Next Article in Journal
A Patient-Centered Ethical Framework for Irritable Bowel Syndrome Care: Communication, Trust, Nutrition-Sensitive Care, and Self-Management
Previous Article in Journal
Phytochemistry, Bioavailability, and Molecular Mechanisms Underlying Multitarget Anticancer Activity of Aloe vera
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Nutrients
Volume 18
Issue 12
10.3390/nu18122035
nutrients-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Association of Dietary Animal and Plant Protein Composition with All-Cause Mortality: 24-Year Population-Based Cohort Study

by
Federica Prinelli
1,* and
Antonio Giampiero Russo
2
1
Institute of Biomedical Technologies, National Research Council, Via Fratelli Cervi 93, 20054 Segrate, MI, Italy
2
Epidemiology Unit, Agency for Health Protection of the Metropolitan City of Milan, Via Conca del Naviglio, 45, 20123 Milano, MI, Italy
*
Author to whom correspondence should be addressed.
Nutrients 2026, 18(12), 2035; https://doi.org/10.3390/nu18122035
Submission received: 18 May 2026 / Revised: 16 June 2026 / Accepted: 18 June 2026 / Published: 22 June 2026
(This article belongs to the Section Nutritional Epidemiology)
Browse Figures
Review Reports Versions Notes

Abstract

Background: This study examined the associations of dietary animal (AP) and plant protein (PP) with all-cause mortality in an Italian population and assessed the potential effect modification by sex, smoking status, and overweight/obesity (defined as BMI ≥ 25 kg/m2). Methods: This longitudinal population-based study included 1350 adults (50.2% females), aged 40–74 years, recruited between 1991 and 1995, who were followed for all-cause mortality through the regional mortality registry until 2015. Dietary data were collected using a food frequency questionnaire, and protein composition was analysed within a compositional data analysis framework, modelling the balance of AP and PP within the overall macronutrient composition. The associations of protein balances with all-cause mortality were estimated using Cox proportional hazards models adjusted for potential confounders. Effect modification was evaluated through stratified analyses. Results: During follow-up, 405 deaths occurred. A greater AP relative to other macronutrients was associated with higher mortality overall (hazard ratio (HR): 1.37; 95% confidence interval (CI): 1.00–1.87) and in males (HR: 1.57; 95% CI: 1.05–2.33). In stratified analyses, these associations were restricted to ever smokers overall (HR 2.06, 95% CI 1.32–3.20), males (HR 1.90, 95% CI 1.18–3.06), females (HR 3.29, 95% CI 1.03–10.54), and to participants with normal weight (HR 1.91, 95% CI 1.07–3.41). No overall association was observed for PP. Among females, PP was associated with lower mortality in those with normal weight. Conclusions: The associations between AP and PP and mortality differed by sex, smoking status, and adiposity, supporting more tailored dietary recommendations.

1. Introduction

Optimal nutritional balance across the life course is crucial for health and longevity [1], as supported by experimental evidence demonstrating that both caloric intake and dietary macronutrient composition influence lifespan in model organisms [2,3]. In humans, epidemiological studies similarly indicate that macronutrient distributions and overall energy intake predict mortality risk and long-term health outcomes [4,5,6]. Among macronutrients, dietary proteins have received growing attention because of their potential role in shaping health trajectories [7]. In particular, plant-based dietary patterns—typically lower in saturated fats and dietary cholesterol and richer in fibre, antioxidants, and plant proteins—have been associated with a reduced risk of adverse health outcomes, including diabetes, cardiovascular diseases, and cancer [8].
Several prospective cohort studies have examined the association between protein consumption and all-cause mortality, but findings remain inconsistent, particularly when distinguishing between animal protein (AP) and plant protein (PP).
Large cohorts such as the Kuopio Ischaemic Heart Disease Risk Factor Study I [9], the Isfahan Cohort Study [10], and the Nurses’ Health Study & Health Professionals Follow-up Study [11] reported higher mortality associated with AP and lower mortality when PP replaced AP, particularly from red or processed meat. Similar patterns were observed in the US NIH–AARP Diet and Health Study [12]. In contrast, other cohorts—including EPIC Italy [13] and two Japanese prospective studies [14,15], reported null or mixed associations, although substitution analyses often still indicated benefits of PP, particularly among individuals with unhealthy lifestyles or poor dietary patterns. The evidence from pooled analyses further reinforces the importance of protein sources [16,17,18].
Emerging findings also suggest that the health effects of AP and PP vary across subgroups defined by smoking, adiposity, or other lifestyle factors [10,11,13]. However, this evidence remains limited, and potential effect modification has not been systematically investigated.
A major methodological challenge in this field is the compositional nature of dietary data. Because macronutrients are parts of a constrained whole, standard regression approaches for energy adjustment suffer from perfect collinearity and may yield biased or non-interpretable estimates. Substitution models, which omit one or more macronutrients, also have limitations because the interpretation of coefficients depends on the omitted component [19]. To overcome these issues, compositional data analysis (CoDA) has been proposed for modelling nutrient compositions using log-ratio transformations, allowing the estimation of the health effects of specific components while accounting for the relative structure of the diet [20,21,22,23].
To our knowledge, no previous study has applied CoDA to investigate how the relative balance of AP and PP is associated with all-cause mortality. Moreover, evidence on sex-specific associations remains limited [24], and potential modifying roles of smoking and adiposity, strong determinants of mortality, have not been adequately explored. To address these gaps, we examined the association between the relative composition of dietary AP and PP and all-cause mortality in a population-based cohort of middle-aged adults, applying a CoDA framework. We further assessed effect modifications by sex, smoking status, and adiposity.

2. Materials and Methods

2.1. Study Design, Setting and Population

Data were used from a population of individuals aged 40–74 years drawn from the residents’ registry of two districts in Lombardy, Italy. The BEST-FU study methodology has been described previously [25,26]. In brief, the population-based cohort study, conducted from 1991 to 1995, included an in-person examination and the completion of a structured questionnaire. Of the 2882 people initially invited to participate, 1693 agreed, resulting in a response rate of 58.7%. During the hospital visit, participants were asked to complete a series of questionnaires regarding their medical and family history, as well as their past and current drug use and lifestyle habits. They also underwent a general clinical assessment, including anthropometric measurements. In 2015, the health status of participants was retrieved by linking the cohort database with the electronic health records of the regional mortality registries of the Agency for Health Protection (ATS) of Milan (Lombardy Region, Italy), using the fiscal code as a unique identifier. Of the 1604 individuals retrieved from the registries, 1350 (672 males and 678 females) were included in the present analysis after excluding those with unavailable dietary information at baseline (n = 254) (Figure 1).

2.2. Dietary Assessment

The study collected information on participants’ typical dietary intake using a quantitative food frequency questionnaire (FFQ)—a modified version of the questionnaire used in the Nurses’ Health Study [27]—which was administered by a trained interviewer. The questionnaire consisted of a list of 158 food and beverage items. The participants were asked to report how often they consumed each item on a seven-level scale ranging from ‘never’ to ‘four to five times per day’ in the year before study entry. To estimate portion sizes, a picture booklet containing photos of foods of three different sizes (small, medium and large) was used. Daily food intake (grams/day) was calculated from the reported frequency of consumption and portion sizes. Reported food intake was then converted into energy and macronutrient intakes, including AP, PP, starch (ST), simple sugars (SSs), saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs) and polyunsaturated fatty acids (PUFAs), using the Italian food composition database compiled for epidemiological studies [28]. Total energy and energy contribution of each macronutrient were calculated using this database and standard energy conversion factors.

2.3. Ascertainment of All-Cause Death

All individuals retrieved from the regional mortality registries of the Agency for Health Protection (ATS) of Milan (Lombardy Region, Italy) were followed up prospectively for all-cause mortality from the start of the study (1 January 1991–1995) until 1 May 2015 (the end of the follow-up period) [25,26].

2.4. Covariate Assessment

The physicians conducted interviews with each participant to gather data on their sociodemographic background, personal and family medical history, and past or current medication use and lifestyle-related habits such as physical activity, smoking status, and alcohol consumption. A physical examination was performed at the start of the study to measure body weight and height, from which body mass index (BMI) was calculated (weight (kg)/height (m2)) [25]. Educational level was defined as the number of years of formal education completed and categorised into two groups: primary school or less and middle and/or high school or higher. Civil status was categorised as either unmarried (including widowed, separated, divorced, or single) or married. Occupation was classified as retired, employed, homemaker, or other. The participants were categorised as either physically active (engaging in at least one sport per week) or inactive. Smoking status was categorised as never or ever smoking (including former and current smokers). The participants were categorised based on their alcohol consumption as follows: (i) abstainers, (ii) moderate drinkers (≤30 g/day), and (iii) heavy drinkers (>30 g/day).

2.5. Dietary Compositional Data Analysis (CoDA)

To take into account the compositional nature of dietary data, i.e., that macronutrients are parts of a whole and therefore convey relative rather than absolute information, macronutrients were analysed using the CoDA approach based on log-ratio transformations [29]. The application of CoDA to dietary data has been extensively discussed in the literature [21,22,23,30,31]. We considered the seven-part macronutrient composition: AP, PP, ST, SS, SFA, MUFA, and PUFA. Our interest was to evaluate the relative dominance of AP and PP in relation to all other macronutrients. Although the pivot balance approach is generally used for this purpose, the additive log-ratio (alr) transformation provides equivalent information when estimating the dominance of a given part relative to the others [32]. For a D-part composition ( x 1 , x 2 , …, x D ), the alr is defined as:
a l r x l n x 1 x D l n x 2 x D , . . . , l n x D − 1 x D
In our case, six alr coordinates were calculated [30,33] by using ST as the denominator, and each alr term was included as a predictor in the regression model.
Let y be the outcome. The model can be expressed as:
f y = B 0 + B 1 · l n A P S T + B 2 · l n P P S T + B 3 · l n S S S T + B 4 · l n S F A S T + B 5 · l n M U F A S T + B 6 · l n P U F A S T + ∑ i B i · z i
where z i are the control variables.
The coefficient B1, for example, represents the expected change in the outcome when AP increases and ST decreases, while keeping all of the other terms in the model constant. This necessarily means that PP, SS, SFA, MUFA, and PUFA also decrease by the same factor as ST.
To estimate the relative dominance of ST, a second set of six alr coordinates was computed using AP as the denominator, and the regression model was re-estimated accordingly [30]. In the present work, we focused on protein relative dominance, so the results for the other macronutrients are not presented among the main findings but are reported in Supplementary Table S1.

2.6. Statistical Analysis

The characteristics of the study population are described using numbers and percentages. Person-years were calculated from baseline until the occurrence of death, loss to follow-up, or the end of the follow-up period, whichever occurred first. A Cox proportional hazards regression model was used to estimate the hazard ratio (HR) and the corresponding 95% confidence intervals (CIs) for the association between alr coordinates and all-cause mortality, using age as the time scale. Schoenfeld’s residuals test confirmed that the ratio of hazards remained proportional and constant over time. HRs and 95% CIs were calculated for all individuals, as well as for males and females. Models were adjusted for birth cohort (in 10-year intervals), educational level, civil status, occupation, smoking status, physical activity, alcohol consumption, daily fibre intake (continuous, in grams) and total daily energy intake excluding alcohol (continuous, in kilocalories). We investigated the potential modifying effects of sex, smoking status (ever vs. never), and adiposity status (using BMI cut-off value of 25 kg/m2) on the association between dietary protein and mortality by evaluating stratum-specific estimates. To avoid possible systematic selection of the study population, we performed a sensitivity analysis by comparing the characteristics of the participants included in the analysis (n = 1350) with those who were excluded due to missing dietary information at baseline (n = 254) (Supplementary Table S2). To ensure that the observed associations were not confounded by dietary changes resulting from undiagnosed chronic diseases at baseline, we also repeated the main analysis, excluding deaths that occurred within the first two years of follow-up (n = 16) (Supplementary Table S3). All analyses were conducted using Stata 15 (StataCorp, College Station, TX, USA).

3. Results

3.1. Description of Cohort

Table 1 shows the baseline characteristics of the 1350 individuals analysed. The mean age of the sample was 57 years (SD 8). Most participants had completed primary school or less, were employed or homemakers, were married, and reported low levels of physical activity. After a median follow-up period of 21 years, 405 deaths were recorded. Compared with females, males were significantly older, better educated, more likely to be employed, married, and more likely to have smoked heavily or consumed alcohol in large quantities. The number of deaths was higher in males than in females: 39% vs. 21%.
Figure 2 shows the daily percentage of energy intake from macronutrients. Proteins represent 16.9% of total energy intake, with 10.9% coming from animal sources and 6% from plant sources. Carbohydrates represent 55.5% of energy intake (31.1% from ST and 24.4% from SS), while total fats represent 27.6% (10% from SFA, 13% from MUFA and 4.5% from PUFA. We observed that males consumed more daily energy, starch and total fat, and fewer SS. No differences in animal or protein % were observed between sexes.

3.2. Association Analysis

The forest plot (Figure 3) presents the hazard ratios from Cox regression models assessing the association between alr and mortality. In the overall sample, a higher dominance of AP was associated with an increased risk of all-cause mortality (HR = 1.37, 95% CI 1.00, 1.87). Sex-stratified analysis revealed clear effect modifications: AP was associated with increased mortality only in males (HR = 1.57, 95% CI 1.05–2.33), whereas no association was observed in females. There was no overall association with PP, and no sex-specific differences were detected. No other significant associations were observed considering the other macronutrients (Supplementary Table S1).
When we stratified the analysis by smoking status, we observed that AP was associated with a markedly higher mortality risk among ever smokers (current and former). In the total population, ever smokers showed a two-fold higher risk (HR 2.06, 95%CI 1.32, 3.20), with consistent associations in males (HR 1.90, 95% CI 1.18, 3.06) and females (HR 3.29, 95% CI 1.03, 10.54). Stratification by BMI status revealed that AP was significantly associated with increased mortality only among individuals with a BMI < 25 kg/m2 (HR 1.91, 95% CI 1.07, 3.41) (Table 2).
Considering PP, sex-stratified analysis revealed a complex pattern, with associations differing markedly by adiposity among females. Specifically, PP was associated with lower mortality among lean females (BMI < 25 kg/m2) (HR 0.05, 95%CI 0.003, 0.94), whereas a non-significant trend toward higher mortality was observed among females with BMI ≥ 25 kg/m2 (Table 3).
Sensitivity analyses revealed that participants excluded from the analysis were significantly younger, more likely to be male and employed, and more likely to be former smokers and heavier drinkers than those included. However, there was no difference in death status (Supplementary Table S2). To minimise reverse causation due to undiagnosed chronic illnesses, we excluded deaths within the first two years of follow-up, and the results remained essentially unchanged (Supplementary Table S3).

4. Discussion

In a cohort of adults living in northern Italy who were followed for more than two decades, shifting the dietary composition toward an increase in animal protein was associated with a 37% increased risk of all-cause mortality. No association was found for plant protein. However, further stratified analysis revealed interaction effects by sex, smoking status, and adiposity, as discussed below.

4.1. Animal Proteins

We found that the association with AP was limited to males and to those who were current or former smokers and had a BMI below 25 kg/m2, regardless of sex.
Our findings were corroborated by data from previous studies reporting a positive association between AP and mortality, even though they used different approaches to analyse the relationship between protein intake [11,34,35] and mortality. For example, in the Kuopio Ischaemic Heart Disease Risk Factor Study I [9], the authors reported that a higher ratio of animal to plant protein in the diet, as well as higher meat intake, was associated with increased mortality risk. Naghshi et al. [16], in a systematic review and meta-analysis, reported that higher total protein intake was associated with reduced all-cause mortality. Similarly, Chen [6] analysed data from the Rotterdam Study and conducted a meta-analysis of prospective cohorts, concluding that higher AP intake was associated with increased all-cause and cardiovascular mortality. Other recent meta-analyses showed that substitution of AP with PP is associated with lower mortality [17,18].
Consistent with our findings, Haghighatdoost and colleagues [10] demonstrated that increased AP intake was associated with an increased risk of mortality in the Isfahan cohort study, especially among individuals with unhealthy lifestyles such as smoking, heavy drinking, obesity, or physical inactivity. Similarly, in the Nurses’ Health Study & Health Professionals Follow-up Study, Song et al. [11] also reported that high AP intake was linked to increased cardiovascular mortality. In contrast, high PP intake was protective—particularly among individuals with at least one unhealthy lifestyle factor based on smoking, heavy alcohol intake, overweight/obesity, and physical inactivity—with substitution of PP for AP, especially from processed red meat, associated with lower mortality.
Even though food-group analyses were beyond the scope of the present study, it is plausible that the observed association partly reflects the food matrices in which animal proteins are commonly used. In many dietary patterns, animal proteins, particularly when derived from red and processed meat, co-occur with harmful components with pro-oxidant and carcinogenic activities (e.g., N-nitroso compounds) [36,37] and that have been linked to increased risks of coronary heart disease, cancer, and mortality. Cigarette smoking represents an additional source of carcinogens and reactive oxidant species. When combined with high animal proteins, smoking may amplify oxidative stress and inflammation, increasing mortality risk [38]. Additionally, smokers often exhibit poorer dietary patterns, including lower intake of antioxidant-rich foods, which reduces their capacity to counteract the oxidative damage generated by certain animal-derived foods. This lack of protection can exacerbate the risk of disease and mortality associated with smoking [39]. A diet rich in animal proteins may further exacerbate this imbalance by increasing the oxidative load without providing compensatory antioxidant protection in smokers [40,41].

4.2. Plant Proteins

The divergent associations observed for PP in females, with an increased risk in lean females (BMI < 25 kg/m2) and the trend for an increased risk in those with overweight/obesity (BMI ≥ 25 kg/m2), suggest that adiposity modifies the metabolic response to plant-derived proteins. To our knowledge, no prior studies have identified these specific associations. However, a cross-sectional study of Indonesian females of reproductive age revealed a positive correlation between higher PP intake (mainly from grains) and BMI [42]. In contrast, a study of Australian adults found that PP intake was inversely associated with BMI and waist circumference in males but not in females after adjusting for energy intake [43]. Compared with lean females, individuals with overweight/obesity are characterised by chronic low-grade inflammation, insulin resistance, and an altered metabolic profile. We speculate that these conditions influence the way dietary proteins and other nutrients are processed, potentially modifying the physiological response to plant-based diets. As a result, the benefits generally attributed to higher plant protein intake may be attenuated in females with excess adiposity [44]. Another possible explanation is interindividual variability in gut microbiota function, which affects the biotransformation of the non-protein compounds present in plant-based foods (e.g., phytoestrogens). Differences in microbial metabolism have been linked to variability in cardiometabolic risk, and some evidence suggests that such variability is more pronounced among individuals with overweight or obesity, limiting the potential benefits of plant-rich diets in this group [45,46]. In addition, plant-based diets might mitigate the genetic risk of obesity on cardiometabolic markers more effectively in genetically susceptible females [47]. However, because this subgroup analysis involved relatively small numbers of participants and deaths, resulting in imprecise estimates with wide confidence intervals, these findings should be interpreted cautiously and considered hypothesis-generating.

4.3. Strengths and Limitations

Notably, there are several limitations in interpreting our findings. Dietary intake was assessed only once at baseline and was assumed to reflect habitual midlife intake. Consequently, changes in dietary habits over time may have led to non-differential exposure misclassification for both absolute intake and macronutrient composition. Such misclassification may attenuate the associations between macronutrient and mortality, particularly if the relative contributions of animal proteins, plant proteins, and other macronutrients changed over time, thereby biasing the estimates toward the null.
Furthermore, selection bias is a potential concern, as excluding slightly younger, male, employed, smokers and alcohol drinkers participants may limit the generalisability of our findings. However, the exclusion of higher-risk individuals may have attenuated the associations and shifted the results towards the null.
Furthermore, residual measured or unmeasured confounding cannot be entirely ruled out because of factors not considered in this analysis (e.g., comorbidities, medications, broader socioeconomic indicators, and changes in lifestyle over time), despite controlling for several potential risk conditions.
Finally, the borderline association of PP with increased mortality observed in females with overweight/obesity may be partly explained by reverse causation, as those with conditions such as diabetes or cardiovascular disease often adopt such dietary patterns on medical advice. In such cases, mortality risk would likely reflect the underlying illness rather than diet itself. By contrast, in lean females, a higher PP balance may simply reflect a generally healthier lifestyle, including greater consumption of vegetables and higher physical activity. However, because the associations remained after excluding early deaths and were independent of fibre intake and other lifestyle factors, reverse causation is unlikely to fully explain our findings.
One of the strengths of this study is its use of the CoDA approach to study the relationship between protein balance and mortality, a technique that has not previously been used to examine this association. Unlike traditional regression, which struggles with the compositional nature and collinearity of macronutrients, often yielding inconsistent or spurious results depending on which nutrient is excluded, CoDA treats nutrients in relation to one another. This enables proportional changes to be estimated without distorting their compositional balance.
Further strengths of this study include its prospective, population-based design, which limits the potential for reverse causation, and its long follow-up period. Additionally, detailed information on potential risk factors for mortality was collected at baseline, enabling rigorous adjustment for confounding factors in the analysis.

5. Conclusions

This study provides the first compositional analysis of dietary protein in relation to mortality, suggesting that the balance between protein sources plays a role in long-term health. By applying a compositional data analysis framework, our findings highlight the importance of evaluating dietary components in relation to one another rather than in isolation, offering a more comprehensive understanding of dietary effects.
Although the sex-, adiposity-, and smoking-specific associations we observed are novel, they should be interpreted with caution. Future longitudinal and experimental studies in different populations are needed to confirm our results.
From a public health perspective, however, these findings indicate that the balance of dietary protein sources could be relevant for implementing more targeted preventive strategies aimed at reducing long-term adverse health outcomes, particularly among individuals with less-healthy behaviours.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu18122035/s1, Table S1: Hazard ratios for all-cause mortality associated with remaining macronutrients. Table S2: Characteristics of the study population included and excluded from the analysis (n = 1604); Table S3: Hazard ratios for all-cause mortality associated with animal and plant protein (alr), stratified by smoking and adiposity after exclusion of participants who died within 2 years of follow-up (n = 16).

Author Contributions

Conceptualisation, F.P. and A.G.R.; methodology, F.P. and A.G.R.; formal analysis, F.P.; data curation, F.P.; writing—original draft preparation, F.P.; writing—review and editing, A.G.R.; visualisation, F.P.; supervision, F.P. and A.G.R.; project administration, F.P. All authors have read and agreed to the published version of the manuscript.

Funding

The original BEST study was supported by the research programme grant funding from the CNR-targeted project “Prevention and Control of Diseases factors—FATMA”.

Institutional Review Board Statement

The original BEST study was conducted in accordance with the Declaration of Helsinki guidelines, and all procedures involving human participants were approved by the Italian National Research Council’s Research Ethics Committee. Ethical approval for the BEST-FU study was obtained from the Ethics Committee of AO ‘G. Salvini’, Garbagnate Milanese (protocol approval number 14295-2012, 29 March 2012) and CET Lombardia 1 (protocol approval number CET 409-2025, 24 November 2025).

Informed Consent Statement

In the original BEST study (1991–1995), verbal informed consent to participate and to allow data collection was obtained from all participants, witnessed by study staff and documented in the study database, in accordance with the national regulations in force at that time. For the present mortality follow-up, re-contacting participants to obtain additional written consent was not feasible because many were deceased or untraceable. The current use of baseline data and linkage to mortality registries was therefore reviewed and approved by the local Ethics Committee (San Raffaele CET, as cited in the manuscript) and the Data Protection Officer, in accordance with GDPR Article 9(2)(j) and Article 110 of the Italian Data Protection Code (Legislative Decree 196/2003, as amended) and applicable national legislation, which allow the processing of health data for scientific research when appropriate safeguards are in place and when renewed consent cannot reasonably be obtained.

Data Availability Statement

The data described in the manuscript, code book, and analytic code will be made available upon request of pending application and approval.

Acknowledgments

The authors would like to thank the original investigators, the study staff, and the participants for their contributions. Special thanks go to Maria Lea Correa Leite and Massimo Musicco for their previous work on this project, as well as for their invaluable and unwavering support.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
CoDACompositional data analysis
BEST-FUBollate Eye Study Follow-up
ATSAgency for Health Protection
FFQFood-frequency questionnaire
APAnimal protein
PPplant protein
STStarch
SSSimple sugar
SFASaturated fatty acids
MUFAMonounsaturated fatty acids
PUFAPolyunsaturated fatty acids
BMIBody mass index
alrAdditive log ratio
HRHazard ratio
CIConfidence interval

References

  1. Senior, A.M.; Nakagawa, S.; Raubenheimer, D.; Simpson, S.J. Global associations between macronutrient supply and age-specific mortality. Proc. Natl. Acad. Sci. USA 2020, 117, 30824–30835. [Google Scholar] [CrossRef] [PubMed]
  2. Le Couteur, D.G.; Solon-Biet, S.; Cogger, V.C.; Mitchell, S.J.; Senior, A.; De Cabo, R.; Raubenheimer, D.; Simpson, S.J. The impact of low-protein high-carbohydrate diets on aging and lifespan. Cell. Mol. Life Sci. 2016, 73, 1237–1252. [Google Scholar] [PubMed]
  3. Nakagawa, S.; Lagisz, M.; Hector, K.L.; Spencer, H.G. Comparative and meta-analytic insights into life extension via dietary restriction. Aging Cell 2012, 11, 401–409. [Google Scholar] [PubMed]
  4. Kitada, M.; Ogura, Y.; Monno, I.; Koya, D. The impact of dietary protein intake on longevity and metabolic health. eBioMedicine 2019, 43, 632–640. [Google Scholar] [CrossRef] [PubMed]
  5. Chen, Z.; Glisic, M.; Song, M.; Aliahmad, H.A.; Zhang, X.; Moumdjian, A.C.; Gonzalez-Jaramillo, V.; van Der Schaft, N.; Bramer, W.M.; Ikram, M.A.; et al. Dietary protein intake and all-cause and cause-specific mortality: Results from the Rotterdam Study and a meta-analysis of prospective cohort studies. Eur. J. Epidemiol. 2020, 35, 411–429. [Google Scholar] [PubMed]
  6. Seidelmann, S.B.; Claggett, B.; Cheng, S.; Henglin, M.; Shah, A.; Steffen, L.M.; Folsom, A.R.; Rimm, E.B.; Willett, W.C.; Solomon, S.D. Dietary carbohydrate intake and mortality: A prospective cohort study and meta-analysis. Lancet Public Health 2018, 3, e419–e428. [Google Scholar] [CrossRef] [PubMed]
  7. Rodriguez, N.R. Introduction to Protein Summit 2.0: Continued exploration of the impact of high-quality protein on optimal health. Am. J. Clin. Nutr. 2015, 101, 1317S–1319S. [Google Scholar] [CrossRef] [PubMed]
  8. Wang, Y.; Liu, B.; Han, H.; Hu, Y.; Zhu, L.; Rimm, E.B.; Hu, F.B.; Sun, Q. Associations Between Plant-Based Dietary Patterns and Risks of Type 2 Diabetes, Cardiovascular Disease, Cancer, and Mortality—A Systematic Review and Meta-analysis. Nutr. J. 2023, 22, 46, Erratum in Nutr. J. 2024, 23, 6. [Google Scholar] [PubMed]
  9. Virtanen, H.E.K.; Voutilainen, S.; Koskinen, T.T.; Mursu, J.; Kokko, P.; Ylilauri, M.P.; Tuomainen, T.P.; Salonen, J.T.; Virtanen, J.K. Dietary proteins and protein sources and risk of death: The Kuopio Ischaemic Heart Disease Risk Factor Study. Am. J. Clin. Nutr. 2019, 109, 1462–1471. [Google Scholar] [CrossRef] [PubMed]
  10. Haghighatdoost, F.; Mohammadifard, N.; Zakeri, P.; Najafian, J.; Sadeghi, M.; Roohafza, H.; Sarrafzadegan, N. Differences in all-cause mortality risk associated with animal and plant dietary protein sources consumption. Sci. Rep. 2023, 13, 3396. [Google Scholar] [CrossRef] [PubMed]
  11. Song, M.; Fung, T.T.; Hu, F.B.; Willett, W.C.; Longo, V.D.; Chan, A.T.; Giovannucci, E.L. Association of Animal and Plant Protein Intake With All-Cause and Cause-Specific Mortality. JAMA Intern. Med. 2016, 176, 1453–1463. [Google Scholar] [PubMed]
  12. Huang, J.; Liao, L.M.; Weinstein, S.J.; Sinha, R.; Graubard, B.I.; Albanes, D. Association Between Plant and Animal Protein Intake and Overall and Cause-Specific Mortality. JAMA Intern. Med. 2020, 180, 1173–1184. [Google Scholar] [CrossRef] [PubMed]
  13. Fontana, L.; Sieri, S.; Ricceri, F.; Agnoli, C.; Pala, V.; Masala, G.; Saieva, C.; Catalano, A.; Macciotta, A.; Tumino, R.; et al. Dietary intake of animal and plant proteins and risk of all cause and cause-specific mortality: The Epic-Italy cohort. Nutr. Healthy Aging 2022, 6, 257–268. [Google Scholar] [CrossRef]
  14. Budhathoki, S.; Sawada, N.; Iwasaki, M.; Yamaji, T.; Goto, A.; Kotemori, A.; Ishihara, J.; Takachi, R.; Charvat, H.; Mizoue, T.; et al. Association of Animal and Plant Protein Intake With All-Cause and Cause-specific mortality in a japanese cohort. JAMA Intern. Med. 2019, 179, 1509–1518. [Google Scholar] [CrossRef] [PubMed]
  15. Kurata, H.; Meguro, S.; Abe, Y.; Sasaki, T.; Asakura, K.; Arai, Y.; Itoh, H. Dietary protein intake and all-cause mortality: Results from The Kawasaki Aging and Wellbeing Project. BMC Geriatr. 2023, 23, 479. [Google Scholar] [CrossRef] [PubMed]
  16. Naghshi, S.; Sadeghi, O.; Willett, W.C.; Esmaillzadeh, A. Dietary intake of total, animal, and plant proteins and risk of all cause, cardiovascular, and cancer mortality: Systematic review and dose-response meta-analysis of prospective cohort studies. BMJ 2020, 370, m2412. [Google Scholar] [CrossRef] [PubMed]
  17. Barrantes-Espinola, G.; Alberti, P.; Jaramillo-Morales, E.; Rodriguez, A.; López-Moreno, M. Isocaloric substitution of animal protein with plant protein and its impact on all-cause, cardiovascular, and cancer mortality: A systematic review and meta-analysis. Clin. Nutr. 2026, 61, 106654. [Google Scholar] [CrossRef] [PubMed]
  18. Wallerer, S.; Papakonstantinou, T.; Morze, J.; Stadelmaier, J.; Kiesswetter, E.; Gorenflo, L.; Barbaresko, J.; Szczerba, E.; Neuenschwander, M.; Bell, W.; et al. Association between substituting macronutrients and all-cause mortality: A network meta-analysis of prospective observational studies. eClinicalMedicine 2024, 75, 102807. [Google Scholar] [CrossRef] [PubMed]
  19. Song, M.; Giovannucci, E. Substitution analysis in nutritional epidemiology: Proceed with caution. Eur. J. Epidemiol. 2018, 33, 137–140. [Google Scholar] [CrossRef] [PubMed]
  20. Leite, M.L. Applying compositional data methodology to nutritional epidemiology. Stat. Methods Med. Res. 2016, 25, 3057–3065. [Google Scholar] [CrossRef] [PubMed]
  21. Corrêa Leite, M.L. Compositional data analysis as an alternative paradigm for nutritional studies. Clin. Nutr. ESPEN 2019, 33, 207–212. [Google Scholar] [CrossRef] [PubMed]
  22. Leite, M.L.C.; Prinelli, F. A compositional data perspective on studying the associations between macronutrient balances and diseases. Eur. J. Clin. Nutr. 2017, 71, 1365–1369. [Google Scholar] [CrossRef] [PubMed]
  23. Djusse, M.E.; Prinelli, F.; Camboni, T.; Ceccarani, C.; Consolandi, C.; Conti, S.; Dall’aSta, M.; Danesi, F.; Laghi, L.; Curatolo, F.M.; et al. Dietary habits and vaginal environment: Can a beneficial impact be expected? Front. Cell. Infect. Microbiol. 2025, 15, 1582283. [Google Scholar] [CrossRef] [PubMed]
  24. Hayashi, K.; Nakamura, M.; Tsujiguchi, H.; Hara, A.; Suzuki, K.; Miyagi, S.; Takazawa, C.; Zhao, J.; Camara, J.; Marama, T.; et al. Sex Differences in the Relationship Between Nutrient Intake and Mortality Results of the Shika Cohort Study. Nutrients 2025, 17, 755. [Google Scholar] [CrossRef] [PubMed]
  25. Prinelli, F.; Yannakoulia, M.; Anastasiou, C.A.; Adorni, F.; Di Santo, S.G.; Musicco, M.; Scarmeas, N.; Leite, M.L.C. Mediterranean diet and other lifestyle factors in relation to 20-year all-cause mortality: A cohort study in an Italian population. Br. J. Nutr. 2015, 113, 1003–1011. [Google Scholar] [CrossRef] [PubMed]
  26. Prinelli, F.; Adorni, F.; Leite, M.L.C.; Pettenati, C.; Russo, A.; Di Santo, S.; Musicco, M. Different Exposures to Risk Factors Do Not Explain the Inverse Relationship of Occurrence Between Cancer and Neurodegenerative Diseases: An Italian Nested Case-control Study. Alzheimer Dis. Assoc. Disord. 2018, 32, 76–82. [Google Scholar] [PubMed]
  27. Willet, W.C. Nurses’ Health Study Dietary questionnaire: Appendix 5-1. In Nutritional Epidemiology, 2nd ed.; Oxford University Press: New York, NY, USA, 1998; pp. 95–97. [Google Scholar]
  28. Salvini, S.; Parpinel, M.; Gnagnarella, P.; Maisonneuve, P.; Turrini, A. Banca Dati di Composizione Degli Alimenti per Studi Epidemiologici in Italia (Datase for Food Composition for Epidemiological Studiesin Italy); IEO s.n.: Milan, Italy, 1998. [Google Scholar]
  29. Pawlowsky-Glahn, V.; Egozcue, J.J.; Tolosana-Delgado, R. Modelling and Analysis of Compositional Data; John Wiley & Sons: Hoboken, NJ, USA, 2015. [Google Scholar] [CrossRef]
  30. Leite, M.L.C. Log-ratio transformations for dietary compositions: Numerical and conceptual questions. J. Nutr. Sci. 2021, 10, e97. [Google Scholar] [CrossRef] [PubMed]
  31. Leite, M.L.C. Orthonormal balances as a means of characterizing dietary exposure. Nutr. Res. 2020, 81, 90–96. [Google Scholar] [CrossRef] [PubMed]
  32. Coenders, G.; Pawlowsky-Glahn, V. On interpretations of tests and effect sizes in regression models with a compositional predictor. SORT Stat. Oper. Res. Trans. 2020, 44, 201–220. [Google Scholar] [CrossRef]
  33. Comas-Cufí, M.; Thió-Henestrosa, S. CoDaPack 2.0, 2011. Universitat de Girona. Available online: https://ima.udg.edu/codapack/ (accessed on 1 May 2026).
  34. Trichopoulou, A.; Psaltopoulou, T.; Orfanos, P.; Hsieh, C.C.; Trichopoulos, D. Low-carbohydrate-high-protein diet and long-term survival in a general population cohort. Eur. J. Clin. Nutr. 2007, 61, 575–581. [Google Scholar] [CrossRef] [PubMed]
  35. Kelemen, L.E.; Kushi, L.H.; Jacobs, D.R., Jr.; Cerhan, J.R. Associations of dietary protein with disease and mortality in a prospective study of postmenopausal women. Am. J. Epidemiol. 2005, 161, 239–249. [Google Scholar] [CrossRef] [PubMed]
  36. Etemadi, A.; Sinha, R.; Ward, M.H.; Graubard, B.I.; Inoue-Choi, M.; Dawsey, S.M.; Abnet, C.C. Mortality from different causes associated with meat, heme iron, nitrates, and nitrites in the NIH-AARP Diet and Health Study: Population based cohort study. BMJ 2017, 357, j1957. [Google Scholar] [CrossRef] [PubMed]
  37. Bastide, N.M.; Pierre, F.H.; Corpet, D.E. Heme iron from meat and risk of colorectal cancer: A meta-analysis and a review of the mechanisms involved. Cancer Prev. Res. 2011, 4, 177–184. [Google Scholar] [CrossRef] [PubMed]
  38. Sinha, R.; Cross, A.J.; Graubard, B.I.; Leitzmann, M.F.; Schatzkin, A. Meat intake and mortality: A prospective study of over half a million people. Arch. Intern. Med. 2009, 169, 562–571. [Google Scholar] [CrossRef] [PubMed]
  39. Dallongeville, J.; Marécaux, N.; Fruchart, J.-C.; Amouyel, P. Cigarette smoking is associated with unhealthy patterns of nutrient intake: A meta-analysis. J. Nutr. 1998, 128, 1450–1457. [Google Scholar] [CrossRef] [PubMed]
  40. Okubo, H.; Shaheen, S.O.; Ntani, G.; Jameson, K.A.; Syddall, H.E.; Sayer, A.A.; Dennison, E.M.; Cooper, C.; Robinson, S.M.; the Hertfordshire Cohort Study Group. Processed meat consumption and lung function: Modification by antioxidants and smoking. Eur. Respir. J. 2014, 43, 972–982. [Google Scholar] [CrossRef] [PubMed]
  41. Margetts, B.M.; Jackson, A.A. Interactions between people’s diet and their smoking habits: The dietary and nutritional survey of British adults. BMJ 1993, 307, 1381–1384. [Google Scholar] [CrossRef] [PubMed]
  42. Sistia, F.; Khusun, H.; Februhartanty, J. Plant protein consumption is associated with body mass index among women of reproductive age in Indonesia. Front. Nutr. 2023, 10, 1243635. [Google Scholar] [CrossRef] [PubMed]
  43. Arini, H.R.B.; Leech, R.M.; Tan, S.-Y.; McNaughton, S.A. Plant and animal protein intakes are differentially associated with diet quality and obesity: Findings from the Australian National Nutrition and Physical Activity Survey of Australian Adults. Br. J. Nutr. 2025, 133, 1133–1144. [Google Scholar] [CrossRef] [PubMed]
  44. Zhu, W.; Gregory, J.C.; Org, E.; Buffa, J.A.; Gupta, N.; Wang, Z.; Li, L.; Fu, X.; Wu, Y.; Mehrabian, M.; et al. Gut microbial metabolite TMAO enhances platelet Hyperreactivity and thrombosis risk. Cells 2016, 165, 111–124. [Google Scholar] [CrossRef] [PubMed]
  45. Frankenfeld, C.L. Cardiometabolic risk and gut microbial phytoestrogen metabolite phenotypes. Mol. Nutr. Food Res. 2017, 61, 1500900. [Google Scholar]
  46. Liu, Z.-M.; Ho, S.C.; Chen, Y.-M.; Liu, J.; Woo, J. Cardiovascular Risks in Relation to Daidzein Metabolizing Phenotypes among Chinese Postmenopausal Women. PLoS ONE 2017, 9, e87861. [Google Scholar]
  47. Hajmir, M.M.; Mirzababaei, A.; Abaj, F.; Aali, Y.; Samadi, M.; Mirzaei, K. Interaction of genetic risk score (GRS) and Plant-Based diet on atherogenic factors and body fat distribution indices among women with overweight and obesity: A cross-sectional study. Sci. Rep. 2025, 15, 24405. [Google Scholar] [PubMed]
Nutrients 18 02035 g001
Figure 1. Flow chart describing steps involved in establishing final study sample.
Figure 1. Flow chart describing steps involved in establishing final study sample.
Nutrients 18 02035 g001
Nutrients 18 02035 g002
Figure 2. Mean daily % of energy intake from macronutrients. ** p-value < 0.001. SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid.
Figure 2. Mean daily % of energy intake from macronutrients. ** p-value < 0.001. SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid.
Nutrients 18 02035 g002
Nutrients 18 02035 g003
Figure 3. Forest plot for association between AP and PP and mortality. Abbreviations: AP, animal protein; PP, plant protein; BMI, body mass index (kg/m2); HR, hazard ratio; CI, confidence interval. Models include terms for birth cohort, educational level, civil status, occupation, smoking status, physical activity, alcohol consumption, daily intake of fibre (g, continuous), and total daily energy intake excluding alcohol intake (Kcal, continuous).
Figure 3. Forest plot for association between AP and PP and mortality. Abbreviations: AP, animal protein; PP, plant protein; BMI, body mass index (kg/m2); HR, hazard ratio; CI, confidence interval. Models include terms for birth cohort, educational level, civil status, occupation, smoking status, physical activity, alcohol consumption, daily intake of fibre (g, continuous), and total daily energy intake excluding alcohol intake (Kcal, continuous).
Nutrients 18 02035 g003
Table 1. Baseline participants’ characteristics by sex.
Table 1. Baseline participants’ characteristics by sex.
All
(n = 1350)		Males
(n = 672, 49.8%)		Females
(n = 678, 50.2%)		p-Value
Baseline Characteristicsn (%)n (%)n (%)
Birth cohort (calendar years) 0.006
  <1920107 (7.9)70 (10.4)37 (5.5)
  1920–1929388 (28.7)191 (28.4)197 (29.1)
  1930–1939611 (45.3)300 (44.6)311 (45.9)
  ≥1940244 (18.1)111 (16.5)133 (19.6)
Age (mean, SD)57 (8)58 (8.3)57 (7)0.050
Educational level <0.001
  Primary school or less773 (57.3)331 (49.3)442 (65.2)
  Middle/high school or graduate577 (42.7)341 (50.7)236 (34.8)
Occupation <0.001
  Retired531 (39.3)307 (45.7)224 (33.0)
  Employed467 (34.6)300 (44.6)167 (24.6)
  Homemakers240 (17.8)1 (0.2)239 (35.3)
  Other112 (8.3)64 (9.5)48 (7.1)
Civil status <0.001
  Unmarried202 (15.0)52 (7.7)150 (22.1)
  Married1148 (85.0)620 (92.3)528 (77.9)
BMI (mean, SD)27.02 (4.1)26.9 (3.4)27.12 (4.7)0.362
  BMI < 25 kg/m2447 (33.1)198 (29.5)249 (36.7)0.005
  BMI ≥ 25 kg/m2903 (66.9)474 (70.5)429 (63.3)
Weekly physical activity 0.073
  Inactive1085 (80.4)527 (78.4)558 (82.3)
  Active265 (19.6)145 (21.6)120 (17.7)
Smoking status <0.001
  Never smokers695 (51.5)197 (29.3)498 (73.5)
  Ever smokers655 (48,5)475 (70.7)180 (26,8)
Energy intake (Kcal)3022.9 (2839.8)3329.7 (1295.3)2718.8 (989.8)<0.001
Alcohol consumption <0.001
  Abstainer407 (30.2)89 (13.2)318 (46.9)
  Moderate drinkers298 (22.1)178 (26.5)120 (17.7)
  Heavy drinkers645 (47.8)405 (60.3)240 (35.4)
Deaths405 (30)262 (39.0)143 (21.1)<0.001
Abbreviation: BMI, body mass index (kg/m2). Numbers are frequencies and percentages if unless specified.
Table 2. Hazard ratios for all-cause mortality associated with animal protein (alr), stratified by smoking and adiposity.
Table 2. Hazard ratios for all-cause mortality associated with animal protein (alr), stratified by smoking and adiposity.
Males
(n = 672, 49.8%)		Females
(n = 678, 50.2%)		All
(n = 1350)
Animal ProteinHR (95% CI)HR (95% CI)HR (95% CI)
Smoking habit
  Never smokers0.91 (0.40, 2.06)0.80 (0.47, 1.37)0.86 (0.55, 1.35)
  Ever smokers1.90 (1.18, 3.06)3.29 (1.03, 10.54)2.06 (1.32, 3.20)
BMI status
  BMI < 25 kg/m21.38 (0.68, 2.79)2.57 (0.92, 7.20)1.91 (1.07, 3.41)
  BMI ≥ 25 kg/m21.62 (0.97, 2.71)0.75 (0.42, 1.33)1.22 (0.83, 1.78)
Abbreviation: BMI, body mass index (kg/m2); HR, hazard ratio; CI, confidence interval. Models include terms for birth cohort, educational level, civil status, occupation, smoking status, physical activity, alcohol consumption, daily intake of fibre (g, continuous), and total daily energy intake excluding alcohol intake (Kcal, continuous).
Table 3. Hazard ratios for all-cause mortality associated with plant protein (alr), stratified by smoking and adiposity.
Table 3. Hazard ratios for all-cause mortality associated with plant protein (alr), stratified by smoking and adiposity.
Males
(n = 672, 49.8%)		Females
(n = 678, 50.2%)		Total
(n = 1350)
Plant ProteinHR (95% CI)HR (95% CI)HR (95% CI)
Smoking status
  Never smokers0.76 (0.11, 5.02)1.49 (0.42, 5.26)1.43 (0.49, 4.21)
  Ever smokers1.14 (0.29, 5.28)0.98 (0.08, 12.07)1.05 (0.31, 3.53)
BMI status
  BMI < 25 kg/m21.06 (0.14, 8.22)0.05 (0.003, 0.94)0.47 (0.09, 2.56)
  BMI ≥ 25 kg/m20.63 (0.18, 2.20)3.06 (0.91, 10.28)1.37 (0.56, 3.33)
Abbreviation: BMI, Body mass Index (kg/m2); HR, hazard ratio; CI, confidence intervals. Models include terms for birth cohort, educational level, civil status, occupation, smoking status, physical activity, alcohol consumption, daily intake of fibre (g, continuous), and total daily energy intake excluding alcohol intake (Kcal, continuous).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Prinelli, F.; Russo, A.G. Association of Dietary Animal and Plant Protein Composition with All-Cause Mortality: 24-Year Population-Based Cohort Study. Nutrients 2026, 18, 2035. https://doi.org/10.3390/nu18122035

AMA Style

Prinelli F, Russo AG. Association of Dietary Animal and Plant Protein Composition with All-Cause Mortality: 24-Year Population-Based Cohort Study. Nutrients. 2026; 18(12):2035. https://doi.org/10.3390/nu18122035

Chicago/Turabian Style

Prinelli, Federica, and Antonio Giampiero Russo. 2026. "Association of Dietary Animal and Plant Protein Composition with All-Cause Mortality: 24-Year Population-Based Cohort Study" Nutrients 18, no. 12: 2035. https://doi.org/10.3390/nu18122035

APA Style

Prinelli, F., & Russo, A. G. (2026). Association of Dietary Animal and Plant Protein Composition with All-Cause Mortality: 24-Year Population-Based Cohort Study. Nutrients, 18(12), 2035. https://doi.org/10.3390/nu18122035

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop