1. Objectives

To evaluate the effects and acceptability of oral protein and energy supplements in elderly people at risk of malnutrition in both community and hospital settings

2. How studies were identified

The following databases were searched in October to December 2007:

• Cochrane Central Register of Controlled Trials (Issue 4, 2007)
• MEDLINE
• EMBASE
• CINAHL
• BIOSIS
• CAB Abstracts
• Current Controlled Trials
• Healthstar (to March 2001)

Relevant books, reference lists, and journals were handsearched, and the authors directly contacted researchers and manufacturers in the field

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials and quasi-randomized randomized controlled trials

3.2 Study participants

Adults with a minimum average age of 65 years

(Studies that exclusively enrolled elderly people in critical care or recovering from cancer treatment who may have had specific nutritional needs relating to their condition were excluded)

3.3 Interventions

Oral protein and energy (carbohydrate and/or fat) supplements, with or without added micronutrients, provided for a minimum of two weeks’ duration, compared to routine care

(Oral protein supplements could include commercial sip feeds, milk-based supplements, or the fortification of normal food)

3.4 Primary outcomes

• All-cause mortality
• Morbidity, number of people with complications (e.g., pressure sores, deep vein thrombosis, respiratory infections, urinary tract infections)
• Functional status (e.g., cognitive function, muscle function, mobility, ability to perform activities of daily living)

Secondary outcomes included participants’ perceived quality of life, length of hospital stay, number of primary care contacts, level of care and support required, number of hospital/care home readmissions, nutritional status (percentage weight change, percentage change in arm muscle circumference), percentage change in dietary intake (protein, energy), compliance with the intervention, economic outcomes, and adverse effects of supplementation

4. Main results

4.1 Included studies

Sixty-two randomized controlled trials, enrolling 10,187 participants, were included in this review

• Forty trials enrolled participants without a particular health condition, and other studies included those with hip fracture, stroke, congestive heart failure, chronic obstructive pulmonary disease (COPD), or recent surgery
• Sixty-percent of included participants underwent nutritional screening and were classified as undernourished or at nutritional risk
• Interventions aimed to provide an additional 10 to 50 g protein/day; <400 kcal/day was provided in 20 trials, ≥400 kcal/day was provided 32 trials, and the energy provided was not reported in ten trials
• Intervention duration was <35 days for 17 trials, ≥35 days in 37 trials, from admission to discharge in five trials, and unclear for two trials

4.2 Study settings

• Australia (6 trials), Belgium, Canada (6 trials), China (2 trials), Denmark (2 trials), Germany (2 trials), France (4 trials), Italy, the Netherlands (7 trials), Spain, Sweden (5 trials), Switzerland (2 trials), the United Kingdom of Great Britain and Northern Ireland (17 trials), the United States of America (5 trials), and multinational
• Twenty-six studies (71% of participants) were conducted in hospitalized in-patients with acute medical conditions; 15 trials (14% of participants) were conducted in long-stay facilities such as nursing homes; and 21 studies (15% of participants) were conducted in people living in their home in the community

4.3 Study settings

How the data were analysed
Oral protein and energy supplements were compared to routine care. Fixed effect meta-analysis was used to generate risk ratios (RR) and corresponding 95% confidence intervals (CI) for dichotomous outcomes, and mean differences (MD) and 95% CI for continuous outcomes. Where substantial heterogeneity was detected (I²>50%), random effects meta-analysis was used. Sensitivity analyses excluding studies of poor methodological quality or with industry funding were conducted. To investigate potential sources of heterogeneity, the following subgroup analyses were conducted:

• Baseline nutritional status (nourished, undernourished)
• Health status (well, unwell)
• Mean age (<75 years, ≥75 years)
• Amount of kilocalories provided in the supplement (<400 kcal/day, ≥400 kcal/day)
• Duration of the intervention (<35 days, ≥35 days)
• Diagnostic group (post hoc)

Results
Primary outcomes
Mortality
In pooled analysis of 42 trials including 8031 participants, oral protein and energy supplementation had no significant effect on the risk of mortality in comparison with routine care (RR 0.92, 95% CI [0.81 to 1.04], I²=0%). When restricted to trials including undernourished participants only, the effect became significant (RR 0.79, 95% CI [0.64 to 0.90], 25 trials/2461 participants). The effect approached significance when considering only those trials in which ≥400 kcal/day was offered (RR 0.89, 95% CI [0.78 to 1.00], p=0.059; 24 trials/7307 participants). Post hoc subgroup analysis including all participants with a diagnosed geriatric condition produced a statistically significant reduction in risk of mortality (RR 0.78, 95% CI [0.62 to 0.98], 23 trials/2701 participants). No other subgroup analyses or sensitivity analyses produced meaningfully different results.

Morbidity
Among 6225 participants in 24 trials, the risk of complications (infective, pressure sores, too ill to continue trial, exacerbation of COPD, hospital readmission, antibiotic prescribed, incomplete wound healing, anaesthetic, gastrointestinal, urinary) was significantly reduced among those receiving oral protein and energy supplements (RR 0.86, 95% CI [0.75 to 0.99]). In subgroup analysis by diagnostic group, the effect was rendered non-significant except among those with hip fractures (RR 0.60, [0.40 to 0.91], 6 trials/298 individuals).

Functional status
Handgrip strength was not significantly improved with supplementation (MD 0.06 [units not reported], [-0.60 to 0.72], 7 trials/535 participants). No other data on functional status were able to be combined in meta-analysis. A significant benefit of protein and energy supplementation was reported in individual studies on falls (0% versus 21%; p=0.05), activity rating (p<0.05), quadriceps muscle power at three months (p=0.03), functional ability, activities of daily living (in 1/11 trials reporting this outcome: p<0.01), calf circumference, and lung function (2/4 trials). No beneficial effect was seen for six metres walk, five time chair rise, six stair climb, timed up-and-go test, physio-independent mobility, mobility, walking distance, walking velocity, dynamic strength, or cognitive function.

Adverse effects
Adverse effects were reported in 12 of the 18 trials including this outcome, although formal comparisons were not made between intervention and control groups. Common side effects of supplementation included nausea and diarrhoea, and other gastrointestinal adverse effects reported included vomiting and discomfort.

Additional outcomes
Quality of life was improved with supplementation in three out of 16 trials reporting on this outcome: the physical and social scores in the SF36 questionnaire were improved in one study, participants reported feeling better in one study (p=0.029), and fewer mobility problems were reported in another study (p=0.022). Length of hospital stay was not significantly reduced with protein and energy supplementation in pooled analysis of 15 trials including 5735 individuals (MD -0.75 days, 95% CI [-2.84 to 1.34]). In 45 trials including 3058 participants, percentage weight change was improved with protein and energy supplementation (MD 2.15%, 95% CI [1.80 to 2.49]). This effect remained statistically significant in the diagnostic subgroups geriatric conditions, chest conditions, and congestive heart failure. Arm muscle circumference was increased overall (MD 1.20%, 95% CI [0.45 to 1.96], 16 trials/1382 participants) as well as in the diagnostic subgroup hip fracture (MD 2.92%, [0.16 to 5.68], 1 trial/10 participants). In 32 out of 41 trials reporting on nutritional intake, an improvement was found in protein and/or energy intake with supplementation. Compliance varied between trials, with several trials excluding non-compliers. In one eight-week trial there was no difference between treatment groups at 24 weeks follow-up in heath care or social service costs. No other outcomes were reported in the included trials.

5. Additional author observations*

Many trials were of poor methodological quality, with 43 of the 62 included trials at unclear or high risk of allocation concealment bias. The greatest majority of trials were conducted in hospital in-patient settings (26/62 trials), contributing 71% of all participants in the review.

While mortality was not reduced with protein and energy supplementation overall, a reduced risk was found among those participants who were undernourished at baseline and among those with a geriatric condition, who were generally hospitalized. In addition, weight gain and nutritional intake appeared to be improved with protein and energy supplementation. While protein and energy supplements are widely used in hospitals, few trials attempted to alter nutrient density or diversity of the normal diet, which may be preferable to supplements. In addition, some trials reported problems such as lack of adherence and gastrointestinal disturbance with supplements.

Further trials of oral protein and energy supplements among older people at risk of malnutrition living in the community are needed, as are trials in clearly defined patient groups. Future trials should be of high methodological quality, be sufficiently powered and of adequate length, and include outcomes relevant to participants, such as functional outcomes and quality of life.