Adherence to a Mediterranean-like Diet as a Protective Factor Against COPD: A Nested Case-Control Study

Abstract

A diet rich in nutrients has been suggested to have protective effects against the development of chronic obstructive pulmonary disease (COPD). Since the traditional Mediterranean diet is high in nutrients, including antioxidants, vitamins, and minerals, it is of interest to study as a protective factor against COPD. Our aim was therefore to study its associations with development of COPD using population-based prospective data from the Västerbotten Intervention Programme (VIP) cohort. Data on diet from 370 individuals, who later visited the Department of Medicine at the University Hospital, Umeå, Sweden, with a diagnosis of COPD, were compared to 1432 controls. Adherence to a Mediterranean diet was assessed by a modified version of the Mediterranean diet score (MDS). Cases were diagnosed with COPD 11.1 years (mean) (standard deviation [SD] 4.5 years) after first stating their dietary habits in the VIP at a mean age of 55.5 years (SD 6.6 years). Higher MDS was associated with a higher level of education and not living alone. After adjustment for co-habiting and education level, individuals with an intermediate MDS and those with the highest MDS had a lower odds of developing COPD (odds ratio [OR] 0.73, 95% confidence interval [CI] 0.56–0.95; OR 0.56, 95% CI 0.37–0.86, respectively). These results remained also after adjustment for smoking intensity, i.e., numbers of cigarettes smoked per day (OR 0.73, 95% CI 0.53–0.99; OR 0.59, 95% CI 0.35–0.97), respectively). To conclude, adherence to a Mediterranean-like diet seems to be inversely associated with the development of COPD.

Keywords:

• COPD
• dietary habit
• Mediterranean diet
• nutrients

Introduction

According to the Global Initiative for Chronic Obstructive Lung Disease (GOLD), chronic obstructive pulmonary disease (COPD) is estimated to become the third leading cause of mortality by 2030 (1). Whilst there are genetic predispositions to develop COPD, the most well-known is the α1‐antitrypsin deficiency, which affects between 0.02% and 0.05% of the population, smoking or exposure to other hazardous fumes, and dust are the major risk factors for COPD (2,3). Approximately 50% of the smokers develop COPD (4), but since not everyone exposed to smoking or hazardous fumes/dust develop COPD, protective factors related to lifestyle, such as diet, may exist. Previous studies on diet have focused mainly on specific dietary micro- and macronutrients, including antioxidants, vitamins, fatty acids, meat, dietary fibers, fruit, and vegetables. Although the results of these studies differ, it appears that consuming a diet high in antioxidants, vitamins, omega-3 fatty acids, fiber, fruits, and vegetables and low in processed meat offers some protection against developing COPD (5–11).

Recently, there has been a shift toward studying the effect of diet on diseases by examining dietary patterns instead of isolated nutrients, due to the complexity of dietary habits (12). The traditional Mediterranean diet is especially interesting, as it is characterized by a high intake of vegetables, legumes, fruits and nuts, unrefined cereals, olive oil, and fish as well as a low intake of saturated fats, meat, and poultry and a low-to-moderate intake of dairy products (13). The Mediterranean diet concept was first brought to light by Ancel Keys and The Seven Countries Study of cardiovascular disease (14) and has been proven to reduce both morbidity and mortality in numerous diseases (15). However, to our knowledge, no study has been performed to investigate whether a Mediterranean diet may be a preventive factor against the development of COPD.

Our aim was, therefore, to investigate if a Mediterranean-style dietary pattern is associated with the development of COPD in the large cohort of the Västerbotten Intervention Program (VIP) in northern Sweden.

Methods

Study area and settings

Due to a high incidence of cardiovascular diseases in the Västerbotten County in northern Sweden, with approximately 260,000 citizens, a health-screening and intervention project was started in 1985 (16) and gradually extended to the Västerbotten Intervention Programme (VIP). Since 1991, the program invited all citizens aged 30, 40, 50, and 60 years throughout the county; individuals aged 30 years were not included after 1996. The participation rate for the VIP varied between 48% and 57% of the target population during the first years; however, it gradually increased to 72% in recent years (17). By the end of 2011, approximately 30% of the studied individuals had participated in the VIP more than once (18). As part of the VIP, background data, such as education and marital status, and data on lifestyle, such as smoking and dietary habits, are investigated through the use of questionnaires. In addition, participants are given the opportunity to donate their data to the Northern Sweden Health and Disease Study (NSHDS) for future research purposes.

Identification of cases and controls

Cases were identified through registered visits with a diagnosis of COPD at the Department of Medicine, Division of Respiratory Medicine and Allergy, University Hospital, Umeå, between 2007 and 2013, using International Classification of Diseases (ICD)-codes J44.0, J44.1, and J44.9 according to clinical practice. The use of J44.x for epidemiological studies of COPD in a Swedish context has been validated with acceptable validity for epidemiological research (19). The resulting cases were searched in the NSHDS for visits in the VIP before their first registered diagnosis at the University Hospital during 1991–2013. For each case found, five controls matched for smoking, age, sex, place of residence, time of examination in the VIP (±2.5 years), and version of questionnaires used were selected using a sampling scheme with replacement (WR scheme).

Study variables and handling of data

Smoking habits in the VIP questionnaires were labeled as never smoker, current daily smoker, current occasional smoker, previous daily smoker, and previous occasional smoker; these were also used as matching variables. In some versions of the questionnaires used in the VIP, additional quantitative data on smoking were requested (i.e., number of cigarettes smoked per day), thereby providing data on smoking intensity. Education level was evaluated as a dichotomous variable with higher education defined as studies at college or university. Participants were classified as either living alone or cohabitating; the former was grouped as unmarried, widow/widower, or divorced/separated, and the latter as married, cohabitating, or remarried.

Dietary habits were assessed through the use of a semi-quantitative food frequency questionnaire (FFQ). Until 1996, the form consisted of 84 questions, but since 1997, the questions have been merged to range between 64 and 66 questions. The longer version of the FFQ has been validated, (20) and comparisons between the versions have shown that ranking is similar, although absolute levels tend to be slightly lower when using the shorter version (21). The FFQ assesses the intake frequency of different foodstuffs on a nine-level scale ranging from never to four or more times per day and contains examples of portion sizes. These frequencies and portion sizes were used to calculate daily energy and nutritional intake based on the reference databases of the Swedish National Food Administration (22). Nutritional intake was adjusted with regard to energy intake using the residual method (23).

For cases and controls, a food intake level (FIL) was calculated as energy intake divided by basal metabolic rate, the latter according to Schofield (24). Individuals who did not state portion size in the FFQ (n = 320; 92 cases and 228 controls), thereby rendering calculations of energy intake impossible, were excluded. Furthermore, individuals with a FIL below the bottom 1st and above the upper 99th percentiles in the total VIP population were classified as unreliable under and over-reporters and were also excluded from the study. Individuals who stated that they were occasional or previous occasional smokers were excluded due to issues with the classification of their exposure, i.e., only cases and controls who stated that they were current smokers, ex-smokers, or never smokers were included in the study. If cases had several visits in the VIP before their diagnosis of COPD (n = 65), the visit closest to the mean time between the VIP and the diagnosis of COPD for the total cohort was used to decrease the heterogenicity of the studied group. After selection according to the criteria above, the final cohort consisted of 370 cases and 1432 controls.

Definition of the Mediterranean diet

Adherence to a Mediterranean dietary pattern was scored based on the Mediterranean diet score (MDS) established by Trichopoulou et al. (13,25) with modifications by Tognon et al. (26). The modified MDS is based on food components typical of a Mediterranean diet, where a favorable gender-specific intake above or below the median was scored one point, together with a favorable intake of alcohol.

The following components were scored one point each: (1) a ratio between intake of mono-/polyunsaturated and saturated fats above the median; (2) intake of vegetables and potatoes above the median; (3) intake of fruit above the median; (4) intake of wholegrain cereals above the median; (5) intake of fish above the median; (6) intake of meat and meat products below the median; (7) intake of dairy products below the median; and (8) an intake of alcohol between 5–25 g/day for women and 10–50 g/day for men. The gender-specific median intake of the components is presented in Supplementary Table S1. Since the shorter version of the FFQ tends to yield slightly lower intake levels (21), the above scoring system was not only gender-specific but also specific to the version of FFQ used. The final modified MDS score ranged from 0 to 8, with a higher score indicating greater adherence to a Mediterranean-like diet.

Statistical method

Odds ratios (ORs) for developing COPD were estimated using conditional logistic regression analysis with 95% confidence intervals (CIs). ORs were calculated by adherence to a Mediterranean diet using a three-grade scale as in the study by Sjögren et al. (27), categorizing individuals as low-adherent (0–2 points), medium-adherent (3–5 points), and high-adherent (6–8 points). In the calculation of ORs, the lowest adherence was used as the reference. Adjustments were made to models based on a scientific and clinical rationale. To evaluate the importance of the different food components contained in the modified MDS, ORs were calculated based on the scoring of each component, i.e., above or below the median intake using conditional logistic regression. In a sensitivity analysis, the dietary habits stated at the first available visit in the VIP at least 10 years before the diagnosis of COPD were analyzed. All final statistical calculations were performed using Stata (version 13.1, StataCorp, Lakeway, TX, USA), and significance was set at a p-value <0.05. All study participants provided informed consent through the original data collection in the VIP cohort, and the procedures of this study were approved by the Regional Ethical Review Board at the University Hospital, Umeå, in accordance with the Declaration of Helsinki and the Swedish Law on personal data act (PuL).

Results

Descriptive data for the 370 individuals who later visited the Department of Medicine, University Hospital, Umeå with a diagnosis of COPD (referred to as cases) and the 1432 matched controls are presented in Table 1. Data from the VIP were collected on average 11.1 years (SD 6.6 years) before the first registered medical visit. There was a significant difference between the cases and controls with respect to education level as well as living arrangements (p = 0.001 and p < 0.001, respectively).

Table 1. Descriptive data for 370 individuals who developed COPD and 1432 controls matched for age, sex, and smoking habit.^a

Variable	Cases (n = 370)	Controls (n = 1432)	OR (95% CI)^b
Age at screening, years, mean (min–max)	55 (30–61)	55 (30–61)
Time to first visit to Respiratory Department, Umeå, mean (SD)	11.1 (4.5)
Female, n (%)	213 (58)	817 (57)
Current smokers, n (%)	269 (73)	1040 (73)
Ex-smokers, n (%)	83 (22)	322 (22)
Never smokers, n (%)	18 (5)	70 (5)
Body mass index, kg/m^2, mean (SD)	25.7 (4.9)	26.0 (3.9)	0.98 (0.95–1.01)
Higher education, n (%)	42 (11)	258 (18)	0.56 (0.39–0.80)
Living alone, n (%)	86 (23)	224 (16)	1.69 (1.26–2.26)
Low adherence to MDS, n (%)	116 (31)	327 (23)	–
Medium adherence to MDS, n (%)	216 (58)	896 (63)	–
High adherence to MDS, n (%)	38 (10)	209 (15)	–

^a Smoking status was defined at the time of examination in the Västerbotten Intervention Programme.

^b Unadjusted OR with 95% CI for the development of chronic obstructive lung disease calculated using conditional logistic regression analyses.

OR: odds ratio; CI: confidence interval; SD: standard deviation; MDS: Mediterranean diet score.

Individuals with higher education scored higher on the modified MDS (mean [standard deviation (SD)] 4.2 [1.5]) than those with a lower education level (3.6 [1.6]; p < 0.001). Individuals living alone scored lower on the modified MDS (mean [SD] 3.5 [1.6]) than those living with someone (3.7 [1.6]; p < 0.05).

In unadjusted analyses, individuals with medium and high adherence to the modified MDS exhibited significant lower ORs in a dose-response manner to develop COPD than those with low adherence (Table 2). The significance and dose-response relationship remained in differently adjusted models. In the sensitivity analysis restricted to the first available visit in the VIP at least 10 years (243 cases; mean time between assessment of dietary habits and diagnosis 14.4 years; range 10.0–21.8 years) before the diagnosis of COPD, similar findings were observed (Supplementary Table S2).

Table 2. Mediterranean diet score (MDS) and odds ratios (95% confidence interval) for developing chronic obstructive pulmonary disease among 370 cases and 1431 controls matched for age, sex, and smoking habit.^a

	Crude	Adjusted, model 1^b	Adjusted, model 2^c	Adjusted, model 3^d	Adjusted, model 4^e
Cases (n)	370	370	367	286	284
Modified MDS
Low adherence 0–2	1 (reference)	1 (reference)	1 (reference)	1 (reference)	1 (reference)
Medium adherence 3–5	0.68 (0.53–0.88)	0.71 (0.55–0.92)	0.73 (0.56–0.95)	0.67 (0.50–0.91)	0.73 (0.53–0.99)
High adherence 6–8	0.50 (0.33–0.76)	0.54 (0.36–0.82)	0.56 (0.37–0.86)	0.52 (0.31–0.87)	0.59 (0.35–0.97)

^a Smoking status was defined at the time of examination in the Västerbotten Intervention Programme.

^b Adjusted for higher education.

^c Adjusted for higher education and living alone.

^d Adjusted for numbers of cigarettes smoked per day.

^e Adjusted for higher education, living alone and numbers of cigarettes smoked per day.

Bold font indicate statistically significant result p<0.05.

In the analysis of the individual scores for food items comprising the modified MDS and the odds to develop COPD, significant effects could only be observed for the consumption of fruit (Table 3).

Table 3. Consumption of different food groups and odds ratios (95% confidence intervals) for developing chronic obstructive pulmonary disease.

Single MDS component	OR (95% CI)
Intake above the median
(MUFA + PUFA): SFA	0.81 (0.64–1.03)
Vegetables and potatoes	0.81 (0.64–1.02)
Fruit	0.71 (0.57–0.90)
Wholegrain cereals	0.88 (0.70–1.11)
Fish	0.88 (0.70–1.10)
Intake below the median
Meat and meat products	0.94 (0.75–1.18)
Dairy	0.88 (0.70–1.11)
Intake within limits
Alcohol	0.84 (0.62–1.15)

MUFA: monounsaturated fat; PUFA: polyunsaturated fat; SFA: saturated fat; OR: odds ratio; CI: confidence interval.

Bold font indicate statistically significant result p<0.05.

Discussion

In this study, adherence to a Mediterranean-like diet was shown to be inversely associated with the development of COPD. Moreover, a higher adherence was associated with a lower odds for developing COPD. Adherence to a Mediterranean diet was associated with higher education and not living alone. Our findings also highlight the importance of studying the diet as a whole, since only fruit had a significant effect on the odds when analyzing individual components contained in the modified MDS score in this population.

Previous prospective studies examining dietary patterns and the risk of developing COPD have focused on either prudent/healthy or Western diets (9,10,28,29). In these studies, a prudent/healthy diet lowered the risk of developing COPD, whereas a Western diet increased the risk. In a meta-analysis on this topic, a healthy/prudent diet was concluded to decrease the risk of developing COPD (OR 0.55, 95% CI 0.46–0.66), whereas a Western diet increased the risk (OR 2.12, 95% CI 1.64–2.74) (30). Since healthy/prudent dietary patterns have characteristics partly overlapping with a Mediterranean diet (12,31), our results concur with the findings of these studies. In the analyses of individual components contained in the modified MDS score, only fruit intake was significantly associated with a decreased odds of developing COPD. This finding highlights the complexity of assessing dietary habits and the value of studying the effect of the diet beyond isolated nutrients. The phenomenon of observing effects from the resulting MDS and not from isolated components of the score was previously described by Trichopoulou et al. in the original MDS in 1995 (13).

As described by Romieu and Trenga (32), there might be several explanatory theories regarding the effects of diet on the development of COPD. For example, the highly oxidative environment of the lung could be protected by antioxidants provided by diet. Moreover, the anti-inflammatory properties of omega-3 fatty acids, as well as the pro-inflammatory properties of omega-6 fatty acids, might be of importance, since inflammation plays an important role in COPD (33). Vegetables and fruits are high in antioxidants, and fish and nuts are high in omega-3 fatty acids, whereas red meat is high in omega-6 fatty acids. The protective effect of vegetables, fruits, and fish and the deleterious effect of red meat in the development of COPD have been demonstrated in several studies (5,6,11,34–40). The effect of processed meat such as sausage and cured meat has also garnered attention due to their high content of food additives such as nitrites, and it has been proposed that these additives may cause oxidative stress in the lung (39). In a recent meta-analysis, it was concluded that each 50 grams of increase in the intake of processed meat per week was associated with an 8% increased risk of developing COPD (41).

The main strength of this study is the use of comprehensive and validated prospective data collected equally among thoroughly matched controls and cases in the VIP cohort. Although the diagnosis of COPD when using ICD coding for identification of the cases was not verified by spirometry, individuals with COPD who visit a specialist care unit usually suffer from advanced disease, mainly spirometric grade GOLD 3 or 4, according to the GOLD criteria (1). A misdiagnosis of the cases is therefore unlikely, but it may introduce a bias toward a more severe disease among the cases. Another limitation with the lack of spirometry in the studied population is that it cannot be ruled out that there are controls that do have COPD. However, this should mainly introduce a statistical type-II error and will decrease the power and possibilities to find a real difference. Since COPD is a slowly progressing disease, and patients are most often not diagnosed until symptoms are evident, slowly emerging symptoms may have already affected the individuals’ lifestyle, including dietary habits, at the time of participation in the VIP examinations. Nonetheless, we were able to verify our findings in the sensitivity analyses, which were restricted to dietary habits more than 10 years before the diagnosis. Last, although we have adjusted for education level and living alone, we cannot rule out that there still is residual confounding, or other unmeasured factors related to lifestyle that may contribute to the results.

Conclusion

Among the population of the Västerbotten County, a Mediterranean-like diet is inversely associated with the development of COPD, defined as a visit to a specialist health-care clinic with a registered diagnosis code for COPD. Adherence to this diet was associated with higher education level and not living alone. Taken together, the present study suggests that healthy dietary habits might be of value to investigate further as a protective factor against the development of COPD, especially among individuals with lower education and those living alone.

Declaration of interest

The authors report no conflict of interest.

Acknowledgments

We gratefully acknowledge the Northern Sweden Medical Research Bank and Västerbotten County for collecting and providing access to the data. Construction of the diet database has been supported by the Swedish Research Council for Health, Working Life and Welfare (FORTE) and The Swedish Research Council.