Study Overview
Objective. To determine if an intervention directed toward early detection of pulmonary exacerbations using electronic home monitoring of spirometry and symptoms would result in slower decline in lung function.
Design. Multicenter, randomized, nonblinded 2-arm clinical trial.
Setting and participants. The study was conducted at 14 cystic fibrosis centers in the United States between 2011 and 2015. Cystic fibrosis patients (stable at baseline, FEV1 > 25% predicted) at least 14 years old (adolescent and adults) were included and randomized 1:1 to either an early intervention arm or usual care arm.
Intervention. The intervention arm used home-based spirometers and patient-reported respiratory symptoms using the Cystic Fibrosis Respiratory Symptoms Diary (CFRSD), which was to be completed twice weekly and collected by the central AM2 system. This AM2 system alerted sites to contact patients for an acute pulmonary exacerbation evaluation when FEV1 values fell by greater than 10% from baseline or CFRSD worsened from baseline in two or more of eight respiratory symptoms. The usual care arm patients had quarterly CF visits and/or acute visits based on their need.
Main outcome measures. The primary outcome variable was the 52-week change in FEV1 volume in liters. Secondary outcome variables were changes in CFQ-R (Cystic Fibrosis Questionnaire, revised), CFRSD, FEV1 % predicted, FVC in liters, FEF25-75%, time to first acute pulmonary exacerbation, time from first pulmonary exacerbation to subsequent pulmonary exacerbation, number of hospitalization days, number of hospitalizations, percent change in prevalence of Pseudomonas or Staphylococcus aureus and global assessment of protocol burden score.
Main results. A total of 267 patients were randomized. The results were analyzed using intention-to-treat analysis. There was no significant difference between study arms in 52-week mean change in FEV1 slope (mean slope difference, 0.00 L, 95% confidence interval, –0.07 to 0.07; P = 0.99). The early intervention arm subjects detected exacerbations sooner and more frequently than usual care arm subjects (time to first exacerbation hazard ratio, 1.45; 94% confidence interval, 1.09 to 1.93; P = 0.01). Adverse events were not significantly different between treatment arms.
Conclusion. An intervention of electronic home monitoring of patients with CF was able to detect more exacerbations than usual care, but this did not result in slower decline in lung function.
Commentary
Establishing efficacy and safety of home monitoring is a popular research topic in the current era of information technology. Most data to date has come from chronic adult disease such as heart failure, diabetes, or COPD [1]. While relatively rare, CF is a chronic lung disease that could potentially benefit from home monitoring. This is supported by previous evidence suggesting that up to a quarter of pulmonary exacerbations in CF patients result in worsened baseline lung function [2]. Close monitoring of symptoms and FEV1 using home monitoring was hypothesized to improve management and long-term function in this population. Indeed, in children with CF, electronic home monitoring of symptoms and lung function was able to detect pulmonary exacerbations early [3]. Frequency of monitoring is widely variable between centers, and some suggest aggressive monitoring of CF provides better clinical outcomes [4]. Current CF guidelines do not make specific recommendations regarding frequency of monitoring.
In this study, Lechtzin et al attempted to determine if the early detection of acute pulmonary exacerbations in CF patients by home monitoring and treatment would prevent progressive decline in lung function. This multicenter randomized trial was conducted at large CF centers in the US with a total cohort of 267 patients. The study had a mean follow-up time of 46.8 weeks per participant in the intervention arm and a mean follow-up time of 50.9 weeks per participant in the usual care arm. Given the predefined follow-up length (52 weeks) the primary outcome of FEV1 in liters was deemed sensitive enough to detect a decline of lung function. However the discrepancy between follow-up times with the intervention group having a 4.1-week shorter mean follow-up than the usual care could have influenced the interpretation of the results. Additionally, a large percentage of these patients were clinically stable at initial enrollment, with an average FEV1 % predicted of 79.5%. The stability of initial participants raises questions as to the efficacy of home monitoring in CF patient with moderate to severe lung disease. Mostly importantly, due to the nature of intervention the study could not be blinded, which could have substantially increased anxiety and self-awareness of patients in reporting their symptoms in the intervention arm.
Currently, an established consensus definition of pulmonary exacerbations of CF is lacking. Previous studies have proposed several different criteria of acute pulmonary exacerbations. Most proposed definitions depend on symptom changes such as cough, sputum, chest pain, shortness of breath, fatigue and weight-loss, making the definition less specific or objective.
The number of acute visits in the intervention arm was significantly higher than that in the usual care arm (153 vs 64). Despite a higher number of visits with intervention group, a significant number of these visits did not lead to a diagnosis of acute pulmonary exacerbation. Reportedly, 108 acute visits met protocol-defined pulmonary exacerbation and 29 acute visits did not meet protocol-defined pulmonary exacerbation in the intervention arm compared to 44 and 12 respectively in the usual care arm of the study. Given that the groups had similar baseline demographics and were randomized appropriately, one would expect that the number of acute visits severe enough to meet protocol-defined criteria as a pulmonary exacerbation would be similar in both groups. However, the absolute number of protocol-defined pulmonary exacerbations was far greater in the intervention group. Therefore, one could question the clinical significance of what was defined as acute pulmonary exacerbation. Potentially, the elevation of the absolute number of protocol-defined pulmonary exacerbations in the intervention group was simply due to increased surveillance. If the former were correct, one would expect the lack of identification/treatment of a significant number of pulmonary exacerbations in the usual care group would have led to a larger decline in FEV1 after 52 weeks than was seen in the results when compared to the intervention group. Given that the results of the study indicate no significant difference in change in FEV1 between study arms, perhaps the studied parameters in the intervention group were overly sensitive.
Of note, the usual care arm did have a statistically significant higher rate of hospitalizations and IV antibiotic use, suggesting that early identification of acute visits can identify patients earlier in the course of an acute pulmonary exacerbation and prevent higher level of care, though at the expense of more acute event “false positives,” or over-diagnosis. This trade-off may not result in cost saving, though this was not a consideration of this study. Additionally, there was likely difference in treatment, as treatment was not standardized, with potential implications for the validity of results.
The early intervention protocol was not only shown to lead to increased visits with no benefit in lung function decline, but as one may expect, also proved to be remarkably burdensome to many patients compared to the usual care protocol. Entering data on a weekly basis (or perhaps even monthly) was found to be burdensome in many remote-monitoring trials [5]. This may be especially apparent in a younger age group: in this study the average age of the study population was between 18 and 30 years of age. It can be hypothesized that this age group may not have enough responsibility, time, or enthusiasm to participate in home monitoring. Home monitoring maybe more effective in a disease condition where the average age is older or in a pediatric population in whom the parents oversee the care of the patient or have more time and receive subjective benefit from home monitoring services.
Less may be sufficient. The current study suggests that the home monitoring in CF may increase medical expense and unnecessary antibiotic use with no improvement in lung function. It is difficult to assess from this study the impact that the burden of home monitoring would have on clinical outcomes, however, previous meta-analysis of data studying COPD populations using home monitoring system, interestingly, also had increased health service usage and even led to increase in mortality in the intervention group compared with usual care group [1,6].
Perhaps the negative result of current study is due to the oftentimes variable definitions of and management algorithms for pulmonary exacerbations rather than the home monitoring system itself. Limited evidence exists for optimal threshold identification [7]. Aggregated, large amounts of data gathered by telemonitoring have not been proven to be used effectively. Moreover, as mentioned, a clear definition and management guidelines for pulmonary exacerbation are lacking. As a next step, studies are ongoing to evaluate how to use the collected data without increasing harm or cost. This could utilize machine learning or developing a more specific model defining and predicting pulmonary exacerbations as well as standardized indications for antibiotic therapy and hospitalization.