseases or family history of cancer (all Pinteraction>0.05).
Table S7. Subgroup analyses for three plant-based diet indices and risks of EC.
| Variables | PDI | uPDI | hPDI | ||||||||
| HR (95% CI)* | P | Pinteraction | HR (95% CI)* | P | Pinteraction | HR (95% CI)* | P | Pinteraction | |||
| PDI, overall plant-based diet index; EC, esophageal cancer; uPDI, unhealthful plant-based diet index; hPDI, healthful plant-based diet index; HR, hazard ratio; 95% CI, 95% confidence interval; *, HR for quartiles of PDI, hPDI, and uPDI. The HR and 95% CI of Quartiles 2, 3, 4 were calculated with reference to Quartile 1. | |||||||||||
| Gender | |||||||||||
| Male | 1.00 | 1.00 | 1.00 | ||||||||
| 0.86 (0.54−1.38) | 0.528 | 1.21 (0.70−2.09) | 0.499 | 0.72 (0.46−1.15) | 0.170 | ||||||
| 0.91 (0.57−1.46) | 0.703 | 1.25 (0.70−2.24) | 0.452 | 0.63 (0.38−1.07) | 0.089 | ||||||
| 0.66 (0.36−1.20) | 0.173 | 1.71 (0.99−2.93) | 0.053 | 0.56 (0.32−0.97) | 0.040 | ||||||
| Female | 1.00 | 0.848 | 1.00 | 0.123 | 1.00 | 0.935 | |||||
| 1.19 (0.58−2.45) | 0.626 | 0.93 (0.41−2.12) | 0.863 | 0.57 (0.29−1.12) | 0.103 | ||||||
| 1.14 (0.54−2.38) | 0.730 | 2.63 (1.26−5.48) | 0.010 | 0.54 (0.27−1.08) | 0.082 | ||||||
| 1.15 (0.51−2.58) | 0.736 | 2.09 (0.94−4.64) | 0.070 | 0.39 (0.18−0.81) | 0.012 | ||||||
| Age at baseline (year) | |||||||||||
| 40−49 | 1.00 | 1.00 | |||||||||
| 0.82 (0.35−1.94) | 0.658 | − | − | 0.96 (0.45−2.05) | 0.926 | ||||||
| 0.59 (0.23−1.54) | 0.284 | − | − | 0.60 (0.23−1.54) | 0.289 | ||||||
| 1.00 (0.39−2.53) | 0.999 | − | − | 0.18 (0.04−0.79) | 0.023 | ||||||
| 50−59 | 1.00 | 0.397 | 1.00 | − | 1.00 | 0.188 | |||||
| 1.01 (0.59−1.73) | 0.966 | 1.01 (0.54−1.87) | 0.981 | 0.63 (0.38−1.04) | 0.073 | ||||||
| 1.33 (0.80−2.22) | 0.278 | 1.76 (0.96−3.20) | 0.067 | 0.57 (0.33−0.98) | 0.044 | ||||||
| 0.84 (0.44−1.63) | 0.614 | 2.11 (1.19−3.77) | 0.011 | 0.45 (0.25−0.81) | 0.007 | ||||||
| 60−69 | 1.00 | 1.00 | 1.00 | ||||||||
| 0.99 (0.46−2.15) | 0.980 | 0.49 (0.20−1.20) | 0.116 | 0.52 (0.20−1.35) | 0.181 | ||||||
| 0.60 (0.25−1.45) | 0.253 | 1.00 (0.44−2.28) | 0.994 | 0.74 (0.30−1.83) | 0.519 | ||||||
| 0.50 (0.16−1.56) | 0.234 | 0.58 (0.22−1.48) | 0.254 | 1.04 (0.46−2.34) | 0.930 | ||||||
| Marital status | |||||||||||
| Married | 1.00 | 1.00 | 1.00 | ||||||||
| 1.04 (0.69−1.58) | 0.836 | 1.22 (0.76−1.97) | 0.415 | 0.69 (0.46−1.04) | 0.073 | ||||||
| 1.01 (0.66−1.54) | 0.964 | 1.85 (1.15−2.99) | 0.012 | 0.62 (0.40−0.95) | 0.030 | ||||||
| Unmarried/divorced/ widowed |
0.90 (0.55−1.47) | 0.673 | 1.94 (1.21−3.12) | 0.006 | 0.48 (0.30−0.76) | 0.002 | |||||
| − | − | − | − | − | − | − | − | − | |||
| − | − | − | − | − | − | ||||||
| − | − | − | − | − | − | ||||||
| Educational status | |||||||||||
| Elementary school and below |
1.00 | 1.00 | 1.00 | ||||||||
| 0.90 (0.50−1.62) | 0.721 | 1.09 (0.59−1.99) | 0.784 | 0.60 (0.32−1.10) | 0.098 | ||||||
| 1.00 (0.56−1.78) | 0.994 | 1.48 (0.80−2.75) | 0.215 | 0.77 (0.43−1.39) | 0.386 | ||||||
| 0.77 (0.38−1.56) | 0.473 | 0.98 (0.50−1.92) | 0.952 | 0.62 (0.34−1.13) | 0.118 | ||||||
| Middle school and above | 1.00 | 0.965 | 1.00 | 0.093 | 1.00 | 0.435 | |||||
| 0.98 (0.58−1.66) | 0.935 | 1.22 (0.62−2.43) | 0.564 | 0.74 (0.45−1.20) | 0.221 | ||||||
| 0.91 (0.53−1.58) | 0.743 | 1.97 (1.00−3.89) | 0.049 | 0.49 (0.27−0.89) | 0.020 | ||||||
| 0.80 (0.42−1.54) | 0.509 | 2.95 (1.56−5.60) | 0.001 | 0.40 (0.21−0.78) | 0.007 | ||||||
| No. of family members | |||||||||||
| 1−3 | 1.00 | 1.00 | 1.00 | ||||||||
| 1.06 (0.60−1.90) | 0.835 | 0.94 (0.51−1.75) | 0.853 | 0.77 (0.43−1.39) | 0.389 | ||||||
| 1.06 (0.58−1.92) | 0.854 | 1.24 (0.66−2.33) | 0.507 | 0.81 (0.44−1.48) | 0.489 | ||||||
| 0.91 (0.45−1.84) | 0.785 | 1.09 (0.56−2.11) | 0.793 | 0.82 (0.44−1.51) | 0.519 | ||||||
| 4−6 | 1.00 | 0.970 | 1.00 | 0.263 | 1.00 | 0.197 | |||||
| 0.88 (0.52−1.50) | 0.639 | 1.36 (0.70−2.65) | 0.365 | 0.63 (0.38−1.04) | 0.071 | ||||||
| 0.92 (0.54−1.57) | 0.773 | 2.32 (1.20−4.49) | 0.012 | 0.48 (0.27−0.85) | 0.013 | ||||||
| 0.72 (0.37−1.37) | 0.317 | 2.73 (1.45−5.14) | 0.002 | 0.31 (0.16−0.60) | 0.001 | ||||||
| Annual income (CNY/person) | |||||||||||
| ≤5,000 | 1.00 | 1.00 | 1.00 | ||||||||
| 1.06 (0.58−1.94) | 0.859 | 0.77 (0.38−1.54) | 0.462 | 0.59 (0.31−1.11) | 0.101 | ||||||
| 0.99 (0.53−1.85) | 0.985 | 0.86 (0.41−1.78) | 0.676 | 0.62 (0.32−1.22) | 0.168 | ||||||
| 0.75 (0.33−1.70) | 0.490 | 1.29 (0.67−2.49) | 0.452 | 0.63 (0.32−1.23) | 0.176 | ||||||
| >5,000 | 1.00 | 0.912 | 1.00 | 0.189 | 1.00 | 0.746 | |||||
| 0.87 (0.52−1.45) | 0.587 | 1.44 (0.79−2.62) | 0.239 | 0.73 (0.45−1.18) | 0.203 | ||||||
| 0.92 (0.55−1.54) | 0.762 | 2.50 (1.38−4.53) | 0.002 | 0.61 (0.36−1.05) | 0.075 | ||||||
| 0.80 (0.44−1.45) | 0.460 | 2.18 (1.19−3.99) | 0.012 | 0.43 (0.24−0.78) | 0.005 | ||||||
| BMI (kg/m2) | |||||||||||
| <23.9 | 1.00 | 1.00 | 1.00 | ||||||||
| 0.97 (0.61−1.56) | 0.906 | 1.19 (0.70−2.01) | 0.526 | 0.77 (0.48−1.24) | 0.275 | ||||||
| 0.89 (0.54−1.45) | 0.631 | 1.30 (0.74−2.28) | 0.357 | 0.84 (0.52−1.38) | 0.499 | ||||||
| 0.85 (0.48−1.50) | 0.581 | 1.49 (0.86−2.56) | 0.153 | 0.63 (0.37−1.08) | 0.090 | ||||||
| ≥24 | 1.00 | 0.808 | 1.00 | 0.082 | 1.00 | 0.098 | |||||
| 0.90 (0.45−1.81) | 0.770 | 0.89 (0.36−2.19) | 0.793 | 0.54 (0.28−1.03) | 0.063 | ||||||
| 1.15 (0.59−2.25) | 0.673 | 2.61 (1.17−5.80) | 0.019 | 0.27 (0.11−0.65) | 0.004 | ||||||
| 0.70 (0.28−1.70) | 0.426 | 2.58 (1.16−5.72) | 0.020 | 0.33 (0.15−0.72) | 0.005 | ||||||
| Smoking status | |||||||||||
| Yes | 1.00 | 1.00 | 1.00 | ||||||||
| 0.86 (0.48−1.52) | 0.597 | 1.52 (0.70−3.31) | 0.289 | 0.86 (0.49−1.50) | 0.596 | ||||||
| 0.66 (0.35−1.26) | 0.208 | 1.42 (0.62−3.25) | 0.408 | 0.53 (0.26−1.10) | 0.090 | ||||||
| 0.74 (0.34−1.58) | 0.433 | 1.91 (0.89−4.11) | 0.098 | 0.33 (0.13−0.83) | 0.019 | ||||||
| No | 1.00 | 0.562 | 1.00 | 0.323 | 1.00 | 0.354 | |||||
| 1.06 (0.62−1.81) | 0.831 | 0.90 (0.51−1.60) | 0.718 | 0.57 (0.33−0.96) | 0.035 | ||||||
| 1.23 (0.73−2.07) | 0.428 | 1.84 (1.07−3.16) | 0.028 | 0.64 (0.38−1.08) | 0.094 | ||||||
| 0.89 (0.48−1.65) | 0.704 | 1.76 (1.01−3.09) | 0.047 | 0.57 (0.34−0.96) | 0.034 | ||||||
| History of digestive diseases | |||||||||||
| Yes | 1.00 | 1.00 | 1.00 | ||||||||
| 0.70 (0.25−1.99) | 0.506 | 0.46 (0.11−1.87) | 0.277 | 0.17 (0.02−1.34) | 0.093 | ||||||
| 0.63 (0.21−1.89) | 0.407 | 0.60 (0.15−2.49) | 0.486 | 0.95 (0.33−2.75) | 0.925 | ||||||
| 0.32 (0.07−1.52) | 0.152 | 1.54 (0.53−4.51) | 0.427 | 0.67 (0.21−2.13) | 0.493 | ||||||
| No | 1.00 | 0.578 | 1.00 | 0.207 | 1.00 | 0.194 | |||||
| 1.01 (0.66−1.54) | 0.962 | 1.26 (0.77−2.05) | 0.355 | 0.73 (0.49−1.09) | 0.125 | ||||||
| 1.05 (0.68−1.60) | 0.839 | 1.88 (1.15−3.07) | 0.012 | 0.58 (0.36−0.91) | 0.018 | ||||||
| 0.92 (0.55−1.52) | 0.740 | 1.82 (1.11−2.98) | 0.018 | 0.48 (0.30−0.77) | 0.003 | ||||||
| Family history of cancer | |||||||||||
| Yes | 1.00 | 1.00 | 1.00 | ||||||||
| 1.16 (0.55−2.45) | 0.700 | 0.99 (0.43−2.31) | 0.989 | 0.59 (0.29−1.18) | 0.137 | ||||||
| 1.41 (0.68−2.92) | 0.362 | 1.90 (0.86−4.22) | 0.114 | 0.66 (0.33−1.33) | 0.245 | ||||||
| 1.52 (0.68−3.40) | 0.310 | 2.16 (1.00−4.67) | 0.049 | 0.44 (0.20−0.97) | 0.042 | ||||||
| No | 1.00 | 0.214 | 1.00 | 0.690 | 1.00 | 0.916 | |||||
| 0.90 (0.57−1.42) | 0.649 | 1.17 (0.69−2.01) | 0.558 | 0.73 (0.46−1.16) | 0.180 | ||||||
| 0.82 (0.51−1.32) | 0.421 | 1.56 (0.90−2.72) | 0.114 | 0.60 (0.35−1.00) | 0.052 | ||||||
| 0.57 (0.31−1.05) | 0.071 | 1.61 (0.93−2.80) | 0.087 | 0.53 (0.31−0.91) | 0.022 | ||||||
Sensitivity analyses
According to the sensitivity analyses, when we excluded individuals who developed EC within one year of follow-up, the results were the same as those of the main analysis, and we found an inverse association between hPDI and EC (HR comparing extreme quartiles: 0.50, 95% CI: 0.32−0.79) and a positive association between uPDI and EC (HR comparing extreme quartiles: 1.73, 95% CI: 1.09−2.75) (Supplementary Table S8). We further excluded individuals who had diabetes, stroke, or myocardial infarction, and the results also remained robust (Supplementary Table S9). This inverse association was also observed for hDPI (HR comparing extreme quartiles: 0.50, 95% CI: 0.32−0.77). Conversely, uPDI was positively associated with EC (HR comparing extreme quartiles: 1.81, 95% CI: 1.16−2.82). We also excluded individuals who were diagnosed with esophageal adenocarcinoma (EAC) or other histological subtypes (Supplementary Table S10). We observed an inverse association for hDPI (HR comparing extreme quartiles: 0.50, 95% CI: 0.32−0.79) and a positive association for uPDI with ESCC (HR comparing extreme quartiles: 1.79, 95% CI: 1.14−2.82). To account for the effect of alcoholic beverages, alcoholic beverages were excluded from the PDIs, and the results were robust (Supplementary Table S11). hPDI was inversely associated with EC incidence (HR comparing extreme quartiles: 0.55, 95% CI: 0.37−0.82). uPDI was positively associated with EC incidence (HR comparing extreme quartiles: 1.65, 95% CI: 1.07−2.54).
Table S8. Sensitivity analysisa of association between three plant-based diet indices and risks of EC.
| Variables | Sample (cases) | Not adjusted model | Minimally adjusted model | Fully adjusted model | |||||
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | ||||
| a, Excluded individuals who developed EC within one year of follow-up; EC, esophageal cancer; Not adjusted model, PDI, hPDI, and uPDI are included separately; Minimally adjusted model, adjusted age and gender; Fully adjusted model, adjusted for age, gender, marital status, educational status, number of family members, annual income, smoking status, history of digestive diseases, and family history of cancer, and BMI categories. PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index; HR, hazard ratio; 95% CI, 95% confidence interval. | |||||||||
| PDI by quartiles | |||||||||
| Quartile 1 | 3,790 (44) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,320 (50) | 1.01 (0.67−1.52) | 0.961 | 1.06 (0.71−1.59) | 0.771 | 1.06 (0.71−1.60) | 0.765 | ||
| Quartile 3 | 4,172 (47) | 0.98 (0.65−1.48) | 0.932 | 1.04 (0.69−1.57) | 0.847 | 1.06 (0.70−1.61) | 0.773 | ||
| Quartile 4 | 2,892 (25) | 0.76 (0.46−1.24) | 0.271 | 0.84 (0.52−1.38) | 0.502 | 0.89 (0.54−1.45) | 0.635 | ||
| Per 10 score increase | 15,174 (166) | 0.77 (0.53−1.13) | 0.184 | 0.84 (0.57−1.24) | 0.377 | 0.88 (0.59−1.30) | 0.520 | ||
| hPDI by quartiles | |||||||||
| Quartile 1 | 4,480 (72) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 3,662 (39) | 0.66 (0.45−0.98) | 0.038 | 0.70 (0.48−1.04) | 0.078 | 0.71 (0.48−1.05) | 0.087 | ||
| Quartile 3 | 3,267 (30) | 0.57 (0.37−0.87) | 0.010 | 0.62 (0.41−0.96) | 0.030 | 0.64 (0.41−0.98) | 0.040 | ||
| Quartile 4 | 3,765 (25) | 0.42 (0.26−0.65) | 0.000 | 0.48 (0.30−0.76) | 0.002 | 0.50 (0.32−0.79) | 0.003 | ||
| Per 10 score increase | 15,174 (166) | 0.37 (0.24−0.58) | <0.001 | 0.44 (0.28−0.68) | <0.001 | 0.45 (0.29−0.72) | 0.001 | ||
| uPDI by quartiles | |||||||||
| Quartile 1 | 4,130 (30) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,851 (44) | 1.24 (0.78−1.97) | 0.366 | 1.17 (0.74−1.86) | 0.508 | 1.13 (0.71−1.80) | 0.603 | ||
| Quartile 3 | 3,097 (45) | 1.98 (1.25−3.15) | 0.004 | 1.81 (1.14−2.88) | 0.012 | 1.72 (1.08−2.75) | 0.022 | ||
| Quartile 4 | 3,196 (47) | 2.05 (1.30−3.24) | 0.002 | 1.81 (1.14−2.86) | 0.012 | 1.73 (1.09−2.75) | 0.021 | ||
| Per 10 score increase | 15,174 (166) | 2.20 (1.44−3.32) | <0.001 | 1.93 (1.27−2.93) | 0.002 | 1.85 (1.21−2.84) | 0.005 | ||
Table S9. Sensitivity analysisb of association between three plant-based diet indices and risks of EC.
| Variables | Sample (cases) | Not adjusted model | Minimally adjusted model | Fully adjusted model | |||||
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | ||||
| b, Excluded individuals who had diabetes, stroke, or myocardial infarction, to account for the possibility that their usual diet changed due to the diagnosis of these chronic diseases; EC, esophageal cancer; Not adjusted model, PDI, hPDI, and uPDI are included separately; Minimally adjusted model, adjusted age and gender; Fully adjusted model, adjusted for age, gender, marital status, educational status, number of family members, annual income, smoking status, history of digestive diseases, and family history of cancer, and BMI categories; PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index; HR, hazard ratio; 95% CI, 95% confidence interval. | |||||||||
| PDI by quartiles | |||||||||
| Quartile 1 | 3,757 (50) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,278 (51) | 0.91 (0.61−1.34) | 0.621 | 0.95 (0.64−1.41) | 0.807 | 0.96 (0.65−1.42) | 0.827 | ||
| Quartile 3 | 4,136 (39) | 0.90 (0.61−1.34) | 0.600 | 0.95 (0.64−1.42) | 0.817 | 0.97 (0.65−1.44) | 0.890 | ||
| Quartile 4 | 2,862 (26) | 0.69 (0.43−1.12) | 0.131 | 0.77 (0.48−1.24) | 0.286 | 0.80 (0.50−1.30) | 0.373 | ||
| Per 10 score increase | 15,033 (176) | 0.74 (0.51−1.07) | 0.111 | 0.81 (0.55−1.17) | 0.257 | 0.84 (0.57−1.22) | 0.358 | ||
| hPDI by quartiles | |||||||||
| Quartile 1 | 4,438 (78) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 3,620 (40) | 0.63 (0.43−0.92) | 0.016 | 0.67 (0.46−0.98) | 0.038 | 0.67 (0.46−0.99) | 0.043 | ||
| Quartile 3 | 3,237 (31) | 0.54 (0.36−0.83) | 0.004 | 0.60 (0.39−0.91) | 0.015 | 0.61 (0.40−0.92) | 0.020 | ||
| Quartile 4 | 3,738 (27) | 0.41 (0.27−0.64) | 0.000 | 0.48 (0.31−0.74) | 0.001 | 0.50 (0.32−0.77) | 0.002 | ||
| Per 10 score increase | 15,033 (176) | 0.35 (0.23−0.54) | <0.001 | 0.41 (0.26−0.63) | <0.001 | 0.42 (0.27−0.66) | <0.001 | ||
| uPDI by quartiles | |||||||||
| Quartile 1 | 4,109 (32) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,796 (46) | 1.22 (0.77−1.91) | 0.395 | 1.15 (0.73−1.80) | 0.548 | 1.12 (0.71−1.76) | 0.623 | ||
| Quartile 3 | 3,063 (46) | 1.90 (1.21−2.99) | 0.005 | 1.74 (1.11−2.73) | 0.017 | 1.67 (1.06−2.62) | 0.028 | ||
| Quartile 4 | 3,065 (52) | 2.13 (1.37−3.31) | 0.001 | 1.88 (1.21−2.92) | 0.005 | 1.81 (1.16−2.82) | 0.009 | ||
| Per 10 score increase | 15,033 (176) | 2.24 (1.50−3.36) | <0.001 | 1.97 (1.31−2.96) | 0.001 | 1.90 (1.26−2.88) | 0.002 | ||
Table S10. Sensitivity analysisc of association between three plant-based diet indices and risks of EC.
| Variables | Sample (cases) | Not adjusted model | Minimally adjusted model | Fully adjusted model | |||||
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | ||||
| c, Excluded individuals who were diagnosed as EAC or other histological subtypes; EC, esophageal cancer; We explored the associations for PDI, hPDI, and uPDI with ESCC; Not adjusted model, PDI, hPDI, and uPDI are included separately; Minimally adjusted model, adjusted age and gender; Fully adjusted model, adjusted for age, gender, marital status, educational status, number of family members, annual income, smoking status, history of digestive diseases, and family history of cancer, and BMI categories; PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index; HR, hazard ratio; 95% CI, 95% confidence interval; EAC, esophageal adenocarcinoma. | |||||||||
| PDI by quartiles | |||||||||
| Quartile 1 | 3,795 (49) | Ref | - | Ref | - | Ref | - | ||
| Quartile 2 | 4,318 (48) | 0.87 (0.58−1.30) | 0.495 | 0.92 (0.61−1.36) | 0.664 | 1.06 (0.66−1.68) | 0.816 | ||
| Quartile 3 | 4,171 (46) | 0.86 (0.58−1.29) | 0.474 | 0.92 (0.61−1.37) | 0.668 | 1.65 (1.04−2.62) | 0.034 | ||
| Quartile 4 | 2,891 (24) | 0.65 (0.40−1.07) | 0.088 | 0.73 (0.45−1.19) | 0.203 | 1.79 (1.14−2.82) | 0.012 | ||
| Per 10 score increase | 15,175 (167) | 0.70 (0.48−1.03) | 0.072 | 0.77 (0.52−1.13) | 0.177 | 0.80 (0.54−1.17) | 0.247 | ||
| hPDI by quartiles | |||||||||
| Quartile 1 | 4,482 (74) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 3,661 (38) | 0.63 (0.42−0.93) | 0.020 | 0.67 (0.45−0.99) | 0.044 | 0.67 (0.46−1.00) | 0.050 | ||
| Quartile 3 | 3,266 (29) | 0.54 (0.35−0.82) | 0.005 | 0.59 (0.38−0.91) | 0.016 | 0.60 (0.39−0.92) | 0.020 | ||
| Quartile 4 | 3,766 (26) | 0.42 (0.27−0.66) | 0.000 | 0.49 (0.31−0.76) | 0.002 | 0.50 (0.32−0.79) | 0.003 | ||
| Per 10 score increase | 15,175 (167) | 0.35 (0.22−0.54) | <0.001 | 0.41 (0.26−0.64) | <0.001 | 0.42 (0.27−0.67) | <0.001 | ||
| uPDI by quartiles | |||||||||
| Quartile 1 | 4,131 (31) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,849 (42) | 1.14 (0.72−1.82) | 0.568 | 1.08 (0.68−1.72) | 0.741 | 1.06 (0.66−1.68) | 0.816 | ||
| Quartile 3 | 3,096 (44) | 1.88 (1.19−2.97) | 0.007 | 1.71 (1.08−2.72) | 0.022 | 1.65 (1.04−2.62) | 0.034 | ||
| Quartile 4 | 3,099 (50) | 2.11 (1.35−3.31) | 0.001 | 1.86 (1.19−2.91) | 0.007 | 1.79 (1.14−2.82) | 0.012 | ||
| Per 10 score increase | 15,175 (167) | 2.26 (1.50−3.42) | <0.001 | 1.98 (1.31−3.01) | 0.001 | 1.92 (1.26−2.93) | 0.003 | ||
Table S11. Sensitivity analysisd of association between three plant-based diet indices and risks of EC.
| Variables | Sample (cases) | Not adjusted model | Minimally adjusted model | Fully adjusted model | |||||
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | ||||
| d, Excluded alcoholic beverage for plant-based dietary patterns and adjusted it in the model; EC, esophageal cancer; Not adjusted model, PDI, hPDI, and uPDI are included separately; Minimally adjusted model, adjusted age and gender; Fully adjusted model, adjusted for age, gender, marital status, educational status, number of family members, annual income, smoking status, alcohol intake, history of digestive diseases, and family history of cancer, and BMI categories; PDI, overall plant-based diet index; hPDI, healthful plant-based diet index; uPDI, unhealthful plant-based diet index; HR, hazard ratio; 95% CI, 95% confidence interval. | |||||||||
| PDI by quartiles | |||||||||
| Quartile 1 | 3,098 (32) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 4,023 (43) | 1.05 (0.67−1.66) | 0.828 | 1.06 (0.67−1.68) | 0.789 | 1.02 (0.65−1.62) | 0.920 | ||
| Quartile 3 | 4,372 (60) | 1.35 (0.88−2.08) | 0.170 | 1.33 (0.87−2.05) | 0.189 | 1.27 (0.82−1.95) | 0.287 | ||
| Quartile 4 | 3,691 (41) | 1.10 (0.69−1.75) | 0.687 | 1.06 (0.67−1.69) | 0.800 | 0.96 (0.59−1.55) | 0.858 | ||
| Per 10 score increase | 15,184 (176) | 1.03 (0.71−1.49) | 0.885 | 0.98 (0.68−1.43) | 0.931 | 0.90 (0.61−1.32) | 0.597 | ||
| hPDI by quartiles | |||||||||
| Quartile 1 | 3,474 (56) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 3,481 (35) | 0.63 (0.41−0.95) | 0.030 | 0.64 (0.42−0.97) | 0.036 | 0.64 (0.42−0.97) | 0.036 | ||
| Quartile 3 | 3,579 (42) | 0.73 (0.49−1.09) | 0.127 | 0.72 (0.49−1.08) | 0.115 | 0.70 (0.47−1.05) | 0.085 | ||
| Quartile 4 | 4,650 (43) | 0.58 (0.39−0.86) | 0.007 | 0.58 (0.39−0.86) | 0.007 | 0.55 (0.37−0.82) | 0.004 | ||
| Per 10 score increase | 15,184 (176) | 0.51 (0.32−0.81) | 0.005 | 0.51 (0.32−0.81) | 0.004 | 0.47 (0.29−0.75) | 0.002 | ||
| uPDI by quartiles | |||||||||
| Quartile 1 | 3,466 (36) | Ref | − | Ref | − | Ref | − | ||
| Quartile 2 | 3,239 (30) | 0.89 (0.55−1.44) | 0.625 | 0.89 (0.55−1.44) | 0.631 | 0.93 (0.57−1.52) | 0.781 | ||
| Quartile 3 | 4,861 (54) | 1.06 (0.69−1.61) | 0.800 | 1.08 (0.71−1.65) | 0.711 | 1.16 (0.76−1.78) | 0.495 | ||
| Quartile 4 | 3,618 (56) | 1.46 (0.96−2.22) | 0.077 | 1.57 (1.03−2.38) | 0.062 | 1.65 (1.07−2.54) | 0.024 | ||
| Per 10 score increase | 15,184 (176) | 1.54 (1.01−2.33) | 0.044 | 1.26 (0.88−1.79) | 0.034 | 1.74 (1.13−2.66) | 0.011 | ||
Individual food groups of plant-based diets and risk of EC
Correlation coefficients among different food groups ranged from −0.45 to 0.43 (Supplementary Figure S1). Supplementary Table S12 shows the results of multivariable Cox regression analyses of the associations between individual food groups of plant-based diets and the risk of EC. The analyses of individual food groups showed that higher intakes of fresh vegetable (HR: 0.68, 95% CI: 0.49−0.93) and legumes (HR: 0.82, 95% CI: 0.70−0.97) were associated with a lower risk of EC. In contrast, a higher intake of alcoholic beverages was associated with a greater risk of EC (HR: 1.49, 95% CI: 1.00−2.22).
Figure S1.
Spearman correlation analysis among food groups.
Table S12. HR (95% CI) of individual food group with risk of EC.
| Food group | Frequency | HR | 95% CI | P |
| Adjusted for age, gender, marital status, educational status, number of family members, annual income, smoking status, history of digestive diseases, family history of cancer, and BMI categories. HR, hazard ratio; 95% CI, 95% confidence interval; EC, esophageal cancer. | ||||
| Fresh vegetable | <1 time/d | 1 | ||
| Daily | 0.68 | 0.49−0.93 | 0.015 | |
| Fresh fruit | <1 time/d | 1 | ||
| Daily | 0.83 | 0.50−1.40 | 0.492 | |
| Meat/egg/dairy | <1 time/d | 1 | ||
| Daily | 0.54 | 0.24−1.22 | 0.139 | |
| Legumes | <1 time/d | 1 | ||
| Daily | 0.82 | 0.70−0.97 | 0.017 | |
| Spring onion, garlic |
<1 time/d | 1 | ||
| Daily | 1.52 | 0.65−3.27 | 0.275 | |
| Staple food | <1 time/d | 1 | ||
| Daily | 0.43 | 0.11−1.73 | 0.234 | |
| Dried fruit and vegetable |
<1 time/d | 1 | ||
| Daily | 0.92 | 0.67−1.25 | 0.592 | |
| Pickled vegetable | <1 time/d | 1 | ||
| Daily | 1.26 | 0.88−1.81 | 0.206 | |
| Fried food | <1 time/d | 1 | ||
| Daily | 0.94 | 0.55−1.61 | 0.831 | |
| Hot food | <1 time/d | 1 | ||
| Daily | 1.26 | 0.92−1.72 | 0.151 | |
| Mildewed food | <1 time/d | 1 | ||
| Daily | 1.27 | 0.18−9.08 | 0.812 | |
| Tea | No | 1 | ||
| Yes | 0.86 | 0.60−1.21 | 0.381 | |
| Alcoholic beverage | No | 1 | ||
| Yes | 1.49 | 1.00−2.22 | 0.047 | |
Discussion
In this prospective cohort study involving Chinese males and females, a healthy plant-based diet captured by hPDI was inversely associated with the risk of EC, particularly ESCC. Conversely, an unhealthful plant-based diet, indicated by the uPDI, was found to be associated with a greater risk of EC. The associations were consistent across the age at baseline, gender, marital status, educational status, number of family members, annual income, smoking status, BMI, history of digestive diseases and family history of cancer. These findings underscore the significance of the healthiness and quality of plant-based diets in the development of EC, particularly ESCC. Notably, this is the first comprehensive study to investigate the association between PDI or hPDI or uPDI and EC, focusing on a Chinese cohort.
Numerous studies have explored the connection between plant-based diets and EC, but the results have been inconclusive, potentially due to the omission of considerations regarding the healthiness and quality of plant-based foods. In a previous study conducted in China, the vegetable and fruit pattern was shown to be associated with a reduced risk of ESCC (26). However, diet was not stratified by the health status of the plant foods. Similarly, other healthy dietary patterns, such as Mediterranean diets and Healthy Eating Indices, have been linked to a lower risk of EC. A higher adherence to the Mediterranean diet was associated with a reduced risk of ESCC in European cohorts (19). The Healthy Eating Index-2005 was inversely associated with the risk of EC in the U.S. National Institutes of Health (NIH)-AARP Diet and Health Study (27). Moreover, a recent meta-analysis reported inverse associations between various types of plant-based diets and EC, although these associations did not account for the quality or healthiness of plant-based foods (28).
A recent US population-based prospective cohort study showed that hPDI was associated with a reduced risk of EC, but this difference was not statistically significant (16). In contrast, our analysis revealed a significant inverse association between hPDI and EC, especially ESCC. Differences in the distribution of tumor subtypes across populations may partly explain these differentials (29). ESCC and EAC are the most common histological subtypes of EC. The distributions of ESCC and EAC differ, with EAC incidence rates being the highest in Northern Europe and North America and ESCC rates being the highest in Eastern and South-Central Asia and South Africa. In addition, our PDIs mainly included the dietary characteristics of the high-risk sites, which may be a more important reason for the inconsistency between results of this study and those of previous studies. Prior research has modified the PDIs framework by incorporating local Chinese dietary attributes (14,15).
In the 21st century, the typical rural Chinese diet predominantly comprises plant-based foods (30,31). In this context, fried foods primarily consist of items such as pancakes and fried dough sticks, while hot foods mainly encompass hot soups and noodles. Mildewed foods are often related to mouldy grains (32,33). Since earlier research has associated these foods with an increased risk of EC, they are categorized as components of an unhealthy plant-based diet group (20). The definition of staple food was based on the dietary characteristics of the two high-risk areas of China. During the survey period, the traditional staple foods in Linzhou and Feicheng were mixed grains, mainly consisting of corn and wheat (34). Early studies also noted that a high whole-grain diet is a dietary characteristic of residents in high-risk areas of EC in China (31). Based on these considerations, staple foods are included in healthy plant foods. The consumption of meat in the questionnaire referred to local common meats (pork, beef, rabbit, chicken, or duck) (35). Because the consumption of fish in Linxian and Feicheng, inland regions, is low, fish were not investigated as separate foods.
The observed associations of hPDI and uPDI with the risk of EC may be attributed to the individual food groups. According to our component analyses, consumption of vegetables and legumes exhibited an inverse association with the risk of EC. Vegetables are rich sources of dietary fibre, phytochemicals (such as flavonoids), and vitamins (such as vitamin C) and are known for their anti-inflammatory and antioxidant abilities. These bioactive compounds can mitigate chronic inflammation and DNA oxidative damage, both of which have been implicated in promoting cancer initiation and progression (36,37). The analysis of food components revealed that the consumption of legumes was linked to a reduced risk of EC, which aligns with findings from previous research. Since legumes exhibit very high trypsin inhibitory activity, these protease inhibitors strongly drive esophageal mucosa proliferation via epidermal growth factor and transforming growth factor-α, which are usually degraded in the presence of proteases (38). These constituents are known to exert anticancer effects through the reduction of oxidative stress and damage, as well as modulation of epigenetic aberrations. Taken together, these findings suggest that healthy plant-based diets might improve the health of the esophagus and contribute to the prevention of developing EC. Alcoholic beverages showed a weak positive association with the risk of EC in the component analyses. This is related to the production of reactive oxygen species by enzymes induced by the alcohol metabolite acetaldehyde and chronic alcohol consumption (39). Therefore, the healthiness and quality of plant foods should be emphasized in dietary recommendations for the general population.
The strengths of this study include the use of prospective cohorts, a large sample size, and up to 17 years of follow-up. Additionally, this is the first study to evaluate the relationship of PDIs with EC and ESCC incidence, focusing on a Chinese cohort. This study has several limitations. First, it would be difficult to compare the PDIs in this study with the original PDIs. The food classification for the calculation of the PDIs score is based on the dietary habits of high-risk areas for EC in China. In addition, we used intake frequency rather than servings or quintiles of intake per day to score the diets. Several studies have provided evidence supporting the reliability and validity of utilizing nonquantitative food frequency questionnaires for assessing dietary patterns (21-23). However, it is important to note that the food questionnaire used in our study might not be standardized to the extent required for direct comparisons with other nutritional studies. We were also unable to estimate or adjust for total energy intake. Second, dietary information was collected only at the baseline, and individuals may have altered their dietary habits during the 17-year follow-up period. Therefore, our results might be subject to nondifferential bias. Nonetheless, it has been revealed that the approach using only baseline information leads to a weaker association than the approach using cumulative averages (40). Third, although information on BMI and tobacco smoking status was available, data on physical activity and other lifestyle factors were lacking. Finally, the compliance rate of endoscopic screening was 49.36%, which may have resulted in selection bias. However, it is difficult to persuade asymptomatic residents to participate in an invasive screening program. In a population-based cohort, a compliance rate of 49.36% was already a tough goal (41).
Conclusions
This study established a significant association between a healthy plant-based dietary pattern and a reduced risk of EC, particularly ESCC. Conversely, an unhealthy plant-based dietary pattern was associated with a greater risk of EC, including ESCC. These findings emphasize the importance of improving the quality of plant-based diets for EC prevention.
Acknowledgements
This study was supported by grants from the Beijing Nova Program (No. Z201100006820069); CAMS Innovation Fund for Medical Sciences (CIFMS, No. 2021-I2M-1-023, 2021-I2M-1-010) and Talent Incentive Program of Cancer Hospital Chinese Academy of Medical Sciences (Hope Star).
Contributor Information
Fen Liu, Email: liufen05@ccmu.edu.cn.
Shaoming Wang, Email: wangshaoming@cicams.ac.cn.
Wenqiang Wei, Email: weiwq@cicams.ac.cn.
References
- 1.Morgan E, Soerjomataram I, Rumgay H, et al The global landscape of esophageal squamous cell carcinoma and esophageal adenocarcinoma incidence and mortality in 2020 and projections to 2040: new estimates from GLOBOCAN 2020. Gastroenterology. 2022;163:649–58.e2. doi: 10.1053/j.gastro.2022.05.054. [DOI] [PubMed] [Google Scholar]
- 2.Ferlay JEM, Lam F, Colombet M, et al. Global Cancer Observatory: Cancer Today. Lyon, France: International Agency for Research on Cancer. Available online: https://gco.iarc.fr/today
- 3.Dinu M, Abbate R, Gensini GF, et al Vegetarian, vegan diets and multiple health outcomes: A systematic review with meta-analysis of observational studies. Crit Rev Food Sci Nutr. 2017;57:3640–9. doi: 10.1080/10408398.2016.1138447. [DOI] [PubMed] [Google Scholar]
- 4.Key TJ, Papier K, Tong TYN Plant-based diets and long-term health: findings from the EPIC-Oxford study. Proc Nutr Soc. 2022;81:190–8. doi: 10.1017/S0029665121003748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sun L, Zhang Z, Xu J, et al Dietary fiber intake reduces risk for Barrett’s esophagus and esophageal cancer. Crit Rev Food Sci Nutr. 2017;57:2749–57. doi: 10.1080/10408398.2015.1067596. [DOI] [PubMed] [Google Scholar]
- 6.Gaesser GA Whole grains, refined grains, and cancer risk: A systematic review of meta-analyses of observational studies. Nutrients. 2020;12:3756. doi: 10.3390/nu12123756. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.National Health Commission of the People’s Republic of China National guidelines for diagnosis and treatment of esophageal carcinoma 2022 in China (English version) Chin J Cancer Res. 2022;34:309–34. doi: 10.21147/j.issn.1000-9604.2022.04.01. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Satija A, Bhupathiraju SN, Rimm EB, et al Plant-based dietary patterns and incidence of type 2 diabetes in US men and women: Results from three prospective cohort studies. PLoS Med. 2016;13:e1002039. doi: 10.1371/journal.pmed.1002039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Satija A, Bhupathiraju SN, Spiegelman D, et al Healthful and unhealthful plant-based diets and the risk of coronary heart disease in U. S. adults. J Am Coll Cardiol. 2017;70:411–22. doi: 10.1016/j.jacc.2017.05.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Heianza Y, Zhou T, Sun D, et al Healthful plant-based dietary patterns, genetic risk of obesity, and cardiovascular risk in the UK biobank study. Clin Nutr. 2021;40:4694–701. doi: 10.1016/j.clnu.2021.06.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Loeb S, Fu BC, Bauer SR, et al Association of plant-based diet index with prostate cancer risk. Am J Clin Nutr. 2022;115:662–70. doi: 10.1093/ajcn/nqab365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zhong GC, Li Z, You AJ, et al Plant-based diets and the risk of pancreatic cancer: a large prospective multicenter study. Am J Clin Nutr. 2023;117:235–42. doi: 10.1016/j.ajcnut.2022.11.013. [DOI] [PubMed] [Google Scholar]
- 13.World Cancer Research Fund/American Institute for Cancer Research. Continuous Update Project expert report (2018). Diet, nutrition, physicalactivity and oesophageal cancer. Available online: https://www.wcrf.org/wp-content/uploads/2021/02/oesophageal-cancer-report.pdf
- 14.Zhu A, Chen H, Shen J, et al Interaction between plant-based dietary pattern and air pollution on cognitive function: a prospective cohort analysis of Chinese older adults. Lancet Reg Health West Pac. 2022;20:100372. doi: 10.1016/j.lanwpc.2021.100372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhu A, Yuan C, Pretty J, et al Plant-based dietary patterns and cognitive function: A prospective cohort analysis of elderly individuals in China (2008-2018) Brain Behav. 2022;12:e2670. doi: 10.1002/brb3.2670. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kim J, Khil J, Kim H, et al Plant-based dietary patterns and the risk of digestive system cancers in 3 large prospective cohort studies. Eur J Epidemiol. 2023;38:617–27. doi: 10.1007/s10654-023-01007-2. [DOI] [PubMed] [Google Scholar]
- 17.Wei WQ, Hao CQ, Guan CT, et al Esophageal histological precursor lesions and subsequent 8.5-year cancer risk in a population-based prospective study in China. Am J Gastroenterol. 2020;115:1036–44. doi: 10.14309/ajg.0000000000000640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Dawsey SM, Lewin KJ, Wang GQ, et al Squamous esophageal histology and subsequent risk of squamous cell carcinoma of the esophagus. A prospective follow-up study from Linxian, China. Cancer. 1994;74:1686–92. doi: 10.1002/1097-0142(19940915)74:6<1686::aid-cncr2820740608>3.0.co;2-v. [DOI] [PubMed] [Google Scholar]
- 19.Schulpen M, Peeters PH, van den Brandt PA Mediterranean diet adherence and risk of esophageal and gastric cancer subtypes in the Netherlands Cohort Study. Gastric Cancer. 2019;22:663–74. doi: 10.1007/s10120-019-00927-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pan D, Su M, Zhang T, et al A distinct epidemiologic pattern of precancerous lesions of esophageal squamous cell carcinoma in a high-risk area of Huai’an, Jiangsu Province, China. Cancer Prev Res (Phila) 2019;12:449–62. doi: 10.1158/1940-6207.CAPR-18-0462. [DOI] [PubMed] [Google Scholar]
- 21.Mohammadifard N, Sajjadi F, Maghroun M, et al Validation of a simplified food frequency questionnaire for the assessment of dietary habits in Iranian adults: Isfahan Healthy Heart Program, Iran. ARYA Atheroscler. 2015;11:139–46. [PMC free article] [PubMed] [Google Scholar]
- 22.Saeedi P, Skeaff SA, Wong JE, et al Reproducibility and relative validity of a short food frequency questionnaire in 9-10 year-old children. Nutrients. 2016;8:271. doi: 10.3390/nu8050271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wong JE, Parnell WR, Black KE, et al Reliability and relative validity of a food frequency questionnaire to assess food group intakes in New Zealand adolescents. Nutr J. 2012;11:65. doi: 10.1186/1475-2891-11-65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ashfield-Watt PA, Welch AA, Day NE, et al Is ‘five-a-day’ an effective way of increasing fruit and vegetable intakes. Public Health Nutr. 2004;7:257–61. doi: 10.1079/PHN2003524. [DOI] [PubMed] [Google Scholar]
- 25.Chen C, Kong L. Guidelines for the prevention and control of overweight and obesity in Chinese adults. Beijing: People’s Medical Publishing House, 2006.
- 26.Liu X, Wang X, Lin S, et al Dietary patterns and the risk of esophageal squamous cell carcinoma: A population-based case-control study in a rural population. Clin Nutr. 2017;36:260–6. doi: 10.1016/j.clnu.2015.11.009. [DOI] [PubMed] [Google Scholar]
- 27.Li WQ, Park Y, Wu JW, et al Index-based dietary patterns and risk of esophageal and gastric cancer in a large cohort study. Clin Gastroenterol Hepatol. 2013;11:1130–6.e2. doi: 10.1016/j.cgh.2013.03.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Zhao Y, Zhan J, Wang Y, et al The relationship between plant-based diet and risk of digestive system cancers: A meta-analysis based on 3,059,009 subjects. Front Public Health. 2022;10:892153. doi: 10.3389/fpubh.2022.892153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Arnold M, Ferlay J, van Berge Henegouwen MI, et al Global burden of oesophageal and gastric cancer by histology and subsite in 2018. Gut. 2020;69:1564–71. doi: 10.1136/gutjnl-2020-321600. [DOI] [PubMed] [Google Scholar]
- 30.Hong Y, Jiang F, Jin X, et al Present status and future prospect in the research of dietary nutrition of domestic and overseas residents. Nong Chan Pin Jia Gong (in Chinese) 2019;2:66–70. doi: 10.16693/j.cnki.1671-9646(X).2019.01.052. [DOI] [Google Scholar]
- 31.Abnet CC, Lai B, Qiao YL, et al Zinc concentration in esophageal biopsy specimens measured by x-ray fluorescence and esophageal cancer risk. J Natl Cancer Inst. 2005;97:301–6. doi: 10.1093/jnci/dji042. [DOI] [PubMed] [Google Scholar]
- 32.Song C, Zhang X, Wang X, et al Investigation results of acrylamide content in fried and baked foods in 9 cities of southeast Henan Province in 2020. An Hui Yu Fang Yi Xue Za Zhi (in Chinese) 2023;29:134–6, 176. doi: 10.19837/j.cnki.ahyf.2023.02.011. [DOI] [Google Scholar]
- 33.Wang Z, Yan X, Gao J, et al Experiment on grain mildew prevention in high incidence area of esophageal cancer. Zhongguo Huan Jing Ke Xue (in Chinese) 1996;1:31–3. [Google Scholar]
- 34.Zou X, Zheng S, Wang W, et al Dietary survey among the residents in Linzhou, a high risk area for esophageal cancer. Shi Yong Zhong Liu Za Zhi (in Chinese) 2003;18:148–50. doi: 10.13267/j.cnki.syzlzz.2003.02.030. [DOI] [Google Scholar]
- 35.Tran GD, Sun XD, Abnet CC, et al Prospective study of risk factors for esophageal and gastric cancers in the Linxian general population trial cohort in China. Int J Cancer. 2005;113:456–63. doi: 10.1002/ijc.20616. [DOI] [PubMed] [Google Scholar]
- 36.Jeurnink SM, Büchner FL, Bueno-de-Mesquita HB, et al Variety in vegetable and fruit consumption and the risk of gastric and esophageal cancer in the European Prospective Investigation into Cancer and Nutrition. Int J Cancer. 2012;131:E963–73. doi: 10.1002/ijc.27517. [DOI] [PubMed] [Google Scholar]
- 37.Sarker U, Rabbani MG, Oba S, et al Phytonutrients, colorant pigments, phytochemicals, and antioxidant potential of orphan leafy amaranthus species. Molecules. 2022;27:2899. doi: 10.3390/molecules27092899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Sewram V, Sitas F, O’Connell D, et al Diet and esophageal cancer risk in the Eastern Cape Province of South Africa. Nutr Cancer. 2014;66:791–9. doi: 10.1080/01635581.2014.916321. [DOI] [PubMed] [Google Scholar]
- 39.Vingeliene S, Chan DSM, Vieira AR, et al An update of the WCRF/AICR systematic literature review and meta-analysis on dietary and anthropometric factors and esophageal cancer risk. Ann Oncol. 2017;28:2409–19. doi: 10.1093/annonc/mdx338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hu FB, Stampfer MJ, Rimm E, et al Dietary fat and coronary heart disease: a comparison of approaches for adjusting for total energy intake and modeling repeated dietary measurements. Am J Epidemiol. 1999;149:531–40. doi: 10.1093/oxfordjournals.aje.a009849. [DOI] [PubMed] [Google Scholar]
- 41.Wang M, Hao C, Xie S, et al Efficacy of endoscopic treatment on patients with severe dysplasia/carcinoma in situ of esophageal squamous cell carcinoma: A prospective cohort study. Chin J Cancer Res. 2019;31:357–65. doi: 10.21147/j.issn.1000-9604.2019.02.10. [DOI] [PMC free article] [PubMed] [Google Scholar]