Home mechanical ventilation: quality of life patterns after six months of treatment

Background It has been shown that home mechanical ventilation improves quality of life, but it has not been widely studied which particular patient groups benefit the most from starting this type of therapy. The purpose of this prospective observational study was to evaluate quality of life change patterns 6 months after initiation of home mechanical ventilation in patients suffering from chronic respiratory failure using patient reported outcomes. Methods We enrolled 74 chronic respiratory failure patients starting invasive or noninvasive home mechanical ventilation through the Semmelweis University Home Mechanical Ventilation Program. Quality of life was evaluated at baseline and at 6 months after initiation of home mechanical ventilation using the Severe Respiratory Insufficiency Questionnaire. Results Overall quality of life showed 10.5% improvement 6 months after initiation of home mechanical ventilation (p < 0.001). The greatest improvement was observed in Respiratory complaint (20.4%, p = 0.015), Sleep and attendant symptoms (19.3%, p < 0.001), and Anxiety related subscales (14.4%, p < 0.001). Interface (invasive versus noninvasive ventilation) was not associated with improvement in quality of life (p = 0.660). Severely impaired patients showed the greatest improvement (CC = -0.328, p < 0.001). Initial diagnosis contributed to the observed change (p = 0.025), with chronic obstructive pulmonary disease and obesity hypoventilation syndrome patients showing the greatest improvement, while amyotrophic lateral sclerosis patients showed no improvement in quality of life. We found that patients who were started on long term ventilation in an acute setting, required oxygen supplementation and had low baseline quality of life, showed the most improvement during the six-month study period. Conclusions Our study highlights the profound effect of home mechanical ventilation on quality of life in chronic respiratory failure patients that is indifferent of ventilation interface but is dependent on initial diagnosis and some baseline characteristics, like acute initiation, oxygen supplementation need and baseline quality of life. Trial registration This study was approved by and registered at the ethics committee of Semmelweis University (SE TUKEB 251/2017; 20th of December, 2017).


Background
Chronic respiratory failure affects many people, causing hypoxia, hypercapnia, secondary symptoms, diminished health related quality of life (HRQL) and adverse outcomes. Long term mechanical ventilation, which can be supplied through a noninvasive (mask) or invasive interface (tracheostomy) if noninvasive ventilation is contraindicated or not feasible (e.g. in the case of bulbar symptoms), improves outcomes in many different types of chronic respiratory failure [1][2][3]. Home mechanical ventilation (HMV) reduces costs, infection rates, optimizes medical care utilization and perhaps most importantly for the patients, improves HRQL [4]. Hence home mechanical ventilation has been an increasing practice worldwide [5,6]. As guidelines do not include specific ventilation goals and settings, practice varies greatly in different regions [7]. Because of this, HRQL measurements are important tools for quality control, optimization of therapy and even outcome prediction [8].
HRQL is usually measured through questionnaires and as general health quality surveys might be inaccurate in specific diseases groups, there have been attempts to establish a more focused questionnaire useful in chronic respiratory failure patients [9][10][11]. Perhaps the most efficient questionnaire for this purpose designed up until now is the SRI Questionnaire, which is a multimodal tool with high psychometric properties, specifically designed to evaluate HRQL in patients battling chronic respiratory conditions [12]. The SRI Questionnaire has been validated in several different conditions and has been proven to be a superior HRQL evaluation tool for patients receiving home mechanical ventilation [9,13]. The questionnaire has been used to evaluate HRQL in both invasively and noninvasively ventilated patients and has been validated in several different languages, including Hungarian [13][14][15][16][17][18][19][20][21].
The effect of starting home mechanical ventilation on patients' HRQL has been studied before, but these studies either used a general HRQL survey [22] or focused only on a well-defined, selected population often excluding patients with tracheostomas, or patients that were recruited after inability to wean or after an acute worsening of their chronic condition [13]. As a considerable proportion of home mechanical ventilation patients are started on long term therapy after acute exacerbation or failure to wean from ventilatory support, these studies might not reflect real-life populations, as more severe patients are underrepresented (see Table 1). Studies have also shown that HRQL differs greatly in chronic respiratory failure patients with different diseases (see Table 1) and we can expect that the change in HRQL induced by home mechanical ventilation also varies [7,10,13,14]. The factors that can possibly affect change in HRQL are baseline characteristics, initial diagnosis, initial HRQL, type of interface used for ventilation, duration of ventilation and lung function test parameters. Studies performed on specific patient groups might have different follow-up plans, and their HRQL changes might not be comparable.
To our knowledge, no prospective study has examined one specific HMV protocol and its HRQL improving effects in a large, unselected case mix population.
The aim of this study was to describe the effect of initiation of protocolized, optimally conducted home mechanical ventilation with standardized follow-up in a real-life, mixed case group of chronic respiratory patients using a disease specific HRQL survey and to evaluate expected HRQL change patterns for these patients.

Study design and participants
The purpose of this prospective observational follow-up study was to evaluate, using the Hungarian validated version of the Severe Respiratory Insufficiency Questionnaire [21], HRQL change 6 months after the initiation of home mechanical ventilation in patients suffering from chronic respiratory failure, and identify possible factors influencing said changes. We enrolled patients diagnosed with chronic respiratory failure in need for long term mechanical ventilation, who were treated through the Semmelweis University Home Mechanical Ventilation Program from January 2014 to December 2018. Patients were referred to the Program either through elective workup for a chronic condition (elective initiation) or after recovery from a previously unknown chronic condition (acute initiation). All patients were enrolled based on work up during stable conditions and diagnosed according to available international guidelines [1][2][3]. Patients who were unable to complete the questionnaire were excluded. Patients who deceased before the study completion were noted as lost to follow up. Written informed consent was obtained from all patients included in the study. The study was approved by the ethics committee of Semmelweis University.

Home mechanical ventilation initiation and follow-up
Home mechanical ventilation was initiated according to the Semmelweis University Home Mechanical Ventilation Program Guideline either after discharge from an acute hospitalization or during an elective hospital admission. Given the current reimbursement system utilized in Hungary, patients have the benefit of receiving patient tailored ventilation plans and equipment supply, including cough assisting devices, if needed. Mechanical ventilation was supplied through A40 or Trilogy 100 home mechanical ventilators (Koninklijke Philips N.V., Patients received individually tailored therapy with adequate interface (nasal, full, total face masks or cuffed tracheostomy tubes with reusable inner cannula), personalized daily ventilation plans and cough assisting device (CoughAssist T70, Koninklijke Philips N.V., Amsterdam) and/or tracheal suction devices if peak expiratory flows were below 2.5 L/s, suggesting insufficient coughing. Home mechanical ventilation was overseen by a voluntary family member or caregiver trained by our institution in skills specific to the patient's treatment plan. Patients did not receive other institutionalized assistance. Optimal care was achieved by frequent physician follow up to achieve high compliance with therapy and continuously maintained treatment goals. Patients were followed up monthly or bimonthly by a physician of the program, trained in home mechanical ventilation. Data were not routinely collected for the purpose of this study during these physician visits.

Data collected
Demographic data (age, sex, initial diagnosis), treatment characteristics (initiation type, interface, daily ventilation need, O 2 supplementation need), arterial blood gas values and lung function tests were collected at baseline, ventilator settings and blood gas values (if feasible) were collected at 6 months. Arterial blood gas sampling was performed minimum 15 min after discontinuing ventilation and/or oxygen supplementation, unless patients were ventilator dependent and could not be disconnected for even short periods of time. Lung function tests were performed with the Piston PinkFlow meter (Piston Ltd., Budapest, Hungary). Initial diagnosis leading to chronic respiratory failure and long-term mechanical ventilation need was identified as chronic obstructive pulmonary disease (COPD), restrictive chest wall disease (RCWD), obesity hypoventilation syndrome (OHS), slowly progressing neuromuscular disease (NMD) or progressive neuromuscular disease (amyotrophic lateral sclerosis) (ALS).

HRQL assessment
HRQL was assessed at baseline and 6 months after initiation of home mechanical ventilation using the Hungarian version of the SRI Questionnaire. The questionnaire consists of 49 statements and a five-point Likert scale. The results are interpreted in 7 different subscales (SRI-RC: Respiratory complaints, SRI-PF: Physical functioning, SRI-AS: Attendant symptoms and sleep, SRI-SR: Social relationships, SRI-AX: Anxiety, SRI-WB: Psychological well-being and SRI-SF: Social functioning) and the Summary Score (SRI-SS), with a value between 0 to 100, with 100 being the highest score. Patients were asked to fill out self-administered questionnaires and were assisted by caregivers if physical limitation or eyesight problems prohibited participation. The baseline questionnaire was filled out and collected during participating patients' hospital stay, while the follow-up questionnaires were collected by the visiting physician. Patients referred to our Program after acute hospitalization were asked to base their baseline answers on the last month prior to their acute illness to reflect chronic health status.

Statistical analysis
Continuous variables are described with means and standard deviation (±SD) and categorical variables with frequencies and percentages (n and %). SRI scores were analyzed with Paired t-test or, in case of non-normally distributed data, Signed Rank Test. Characteristics were compared with t-test or, in case of non-normally distributed data, Mann-Whitney U Rank sum test. Diagnostic

Patients
Out of the 75 patients enrolled, 2 were excluded because they were unable to complete HRQL forms due to cognitive impairment. One patient was deceased before study completion, and 6 patients were excluded because of missing data. A total of 66 patients completed the study. Baseline characteristics of the patients are listed in Table 2.
Indication for invasive ventilation was bulbar symptoms in 10 (71.4%) and more than 16 h of ventilation in 4 (28.6%) of the 14 cases in the study. Invasively ventilated patients had higher daily ventilation need Patients initiated after acute hospitalization had no significant differences in baseline ventilation need, supplementary oxygen need, lung function or blood gas values compared to those enrolled electively.

Baseline HRQL
Overall SRI score was 57.7 (±14.4) and several subscales showed values under 60, corresponding to limited HRQL with SRI-RC, −PF, −AS, −WB and -SF being the aspects with the lowest values (Fig. 1).

Differences in baseline HRQL scores by interface and diagnosis
As can be seen from Fig. 1, none of the baseline SRI subscales differed in patients treated through invasive or noninvasive interface. Baseline SRI-AS and -AX subscales were significantly associated with initial diagnosis (p = 0.048 and p = 0.018 respectively). The SRI-AS scores were the lowest in OHS and ALS patients, while SRI-AX scores were the lowest in COPD and ALS patients (Fig. 1).

Differences in baseline HRQL scores by initiation type and O 2 need
In addition, there was no difference in baseline SRI scores in patients initiated acutely compared to those initiated electively, or in patients needing O 2 supplementation compared to patients who did not. Furthermore, SRI subscales showed no correlation with initial blood gas values, lung function test parameters, or hours of ventilation need (data not shown).
Overall HRQL change at 6-month follow-up There was an 10.5% overall improvement of SRI summary scores from baseline to six-months (57.7 ± 14.4 vs. 68.2 ± 15.8, p < 0.001). All SRI subscales showed significant improvement during the first 6 months of home mechanical ventilation (Table 3.)

Factors effecting HRQL change HRQL change by interface and diagnosis
As can be seen from Fig. 2, the choice of interface did not affect change in SRI subscales (p = 0.660). Changes in SRI-RC, −PF, −SF and -SS subscales were significantly influenced by initial diagnosis (p = 0.025, p < 0.001, p = 0.002 and p = 0.025 respectively).
When further analyzing HRQL changes within diagnostic groups, we found that different diagnostic groups had different HRQL change patterns, which is visualized in the bubble chart depicting relative changes in different patient groups (Fig. 3). As can be seen from the figure, the patients benefiting most from HMV were COPD patients, while OHS patients improved across the most SRI subscales. SRI-RC subscale improved in all groups but ALS patients, and most prominently in COPD and OHS patients. SRI-PF and -AX scores improved only in COPD and OHS patients. SRI-SR subscales showed no change in any of the patient groups, except for OHS patients. SRI-WB and-SF improved only in OHS, and SRI-SF actually declined significantly in ALS patients. Overall SRI-SS scores improved in all patient groups except for ALS (see also Additional file 1).

HRQL change by initiation type, O 2 need, and baseline HRQL
Summary score improved significantly more in patients initiated acutely compared to patients initiated electively (12.3 ± 16.8 vs. 7.4 ± 9.4, p = 0.029). Change in SRI subscales showed no correlation with initial blood gas values, lung function test parameters or hours of ventilation need (data not shown). In patients using O 2 supplementation SRI-PF, −SF and -SS subscales improved significantly compared to patients that did not require  (Fig. 4). This phenomenon was most prominent in SRI-RC and SRI-AS scales.

Discussion
The aim of our study was to assess HRQL change induced in a mixed case chronic respiratory failure patient population in one HMV center. Our cohort included patients with a variety of underlying conditions, ventilated both invasively and noninvasively, initiated acutely and electively, resulting in a real-life study population. Frequent follow up visits ensured optimal compliance and maintained treatment goals, ensuring that HRQL changes were comparable despite the mixed case study population.
We found that overall HRQL improved significantly in the first 6 months of home mechanical ventilation and the grade of HRQL change was influenced by initial diagnosis. Furthermore, we also found that patients with worse initial HRQL, initiated acutely and needing O 2 supplementation, had a greater improvement in HRQL and that HRQL did not improve in ALS patients but was maintained despite significant progression of the disease.
The group of patients recruited for this study had baseline characteristics similar to home mechanical ventilation populations described before (see Tables 1 and  2). Noninvasively ventilated patient populations have been known to be characterized by a daily ventilation need of 6.5 to 9 h, FVC% values of 42-62%, pO 2 of 66-73 Hgmm, and pCO 2 of 44-59, while invasively ventilated patients have been reported to have higher daily ventilation needs (18-20 h), with O 2 supplementation needed in 75% of patients [12][13][14]22]. In our study    [12][13][14]22]. The majority of patients were initiated on HMV after resolution of an acute exacerbation of a previously undiagnosed chronic condition, which might be explained by possible insufficient diagnostic algorithms and screening of patient populations with higher risk of chronic respiratory failure. Baseline HRQL in our study population was similar (SRI-SS: 57.7 ± 14.4) to values described in patients suffering from chronic respiratory failure (see Table 1), with the subscales of Respiratory complaints, Attendant symptoms and sleep, Physical functioning, Social functioning and Psychological well-being being the aspects most diminished. To our knowledge, no previous study compared patients ventilated invasively versus patients ventilated noninvasively, although patients ventilated through a tracheostomy are generally assumed to have diminished HRQL [14,34]. In our study population none of the baseline SRI subscales differed in patients treated through invasive or noninvasive interface, in patients initiated acutely or electively or in patients needing O 2 supplementation compared to patients who did not, and SRI subscales showed no correlation with initial blood gas values, lung function test parameters or hours of ventilation need, pointing to the fact that HRQL is Fig. 2 Bar graphs of changes in the scores of SRI subscales at the 6-month follow-up in the whole study group, by interface and by diagnostic groups. Boxes represent means, error bars represent standard error. Significant differences are marked with asterisk. COPD: chronic obstructive pulmonary disease, RCWD: restrictive chest wall disease, OHS: obesity hypoventilation syndrome, NMD: neuromuscular disease, ALS: amyotrophic lateral sclerosis, RC: Respiratory complaints, PF: Physical functioning, AS: Attendant symptoms and sleep, SR: Social relationships, AX: Anxiety WB: Psychological well-being, SF: Social functioning, SS: Summary Score not merely influenced by the severity of respiratory failure, although the relatively low number of invasively ventilated patients might contribute to this finding.
We found that quality of life increased significantly during the six-month observation period (SRI summary score changed from 57.7 ± 14.4 vs. 68.2 ± 15.8, p < 0.001), which is similar to changes reported by studies in more selective patient populations (see Table 1). HRQL subscales that were most robustly improved were Respiratory complaints and Attendant symptoms and sleep, which are indeed the main goals of long-term mechanical ventilation. Anxiety subscales also showed significant improvement, which is most likely a consequence of the improvement of the two prior scales. Social relations and Social functioning showed no change, which suggests that these subscales are influenced by more complex disease attributes and are not solely related to respiratory symptoms. The change in Physical functioning also showed no significant overall improvement, which is understandable since several patient groups had stable or progressive neuromuscular impairment. When looking at specific disease groups, we found that diseases where neuromuscular involvement was not present (COPD and OHS), Physical functioning subscale actually improved significantly and considerably (16.2 ± 12.2, p = 0.002 and 19.8 ± 24.9, p < 0.001 respectively).
The overall SRI HRQL change found in our study is similar to changes previously reported in a study using solely noninvasive ventilation [13]. Furthermore, our study found higher post-ventilation HRQL values than a study in a population ventilated invasively [14]. As an additional interesting finding, our study suggests that an increase in HRQL does not depend on interface [35], meaning HRQL can both be increased through invasive and noninvasive ventilation. This is important because there is a general fear towards tracheostomy in patients requiring long term mechanical ventilation [36]. The results of our study provide evidence that significant improvement might be achieved in HRQL even through an invasive interface. This can be attributed to patienttailored home mechanical ventilation therapy, which is feasible under current reimbursement system in Hungary and permits complex ventilation and physiotherapy plans for patients, despite vastly different needs. The other possible reason for the consistently improved outcome might be the frequent follow-ups provided by our program. Current international guidelines do not specify the frequency of follow up for home mechanically ventilated patients [1][2][3], but most centers reduce the frequency of home or ambulatory visits to twice a year after successful initiation [4,7]. In our patient population, however, follow-up was provided monthly for patients ventilated invasively and monthly or bimonthly for stable patients ventilated through a noninvasive interface.
When looking at other factors potentially influencing HRQL change, we found that HRQL increase was the progression. This is corroborated by the study previously published by our group that describes the current characteristics and prevalence of HMV in Hungary [37], which found that prevalence of home mechanical ventilation is lower than that found in other internationally published data, because many ventilator-dependent patients might not receive therapy in time. This finding is especially important, since there is still no clear data to help us differentiate between long term therapeutic home mechanical ventilation and palliative care with ventilation. The results of our study suggest that even patients with very poor HRQL benefit greatly from initiation of home mechanical ventilation.
Our results regarding the HRQL influencing effect of initial diagnosis corroborate the findings published previously [13,14]. Based on our findings, COPD patients can expect the most benefit in HRQL increase, which is important, since long term home mechanical ventilation is still controversial in this indication [38][39][40], although it should be noted that COPD is a heterogenous disease group and our study did not analyze disease phenotypes. Moreover, when analyzing patient groups, we found that OHS patients improved most consistently, on par with recently published guidelines [41]. Given that this disease effects a younger population and its main symptoms are debilitating daytime sleepiness and hypercapnia [42], this HRQL improvement has important socioeconomical consequences. It is also important to note that recent guidelines indicate most OHS patients might be managed with initial CPAP therapy. Prospective HRQL change studies might verify whether optimal improvement can be achieved by this approach.
In our study, the only patient group not showing improved HRQL scores were the ALS patients, which is on par with literature describing this progressive neuromuscular disease [43], although noninvasive ventilation has been shown to improve HRQL and even survival for a shorter period of time [44]. In our study, spanning a longer period, it is especially conspicuous that ALS patients were the only ones to show significant detriment in one subscale (SRI Social functioning change), suggesting that these patients often become isolated during the progression of their disease. However, it is noteworthy that despite HRQL subscale changes due to disease progression, overall Summary scores remained stable, mainly because Attendant symptoms and sleep, Anxiety and Respiratory complaints subscales counterbalance the decrease in Social functioning and Physical functioning subscales. This suggests that home mechanical ventilation might be a valid palliative care technique in ALS even in cases where significant lifespan or HRQL improvement can't be proven.
The main limitation of our study is the sample size that might be considered small; however previously published studies on HMV patients utilized similar sample sizes [11,13,14,21,23]. Our study was a single center, prospective HRQL change investigation under optimized care using the SRI Questionnaire, providing valuable "real-life" information on a mixed case population previously not extensively described in the literature. Another limitation is that the study was not designed to identify outcome differences in patients with different interfaces or initial diagnosis, nor was powered for multivariate analysis to exactly define confounding factors influencing subscale changes, as subgroup analysis would have resulted in groups that would have been too small to deduct relevant clinical correlations.

Conclusions
Starting home mechanical ventilation is accompanied by improved quality of life in several patient groups suffering from chronic respiratory failure. Our findings suggest that HRQL improvement is independent of classic markers of the severity of chronic respiratory failure (e.g., baseline lung function, arterial blood gas value, hours of ventilation need) or interface used for ventilation, but it is dependent on the type of disease causing the chronic respiratory failure, initiation type (acute versus elective), initial HRQL and O 2 supplementation need. Our results further suggest that acutely initiated, O 2 dependent COPD and OHS patients with low initial HRQL can expect the most benefit, while ALS patients can expect maintenance of overall HRQL despite ongoing neurological deterioration. Further prospective studies are needed to confirm our findings and to identify whether HRQL changes correlate with disease progression and mortality, and how often HRQL needs to be measured during the follow up of patients receiving home mechanical ventilation in order to better improve the lives of patients suffering from chronic respiratory failure.
Additional file 1. HRQL subscale changes according to diagnostic groups. Contains exact numerical HRQL subscale changes according to diagnostic groups with significant values noted.