Participants: This population-based cross-sectional study was performed from May to October 2011 in Tehran, Iran. The detailed report on the sampling method and data collection procedure has been published previously [23]. In total, 300 elderly people (150 men and 150 women) with aged ≥ 55 years were enrolled by the use of cluster random sampling method in district 6 of Tehran. The head of each cluster was selected based on a ten-digit postal code. To ensure the homogeneity of our sample, individuals whose potential cause of sarcopenia were factors other than aging were not invited. Indeed, people who were susceptible to sarcopenia due to secondary causes [24], including those who were unable to move, subjects with artificial limbs or limb prostheses, or individuals with debilitating diseases that predispose the person to sarcopenia (e.g malignancy, organ failure) were not included in the study.
Dietary intake assessment: Usual dietary intakes of the study participants was assessed using a 117-item Food Frequency Questionnaire (FFQ); the validity and reliability of this questionnaire was reported in previous studies [23, 25]. The questionnaire consisted of a list of foods with a specific portion size. Participants were able to report their consumption frequency based on daily, weekly or monthly basis for each food item. The questionnaire was filled by a trained nutritionist through face-to-face interview. After completing the FFQ, the frequency of each food item was converted to grams per day considering the household measures of portion sizes. Daily energy and nutrients intake of each participant was calculated by using Nutritionist IV software with a modified food composition database based on the US Department of Agriculture.
Construction of Dietary Inflammatory Index: The method of Shivappa et al was used to compute the Dietary Inflammatory Index (DII) [14]. DII score was calculated using 29 dietary parameters because some parameters suggested in the original scoring method were not available in our dataset in Iran. Dietary parameters included: energy, carbohydrate, fat, protein, fiber, cholesterol, mono-unsaturated fatty acids (MUFAs), polyunsaturated fatty acids (PUFAs), saturated fats (SFAs), vitamin B12, pyridoxine (B6), folic acid, niacin (B3), riboflavin (B2), thiamin (B1), vitamin A, C, D, E, b-carotene, zinc, selenium, magnesium, iron, caffeine, pepper, onion, garlic and green/black tea. Residual method was used to obtain the energy-adjusted amounts for all nutrients [26]. Then, to get z-score, the “standard global mean” was subtracted from the quantity of food and it was divided by the “global standard deviation”. Standard global means and SDs for each dietary item were obtained from Shivappa et al [14]. To decrease skewness, this value was converted to a centered percentile score. Then, this score was multiplied by the effect score for each of the food items obtained from Shivappa et al [14]. Finally, to compute a total DII score for each participant, we summed the DII score obtained from all individual dietary parameters. The most negative score implies the maximum anti-inflammatory diet, while the most positive score implies the maximum pro-inflammatory diet.
Assessment of Sarcopenia: Based on the European Working Group on Sarcopenia (EWGSOP) definition [24], sarcopenia was determined by considering the combination of both low muscle mass and low muscle function (either strength or performance). The muscle mass was measured as the ratio of an individual’s total lean mass of legs and arms (also named Appendicular Skeletal Muscle or ASM) [27] to their squared height (ASM/height²). ASM was calculated with a DXA scanner (Discovery W S/N 84430). Based on EWGSOP, low muscle mass was considered as the amount of muscle mass less than 5.45 kg/m² for women and 7.26 kg/m² for men [24].
A handgrip test was used to measure muscle strength. The handgrip test was assessed by a pneumatic instrument that is a squeeze bulb dynamometer (c7489-02 Rolyan) calibrated in pound per square inch (psi). The handgrip strength (maximum voluntary contractions) was calculated three times for each right and left hand with a 30-second rest between measurements. Then, the average of these three measurements for each hand was calculated. Finally, the average number was obtained based on the sum of mean of both hands and this was considered as muscle strength. Sex and age-specific cutoff points recommended by Merkies et.al was then used to identify low muscle strength [28]. To measure muscle performance, a 4-Meter walk gait speed test was applied [24]. Participants who had gait speeds less than 0.8 m/s were recognized as low muscle performance [24].
Assessment of other variables: Information about general characteristics of participant including age, sex, socio-economic status, medical history, medication use, smoking habits, and alcohol consumption were collected by a pre-tested questionnaire. The physical activity level in this study was examined by a trained interviewer using the short form of the International Physical Activity Questionnaire (IPAQ), its validity has previously been examined [25]. Measures of physical activity for each participant was expressed as metabolic equivalent-hour per week (MET-h/week) based on IPAQ’s guideline [29]. Weight was measured using a digital scale while participants were minimally clothed. Height was measured by a wall tape meter in standing position without shoes. Waist circumference was measured in the middle of the lower rib margin and iliac crest while participants were stand up and normally breathe. Weight (kg) divided by height squared (m2) was used to calculate body mass index (BMI).
Statistical analysis: Subjects were classified according to the tertiles of DII score. General characteristics of study participants across tertiles of DII score were compared using Chi-square for categorical variables and ANOVA for continuous variables. Age- sex-, and energy-adjusted dietary intakes of participants were computed using General Linear Model and compared using ANCOVA across tertile categories of DII score. Multivariable logistic regression was conducted to find the relationship between inflammatory potential of the diet and odds of sarcopenia. In these analyses, several confounders were controlled. First, the association was adjusted for age (continuous), sex (male/female) and energy intake (kcal/d). Then, further controlling was done for physical activity (MET-h/wk), smoking (yes/no), alcohol consumption (yes/no), medication use (statin, corticosteroid, estrogen, testosterone), and positive history of chronic disease means subjects that affected by one of the following diseases: asthma, arthritis, myocardial infarction, cerebrovascular accident, and diabetes. In all these analyses, the bottom tertile was considered as the reference category and the odds ratio for sarcopenia in other categories was calculated. To evaluate the linear trend across tertiles of the DII score, the tertile categories were considered as an ordinal variable in the models. All analyses were done using SPSS (version 26). P-values were considered significant at <0.05.