Discussion
A major finding regarding BOS is the decrease of lung function, especially reduced airway function including FEV1 and FEF25-75%. However, no study has examined the relationship between prognosis and lung function change in patients with BOS. Our current study is the first to discover the clinical implications of lung function change in BOS, especially during the first 3 months after BOS diagnosis. In the current study, 21 children were diagnosed with BOS, treated, and monitored with regular PFTs prospectively on the basis of our protocol, and accumulated data were analyzed retrospectively. The good prognosis groups (Groups 2 and 4) showed different changes in FEV1% pred and FEF25-75% pred during the first 3 months after BOS diagnosis when compared to the poor prognosis groups (Groups 1 and 3).
Because BOS is characterized by the narrowing of small airways via a fibroproliferative process 23,24, the reduction in FEV1 and FEF25-75% may be more prominent than in FVC, as in our current analysis. In a cohort study of adult patients with BOS, lung function trajectory was examined, and it concluded that FEV1 decreased rapidly 6 months before BOS diagnosis and stabilized after BOS diagnosis 19. Also, a study examining BOS after lung transplantation25 concluded that subjects with higher FEV1 or 6MWT values (measured at 3, 6, and 12 months and annually after lung transplantation) had a better chance of survival and a lower risk of developing BOS. However, these studies focused on pulmonary function before BOS diagnosis and did not show a relationship between lung function trajectory after BOS diagnosis and BOS prognosis. In the current study, we analyzed spirometric parameters over time after BOS diagnosis, and we investigated impulse oscillometric values, 6MWT and FeNO values at BOS diagnosis.
The pathogenesis of BOS after HSCT is associated with alloreactivity (e.g. graft-versus-host disease) 26,27, and systemic steroids and other anti-inflammatory agents are typically used to treat BOS; however, peribronchial inflammation can still progress into fibrosis causing a poor prognosis 18. Our current study showed that a rapid decline in FEV1 in the first 3 months after BOS correlates with poor prognosis (i.e. death or the need for a lung transplant or O2 therapy). Moreover, the slope of FEV1% pred level change during the first 3 months after BOS diagnosis determined BOS prognosis in Kaplan-Meier estimates even though all patients with BOS received the same treatment. Our first hypothesis that may explain this phenomenon is that airway inflammation in patients with a rapid FEV1 decline may be too severe to be controlled by anti-inflammatory treatment. Our second hypothesis posits that the rate of lung function reduction after BOS development in some patients is faster because it rapidly progresses into fibrosis in the airway, which can no longer respond to anti-inflammatory treatment. This progression may depend on an individual’s susceptibility to fibrosis after injury in the airway28. This study is the first to show that changes in pulmonary function during early disease processes can predict BOS prognoses. New treatment modalities during the first 3 months after BOS diagnosis are needed for patients with a poor prognosis.
This study found different changes in FEV1% pred and FEF25-75% pred during the first 3 months after BOS diagnosis between good and poor prognosis groups. However, changes in PFTs could not be compared across all periods because of death or lung TPL. Lung function likely declines persistently in patients with a poor prognosis, and if their data were included, a significant difference in the change of FEV1% pred after the first 3 months after BOS would likely occur. However, the data from subjects with a poor prognosis (Groups 1 and 3) were all present except for one subject until 6 months after BOS diagnosis. Furthermore, PFT changes from 3 to 6 months after BOS diagnosis were not significant. Therefore, the first 3 months after BOS diagnosis should be considered a meaningful period for determining prognosis. In addition, Figure 1a shows that FEV1 in Group 1 tended to increase after 18 months from BOS diagnosis, but not in a significant manner. However, this analysis was distorted because subjects with low lung functions died early. In this group, only two subjects completed PFTs during the 24 months after BOS diagnosis, and others dropped out due to death or lung TPL within 24 months after BOS diagnosis. We analyzed the PFT trajectories of two subjects who consistently performed PFTs until 24 months after BOS diagnosis (E-table 5, E-figure 3). Both subjects eventually received lung TPL because of decreasing lung function. One received lung TPL at 34.4 months after BOS diagnosis and the other received lung TPL at 68.5 months after BOS diagnosis. E-figure 3 illustrates that FEV1 steadily decreased during the first 24 months after BOS diagnosis in both subjects.
In this study, subjects performed IOS upon BOS diagnosis, and we found that the Xrs5% value in Group 3 was significantly higher than in Group 4. Reactance implies tissue elastance and inertance, especially at lower frequencies (e.g. 5 Hz), with a close association with capacitance, which better reflects the elasticity of the lung periphery29. No study has clarified the effect of IOS on the prognosis of BOS after HSCT. Studies using patients with diffuse interstitial lung disease and emphysema showed pulmonary fibrosis and emphysema can cause changes in Xrs5% pred due to lung stiffness, hyperinflation and a loss of lung elastic recoil, respectively30,31. Another study examining children with postinfectious bronchiolitis obliterans (PIBO) showed greater differences in Xrs5 than Rrs5 in children with PIBO compared to other groups 32. These findings are consistent with our current findings, and it suggests that Xrs5% pred may better reflect small airway obstruction than Rrs5% pred due to peribronchial fibrosis of lung periphery in children with PIBO and BOS 29-31. This study showed that IOS can predict BOS prognosis in young children who cannot undergo spirometry.
We examined CT results, which are another crucial factor in diagnosing BOS. Some studies attempted to predict lung function via CT scan in BOS patients, but this is challenging in young children because no consensus exists regarding reference values that define air trapping in chest CTs. In a pilot feasibility study of children (6-17 years of age), quantitative computed tomography assessments in children with BOS showed a correlation between their lung function and air trapping as defined by the individualized threshold ([attenuation values of normal lung parenchyma + attenuation values of air trapping area] / 2)33. In our current analysis, we used this as a cut-off value to distinguish air trapping lungs from normal lungs on CT scans. The percent of air trapping total lung volume according to CT measurements (air trapping volume/total lung volume) at 3, 6, and 12 months after BOS diagnosis was calculated, and Groups 1 and 2 as well as Groups 3 and 4 were compared. The change in air trapping during the first 3 months after BOS diagnosis was greater in poor prognosis groups (Groups 1 and 3), but the result was not significant. This suggests that the critical period for determining the prognosis of BOS is the first 3 months after BOS diagnosis, which agrees with the analysis of lung function change during BOS. In addition, these results suggest that PFTs are more sensitive to predict the prognosis of BOS even though CT scans may be helpful in tracking the prognosis of BOS. However, we could not perform expiration chest CTs to identify air trapping because children with dyspnea had difficulty completing expiration CTs, and frequent radiation exposure was a concern 34.
The major limitations of the current study are a small sample size and the use of different observation periods for each subject. However, there are very few studies examining pulmonary function in children diagnosed with BOS because of its low prevalence after hematopoietic stem cell transplantation. This is also the first study to show that changes in pulmonary function after BOS diagnosis can predict the prognosis. In the survival analysis, we also showed that PFT changes during the first 3 months after BOS diagnosis in children may influence their prognosis. Furthermore, when the observation periods of all subjects were unified to 24 months after BOS diagnosis in the analysis, there was a significant difference in the change of FEV1during the first 3 months after BOS diagnosis between Groups 3 and 4 (E-table 6). As another possibility to affect the results, the difference in the diagnosis time of BOS after HSCT or the difference in lung function change before BOS diagnosis among study subjects could be considered to affect the change in lung function after diagnosis of BOS. However, lung function changes during 6 months before BOS diagnosis had no effect on the prognosis (data was not shown).
In conclusion, the change of FEV1 during the first 3 months after BOS diagnosis was significantly different between good and poor prognosis groups. Our current analysis shows that the phase right after BOS diagnosis is the most critical in determining the prognosis of BOS. These results suggest that an active intervention strategy is needed during the first 3 months after BOS diagnosis to improve its prognosis