We demonstrated that dyspnea and perceived leg exertion ratings obtained using the Borg CR-10 scale closely tracked analogous ratings obtained using our Dalhousie scales in adult, rising more steeply in patients with pulmonary disease than in healthy controls but exponentially in both groups during incremental exercise. The coherence between measurements and similar trajectories obtained with either scale as exercise intensity grew implies that the Dalhousie pictorial scales can be used in adult populations with equal confidence and are arguably superior to the Borg scale in adults who have difficulty understanding the latter, e.g., patient 31.
We created the Dalhousie scales primarily as a tool to measure degree of dyspnea including the overall sense of breathing effort or discomfort and secondly perceived exertion with specific focus on the legs—the principal muscle groups involved in most ergometry systems. In his thesis, Borg stated the concept of overall perceived exertion could be regarded as a gestalt made up of perceptions from several important cues [3]. He later developed a category (C) scale with ratio (R) properties by adding numbers along his line at specific distances and adding descriptive adjectives to this non-linear numeric scale [4]. We submit that our pictures serve as the equivalent of, or replace, the descriptive categories for subjects who cannot grasp these sometimes-subtle, quantitative semantics. We further speculate that they embody the “qualitatively distinct sensations” [1] and the gestalt made up of perceptions from several important cues [3]. The larger variability in ratings obtained by the two scales seen during heavy exercise in canonical plots could simply be due to the additional gradations of sensation available on a CR-10 scale, but these extras matter little if a subject cannot distinguish “very strong” (7) from “extremely strong” (10). Killian argued that the Borg CR-10 scale adheres to principles of an absolute scale with ratio properties [14]. The Dalhousie Dyspnea and Perceived Exertions Scales function at least as interval scales with perceptual anchors (“from nothing at all, to the hardest breathing imaginable”), and our findings confirm their ratio properties as well. The novel finding of this study was that ratings of dyspnea and leg exertion derived from either Borg or Dalhousie pictorial scale conformed better to a trajectory that includes a delay term than to the expected simple power function.
Borg found it necessary to include one and sometimes two extra basic constants in his original equation (for the absolute threshold, or describing basic “noise”) [3, 4]. Borg and Kaijser ignored this delay term (eqtn.1 in their paper) [2], whereas Killian et al. noted thresholds in relative power output below which leg effort and dyspnea did not change appreciably from resting level [9]. This delay amounted to somewhere between 20 and 65 % (longest with Borg CR-10) of healthy subjects’ Wmax and accounts for some of the variability in dyspnea and perceived exertion ratings. It was markedly different between our healthy subjects and those with respiratory disease (Table 3) but we did not seek matched controls for purposes of this study since our aims were to compare scales, not healthy controls vs patients. In general, trajectories rose quite steeply in pulmonary patients from the outset, markedly different than growth function(s) in healthy controls. A model that accounts for variable delays and variability in rate of rise will be intuitively superior to a simple power function. The delay model that includes a linear and a quadratic term had the lowest RMSE and AIC, though coefficients for the quadratic terms were small. While ratio scales in general function well empirically, alternative modeling of the sensory–perceptual function, specifically incorporating a delay term, can reduce inter-subject variability in ratings.
Dyspnea is a complex sensation that relies on mechanical, chemical, and cortical inputs and incorporates numerous sensations, with a common theme of an imbalance or inappropriateness between neural output to, and mechanical output from, the respiratory pump [15–18]. Clearly, one becomes more aware of one’s breathing and the effort it requires as exercise becomes progressively harder or prolonged, but it does not become unpleasant or perceived as dyspnea so long as afferent signals from the respiratory apparatus remain in harmony with the efferent drive to the muscle(s) of breathing. This balance becomes upset by dynamic hyperinflation with exercise onset in patients with COPD or asthma in order to compensate for flow limitation [19–21] by low lung compliance coupled with flow limitation in ILD [22], whereas increased neural drive rather than mechanical ventilatory displacement appears to play a greater role in obesity [23]. Chemical stimuli arising from blood-gas abnormalities, particularly hypercapnia, both magnify and alter perceived dyspnea (and consequent affective response) [24], a common scenario in patients with respiratory disease. Patients with different disease processes tend to describe their sensation of dyspnea with typical descriptors [25–28]. Our pediatric focus group interviews during scale development differentiated the separate and distinguishable concepts of shortness of breath from increased breathing effort, prompting creation of the “breathing effort” sub-scale [2, 11]. There is now general agreement based on descriptor studies that perceptions of “work and effort,” “air hunger or unsatisfied breaths,” and “chest tightness” are separable qualities of dyspnea, and recent work suggests they are likely distinct [29]. In our study, all sub-scale (breathing effort, throat, or chest) ratings rose more or less in unison. We instructed subjects at the outset that their difficulty breathing might be perceived in the chest or throat, but we did not instruct them to choose one sensation or location over the other. “Descriptor” literature was a contemporaneous phenomenon during our scale development (1998) and work that elucidated mechanisms explaining separate but distinguishable types of dyspnea was only beginning when we conducted exercise tests. We demonstrated that dyspnea rose more steeply and to a greater extent in adults with pulmonary disease using our pictorial scales having receiving generic instructions. Future studies in this population can examine whether re-phrasing instructions to patients prompts them to select one scale over another to describe their sensory-perceptual experience. For example, it seems intuitive that an asthmatic experiencing bronchoconstriction may select the chest constriction scale, but will a COPD patient be drawn toward the throat narrowing scale to convey a sense of or unsatisfied inhalation (a common descriptor offered by adolescents with exercise-induced glottic obstruction) while both gravitate toward the breathing effort scale to describe exercise hyperpnea?
This issue may pose a limitation on applicability of the Dalhousie dyspnea and perceived exertion scales. One could argue that only the breathing effort and perceived leg exertion pictorial scales are all that are required, although one hopes that localization of the site of perceived difficulty might permit separation of patients with different disorders. No patients with cardiac disease were tested in our study, and utility of the Dalhousie scale in this population cannot be presumed. The pathophysiologic bases of dyspnea and perceived exertion scale in this population are arguably more complex [29, 30]. On the other hand, recent studies have considered the multidimensional aspect of dyspnea [30] that comprises three major aspects: a sensory-perceptual domain and a symptom impact both in physical and affective terms. Sensory-perceptual dimension includes ratings of dyspnea intensity and its quality, i.e., “what breathing feels like.” One might argue that there will be less precision differentiating perceived intensity levels denoted by solely by pictures without accompanying numbers. The pictorial nature of the Dalhousie scales implies de facto they are imbued with qualitative as well as quantitative information. It has been recently recognized that dyspnea, like pain, is a complex symptom with both sensory and affective domains, the latter thought to be processed in central limbic neural structures. Functional MRI has shown regions in the CNS where unpleasantness of perceived dyspnea is received, perceived, and processed [31, 32]. Furthermore, even verbal cues in the absence of any breathing load can evoke affective responses, which themselves are influenced by fatigue, anxiety, and somatic hypervigilance [32]. It is not unreasonable to presume that visual cues could do likewise, although we have no data for or against this notion. The Dalhousie pictorial scales were developed to quantify and discriminate qualitative aspects of the sensory perception of effort and difficulty associated with breathing in health and disease. Although the ratings are modulated by affective state in individual subjects, the pictorial scales were not developed to specifically assess the affective domain of dyspnea; further research would be necessary to develop pictorial scales to quantify the affective dimension of dyspnea. Finally, the perceived exertion scale is predicated on large-muscle, leg exercise—either treadmill or cycle ergometry. Arm cranking exercise is seldom performed in children (unless paraplegic), and thus, we convened no focus group to create the analogous arm-scale for this modality. Such would have to be created de novo if one wished to study perceived exertion with arm exercise. Such multiplicity of scales may render the entire pictorial concept unwieldy, but we believe this limitation will be more than compensated by the richness of information they yield.
