Study on blueberries, protein and exercise for improving frailty and cardiovascular disease (STRONG): protocol for a parallel-group randomised controlled superiority trial in older Nova Scotians

Abstract

Introduction

As frailty and cardiovascular disease are closely linked, lifestyle interventions to reduce frailty could improve cardiovascular health in at-risk individuals. Previous clinical trials have shown that exercise reduces frailty, both individually and in combination with dietary protein. Blueberries can improve vascular health and protect against cellular senescence. The objective of the study on blueberries, protein and exercise for improving frailty and cardiovascular disease (STRONG) is to determine whether a year-long treatment including blueberries, protein and exercise will reduce frailty and improve cardiovascular health in older adults of both sexes.

Methods and analysis

STRONG is a pragmatic randomised controlled trial 1 year in length with an additional year of follow-up. We will recruit 240 adults aged ≥65 years (120 females and 120 males) in Nova Scotia, Canada. Participants will be randomly assigned to the STRONG intervention group (1 cup of blueberries/day, 30 g of protein powder/day and three 60-min multimodal exercise sessions with an exercise physiologist/week) or the control group (usual care). Measured outcomes will include changes in frailty (eg, frailty scores, balance, falls, changes in life-space, social engagement), cardiovascular disease risk factors (eg, blood pressure, lipid profile, Framingham risk score), cardiovascular health (eg, 6-min walk test, cardiac function measures), inflammatory markers, physical fitness, diet quality and health-related quality of life. All measures will be collected at baseline and 12 months, with many also repeated at the 3-, 6-, 9- and 24-month time points.

Ethics and dissemination

This study has received approval from the Nova Scotia Health Research Ethics Board (ROMEO #1031656) and is being conducted according to Consolidated Standards of Reporting Trials (CONSORT) and Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines. The findings will be disseminated through peer-reviewed publications, academic and community presentations, lay materials on our research team websites and social media and summaries/briefings to stakeholders that will be directed toward registered dietitians, exercise physiologists and clinicians.

Trial registration number

NCT06693271.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• STRONG is a randomised controlled trial with a year-long intervention duration and a large and sex-balanced sample size.

• The multidomain intervention of blueberries, protein and multimodal exercise programme is a novel and relatively simple, low-cost treatment that could be prescribed by healthcare professionals to older patients.

• Despite three self-reported diet-related measures to track blueberry and protein consumption, the extent of adherence to the nutrition component cannot be perfectly verified, and participants may grow tired of eating blueberries and protein daily.

• While it is a strength that the exercise workouts are supervised and personalised, participants may have difficulty attending three exercise sessions per week for a full year.

• STRONG will have repeated measurements of many outcome variables related to frailty, cardiovascular disease, lifestyle and well-being.

Introduction

Cardiovascular disease (CVD) is a major cause of disability and death worldwide.^{1 2} Despite recent declines in CVD mortality, increasing numbers of adults are living with various types of CVD,² especially older adults. The risk of developing CVDs, including coronary heart disease, heart failure and arrhythmias, increases dramatically with age in both sexes.^3–5 However, the disease profiles differ in a sexually dimorphic way, with females and males aged 65+ years being at different risks of developing certain CVDs and often experiencing different outcomes.^3–7

Frailty is a state of physiological decline or accelerated ageing that increases vulnerability to stressors, especially by reducing the ability to withstand potentially damaging stress (robustness) or repair damage that arises (resilience).^{8 9} In older adults with CVD, frailty robustly increases the risk of major cardiovascular events, hospitalisation and death.^10–14 For example, frail people with atrial fibrillation are much more likely to experience stroke, anticoagulant bleeding complications, unplanned hospitalisation and death than age-matched people who are not frail.^{15 16} Similarly, the chance of developing heart failure and its complications is higher in frail older people than in their fit counterparts.^17–19 Higher levels of cardiovascular complications and mortality are even seen in frail individuals with no history of CVD,¹³ providing compelling evidence that frailty influences the risk and expression of CVD.^{13 20 21} Since many of the proposed mechanisms of frailty, such as chronic inflammation and oxidative stress, are also strongly associated with CVD in older people,^{22 23} the Geroscience Hypothesis proposes that therapies directed at one or more fundamental frailty mechanisms to reduce the degree of frailty can delay or prevent age-related diseases such as CVDs.^{24 25}

Growing evidence shows that the degree of frailty can be attenuated by lifestyle interventions of exercise and diet.²⁶ Systematic reviews and meta-analyses of studies in older adults at risk of frailty have generally concluded that resistance training and aerobic exercise reduce frailty,^27–30 and studies suggest this may occur through a reduction in chronic inflammation, particularly IL-6 and TNFα.^31–33 Other benefits of regular exercise include improved aerobic fitness, cardiac repair, lower blood pressure, improved metabolic function and enhanced insulin sensitivity, all of which help lower the degree of frailty.^{31 32 34 35}

Evidence demonstrates that the addition of protein supplementation to exercise reduces frailty, particularly in females.^{36 37} This is likely due, at least in part, to the preservation of muscle mass to maintain mobility and reduction of pro-inflammatory cytokines (IL-6, MCP-1).^{38 39} Additionally, dietary interventions with foods such as blueberries that are high in antioxidants have been proposed as potential treatments to decrease the oxidative stress and inflammation associated with sarcopenia in frail older people.^{22 40–42} Blueberries provide potent anti-inflammatory and antioxidant properties that can improve vascular function, lower blood pressure and reduce inflammatory biomarkers.^{22 23 42} Therefore, targeting the pathways of inflammation and oxidative stress that jointly affect frailty and CVD, we have developed a protocol to test the effect of a multidomain intervention in the study on blueberries, protein and exercise for improving frailty and CVD (STRONG) in older adults (figure 1). We hypothesise that compared with usual care, the STRONG intervention will have beneficial effects on indicators of frailty and cardiovascular health in older adults, and that these effects will be sex-specific and at least partially mediated by changes in chronic inflammation.

Figure 1. The STRONG intervention: Combining protein, blueberries and resistance/aerobic exercise may reduce frailty and improve CV health, further reducing frailty. CV, cardiovascular; IL-6, interleukin-6; TNF-α, tumor necrosis factor alpha.

Methods and analysis

Objectives

The primary objective of STRONG is to determine whether the intervention reduces frailty and improves cardiovascular health in older adults. The secondary objective is to determine whether any effects of this intervention are mediated or modified, in part, by effects on markers of chronic inflammation. Since older women are frailer than older men,^43–45 and it is unclear whether the relationship between frailty and CVD differs between the sexes, we will examine whether any effects are sex-specific.

Trial design

STRONG is a pragmatic randomised controlled trial in community-dwelling older adults that is 1 year in length with an additional year of follow-up. This is a parallel-group superiority trial in which the intervention group will receive the STRONG treatment (blueberries, protein and exercise), and the control group will receive usual care (figure 2) in a 1:1 ratio with equal numbers of females and males in each group. The study has been registered on ClinicalTrials.gov (NCT06693271) (table 1).

Figure 2. STRONG study design, procedures and stages.

Table 1. Structured summary of STRONG trial design and methods with items from the WHO Trial Registration Data Set.

Public title	Study on blueberries, protein and exercise for improving frailty and cardiovascular disease (STRONG)
Scientific title	Study on blueberries, protein and exercise for improving frailty and cardiovascular disease (STRONG)
Contact for public inquiries	STRONG@nshealth.ca
Contact for scientific inquiries	Drs. Scott Grandy (scott.grandy@dal.ca), Susan Howlett (susan.howlett@dal.ca), Kenneth Rockwood (kenneth.rockwood@nshealth.ca), Leah Cahill (leah.cahill@dal.ca)
Primary registry and identifying number	NCT06693271 (clinicaltrials.gov)
Sources of monetary support, funder contact info	River Philip Foundation, https://riverphilipfoundation.ca/contact-us
Sources of material support	None
Country of recruitment	Canada
Health conditions or problems studied	Frailty and cardiovascular disease in older adults (aged 65+years)
Study type	A pragmatic parallel-group randomised controlled trial
Sample size	n=240
Study intervention	Treatment group: STRONG intervention (n=120, equal number of each sex) Participants receive the usual standard of care with the addition of: 30 g/day (one scoop) of colourless, tasteless protein powder for 12 months 1 cup (150 grams) of blueberries per day for 12 months Participation in three 60- min multimodal exercise sessions per week for 12 months Control group: Usual medical care (n=120, equal number of each sex) Participants receive the usual standard of care
Inclusion and exclusion criteria	Inclusion criteria: Must be 65 years of age or above Must be willing to participate in a low to moderate exercise programme Exclusion criteria Medical and cognitive conditions that prevent participation in the intervention Allergies or dislikes of blueberries and/or protein powder
Date of first enrolment (planned)	December 2025
Primary outcomes	Frailty (frailty index, clinical frailty scale, pictorial fit-frail scale) Cardiovascular risk (blood lipid profile, blood pressure, Reynold’s risk score, Framingham risk score, ejection fraction and global longitudinal strain measured by echocardiography) Inflammatory and anti-inflammatory markers (IL-1β, TNFα, IL-6, IL-10)
Key secondary outcomes	Physical fitness (6-min walk test) Diet quality (measured by diet intake from the ASA24) Household food work (measured by time-use for food work and food work-related stress measured from the FIA-Q10) Quality of life (EQ-5D-5L)
Ethics review	This study has been approved by the Nova Scotia Health Research Ethics Board (ROMEO #1031656)
Individual trial participant data sharing statement	Deidentified data available on reasonable request.
Protocol version	V.6, revised June 2026.

ASA24, Automated Self-Administered 24-Hour Dietary Recall; CFS, Clinical Frailty Scale; EQ-5D-5L, five- level EuroQol five- dimensional questionnaire; FI, frailty index; FIA-Q10, Foodwork Interactional Assessment 10 Questions; GLS, global longitudinal strain; IL-6, interleukin-6; IL-10, interleukin-10; IL-1β, interleukin-1 beta; TNFα, tumor necrosis factor alpha.

Setting

The study site is the Physical Activity and Cancer (PAC) Lab in the Queen Elizabeth II Health Sciences Centre in Halifax, Nova Scotia, Canada. Nova Scotia has especially high rates of people living with various types of CVD, due in part to its rapidly ageing population.^{2 46}

Patient and public involvement

The STRONG Patient Public Partner Panel is comprised of five older adults living in Nova Scotia who became involved in STRONG at the protocol development stage when they provided ideas for the feasibility of the intervention and the recruitment and retention of participants (such as locations to place recruitment posters, communication check-in schedules and methods). They were also involved in the development of participant-facing study materials and will continue to be engaged throughout the study stages, including the contextualisation and dissemination of study findings.

Participants and recruitment

To be eligible to participate in this study, individuals must meet two inclusion criteria: (1) they must be at least 65 years of age and (2) they must be willing to participate in a low- to moderate-exercise programme. Individuals are ineligible to participate in this study if they have medical and cognitive conditions that prevent participation in the intervention or if they have allergies or dislikes to blueberries or protein powder. All medications and treatments will be permitted before and during the study.

Participants will be referred from several relevant clinics at Nova Scotia Health, including Nova Scotia Health’s Geriatric Ambulatory Care Clinics, Women’s Heart Health Clinic and the Nova Scotia Rehabilitation Centre. Participants may also self-refer through Nova Studies Connect, an online platform that connects Nova Scotians with research opportunities and posters placed in the community (eg, seniors’ centres, hospitals, pharmacies, public libraries and grocery stores) and social media.

Sample size and statistical power

The study will aim to recruit 240 participants (120 in each trial arm, 60 males and 60 females per arm). Sample size calculations were conducted for 0.80 power for three primary outcomes stratified by sex. First, we calculated that we require a sample size of at least 180 (90 male and 90 female participants) to have 80% power to detect a 0.07 change in frailty as measured by a frailty index (FI) with an SD of 0.1. This represents a moderately large effect size; a clinically meaningful difference in the FI has been observed to be as small as 0.03.^47–49 We also require at least 180 participants to have 80% power to detect a difference in HDL cholesterol of 0.26 mmol (an increase of 0.26 mmol has been associated with a 30% decreased risk of coronary artery disease),⁵⁰ assuming an SD of 0.41 mmol.⁵¹ However, we require a larger sample size of 55 per group (total=220) to detect a change in inflammatory biomarkers such as IL-6, as in the study by Hubbard et al.⁵² Therefore, although if we had only one primary outcome, it would be the FI, we aim to recruit 240 participants to allow for the mechanistic study of underlying pathways such as inflammation and to account for an 8% loss to follow-up or adherence as observed in similar studies.^{37 53 54}

Randomisation

The unit of randomisation is the individual participant, and the allocation ratio is 1:1. Block randomisation will be used to balance the intervention and control groups by sex and ensure equal numbers in each group while allowing continuous enrolment to occur, with block sizes randomised to n=2, n=4 and n=6.⁵⁵ Blocking will be performed using the randomizeR package in R by a third-party biostatistician.⁵⁶ Prior to the study, the biostatistician will generate sequentially numbered folders to assign participants to the intervention or control groups. Only the biostatistician will have access to the files with the sequence. The group assignment will be placed in sealed envelopes to be given to the study coordinator to hand to the participants after the collection of their baseline data.

Blinding

Due to the nature of the intervention, participants and members of the data collection team administering the intervention cannot be blinded to the treatment groups. Baseline measurements will be taken before participants are randomised so that participants and assessors are blinded during baseline data collection.³⁵ Whenever possible, data collectors who are not administering the intervention will be blinded. For example, the sonographers will be blinded. The study statistician and data analysis team will be blinded to allocation during analysis.⁵⁷

Intervention and comparator

The STRONG treatment consists of 30 g/day of a colourless, tasteless whey protein isolate powder supplement, 1 cup of blueberries per day and three 60-min multimodal exercise sessions per week. This will be compared with a control group who will receive their usual (standard) medical care, as in most pragmatic trials that are performed to determine if an intervention can improve current practice.^{58 59} Usual medical care will include participants’ usual diet and exercise routines, including any prescribed by their healthcare teams (it may or may not include diet and exercise recommendations from their healthcare professionals) and may differ between participants. Concomitant care will be permitted during the trial. Changes to participants’ health and medications will be recorded at study assessments.

Participants in the STRONG treatment group can consume the blueberries and protein powder in any manner they wish; the intake method is not prescriptive. Whey protein was chosen over the other common milk protein, casein, for several reasons. A 2020 systematic review of the effects of cow-milk protein supplementation among older adults reported that the beneficial effect is stronger for whey protein supplementation due to its high leucine content.⁶⁰ It is also easier to mix with liquids. More specifically, we selected whey protein isolate due to its processing that removes fat and lactose, making it suitable for a study not testing fat intake and for people who are lactose intolerant. An alternate suitable protein powder, such as pea protein isolate powder, pumpkin, rice, etc., is available for participants who do not eat milk products. Studies of protein supplementation in older adults range from 15 g/day to 60 g/day, but a systematic review and meta-analysis of different doses of whey supplementation on inflammatory cytokines in older adults reported that only an intervention dose of ≥30 g/day led to significant decreases in serum IL-6 levels.⁶¹ The planned daily dose of 30 g provides ~25 g of protein and is a substantial amount of protein powder without being too burdensome and will be consumed in addition to other proteins that are part of participants’ usual diets.

The dose of 1 cup/day of blueberries is based on the longest blueberry trial to date (6 months), in which Curtis et al observed that a daily intake of 1 cup of blueberries improved measures of vascular health, including endothelial function, systemic arterial stiffness, nitric oxide bioactivity, high-density lipoprotein cholesterol and reduced cyclic guanosine monophosphate concentrations in older adults with metabolic syndrome.⁵³ By contrast, a half cup per day had no effect on any biomarkers.⁵³

Participants will take part in three 60-min multimodal exercise sessions per week. Each session will include a warm-up and cool-down (~5 min each), 5–20 min of aerobic exercise (depending on fitness level) and 30–45 min of resistance training. Exercise intensity will be monitored using the Category Ratio 10 (CR10) rating of perceived exertion scale.^{62 63} All participants will be given a copy of the CR10 scale and taught how to use it. Early in the intervention and for participants with low fitness levels, exercise should elicit a CR10 score of 3–5, which indicates light to moderate exertion. Exercise will elicit higher scores (up to 10, which corresponds to maximum effort) as fitness levels and familiarity with the exercise programme increase over the 12-month intervention. All exercise programmes will be designed specifically for the individual and supervised by a clinical exercise physiologist. Exercise sessions will be attended in person, except in circumstances in which participants cannot travel. In these cases, virtual sessions will be delivered via a secure Nova Scotia Health Zoom link.

Adherence and retention strategies

We have taken several steps to help facilitate participants’ adherence to the STRONG protocol. All participants are required to attend in-person study visits for the collection of outcome measures. Control group retention will be enhanced by giving the option to review their assessment results at each timepoint. Participants in the control group will be offered a $50 grocery store gift card at each assessment to compensate them for their time. This amount aligns with the current living wage in Nova Scotia and is fair without constituting undue inducement.^{64 65} Participants will also be reimbursed parking fees and bus tickets.

Participants in the STRONG treatment group will be provided with the protein powder and blueberries (usually frozen wild blueberries) during their exercise sessions. At the start of the study, participants will also be supplied with a 1-cup measure for the blueberries, a scoop for the protein powder, a personal blender and a booklet with ideas for how to consume the blueberries and protein powder. A video demonstrating how to make a smoothie using protein powder and blueberries is available for participants. To assist with adherence to the STRONG treatment, both the nutrition and physical activity interventions can be adapted to meet participants’ preferences. If participants do not want to continue with the whey protein, we can provide an alternate protein powder such as pea protein powder in its place. Should participants express that they are tired of eating 1 cup of blueberries per day, the study team will consider allowing participants to occasionally substitute blueberries with other berries that provide similar nutritional benefits. If the diet component causes adverse side effects at the beginning of the study, we will suggest that participants start with a lower dose of blueberries and protein powder and gradually increase the dose each day. Workouts will be individualised by a clinical exercise physiologist for each participant based on their medical history and baseline fitness assessment. Exercise sessions will be adjusted over the year to ensure progression as well as to account for changes in health and fitness. For example, the exercise physiologist can adjust the exercise or the difficulty level if a participant is experiencing pain, soreness or fatigue after their exercise sessions. If participants are unable to attend an exercise session at the PAC Lab due to extenuating circumstances, exercise sessions can also be conducted virtually and the nutrition component can be delivered to participants’ homes.

Adherence assessment

Deviation in STRONG intervention uptake and reasons for non-adherence will be assessed through returned nutrition checklists, the nutrition check-in, the Automated Self-Administered 24-Hour Dietary Recall (ASA24) and through recording attendance at workouts. The nutrition checklist tracks participants’ daily protein powder and blueberry consumption. The nutrition check-in consists of eight questions to assess how the study intervention is being experienced by the participants. It is completed after the first week and after the first month by phone call with a member of our research team. Hereafter, participants will switch to filling out the nutrition check-in form on paper or online monthly until the end of the STRONG treatment period through REDCap unless a participant requests a phone call instead. Compliance with the exercise component of the STRONG treatment will be assessed through attendance at workouts as recorded by the exercise physiologist and/or the study coordinator (including completion of exercises). For missed exercise sessions, the reasons for lack of treatment compliance will be noted. For the per-protocol analysis, adherence to the STRONG treatment will be defined as completing the full STRONG nutrition dose 5 out of 7 days on average for a year (~70%) and completing 70% of workouts averaged over the year.⁶⁶ The intention-to-treat analysis will include all participants in their intervention group regardless of their level of adherence.

Study outcome measures

Participants will receive a series of assessments as scheduled in table 2, which also provides assessment tool names and methods. All assessments will be performed at baseline and 12 months, and most are also assessed at 3 months, 6 months, 9 months and 1-year post-study (24 months). Primary outcomes are frailty (measured using an FI, the clinical frailty scale and the pictorial fit-frail scale), cardiovascular risk factors (blood lipid profile, blood pressure, Reynolds and Framingham risk scores and ejection fraction and global longitudinal strain (GLS) measured by echocardiography) and inflammatory/anti-inflammatory markers (IL-1β, TNFα, IL-6 and IL-10). The primary outcomes were chosen because they are indicators of frailty level and cardiovascular health, which are the health issues being targeted by this intervention. Secondary outcomes are physical function (eg, functional fitness measured by the 6-min walk test, upper body strength measured by a hand grip dynamometer, lower body strength measured by the sit-to-stand test, balance measured by the single-leg balance test, mobility measured by the sit-and-reach test and the shoulder girdle test), diet quality (measured by diet intake from the ASA24), household food work (measured by time-use for food work and food work-related stress measured from the FIA-Q10) and quality of life (measured by the EQ-5D-5L). As additional outcomes, we will record incident outcomes related to frailty (eg, impaired balance, falls, hip fractures, changes in life-space, social engagement) and CVD (eg, angina, heart attack, stroke) in case there are enough events to analyse as a study outcome.

Table 2. The schedule of assessments in STRONG occurring during study visits.

Assessment	Baseline	3 months	6 months	9s month	12 months	1 year post
General
Blood collection*	♥				♥	♥
Health history form	♥
Changes to health status and medication(s)		♥	♥	♥	♥	♥
Anthropometrics – height, weight, waist girth (85)	♥	♥	♥	♥	♥	♥
Vitals – resting heart rate, resting blood pressure (85)	♥	♥	♥	♥	♥	♥
EQ-5D-5L (quality of life) (86)	♥	♥	♥	♥	♥	♥
Physical activity
Functional fitness – 6-min walk test(85)	♥	♥	♥	♥	♥	♥
Grip strength (right and left) (85)	♥	♥	♥	♥	♥	♥
30-s sit-to-stand task(87)	♥	♥	♥	♥	♥	♥
Single leg balance (right and left)(85)	♥	♥	♥	♥	♥	♥
Shoulder range of motion (shoulder girdle test)(85)	♥	♥	♥	♥	♥	♥
Hamstring flexibility (sit and reach test)(85)	♥	♥	♥	♥	♥	♥
Physical activity scale for the elderly (PASE) International sedentary assessment tool (ISAT)(88)	♥	♥	♥	♥	♥	♥
Physical activity readiness questionnaire (PAR-Q+)(89)	♥	♥	♥	♥	♥	♥
Nutrition/diet
ASA24 dietary assessment (90)	♥		♥		♥	♥
Household foodwork interactional assessment 10 questions (FIA-Q10) (91)	♥		♥		♥	♥
Frailty and ageing
Frailty index†(92)	♥	♥	♥	♥	♥	♥
Clinical frailty scale (CFS) (93)	♥	♥	♥	♥	♥	♥
Pictorial fit-frail scale (PFFS) (94)	♥	♥	♥	♥	♥	♥
Cognitive function FreeCog Questionnaire (95)	♥	♥	♥	♥	♥	♥
Cardiovascular
Echocardiogram	♥				♥
CVD risk score (Framingham, Reynolds)‡(97,98)	♥				♥	♥

♥=assessment is occurring.

^*Blood collection – blood lipids, glucose, HbA1c, electrolytes, enzymes, haematological profiles and systemic inflammation will be analysed.

^† Will be created using assessment data collected during the study.

^‡Will be calculated using results from blood analysis.

ASA24, Automated Self-Administered 24-Hour Dietary Recall; CVD, cardiovascular disease; EQ-5D-5L, five-level EuroQol five-dimensional questionnaire; FI, frailty index.

An FI will be constructed based on deficit accumulation using established techniques.^{67 68} Variables to be used as deficits will be obtained from the STRONG assessments (table 2) as presented in table 3. For example, the health history form contains information about diverse bodily systems, including respiratory, cardiovascular, sensory and urogenital systems, as well as self-rated health, medication use and other conditions. This form contains up to 55 potential deficits that can be used to create an FI (table 3). Using a protocolised data-driven approach to deficit selection, we will determine which health attributes count as deficits using the following standard criteria^{67 68}: (1) deficit prevalence must increase with age, (2) deficit should relate to adverse health outcomes, (3) deficit should not be present in all individuals at a young age, (4) deficits should encompass a range of bodily systems and (5) each deficit should have <5% missing data. This data-driven approach is based on the distribution of scores for each deficit to determine which items fulfil these criteria. At least 30 variables are needed to ensure a stable FI,⁶⁷ but we have many more potential deficits available (table 3). This will allow us to screen variables for suitability by removing any items from the FI that might also be an outcome in our study, as recommended.⁶⁸

Table 3. Variables for construction of a frailty index (FI).

Assessment	Potential deficits
Health history form	Self-rated health Respiratory disorders Mouth and soft tissue disorders Cardiovascular health Urogenital conditions Disorders of the special senses Inflammatory and immune disorders Medication use Other (up to 55 potential deficits)
Anthropomorphic measures	Height Weight Waist circumference Body mass index
Vital signs	Heart rate Systolic blood pressure Diastolic blood pressure Pulse pressure (calculated)
EQ-5D-5L (quality of life)	Mobility Self-care Usual activities Pain/discomfort Anxiety/depression
Physical fitness	6-min walk test Hand grip strength Shoulder range of motion 30-s sit-to-stand test Sit and reach test/hamstring flexibility One-leg stance
Physical activity scale for the elderly (PACE)	Leisure time activity Household activity Work-related activity Other (up to 10 deficits)
International sedentary assessment tool (ISAT)	Time spent sitting on weekdays Time spent sitting on the weekend
Physical activity readiness questionnaire (PAR-Q+)	Chest pain at rest or during physical activity Dizziness at rest or during activity Loss of consciousness at any time Past medical history (up to 41 potential deficits)
Cognitive function FreeCog questionnaire	General knowledge Orientation in time and place Memory Calculation Executive function (eg, response to an emergency) Other cognitive and executive function measures (up to 17 potential deficits)
Blood collection	Lipids Random glucose Haemoglobin A1C Electrolytes Enzymes Proteins Haematological profile Other (up to 34 potential deficits)

At least 30 variables are needed to ensure a stable FI but we have many more potential deficits available. This will allow us to screen variables for suitability and to remove any items from the FI that will also be analysed as an outcome in our study, as recommended.

Using a protocolised data-driven approach to deficit selection, we will determine which health attributes count as deficits using the following standard criteria : (1) deficit prevalence must increase with age, (2) deficit should relate to adverse health outcomes, (3) deficit should not be present in all individuals at a young age, (4) deficits should encompass a range of bodily systems and (5) each deficit should have <5% missing data.

EQ-5D-5L, five-level EuroQol five-dimensional questionnaire.

Venipuncture will be performed by a nurse/phlebotomist to collect approximately 45 mL of non-fasting blood samples. Half of the blood sample will be sent to the Nova Scotia Health hospital laboratory for lipid and glucose analysis. The remaining ~20 mL of blood will be processed and stored as plasma at −80 °C for up to 10 years for measurement of inflammatory biomarkers, in case any sample requires re-analysis, and for additional biomarkers measured in the future if consent for future research is obtained. Blood samples will be collected in a non-fasting state because at the time of blood draw, participants will also undergo physical assessments involving exertion and exercise testing; requiring an 8–12 hour fast prior to the study visit could increase the risk of hypoglycaemia and vasovagal events during testing, which is an unacceptable safety risk. All participants will be assessed under the same protocol, and the timing of meals is not expected to differ systematically by group. Non-fasting measurements are increasingly accepted in clinical and research settings.⁶⁹ Non-fasting samples may also improve participant compliance⁷⁰ and better reflect usual physiological conditions, since individuals spend most of their time in a non-fasted state.⁷¹ Blood pressure (four or five recordings, including one standing recording) will be recorded by study staff at each time point with an electronic sphygmomanometer. Cardiac function will be assessed with a portable echocardiography machine by a registered sonographer. Left ventricular function will be assessed using transthoracic echocardiography (TTE) (Vivid iqIQ Ultra Edition Premium, GE Medical Systems, Milwaukee, WI; standard multi-frequency transducer & EchoPACac, V.2046). TTE will be conducted by professional sonographers at each site, and all images will be analysed by a trained study staff member. Left ventricular dimensions and ejection fraction (LVEF) will be determined using the 2D parasternal and apical view images according to the American Society of Echocardiography guidelines.⁷² LVEF will be calculated using the modified biplane Simpson’s method.⁷³ Tissue Doppler imaging will be performed at the basal segment of the lateral mitral annulus to evaluate longitudinal endocardial velocity. The measured indices will include systolic (S’), early diastolic (e’) and late diastolic (a’) velocities. GLS will be calculated using automated function imaging using the apical long-axis, 4-chamber and 2-chamber views [75,76].

Data collection and management

Study assessors receive training and will undergo a test data collection round with the STRONG Patient Public Partner Panel members. Data from the data collection forms will be entered into REDCap, a secure web-based application that allows data to be stored securely and locally on Nova Scotia Health servers [79,80]. Efforts will be made to prevent missing values at the data collection stage. We will follow-up quickly with participants who do not attend their scheduled study visits for in-person data collection. If participants are unable to come to an in-person appointment for reasons that cannot be solved by simply rescheduling (eg, long-term illness, transportation problems), we will have them complete the questionnaires and some assessments virtually and travel to their home for the other components of the assessment if time and finances permit. We aim for a data cut-off window of 10 business days within each predefined study milestone, but extensions may be made on a case-by-case basis.

All data will be entered into REDCap for secure data storage by the study coordinator, trainees, research associate or exercise physiologist. Physical case report forms will be used to record all outcome measures captured during in-person study visits (eg, weight, height, blood pressure). When transferring the information from the paper source into REDCap, we will employ a double data entry check method by two individual team members for the first 10 participants and every 20^th participant afterwards. Range checks for data values will be performed after the first 10 participants and at 3-month intervals. When applicable, our REDCap instruments have also activated the available settings (eg, validation, max/min values, variable types) to help ensure the quality of the data being entered into the system. Prior to their use, all REDCap instruments have undergone thorough testing by our study coordinator and research associate. Exported REDCap data will be cleaned using Microsoft Excel and Stata, and an audit trail will be maintained of all corrections made to the dataset. For validated or standardised questionnaires, such as the ASA24 [81], data will be cleaned according to recommended procedures.

Statistical analysis

Interim analyses for safety and sample size will be conducted. Descriptive analysis will be performed for the variables collected at each time point, with the distribution of continuous variables examined. The effectiveness of the intervention versus control for each outcome measure will be compared using repeated measures mixed models in both an intention-to-treat analysis and a per-protocol analysis defined by adherence to the study intervention. Continuous variables such as the FI, some physical fitness measures, dietary intakes, questionnaire scores and biomarker levels will also be analysed by examining the mean change from baseline to the time points of interest (eg, 3 months, 6 months, 9 months, 12 months, 1-year post-study, depending on the measurement frequency for each variable). Mean values will be plotted as a function of time for each intervention group (control group and treatment group). Serum concentrations of biomarkers such as cytokines will be plotted as a function of frailty scores for each intervention group; biomarkers will not be included in the FI for this analysis if they meet the criteria for inclusion in the FI. Relationships between biomarkers and cardiovascular risk factors will be assessed through repeated measures linear mixed effects regression models. For categorical outcomes, if the number of events is too low to perform regression models, they will be analysed descriptively. Covariables in adjusted models will include sex and any characteristics that are not balanced between the treatment and control groups at baseline. If missing data are present, we will describe patterns and predictors of missingness to assess the plausibility of the missing-at-random assumption. Primary analyses will use multiple imputation by chained equations, including covariables in the imputation model, with ≥20 imputations and estimates combined using Rubin’s rules. Results will be compared with complete-case analyses and sensitivity analyses under alternative missing-not-at-random assumptions. Statistical significance will be set at p=0.05. Analyses will be run within males and females separately for comparison as well as all together.

As well as stratifying by sex, additional subgroup analyses will examine the effect in separate groups defined by important demographic and risk factors such as smoking status (yes/no), diabetes (yes/no), obesity (yes/no), physical activity level (meeting guidelines: yes/no), dietary intake (meeting guidelines: yes/no), diet quality (low/medium/high), health-related quality of life (no problems, any problems), previous CVD at baseline (yes/no) and baseline degree of frailty (high/not high). These subgroup analyses are likely to be underpowered and so are considered exploratory. We will also conduct a subgroup analysis by level of compliance with the STRONG treatment.

Potential harms

Potential harms are noted on the consent form and intervention instructions. The exercise sessions might result in short-term fatigue, muscle soreness, stiffness, or injury. The diet intervention could cause short-term digestive discomfort as participants adjust to the change. Eating blueberries and protein powder can potentially cause gas if inadequate amounts of fluids (ie, water) are consumed, and stools may become a different colour (darker, bluer) from the blueberries, but these are not dangerous. Blueberries can stain fabrics and temporarily colour the mouth and fingers. The consent form and intervention instructions note that the daily dose of blueberries provides 15 g of carbohydrate and needs to be included in the daily carbohydrate intake plan of participants with diabetes (the protein powder does not contain carbohydrate). The consent form and intervention instructions also note that blueberries are perishable and provide instructions regarding food safety. Dietary interventions pose the risk of unknown allergies, so we provide instructions on what to do if a participant experiences an allergic reaction. The nutrition check-in sheet assesses whether participants experience any negative effects or discomforts related to consuming the blueberries or protein powder. Similarly, the exercise physiologist will check-in on participants at the beginning of each workout session in terms of general discomforts related to their physical activity. Any additional harm assessment is initiated and defined by participants’ reports. Study staff are aware of when to consult expert members of the research team (eg, nurse, dietitian, physician) for participants’ reported harms and when it is appropriate to complete an adverse event report. They are also trained in what to do if a participant becomes injured or sick during a workout or assessment. Should any participants become ill or injured as a result of the study, necessary medical care will be provided that is consistent with provincial health coverage. Participants may receive results on their STRONG outcome assessments that need to be followed up with their healthcare providers. Otherwise, there will be no care required for participants after completion of the study.

Data and trial monitoring

Due to the low risk associated with this study, there is no formal data monitoring committee. The initiation of harm assessment is per participant report. Adverse events will be recorded with a detailed outline regarding the event (eg, start and stop date, severity, outcome). The study team reviews all adverse events, and the steering committee (which is independent from the funder and composed of LEC, SEH, SG and KR) will track the number of adverse events and will make the decision to modify or end the intervention if needed. If required to make these decisions, interim analyses will be conducted.

Withdrawal

Participants will discontinue the intervention in a safety exclusion from the STRONG study (withdrawal) if the study doctor (a steering committee member) determines that their health is at risk from proceeding with the STRONG treatment. New participants will be recruited to replace participants who are withdrawn in the first half of the study duration. If participants are withdrawn in the second half of the study duration, withdrawn participants will be asked to continue with their regular study visits to capture outcome measures at the planned time points.

Ethics and dissemination

Ethical approval has been obtained from the Nova Scotia Health Research Ethics Board (#1031656). Any potential modifications to the protocol will be decided on by the steering committee, submitted to the REB for approval and communicated to relevant parties, including ClinicalTrials.gov. All procedures performed will be in accordance with the ethical standards of the Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans - TCPS (2022) [82]. Formal consent will be obtained from participants by the study team after the participant has reviewed the informed consent form in REDCap. The participant will have the opportunity to ask questions. Prospective participants may take as long as they need to consider the study and may review the informed consent form (see online supplemental file 1) with their physician or others as needed. The study will not be recruiting participants who lack decisional capacity. The informed consent form asks for additional (optional) consent of participants for future secure linkage of their STRONG data with administrative health data and future biomarker measurement.

An identification (ID) number will be assigned to each participant to conceal their identity. No personal identifiers other than the study ID will be entered on any study documentation. Trial monitoring, audits and regulatory inspection will be permitted by providing direct access to source data and documentation. Hard copies of study documents will be stored in a locked filing cabinet until they are scanned and uploaded for electronic storage; at this time the paper copy will be shredded. At the end of the 10-year retention period, a trained IT professional will securely destroy the electronic study records, and blood samples will be destroyed per institutional policies.

Knowledge emanating from this study will be transferred by our research team in collaboration with institutional media relations offices with guidance from our patient public partners. We will take a multi-pronged approach with disseminating our findings to scientific and public audiences, such as through peer-reviewed publications, conference and research day presentations, community presentations, lay documents on our research team websites and infographics for social media. We will also provide our stakeholders (registered dietitians, exercise physiologists and clinicians) with summaries or briefings that are geared towards healthcare professionals.

Discussion

STRONG is the first multidomain clinical trial to investigate the synergistic effects of the three lifestyle interventions on the degree of frailty in older adults. It is also unique in that it will determine whether reducing frailty leads to improved cardiovascular health, whether this differs between the sexes and whether these effects are linked to a reduction in plasma markers of chronic inflammation. STRONG’s methodology includes a long duration, repeated measurements of many outcome variables related to frailty, CVD, lifestyle and well-being and a large and sex-balanced sample size allowing for sex-specific findings to be observed.

Blueberries, protein powder and exercise are relatively simple interventions that can be prescribed by healthcare professionals to older patients. The protocol for STRONG has been developed using best practices to be rigorous but also feasible and not too burdensome to participants. While potential recruitment and retention challenges include the participants growing tired of eating the food portion daily, studies with a similar whey protein supplementation report compliance rates of 94.2%^{37 54} and the longest blueberry trial to date (6 months) used the same blueberry ‘dose’ of 1 cup per day and reported high treatment compliance of 94.1%.⁵³ However, a limitation of all nutrition studies where the intervention is consumed outside of the laboratory is that the exact extent of adherence to the nutrition component cannot be completely verified. Another possible study challenge is the attendance of study exercise sessions three times per week. While a systematic review reported older adults’ adherence to exercise programmes to be mixed, rates were reported to be higher for supervised exercise interventions [83]. In addition to supervision by an exercise physiologist, STRONG also tailors its exercise sessions to each participant within the treatment group. A personalised approach to exercise for older adults has been reported to be beneficial for enhancing physical outcomes [84]. Pragmatic in design, STRONG strives to emulate a real-life setting in which the STRONG treatment would be recommended or prescribed to a community-dwelling older adult, and thus it allows for flexibility and personalisation with treatment details such as the time of day, the exact recipe for consuming the blueberries and protein powder and other such factors purposefully remaining uncontrolled. However, a limitation of this design is that the intervention and control groups will differ in level of contact and support.

CVD and frailty each pose a significant burden, straining healthcare systems and peoples’ well-being. Interventions that modify lifestyle can reduce both frailty and the burden of CVD on individual health outcomes and the broader healthcare system. This project will yield insights into the combined effects of protein powder, blueberries and physical activity on cardiovascular health and frailty, allowing for mechanistic and sex-specific analyses, thereby offering potential evidence-based recommendations for older adults of both sexes. Our proposed research studies a medical management issue that could help prevent frailty, CVD and mortality in millions of people worldwide.