Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative condition, respiratory failure being the commonest cause of death. Quality of life and survival can be improved by supporting respiratory function with non-invasive ventilation. Transcutaneous carbon dioxide monitoring is a non-invasive method of measuring arterial carbon dioxide levels enabling simple and efficient screening for respiratory failure. The aim of this study was to validate the accuracy of carbon dioxide level recorded transcutaneously with a TOSCA 500 monitor. It is a prospective, observational study of 40 consecutive patients with ALS, recruited from a specialist ALS clinic. The partial pressure of carbon dioxide (PCO(2)) in each patient was determined by both transcutaneous monitoring and by an arterialized ear lobe capillary blood sample. The carbon dioxide (CO(2)) levels obtained with these two methods were compared by Bland-Altman analysis. The results showed that the mean difference between arterialized and transcutaneous readings was - 0.083 kPa (SD 0.318). The Bland-Altman limits of agreement ranged from 0.553 to - 0.719 kPa. The difference was < 0.5 kPa in 90% of the recordings. Four of the 40 measurements had a difference of > 0.5 kPa, with a maximum recorded difference of 0.95 kPa. In conclusion, non-invasive carbon dioxide monitoring using a TOSCA monitor is a useful clinical tool in neurology practice. Users should be aware of the possibility of occasional inaccurate readings. A clinically unexpected or incompatible reading should be verified with a blood gas analysis, especially when a decision to provide ventilatory support is required.