Advice on diet for adults with malnutrition that is the result of disease

Review question

Can dietary advice with or without oral nutritional supplements (ONS) improve disease-related malnutrition in adults?


Ill people often have a poor appetite or feel sick because of medicines or other treatments and eat less than usual. Eating less over a longer time can cause weight loss, malnutrition, more health problems and death. Healthcare professionals may offer advice about dietary changes to help people to re-establish good eating habits. They might recommend high-protein and high-energy foods so that these people can gain weight and improve their nutrition and general health. It is common for sick people to be offered ONS with or without advice about changing their food intake.

To find the best answer to our review question, we looked for studies that compared five different treatment options: dietary advice compared with no advice; dietary advice compared with ONS; dietary advice plus ONS compared with dietary advice; dietary advice plus ONS if appropriate compared with no dietary advice; and dietary advice plus ONS compared with no dietary advice and no ONS. To make these comparisons fair, we looked for randomised controlled trials (RCTs), where the people taking part had an equal chance (like the flip of a coin) of being in either group that was being compared. 

Search date

The evidence is current to: 01 March 2021.

Study characteristics

We found 94 studies (with a total of 10,284 people) that we could include in our review. Although older people have a higher risk of malnutrition, the people in these studies ranged from 17 to over 80 years of age and they were living either at home, in the community, or in hospital. They had a wide range of health conditions, including cancer, dementia and kidney disease. The studies reported on the participants for the length of their hospital stay or in some people in the community for up to six and a half years.

Key results

There is no evidence that any of the treatments affected how long many of the people in the studies lived. They did report some positive changes in energy intake (measured in calories), protein intake, weight, muscle bulk and quality of life. There were some reductions in complications and the length of time spent in hospital. However, there is no clear evidence about which treatment is the most helpful or the time it takes to achieve any benefit. Few studies reported results separately for men and women and so we cannot comment on whether there were any overall differences by sex. No studies recorded information about adverse events (harms) so we cannot offer a summary about possible harms.

More research is needed to work out the best ways to help people who are losing weight because of illness in order to improve their clinical outcomes and quality of life.

Certainty of the evidence

Overall we rated the certainty of the evidence as low for most results, which means that we cannot be confident about the findings we report. There were several reasons for this. Some of the treatment comparisons that we looked at had only a few studies  and some of those had small numbers of participants. There were problems with the design of some studies that may have affected the results. Some people knew which treatment they were receiving. We think this may influence the way that they reported some changes, e.g. their energy and protein intake, body weight and quality of life. We think that the way the decision about which group a person went into at the start of the study may have affected the  results for some outcomes, e.g. change in weight, change in muscle bulk and mortality.

We needed to see particular results to help us understand whether adults living with disease-related malnutrition can improve their survival, weight and general quality of life if they receive advice about diet with or without ONS. None of the studies reported all of the results that we needed to do this. We were not able to estimate whether participants gain any benefits from the treatments, such as shortening the length of hospital stay, lowering the risk of readmission to hospital or developing complications. The low certainty of evidence, with no evidence in many areas, means we cannot  make statements about any benefits and the possible disadvantages of these treatments despite the fact they are being used extensively in clinical practice. We recommend that future studies should be designed to measure these important patient-centred and healthcare outcomes as well as any potential harms.

Authors' conclusions: 

We found no evidence of an effect of any intervention on mortality. There may be weight gain with DA and with DA plus ONS in the short term, but the benefits of DA when compared with ONS are uncertain. The size and direction of effect and the length of intervention and follow-up required for benefits to emerge were inconsistent for all other outcomes.  There were too few data for many outcomes to allow meaningful conclusions. Studies focusing on both patient-centred and healthcare outcomes are needed to address the questions in this review.

Read the full abstract...

Disease-related malnutrition has been reported in 10% to 55% of people in hospital and the community and is associated with significant health and social-care costs. Dietary advice (DA) encouraging consumption of energy- and nutrient-rich foods rather than oral nutritional supplements (ONS) may be an initial treatment.


To examine evidence that DA with/without ONS in adults with disease-related malnutrition improves survival, weight, anthropometry and quality of life (QoL).

Search strategy: 

We identified relevant publications from comprehensive electronic database searches and handsearching.

Last search: 01 March 2021.

Selection criteria: 

Randomised controlled trials (RCTs) of DA with/without ONS in adults with disease-related malnutrition in any healthcare setting compared with no advice, ONS or DA alone.

Data collection and analysis: 

Two authors independently assessed study eligibility, risk of bias, extracted data and graded evidence.

Main results: 

We included 94, mostly parallel, RCTs (102 comparisons; 10,284 adults) across many conditions possibly explaining the high heterogeneity.  Participants were mostly older people in hospital, residential care and the community, with limited reporting on their sex. Studies lasted from one month to 6.5 years. 

DA versus no advice - 24 RCTs (3523 participants)

Most outcomes had low-certainty evidence. There may be little or no effect on mortality after three months, RR 0.87 (95% confidence interval (CI) 0.26 to 2.96), or at later time points. We had no three-month data, but advice may make little or no difference to hospitalisations, or days in hospital after four to six months and up to 12 months. A similar effect was seen for complications at up to three months, MD 0.00 (95% CI -0.32 to 0.32) and between four and six months. Advice may improve weight after three months, MD 0.97 kg (95% CI 0.06 to 1.87) continuing at four to six months and up to 12 months; and may result in a greater gain in fat-free mass (FFM) after 12 months, but not earlier. It may also improve global QoL at up to three months, MD 3.30 (95% CI 1.47 to 5.13), but not later.

DA versus ONS - 12 RCTs (852 participants)

All outcomes had low-certainty evidence. There may be little or no effect on mortality after three months, RR 0.66 (95% CI 0.34 to 1.26), or at later time points. Either intervention may make little or no difference to hospitalisations at three months, RR 0.36 (95% CI 0.04 to 3.24), but ONS may reduce hospitalisations up to six months. There was little or no difference between groups in weight change at three months, MD -0.14 kg (95% CI -2.01 to 1.74), or between four to six months. Advice (one study) may lead to better global QoL scores but only after 12 months. No study reported days in hospital, complications or FFM.

DA versus DA plus ONS - 22 RCTs (1286 participants)

Most outcomes had low-certainty evidence. There may be little or no effect on mortality after three months, RR 0.92 (95% CI 0.47 to 1.80) or at later time points. At three months advice may lead to fewer hospitalisations, RR 1.70 (95% CI 1.04 to 2.77), but not at up to six months. There may be little or no effect on length of hospital stay at up to three months, MD -1.07 (95% CI -4.10 to 1.97). At three months DA plus ONS may lead to fewer complications, RR 0.75 (95% CI o.56 to 0.99); greater weight gain, MD 1.15 kg (95% CI 0.42 to 1.87); and better global QoL scores, MD 0.33 (95% CI 0.09 to 0.57), but this was not seen at other time points. There was no effect on FFM at three months.

DA plus ONS if required versus no advice or ONS - 31 RCTs (3308 participants)

Evidence was moderate- to low-certainty. There may be little or no effect on mortality at three months, RR 0.82 (95% CI 0.58 to 1.16) or at later time points. Similarly, little or no effect on hospitalisations at three months, RR 0.83 (95% CI 0.59 to 1.15), at four to six months and up to 12 months; on days in hospital at three months, MD -0.12 (95% CI -2.48 to 2.25) or for complications at any time point. At three months, advice plus ONS probably improve weight, MD 1.25 kg (95% CI 0.73 to 1.76) and may improve FFM, 0.82 (95% CI 0.35 to 1.29), but these effects were not seen later. There may be little or no effect of either intervention on global QoL scores at three months, but advice plus ONS may improve scores at up to 12 months.

DA plus ONS versus no advice or ONS - 13 RCTs (1315 participants)

Evidence was low- to very low-certainty. There may be little or no effect on mortality after three months, RR 0.91 (95% CI 0.55 to 1.52) or at later time points. No study reported hospitalisations and there may be little or no effect on days in hospital after three months, MD -1.81 (95% CI -3.65 to 0.04) or six months. Advice plus ONS may lead to fewer complications up to three months, MD 0.42 (95% CI 0.20 to 0.89) (one study). Interventions may make little or no difference to weight at three months, MD 1.08 kg (95% CI -0.17 to 2.33); however, advice plus ONS may improve weight at four to six months and up to 12 months. Interventions may make little or no difference in FFM or global QoL scores at any time point.