Abstract
BACKGROUND
The objective was to compare the dietary intakes of individuals with type 1 diabetes (T1D) to individuals without diabetes in China.
METHODS
Data are from 1) the 3C Nutrition Ancillary Study, a cross-sectional study of individuals with T1D in China, and 2) the China Health and Nutrition Survey. Dietary intake in both samples was assessed using three 24-hour recalls. ANCOVA and multivariable logistic regression, adjusted for sex, age, and urban-rural residence, were used to assess differences in nutrient and food group intake between participants without diabetes (n=1059) and participants with T1D (n=97), who were stratified by insulin regimen (basal-bolus, n=49, versus fixed, n=48).
RESULTS
Participants with T1D had a lower percentage of energy from carbohydrates, higher vegetable intake, and were more likely to consume low-fat cakes and fungi/seaweed compared to participants without diabetes (all p<0.05). Distinguishing characteristics of insulin regimen groups also emerged. Participants on fixed regimens had higher intakes of wheat and were less likely to consume fruit and more likely to consume high-fat cakes and dairy compared to participants without diabetes (all p<0.05). Participants on basal-bolus regimens were less likely to consume fried foods and more likely to consume fish/shellfish compared to participants without diabetes (all p<0.05).
CONCLUSIONS
Differences in dietary intake between participants with and without T1D in China suggest that dietary modifications are common and reflect carbohydrate-conscious nutrition recommendations for individuals with T1D. Future research should focus on the health effects of these modifications.
Keywords: human nutrition, carbohydrate, epidemiology, type 1 diabetes, China
INTRODUCTION
Modifications to dietary intake have consistently been an essential component of type 1 diabetes (T1D) treatment. In a considerable shift from the prescriptive diets characteristic of much of the 20th century,1–3 in 1994, the American Diabetes Association (ADA) published nutrition recommendations emphasizing individualization of dietary advice with a focus on the effects of nutrition therapy on metabolic control.4 The most recent ADA recommendations reiterate that there is no “one-size-fits-all” diet for individuals with diabetes and that food choices should only be limited when supported by scientific evidence.5
There is a dearth of information on the dietary intakes of individuals with T1D from low- and middle-income countries where economic development and urbanization have resulted in a transition from traditional foods to foods high in saturated fat and sugar, and low in fiber.6,7 Whether this nutrition transition has also affected the dietary intakes and self-management practices of individuals with T1D in these countries has yet to be explored.
A recent study conducted in Shanghai reported a mean annual increase in the incidence rate of T1D of 14.2% between 1997 and 2011, suggesting that the burden of T1D in China is increasing.8 China has also undergone dramatic changes in diet over the past 20 years.9–12 The objective of this study was to compare the dietary intakes of individuals with T1D in China to those of individuals without diabetes in China. Such a comparison will improve our understanding of how patients with T1D are modifying their diet relative to the general population in order to manage their condition and will inform future interventions to improve the lives of individuals with T1D in China.
MATERIALS AND METHODS
Study samples
The 3C Study was an epidemiological study of the coverage, cost, and care of T1D in China.13 The 3C Nutrition Ancillary Study (3CNAS) was conducted between February and August 2013, on average, 1.6±0.2 years (range: 1.2–2.0 years) after the 3C Study, and expanded the 3C Study to include comprehensive information on dietary intake. 3C Study participants who met the following criteria were eligible for 3CNAS: Beijing resident, ≥12 years old, no severe diabetes complications (i.e. advanced micro- and macrovascular complications, including nephropathy and stroke), and in-service telephone number available. A total of n=195 3C Study participants met these criteria. Of these, n=72 (37%) refused to participate and n=23 (12%) dropped out before the 3CNAS visit. The final sample size was n=100. There were no differences in diabetes duration, sex, education, or residence status between those who participated and those who refused or dropped out (all p>0.05). However, those who participated tended to be older (p=0.001), have higher household incomes (p=0.05), and be retired/unemployed (p=0.06) and married/cohabitating (p=0.02).
Data on individuals without T1D residing in Beijing came from the most recent (2011) China Health and Nutrition Survey (CHNS). CHNS is an ongoing open cohort of the health and nutritional status of the Chinese population.14 For the first time in 2011, 24 communities from Beijing were included in the survey. CHNS participants who met the following criteria were eligible for this analysis: Beijing resident, ≥12 years old, no diagnosed diabetes, and dietary data available. The final sample size was n=1059.
All procedures were approved by The University of North Carolina Office of Human Research Ethics and for 3CNAS, the Peking University Biomedical Institutional Review Board, and for CHNS, the China National Institute of Nutrition and Food Safety, China Center for Disease Control and Prevention. All participants provided informed consent (≥18 years old) or parent permission and participant assent (12 to <18 years old).
Dietary intake assessment
Dietary intake in 3CNAS was assessed using three (two weekdays and one weekend day) telephone administered 24-hour recalls assisted by food records. Telephone administration was used to overcome the cultural stigma associated with T1D in China, which precluded visiting the homes of participants. This method was validated against in-person administration in a subset of participants.15 At the end of the 3CNAS visit, participants were trained by dietitians to record their dietary intake on food records provided in an introductory packet. Emphasis was placed on estimating portion sizes using food samples, an electronic scale, and a culturally appropriate portion size picture guide. On average, beginning 3.0±2.8 days after this visit, participants completed the three 24-hour recalls.
Dietary intake in CHNS was assessed using three consecutive in-person administered 24-hour recalls randomly allocated to start between Monday and Sunday in combination with a household food inventory conducted over the same 3-day period.10 Food models and picture aids were used by trained interviewers to assist with estimating portion sizes. For dishes prepared at home, the recipe components were taken from the household food inventory and portion sizes were based on the proportion of the dish reportedly consumed by the participant.
For both 3CNAS and CHNS, the 24-hour recall food lists were converted into nutrients using the Chinese Food Composition Tables.16,17 Nutritionally meaningful food groups were derived by a collaborative working group that included researchers at The University of North Carolina, Chapel Hill, and the China National Institute of Nutrition and Food Safety.10 Nutrients were specified continuously as energy densities.18 Due to the non-normal distribution of some food groups largely stemming from large proportions of non-consumers, food groups were specified as binary variables. Food groups with ≥80% consumers were dichotomized as below versus above the median of the corresponding food group distribution in g/1000 kcal among participants without diabetes, and food groups with <80% consumers were dichotomized as non-consumers versus consumers.
Covariate assessment
Self-reported demographic data were collected via interviewer-administered questionnaires during the 3C Study visit and the CHNS visit. Both studies used standardized protocols to measure weight and height. Underweight and overweight for adults (≥18 years) were defined according to the WHO’s recommendation as underweight, BMI <18.5 kg/m2, and overweight, BMI ≥25 kg/m2,19 and for adolescents (12 to <18 years) using the sex- and age-specific cut-points recommended by the International Obesity Task Force.20,21
Information on insulin administration was collected during the 3CNAS visit. Two insulin regimens were defined as follows: (1) “basal-bolus,” including continuous subcutaneous infusion and regimens with ≥3 daily injections that included glargine or detemir, and (2) “fixed,” including regimens with ≥3 daily injections with any insulin types excluding glargine and detemir and regimens with 1–2 injections per day of any insulin types.
Statistical analysis
Descriptive statistics were used to summarize sample characteristics, and differences between participants with and without diabetes were evaluated using ANOVA for continuous variables and chi-square tests for categorical variables.
Potential confounders, including age, sex, marital status, household income, residence status, education, occupation, medical insurance coverage, smoking status, and BMI status, were explored using a directed acyclic graph22 and formally evaluated by estimating their associations with both the exposure (diabetes status) and outcomes (nutrients and food groups). We found that BMI status was a collider in the directed acyclic graph analysis, being both the result of diabetes status and dietary intake. We therefore did not include BMI status in our final models as adjustment for colliders results in biased estimates.23,24 The final adjustment set included sex, age (<22 years versus ≥22 years), and residence status (urban versus rural) as these factors were associated with the exposure (Table 1) and the outcomes (data not shown). Because n=3 participants with T1D were missing residence status, the final sample size for this analysis was n=1059 without diabetes and n=97 with T1D (n=49 on basal-bolus insulin regimens and n=48 on fixed insulin regimens).
Table 1.
Comparison of demographic and clinical characteristics between participants without diabetes and those with type 1 diabetes, combined and stratified by insulin regimen
| No diabetes (n=1059) | T1D (n=97) | T1D
|
p combined T1D † | p stratified T1D ‡ | ||
|---|---|---|---|---|---|---|
| Basal-bolus insulin (n=49) | Fixed Insulin (n=48) | |||||
| Age (years) | 43.1±15.3 | 41.3±16.3 | 41.9±17.4 | 40.7±15.3 | 0.28 | 0.52 |
| Age | ||||||
| <22 years | 8.3 (88) | 14.4 (14) | 16.3 (8) | 12.5 (6) | 0.04 | 0.10 |
| ≥22 years | 91.7 (971) | 85.6 (83) | 83.7 (41) | 87.5 (42) | ||
| Diabetes duration (years) | N/A | 11.6±9.3 | 13.6±10.9 | 9.5±6.9 | N/A | 0.03 |
| Sex | ||||||
| Female | 53.2 (563) | 46.4 (45) | 57.1 (28) | 35.4 (17) | 0.20 | 0.04 |
| Male | 46.8 (496) | 53.6 (52) | 42.9 (21) | 64.6 (31) | ||
| Marital status | ||||||
| Married/cohabitating | 88.0 (871) | 59.6 (56) | 52.2 (24) | 66.7 (32) | <0.001 | <0.001 |
| Single/divorced/widowed | 12.0 (119) | 40.4 (38) | 47.8 (22) | 33.3 (16) | ||
| Household income | ||||||
| <3000 RMB/month | 22.4 (234) | 19.0 (18) | 12.8 (6) | 25.0 (12) | 0.002 | 0.009 |
| 3000 – <5000 RMB/month | 25.3 (264) | 23.2 (22) | 23.4 (11) | 22.9 (11) | ||
| 5000 – <10,000 RMB/month | 39.9 (417) | 31.6 (30) | 34.0 (16) | 29.2 (14) | ||
| ≥10,000 RMB/month | 12.4 (130) | 26.3 (25) | 29.8 (14) | 22.9 (11) | ||
| Residence | ||||||
| Urban | 78.9 (836) | 84.5 (82) | 95.9 (47) | 72.9 (35) | 0.19 | 0.008 |
| Rural | 21.1 (223) | 15.5 (15) | 4.1 (2) | 27.1 (13) | ||
| Highest attained level of education | ||||||
| <University | 48.0 (506) | 32.3 (31) | 25.0 (12) | 39.6 (19) | 0.01 | 0.02 |
| Junior college | 12.4 (131) | 14.6 (14) | 14.6 (7) | 14.6 (7) | ||
| ≥University | 39.6 (417) | 53.1 (51) | 60.4 (29) | 45.8 (22) | ||
| Occupation | ||||||
| Non-government worker | 37.5 (364) | 30.9 (30) | 28.6 (14) | 33.3 (16) | <0.001 | <0.001 |
| Government worker | 22.8 (221) | 17.5 (17) | 16.3 (8) | 18.8 (9) | ||
| Student | 1.8 (17) | 17.5 (17) | 16.3 (8) | 18.8 (9) | ||
| Farmer | 1.4 (14) | 7.2 (7) | 4.1 (2) | 10.4 (5) | ||
| Retired or not working | 36.5 (354) | 26.8 (26) | 34.7 (17) | 18.8 (9) | ||
| Medical insurance coverage | ||||||
| Yes | 94.7 (1000) | 95.9 (93) | 95.9 (47) | 95.8 (46) | 0.62 | 0.88 |
| No | 5.3 (56) | 4.1 (4) | 4.1 (2) | 4.2 (2) | ||
| Smoking status | ||||||
| Non-smoker | 79.1 (834) | 80.4 (78) | 85.7 (42) | 75.0 (36) | 0.77 | 0.41 |
| Smoker | 20.9 (220) | 19.6 (19) | 14.3 (7) | 25.0 (12) | ||
| BMI Status § | ||||||
| Underweight | 3.1 (33) | 16.5 (16) | 14.3 (7) | 18.8 (9) | <0.001 | <0.001 |
| Normal weight | 55.0 (580) | 73.2 (71) | 67.4 (33) | 79.2 (38) | ||
| Overweight | 41.8 (441) | 10.3 (10) | 18.4 (9) | 2.1 (1) | ||
Note: T1D-type 1 diabetes.
Values are given as mean ± SD or percentage (n).
p from chi-square test for categorical variables and ANOVA for continuous variables comparing participants without diabetes and participants with type 1 diabetes.
p from chi-square test for categorical variables and ANOVA for continuous variables comparing participants without diabetes, participants with type 1 diabetes on basal-bolus insulin regimens, and participants with type 1 diabetes on fixed insulin regimens.
Underweight and overweight for adults (≥18 years) were defined according to the WHO’s recommendation as underweight, BMI <18.5 kg/m2, and overweight, BMI ≥25 kg/m2, and for adolescents (12 to <18 years) using the sex- and age-specific cut-points recommended by the International Obesity Task Force.
Confounder-adjusted differences were calculated using ANCOVA for continuous variables (nutrients) and multivariable logistic regression for binary variables (food groups). Given that dietary recommendations for T1D management are specific to an individual’s insulin regimen and differ between flexible, basal-bolus regimens and fixed regimens,25 analyses were conducted with 1) all participants with T1D combined and 2) participants with T1D stratified by insulin regimen (basal-bolus versus fixed).
All statistical analyses were conducted using SAS 9.2 (SAS Institute, Cary, North Carolina). Values presented are mean±SD or percentage (n) unless otherwise indicated.
RESULTS
Participants with T1D were more likely to be <22 years old (p=0.04) and single/divorced/widowed (p<0.0001), had higher incomes (p=0.002) and education (p=0.01), and were more likely to be underweight and less likely to be overweight (p<0.0001) compared to participants without diabetes (Table 1). Participants with T1D on fixed insulin regimens were more likely to be male compared to the other two groups (p=0.04), and participants with T1D on basal-bolus insulin regimens were more likely to be urban residents compared to the other two groups (p=0.008).
The macronutrient composition of the diets of participants with and without T1D differed substantially (Table 2). Participants with T1D had a significantly (p=0.01) lower adjusted mean percentage of energy from carbohydrates compared to participants without diabetes: approximately 47% of calories from carbohydrates in those with T1D versus 50% in those without diabetes. Consistent with this observation, participants with T1D had a significantly higher adjusted mean percentage of energy from fat (p=0.04) and protein (p=0.02) compared to participants without diabetes. Stratification by insulin regimen revealed that the differences in carbohydrate and fat were stronger among participants on fixed insulin regimens, while the difference in protein was more evident in participants on basal-bolus insulin regimens. No statistically significant differences were observed in adjusted mean fiber intake.
Table 2.
Comparison of nutrient intake between participants without diabetes and those with type 1 diabetes, combined and stratified by insulin regimen
| No diabetes (n=1059) | T1D (n=97) | T1D
|
p T1D vs. no diabetes † | p basal-bolus vs. no diabetes ‡ | p fixed vs. no diabetes ‡ | ||
|---|---|---|---|---|---|---|---|
| Basal-bolus insulin (n=49) | Fixed Insulin (n=48) | ||||||
| Energy intake (kcal/day) | 1769±18 | 1560±60.4 | 1543±85 | 1578±86 | 0.001 | 0.01 | 0.03 |
| Fat (% kcal) | 33.8±0.3 | 36.0±1.0 | 35.7±1.5 | 36.4±1.5 | 0.04 | 0.21 | 0.08 |
| Carbohydrate (% kcal) | 49.9±0.3 | 47.2±1.0 | 47.5±1.4 | 46.8±1.4 | 0.01 | 0.10 | 0.04 |
| Protein (% kcal) | 15.3±0.1 | 16.2±0.4 | 16.3±0.5 | 16.1±0.5 | 0.02 | 0.06 | 0.15 |
| Fiber (g/1000 kcal) | 8.2±0.1 | 7.4±0.5 | 7.4±0.7 | 7.4±0.7 | 0.13 | 0.27 | 0.27 |
Note: T1D-type 1 diabetes.
Values are given as least square (LS) mean±SD from analysis of covariance, adjusted for sex (male versus female), age (< 22 years versus ≥ 22 years), and residence status (urban versus rural).
p from pair-wise comparison of LS means from analysis of covariance. p > |t| for H0: LS mean(type 1 diabetes) = LS mean(no diabetes).
p from pair-wise comparison of LS means from analysis of covariance. p > |t| for H0: LS mean(basal-bolus or fixed insulin) = LS mean(no diabetes).
Participants with T1D in both insulin regimen groups were more likely to be consumers of low-fat cakes (adjusted OR [95% CI], 3.19 [1.99, 5.10]) and fungi/seaweed (adjusted OR [95% CI], 2.69 [1.76, 4.12]) compared to participants without diabetes (Table 3). They were also significantly more likely to be above the median intake for vegetables (adjusted OR [95% CI], 8.33 [4.37, 15.86]) compared to participants without diabetes.
Table 3.
Comparison of food group intake between participants without diabetes and those with type 1 diabetes, combined and stratified by insulin regimen
| No diabetes (n=1059) | T1D (n=97) | T1D
|
OR (95% CI)
|
||||
|---|---|---|---|---|---|---|---|
| Basal-bolus insulin (n=49) | Fixed Insulin (n=48) | T1D vs. no diabetes (ref) ‡ | Basal-bolus vs. no diabetes (ref) ‡ | Fixed vs. no diabetes (ref) ‡ | |||
| Rice | |||||||
| ≤ Median † (ref) | 50.1 (530) | 56.7 (55) | 55.1 (27) | 58.3 (28) | 0.74 (0.49, 1.13) | 0.81 (0.45, 1.45) | 0.68 (0.38, 1.23) |
| > Median | 50.0 (529) | 43.3 (42) | 44.9 (22) | 41.7 (20) | |||
| Wheat products | |||||||
| ≤ Median † (ref) | 50.1 (530) | 38.1 (37) | 44.9 (22) | 31.3 (15) | 1.66 (1.08, 2.56) | 1.39 (0.78, 2.49) | 2.03 (1.08, 3.81) |
| > Median | 50.0 (529) | 61.9 (60) | 55.1 (27) | 68.8 (33) | |||
| Low-fat cakes | |||||||
| Non-consumer (ref) | 87.0 (921) | 68.0 (66) | 65.3 (32) | 70.8 (34) | 3.19 (1.99, 5.10) | 3.25 (1.74, 6.06) | 3.12 (1.62, 6.03) |
| Consumer | 13.0 (138) | 32.0 (31) | 34.7 (17) | 29.2 (14) | |||
| High-fat cakes | |||||||
| Non-consumer (ref) | 90.9 (963) | 81.4 (79) | 81.6 (40) | 81.3 (39) | 2.25 (1.28, 3.95) | 1.91 (0.89, 4.09) | 2.73 (1.25, 5.95) |
| Consumer | 9.1 (96) | 18.6 (18) | 18.4 (9) | 18.8 (9) | |||
| Fried foods | |||||||
| Non-consumer (ref) | 69.3 (734) | 76.3 (74) | 81.6 (40) | 70.8 (34) | 0.68 (0.42, 1.11) | 0.48 (0.23, 1.00) | 0.93 (0.49, 1.77) |
| Consumer | 30.7 (325) | 23.7 (23) | 18.4 (9) | 29.2 (14) | |||
| Fast food | |||||||
| Non-consumer (ref) | 42.5 (450) | 41.2 (40) | 44.9 (22) | 37.5 (18) | 1.03 (0.68, 1.58) | 0.87 (0.49, 1.55) | 1.23 (0.68, 2.25) |
| Consumer | 57.5 (609) | 58.8 (57) | 55.1 (27) | 62.5 (30) | |||
| Vegetables | |||||||
| ≤ Median † (ref) | 50.1 (530) | 11.3 (11) | 8.2 (4) | 14.6 (7) | 8.33 (4.37, 15.86) | 11.45 (4.06, 32.26) | 6.53 (2.88, 14.78) |
| > Median | 50.0 (529) | 88.7 (86) | 91.8 (45) | 85.4 (41) | |||
| Fruit | |||||||
| Non-consumer (ref) | 26.0 (275) | 32.0 (31) | 18.4 (9) | 45.8 (22) | 0.70 (0.44, 1.13) | 1.20 (0.56, 2.57) | 0.47 (0.26, 0.87) |
| Consumer | 74.0 (784) | 68.0 (66) | 81.6 (40) | 54.2 (26) | |||
| Fungi & seaweed | |||||||
| Non-consumer (ref) | 65.3 (692) | 41.2 (40) | 34.7 (17) | 47.9 (23) | 2.69 (1.76, 4.12) | 3.47 (1.89, 6.36) | 2.10 (1.17, 3.77) |
| Consumer | 34.7 (367) | 58.8 (57) | 65.3 (32) | 52.1 (25) | |||
| Nuts & seeds | |||||||
| Non-consumer (ref) | 66.3 (702) | 56.7 (55) | 57.1 (28) | 56.3 (27) | 1.49 (0.97, 2.29) | 1.31 (0.73, 2.36) | 1.71 (0.94, 3.11) |
| Consumer | 33.7 (357) | 43.3 (42) | 42.9 (21) | 43.8 (21) | |||
| Beans & bean products | |||||||
| Non-consumer (ref) | 25.4 (269) | 19.6 (19) | 14.3 (7) | 25.0 (12) | 1.40 (0.83, 2.36) | 2.10 (0.93, 4.74) | 1.00 (0.51, 1.96) |
| Consumer | 74.6 (790) | 80.4 (78) | 85.7 (42) | 75.0 (36) | |||
| Red meat | |||||||
| ≤ Median † (ref) | 50.0 (529) | 42.3 (41) | 42.9 (21) | 41.7 (20) | 1.32 (0.86, 2.02) | 1.24 (0.69, 2.23) | 1.40 (0.77, 2.54) |
| > Median | 50.1 (530) | 57.7 (56) | 57.1 (28) | 58.3 (28) | |||
| Poultry | |||||||
| Non-consumer (ref) | 64.4 (682) | 69.1 (67) | 65.3 (32) | 72.9 (35) | 0.72 (0.46, 1.14) | 0.79 (0.43, 1.46) | 0.65 (0.34, 1.26) |
| Consumer | 35.6 (377) | 30.9 (30) | 34.7 (17) | 27.1 (13) | |||
| Fish & shellfish | |||||||
| Non-consumer (ref) | 59.1 (626) | 48.5 (47) | 40.8 (20) | 56.3 (27) | 1.52 (1.00, 2.32) | 1.95 (1.08, 3.52) | 1.18 (0.65, 2.13) |
| Consumer | 40.9 (433) | 51.6 (50) | 59.2 (29) | 43.8 (21) | |||
| Eggs | |||||||
| ≤ Median † (ref) | 50.1 (530) | 39.2 (38) | 34.7 (17) | 43.8 (21) | 1.55 (1.01, 2.38) | 1.76 (0.96, 3.21) | 1.38 (0.77, 2.48) |
| > Median | 50.0 (529) | 60.8 (59) | 65.3 (32) | 56.3 (27) | |||
| Milk & milk products | |||||||
| Non-consumer (ref) | 34.0 (360) | 17.5 (17) | 14.3 (7) | 20.8 (10) | 2.39 (1.35, 4.25) | 2.19 (0.95, 5.05) | 2.57 (1.19, 5.53) |
| Consumer | 66.0 (699) | 82.5 (80) | 85.7 (42) | 79.2 (38) | |||
| Sugar-sweetened beverages | |||||||
| Non-consumer (ref) | 89.9 (952) | 90.7 (88) | 85.7 (42) | 95.8 (46) | 0.85 (0.41, 1.74) | 1.24 (0.54, 2.87) | 0.40 (0.10, 1.70) |
| Consumer | 10.1 (107) | 9.3 (9) | 14.3 (7) | 4.2 (2) | |||
Note: T1D-type 1 diabetes; ref-referent.
Values are given as percentage (n) or OR (95% CI).
Median defined according to distribution in g/1000 kcal among participants without diabetes.
OR (95% CI) from multivariable logistic regression, adjusted for sex (male versus female), age (< 22 years versus ≥ 22 years), and residence status (urban versus rural).
Participants with T1D on basal-bolus insulin regimens had several distinct dietary modifications (Table 3): they were less likely to be consumers of fried foods (adjusted OR [95% CI], 0.48 [0.23, 1.00]) and more likely to be consumers of fish/shellfish (adjusted OR [95% CI], 1.95 [1.08, 3.52]) compared to participants without diabetes.
Participants with T1D on fixed insulin regimens also had several distinct dietary modifications (Table 3): they were more likely to be above the median intake for wheat products (adjusted OR [95% CI], 2.03 [1.08, 3.81]), and they were more likely to be consumers of high-fat cakes (adjusted OR [95% CI], 2.73 [1.25, 5.95]) and milk/milk products (adjusted OR [95% CI], 2.57 [1.19, 5.53]). Furthermore, they were less likely to be consumers of fruit (adjusted OR [95% CI], 0.47 [0.26, 0.87]).
DISCUSSION
Individuals with T1D in China had a lower mean percentage of energy from carbohydrates compared to individuals without diabetes. In addition, they had higher intakes of vegetables, fungi/seaweed, and low-fat cakes. Several distinguishing characteristics of insulin regimen groups also emerged: participants on fixed regimens had higher intakes of wheat products and were less likely to consume fruit and more likely to consume high-fat cakes and dairy compared to participants without diabetes. In contrast, participants on basal-bolus regimens were less likely to consume fried foods and more likely to consume fish/shellfish compared to participants without diabetes. Together, these observations suggest that dietary modifications are common among individuals with T1D in China and reflect carbohydrate-conscious nutrition recommendations.
Few studies have systematically compared the dietary intakes of individuals with T1D and the general population. One study conducted in the United States found similar results to our study, reporting that adults with T1D had a higher mean percentage of energy from fat and protein and a lower mean percentage of energy from carbohydrates compared to controls.26 A similarly higher mean percentage of energy from fat compared to controls has also been observed in several small samples of youth with T1D in Europe.27,28
In contrast to the aforementioned studies, a small study conducted in Australia did not find any statistically significant differences in dietary intake between adults with newly diagnosed T1D and controls despite the fact that all of the T1D participants were reportedly following fixed insulin regimens matched to a set diet.29 Two other studies, both conducted in youth with T1D in Europe, also did not find significant differences in the macronutrient composition of the diet between participants with T1D and controls.30,31 In our study, most participants with T1D were on fixed insulin regimens matched to a rigid diet with respect to timing and amount of food,15 and we observed substantial differences in dietary intake between participants with T1D and those without diabetes despite previously reporting limited nutrition education in this group of participants.15 This result may partially stem from the fact that our sample, relative to the aforementioned samples, was older and had longer disease durations. More research is needed to identify the underlying factors contributing to the observed dietary modifications of individuals with T1D in China.
The mean fiber intakes of participants with T1D and those without diabetes were well below the recommended level of 14 g/1000 kcal.5 While there was no statistically significant difference between these means after adjustment for confounders, the fiber intake of participants with T1D was, on average, approximately 0.8 g/1000 kcal lower than that of participants without diabetes. Given that fiber intake is inversely associated with all-cause mortality and cardiovascular disease risk among individuals with diabetes,32,33 identifying strategies for improving dietary fiber intake should be an important goal of future research.
Individuals with T1D in both insulin regimen groups had higher intakes of vegetables, fungi/seaweed, and low-fat cakes compared to individuals without diabetes. The observation that they were more likely to consume fungi/seaweed may be the result of recommendations during diabetes education courses to eat more sugar-free foods, of which fungi are used as an example. The increased intake of low-fat cakes, which includes biscuits/crackers, likely reflects the common use of these food items to treat hypoglycemia in China.
We observed that individuals with T1D on fixed insulin regimens were increasing their wheat product intake relative to individuals without diabetes. They also restricted their rice intake relative to individuals without diabetes, though it was not statistically significant. These observations may reflect the fact that physicians in China, when discussing the effects of carbohydrates on blood glucose, often use rice as an example. In the absence of additional nutrition education, we hypothesize that patients may subsequently equate rice with high blood glucose and replace rice with wheat products for their staple food. Similarly, we observed that individuals with T1D on fixed insulin regimens were more likely to consume high-fat cakes compared to individuals without diabetes, perhaps because sugar-free cakes tend to be higher in fat and patients focus on low-sugar, low-carbohydrate foods rather than the totality of nutritional information. Finally, individuals with T1D on fixed insulin regimens (but not those on basal-bolus insulin regimens) restricted fruit intake, a phenomenon also reported in the United States,34 and again, likely the result of a carbohydrate focus rather than a healthy diet focus. Overall, the lack of physician time and dietitian involvement in T1D care in China may be contributing to patient misconceptions relating to nutrition. In-depth, qualitative research into this phenomenon may prove to be informative for future nutrition interventions in this population.
Participants with T1D on basal-bolus insulin regimens had a generally healthier diet than the other groups, consuming fewer fried foods and more fish/shellfish. They also had higher intakes of beans and eggs relative to individuals without diabetes, though these differences were not statistically significant. These observations are consistent with the higher protein intake found in this group. Together, these results suggest that individuals with T1D in China who are on basal-bolus insulin regimens may be more motivated and/or have higher adherence to self-management relating to nutrition.
Participants with T1D in both insulin regimen groups had significantly lower energy intake compared to participants without diabetes. They were also more likely to be underweight and less likely to be overweight compared to individuals without diabetes. This observation is consistent with a study conducted in Guangdong in southern China, which reported underweight and overweight prevalences of 19.6% and 11.8%, respectively, among individuals with T1D.35 Several factors may be underlying these results including the fact that most participants with T1D in 3CNAS reported a rigid diet with respect to timing and amount of food, including those on basal-bolus insulin regimens,15 and this may prevent overeating and weight gain. Furthermore, the diabetes duration in this sample was relatively long and though 3C Study participants with self-reported advanced micro- and macrovascular complications were not eligible for 3CNAS, it is possible that the natural progression of T1D and undiagnosed complications contributed to the relatively high prevalence of underweight.
This study focused on individuals with T1D in a large urban area in northern China. Therefore, results may not be generalizable to more rural areas or southern China where the diet differs substantially from the north. Additional challenges of this study were differences in dietary assessment methods between 3CNAS and CHNS. While interviewers from both studies were trained by staff from the China Center for Disease Control and Prevention, and the 24-hour recall questionnaire and food composition tables were the same between studies, we cannot rule out that some of the differences observed between individuals with and without diabetes were the result of slight differences in dietary assessment methods rather than true differences between these populations. Related to this, dietary intake was self-reported and therefore subject to differential misclassification as individuals with T1D may be more accustomed to reporting dietary intake. Finally, participants with T1D on basal-bolus insulin regimens were more urban and therefore may have more modern lifestyles than the other two groups. Although we controlled for residence status in the analysis, we cannot rule out residual confounding by unmeasured correlates of a more urbanized, modern lifestyle.
The substantial differences in macronutrient content, particularly carbohydrates, and food groups between individuals with T1D in China and those without diabetes suggest that dietary modifications are common and reflect carbohydrate-conscious nutrition recommendations for individuals with T1D. Future research should focus on the effects of these modifications on quality of life and health outcomes.
Acknowledgments
3CNAS is indebted to the individuals whose participation made this study possible, and to our interviewers at Peking University People’s Hospital: Yang Xiao, Jing Lv, Wenjia Yang, Jia Liu, Han Feifei, and Lihua Zhang. We would like to thank Xueying Zheng of the Third Affiliated Hospital of Sun Yat-Sen University for her contribution to the interpretation of the results. We would also like to acknowledge Helen McGuire, Principle Investigator of the Care and Education arm of the 3C Study: Coverage, Cost, and Care of Type 1 Diabetes in China. The 3C Study was a collaborative effort of the International Diabetes Federation and the Chinese Diabetes Society.
Footnotes
Data collection was conducted at the Department of Endocrinology and Metabolism, Peking University People’s Hospital, Beijing, China. Data analysis was conducted at the Department of Nutrition, University of North Carolina, Chapel Hill, North Carolina, United States.
AUTHOR CONTRIBUTIONS
All authors were involved in developing the study design and analysis plan. LMJ and WL coordinated and supervised the field work. SD calculated the nutrient and food group intakes. LMJ managed the data, performed all statistical analyses, and wrote the initial draft of the manuscript. All authors contributed to interpreting the data and critically reviewing the manuscript. All authors approved the final version submitted for publication.
CONFLICT OF INTEREST AND FUNDING DISCLOSURE
3CNAS was supported by funding from the Sanofi Global Scholars Program, the Fogarty International Center of the U.S. National Institutes of Health (5D43TW009077), and the National Center for Advancing Translational Sciences (ULTR000083). None of the aforementioned funding sources had a role in the design, analysis, or writing of this article. No potential conflicts of interest relevant to this article were reported.
References
- 1.Leeds AR. The dietary management of diabetes in adults. Proc Nutr Soc. 1979;38:365–71. doi: 10.1079/PNS19790061. [DOI] [PubMed] [Google Scholar]
- 2.Joslin E. The diabetic diet. J Am Diet Assoc. 1927;3:89–92. [Google Scholar]
- 3.Caso EK. Calculation of diabetic diets. J Am Diet Assoc. 1950;26:575–83. [PubMed] [Google Scholar]
- 4.Franz MJ, Horton ESS, Bantle JP, Beebe CA, Brunzell JD, Coulston AM, Henry RR, Hoogwerf BJ, Stacpoole PW. Nutrition principles for the management of diabetes and related complications. Diabetes Care. 1994;17:490–518. doi: 10.2337/diacare.17.5.490. [DOI] [PubMed] [Google Scholar]
- 5.Evert AB, Boucher JL, Cypress M, Dunbar SA, Franz MJ, Mayer-Davis EJ, et al. Nutrition therapy recommendations for the management of adults with diabetes. Diabetes Care. 2013;36:3821–42. doi: 10.2337/dc13-2042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Popkin B. Nutritional patterns and transitions. Popul Dev Rev. 1993;19:138–57. [Google Scholar]
- 7.Popkin BM. The nutrition transition in low-income countries: an emerging crisis. Nutr Rev. 1994;52:285–98. doi: 10.1111/j.1753-4887.1994.tb01460.x. [DOI] [PubMed] [Google Scholar]
- 8.Zhao Z, Sun C, Wang C, Li P, Wang W, Ye J, et al. Rapidly rising incidence of childhood type 1 diabetes in Chinese population: epidemiology in Shanghai during 1997–2011. Acta Diabetol. doi: 10.1007/s00592-014-0590-2. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 9.Popkin BM, Keyou G, Zhai F, Guo X, Ma H, Zohoori N. The nutrition transition in China: a cross-sectional analysis. Eur J Clin Nutr. 1993;47:333–46. [PubMed] [Google Scholar]
- 10.Popkin BM, Lu B, Zhai F. Understanding the nutrition transition: measuring rapid dietary changes in transitional countries. Public Health Nutr. 2002;5:947–53. doi: 10.1079/PHN2002370. [DOI] [PubMed] [Google Scholar]
- 11.Du S, Lu B, Zhai F, Popkin BM. A new stage of the nutrition transition in China. Public Health Nutr. 2002;5:169–74. doi: 10.1079/PHN2001290. [DOI] [PubMed] [Google Scholar]
- 12.Zhai F, Wang H, Du S, He Y, Wang Z, Ge K, Popkin BM. Prospective study on nutrition transition in China. Nutr Rev. 2009;67:S56–61. doi: 10.1111/j.1753-4887.2009.00160.x. [DOI] [PubMed] [Google Scholar]
- 13.McGuire H, Kissimova-Skarbek K, Whiting D, Ji L. The 3C Study: coverage cost and care of type 1 diabetes in China—study design and implementation. Diabetes Res Clin Pract. 2011;94:307–10. doi: 10.1016/j.diabres.2011.10.016. [DOI] [PubMed] [Google Scholar]
- 14.Popkin BM, Du S, Zhai F, Zhang B. Cohort Profile: The China Health and Nutrition Survey—monitoring and understanding socio-economic and health change in China, 1989–2011. Int J Epidemiol. 2010;39:1435–40. doi: 10.1093/ije/dyp322. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jaacks LM, Liu W, Ji L, Mendez M, Du S, Crandell J, Rosamond W, Mayer-Davis EJ. Diabetes nutrition therapy and dietary intake among individuals with type 1 diabetes in China. Diabet Med. doi: 10.1111/dme.12634. [In press] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Institute of Nutrition and Food Hygiene, editor. Chinese Food Composition Tables, Book 1. 2. People’s Medical Publishing House; Beijing: 2009. [Google Scholar]
- 17.Institute of Nutrition and Food Hygiene, editor. Chinese Food Composition Tables, Book 2. 1. People’s Medical Publishing House; Beijing: 2004. [Google Scholar]
- 18.Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. Am J Clin Nutr. 1997;65:S1220–8. doi: 10.1093/ajcn/65.4.1220S. [DOI] [PubMed] [Google Scholar]
- 19.World Health Organization and Food and Agriculture Organization. Diet, nutrition and the prevention of chronic diseases: report of a joint WHO/FAO expert consultation, Geneva, 28 January–1 February 2002. World Health Organ Tech Rep Ser. 2003;916:1–149. [PubMed] [Google Scholar]
- 20.Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ. 2000;320:1240–3. doi: 10.1136/bmj.320.7244.1240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Cole TJ, Flegal KM, Nicholls D, Jackson AA. Body mass index cut offs to define thinness in children and adolescents: international survey. BMJ. 2007;335:194–201. doi: 10.1136/bmj.39238.399444.55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10:37–48. [PubMed] [Google Scholar]
- 23.Howards PP, Schisterman EF, Poole C, Kaufman JS, Weinberg CR. “Toward a clearer definition of confounding” revisited with directed acyclic graphs. Am J Epidemiol. 2012;176:506–11. doi: 10.1093/aje/kws127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Weinberg CR. Toward a clearer definition of confounding. Am J Epidemiol. 1993;137:1–8. doi: 10.1093/oxfordjournals.aje.a116591. [DOI] [PubMed] [Google Scholar]
- 25.Smart CC. Nutritional management in children and adolescents with diabetes. Pediatr Diabetes. 2009;10:100–17. doi: 10.1111/j.1399-5448.2009.00572.x. [DOI] [PubMed] [Google Scholar]
- 26.Snell-Bergeon JK, Chartier-Logan C, Maahs DM, Ogden LG, Hokanson JE, Kinney GL, Eckel RH, Ehrlich J, Rewers M. Adults with type 1 diabetes eat a high-fat atherogenic diet that is associated with coronary artery calcium. Diabetologia. 2009;52:801–9. doi: 10.1007/s00125-009-1280-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Overby NC, Flaaten V, Veierod MB, Bergstad I, Margeirsdottir HD, Dahl-Jorgensen K, Andersen LF. Children and adolescents with type 1 diabetes eat a more atherosclerosis-prone diet than healthy control subjects. Diabetologia. 2007;50:307–16. doi: 10.1007/s00125-006-0540-9. [DOI] [PubMed] [Google Scholar]
- 28.Lodefalk M, Aman J. Food habits, energy and nutrient intake in adolescents with type 1 diabetes mellitus. Diabet Med. 2006;23:1225–32. doi: 10.1111/j.1464-5491.2006.01971.x. [DOI] [PubMed] [Google Scholar]
- 29.Tahbaz F, Kreis I, Calvert D. An audit of diabetes control, dietary management and quality of life in adults with type 1 diabetes mellitus, and a comparison with nondiabetic subjects. J Hum Nutr Diet. 2006;19:3–11. doi: 10.1111/j.1365-277X.2006.00668.x. [DOI] [PubMed] [Google Scholar]
- 30.Pietiläinen KH, Virtanen SM, Rissanen A, Rita H, Mäenpää J. Diet, obesity, and metabolic control in girls with insulin dependent diabetes mellitus. Arch Dis Child. 1995;73:398–402. doi: 10.1136/adc.73.5.398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Schober E, Langergraber B, Rupprecht G, Rami B. Dietary intake of Austrian diabetic children 10 to 14 years of age. J Pediatr Gastroenterol Nutr. 1999;29:144–7. doi: 10.1097/00005176-199908000-00009. [DOI] [PubMed] [Google Scholar]
- 32.Slavin JL. Position of the American Dietetic Association: health implications of dietary fiber. J Am Diet Assoc. 2008;108:1716–31. doi: 10.1016/j.jada.2008.08.007. [DOI] [PubMed] [Google Scholar]
- 33.Wheeler ML, Dunbar SA, Jaacks LM, Karmally W, Mayer-Davis EJ, Wylie-Rosett J, Yancy WS., Jr Macronutrients, food groups, and eating patterns in the management of diabetes: a systematic review of the literature, 2010. Diabetes Care. 2012;35:434–45. doi: 10.2337/dc11-2216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Mehta SN, Haynie DL, Higgins LA, Bucey NN, Rovner AJ, Volkening LK, Nansel TR, Laffel LM. Emphasis on carbohydrates may negatively influence dietary patterns in youth with type 1 diabetes. Diabetes Care. 2009;32:2174–6. doi: 10.2337/dc09-1302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Li J, Yang D, Yan J, Huang B, Zhang Y, Weng J. Secondary Diabetic ketoacidosis and severe hypoglycaemia in patients with established type 1 diabetes mellitus in China: a multicentre registration study. Diabetes Metab Res Rev. 2014;30:497–504. doi: 10.1002/dmrr.2547. [DOI] [PubMed] [Google Scholar]