**2. Outcome measures for respiratory physiotherapy**

Respiratory physiotherapy in CF involves a wide range of interventions and among them airway clearance techniques and exercise training are recognised as the most important. The primary aim of airway clearance techniques is to relieve the airway obstruction by promoting the normal mucociliary clearance mechanism of the lungs and facilitating expectoration, thus reducing the risk of infection and inflammation. A variety of airway clearance techniques have been developed. Some involve airway oscillation, some are independently performed and others require electricity or physical assistance [5]. Exercise training is advocated as an important package of care delivered to patients with CF [5], since exercise intolerance has been associated with reduced survival [45]. Observed benefits of exercise training include slow pulmonary function decline [46], reduced dyspnoea and improved exercise capacity, muscle strength and health-related quality of life (HRQoL) [47].

Although adhering to airway clearance techniques [7] and exercise training is generally regarded as beneficial for patients with CF, there is no consensus about the superiority of one technique over the other [5]. Methodological issues, such as the underpowered samples or the selection of the outcome measures, may explain this lack of evidence and thus hamper the establishment of the effectiveness of these interventions. A clear example of this issue has been the use of forced expiratory volume in the first second (FEV1) as a gold standard to assess the impact of the mentioned interventions for many years, which is currently considered as not a sensitive measure to be used in respiratory physiotherapy [5].

Most of the clinically available outcome measures are not specifically related to the physio‐ therapy intervention employed and may be affected by other factors. This means that there is no gold standard outcome measure that is specifically related to respiratory physiotherapy interventions. Moreover, there are many doubts about the accuracy, reliability, sensitivity and validity of the current measures. Given this problematic situation, several researchers have been investigating the potential of other objective, simple and non-invasive measures to be used as outcome measures in respiratory physiotherapy.

The outcome measures most commonly used by respiratory physiotherapists to monitor their interventions and evaluate their practice are: FEV1, respiratory sounds, sputum weight, measures of oxygenation, chest radiography, dyspnoea, exercise capacity and HRQoL. Computerised respiratory sounds, lung ultrasound, fat-free mass, inspiratory muscle strength and endurance, physical activity and burden of treatment are some examples of these emerging outcome measures to assess and monitor respiratory physiotherapy interventions in CF. Each one of these outcome measures, their strengths and weaknesses are presented in detail below according to their novelty in the field (i.e., conventional and emerging).

### **2.1. Conventional outcome measures**

physiotherapy is a key element of care for patients with CF, as it aims at both rehabilitation and prevention [2]. Specifically, respiratory physiotherapy is used to deal with the progressive loss of pulmonary function accompanied by symptoms of cough, excessive sputum produc‐ tion, dyspnoea, exercise intolerance, reduced functionality and impaired quality of life. To respond to these multiple problems and needs, respiratory physiotherapy involves a range of strategies and techniques, such as airway clearance, exercise training and breathlessness management, which have an overall aim of reducing the progression of the disease [3, 4]. However, there is a lack of evidence to suggest the superiority of one technique over the other [5] and to determine which strategies promote the adherence of this population to regular physical activity [6]. Nevertheless, the lack of evidence does not mean lack of benefit. Instead, methodological issues, such as the selection of the outcome measures, may hamper the

Respiratory physiotherapists use several outcome measures to monitor and evaluate their interventions. Most of the clinically available outcome measures are not specific for the physiotherapy intervention employed and may be affected by other factors. This means that there are no gold standard outcome measures specifically related to respiratory physiotherapy interventions. Thus, in all areas of respiratory physiotherapy, one of the barriers to generate evidence has been the lack of accurate, reliable, sensitive and valid outcome measures. To

This chapter starts by providing an overview of the problem. It then presents and discusses the strengths and weaknesses of the commonly used clinical outcome measures and other measures that have been gaining interest in the assessment and monitoring of respiratory physiotherapy interventions in CF. The chapter ends with a brief conclusion. A comprehensive overview of the available outcome measures is sought to be provided, with particular emphasis on their strengths and limitations that should be recognised when interpreting the results.

Respiratory physiotherapy in CF involves a wide range of interventions and among them airway clearance techniques and exercise training are recognised as the most important. The primary aim of airway clearance techniques is to relieve the airway obstruction by promoting the normal mucociliary clearance mechanism of the lungs and facilitating expectoration, thus reducing the risk of infection and inflammation. A variety of airway clearance techniques have been developed. Some involve airway oscillation, some are independently performed and others require electricity or physical assistance [5]. Exercise training is advocated as an important package of care delivered to patients with CF [5], since exercise intolerance has been associated with reduced survival [45]. Observed benefits of exercise training include slow pulmonary function decline [46], reduced dyspnoea and improved exercise capacity, muscle

Although adhering to airway clearance techniques [7] and exercise training is generally regarded as beneficial for patients with CF, there is no consensus about the superiority of one

establishment of the effectiveness of the respiratory physiotherapy techniques.

overcome this problematic issue, new measures have been emerging.

38 Cystic Fibrosis in the Light of New Research

**2. Outcome measures for respiratory physiotherapy**

strength and health-related quality of life (HRQoL) [47].

### *2.1.1. Forced expiratory volume in the first second*

The most common pulmonary function test performed to assess respiratory physiotherapy interventions is the forced spirometry, i.e., the volume and/or flow of air that can be inhaled and exhaled as a function of time. The procedure consists in three distinct phases: (1) maximal inspiration followed by an expiration at functional residual capacity; (2) a ''blast'' of exhalation; and (3) continued complete exhalation until the end of test [8]. First, the patient should exhale until he or she reaches the functional residual capacity and then be instructed to inhale rapidly and completely. In this phase, the mouthpiece should be placed in the patient's mouth and indications should be given for the patient to blow as much and as fast as possible and to keep blowing until totally emptying the lungs. Spirometry has been described as a cost-effective, simple, reliable, valid and easy-to-interpret bedside measure [8]. The most used pulmonary function parameter is the forced expiratory volume in the first second (FEV1), followed by the forced vital capacity (FVC) and the ratio between FEV1/FVC. Measurements are taken considering patient's gender, age, height, weight and race and are then compared with predicted values.

Spirometry, namely FEV1, has been used to assess the effectiveness of respiratory physiother‐ apy interventions. However, contradictory findings have emerged. Pfleger et al. (1992) found significant improvements in FEV1 after autogenic drainage and high-pressure PEP-mask (n=14, Cohen's dz 0.09 and 0.22) [9]. Jarad, Powell and Smith (2010) reported a statistically significant reduction in FEV1 following hydro-acoustic therapy (n=19, Cohen's dz=0.12) or flutter (n=19, Cohen's dz=0.06) [10]. Nevertheless, these changes returned to baseline on the second study day [10].

In two recent reviews where conventional physiotherapy [11] and the active cycle of breathing techniques (ACBT) [12] were compared with other airway clearance techniques, no significant differences were observed between the techniques in terms of pulmonary function measured with spirometry [11]. Nevertheless, when oscillating devices for airway clearance were used, significant results were observed in pulmonary function, the FEV1 being the primary outcome measure more frequently reported [13].

Findings in the literature about the effectiveness of respiratory physiotherapy interventions in CF remain controversial when FEV1 is considered as the outcome measure. This is in part due to the fact that accuracy and sensitivity of spirometry depends on many factors that are difficult to control and not related to the intervention itself. Some examples of these factors are the transducer characteristics, presence or absence of an in-line filter, presence or absence of a display, patient's mood and motivation to cooperate, relationship between the patient and the technician, among others. Therefore, spirometry might be unsuitable or its reliability may be affected in a number of situations, for example if the equipment or settings change, if patients are unwilling or unable to collaborate (e.g. children, people with dementia), or if pain or discomfort is present. Hence, this measure should be routinely used to characterise the pulmonary function of patients with CF, but not to assess the effectiveness of respiratory physiotherapy interventions.

### *2.1.2. Respiratory sounds*

Lung auscultation, performed with conventional stethoscopes, is one of the oldest and most used techniques to diagnose and monitor respiratory diseases [14, 15]. It consists in acquiring acoustic signals from the lung structures during spontaneous or controlled volume or flow breathing, and classifying the respiratory sounds as normal or abnormal (e.g., adventitious respiratory sounds, such as crackles and wheezes) [15]. Auscultation is recognised as an efficient and safe method for the early detection of respiratory diseases as it is non-invasive, practical, low cost and easy to apply in all clinical settings and patients, irrespective of patients' age and severity of the disease [14-16].

The efficiency of this method depends on the hearing ability of the health professionals [17], their capacity to memorise different sound patterns [18] and on the quality of the acoustic properties of the stethoscope being used [17]. Considering these limitations and subjectivity, reliability studies have been performed. In CF there are no reliability studies using conven‐ tional stethoscopes. However, in other respiratory diseases, poor to fair correlations between different raters have been reported, either in taped recorded sounds (kappa=0.26 and coeffi‐ cient of reliability of less than 60%) [19, 20] and real-time auscultation (–0.02*<*kappa*<*0.77) [21]. Using digital stethoscopes, one study conducted in adult patients with CF assessed the interrater agreement between real-time manual annotation of respiratory sounds and automatic detection of respiratory sounds through a computerised system [22]. Poor to moderate correlations were found (–0.20*<*kappa*<*0.60) [22]. Similar results were found in children with respiratory diseases (–0.08*<*kappa*<*0.86) [23–25]. Direct comparisons between conventional and digital stethoscopes showed that, although digital stethoscopes have a better sound quality [26], the performance of health professionals to detect respiratory sounds is not enhanced by its use [27].

It is also important to note that, in all the studies, the recognition of crackles was always less accurate than the recognition of wheezes. As respiratory sounds in CF are mainly characterised by the presence of crackles [28] and conventional auscultation has provided poor reliability results, particularly for detection of crackles, it can be concluded that respiratory sounds are not a reliable outcome measure for CF diagnosis and monitoring. Also due to its poor relia‐ bility, conventional auscultation has not been used as an outcome measure for respiratory physiotherapy interventions in recent research [7]. Nevertheless, due to its simplicity and wide availability in all clinical settings, it is still recurrently used in clinical practice by physiothera‐ pists to monitor patients with CF and to define therapeutic approaches [3]. Despite these limitations, the advantages of using lung auscultation should not be overlooked, and currently significant research efforts are being conducted to create equipment capable of overcoming the subjectivity associated with conventional auscultation while preserving its main advan‐ tages (i.e., portability, patients' minimal cooperation and cost-effectiveness) [16].

### *2.1.3. Sputum weight*

Cohen's dz 0.09 and 0.22) [9]. Jarad, Powell and Smith (2010) reported a statistically significant reduction in FEV1 following hydro-acoustic therapy (n=19, Cohen's dz=0.12) or flutter (n=19, Cohen's dz=0.06) [10]. Nevertheless, these changes returned to baseline on the second study

In two recent reviews where conventional physiotherapy [11] and the active cycle of breathing techniques (ACBT) [12] were compared with other airway clearance techniques, no significant differences were observed between the techniques in terms of pulmonary function measured with spirometry [11]. Nevertheless, when oscillating devices for airway clearance were used, significant results were observed in pulmonary function, the FEV1 being the primary outcome

Findings in the literature about the effectiveness of respiratory physiotherapy interventions in CF remain controversial when FEV1 is considered as the outcome measure. This is in part due to the fact that accuracy and sensitivity of spirometry depends on many factors that are difficult to control and not related to the intervention itself. Some examples of these factors are the transducer characteristics, presence or absence of an in-line filter, presence or absence of a display, patient's mood and motivation to cooperate, relationship between the patient and the technician, among others. Therefore, spirometry might be unsuitable or its reliability may be affected in a number of situations, for example if the equipment or settings change, if patients are unwilling or unable to collaborate (e.g. children, people with dementia), or if pain or discomfort is present. Hence, this measure should be routinely used to characterise the pulmonary function of patients with CF, but not to assess the effectiveness of respiratory

Lung auscultation, performed with conventional stethoscopes, is one of the oldest and most used techniques to diagnose and monitor respiratory diseases [14, 15]. It consists in acquiring acoustic signals from the lung structures during spontaneous or controlled volume or flow breathing, and classifying the respiratory sounds as normal or abnormal (e.g., adventitious respiratory sounds, such as crackles and wheezes) [15]. Auscultation is recognised as an efficient and safe method for the early detection of respiratory diseases as it is non-invasive, practical, low cost and easy to apply in all clinical settings and patients, irrespective of patients'

The efficiency of this method depends on the hearing ability of the health professionals [17], their capacity to memorise different sound patterns [18] and on the quality of the acoustic properties of the stethoscope being used [17]. Considering these limitations and subjectivity, reliability studies have been performed. In CF there are no reliability studies using conven‐ tional stethoscopes. However, in other respiratory diseases, poor to fair correlations between different raters have been reported, either in taped recorded sounds (kappa=0.26 and coeffi‐ cient of reliability of less than 60%) [19, 20] and real-time auscultation (–0.02*<*kappa*<*0.77) [21]. Using digital stethoscopes, one study conducted in adult patients with CF assessed the interrater agreement between real-time manual annotation of respiratory sounds and automatic detection of respiratory sounds through a computerised system [22]. Poor to moderate

day [10].

measure more frequently reported [13].

40 Cystic Fibrosis in the Light of New Research

physiotherapy interventions.

age and severity of the disease [14-16].

*2.1.2. Respiratory sounds*

Mucus is transported from the bronchial airways towards the exterior by mucociliary clear‐ ance, spontaneous cough and through a range of airway clearance techniques, such as directed huffs and coughs. Subsequently, secretions are either expectorated or swallowed. During respiratory physiotherapy sessions, patients are encouraged to expectorate to a cup. Sputum volume or weight can then be used as an outcome measure for respiratory physiotherapy.

While sputum volume has shown to be difficult to determine with precision, sputum weight (either dry or wet) has shown to be more accurate [29]. In a recent Cochrane review on the effectiveness of respiratory physiotherapy interventions in respiratory diseases, five of the eight studies used sputum weight as an outcome measure, out of which four were conducted in patients with CF [7]. Mortensen et al. (1991) found that patients expectorated 8–8.6g of sputum weight after airway clearance techniques and 0 g during a control day (range 0–2.1 g) [30]. In Pfeger et al. (1992), the mean weight of expectorated mucus with spontaneous cough was approximately 17g whereas with airway clearance techniques ranged from 34 to 45g (n=14) [9]. Although no mean and standard deviation was provided, Rossman et al. (1982) also found a higher volume of expectorated secretions during the different forms of chest physiotherapy compared to a control session (n=6) [31]. However, Jarad et al. (2010) found no significant differences in wet or dry sputum weight between a positive expiratory pressure (PEP) device (flutter) or a placebo [10].

Another study conducted by Osman et al. (2010) compared high-frequency chest wall oscillation with conventional airway clearance techniques in 29 patients with CF using wet sputum weight. Expectorated sputum in a single session was significantly different between the techniques (p<0.001; Cohen's dz=0.72) [32].

The problem with sputum weight remains that the lack of expectoration during respiratory physiotherapy techniques does not mean that airway clearance techniques are not effective. It is very common to expectorate after respiratory physiotherapy sessions or to swallow secretions even during the session, which means that sputum weight expectorated during a session may underestimate the effect of airway clearance techniques. On the other hand, expectoration may be contaminated with saliva [30]. Sputum weight can, therefore, be both over- and under-estimated and it is not surprising that contradictory findings have been found in the literature.

Although simple to collect and measure, sputum collected during and following the treatments is not considered a reliable measure of alveolar recruitment, airway clearance or even sensitive to small changes. In fact, this outcome measure has been frequently questioned [29, 30, 33–35] and is no longer considered as a valid outcome measure for studies in airway clearance techniques [36].

### *2.1.4. Measures of oxygenation*

Blood gas measurements are used to evaluate a person's lung function and acid–base balance, i.e., it measures the amount of oxygen and carbon dioxide that is in the blood and determines its levels of acidity (pH). The test results provide information about the partial pressure of oxygen (PaO2), partial pressure of carbon dioxide (PaCO2) and hydrogen ion activity (pH) in arterial blood, as well as indices of bicarbonate concentration, base excess and oxygen saturation. The analysis is performed on the blood collected from a person's artery and is therefore, an invasive and relatively complex procedure. Although it is possible to measure oxygenation more accurately, given the nature of the procedure, it is not practical to be performed on a routine basis to monitor respiratory physiotherapy treatments. Hence, transcutaneous pulse oximetry (reported as the percentage of saturation of arterial haemoglo‐ bin by oxygen [SaO2]) has become the most common method in clinical practice and research for measuring oxygenation [37].

Peripheral oxygen saturation is commonly used as it is simple to perform via pulse oximeters and finger probes. Pulse oximeters monitor the saturation of haemoglobin with oxygen (i.e., oxyhaemoglobin). This is possible because blood changes its colour as haemoglobin absorbs various amounts of light depending on its saturation with oxygen. Oxyhaemoglobin absorbs greater amounts of infrared light and does not absorb much red light, but as the haemoglobin oxygen saturation drops, more and more red light is absorbed and the blood becomes darker. Hence, pulse oximeters emit two wavelengths of light, red at 660 nm and near-infrared at 940 nm from the finger probe [38]. The pulse oximeter directly senses the absorption of light and translates it through complex signal processing to a function of the arterial oxygen saturation. A microprocessor integrates the data and, through an elaborate calibration algorithm based on human volunteer data, the oxygen saturation can be estimated. This measure allows constant monitoring of heart rate as well, being ideal to monitor patient's safety during interventions.

Nonetheless, this measure can provide unreliable readings due to several factors such as haemoglobin level, arterial blood flow to the vascular bed, oximetry sensor location and temperature, fluorescent or direct sunlight, jaundice, discoloration of the nail bed, nail polish, bruising under the nail, motion artefacts, intravascular dyes and skin pigmentation. All these are discussed in detail elsewhere [38].

In two Cochrane reviews about respiratory physiotherapy techniques, insufficient evidence was found regarding peripheral oxygen saturation as an outcome measure [7,12]. Osman et al. (2010) compared high-frequency chest wall oscillation with conventional airway clearance techniques in 29 patients with CF. It was found that, compared to the baseline, non-significant changes in peripheral oxygen saturation were observed during or after any of the treatments applied (Cohen's dz from 0 to 0.28) [32]. However, in a recent study, the addition of noninvasive ventilation to chest physiotherapy resulted in a significant reduction in the proportion of treatment time with peripheral oxygen saturation below 90% (p<0.001) [39].

Controversial evidence exists on the potential of peripheral oxygen saturation as an adequate measure to assess the effectiveness of respiratory physiotherapy. Hence, this measure seems ideal to monitor patients' safety but it may not present the required specificity and/or sensi‐ tivity to assess changes caused by respiratory physiotherapy interventions.

### *2.1.5. Dyspnoea*

The problem with sputum weight remains that the lack of expectoration during respiratory physiotherapy techniques does not mean that airway clearance techniques are not effective. It is very common to expectorate after respiratory physiotherapy sessions or to swallow secretions even during the session, which means that sputum weight expectorated during a session may underestimate the effect of airway clearance techniques. On the other hand, expectoration may be contaminated with saliva [30]. Sputum weight can, therefore, be both over- and under-estimated and it is not surprising that contradictory findings have been found

Although simple to collect and measure, sputum collected during and following the treatments is not considered a reliable measure of alveolar recruitment, airway clearance or even sensitive to small changes. In fact, this outcome measure has been frequently questioned [29, 30, 33–35] and is no longer considered as a valid outcome measure for studies in airway clearance

Blood gas measurements are used to evaluate a person's lung function and acid–base balance, i.e., it measures the amount of oxygen and carbon dioxide that is in the blood and determines its levels of acidity (pH). The test results provide information about the partial pressure of oxygen (PaO2), partial pressure of carbon dioxide (PaCO2) and hydrogen ion activity (pH) in arterial blood, as well as indices of bicarbonate concentration, base excess and oxygen saturation. The analysis is performed on the blood collected from a person's artery and is therefore, an invasive and relatively complex procedure. Although it is possible to measure oxygenation more accurately, given the nature of the procedure, it is not practical to be performed on a routine basis to monitor respiratory physiotherapy treatments. Hence, transcutaneous pulse oximetry (reported as the percentage of saturation of arterial haemoglo‐ bin by oxygen [SaO2]) has become the most common method in clinical practice and research

Peripheral oxygen saturation is commonly used as it is simple to perform via pulse oximeters and finger probes. Pulse oximeters monitor the saturation of haemoglobin with oxygen (i.e., oxyhaemoglobin). This is possible because blood changes its colour as haemoglobin absorbs various amounts of light depending on its saturation with oxygen. Oxyhaemoglobin absorbs greater amounts of infrared light and does not absorb much red light, but as the haemoglobin oxygen saturation drops, more and more red light is absorbed and the blood becomes darker. Hence, pulse oximeters emit two wavelengths of light, red at 660 nm and near-infrared at 940 nm from the finger probe [38]. The pulse oximeter directly senses the absorption of light and translates it through complex signal processing to a function of the arterial oxygen saturation. A microprocessor integrates the data and, through an elaborate calibration algorithm based on human volunteer data, the oxygen saturation can be estimated. This measure allows constant monitoring of heart rate as well, being ideal to monitor patient's safety during

Nonetheless, this measure can provide unreliable readings due to several factors such as haemoglobin level, arterial blood flow to the vascular bed, oximetry sensor location and

in the literature.

techniques [36].

*2.1.4. Measures of oxygenation*

42 Cystic Fibrosis in the Light of New Research

for measuring oxygenation [37].

interventions.

Dyspnoea is defined as "a subjective experience of breathing discomfort that consists of qualitatively distinct sensations that vary in intensity" [40]. In a study conducted with 123 patients with CF, dyspnoea was reported to be present in 74% of the patients [41] and therefore, it is an important problem to consider in this population. As a subjective experience, adequate assessment depends on patient self-report. Depending on the circumstances in which breath‐ ing discomfort occurs and the history of similar sensations, dyspnoea may be perceived as a threat associated with anxiety, fear or depression, and it may be viewed as a sign of disease. This symptom has multidimensional aspects involving physiological, psychological, social and environmental factors that result in a behavioural response. The assessment of multidi‐ mensional aspects of dyspnoea has assumed significance in recent years [40].

Dyspnoea has been assessed using scales and questionnaires, however, important differences exists between these instruments, particularly in what they measure (e.g., one instrument may ask what breathing feels like, whereas another may ask how distressing it is or how it impacts on patient's performance or quality of life), in the rating task (i.e., what patients are instructed to rate), and in whether measurements are performed on real time or involve the recall of a specific episode. These differences make comparisons across studies difficult.

In one study exploring the effectiveness of non-invasive ventilation during chest physiother‐ apy [39], it was found that dyspnoea scores, assessed with the Borg scale, increased in slightly lesser extent following chest physiotherapy assisted with non-invasive ventilation than chest physiotherapy alone (2.26±1.96 vs. 2.69±1.82, p=0.02, Cohen's dz=0.23). Marques (2008), however, found that dyspnoea assessed with the modified Borg scale was not significantly different after one session of airway clearance techniques (Cohen's dz=0.06) [42]. Enright et al. (2004) investigated the effect of an 8-week inspiratory muscle training (n=19) and also did not find significant differences is dyspnoea, using the dyspnoea domain of the Chronic Respiratory Disease Questionnaire (Cohen's dz 0.07 to 0.37) [43].

A combination of unidimensional and multidimensional tools will probably be the best approach for clinical assessment. Measurement of dyspnoea has to be seen in context, taking into consideration the patient's history, physical examination and diagnostic tests [44]. As dyspnoea is a symptom perceived by patients, it has the potential to contribute for early management of the disease, for adjusting the respiratory physiotherapy interventions and, thus, for improving outcomes in patients.

### *2.1.6. Chest radiography*

Several imaging techniques are available to diagnose and monitor a respiratory condition. Within the respiratory field, chest radiography is the most commonly used. Although relatively simple to perform, and in itself being relevant and reliable, the measure presents several important limitations to be used as an outcome measure for respiratory physiotherapy. The interpretation of chest radiography imaging is somewhat complex and presents high levels of inter-observer subjectivity [45], the reports detailing either the presence or the absence of any abnormalities are not commonly available immediately after the exam, its portability is limited to places where health and safety radiation protection standards can be ensured, as considerable doses of radiation are involved and is difficult to perform in non-collaborative populations such as children or people with dementia. These factors prevent the use of this measure to monitor respiratory physiotherapy patients with the required frequency.

In two long-term studies (n=36 and n=32), conducted for approximately one year comparing different respiratory physiotherapy airways clearance techniques in patients with CF, chest radiography was used as an outcome measure. However, no significant differences were shown between interventions [46, 47]. In the most recent Cochrane reviews [11, 12] about physiotherapy in CF, chest radiography is no longer reported as an outcome measure in the included studies.

It therefore, seems that chest radiography is ideal for contributing to diagnosis and providing a measure of improvement or deterioration over time but might not be the most appropriate to be used as an outcome measure for respiratory physiotherapy.

### *2.1.7. Exercise capacity*

The assessment of exercise capacity, expressed as the maximal workload achievable or the peak oxygen consumption on a progressive maximal test, has been used to measure patients' functional capacity and limitation, facilitates a safe exercise prescription and identifies changes in patients' performance as a result of an intervention [48, 49]. There are two types of clinically applicable tests to assess exercise capacity in patients with CF: laboratory-based tests and field tests [50, 51].

Laboratory-based tests, such as the cardiopulmonary exercise testing (CPET) [48, 49], are the gold standard for evaluating the causes of exercise intolerance in patients with respiratory diseases, as they provide a comprehensive assessment of the physiological responses to exercise involving the respiratory, cardiovascular and musculoskeletal systems [52]. CPET consists of a progressive exercise performed at increasing levels of intensity either on a treadmill or cycle ergometer, with simultaneous monitoring of breath-by-breath measures of airflow (e.g., VO2, VCO2) along with heart rate, peripheral oxygen saturation and exerciserelated symptom responses (e.g. dyspnoea, leg discomfort) [49]. There are a number of protocols that can be used to perform a CPET in patients with CF, but the most commonly reported is the Godfrey protocol [53]. The CPET has been shown to be reproducible in young and adult patients with CF [54, 55]. In patients with CF (n=23) enrolled in a strength and endurance training for 6 months, significant differences were observed in VO2 peak between 3 and 6 months (Cohen's dz=2.01) and between 18 and 24 months (Cohen's dz=2.22) [56]. Despite its sensitivity to change, the application of CPET is still limited in clinical settings [57]. This may be attributable to the need for expensive equipment and technical expertise. For this reason, field tests are more commonly used by physiotherapists to assess changes in exercise capacity [58].

find significant differences is dyspnoea, using the dyspnoea domain of the Chronic Respiratory

A combination of unidimensional and multidimensional tools will probably be the best approach for clinical assessment. Measurement of dyspnoea has to be seen in context, taking into consideration the patient's history, physical examination and diagnostic tests [44]. As dyspnoea is a symptom perceived by patients, it has the potential to contribute for early management of the disease, for adjusting the respiratory physiotherapy interventions and,

Several imaging techniques are available to diagnose and monitor a respiratory condition. Within the respiratory field, chest radiography is the most commonly used. Although relatively simple to perform, and in itself being relevant and reliable, the measure presents several important limitations to be used as an outcome measure for respiratory physiotherapy. The interpretation of chest radiography imaging is somewhat complex and presents high levels of inter-observer subjectivity [45], the reports detailing either the presence or the absence of any abnormalities are not commonly available immediately after the exam, its portability is limited to places where health and safety radiation protection standards can be ensured, as considerable doses of radiation are involved and is difficult to perform in non-collaborative populations such as children or people with dementia. These factors prevent the use of this

measure to monitor respiratory physiotherapy patients with the required frequency.

In two long-term studies (n=36 and n=32), conducted for approximately one year comparing different respiratory physiotherapy airways clearance techniques in patients with CF, chest radiography was used as an outcome measure. However, no significant differences were shown between interventions [46, 47]. In the most recent Cochrane reviews [11, 12] about physiotherapy in CF, chest radiography is no longer reported as an outcome measure in the

It therefore, seems that chest radiography is ideal for contributing to diagnosis and providing a measure of improvement or deterioration over time but might not be the most appropriate

The assessment of exercise capacity, expressed as the maximal workload achievable or the peak oxygen consumption on a progressive maximal test, has been used to measure patients' functional capacity and limitation, facilitates a safe exercise prescription and identifies changes in patients' performance as a result of an intervention [48, 49]. There are two types of clinically applicable tests to assess exercise capacity in patients with CF: laboratory-based tests and field

Laboratory-based tests, such as the cardiopulmonary exercise testing (CPET) [48, 49], are the gold standard for evaluating the causes of exercise intolerance in patients with respiratory

to be used as an outcome measure for respiratory physiotherapy.

Disease Questionnaire (Cohen's dz 0.07 to 0.37) [43].

thus, for improving outcomes in patients.

*2.1.6. Chest radiography*

44 Cystic Fibrosis in the Light of New Research

included studies.

*2.1.7. Exercise capacity*

tests [50, 51].

Field tests are simple clinical exercise tests that do not require expensive equipment [59] and, thus, are suitable to be used at different settings of respiratory physiotherapy practice, such as hospitals, private practices, at home or in community environments. In patients with CF, the field tests most frequently used are the 6-minute walk test (6MWT), the 3-min step test (3MST) and the modified shuttle test (MST) [51, 58]. The 6MWT measures the maximal distance that a patient can walk in 6 minutes over a marked course (usually a corridor) following a standardised protocol [60]. The 6MWT has been recommended for patients with CF [60] and is valid and reliable for assessing exercise capacity in children [61, 62] and adults [63] with CF. The product of the distance walked during a 6MWT and body mass have been reported to correlate with aerobic capacity (VO2max) in children with CF [64], thus supporting the use of this test in clinical settings. Furthermore, the 6MWT was found to be valuable for identifying patients who might experience oxygen desaturation and physical impairment in daily activities [65]. For this reason, the 6MWT has been considered more reflective of activities of daily living than other walk tests [59]. Numerous reference equations of the 6MWT are available for children [66–69] and adults [70].

The 3MST is a simple test that requires the patient to step at a rate of 30 steps per min on and off a step with 15 cm (6 inch) of height, during 3 minutes. This test was developed for children with CF by Balfour-Lynn et al. (1998) [71] based on the original Master two-step exercise test (1929) [72]. One major advantage of the 3MST is that it does not depend on patient's motivation, since the cadence of steps is fixed and determined by a metronome. One limitation concerns to the fact that, as step height and rate are kept constant, the workload varies between patients depending on their height and lower limb length [51, 58]. The 3MST has been found to be repeatable [71] and sensitive to changes in pulmonary function and peripheral oxygen saturation in children with CF following a course of intravenous antibiotics for acute respira‐ tory exacerbations [73]. This is an important finding since even physiotherapists without access to a formal exercise laboratory may evaluate the response to an intervention using a simple field test. When compared to the 6MWT, the 3MST elicited a significantly greater change in heart rate and breathlessness in children with CF, along with a comparable [71] or higher [74] fall in oxygen saturation, suggesting that this test is a more vigorous exercise challenge. Nevertheless, Narang et al. (2003) showed that, in children with mild disease, important information such as the exercise-related desaturation that occurs at higher exercise levels may be missed with this test [75].

While the 6MWT and 3MST may be submaximal tests for patients with mild-to-moderate CF, the MST allows maximal exercise capacity to be reached [58]. The MST [76] consists of a validated adaptation of the Incremental Shuttle Walk Test [77], specifically developed for adults with CF. This adaptation allowed individuals to walk and run at increasing speeds over 15 levels with a maximum speed of 10.2 km/h [76]. Bradley et al. (1999, 2000) studied the validity, reliability and sensitivity of the MST in adults with CF [76, 78]. When compared to a laboratory-based treadmill test, they showed that 90% of the variation of directly measured VO2 peak was explained by the variation in MST performance [76]. Moreover, the MST was effective in evoking a symptom-limited exercise response (i.e., peak heart rate and peak rating of perceived dyspnoea) similar to what is found in treadmill tests [76]. Thus, it may be a valid alternative when laboratory-based exercise testing cannot be performed. The validity of the MST has also been assessed in children with CF (correlation of the MST distance with VO2peak, r=0.663 p<0.01) [79]. In adults with CF, the MST was found to be reliable (correlation between trials for distance completed and symptoms reported, r=0.99, p<0.01) and repeatable (coeffi‐ cients of repeatability: distance completed, four shuttles; peak heart rate, 6 beats/min; peak oxygen saturation, 4%) [78]. The sensitivity of the test was assessed by measuring the change in MST performance after 2 weeks of antibiotic therapy in patients admitted to hospital with acute respiratory exacerbations. An effect size of 1.18 was achieved, suggesting that this is a highly sensitive measure [78]. This test has also been used to assess changes in exercise capacity after 2 months of exercise training and significant differences were found (median number of shuttles: from 100 (range 21–150) to 105 (44–150)) [80]. Despite the good measurement properties, it was recently argued that the 15-level MST developed by Bradley et al. (1999) [76] still remained sub-maximal for some patients with CF and, thus, levels were extended to 25 in order to create a truly maximal test [81].

There is no "best" exercise test. The selection of the test will depend upon the aspect of exercise capacity of interest, availability of resources (i.e., time to perform the test, staff and equipment) and patient's characteristics (e.g., age, disease stage) [50, 51]. For example, if the aetiology of a patient's reduced exercise capacity is of interest, a laboratory-based test would be more appropriate, while a field test could be used in large population studies. In addition, a young child (< 7 years old) may not be able to cooperate sufficiently for a formal laboratory exercise testing. Physiotherapists have the important role of selecting the best exercise test for a specific individual according to the specific question being asked and the specificities of each test.

### *2.1.8. Health-related quality of life*

Health-related quality of life (HRQoL) has been extensively studied in patients with CF, especially in the last decades due to the improvement in patients' life expectancy [82]. Several instruments have been used to assess HRQoL in CF, either generic [83-85] or disease-specific [85-87]. One disadvantage of generic questionnaires is that they lack the sensitivity needed to assess areas of functioning that are critically important for patients with CF [88]. Diseasespecific questionnaires were developed in an attempt to better understand CF specific issues in clinical practice and research.

The most commonly used disease-specific questionnaires include the Cystic Fibrosis Ques‐ tionnaire (CFQ) and the Cystic Fibrosis Quality of Life questionnaire. The CFQ was developed for assessing HRQoL of patients with CF and encompasses general domains of HRQoL (physical functioning, role functioning, vitality, health perceptions, emotional functioning and social functioning), as well as domains specific to CF (body image, eating disturbances, treatment burden, and respiratory and digestive symptoms) [89]. Three versions of the instrument have been developed: one for adolescents aged 14 years or older and adults (CFQ-Teen/Adult), and two for assessing children 6–13 years old, one to be completed by the child (CFQ-Child) and the other by parents (CFQ-Parent) [89, 90]. Each version takes around 15 minutes to fill in. The different versions of the CFQ questionnaire have shown good psycho‐ metric properties (validity [88-90], reliability [89, 90], internal consistency [88-90] and respon‐ siveness [90]). Hebestreit et al. (2010) found that patients with CF (n=23) enrolling in strength and endurance training for 6 months improved their subjective health perception (CFQ domain) significantly between months 3 and 6 (Cohen's dz=0.92) [56]. Schmidt et al. (2011) evaluated a 12-week individually tailored unsupervised aerobic exercise programme in 14 patients with CF and also found significant differences in emotional functioning and treatment burden domains of the CFQ (Cohen's dz of 0.29 and 1.03) [91]. Thus, this questionnaire may be used in clinical practice to assess HRQoL of patients with CF, document the progression of disease and explore the effects of respiratory physiotherapy interventions.

The Cystic Fibrosis Quality of Life questionnaire was specifically developed for adolescents (14 years old or older) and adults, presenting good validity (concurrent and discriminate), internal consistency, test–retest reliability results and responsiveness [92]. One advantage of this instrument is the inclusion of domains concerning wider impacts of the disease on patient's lives (e.g., interpersonal relationships, career issues and future concerns), which are not found in other CF questionnaires. Though, it is not adaptable to children.

In sum, disease-specific instruments present good psychometric properties and seem appro‐ priate for several different applications in patients with CF, such as to identify problems and intervene on an individual basis, compare different CF groups, or detect changes in patients' HRQoL as a result of disease progression or interventions.

### **2.2. Emerging outcome measures**

heart rate and breathlessness in children with CF, along with a comparable [71] or higher [74] fall in oxygen saturation, suggesting that this test is a more vigorous exercise challenge. Nevertheless, Narang et al. (2003) showed that, in children with mild disease, important information such as the exercise-related desaturation that occurs at higher exercise levels may

While the 6MWT and 3MST may be submaximal tests for patients with mild-to-moderate CF, the MST allows maximal exercise capacity to be reached [58]. The MST [76] consists of a validated adaptation of the Incremental Shuttle Walk Test [77], specifically developed for adults with CF. This adaptation allowed individuals to walk and run at increasing speeds over 15 levels with a maximum speed of 10.2 km/h [76]. Bradley et al. (1999, 2000) studied the validity, reliability and sensitivity of the MST in adults with CF [76, 78]. When compared to a laboratory-based treadmill test, they showed that 90% of the variation of directly measured VO2 peak was explained by the variation in MST performance [76]. Moreover, the MST was effective in evoking a symptom-limited exercise response (i.e., peak heart rate and peak rating of perceived dyspnoea) similar to what is found in treadmill tests [76]. Thus, it may be a valid alternative when laboratory-based exercise testing cannot be performed. The validity of the MST has also been assessed in children with CF (correlation of the MST distance with VO2peak, r=0.663 p<0.01) [79]. In adults with CF, the MST was found to be reliable (correlation between trials for distance completed and symptoms reported, r=0.99, p<0.01) and repeatable (coeffi‐ cients of repeatability: distance completed, four shuttles; peak heart rate, 6 beats/min; peak oxygen saturation, 4%) [78]. The sensitivity of the test was assessed by measuring the change in MST performance after 2 weeks of antibiotic therapy in patients admitted to hospital with acute respiratory exacerbations. An effect size of 1.18 was achieved, suggesting that this is a highly sensitive measure [78]. This test has also been used to assess changes in exercise capacity after 2 months of exercise training and significant differences were found (median number of shuttles: from 100 (range 21–150) to 105 (44–150)) [80]. Despite the good measurement properties, it was recently argued that the 15-level MST developed by Bradley et al. (1999) [76] still remained sub-maximal for some patients with CF and, thus, levels were extended to 25 in

There is no "best" exercise test. The selection of the test will depend upon the aspect of exercise capacity of interest, availability of resources (i.e., time to perform the test, staff and equipment) and patient's characteristics (e.g., age, disease stage) [50, 51]. For example, if the aetiology of a patient's reduced exercise capacity is of interest, a laboratory-based test would be more appropriate, while a field test could be used in large population studies. In addition, a young child (< 7 years old) may not be able to cooperate sufficiently for a formal laboratory exercise testing. Physiotherapists have the important role of selecting the best exercise test for a specific individual according to the specific question being asked and the specificities of each test.

Health-related quality of life (HRQoL) has been extensively studied in patients with CF, especially in the last decades due to the improvement in patients' life expectancy [82]. Several instruments have been used to assess HRQoL in CF, either generic [83-85] or disease-specific

be missed with this test [75].

46 Cystic Fibrosis in the Light of New Research

order to create a truly maximal test [81].

*2.1.8. Health-related quality of life*

### *2.2.1. Computerised respiratory sounds*

Computerised auscultation consists of recording patient's respiratory sounds with a digital stethoscope and automatically analysing and classifying them based on specific signal characteristics [93]. Computerised auscultation allows to objectively detect, characterise and identify both normal and adventitious (i.e., crackles and wheezes) respiratory sounds within the breathing cycle [93].

To assess computerised respiratory sounds, the patient is positioned in the sitting or supine position (for long-term assessments) and instructed to breath from his/her mouth. Successive or simultaneous recordings are taken from the trachea plus six chest locations (i.e., right and left: anterior, lateral and posterior positions), using a single or a multichannel equipment (Figures 1 and 2) [94].

**Figure 1.** Single-channel equipment for computerised respiratory sound analysis.

**Figure 2.** Multi-channel equipment for computerised respiratory sound analysis.

Seven to ten respiratory cycles at tidal breathing or at a flow of 1 to 1.5 L should be recorded to ensure the stability of sound and quality of the analysis [94]. Then, the sound is filtered using a combination of low-pass and high-pass filters in cascade to reduce sound artefacts (e.g., muscle, heart sounds and movement frequencies) and specialised algorithms for respiratory sound detection and analysis are implemented [94].

A wide range of analysis methods exists and new ones are continuously being proposed. However, the ones with more evidence are time-frequency analysis, fast Fourier and wavelet transformations, neural networks, periodogram and auto- regressive models [93, 95-97]. Specifically for patients with CF, a time-frequency wheeze detector has already been validated, demonstrating high levels of sensitivity (77.2%) and specificity (98.4%) [98]. Nevertheless, algorithms for the detection of crackles in patients with CF were not found and, thus, further studies are needed.

To assess computerised respiratory sounds, the patient is positioned in the sitting or supine position (for long-term assessments) and instructed to breath from his/her mouth. Successive or simultaneous recordings are taken from the trachea plus six chest locations (i.e., right and left: anterior, lateral and posterior positions), using a single or a multichannel equipment

**Figure 1.** Single-channel equipment for computerised respiratory sound analysis.

**Figure 2.** Multi-channel equipment for computerised respiratory sound analysis.

sound detection and analysis are implemented [94].

Seven to ten respiratory cycles at tidal breathing or at a flow of 1 to 1.5 L should be recorded to ensure the stability of sound and quality of the analysis [94]. Then, the sound is filtered using a combination of low-pass and high-pass filters in cascade to reduce sound artefacts (e.g., muscle, heart sounds and movement frequencies) and specialised algorithms for respiratory

(Figures 1 and 2) [94].

48 Cystic Fibrosis in the Light of New Research

The reliability of computerised respiratory sounds has been mainly studied in adults with CF [99] and children with asthma, pneumonia [100] and bronchiolitis [23]. In adults with CF, one study assessed the test–retest reliability of two main parameters of crackles (i.e., the initial defection width and the two cycle duration). The intra-subject reliability of crackle parameters was found to be 'good' to 'excellent' (0.76<intraclass correlation coefficients<0.94) with no systematic bias. The smallest real difference found for the initial defection width was between 0.30 and 0.66 ms, and for the two-cycle duration between 1.57 and 2.42 ms. The reliability of wheezes has only been assessed in children with bronchiolitis, presenting moderate to good agreement (0.77*<*kappa*<*0.79) [99]. The reliability of computerised respiratory sounds to assess wheezing in adults and its overall reliability in children with CF is still unexplored.

Due to its simplicity and increasing reliability, computerised respiratory sounds have been used as outcome measures for pharmacological and respiratory physiotherapy interventions in children and adult patients with several respiratory diseases, including CF [93]. Marques et al. (2008) investigated the effect of one single session of respiratory physiotherapy using the ACBT in 17 adult patients with CF [42]. The initial defection width and two cycle duration of crackles were analysed. Considering the mean of all participants, no significant differences were found in the analysed parameters. However, when the individual data of each participant was considered, significant changes were observed in the initial defection width of 9 patients (53%) and in the two-cycle duration of 10 participants (59%).

Considering the detection of wheezes and their analysis, one study assessing the sensitivity of computerised respiratory sounds to defect bronchial hyperactivity in 23 children with CF, following an induced methacholine challenge, found a sensitivity of 50% and a specificity of 100% [101]. Studies in patients with lower respiratory tract infection, a common form of exacerbation in CF [102], have shown that computerised respiratory sounds are effective in detecting changes following pharmacological and respiratory physiotherapy interventions [93]. Small to large effects were found in the number of crackles (Cohen's dz: 0.14 to 1.65), peak frequency (Cohen's dz: 0.11 to 0.47), two-cycle duration (Cohen's dz: 0.83 to 0.85) and initial/ largest deflection width (Cohen's dz: 0.38 to 1.25). Better results have been found in the detection and characterisation of wheezes, with medium to large effect sizes for the number and occupation rate of wheezes (Cohen's dz: 0.34 to 4.30) [93].

One of the disadvantages of computerised respiratory sounds concerns the complexity and costs associated with the equipment and subsequent sound analysis. Nevertheless, an emerging body of health and engineer researchers have been gathering efforts to produce more simple and efficient hardware/software that can be used in clinical practice [101, 103].

Despite the scarce evidence of computerised respiratory sounds in CF, it seems that this measure might offer potential to assess the short- and long-term effects of respiratory physi‐ otherapy interventions in these patients. However, more research is still required to determine the parameters of computerised respiratory sounds (i.e., number, frequency, duration) that are more sensitive to change after an intervention and to establish the reference values that will allow physiotherapists to interpret with confidence the results obtained from the computerised auscultation.
