METHODOLOGY: A prospective comparison of subjective global assessment (SGA), and anthropometry (mid-arm muscle circumference, MAMC; triceps skinfold thickness, TST) between elderly stroke patients on long-term NG feeding and matched controls was performed. Selected clinicians and carers of patients were interviewed to assess their knowledge and attitudes to gastrostomy feeding.

RESULTS: 140 patients (70 NG, 70 oral) were recruited between September 2010 and February 2011. Nutritional status was poorer in the NG compared to the oral group (SGA grade C 38.6% NG vs 0% oral, p<0.001; TST males 10.7 + 3.7 mm NG vs 15.4 + 4.6 mm oral, p<0.001; MAMCmales 187.9 + 40.4 mm NG vs 228.7 + 31.8 mm oral, p<0.001). 45 (64.3%) patients on long-term NG feeding reported complications, mainly consisting of dislodgement (50.5%), aspiration of feed content (8.6%) and trauma from insertion (4.3%). Among 20 clinicians from relevant speciliaties who were interviewed, only 11 (55%) clinicians would routinely recommend a PEG. All neurologists (100%) would recommend a PEG, whilst the response was mixed among non-neurologists. Among carers, lack of information (47.1%) was the commonest reason stated for not choosing a PEG.