Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Feb 1.
Published in final edited form as: Am J Prev Med. 2017 Feb;52(2):237–248. doi: 10.1016/j.amepre.2016.10.041

Dietary Protein Sources and All-Cause and Cause-Specific Mortality: The Golestan Cohort Study in Iran

Maryam S Farvid 1,2,3, Akbar F Malekshah 2,4, Akram Pourshams 2,5, Hossein Poustchi 4,5, Sadaf G Sepanlou 2,4, Maryam Sharafkhah 2,5,6, Masoud Khoshnia 7, Mojtaba Farvid 8, Christian C Abnet 9, Farin Kamangar 10, Sanford M Dawsey 9, Paul Brennan 11, Paul D Pharoah 12, Paolo Boffetta 13, Walter C Willett 1,14,15, Reza Malekzadeh 2,4,5
PMCID: PMC5360102  NIHMSID: NIHMS835649  PMID: 28109460

Abstract

Introduction

Dietary protein comes from foods with greatly different compositions that may not relate equally with mortality risk. Few cohort studies from non-Western countries have examined the association between various dietary protein sources and cause-specific mortality. Therefore, the associations between dietary protein sources and all-cause, cardiovascular disease, and cancer mortality were evaluated in the Golestan Cohort Study in Iran.

Methods

Among 42,403 men and women who completed a dietary questionnaire at baseline, 3,291 deaths were documented during 11 years of follow up (2004–2015). Cox proportional hazards models estimated age-adjusted and multivariate-adjusted hazard ratios (HRs) and 95% CIs for all- cause and disease-specific mortality in relation to dietary protein sources. Data were analyzed from 2015 to 2016.

Results

Comparing the highest versus the lowest quartile, egg consumption was associated with lower all-cause mortality risk (HR=0.88, 95% CI=0.79, 0.97, ptrend=0.03). In multivariate analysis, the highest versus the lowest quartile of fish consumption was associated with reduced risk of total cancer (HR=0.79, 95% CI=0.64, 0.98, ptrend=0.03) and gastrointestinal cancer (HR=0.75, 95% CI=0.56, 1.00, ptrend=0.02) mortality. The highest versus the lowest quintile of legume consumption was associated with reduced total cancer (HR=0.72, 95% CI=0.58, 0.89, ptrend=0.004), gastrointestinal cancer (HR=0.76, 95% CI=0.58, 1.01, ptrend=0.05), and other cancer (HR=0.66, 95% CI=0.47, 0.93, ptrend=0.04) mortality. Significant associations between total red meat and poultry intake and all- cause, cardiovascular disease, or cancer mortality rate were not observed among all participants.

Conclusions

These findings support an association of higher fish and legume consumption with lower cancer mortality, and higher egg consumption with lower all-cause mortality.

INTRODUCTION

Red meat and processed meat consumption may contribute to increased risk of several chronic diseases such as diabetes,1 cardiovascular disease (CVD)1, and certain cancers.2–6 Prospective studies have noted increased mortality risk among men and women with high total and processed red meat consumption,7,8 whereas poultry and fish are often considered healthier substitutes for red meat. Current evidence supports the relationship between fish consumption and lower all-cause or coronary heart disease (CHD) mortality risk,9,10 whereas the role of fish intake in cancer mortality is unclear.11–13 Findings from prospective studies relating poultry to mortality risk are inconsistent.12,14,15 The roles of other dietary protein sources such as eggs or legumes in mortality risk are less clear.16,17 Further, most of the results in relation to red meat consumption and mortality were from Western countries, where high red meat consumption is common. However, Iranian people consume less red meat: 9.2 kg per capita compared with 65.3 and 58.3 kg per capita among American or European populations, respectively.18 Therefore, to examine the association between red meat and mortality and to show the effect of other protein sources on mortality, the association of red meat, poultry, fish, eggs, and legumes consumption with mortality risk from all and specific causes was investigated, using data from a prospective cohort study of adult men and women in Golestan, Iran.

METHODS

Study Population

The Golestan Cohort Study is an ongoing prospective population- based study established in 2004 in Iran with enrollment of 50,045 participants (21,234 men and 28,811 women) aged 36–85 years, without previous history of upper gastrointestinal cancer, from Gonbad city and 326 villages in Golestan province. The study design is described in further detail elsewhere.19 This analysis included participants who had data on dietary intake (n=49,112) at baseline. Participants were excluded owing to loss of follow up (n=63); extreme total energy intake (<600 or >4,200 kcal/day, n=436); or prior diagnosis of chronic disease, including cancer, diabetes, CHD, or stroke (n=6,210), leaving 42,403 participants (18,261 men and 24,142 women).

This study was approved by the IRBs of the Digestive Disease Research Center of Tehran University of Medical Sciences, the U.S. National Cancer Institute, the WHO International Agency for Research on Cancer, and the Harvard T.H. Chan School of Public Health. All participants gave written informed consent before enrollment.

Measures

At enrollment, dietary intake was measured by face-to-face interviews using a validated, 116-item food frequency questionnaire specifically designed for this population.20 Total red meat items listed on the food frequency questionnaire included unprocessed (beef or lamb, hamburger) and processed (sausage) red meat; poultry included chicken; fish included tuna, stellate sturgeon, carp, smoked fish, salted fish, and other fish; and legumes included soybeans, beans, lentils, peas, and split peas. Participants were asked about the frequency of food item consumption per day, week, month, or year. The standard serving sizes for these food items were 85 g for cooked beef, lamb, hamburger, chicken, and fish; 45 g for sausage; 100 g for cooked beans, lentils, peas, split peas, and soybeans; and 54 g for eggs. Iranian21 and U.S. Department of Agriculture22 databases were used to calculate daily energy intake.

Data on lifestyle variables and medical history were obtained through face-to-face interviews using a general questionnaire at the beginning of the study. Participants were asked about gender, age, ethnicity, residential history, occupation, education, smoking habits, opium use, alcohol consumption, indicators of SES, and medical history including self-reports of medically diagnosed diabetes mellitus, heart disease, stroke, and cancer. Height and weight were then measured. Blood pressure in both arms was also measured twice at a 10-minute interval. Occupational physical activity was measured in four levels: Level 1 activities defined as sedentary work mostly done while sitting (e.g., driving); Level 2 activities defined as standing or occasional walking (e.g., teaching); Level 3 activities defined as mainly indoor activities causing a mild increase in heart rate and sweating (e.g., housekeeping); and Level 4 activities defined as those causing a significant increase in heart rate and sweating usually performed outdoors (e.g., farming).23 Using multiple correspondence analysis, the wealth score was created based on occupation, house ownership, house structure, house size, having a bath in the residence, as well as a personal car, motorbike, black/white TV, color TV, refrigerator, freezer, vacuum, and washing machine.24

Deaths were reported by family members, friends, or local health workers during annual telephone calls. When a death was reported, a physician visited the house to complete a validated verbal autopsy questionnaire by interviewing the next of kin.25 Relevant medical documents including medical charts, radiography and pathology reports, and hospital discharge reports were gathered from the hospitals or pathology centers. All collected documents were reviewed, and the cause of death was coded according to the ICD-10. In this analysis, deaths were classified as due to CVD (ICD-I00–I99); CHD (ICD-I20–I52); stroke (ICD- I60–I69); cancer (ICD-C00–D48); gastrointestinal cancers (ICD- C15–26); and other cancers (ICD-C00–14, ICD-C30–97, and ICD- D00–48). If a final diagnosis could not be determined for any reason, the cause of death was classified as “unknown.”

Statistical Analysis

Data were analyzed from 2015 to 2016. Person-years of follow up were calculated from the date of completed questionnaires at baseline until death, or last follow-up date (until June 1, 2015), which ever came first. Participants were divided into quintiles according to food group intake. Because of the low consumption of fish and eggs, participants who did not eat fish or eggs were assigned to the first category, and others were divided into tertiles. Cox proportional hazards regression was used to estimate age- adjusted and multivariate-adjusted hazard ratios (HRs) and 95% CIs for all-cause and disease-specific mortality in relation to total red meat, poultry, fish, egg, and legume consumption. Multivariate models were adjusted for age, gender, place of residence, marital status, educational level, ethnicity, cigarette smoking, opium use, BMI, systolic blood pressure, family history of cancer, occupational physical activity, medication, wealth score, alcohol consumption, and total energy intake (footnotes in Tables 2–4 list categorizations). Tests for linear trend were carried out by assigning a median value for each quintile (or quartile) and modeling this as a continuous variable in the model. Missing covariate data, which included five participants for BMI and eight participants for systolic blood pressure, were replaced with median values. To address whether the observed associations were independent of other dietary intake, the analysis additionally controlled for fruit and vegetable or total grain intake. Furthermore, a previous study using data from the Golestan Cohort Study indicated that high consumption of low-fat dairy was associated with reduced mortality risk,26 so the associations were evaluated after additional adjustment for intake of low-fat dairy foods. By including a cross-product interaction term for gender, age, BMI, smoking, or wealth score, and dietary protein sources in the multivariate-adjusted model, this study examined whether the associations between dietary protein sources and all-cause mortality were modified by other mortality risk factors.27–29 The effect of substituting one serving/day of poultry, fish, legumes, or eggs for one serving/day of total red meat was estimated by including these food items simultaneously as continuous variables in the multivariate-adjusted model. The HRs and 95% CIs for the substitution effect were estimated from the difference between the regression coefficients, variances, and covariance.30 Stata, version 12, was used for all analyses. All p-values were two-sided.

Table 2.

Risk for All-cause Mortality According to Quintile of Dietary Protein Sources Consumption

Dietary protein sources Quintile of intake p trend 3 servings/week
1 2 3 4 5
Total red meat
 Median intake, serving/day 0.02 0.07 0.13 0.23 0.43
 No. of deaths/person-year 775/65,241 655/67,685 617/68,133 618/68,990 626/69,818
 Age-adjusted HR (95% CI)a 1 0.90 (0.81, 1.00) 0.89 (0.80, 0.99) 0.91 (0.82, 1.01) 0.95 (0.86, 1.06) 0.82 1.01 (0.94, 1.08)
 Multivariate-adjusted HR (95% CI)b 1 0.96 (0.86, 1.07) 0.98 (0.88, 1.09) 1.01 (0.90, 1.13) 1.04 (0.93, 1.17) 0.25 1.02 (0.96, 1.09)
Poultry
 Median intake, serving/day 0.11 0.33 0.54 0.78 1.33
 No. of deaths/person-year 706/70,172 654/68,732 614/67,769 617/67,029 700/66,165
 Age-adjusted HR (95% CI)a 1 0.95 (0.86, 1.06) 0.90 (0.81, 1.00) 0.89 (0.80, 0.99) 0.90 (0.81, 1.00) 0.06 0.97 (0.93, 1.00)
 Multivariate-adjusted HR (95% CI)b 1 0.97 (0.87, 1.08) 0.96 (0.86, 1.07) 0.98 (0.88, 1.10) 1.02 (0.91, 1.14) 0.54 1.01 (0.98, 1.05)
Fishc
 Median intake, serving/day 0 0.01 0.06 0.19
 No. of deaths/person-year 883/68,316 910/90,608 790/90,242 708/90,701
 Age-adjusted HR (95% CI)a 1 0.88 (0.81, 0.97) 0.82 (0.75, 0.90) 0.77 (0.69, 0.85) <0.0001 0.61 (0.49, 0.75)
 Multivariate-adjusted HR (95% CI)b 1 0.94 (0.86, 1.04) 0.93 (0.84, 1.02) 0.93 (0.83, 1.03) 0.32 0.89 (0.71, 1.12)
Eggsc
 Median intake, serving/day 0 0.06 0.18 0.48
 No. of deaths/person-year 905/69,165 804/82,907 798/95,280 784/92,515
 Age-adjusted HR (95% CI)a 1 0.88 (0.80, 0.97) 0.86 (0.79, 0.95) 0.89 (0.81, 0.98) 0.12 0.94 (0.86, 1.02)
 Multivariate-adjusted HR (95% CI)b 1 0.92 (0.84, 1.02) 0.90 (0.82, 0.99) 0.88 (0.79, 0.97) 0.03 0.91 (0.84, 0.99)
Legumes
 Median intake, serving/day 0.03 0.08 0.12 0.17 0.26
 No. of deaths/person-year 787/69,186 622/68,484 581/67,714 619/67,013 682/67,470
 Age-adjusted HR (95% CI)a 1 0.87 (0.78, 0.96) 0.82 (0.73, 0.91) 0.87 (0.79, 0.97) 0.85 (0.77, 0.94) 0.01 0.80 (0.66, 0.95)
 Multivariate-adjusted HR (95% CI)b 1 0.92 (0.83, 1.02) 0.89 (0.80, 0.99) 0.95 (0.85, 1.06) 0.94 (0.84, 1.04) 0.45 0.93 (0.77, 1.13)
Open in a new tab

Note: Bold values were statistically significant (p<0.05).

a

Age-adjusted model was adjusted for age (years).

b

Multivariate-adjusted model was adjusted for gender; age (years); ethnicity (Turkmen, Persian, others); education (illiterate, ≤5 years, 6–8 years, high school, academic); marital status (married, other); residency (rural, urban); smoking (never, former, current <20 p/year, current 20–30 p/year, current ≥30 p/year); opium use (user, non-user); alcohol (user, non-user); BMI (<18.5,18.5 to <20.0, 20.0 to <22.5, 22.5 to <25.0, 25.0 to <30.0, 30 to <35.0, ≥35.0); systolic blood pressure (quintile); occupational physical activity (sedentary, standing or occasional walking, causing mild increase in heart rate and sweating, causing significant increase in heart rate and sweating); family history of cancer (yes, no); wealth score (quintile); medication (using a fixed-dose combination pill consisting of aspirin, valsartan, atorvastatin, and hydrochlorothiazide; yes/no); and energy intake (quintile).

c

Quartile.

HR, hazard ratio.

Table 4.

Risk for Total Cancer, Gastrointestinal Cancer, and Other Cancer Mortality According to Quintile of Dietary Sources of Protein Consumption

Dietary protein sources Quintile of intake p trend 3 servings/week
1 2 3 4 5
Total red meat
 Total cancer
  Median intake, serving/day 0.02 0.07 0.13 0.23 0.43
  No. of deaths/person-year 191/65,241 193/67,685 153/68,133 156/68,990 166/69,818
  Age-adjusted HR (95% CI)a 1 1.07 (0.87, 1.31) 0.89 (0.72, 1.10) 0.93 (0.75, 1.14) 1.01 (0.82, 1.25) 0.85 1.05 (0.93, 1.18)
  Multivariate-adjusted HR (95% CI)b 1 1.09 (0.89, 1.34) 0.92 (0.73, 1.15) 0.96 (0.76, 1.20) 1.03 (0.82, 1.30) 1.00 1.04 (0.92, 1.17)
 Gastrointestinal cancer
  No. of deaths/person-year 108/65,241 110/67,685 98/68,133 92/68,990 101/69,818
  Age-adjusted HR (95% CI)a 1 1.09 (0.84, 1.42) 1.03 (0.78, 1.36) 0.99 (0.75, 1.31) 1.12 (0.85, 1.47) 0.57 1.06 (0.91, 1.23)
  Multivariate-adjusted HR (95% CI)b 1 1.12 (0.85, 1.47) 1.08 (0.81, 1.43) 1.05 (0.78, 1.41) 1.17 (0.87, 1.59) 0.43 1.05 (0.89, 1.22)
 Other cancer
  No. of deaths/person-year 83/65,241 83/67,685 55/68,133 64/68,990 65/69,818
  Age-adjusted HR (95% CI)a 1 1.04 (0.76, 1.41) 0.71 (0.51, 1.01) 0.84 (0.61, 1.17) 0.87 (0.63, 1.21) 0.33 1.04 (0.86, 1.26)
  Multivariate-adjusted HR (95% CI)b 1 1.05 (0.77, 1.44) 0.72 (0.51, 1.03) 0.85 (0.60, 1.20) 0.85 (0.60, 1.22) 0.34 1.03 (0.84, 1.25)
Poultry
 Total cancer
  Median intake, serving/day 0.11 0.33 0.54 0.78 1.33
  No. of deaths/person-year 200/70,172 169/68,732 152/67,769 159/67,029 179/66,165
  Age-adjusted HR (95% CI)a 1 0.87 (0.71, 1.07) 0.78 (0.64, 0.97) 0.81 (0.66, 1.00) 0.82 (0.67, 1.01) 0.10 0.94 (0.88, 1.01)
  Multivariate-adjusted HR (95% CI)b 1 0.88 (0.72, 1.08) 0.83 (0.67, 1.03) 0.90 (0.73, 1.11) 0.93 (0.75, 1.15) 0.75 0.99 (0.92, 1.06)
 Gastrointestinal cancer
  No. of deaths/person-year 110/70,172 111/68,732 92/67,769 90/67,029 106/66,165
  Age-adjusted HR (95% CI)a 1 1.04 (0.80, 1.35) 0.86 (0.65, 1.14) 0.83 (0.63, 1.09) 0.87 (0.67, 1.14) 0.16 0.94 (0.86, 1.03)
  Multivariate-adjusted HR (95% CI)b 1 1.05 (0.81, 1.38) 0.93 (0.70, 1.22) 0.94 (0.70, 1.24) 1.04 (0.78, 1.37) 0.95 1.00 (0.91, 1.10)
 Other cancer
  No. of deaths/person-year 90/70,172 58/68,732 60/67,769 69/67,029 73/66,165
  Age-adjusted HR (95% CI)a 1 0.66 (0.48, 0.92) 0.69 (0.50, 0.96) 0.79 (0.57, 1.08) 0.77 (0.56, 1.05) 0.36 0.95 (0.85, 1.06)
  Multivariate-adjusted HR (95% CI)b 1 0.67 (0.48, 0.93) 0.72 (0.52, 1.00) 0.85 (0.62, 1.17) 0.81 (0.58, 1.12) 0.61 0.97 (0.87, 1.09)
Fishc
 Total cancer
  Median intake, serving/day 0 0.01 0.06 0.19
  No. of deaths/person-year 240/68,316 253/90,608 203/90,242 163/90,701
  Age-adjusted HR (95% CI)a 1 0.90 (0.75, 1.07) 0.76 (0.63, 0.92) 0.64 (0.52, 0.78) <0.0001 0.38 (0.25, 0.59)
  Multivariate-adjusted HR (95% CI)b 1 0.96 (0.80, 1.14) 0.87 (0.72, 1.06) 0.79 (0.64, 0.98) 0.03 0.60 (0.38, 0.94)
 Gastrointestinal cancer
  No. of deaths/person-year 147/68,316 160/90,608 117/90,242 85/90,701
  Age-adjusted HR (95% CI)a 1 0.94 (0.75, 1.17) 0.73 (0.58, 0.94) 0.56 (0.43, 0.73) <0.0001 0.26 (0.15, 0.47)
  Multivariate-adjusted HR (95% CI)b 1 1.03 (0.82, 1.29) 0.88 (0.68, 1.12) 0.75 (0.56, 1.00) 0.02 0.49 (0.26, 0.89)
 Other cancer
  No. of deaths/person-year 93/68,316 93/90,608 86/90,242 78/90,701
  Age-adjusted HR (95% CI)a 1 0.83 (0.62, 1.11) 0.81 (0.60, 1.09) 0.76 (0.56, 1.03) 0.17 0.64 (0.34, 1.20)
  Multivariate-adjusted HR (95% CI)b 1 0.85 (0.64, 1.14) 0.87 (0.64, 1.18) 0.84 (0.60, 1.17) 0.52 0.80 (0.40, 1.58)
Eggsc
 Total cancer
  Median intake, serving/day 0 0.06 0.18 0.48
  No. of deaths/person-year 235/69,165 189/82,907 239/95,280 196/92,515
  Age-adjusted HR (95% CI)a 1 0.79 (0.65, 0.96) 0.98 (0.82, 1.17) 0.84 (0.69, 1.02) 0.35 0.93 (0.79, 1.09)
  Multivariate-adjusted HR (95% CI)b 1 0.81 (0.67, 0.99) 0.99 (0.83, 1.20) 0.81 (0.67, 0.99) 0.17 0.89 (0.75, 1.05)
 Gastrointestinal cancer
  No. of deaths/person-year 143/69,165 113/82,907 136/95,280 117/92,515
  Age-adjusted HR (95% CI)a 1 0.79 (0.62, 1.01) 0.95 (0.75, 1.20) 0.86 (0.67, 1.09) 0.56 0.94 (0.76, 1.16)
  Multivariate-adjusted HR (95% CI)b 1 0.82 (0.64, 1.05) 0.97 (0.76, 1.23) 0.84 (0.65, 1.08) 0.39 0.91 (0.73, 1.13)
 Other cancer
  No. of deaths/person-year 92/69,165 76/82,907 103/95,280 79/92,515
  Age-adjusted HR (95% CI)a 1 0.79 (0.58, 1.07) 1.03 (0.77, 1.36) 0.82 (0.61, 1.11) 0.44 0.91(0.71, 1.17)
  Multivariate-adjusted HR (95% CI)b 1 0.80 (0.59, 1.09) 1.03 (0.77, 1.37) 0.78 (0.57, 1.07) 0.26 0.86 (0.66, 1.12)
Legumes
 Total cancer
  Median intake, serving/day 0.03 0.08 0.12 0.17 0.26
  No. of deaths/person-year 232/69,186 178/68,484 136/67,714 161/67,013 152/67,470
  Age-adjusted HR (95% CI)a 1 0.84 (0.69, 1.02) 0.64 (0.52, 0.79) 0.77 (0.63, 0.94) 0.65 (0.53, 0.79) <0.0001 0.47 (0.32, 0.68)
  Multivariate-adjusted HR (95% CI)b 1 0.88 (0.73, 1.08) 0.70 (0.57, 0.87) 0.84 (0.68, 1.03) 0.72 (0.58, 0.89) 0.004 0.57 (0.39, 0.84)
 Gastrointestinal cancer
  No. of deaths/person-year 136/69,186 112/68,484 79/67,714 94/67,013 88/67,470
  Age-adjusted HR (95% CI)a 1 0.91(0.71, 1.17) 0.64 (0.49, 0.85) 0.77 (0.59, 1.00) 0.63 (0.48, 0.83) 0.001 0.43 (0.27, 0.70)
  Multivariate-adjusted HR (95% CI)b 1 0.96 (0.75, 1.24) 0.72 (0.54, 0.95) 0.86 (0.66, 1.13) 0.76 (0.58, 1.01) 0.05 0.60 (0.37, 1.00)
 Other cancer
  No. of deaths/person-year 96/69,186 66/68,484 57/67,714 67/67,013 64/67,470
  Age-adjusted HR (95% CI)a 1 0.74 (0.54, 1.01) 0.64 (0.46, 0.89) 0.76 (0.55, 1.04) 0.66 (0.48, 0.91) 0.03 0.53 (0.30, 0.94)
  Multivariate-adjusted HR (95% CI)b 1 0.78 (0.57, 1.06) 0.68 (0.49, 0.95) 0.80 (0.58, 1.10) 0.66 (0.47, 0.93) 0.04 0.52 (0.29, 0.96)
Open in a new tab
a

Age-adjusted model was adjusted for age (years).

b

Multivariate-adjusted model was adjusted for gender; age (years); ethnicity (Turkmen, Persian, others); education (illiterate, ≤5 years, 6–8 years, high school, academic); marital status (married, other); residency (rural, urban); smoking (never, former, current <20 p/year, current 20–30 p/year, current ≥30 p/year); opium use (user, non-user); alcohol (user, non-user); BMI (<18.5,18.5 to <20.0, 20.0 to <22.5, 22.5 to <25.0, 25.0 to <30.0, 30 to <35.0, ≥35.0); systolic blood pressure (quintile); occupational physical activity (sedentary, standing or occasional walking, causing mild increase in heart rate and sweating, causing significant increase in heart rate and sweating); family history of cancer (yes, no); wealth score (quintile); medication (using a fixed-dose combination pill consisting of aspirin, valsartan, atorvastatin, and hydrochlorothiazide; yes/no); and energy intake (quintile).

c

Quartile.

HR, hazard ratio.

RESULTS

During 339,867 person-years of follow up (2004–2015; total, 11 years; median, 8.1 years), 3,291 deaths were documented (1,938 deaths among men and 1,353 deaths among women) in the Golestan Cohort Study. A total of 1,467 deaths were from CVD (764 CHD deaths, 507 stroke deaths, and 196 other CVD deaths) and 859 deaths were caused by cancer (509 gastrointestinal cancer deaths and 350 other cancer deaths). Mean consumption of total red meat was 0.19 serving/day, which included 0.04 serving/day of processed red meat. Participants with higher total red meat intake were more likely to be male, Turkmen, married, and smokers, and to have higher BMI. Also, they were less likely to live in rural areas and perform higher levels of occupational physical activity (Table 1). The percentage of energy intake from each food group according to tertiles of total red meat intake was used to show changes in dietary habits by increasing consumption of total red meat: the higher total red meat intake, the lower percentage of energy intake from grains (Appendix Figure 1, available online).

Table 1.

Characteristics of Participants According to Quintile of Total Red Meat Consumption

Potential risk factors Total red meat
Q1 (n=8,480) Q2 (n=8,481) Q3 (n=8,480) Q4 (n=8,481) Q5 (n=8,481)
M±SD
 Total red meat intake, serving/day 0.02±0.01 0.07±0.02 0.14±0.02 0.23±0.04 0.52±0.30
 Poultry intake, serving/day 0.8±0.8 0.7±0.6 0.7±0.6 0.7 ±0.6 0.6±0.6
 Fish intake, serving/day 0.06±0.13 0.08±0.15 0.09±0.18 0.10±0.15 0.12±0.19
 Egg intake, serving/day 0.14±0.20 0.18±0.22 0.21±0.25 0.23 ±0.25 0.28±0.31
 Legume intake, serving/day 0.13±0.11 0.13 ±0.10 0.13±0.10 0.14±0.10 0.15±0.13
 Total dairy food intake, serving/day 0.9±0.7 1.2±0.8 1.3±0.8 1.5±0.9 1.7±1.0
 Total grain intake, serving/day 12.1±5.0 12.4±4.8 12.7±4.8 12.9±4.8 13.4±4.9
 Total fruit intake, serving/day 1.1±1.0 1.6±1.3 1.9±1.5 2.1±1.7 2.7±1.9
 Total vegetable intake, serving/day 1.1±0.8 1.3±0.8 1.4±0.9 1.6±0.9 1.8±1.1
 Total energy intake, kcal/day 1,891±542 2,060±525 2,174±528 2,265±525 2,471±565
 Age (years) 53.1±9.2 52.0±8.8 51.4±8.7 51.0±8.6 50.6±8.4
 BMI (kg/m2) 25.6±5.4 26.4±5.4 26.7±5.4 26.7 ±5.3 26.9±5.3
Percentage
 Gender (male) 37 39 42 45 53
 Ethnicity (Turkmen) 54 77 81 82 82
 Marital status (married) 85 88 89 90 91
 Current smokers 12 12 14 15 18
 Opium use 18 15 16 16 17
 Place of residence (rural) 89 84 81 76 73
Occupational physical activity, %
 Level 1 3 3 2 2 3
 Level 2 3 4 5 7 8
 Level 3 51 57 58 56 51
 Level 4 43 36 35 35 38
Open in a new tab

Higher intake of total red meat as well as poultry, fish, and legumes was not associated with all-cause mortality risk (Table 2). The wealth score was responsible for most of the differences in HRs between age-adjusted and multivariate-adjusted models (data not shown). Higher egg consumption was associated with a lower all-cause mortality risk (HR [highest versus lowest] =0.88, 95% CI=0.79, 0.97, ptrend=0.03). Similar point estimates were observed after additional adjustment for fruit and vegetable, total grain food, or low-fat dairy food intake (data not shown). With mutual adjustment for dietary protein sources, a non-significant positive association for total red meat intake was observed (HR=1.08, 95% CI=0.96, 1.22, ptrend=0.08). The association of egg intake was not changed materially (HR=0.88, 95% CI=0.80, 0.98, ptrend=0.045). Significant associations with poultry, legume, or fish intake were not noted.

Furthermore, substituting one serving/day of legumes for one serving/day of total red meat was associated with a lower risk of all-cause mortality (HR=0.68, 95% CI=0.50, 0.93). Substituting one serving/day of eggs for one serving/day of total red meat was associated with a lower risk of all-cause mortality (HR=0.65, 95% CI=0.43, 0.98). Further, substituting one serving/day of fish for one serving/day of total red meat was associated with lower risk of all-cause mortality (HR=0.63, 95% CI=0.47, 0.84). Substituting one serving/day of poultry for one serving/day of total red meat was not significantly associated with risk of all-cause mortality (data not shown).

Total red meat intake was not associated with CVD, CHD, or stroke mortality risk. After additional adjustment for low-fat dairy food consumption, higher red meat consumption was somewhat associated with increased CHD mortality (HR [highest versus lowest] = 1.29, 95% CI=1.00, 1.66, ptrend=0.14), but not CVD mortality (HR=1.12, 95% CI=0.94, 1.34, ptrend=0.08) or stroke mortality risk (HR=1.05, 95% CI=0.78, 1.42, ptrend=0.25). No significant association was observed between poultry, fish, egg, or legume intake and CVD, CHD, or stroke mortality risk (Table 3).

Table 3.

Risk for Cardiovascular Disease, Coronary Heart Disease, and Stroke Mortality According to Quintile of Dietary Protein Sources Consumption

Dietary protein sources Quintile of intake p trend 3 servings/week
1 2 3 4 5
Total red meat
 Cardiovascular disease
  Median intake, serving/day 0.02 0.07 0.13 0.23 0.43
  No. of deaths/person-year 339/65,241 271/67,685 295/68,133 291/68,990 271/69,818
  Age-adjusted HR (95% CI)a 1 0.86 (0.73, 1.01) 0.99 (0.85, 1.16) 1.00 (0.86, 1.17) 0.96 (0.82, 1.13) 0.74 1.03 (0.94, 1.13)
  Multivariate-adjusted HR (95% CI)b 1 0.93 (0.79, 1.09) 1.11 (0.94, 1.31) 1.12 (0.95, 1.33) 1.07 (0.90, 1.28) 0.20 1.04 (0.95, 1.15)
 Coronary heart disease
  No. of deaths/person-year 153/65,241 145/67,685 170/68,133 143/68,990 153/69,818
  Age-adjusted HR (95% CI)a 1 1.01 (0.80, 1.27) 1.24 (1.00, 1.55) 1.07 (0.85, 1.34) 1.17 (0.94, 1.47) 0.19 1.08 (0.96, 1.21)
  Multivariate-adjusted HR (95% CI)b 1 1.09 (0.86, 1.38) 1.37 (1.09, 1.73) 1.16 (0.91, 1.48) 1.22 (0.95, 1.56) 0.29 1.04 (0.91, 1.18)
 Stroke
  No. of deaths/person-year 131/65,241 90/67,685 96/68,133 102/68,990 88/69,818
  Age-adjusted HR (95% CI)a 1 0.75 (0.58, 0.99) 0.86 (0.66, 1.12) 0.94 (0.73, 1.23) 0.85 (0.65, 1.11) 0.71 0.99 (0.83, 1.18)
  Multivariate-adjusted HR (95% CI)b 1 0.78 (0.60, 1.03) 0.96 (0.73, 1.26) 1.07 (0.81, 1.41) 1.00 (0.74, 1.34) 0.41 1.05 (0.89, 1.24)
Poultry
 Cardiovascular disease
  Median intake, serving/day 0.11 0.33 0.54 0.78 1.33
  No. of deaths/person-year 310/70,172 277/68,732 295/67,769 266/67,029 319/66,165
  Age-adjusted HR (95% CI)a 1 0.92 (0.78, 1.08) 0.98 (0.83, 1.15) 0.87 (0.73, 1.02) 0.92 (0.79, 1.08) 0.33 0.97 (0.93, 1.03)
  Multivariate, adjusted HR (95% CI)b 1 0.93 (0.79, 1.10) 1.04 (0.89, 1.22) 0.95 (0.80, 1.12) 1.03 (0.87, 1.21) 0.63 1.01 (0.96, 1.07)
 Coronary heart disease
  No. of deaths/person-year 164/70,172 131/68,732 162/67,769 140/67,029 167/66,165
  Age-adjusted HR (95% CI)a 1 0.82 (0.65, 1.03) 1.02 (0.82, 1.26) 0.87 (0.69, 1.08) 0.93 (0.75, 1.16) 0.79 0.99 (0.92, 1.06)
  Multivariate-adjusted HR (95% CI)b 1 0.83 (0.66, 1.04) 1.06 (0.85, 1.32) 0.92 (0.73, 1.15) 0.97 (0.77, 1.22) 0.90 1.00 (0.93, 1.08)
 Stroke
  No. of deaths/person-year 109/70,172 93/68,732 97/67,769 96/67,029 112/66,165
  Age-adjusted HR (95% CI)a 1 0.88 (0.66, 1.15) 0.91 (0.69, 1.19) 0.88 (0.67, 1.16) 0.89 (0.68, 1.16) 0.53 0.97 (0.89, 1.06)
  Multivariate-adjusted HR (95% CI)b 1 0.89 (0.68, 1.18) 0.98 (0.75, 1.30) 0.99 (0.75, 1.31) 1.06 (0.80, 1.39) 0.47 1.03 (0.94, 1.13)
Fishc
 Cardiovascular disease
  Median intake, serving/day 0 0.01 0.06 0.19
  No. of deaths/person-year 386/68,316 396/90,608 356/90,242 329/90,701
  Age-adjusted HR (95% CI)a 1 0.89 (0.77, 1.02) 0.86 (0.74, 0.99) 0.83 (0.71, 0.96) 0.04 0.73 (0.54, 0.99)
  Multivariate-adjusted HR (95% CI)b 1 0.95 (0.82, 1.09) 0.95 (0.82, 1.10) 0.96 (0.82, 1.13) 0.88 0.98 (0.70, 1.36)
 Coronary heart disease
  No. of deaths/person-year 198/68,316 192/90,608 188/90,242 186/90,701
  Age-adjusted HR (95% CI)a 1 0.83 (0.68, 1.01) 0.86 (0.70, 1.05) 0.89 (0.72, 1.08) 0.72 0.93 (0.61, 1.40)
  Multivariate-adjusted HR (95% CI)b 1 0.85 (0.70, 1.04) 0.90 (0.73, 1.10) 0.93 (0.75, 1.16) 0.94 1.02 (0.65, 1.59)
 Stroke
  No. of deaths/person-year 131/68,316 145/90,608 129/90,242 102/90,701
  Age-adjusted HR (95% CI)a 1 0.99 (0.78, 1.25) 0.95 (0.74, 1.21) 0.79 (0.61, 1.03) 0.05 0.58 (0.34, 1.00)
  Multivariate-adjusted HR (95% CI)b 1 1.11 (0.87, 1.41) 1.15 (0.89, 1.47) 1.03 (0.78, 1.36) 0.90 0.96 (0.54, 1.70)
Eggsc
 Cardiovascular disease
  Median intake, serving/day 0 0.06 0.18 0.48
  No. of deaths/person-year 396/69,165 374/82,907 348/95,280 349/92,515
  Age-adjusted HR (95% CI)a 1 0.95 (0.82, 1.09) 0.88 (0.76, 1.02) 0.93 (0.80, 1.07) 0.41 0.95 (0.84, 1.07)
  Multivariate-adjusted HR (95% CI)b 1 1.01 (0.87, 1.16) 0.93 (0.80, 1.08) 0.92 (0.79, 1.07) 0.22 0.92 (0.81, 1.05)
 Coronary heart disease
  No. of deaths/person-year 190/69,165 192/82,907 193/95,280 189/92,515
  Age-adjusted HR (95% CI)a 1 1.00 (0.82, 1.22) 0.98 (0.80, 1.20) 1.01 (0.82, 1.24) 0.90 1.01 (0.86, 1.20)
  Multivariate-adjusted HR (95% CI)b 1 1.03 (0.84, 1.26) 0.99 (0.80, 1.21) 0.92 (0.74, 1.14) 0.31 0.91 (0.77, 1.09)
 Stroke
  No. of deaths/person-year 152/69,165 134/82,907 104/95,280 117/92,515
  Age-adjusted HR (95% CI)a 1 0.91 (0.72, 1.14) 0.72 (0.56, 0.92) 0.86 (0.67, 1.09) 0.27 0.89 (0.72, 1.10)
  Multivariate-adjusted HR (95% CI)b 1 1.00 (0.79, 1.26) 0.80 (0.62, 1.04) 0.94 (0.73, 1.21) 0.56 0.94 (0.75, 1.17)
Legumes
 Cardiovascular disease
  Median intake, serving/day 0.03 0.08 0.12 0.17 0.26
  No. of deaths/person-year 330/69,186 279/68,484 247/67,714 285/67,013 326/67,470
  Age-adjusted HR (95% CI)a 1 0.93 (0.80, 1.09) 0.83 (0.70, 0.98) 0.96 (0.82, 1.12) 0.97 (0.83, 1.13) 0.98 1.00 (0.76, 1.31)
  Multivariate-adjusted HR (95% CI)b 1 0.99 (0.84, 1.16) 0.89 (0.76, 1.06) 1.03 (0.88, 1.21) 1.05 (0.89, 1.23) 0.40 1.13 (0.85, 1.51)
 Coronary heart disease
  No. of deaths/person-year 160/69,186 155/68,484 127/67,714 149/67,013 173/67,470
  Age-adjusted HR (95% CI)a 1 1.06 (0.85, 1.32) 0.87 (0.69, 1.10) 1.02 (0.82, 1.28) 1.06 (0.86, 1.32) 0.57 1.12 (0.77, 1.62)
  Multivariate-adjusted HR (95% CI)b 1 1.10 (0.88, 1.37) 0.92 (0.72, 1.16) 1.06 (0.85, 1.33) 1.08 (0.86, 1.36) 0.57 1.12 (0.75, 1.67)
 Stroke
  No. of deaths/person-year 120/69,186 86/68,484 88/67,714 109/67,013 104/67,470
  Age-adjusted HR (95% CI)a 1 0.81 (0.61, 1.06) 0.83 (0.63, 1.09) 1.02 (0.79, 1.33) 0.84 (0.65, 1.09) 0.58 0.88 (0.55, 1.39)
  Multivariate-adjusted HR (95% CI)b 1 0.86 (0.65, 1.14) 0.91 (0.69, 1.20) 1.15 (0.88, 1.50) 0.99 (0.75, 1.30) 0.55 1.16 (0.71, 1.89)
Open in a new tab
a

Age-adjusted model was adjusted for age (years).

b

Multivariate-adjusted model was adjusted for gender; age (years); ethnicity (Turkmen, Persian, others); education (illiterate, ≤5 years, 6–8 years, high school, academic); marital status (married, other); residency (rural, urban); smoking (never, former, current <20 p/year, current 20–30 p/year, current ≥30 p/year); opium use (user, non-user); alcohol (user, non-user); BMI (<18.5,18.5 to <20.0, 20.0 to <22.5, 22.5 to <25.0, 25.0 to <30.0, 30 to <35.0, ≥35.0); systolic blood pressure (quintile); occupational physical activity (sedentary, standing or occasional walking, causing mild increase in heart rate and sweating, causing significant increase in heart rate and sweating); family history of cancer (yes, no); wealth score (quintile); medication (using a fixed-dose combination pill consisting of aspirin, valsartan, atorvastatin, and hydrochlorothiazide; yes/no); and energy intake (quintile).

c

Quartile.

HR, hazard ratio.

Higher intake of total red meat and poultry was not associated with total cancer, gastrointestinal, or other cancer mortality (Table 4). In multivariate analysis, high fish consumption was associated with reduced total cancer mortality risk (HR [highest versus lowest] =0.79, 95% CI=0.64, 0.98, ptrend=0.03). This association was unchanged with adjustment for total red meat (HR=0.79, 95% CI=0.64, 0.98, ptrend=0.03); fruit and vegetable (HR=0.81, 95% CI=0.66, 1.01, ptrend=0.07); or low-fat dairy food intake (HR=0.79, 95% CI=0.64, 0.99, ptrend=0.03). Fish intake was also associated with lower gastrointestinal cancer mortality risk (HR [highest versus lowest]=0.75, 95% CI=0.56, 1.00, ptrend=0.02). Intake of eggs was somewhat associated with a reduced risk of total cancer mortality (HR=0.81, 95% CI=0.67, 0.99, ptrend=0.17). The highest versus the lowest quintile of legume intake was associated with 28% lower risk of total cancer mortality (HR=0.72, 95% CI=0.58, 0.89, ptrend=0.004), 24% lower risk of gastrointestinal cancer mortality (HR=0.76, 95% CI=0.58, 1.01, ptrend=0.05), and 34% lower risk of other cancer mortality (HR=0.66, 95% CI=0.47, 0.93, ptrend=0.04). Similar estimates for total cancer mortality were noted after additional adjustment for fruit and vegetable intake (data not shown).

Sensitivity Analysis

In sensitivity analyses, participants who died as a result of external events (i.e., accidents, intoxication, suicide, or other types of injury; n=217) were excluded. The results were similar to total mortality (data not shown). Further, the authors evaluated whether the association between protein-rich food intake and all-cause mortality risk was modified by gender, age, BMI, smoking, and wealth score. The associations between dietary protein sources and all-cause mortality did not differ by gender (Appendix Figure 2, available online), age, or BMI (data not shown). A significant interaction between red meat intake and smoking status was noted (pinteraction=0.04). Red meat consumption was significantly associated with higher all-cause mortality risk in current smokers (HR [highest versus lowest] =1.43, 95% CI=1.11, 1.82, ptrend=0.006) compared with never smokers (HR [highest versus lowest]=0.93, 95% CI=0.80, 1.07,ptrend=0.64) or former smokers (HR [highest versus lowest]=1.06, 95% CI=0.63, 1.78, ptrend=0.87). An inverse association between fish intake and all-cause mortality was also observed among non-smokers (pinteraction=0.05; Appendix Figure 3, available online). Furthermore, the association between red meat intake and all-cause mortality was modified by the wealth score. Comparing the highest versus the lowest consumption of total red meat, all-cause mortality risk was 34% higher among participants who had a higher wealth score (HR=1.34, 95% CI=1.09, 1.65, ptrend=0.02), with no association among participants with a lower wealth score (HR=0.93, 95% CI=0.80, 1.08, ptrend=0.67; pinteraction=0.02; Appendix Figure 4, available online). No other significant interaction was observed.

DISCUSSION

In this prospective community-based cohort study in Iran, high consumption of fish and legumes was associated with lower cancer mortality risk, independent of other dietary factors. Lowered all-cause mortality risk was observed among participants with high egg consumption. Furthermore, high red meat intake was associated with higher mortality risk among current smokers as well as participants with higher SES. Poultry, fish, and legume consumption was not related to reduced CVD or all-cause mortality risk among all participants. However, replacement of red meat with fish, eggs, or legumes was associated with lower risk of all-cause mortality.

Based on data from American and European cohort studies, high red meat consumption, with more emphasis on processed meat, has been reported to be associated with all-cause and specific-cause mortality.17,31–33 However, similar to the results from a pooled analysis of eight Asian prospective cohort studies,12 the current study showed that high total red meat intake was not associated with all-cause, CVD, or cancer mortality risk. Wide variation in the amount of red meat and processed meat intake among studies may lead to a difference in risk. Although two servings/day of total red meat is common in most Western countries, mean consumption of total red meat was only 0.19 serving/day in the current study, with a very small amount of processed red meat (mean intake, 0.04 serving/day). A small amount of red meat intake has been reported in Asian prospective cohort studies as well.12 However, within smoking subgroups, similar to the European Prospective Investigation into Cancer and Nutrition study,31 increased mortality risk was observed among current smokers with high total red meat consumption in the current study. Current smokers might be more susceptible to adverse health effects of red meat than never or former smokers. Smoking may accelerate the carcinogenic effect of red meat owing to its induction of CYP1A2 activity,34 however, residual confounding from smoking could not be ruled out.

Although high fish consumption has been associated with reduced all-cause mortality9 and CHD mortality risk10 in meta-analyses of prospective studies, significant associations were not observed in the multivariate-adjusted model. Differences in fish intake may partly explain disparities in results, and SES may play an important role in mediating the association between fish intake and mortality. In the current study, high fish consumption was associated with lower cancer mortality risk, even after adjusting for the wealth score. Prospective studies show a very wide range of results.11–13 While high fish consumption was not associated with cancer mortality risk in the European Prospective Investigation into Cancer and Nutrition study,11 high fish consumption was associated with increased cancer mortality risk among men in a pooled analysis of eight Asian cohorts12 and decreased cancer mortality risk in British men and women.13 Fish contributing to high intake of n-3 fatty acids may suppress the carcinogenesis process and inhibit tumor growth.35

The relation between bean and soybean consumption and cancer risk has been reported in some cohort studies.36–39 As an important dietary source of fiber, flavonols, and other bioactive constituents, legumes may regulate cell proliferation40 and contribute to the prevention of cancer.41

Eggs are a diverse food consisting of several factors that may potentially influence risk of chronic diseases. Cholesterol that is found in eggs has been hypothesized to increase risk of CHD,42 whereas other nutrients such as vitamins, minerals, carotenoids, and monounsaturated fatty acids in eggs may have beneficial health effects.43–45 In general, there is no consistent association across epidemiologic studies in terms of CVD or overall mortality and egg consumption. Although egg consumption was not associated with CVD and all-cause mortality risk in a recent meta-analysis,16 in the current study, high egg consumption was associated with reduced mortality risk. The inconsistencies among cohort studies might be due to chance or differences in intake. In the Physicians’ Health Study, consuming at least one egg/day was significantly associated with higher all-cause mortality risk,46 whereas the median egg intake in the highest quartile in the current study population was less than half an egg/day.

Limitations

The results need to be interpreted in the context of several limitations. Residual confounding is always of concern in any observational study, as participants with high consumption of red meat or other dietary protein sources may have different lifestyles and SES compared with other participants. Although controlling for potential confounders (except the wealth score) had minimal effects on associations, the possibility of residual confounding by unmeasured or unknown confounders could not be ruled out. Moreover, random measurement error due to within-person variation is possible from use of food frequency questionnaires to assess dietary intake. Further, in the current study, the dietary protein sources were only evaluated at enrollment and it is likely that participants might have altered dietary habits during the follow-up period.

This study has several strengths. The prospective population-based cohort study with a high rate of follow up (> 99%) reduced the potential risk of selection bias. To minimize the possibility of recall bias, participants with pre-existing diabetes, CVD, or cancer at time of enrollment were excluded from the analyses. Furthermore, to the best of the authors’ knowledge, this is the first cohort study in the Middle East that prospectively evaluated the role of dietary protein sources in mortality risk among a population with different socioeconomic and lifestyle patterns.

CONCLUSIONS

The present findings suggest that a high intake of fish and legumes might be associated with decreased cancer mortality risk. In addition, lower mortality risk with egg consumption was observed. Although there was no significant association between total red meat intake and all-cause mortality, the association differed by smoking and SES with increased risk in current smokers and in individuals with higher wealth score. Further studies are needed to investigate underlying mechanisms.

Supplementary Material

1

Acknowledgments

This work was supported by Tehran University of Medical Sciences (Grant No. 82-603), Cancer Research United Kingdom (grant No: C20/A5860), the Intramural Research Program of the U.S. National Cancer Institute at NIH National Institutes of Health (Z01 CP000185-03), and various collaborative research agreements with the International Agency for Research on Cancer. Many individuals have contributed to this study. We wish to thank the study participants for their cooperation over many years and the Behvarz’ working in the study areas for their help. We thank the directors of the public health districts of Gonbad and Kalaleh for their collaboration. We express our special thanks to general physicians, nurses, and nutritionists in the enrollment teams for their collaboration and assistance. We received special support from the Social Security Organization of Iran, Golestan Branch. We have enjoyed the close collaboration of the Golestan health deputies and the Chief of the Gonbad health district. Dr. Maryam Farvid would like to express appreciation to the Japan Pharmaceutical Manufacturers Association for their financial support of the Takemi Fellowship Program.

The authors’ responsibilities were as follows: MSF, AFM, AP, HP, SGS, MS, MK, MF, CCA, FK, SMD, PB, PDP, PB, WCW, and RM designed the research; MSF performed the analyses and wrote the manuscript. All authors provided critical input in the writing of the manuscript and read and approved the final version of manuscript. MSF and RM are the guarantors of this investigation.

Footnotes

No financial disclosures were reported by the authors of this paper.

SUPPLEMENTAL MATERIAL

Supplemental materials associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.amepre.2016.10.041.

References

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS835649-supplement-1.pdf (503.4KB, pdf)

RESOURCES