This study of MND patients attending a multidisciplinary clinic found that 66.7% of the patient cohort were treated with NIV and of this, 44.1% used NIV optimally. It has been shown in previous studies that better compliance and longer use of the machine is associated with prolonged survival with Piepers et al 2006 and Bourke et al 2003, both showing that survival improved with better compliance compared to those who did not tolerate or did not use NIV, while Kleopa et al 1999 showed that those who used NIV for 4 or more hours per day survived longer than those who did not use NIV [7, 12, 17]. When those with bulbar onset MND were compared with those with limb onset MND, the bulbar MND group had better compliance and better machine usage. This may be due to poorer respiratory function and a greater need for NIV amongst the bulbar group. Review of the literature shows varying results regarding compliance and tolerance of NIV in those with bulbar MND with some finding that those with bulbar onset MND had a decreased benefit due to intolerance and that NIV was difficult to establish in bulbar MND [7, 16]. In contrast, other studies have found that the area of MND onset did not affect tolerance .
In our population cohort, those with limb onset MND survived significantly longer than those with bulbar onset MND (33 months v 17 months). This is not unexpected as bulbar onset MND is known to have a worse prognosis . There was a statistically significant longer survival in those who used NIV compared to those who never used NIV (28 months v 12 months). This is in keeping with multiple studies including the only randomised control trial on this topic by Bourke et al 2006 [7, 12-17, 21, 30] Furthermore, there was significantly longer survival in those who used NIV optimally compared with those who never used NIV. This survival advantage has been previously demonstrated in several studies [7, 15, 31, 32]. Of additional interest, those who used NIV regardless of compliance and pattern of onset survived 2 months longer on average than those who optimally used NIV. This may be as a consequence of having smaller patient numbers in the optimal use NIV group (n=41 v n=93) and that this group had a slightly higher proportion of bulbar onset MND patients (51% v 40.4%). When this data was examined for each phenotypic subgroup, it was found that optimal use of NIV in the bulbar onset MND group compared with those that did not use NIV had a statistically significant longer survival (23 months v 8 months). This same survival benefit was not seen for the limb onset MND group. This finding would be in keeping with the results of Kleopa et al  who found that there was a larger survival benefit in those with bulbar onset MND after initiation of NIV as well as other studies with similar findings .
This study also found that those with bulbar onset MND have more marked impairment in PFT measurements at diagnosis compared to those with limb onset MND, with SNIP and PCF values being statistically significant lower in those with bulbar onset MND. This highlights a worse overall prognosis of bulbar onset MND when compared with limb onset MND. It has long been established that the rate of decline of respiratory function is related to mortality. We have shown in this study that the SNIP and PCF values declined significantly between diagnosis and values recorded within 6 months of death. SNIP measurements can better predict respiratory muscle function as they linearly decline with disease progression, can be better performed in advanced disease and are a reliable predictor of mortality and can better indicate earlier respiratory muscle dysfunction than FVC [24-28]. Figure 2 highlights the gradual decline of SNIP and PCF in the year before death of all patients for which this data was available. It could be concluded from this that serial SNIP and PCF measurements, as is carried out in our centre, are an accurate measure of respiratory muscle decline and can help predict mortality.
There is a lack of data regarding the best location for NIV set up for MND patients. As mentioned, our institute aims to set up NIV while the patient is in hospital so that nursing and physiotherapy staff can help with mask adjustment and therefore improve compliance.
Although 70% of the patients for which this data was available had their NIV set up while in hospital compliance amongst this group was lower compared with those set up at home. There may be several reasons for this including that those who are able to be set up at home may have less severe phenotypes and/or may be more willing to embrace the intervention. More research is required in this area, but it is a unique data point to our current study.
Advantages and Limitations One of the limitations of this study was that it was a retrospective study performed over a 5-year period and therefore it was not possible to standardise data collection of each patient. Practise in the clinic changed significantly during this time which influenced the data available. As the clinic became established a physiotherapist, speech and language therapist, neurology consultant and respiratory consultant saw the patient on every visit. This meant that compliance data was analysed and recorded and that the physiotherapist preformed spirometry and recorded SNIP and PCF values at every visit. This was not always the case with earlier clinics which resulted is a lack of PFT data for certain patients. Secondly, in some cases there was missing data especially elements of PFTs. This was due to a combination of factors including lack of availability in the PFT lab, patient difficulty travelling to appointments and patient difficulty carrying out the test. This affected our ability to accurately measure FVC decline over time. A further limitation is that not every patient was set up in NIV in our institute. A subset of patients were set up at other institutes prior to assessment at our institution. This resulted in difficulties retrieving compliance data for some patients. This was particularly the case in earlier clinics, but this issue was subsequently resolved by standardising practice with one primary vendor within the region. Given its observational nature, our study while showing robust associations cannot establish causation, and a variety of factors may have contributed to the observed difference in survival.
One of the main advantages of this study includes the large sample size of 111 patients. This sample is representative of MND patients from all over the south of Ireland. Furthermore, we believe that this is the first study of its kind in an Irish MND population. Another advantage was that there was a high uptake of NIV among our patient cohort with two-thirds of patients who continued to use NIV using it optimally. Finally, we analysed home and inpatient set up in our patient cohort and looked at how this affected compliance which is a factor rarely mentioned in other studies.