A preliminary feasibility study in participants of the ongoing Exercise and Nutrition for Healthy AgeiNg (ENHANce) study was performed with participants included in a period between February 2018 and August 2019. The food diaries of (pre)sarcopenic participants that were screened for the ENHANce study were included in the analysis to describe regular protein intake (n = 51). For analysis of the effect of the protein supplement on regular dietary intake, participants were divided in two groups: the protein supplement group (arm 2, 3, and 4 of ENHANce) or the placebo supplement group (arm 1 and 5 of ENHANce). In total, n = 35 ENHANce participants initiated and completed the ENHANce intervention period as well as a second food diary after 12 weeks of the ENHANce study. These participants were included in the analyses to describe the effect of personalized supplementation on regular protein intake.

The ENHANce study is an ongoing, single-center, placebo-controlled triple blinded randomized controlled trial, that was initiated in February 2018. Details of the methods are described elsewhere (20). In brief, the ENHANce study aims to examine the effect of combined anabolic interventions (exercise and food compounds) on physical functioning in (pre)sarcopenic older adults. The ENHANce study consists of a screening visit (1 month before the start of intervention) and a 12-week intervention period. During the screening visit, fulfillment with the inclusion criteria was checked, participants were asked to complete a food diary during the first week after screening and participants were block randomized (based on gender) into one of the five intervention arms: (1) Exercise intervention + protein placebo + PUFA's placebo; (2) Protein + PUFA's placebo; (3) Exercise intervention + protein + PUFA's placebo; (4) Exercise intervention + protein + PUFA's; (5) Control group: protein placebo + PUFA's placebo. The 12-week intervention period started and ended with an extensive visit including measurements of physical and cognitive abilities (Supplementary Figure 1). The study was approved by KU Leuven/UZ Leuven Ethics Committee (s60763) and registered on ClinicalTrials.gov (NCT03649698).

Participants of present study are community-dwelling older adults (aged ≥ 65 years) living in Belgium. Participants were recruited from Leuven and surrounding area in a number of different ways, including senior citizen organizations, as described in the study protocol (20). The ENHANce participants screened between February 2018 and July 2019 had “presarcopenia,” “sarcopenia,” or “severe sarcopenia” (in this article also referred to as “sarcopenia”) according to the operational definition of the European Working Group on Sarcopenia in Older People (EWGSOP1) (1). From November 2018 onwards, participants had “probable sarcopenia,” “confirmed sarcopenia,” or “severe sarcopenia” (in this article referred to as “sarcopenia”) according to EWGSOP2 (2). The period between November 2018 and July 2019 was a transition period where both definitions were accepted. Other inclusion criteria were: to be able to communicate in Dutch, English or French, no allergy to milk, soy, peanut or peanut oil, Mini-Mental State Examination (MMSE) > 21 (21), no terminal illness with a prognosis <6 months, protein intake lower than 1.5 g·kg⁻¹·d⁻¹ of body mass 1 week after screening, not following a training program equal to or more than twice per week during the last 6 months, no uncontrolled disease or acute cardiovascular problems (no medical contraindication to perform physical activities), glomerular filtration rate > 30 ml ·min⁻¹·(1.73 m²)⁻¹, fasting glycaemia < 126 mg ·dl⁻¹ and no treatment for diabetes and no impairments or diseases that impose problems to study participation according to the investigators. Written informed consent was obtained by the participant before the start of the screening visit.

Participant characteristics including age (years), body mass index (BMI, kg/m²), and risk of malnutrition were assessed. Body weight (SECA, model no 8801321009, SECA UK Ltd.) to the nearest 0.1 kg and height (Harpenden stadiometer, Holtain Ltd., Crosswell, UK) to the nearest 0.1 cm, were measured standing upright, barefoot and in light clothing. The BMI was calculated as weight/height². The Mini Nutritional Assessment—Short Form (MNA-SF) was completed, classifying participants in a range between 0 and 14 points. A score of 0–7 indicates malnourishment and a score of 8–11 identifies persons at risk of malnutrition (22).

Appendicular lean mass was assessed by dual-energy X-ray absorptiometry (DXA, Horizon A scanner, Hologic Inc., Bedford, MA, USA). Low muscle mass was defined by the skeletal muscle index (SMI), which is the appendicular lean mass (kg) divided by height² (m²). Participants with presarcopenia and sarcopenia (EWGSOP1) and confirmed or severe sarcopenia (EWGSOP2) had low muscle mass. The cut-off of low SMI in EWGSOP1 was SMI < 7.26 kg/m² for men and SMI < 5.50 kg/m² for women (1) and in EWGSOP2, the cut-off was SMI < 7.00 kg/m² for men and SMI < 5.50 kg/m² for women (2).

Handgrip strength was measured by hand dynamometry (Jamar 1, TEC Inc., Clifton, NJ, USA) while the participant was seated with the elbow at a 90° angle according the American Society of Hand Therapists (23) for EWGSOP1 measurements. The mean value of three consecutive measurements of the dominant hand [according to Fried et al. (24)] was recorded for scoring purposes. According to the EWGSOP1 criteria, the cut-off values were dependent of the BMI (1). The EWGSOP2 proposed the Southampton protocol to assess handgrip strength six times, alternating sides and the maximal value was reported (25). According to the EWGSOP2, low grip strength was defined as < 16 kg for women and < 27 kg for men according to the EWGSOP2 (2).

Gait speed was assessed over a distance of six meters. Participants were instructed to walk 6 m on usual pace (walking aids allowed). The time was measured over 4 m: from the moment the participant's foot passed the mark of the 1 m line on the ground until the foot passed the mark of the 5 m line, according to the BC guidelines (26). The mean of three tests was calculated for scoring purposes. A cut-off of ≤ 0.8 m·s⁻¹ was used to define low gait speed according to the definition of the EWGSOP (1) and EWGSOP2 (2).

The chair stand test was performed by measuring the time required to rise five consecutive times as quickly as possible from a chair with arms folded in front of the chest (27). The cut-off point for low physical performance was >15 s, as defined according to the EWGSOP2 (2).

Participants completed a 4-day estimated dietary records (EDR) on non-consecutive days including 3 week- and 1 weekend day. Based on the formula of Beaton (28) including the within-subject standard deviation of older (65–85 years) adults living in the north of England (29), a 4-day EDR is appropriate to assess protein intake of an individual. Non-consecutive days and alternating periods were chosen to obtain a better estimate of the day-to-day variability (30). At screening, recorded days were Monday, Wednesday, Friday, Sunday, and at week 12, recorded days were Tuesday, Thursday, Saturday, and Monday. Participants were instructed to meticulously record food and beverage intake for 24 h. Portion sizes were documented in pre-defined household units and determined based on “Maten en Gewichten,” a guide to convert qualitative data relating to the amount (estimates) of food ingested into quantitative data in a uniform manner (31). Time of day, description of food and brand of food were documented in the EDR. Quality of records was assured by reviewing records together with the participant in order to complete the records as accurately as possible.

Dietary intake was computed based on the Belgian and USA Food Composition Table [Nubel, BE (31) and USDA FoodData Central, USA] (32). The estimate of mean protein (g·kg⁻¹·d⁻¹ of body mass, g·day⁻¹ or % of total energy·day⁻¹), fat (% of total energy·day⁻¹), carbohydrates (% of total energy·day⁻¹), energy (kcal·kg⁻¹·d⁻¹ of body mass or kcal·day⁻¹), leucine (g·day⁻¹or mg·kg⁻¹·d⁻¹ of body mass) consumption was calculated as a mean of 4 days. The actual body weight of the participant was used. Snacks were defined as the eating moments between meals. Snack 1 is defined as snacks between breakfast and lunch, snack 2 between lunch and dinner and snack 3 after dinner, before bedtime. Factors of protein quality included the protein source and amino-acids composition, such as leucine. The protein source (plant/animal source) of each dietary protein product was based on the Dutch Food Composition Database (33). Amino-acid composition was based on the USDA food composition database (USDA FoodData Central) (32). The amino-acid composition of a food item was analyzed when the relative protein contribution of this product was 1% or more of the total protein intake per day. Regular dietary intake reflects the protein intake originating from foods while the total intake reflects the dietary intake and the intake from supplements.

The aim of the personalized supplementation was to reach the before mentioned protein recommendation for older adults with acute or chronic diseases of 1.5 g·kg⁻¹·d⁻¹ of body mass (16). Meals were not substituted with a supplement, but a supplement was added to a meal if necessary. The exact amount (to the nearest 0.1 g·kg⁻¹·d⁻¹ of body mass) of protein needed to reach this recommendation was calculated. Each meal, if needed, was supplemented according to the following rule: a combined (diet + supplement) protein intake of at least 25 g protein/meal (breakfast/lunch/dinner) must be reached. Next, in case the protein intake of 1.5 g·kg⁻¹·d⁻¹ of body mass intake per day was not yet reached, the combined (diet + supplement) protein intake per meal was equally increased over the meals (breakfast/lunch/dinner) up to a maximum of 35 g protein·meal⁻¹, but a maximum of 700 kcal·meal⁻¹ is never exceeded [35% of average requirement for energy intake for older adults (34)]. In case the protein intake of 1.5 g·kg⁻¹·d⁻¹ of body mass was not yet reached after applying the above rules, the snack before bedtime (snack 3) was supplemented. In summary, if supplementation was needed, at least one evening snack and maximum three main meals and the evening snack were enriched. For supplementation calculations, the body weight of participants with a BMI < 22 or >27 kg/m² was adjusted: the adjusted body weight of participant with BMI < 22 was the body weight that would result in a BMI of 22 kg/m², the adjusted body weight of participant with BMI > 27 was the body weight that would result in a BMI of 27 kg/m² (35). The supplemented protein (g) was visualized for each individual receiving the protein supplement in Supplementary Figure 2.

Participants were instructed to initiate the intake of the powder supplement (with protein or placebo powder) 5 days before the start of the 12-week intervention period. Participants received instructions on how to dissolve the protein powder and suggestions of drinks and foods to mix with the powder before consumption. For that purpose, each participant received an individualized drinking glass for each meal or snack that had to be supplemented. By means of a line on this drinking glass, the amount of powder to be taken with this meal/snack was indicated. For each meal/snack at home, the participant completed the glass for that meal/snack up to the marked line with the provided powder. Participants in the protein supplement group received Resource Instant Protein (Nestlé, 4.5 g protein per 5 g powder, 9.14% leucine). The placebo product was Resource dextrin maltose (Nestlé, 0 g protein). Nutritional content of both powders is described in Supplementary Table 1. The powder boxes and drinking glasses were prepared by an unblinded researcher, who did not participate in study visits or had contact with the participants. The boxes for protein or placebo powder products were equal looking blank boxes.

Compliance with powder supplement intake was assessed by weighing the amount of powder that was taken from the provided powder box at each study visit (seven times in 12-week period). The compliance was defined as the proportion of amount of powder that was actually removed from the boxes to the prescribed amount of powder that was suggested to be taken.

Fifty-one participants in the ENHANce study completed the screening visit and EDR at screening. Thirty-five participants completing EDR at wk 12 since 16 participants dropped out before completion of the 12-week study period. Fifty percent (n = 8) of the dropouts perceived the study as “too intense,” 19% (n = 3) had a protein intake above 1.5 g g·kg⁻¹·d⁻¹ of body mass, two participants were no longer meeting inclusion criteria, two participants were not willing to comply with study procedures, and one participant was no longer interested in study participation. Twenty of the thirty-five participants received a protein supplement, 15 received a placebo supplement (Figure 1). Study participants at screening visit were 73.6 years old (± 6.5) and 52.9% were female. Mean total fat and total carbohydrate intake were, respectively 36.3 ± 7.6% of total energy·day⁻¹ and 42.5 ± 9.2% of total energy·day⁻¹ (Table 1). Forty-four participants were included according to EWGSOP1 of which 39 participants were presarcopenic and five participants were sarcopenic; seven participants were according to EWGSOP2, one probable sarcopenic and six confirmed sarcopenic.

The mean regular dietary energy intake at screening was 2,000 kcal·day⁻¹, or 29.3 kcal·kg⁻¹·d⁻¹ of body mass (n = 51). In the protein supplement group, supplementation significantly increased the total energy intake (2,110 kcal·day⁻¹, 30.4 kcal·kg⁻¹·d⁻¹ of body mass) compared to the regular dietary intake at wk 12 (1,990 kcal·day⁻¹, 28.9 kcal·kg⁻¹·d⁻¹ of body mass). In the placebo supplement group, regular dietary energy intake (kcal·day⁻¹) did not significantly change, except that the regular dietary intake decreased from screening (29.1 kcal·kg⁻¹·d⁻¹ of body mass) to wk 12 (26.3 kcal·kg⁻¹·d⁻¹ of body mass; p = 0.047; Table 2).

The mean regular dietary protein intake at screening was 72.8 ± 17.6 g·day⁻¹, or 1.06 ± 0.3 g·kg⁻¹·d⁻¹ of body mass (n = 51). Participants receiving a protein supplement (n = 20) showed an increased total protein intake (105 g·day⁻¹, 1.55 ± 0.3 g·kg⁻¹·d⁻¹ of body mass) compared to regular dietary intake at screening (69.1 g·day⁻¹, 1.04 ± 0.2 g·kg⁻¹·d⁻¹ of body mass) or wk 12 (74.7 g·day⁻¹, 1.13 ± 0.4 g·kg⁻¹·d⁻¹ of body mass), whilst participants receiving a placebo supplement (n = 15) did not significantly change regular dietary protein intake (g·d⁻¹ or g·kg⁻¹·d⁻¹ of body mass). Compared to regular dietary intake at screening (15.8 energy%) or wk 12 (15.8 energy%), protein supplementation resulted in an increased total relative protein intake (20.9 energy%; Table 3).

At screening, the regular dietary distribution of protein intake at breakfast, lunch, dinner, and snack times was unequal, and breakfast did not meet the 25 g protein per meal threshold (n = 51, Supplementary Table 2). Protein supplementation resulted in higher total protein intake (28.2 g at breakfast p < 0.001; 36.7 g at dinner, p = 0.004) compared to regular dietary intake at screening and wk 12 at breakfast (11.7 g, p < 0.001; 10.2 g, p < 0.001) and dinner (24.4 g, p = 0.002; 26.2 g, p = 0.001). Dietary protein intake after dinner/before bedtime (snack 3) decreased (p = 0.023) at wk 12 compared to screening (Figure 2; Supplementary Table 2).

The contribution of animal- and plant-based sources to daily protein intake was not significantly different between regular dietary intake at screening (animal 63.5%; plant 36.5%) and wk 12 (animal 61.9%, plant 38.1%; p = 0.358). The daily intake of the individual amino acids (g) did not significantly change between regular dietary intake at screening and wk 12, except a slight increase in cysteine intake (Supplementary Table 3). Daily dietary leucine intake at screening (6.1 ± 1.0 g·day⁻¹, 114 ± 57 mg·kg⁻¹·d⁻¹ of body mass) or wk 12 (6.8 ± 1.6 g·day⁻¹, 102 ± 30 mg·kg⁻¹·d⁻¹ of body mass) increased by supplementation (9.5 ± 2.2 g·day⁻¹, 140 ± 24 mg·kg⁻¹·d⁻¹ of body mass), without affecting regular dietary leucine intake (Table 4).

At screening, the regular dietary distribution of leucine intake at breakfast, lunch, dinner and snack times was unequal, and no meal reached the 2.5 g leucine threshold (Table 4). Protein supplementation resulted in higher total leucine intake (2.46 g at breakfast, 3.33 g at dinner) compared to regular dietary intake at screening at breakfast (0.87 g, p < 0.001) and dinner (1.99 g, p = 0.009) or compared to regular dietary intake at wk 12 at breakfast (0.90 g, p < 0.001) and dinner (2.35 g, p < 0.001; Table 4).

The compliance with protein powder intake ranged from 44 to 100%, and the compliance with placebo powder intake ranged from 49 to 100%. Seventy-five percent (15/20) of the participants with protein powder supplementation and 79% (11/14) of the participants with placebo powder supplementation had a compliance higher than 79%. Compliance of one participant of the placebo powder supplementation group was not calculated as this person did not bring the boxes to the study visits. Figure 3 visualizes the compliance on an individual level.

The results of this study showed that personalized protein supplementation is needed to achieve recommended dietary protein intake, and showed to be achievable to improve protein quantity, quality and distribution in community-dwelling (pre)sarcopenic older adults (16). Since diet is a modifiable risk factor for sarcopenia (65), influencing nutritional status may influence muscle functioning and therefore for example sarcopenia, a syndrome that increases the health care expenditures with $900/person/year (66). In that respect, nutritional adjustments can be considered as relatively inexpensive interventions, for example by dieticians. Consequently, the effect of personalized protein supplementation on outcomes such as sarcopenia or physical functioning should be one of the focal points of further research.