1. Objectives

To evaluate the effects of dietary advice, with and without nutritional supplements, in adults with disease-related malnutrition on survival, weight and anthropometry, and to compare dietary advice with nutritional supplements

2. How studies were identified

The following databases were searched in February 2010:

• Cystic Fibrosis Trials Register
• CENTRAL (The Cochrane Library 2010, Issue 2)
• MEDLINE
• EMBASE
• National Cancer Institute CancerLit
• AMED
• Web of Science
• SCOPUS
• ERIC (to 1998)
• Dissertation Abstracts (to 2000)

Reference lists were also searched and the authors contacted researchers and organizations in the field, including the British Dietetic Association. Results from NICE literature search were also included

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials and quasi-randomized controlled trials

3.2 Study participants

Adults >16 years of age with disease-related malnutrition or at nutritional risk (as judged by study authors or reviewers) due to their clinical condition, treatment, or both

(Studies including pregnant women, people with eating disorders, or set in conditions of food insufficiency were excluded)

3.3 Interventions

Dietary advice compared with no advice, a nutritional supplement, or dietary advice plus a nutritional supplement; and dietary advice, with a nutritional supplement if required, compared with no advice/no supplement

(Dietary advice was provided by a dietician or other health professional and defined as instruction in modification of food intake with the aim of improving nutritional intake)

(Nutritional supplements were defined as whole protein enteral food supplements marketed as clinical products for the management of disease-related malnutrition)

3.4 Primary outcomes

• Mortality
• Morbidity (assessed by risk of hospital admission/readmission and length of hospital stay)
• Measures of nutritional status (change in weight, triceps skinfold thickness and mid-arm muscle circumference, MAMC)

Secondary outcomes included nutritional intake pre- and post-intervention, measures of clinical function (e.g., immune function, cardiac function, respiratory function and other indices of nutritional status), quality of life, and cost

4. Main results

4.1 Included studies

Forty-five randomized controlled trials (50 comparisons), enrolling 3186 adults, were included in this review

• Twelve studies including 1053 participants compared dietary advice with no advice: six in people with cancer, and one each in the elderly, people at risk of osteoporotic fractures, people with Crohn’s disease, people with chronic obstructive pulmonary disease (COPD), people with chronic kidney disease, and people with liver cirrhosis
• Eight studies including 332 participants compared dietary advice with oral nutritional supplements: three in people with cancer, and one each in the elderly, people with HIV infection, people with cystic fibrosis, people with chronic pancreatitis, and people with a femoral neck fracture
• Sixteen studies including 731 participants compared dietary advice with dietary advice plus nutritional supplements: four in people with cancer, three in surgical patients, three in people with HIV infection, three in people with renal failure, and one each in people with COPD, tuberculosis, and benign gastrointestinal disease
• Fourteen studies including 1070 participants compared dietary advice, with supplements if required, to no advice and no supplements: six trials in people with cancer, two in people with COPD, two in frail elderly patients, and one each in people with HIV infection, elderly people, surgical patients, and people undergoing lung transplantation

4.2 Study settings

• Studies were conducted in both hospital and community settings in people with malnutrition or with clinical conditions placing them at high risk of malnutrition

4.3 Study settings

How the data were analysed
Four comparisons were made: i) dietary advice compared with no advice; ii) dietary advice compared with oral nutritional supplements; iii) dietary advice versus dietary advice plus oral nutritional supplements; and iv) dietary advice plus supplements, if required, compared with no advice and no supplements. Fixed effect meta-analysis was used to produce risk ratios (RR) and corresponding 95% confidence intervals (CI) for dichotomous outcomes and mean differences (MD) or standardized mean differences (SMD) and 95% CI for continuous outcomes. Where substantial heterogeneity was detected (I²>33%), random effects meta-analysis was used. To further explore potential sources of heterogeneity, the following subgroup analyses were planned:

• By underlying clinical condition
• By age (under 65 versus over 65)
• By nutritional status at baseline (percentage of malnourished participants versus participants at risk of malnutrition)
• By study setting (hospital versus community versus mixed)

Results
Dietary advice compared with no advice
Primary outcomes
In pooled analysis, dietary advice compared with no advice had no significant effect on mortality at zero to three months follow-up (RR 1.35, 95 CI [0.59 to 3.08]; 6 trials/510 participants), four to six months follow-up (RR 1.73, 95% CI [0.40 to 7.57]; 4 trials/377 participants), or overall (RR 1.43, 95% CI [0.70 to 2.94], 10 trials/887 individuals). The number of people admitted or re-admitted to hospital did not differ by treatment group at four to six months follow-up (RR 0.89, 95% CI [0.52 to 1.50]; 2 trials/177 participants). Overall, weight was improved with dietary advice (MD 1.47 kg, 95% CI [0.32 to 2.61]; 9 trials/733 individuals). Weight change was not significantly different between groups at zero to three months (MD 1.30 kg, 95% CI [-0.82 to 3.42]; 6 trials/451 participants) or four to six months (MD 1.46 kg, 95% CI [-1.03 to 3.95]; 2 trials/190 participants), but was improved with dietary advice at 12 months and over (MD 3.75 kg, 95% CI [0.97 to 6.53]; 1 trial/92 participants). Change in MAMC was improved with dietary advice overall (MD 0.81 cm, 95% CI [0.31 to 1.31]; 2 trials/130 individuals) and at zero to three months follow-up (MD 1.02 cm [0.65 to 1.39]; 1 trial/90 participants), but not at four to six months (MD 0.50 cm, 95% CI [-0.09 to1.09]; 1 trial/40 participants). Change in triceps skinfold thickness was not improved overall (MD 0.15 mm, 95% CI [-1.37 to 1.67]; 3 trials/222 participants), or at any individual time point.

Additional outcomes
Change in energy intake was not improved overall, but was improved with dietary advice at four to six months follow-up (MD 63.70 kcal, 95% CI [55.29 to 72.11]; 1 trial/150 participants). Change in grip strength was not significantly affected by dietary advice in one trial of 40 individuals. In one trial of COPD participants, those receiving dietary advice reported a significantly improved quality of life in the Short Form-36 questionnaire (p=0.003), as well as several other domains of the same questionnaire (pain: p=0.05; vitality: p=0.04; general health: p=0.05), and also significantly improved in several domains in the St George’s Respiratory Questionnaire (activity: p=0.01; impacts: p=0.04; total scores: p=0.01).

Dietary advice compared with oral nutritional supplements
Primary outcomes
Compared with nutritional supplements, dietary advice conferred no significant benefit on the outcome mortality between zero and three months follow-up (RR 0.56, 95% CI [0.24 to 1.31]; 6 trials/411 participants). The number of participants admitted to hospital did not differ significantly between groups in one trial of 50 individuals (RR 0.36, 95% CI [0.04 to 3.24]). Change in weight did not differ between groups during zero to three months follow-up (MD -0.00 kg, 95% CI [-2.42 to 2.42]; 7 trials/399 participants), although heterogeneity was high (I²=95%) and removal of one study from the analysis revealed a greater weight gain in the supplemented group (MD -0.91 kg, 95% CI [-1.60 to -0.23]). MAMC (MD -0.80 cm, 95% CI [-5.29 to 3.69]; 1 trial/46 participants) and triceps skinfold thickness (MD -0.54 mm, 95% CI [-1.41 to 0.33]; 2 trials/95individuals) did not differ between groups.

Additional outcomes
Change in energy intake and change in grip strength from zero to three months follow-up were not different according to treatment group. Quality of life did not differ between dietary advice and supplement groups in four studies reporting on this outcome (data not pooled).

Dietary advice compared with dietary advice plus oral nutritional supplements
Primary outcomes
In pooled analysis of six trials including 517 participants, the risk of mortality at zero to three months follow-up did not differ significantly according to the receipt of nutritional supplements (RR 0.55, 95% CI [0.08 to 3.95]). Overall, the number of people admitted or readmitted to hospital was marginally significantly higher among the group who received dietary advice alone (RR 1.53, 95% [1.00 to 2.34], p=0.049; 2 trials/108 individuals). Change in weight did not differ between groups overall (MD 0.95 kg, 95% CI [-0.03 to 1.93]; 11 trials/636 individuals) or when subgrouped by follow-up time. Overall, change in MAMC was lower in the group receiving dietary advice alone (MD -0.89 cm, 95% CI [-1.35 to -0.43]; 3 trials/492 individuals), and remained significantly lower in subgroup analysis of zero to three months follow-up (MD -0.85 cm, 95% CI [-1.34 to -0.36]; 2 trials/464 individuals), but not in subgroup analysis of four to six months follow-up (MD -1.23 cm, 95% CI [-2.65 to 0.19]; 1 trial/28 participants). Change in triceps skinfold thickness followed the same pattern as MAMC (overall: MD -1.22 mm, 95% CI [-2.34 to -0.09]; 6 trials/386 individuals; zero to 3 months: MD -1.32 mm, 95% CI [-2.51 to -0.12]; 5 trials/358 participants; 4 to 6 months: MD 0.10 mm, 95% CI [-3.99 to 4.19], 1 trial/28 individuals).

Additional outcomes
Change in energy intake did not differ by treatment groups overall or in analyses subgrouped by follow-up time. In pooled analysis of four trials including 308 individuals, change in grip strength was significantly lower in the group receiving dietary advice alone (MD -1.67 kg force, 95% CI [-2.96 to -0.37]). No significant differences between groups were found in five trials reporting on quality of life (data not pooled).

Dietary advice plus supplements if required compared with no advice and no supplements
Primary outcomes
Compared with no nutritional intervention, dietary advice plus supplements if required had no effect on mortality at zero to three (RR 0.95, 95% CI [0.47 to 1.93]; 4 studies/291 participants), four to six (RR 1.04, 95% CI [0.54 to 2.00]; 3 studies/350 individuals), seven to 12 (RR 1.26, 95% CI [0.76 to 2.10]; 3 trials/238 individuals), or 12 months or longer follow-up (RR 0.94, 95% CI [0.84 to 1.05]; 2 trials/317 participants). Change in weight was significantly higher in the dietary advice group at zero to three (MD 1.74 kg, 95% CI [0.53 to 2.95]; 5 trials/232 individuals), but not at other time points (four to six months: MD 1.87 kg, 95% CI [-0.07 to 3.81]; 4 trials/281 individuals; seven to 12 months: MD 0.70 kg, [-0.84 to 2.24]; 1 trial/62 children; 12 months or longer: MD 2.17 kg, 95% CI [-1.20 to 5.54]; 2 trials/77 individuals). Triceps skinfold thickness was significantly improved in the intervention group in one trial including 27 individuals (MD 0.40 mm, 95% CI [0.10 to 0.70]).

Additional outcomes
In one trial including 63 participants, energy intake at seven to 12 months follow-up was increased in the dietary advice group compared with the group receiving no intervention (MD 464.0 kcal, 95% CI [270.07 to 657. 93]), while no significant difference was found overall or at zero to three or four to six months follow-up. Change in grip strength did not differ between treatment and control groups at four to six months follow-up. The outcome quality of life was included in seven trials, none of which reported any significant differences between treatment groups.

5. Additional author observations*

The methodological quality of the trials included in this review was low to moderate, with many included trials being inadequately powered and not blinded. Additionally, 21 of the 45 included trials were at unclear or high risk of allocation concealment bias. Clinical and statistical heterogeneity was found across all analyses in this review, except for those of the outcome mortality.

Limited evidence suggests that in patients with malnutrition or at risk of malnutrition, dietary advice with or without nutritional supplements may improve weight and body composition. Dietary advice, given alone or with nutritional supplements, had no effect on mortality and did not prevent hospital admission or readmission.

Further studies evaluating the effects of dietary advice with or without supplements on nutritional, functional and patient-centred outcomes in both hospital and community settings are needed. Such studies should detail the type, intensity, and duration of the intervention, the change in nutrient intake, and the expertise of the person giving the advice. Inclusion criteria should aim to select a population homogenous in the state of malnutrition at baseline, and the participants’ clinical condition, including stage and treatment, should be clearly described.