Theory and performance of substitution models for estimating relative causal effects in nutritional epidemiology

Substitutions of saturated fat intakes with other macronutrients and foods and risk of NAFLD cirrhosis and all-cause hepatocellular carcinoma: a prospective cohort study

BACKGROUND

Short-term trials have shown a reduction in liver fat when saturated fatty acids (SFAs) are substituted with polyunsaturated fatty acids (PUFA), or with low-glycemic carbohydrates. However, few cohort studies have been conducted to investigate the associations of replacing SFA and SFA-rich foods with different macronutrients and foods in more severe stages of liver disease; nonalcoholic fatty liver disease (NAFLD) cirrhosis and hepatocellular carcinoma (HCC).

OBJECTIVES

To investigate associations between the substitution of SFA and SFA-rich foods with other macronutrients and foods and NAFLD cirrhosis and HCC in a middle-aged to elderly Swedish population of n = 77,059 males and females.

METHODS

Time-to-event analyses were performed to investigate associations between the food and macronutrient substitutions and NAFLD cirrhosis and HCC. Multivariable Cox regression models were constructed to estimate hazard ratios (HRs) with corresponding 95% confidence intervals (CIs). Statistical isocaloric and equal-mass substitutions were performed using the leave-one-out method. Prespecified nutrient and food substitutions of interest were SFA with carbohydrates, SFA with fiber, SFA with PUFA, butter with margarine and vegetable oils, unprocessed red meat with fish, and milk with fermented milk.

RESULTS

Over a median follow-up of 24 y, 566 cases of NAFLD cirrhosis and 205 cases of HCC were registered. Overall, dietary substitutions showed no clear associations with either NAFLD cirrhosis or HCC. Substituting SFA with carbohydrates showed an HR of 0.87 (95% CI: 0.74, 1.02) for HCC and 1.00 (95% CI: 0.89, 1.11) for NAFLD cirrhosis. Substituting milk with fermented milk showed an HR of 0.93 (95% CI: 0.85, 1.01) for HCC and 0.97 (95% CI: 0.92, 1.03) for NAFLD cirrhosis.

CONCLUSIONS

No clear associations were observed between diet and NAFLD cirrhosis or HCC. Although accompanied by low precision, possible lowered risks of HCC by substituting SFA with carbohydrates or milk with fermented milk might be of interest, but needs replication in other cohorts.

Association between dietary macronutrient composition and plasma one-carbon metabolites and B-vitamin cofactors in patients with stable angina pectoris

Elevated plasma concentrations of several one-carbon metabolites are associated with increased CVD risk. Both diet-induced regulation and dietary content of one-carbon metabolites can influence circulating concentrations of these markers. We cross-sectionally analysed 1928 patients with suspected stable angina pectoris (geometric mean age 61), representing elevated CVD risk, to assess associations between dietary macronutrient composition (FFQ) and plasma one-carbon metabolites and related B-vitamin status markers (GC-MS/MS, LC-MS/MS or microbiological assay). Diet-metabolite associations were modelled on the continuous scale, adjusted for age, sex, BMI, smoking, alcohol and total energy intake. Average (geometric mean (95 % prediction interval)) intake was forty-nine (38, 63) energy percent (E%) from carbohydrate, thirty-one (22, 45) E% from fat and seventeen (12, 22) E% from protein. The strongest associations were seen for higher protein intake, i.e. with higher plasma pyridoxal 5'-phosphate (PLP) (% change (95 % CI) 3·1 (2·1, 4·1)), cobalamin (2·9 (2·1, 3·7)), riboflavin (2·4 (1·1, 3·7)) and folate (2·1 (1·2, 3·1)) and lower total homocysteine (tHcy) (-1·4 (-1·9, -0·9)) and methylmalonic acid (MMA) (-1·4 (-2·0, -0·8)). Substitution analyses replacing MUFA or PUFA with SFA demonstrated higher plasma concentrations of riboflavin (5·0 (0·9, 9·3) and 3·3 (1·1, 5·6)), tHcy (2·3 (0·7, 3·8) and 1·3 (0·5, 2·2)) and MMA (2·0 (0·2, 3·9) and 1·7 (0·7, 2·7)) and lower PLP (-2·5 (-5·3, 0·3) and -2·7 (-4·2, -1·2)). In conclusion, a higher protein intake and replacing saturated with MUFA and PUFA were associated with a more favourable metabolic phenotype regarding metabolites associated with CVD risk.

Grilling the data: application of specification curve analysis to red meat and all-cause mortality

OBJECTIVES

To present an application of specification curve analysis-a novel analytic method that involves defining and implementing all plausible and valid analytic approaches for addressing a research question-to nutritional epidemiology.

STUDY DESIGN AND SETTING

We reviewed all observational studies addressing the effect of red meat on all-cause mortality, sourced from a published systematic review, and documented variations in analytic methods (eg, choice of model, covariates, etc.). We enumerated all defensible combinations of analytic choices to produce a comprehensive list of all the ways in which the data may reasonably be analyzed. We applied specification curve analysis to data from National Health and Nutrition Examination Survey 2007 to 2014 to investigate the effect of unprocessed red meat on all-cause mortality. The specification curve analysis used a random sample of all reasonable analytic specifications we sourced from primary studies.

RESULTS

Among 15 publications reporting on 24 cohorts included in the systematic review on red meat and all-cause mortality, we identified 70 unique analytic methods, each including different analytic models, covariates, and operationalizations of red meat (eg, continuous vs quantiles). We applied specification curve analysis to National Health and Nutrition Examination Survey, including 10,661 participants. Our specification curve analysis included 1208 unique analytic specifications, of which 435 (36.0%) yielded a hazard ratio equal to or more than 1 for the effect of red meat on all-cause mortality and 773 (64.0%) less than 1. The specification curve analysis yielded a median hazard ratio of 0.94 (interquartile range: 0.83-1.05). Forty-eight specifications (3.97%) were statistically significant, 40 of which indicated unprocessed red meat to reduce all-cause mortality and eight of which indicated red meat to increase mortality.

CONCLUSION

We show that the application of specification curve analysis to nutritional epidemiology is feasible and presents an innovative solution to analytic flexibility.

Association of postoperative dietary macronutrient content and quality with total weight loss and fat-free mass loss at midterm after sleeve gastrectomy

OBJECTIVES

This study aimed to investigate the associations of macronutrient quantities and qualities with percentage total weight loss and percentage of fat-free mass loss relative to total weight loss in adults undergoing sleeve gastrectomy.

METHODS

This cross-sectional study included 146 patients on postoperative time since sleeve gastrectomy of 2 to 4 y. Diet was assessed using a food frequency questionnaire. Macronutrient quality index, carbohydrate quality index, fat quality index, and healthy plate protein quality index were calculated. The associations of dietary variables with percentage total weight loss and percentage of fat-free mass loss relative to total weight loss were determined using linear regression. Logistic regression was used to estimate the odds of non-response (percentage total weight loss < 25%) and excessive fat-free mass loss (percentage of fat-free mass loss relative to total weight loss > 28%) based on dietary intakes.

RESULTS

Forty-six (31.5%) were non-responders, and 49 (33.6%) experienced excessive fat-free mass loss. The fully adjusted model showed a 0.75 decrease in percentage total weight loss per 5% carbohydrate increase (95% CI, -1.45 to -0.05). The odds of non-response were 53% lower per 5% increase in protein (95% CI, 0.23-0.94). Each 5-g higher intake of fat was associated with 0.29 higher percentage of fat-free mass loss relative to total weight loss (95% CI, 0.03-0.55). The odds of excessive fat-free mass loss were reduced by 5% per gram of fiber intake (95% CI, 0.90-0.99). Each 5% increment in energy intake from protein that was isocalorically substituted for either carbohydrate or fat was associated with lower odds of nonresponse. Macronutrient quality indices had no significant associations.

CONCLUSIONS

Adherence to a high-protein, high-fiber diet after sleeve gastrectomy may enhance surgical success by improving total weight loss and preventing excessive fat-free mass loss.

Modelling the replacement of red and processed meat with plant-based alternatives and the estimated effect on insulin sensitivity in a cohort of Australian adults

Dietary guidelines are increasingly promoting mostly plant-based diets, limits on red meat consumption, and plant-based sources of protein for health and environmental reasons. It is unclear how the resulting food substitutions associate with insulin resistance, a risk factor for type 2 diabetes. We modelled the replacement of red and processed meat with plant-based alternatives and the estimated effect on insulin sensitivity. We included 783 participants (55 % female) from the Childhood Determinants of Adult Health study, a population-based cohort of Australians. In adulthood, diet was assessed at three time points using FFQ: 2004–2006, 2009–2011 and 2017–2019. We calculated the average daily intake of each food group in standard serves. Insulin sensitivity was estimated from fasting glucose and insulin concentrations in 2017–2019 (aged 39–49 years) using homoeostasis model assessment. Replacing red meat with a combination of plant-based alternatives was associated with higher insulin sensitivity (β = 10·5 percentage points, 95 % CI (4·1, 17·4)). Adjustment for waist circumference attenuated this association by 61·7 %. Replacing red meat with either legumes, nuts/seeds or wholegrains was likewise associated with higher insulin sensitivity. Point estimates were similar but less precise when replacing processed meat with plant-based alternatives. Our modelling suggests that regularly replacing red meat, and possibly processed meat, with plant-based alternatives may associate with higher insulin sensitivity. Further, abdominal adiposity may be an important mediator in this relationship. Our findings support advice to prioritise plant-based sources of protein at the expense of red meat consumption.

Substitution analyses of foods with varying fat quality and the associations with all-cause mortality and impact of the FADS-1 genotype in elderly men

PURPOSE

To investigate associations between substitutions of foods varying in fat quality and all-cause mortality in elderly Swedish men and to examine effect measure modification by a gene involved in fatty acid desaturation (rs174550 FADS1).

METHODS

Using Cox-regression models in the ULSAM cohort (n = 1133 men aged 71), we aimed to investigate; (1) Associations between the substitution of a nutrient or food for another on all-cause mortality (primary outcome) and CVD (secondary outcome) and (2) Associations between the addition of various fat-rich foods to the habitual diet and all-cause mortality and CVD. Subgroup analyses based on the rs174550 FADS1 genotype were conducted.

RESULTS

Over a mean follow-up of 11.6-13.7 years, n = 774 died and n = 494 developed CVD, respectively. No clear associations were observed for the vast majority of substitution nor addition models. Adding saturated fatty acids (SFA) on top of the habitual diet was however associated with an increased risk of mortality in men with the CT/CC-genotype [HR (95% CI) 1.44 (1.05, 1.97)]. Post-hoc analyses showed an inverse association of substituting SFA with carbohydrates [HR (95% CI) 0.79 (0.65, 0.97)], which was somewhat stronger in men with the CT/CC-genotype compared to men carrying the TT-genotype.

CONCLUSIONS

Few associations were observed between diet and all-cause mortality and CVD in this population. However, substituting SFA with carbohydrates was associated with lower mortality in post-hoc analyses and adding SFA to the habitual diet increased mortality in men with the CT/CC-genotype. The latter observation is novel and warrants further investigation in larger cohort studies including women.

Ultra-processed food consumption, mediating biomarkers, and risk of chronic obstructive pulmonary disease: a prospective cohort study in the UK Biobank

Background: Ultra-processed food (UPF) is a popular supplement in the UK and other developed countries. However, whether and how UPF intake is associated with chronic obstructive pulmonary disease (COPD) remains unclear. Objective: We aimed to examine the association between UPF consumption and COPD incidence and explore the potential mediating effects of COPD-related biomarkers. Methods: This prospective cohort study included 207 002 participants without COPD at recruitment and completed 24-hour dietary recalls. UPF was defined according to the NOVA classification system. Incident COPD was ascertained using electronic hospital and mortality records. Cox regression models were used to estimate UPF consumption and the subsequent risk of COPD. Substitution analysis was performed to assess the risk of COPD by substituting UPF with an equivalent proportion of unprocessed or minimally processed food (UNPF). Mediation analyses were performed to evaluate the contribution of biomarkers related to the lipid profile, glucose metabolism, and systemic inflammation to the observed associations. Results: During a median follow-up of 13.1 (interquartile range: 12.5-13.9) years, 4670 COPD events were recorded. The adjusted hazard ratio (HR) of COPD in the highest quintile versus the lowest quintile of the UPF consumption proportion (weight percentage of the UPF) was 1.22 (95% confidence interval [CI]: 1.11-1.34). There was a 10% elevated risk of COPD incidence per SD increase in UPF intake (HR: 1.10; 95% CI: 1.08-1.13). Replacing 20% of the UNPF weight with the UPF was associated with a 13% decrease in COPD risk (95% CI: 0.84-0.91). In mediation analyses, biomarkers explained 1.0-10.1% of the association between UPF intake and COPD. Results from stratified and sensitivity analyses further support the robustness of these findings. Conclusions: Elevated UPF consumption was associated with a higher risk of COPD, and this association was primarily mediated by glucose, inflammation, and lipids, whereas substituting UNPF for UPF was associated with a decreased risk of COPD.

Association of Macronutrients Intake with Body Composition and Sarcopenic Obesity in Children and Adolescents: A Population-Based Analysis of the National Health and Nutrition Examination Survey (NHANES) 2011-2018

The association of macronutrients intake with body composition and sarcopenic obesity remains uncertain in children and adolescents. We aimed to explore the association between macronutrients intake and body composition, especially sarcopenic obesity, in children and adolescents residing in the United States. The study utilized data from 5412 participants aged 6-17 years who attended NHANES between 2011 and 2018. Body composition was assessed using DXA, and nutrient intake was based on 24-h recall. Multivariable linear regression and multinomial logistic regression were used. The unweighted prevalence of sarcopenic obesity was 15.6%. A higher percentage of energy (5 %E) from fat was inversely associated with muscle mass but positively associated with fat mass and sarcopenic obesity. Substituting carbohydrate (5 %E) with fat decreased muscle mass by 0.03 (95% CI 0.01 to 0.06) but increased fat mass by 0.03 (95% CI 0.01 to 0.06) and increased the prevalence of sarcopenic obesity by 254% (95% CI 15% to 487%). Replacing protein intake with fat intake also increased the OR of sarcopenic obesity (OR, 2.36 [95% CI 1.18 to 3.18]). In conclusion, a high-fat diet, coupled with low carbohydrate/protein intake, is associated with sarcopenic obesity among children and adolescents. The change in children's diet towards a healthy diet with low fat composition may help prevent sarcopenic obesity. However, randomized clinical trials or longitudinal studies are needed to further validate our findings.

Dietary intake of total, heme and non-heme iron and the risk of colorectal cancer in a European prospective cohort study

BACKGROUND

Iron is an essential micronutrient with differing intake patterns and metabolism between men and women. Epidemiologic evidence on the association of dietary iron and its heme and non-heme components with colorectal cancer (CRC) development is inconclusive.

METHODS

We examined baseline dietary questionnaire-assessed intakes of total, heme, and non-heme iron and CRC risk in the EPIC cohort. Sex-specific multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were computed using Cox regression. We modelled substitution of a 1 mg/day of heme iron intake with non-heme iron using the leave one-out method.

RESULTS

Of 450,105 participants (318,680 women) followed for 14.2 ± 4.0 years, 6162 (3511 women) developed CRC. In men, total iron intake was not associated with CRC risk (highest vs. lowest quintile, HRQ5vs.Q1:0.88; 95%CI:0.73, 1.06). An inverse association was observed for non-heme iron (HRQ5vs.Q1:0.80, 95%CI:0.67, 0.96) whereas heme iron showed a non-significant association (HRQ5vs.Q1:1.10; 95%CI:0.96, 1.27). In women, CRC risk was not associated with intakes of total (HRQ5vs.Q1:1.11, 95%CI:0.94, 1.31), heme (HRQ5vs.Q1:0.95; 95%CI:0.84, 1.07) or non-heme iron (HRQ5vs.Q1:1.03, 95%CI:0.88, 1.20). Substitution of heme with non-heme iron demonstrated lower CRC risk in men (HR:0.94; 95%CI: 0.89, 0.99).

CONCLUSIONS

Our findings suggest potential sex-specific CRC risk associations for higher iron consumption that may differ by dietary sources.