Uncovering the mechanisms of exertional dyspnoea in combined pulmonary fibrosis and emphysema

Ventilatory inefficiency and exertional dyspnoea in CPFE

We confirmed our main original hypothesis that poor ventilatory efficiency (high V′_E/V′_CO2) [32] would emerge as a key functional correlate of exertional breathlessness in CPFE when the confounding effects of hypoxaemia were controlled for [16, 30]. In the present study, a high exertional dyspnoea burden was defined considering its importance relative to leg effort in a given subject and its severity comparative to other patients facing a similar challenge (work rate) [27], i.e. we obtained an index of intra- and inter-subject dyspnoea severity [21]. Exertional dyspnoea characteristically increases when the mechanical output of the respiratory muscles becomes uncoupled from increases in neural respiratory drive [33]. This might happen secondary to excessive ventilatory stimulus as reflect by ventilatory–metabolic uncoupling (high V′_E/V′_CO2) and/or the development of critical constraints to V_T expansion (neuromechanical dissociation) [33]. Our observation that the more dyspnoeic CPFE patients (figure 2f) showed the highest V′_E/V′_CO2 (figure 2b) which coexisted with increased V_D/V_T (figure 2c) and lower P_cCO2 (figure 2d) despite preserved inspiratory reserves (figure 2e) provide support to the notion that heightened ventilatory stimuli (not hypoxaemia (figure 2a) or constrained lung mechanics) largely explained the unduly exertional dyspnoea reported by these patients.

In this context, there is well established evidence that increased dead space ventilation (high V_D/V_T) is a consequence of increased volume of conducting airways (due to dilation or longitudinal growth) [34] and more extensive areas of high ventilation/perfusion ratio in IPF [9], i.e. both the anatomical and physiological dead space are increased [35]. A high dead space requires an increase in V′_E to avoid carbon dioxide (CO₂) retention [36]: this is particularly critical under the stress of exercise when there is an increase in the CO₂ flow from the peripheral muscles to the lungs. The coexistence of a high V_D/V_T with a low P_aCO2 at rest (table 1) and P_cCO2 on exertion (figure 2d) implies that V′_E increased beyond what was required to overcome the “wasted” ventilation in the more dyspnoeic group. Remarkably, the lower resting P_CO2 in this group did not change appreciably with exercise, i.e. the overhead in V′_E was tightly controlled to maintain it close to its resting set-point (figure 2d). Hypocapnia in our patients may have reflected heightened afferent stimuli from multiple sources, e.g. central and peripheral chemoreceptors, sympathetic drive and ergoreceptors, among others [37]. Although it remains to be elucidated what are the precise mechanisms driving hyperventilation in the more dyspnoeic patients, enlarged V_D/V_T may increase the amplitude of intra- and inter-breath CO₂ oscillations thereby stimulating the central chemosensitive neurons [38], particularly in the presence of chronic hypoxaemia [39]. Future studies addressing chemoreceptor sensitivity with hypoxia/hypercapnia tests might provide important insights into this complex issue [40].

It is also noteworthy that V_T increased to a greater extent in the hypocapnic group (table 2). Contrary to patients with “pure” emphysema in whom V_T expansion is usually constrained at high operating lung volumes (leading to hypercapnia) [41], it seems apparent that the restrictive effects of fibrosis allowed greater room for V_T increase in the dyspnoeic patients with CPFE. As a consequence, V_T could readily overcome the enlarged V_D in order to keep P_aCO2 at the low level regulated at rest. These findings demonstrated that the operating characteristics of the mechanical plant (the ventilatory pump) have an important influence in the dynamic interactions between the central controller and the chemical/gas exchange plant in human disease [42].

The structural determinants of ventilatory inefficiency in CPFE

We found a noticeable cross-association between dyspnoea burden, high V_D/V_T and high V′_E/V′_CO2 with emphysema extent in CPFE. Thus, despite the growing evidence that emphysema extent does not affect mortality beyond the expected from the fibrosis burden [3, 6] the present results show that it does carry an important negative effect on morbidity. For instance, CPFE patients desaturate at similar extent than those with IPF (figure 1a) in spite of showing less extensive fibrosis (table 1). Interestingly, increased wasted ventilation was significantly related to the extent of structural abnormalities signalling for enlarged air spaces, such as emphysema and traction bronchiectasis (table 1). Influential microcomputed tomography studies showed that loss of terminal bronchiole usually precede centrilobular emphysema; centrilobular emphysema is then formed distal to surviving bronchioles, supported by extensive collateral ventilation. Functionally, therefore, they work as areas of high ventilation–perfusion [43]. In an autopsy based study of patients with CPFE, vascular rarefaction was observed near thick-walled cystic lesions representing areas of reticulation adjacent to emphysema [44]. As emphysema and fibrosis frequently coexist in a given lung segment, the traction effects of the latter may further enlarge the dead space effect of emphysematous areas [45]. Although this preserves lung volumes [3], it may have important deleterious effects on ventilation distribution. Thus, vascular destruction might be a consequence of both fibrotic and emphysematous processes in areas of admixed emphysema [45]. Collectively, these results seem to provide support to long-term conjectures that these abnormalities are relatively well ventilated but poorly (or not) perfused, i.e. they would work either as “dead space effect” or true dead space units [46]. A note of caution should be made regarding to V_D/V_T by the Bohr–Engoff formulation as hyperventilation may spuriously increase its value [10]. However, the noticeable upward displacement of V′_E/V′_CO2 as a function of P_cCO2 (figure 3) and the larger positive P_(c-ET)CO2 (table 2) provide supporting evidence that a high V_D/V_T was an important mechanism driving poorer ventilatory efficiency in the more dyspnoeic group with CPFE (figure 2) [31].

It also remains possible that part of the higher V_D/V_T in these patients was related to impaired perfusion of relatively preserved lung parenchyma (as has been described in pulmonary fibrosis [35] and COPD [47]) and/or a predominance of the pulmonary vascular phenotype of COPD. The presence of markers of “impaired oxygen delivery/utilisation” in this group (e.g. earlier estimated lactate threshold, lower peak oxygen pulse) (table 2) might reveal worsening pulmonary vasculopathy; however, these findings might merely reflect the consequences of severe physical inactivity experienced by the most dyspnoeic patients. It should be noted that we were unable to find significant associations between echocardiographic indicators of pulmonary hypertension and ventilatory inefficiency or V_D/V_T in the CPFE patients. The lower prevalence of patients with intermediate or high likelihood of pulmonary hypertension compared to previous studies [48] might reflect the lower summative extents of emphysema and PF in our sample. In fact, T_LCO in the range of 20–30% pred is commonly seen in those with higher likelihood of pulmonary hypertension [48]; in contrast, our patients typically showed values >30% pred (table 1). It could be argued that the currently recommended TRV thresholds have limited value in excluding pulmonary hypertension in PF [49]; however, we did not find a single patient with additional echocardiographic signs of pulmonary hypertension. In any case, the relevance of ventilatory inefficiency as a mechanism of exertional dyspnoea might further increase as CPFE progresses as pulmonary hypertension is characteristically associated with increased ventilatory stimulus and high V_D/V_T.

Clinical implications

In order to identify the CPFE patients at higher risk of disablement due to severe exertional dyspnoea, we provided evidence that the physician may rely on selected imaging (high burden of overall and admixed emphysema and traction bronchiectasis) and, in particular, CPET (high V′_E/V′_CO2) variables. Identifying areas of poor pulmonary perfusion relative to ventilation might also prove valuable to identify patients with more extensive wasted ventilation. Although alleviation of exertional hypoxaemia is likely useful to decrease the afferent ventilatory stimuli and improve central haemodynamics, our results indicate that it has a limited potential to improve dyspnoea and exercise tolerance in CPFE. Due to relatively preserved inspiratory reserves (table 2; figure 2e), inhaled bronchodilators are also unlikely to work at the same extent as in COPD. It follows that there is a sounder rationale for strategies to reduce “wasted” ventilation (e.g. surgical or bronchoscopic approaches to reduce localised hyperinflation and emphysema) and/or decrease non-hypoxic afferent stimuli (e.g. exercise training to lessen ergoreceptor activation and lactic acid production, sympathetic modulation) [31, 50]. If future studies show a role for pulmonary vasculopathy in increasing the “wasted” ventilation in non-emphysematous areas, vasodilators (particular if inhaled to minimise the risk of worsening ventilation–perfusion relationship) might prove beneficial. Acting on the central modulation of the symptom may also be helpful in selected patients, e.g. oral or inhaled opiates [31, 50].

Study limitations

Naturally, the present study has some limitations. Due to the complexities involved in a prospective clinical physiology study involving exercise tests a frail population, our sample was necessarily smaller than large imaging-based, retrospective studies [2, 3]. Haemodynamic measurements on exercise might have been useful to confidently rule out exercise-induced pulmonary hypertension, a confounding factor that certainly merits further investigations using invasive CPET. We recognise that by not matching IPF and CPFE groups by resting lung volumes an “advantage” was given to the latter group pertaining to lung mechanics. However, we contend that had we matched the groups by lung volumes (which are characteristically higher in CPFE) [1] IPF patients would be necessarily less hypoxaemic than their counterparts with CPFE. In fact, by showing that a subgroup of CPFE patients had worse dyspnoea than the IPF group despite higher static and operational lung volumes (table 1) on a background of similar hypoxaemia (figures 1a and 2a), our results provided novel insights into the relevance of ventilatory inefficiency to exertional breathlessness in CPFE.

Conclusions

The present study uncovered a hitherto unknown link between ventilatory inefficiency and exertional dyspnoea in CPFE, a key clinical outcome in these patients [1–5]. The heightened ventilatory response to exertion in the more dyspnoeic patients was associated with increased “wasted” ventilation and chronic alveolar hyperventilation, but not the severity of hypoxaemia or the development of inspiratory constraints. A consistent association between ventilatory inefficiency with total and admixed emphysema and traction bronchiectasis, regardless of the likelihood of pulmonary hypertension by echocardiography, indicates that “wasted” ventilation was primarily related to enlarged airspace in these patients. Therapeutic strategies focused in improving ventilatory efficiency and/or lessening hyperventilation might prove particularly valuable to mitigate the devastating consequences of exertional dyspnoea in patients with CPFE.