neutral quality studies - what are these? If they are studies showing no variation by nutritional status, as seems logical from the terms, then there are only 2 studies showing any variation out of the 21 studies identified. This is very weak evidence, and I am not convinced it should be used to develop this recommendation. There should be some information provided on the total numbers of people in the positive versus neutral studies. Is C reactive protein a measure of nutritional status. Albumin level is quoted but no mention of whether it is high or low or some of each which would muddy the picture. Could also represent liver mets so not primarily a marker for malnutrition. "have a reduced survival if they have had more weight loss at baseline". Weight loss is a common measure of extent of disease and therefore one would expect it to be negatively correlated with outcomes. It is not accurate to use it as a proxy for nutritional status.
Neutral quality study relates to the strength of the study design and not the results of the study. Therefore the results actually show that 19/21 studies found that malnutrition had an impact on various patient outcomes, and only 2 studies showed no association. Recommendation statement will therefore not change as the overall volume of evidence supports this. Additional wording added to the summary to make this clearer. A definition of positive, neutral and negative quality will also be added into the methodology section.
Agree that C reactive protein and albumin and weight loss alone should not be used as key measures of nutritional status, however this was what was used in the studies, and this is why many of the studies could not be rated as positive quality for study design due to inappropriate use of outcome measures. Q4 provides further information on using validated nutritional assessment tools.
Teresa Brown, on behalf of the authors.