To the Editor:
Expiratory central airway collapse (ECAC) is defined by the excessive collapse of the central airway lumen during exhalation because of either tracheobronchomalacia or invagination of the posterior membrane (1). ECAC occurs in 5% of smokers and is associated with substantial respiratory morbidity (2). However, the symptoms of ECAC are nonspecific and include dyspnea, wheezing, cough, and difficulty with secretion clearance (3, 4). The diagnosis of ECAC can be made via bronchoscopy or computed tomography (CT) showing central airway collapse during exhalation (1). However, these diagnostic studies are expensive and time-consuming and expose the patient to risks of the procedure (bronchoscopy) or radiation (CT), respectively. There is a need for inexpensive screening strategies to guide further evaluation. Spirometry is a commonly used test for the evaluation of dyspnea, and several case reports, case series, and observational studies suggest that there may be distinct patterns on spirometry that are associated with the presence of ECAC (5–11). We tested whether distinct flow–volume patterns on spirometry can aid the identification of radiologic ECAC.
Methods
We analyzed paired static inspiratory/expiratory CT images of participants enrolled in the COPDGene (Genetic Epidemiology of Chronic Obstructive Pulmonary Disease) study, a multicenter cohort of current and former smokers aged 45 to 80 years (12). CT scans were acquired at end inspiration (total lung capacity) and at end expiration (functional residual capacity). A subset of expiratory scans (13%) was acquired at residual volume. The COPDGene study was approved by the institutional review boards of all 21 participating centers, and all participants provided written informed consent before participation. ECAC status on radiology was defined by ⩾50% reduction (ECAC50) in cross-sectional area from inspiration to end expiration at one or more of three predefined anatomic landmarks (aortic arch, carina, and bronchus intermedius) (2). In sensitivity analyses, we also analyzed ECAC defined by ⩾70% collapse (ECAC70).
Post-bronchodilator expiratory flow–volume curves were recreated using flow data collected every 30 ml and volume data collected every 60 ms. Curves were plotted on a 2:1 aspect ratio with flow on the y-axis and volume on the x-axis (13). Based on the shape of the descending part, the expiratory curves were classified as normal (linear descent), convex (bending away from the x-axis), or concave (bending toward the x-axis). The curves were also labeled as biphasic (rapid decline in respiratory flow to an inflection point, followed by a prolonged gradual descent), notching (rapid midexpiratory dip and return of flow), or flow oscillation (large-amplitude, high-frequency oscillations throughout most of the curve in a “sawtooth” pattern) when these patterns were present (Figure 1). These distinct patterns were selected on the basis of prior associations described with ECAC (5–11). The flow–volume curve readers were blinded to the ECAC status of the participants.
Figure 1.
Representative spirometry patterns. (A) Biphasic: rapid decline in flow to an inflection point followed by a prolonged slow decline. (B) Notching: rapid midexpiratory dip and return of flow. (C) Flow oscillation: large-amplitude, high-frequency oscillations throughout most of the curve.
The proportions of spirometry patterns were compared between those with and without radiologic ECAC using chi-square tests. Logistic regression was used to test the association between each spirometry pattern and ECAC status after adjustment for age, sex, race, body mass index, and forced expiratory volume in 1 second (FEV1) percent predicted. These variables were selected a priori on the basis of previously reported associations with ECAC (1, 13). All analyses were performed using IBM SPSS version 27.0 (IBM Corp.). A two-tailed α-value of 0.05 was deemed statistically significant.
Results
Of 10,300 participants enrolled in COPDGene, 8,440 participants were included in the analyses after those with incomplete CT data (n = 1,480) and those with incomplete raw spirometry data (n = 380) were excluded. The mean age was 59.6 years (standard deviation, 9); 3,974 (47%) were female; 2,646 (31%) were of non-Hispanic Black race; and 3,664 (43%) had airflow obstruction defined by FEV1/forced vital capacity <0.70. Radiologic ECAC was present in 418 (5%) participants using the 50% collapse threshold and in 125 (1.5%) using the 70% collapse threshold. Baseline characteristics of participants are shown in Table 1. Compared with those without ECAC, those with ECAC50 had a higher prevalence of concave (58.3% vs. 70.3%; P < 0.001) and biphasic (8.9% vs. 13.4%; P = 0.002) patterns and a lower prevalence of normal (33% vs. 24.9%; P < 0.001), convex (8.7% vs. 4.8%; P = 0.005), and flow oscillation (2.5% vs. 0.7%; P = 0.023) patterns. There was no difference in the prevalence of notching pattern in non-ECAC versus ECAC50 (4.4% vs. 3.6%; P = 0.422) (Table 1). Compared with non-ECAC, those with ECAC70 had a higher prevalence of concave (58.7% vs. 74.4%; P < 0.001) and notching (4.3% vs. 8.0%; P = 0.046) patterns and a lower prevalence of normal (32.8% vs. 21.6%; P = 0.008). There was no difference in the prevalence of convex (8.5% vs. 4.0%; P = 0.07), biphasic (9.0% vs. 13.6%; P = 0.078), and flow oscillation (2.4% vs. 0.8%) patterns (Table 1). The sensitivity, specificity, and positive and negative likelihood ratios for each of the spirometry patterns for the detection of ECAC are displayed in Table 2.
Table 1.
Baseline characteristics of participants
| Total (N = 8,440) | ECAC50 |
ECAC70 |
|||
|---|---|---|---|---|---|
| Absent (n = 8,022) | Present (n = 418) | Absent (n = 8,315) | Present (n = 125) | ||
| Age, yr | 59.6 (9.0) | 59.3 (8.9) | 65.0 (8.5) | 59.5 (9.0) | 65.0 (8.8) |
| Female sex, n (%) | 3,974 (47.1) | 3,692 (46.0) | 282 (67.5) | 3,886 (47.7) | 88 (70.4) |
| Non-Hispanic Black race, n (%) | 2,646 (31.4) | 2,582 (32.2) | 64 (15.3) | 2,627 (31.6) | 19 (15.2) |
| BMI, kg/m2 | 28.8 (6.2) | 28.7 (6.1) | 31.1 (6.5) | 28.8 (6.2) | 30.7 (5.9) |
| Current smoker, n (%) | 4,378 (51.9) | 4,261 (53.1) | 117 (28.0) | 3,834 (46.1) | 90 (72.0) |
| Pack-years of smoking | 43.8 (25.1) | 43.6 (25.0) | 46.8 (26.7) | 43.7 (25.1) | 45.3 (24.7) |
| FEV1, L | 2.3 (0.9) | 2.3 (0.9) | 1.8 (0.7) | 2.3 (0.9) | 1.8 (0.7) |
| FEV1, % predicted | 76.9 (25.5) | 77.3 (25.5) | 70.1 (25.9) | 77.0 (25.5) | 69.2 (25.1) |
| FVC, L | 3.3 (1.0) | 3.4 (1.0) | 2.8 (0.9) | 3.3 (1.0) | 2.8 (0.9) |
| FVC, % predicted | 87.4 (18.1) | 87.7 (18.1) | 82.9 (18.7) | 87.4 (18.0) | 83.4 (20.3) |
| FEV1/FVC | 0.67 (0.16) | 0.67 (0.16) | 0.63 (0.17) | 0.67 (0.16) | 0.63 (0.17) |
| COPD, n (%) | 3,664 (43.4) | 3,447 (43.0) | 217 (51.9) | 3,598 (43.2) | 66 (52.8) |
| GOLD stage, n (%) | |||||
| No obstruction | 4,776 (56.6) | 4,575 (57.0) | 201 (48.0) | 4,717 (56.7) | 59 (47.2) |
| 1 | 645 (7.7) | 613 (7.6) | 32 (7.7) | 635 (7.6) | 10 (8.0) |
| 2 | 1,599 (19.2) | 1,513 (18.9) | 86 (20.6) | 1,572 (18.9) | 27 (21.6) |
| 3 | 936 (11.2) | 873 (10.9) | 63 (15.1) | 920 (11.1) | 16 (12.8) |
| 4 | 484 (5.8) | 448 (5.6) | 36 (8.6) | 471 (5.7) | 13 (10.4) |
| Chronic bronchitis, n (%) | 1,589 (18.8) | 1,503 (18.7) | 86 (20.6) | 1,564 (18.8) | 25 (20.0) |
| mMRC score | 1.3 (1.4) | 1.3 (1.4) | 1.7 (1.5) | 1.3 (1.4) | 1.9 (1.5) |
| mMRC ⩾2, n (%) | 3,428 (40.6) | 3,216 (40.1) | 212 (50.7) | 3,356 (40.3) | 72 (57.6) |
| SGRQ score | 26.8 (22.8) | 26.6 (22.8) | 30.8 (22.3) | 26.7 (22.8) | 31.9 (21.0) |
| SGRQ score ⩾25, n (%) | 3,842 (45.5) | 3,618 (45.1) | 224 (53.6) | 3,772 (45.4) | 70 (56.0) |
| 6-Min-walk distance, ft | 1,371 (393) | 1,374 (393) | 1,317 (394) | 1,372 (393) | 1,308 (385) |
| Spirometry patterns, n (%) | |||||
| Normal | 2,753 (32.6) | 2,649 (33.0) | 104 (24.9) | 2,726 (32.8) | 27 (21.6) |
| Concave | 4,971 (58.9) | 4,677 (58.3) | 294 (70.3) | 4,878 (58.7) | 93 (74.4) |
| Convex | 715 (8.5) | 695 (8.7) | 20 (4.8) | 710 (8.5) | 5 (4.0) |
| Biphasic | 768 (9.1) | 712 (8.9) | 56 (13.4) | 751 (9.0) | 17 (13.6) |
| Notching | 369 (4.4) | 354 (4.4) | 15 (3.6) | 359 (4.3) | 10 (8.0) |
| Flow oscillation | 200 (2.4) | 197 (2.5) | 3 (0.7) | 199 (2.4) | 1 (0.8) |
Definition of abbreviations: BMI = body mass index; COPD = chronic obstructive pulmonary disease; ECAC50 = expiratory central airway collapse as defined by ⩾50% collapse at end expiration; ECAC70 = expiratory central airway collapse as defined by ⩾70% collapse at end expiration; FEV1 = forced expiratory volume in 1 second; FVC = forced vital capacity; GOLD = Global Initiative for Chronic Obstructive Lung Disease; mMRC = modified Medical Research Council dyspnea scale; SD = standard deviation; SGRQ = St. George’s Respiratory Questionnaire.
All continuous variables expressed as mean (SD), and all categorical variables are expressed as n (%).
Table 2.
Spirometry patterns in radiologic expiratory central airway collapse
| Spirometry Pattern | Sensitivity (95% CI) | Specificity (95% CI) | Positive Likelihood Ratio (95% CI) | Negative Likelihood Ratio (95% CI) | Adjusted Odds Ratio (95% CI) |
|---|---|---|---|---|---|
| ECAC50 | |||||
| Normal | 24.9 (20.8–29.3) | 67.0 (65.9–68.0) | 0.8 (0.6–0.9) | 1.1 (1.1–1.2) | 1.1 (0.8–1.4) |
| Concave | 70.3 (65.7–74.7) | 41.7 (40.6–42.8) | 1.2 (1.1–1.3) | 0.7 (0.6–0.8) | 1.1 (0.8–1.4) |
| Convex | 4.8 (3.0–7.3) | 91.3 (90.7–91.9) | 0.6 (0.4–0.9) | 1.0 (1.0–1.1) | 0.7 (0.4–1.1) |
| Biphasic | 13.4 (10.3–17.0) | 91.1 (90.5–91.7) | 1.5 (1.2–2.0) | 1.0 (0.9–1.0) | 1.2 (0.8–1.7) |
| Notching | 3.6 (2.0–5.9) | 95.6 (95.1–96.0) | 0.8 (0.5–1.4) | 1.0 (0.99–1.03) | 0.9 (0.5–1.6) |
| Flow oscillation | 0.7 (0.2–2.1) | 97.5 (97.2–97.9) | 0.3 (0.1–0.9) | 1.02 (1.01–1.03) | 0.6 (0.2–1.8) |
| ECAC70 | |||||
| Normal | 21.6 (14.7–29.9) | 67.2 (66.2–68.2) | 0.7 (0.5–0.9) | 1.2 (1.1–1.3) | 0.9 (0.6–1.4) |
| Concave | 74.4 (65.8–81.8) | 41.3 (40.3–42.4) | 1.3 (1.1–1.4) | 0.6 (0.5–0.8) | 1.3 (0.8–2.1) |
| Convex | 4.0 (1.3–9.1) | 91.5 (90.8–92.1) | 0.5 (0.2–1.1) | 1.05 (1.01–1.09) | 0.6 (0.2–1.5) |
| Biphasic | 13.6 (8.1–20.9) | 91.0 (90.3–91.6) | 1.5 (1.0–2.4) | 0.95 (0.89–1.02) | 1.0 (0.5–2.0) |
| Notching | 8.0 (3.9–14.2) | 95.7 (95.2–96.1) | 1.9 (1.0–3.4) | 0.96 (0.91–1.01) | 2.3 (1.2–4.5) |
| Flow oscillation | 0.8 (0.0–4.4) | 97.6 (97.3–97.9) | 0.3 (0.1–2.4) | 1.02 (1.00–1.03) | 0.7 (0.1–4.8) |
Definition of abbreviations: CI = confidence interval; ECAC50 = expiratory central airway collapse as defined by ⩾50% collapse at end expiration; ECAC70 = expiratory central airway collapse as defined by ⩾70% collapse at end expiration.
Adjusted odds ratio calculated by logistic regression, adjusting for age, sex, race, body mass index, and the forced expiratory volume in 1 second percent predicted. Bold value represents associations significant at two-sided alpha of 0.05.
For the presence of ECAC50, the unadjusted odds ratios (95% confidence intervals [CI]) for normal, concave, convex, biphasic, notching, and flow oscillation were 0.67 (0.54–0.84), 1.70 (1.37–2.10), 0.53 (0.33–0.84), 1.59 (1.19–2.13), 0.81 (0.48–1.37), and 0.29 (0.10–0.90), respectively. For ECAC70, the unadjusted odds ratios (95% CIs) were 0.57 (0.67–0.87), 2.05 (1.37–3.07), 0.45 (0.18–1.10), 1.59 (095–2.66), 1.93 (1.00–3.71), and 0.33 (0.05–2.37), respectively. After adjustment for age, sex, race, body mass index, and FEV1 percent predicted, none of the spirometry patterns were associated with the presence of ECAC50. Only the notching pattern was associated with ECAC70 (odds ratio, 2.3; 95% CI, 1.2–4.5) (Table 2).
Discussion
In a cohort of current and former smokers at risk for ECAC, after adjusting for underlying lung disease, we did not find patterns on flow–volume curves that are specifically associated with radiologic ECAC at traditional thresholds for collapse. The notching pattern was associated with radiologic ECAC when defined by a 70% threshold for collapse, but the positive and negative likelihood ratios were small.
Several prior reports have suggested associations of specific spirometry patterns with ECAC. The biphasic pattern has previously been described in case series of patients with bronchoscopically confirmed ECAC (6, 7). Flow oscillation, defined as a reproducible sequence of alternating decelerations and accelerations of flow, has been described in case reports of patients with ECAC (10, 11). However, this pattern has also been described in other airway disorders, such as obstructive sleep apnea and disorders of the larynx, and in neuromuscular diseases, such as Parkinson’s disease (14–16). In a study of 47 participants, the notching pattern was described in 54% of those with radiologic ECAC versus 6% in control subjects, suggesting a possible association (8). In an observational cohort of 76 subjects with confirmed ECAC, Majid and colleagues reported biphasic patterns in 19.7%, flow oscillations in 2.6%, and notching patterns in 9.2% of participants (5). These studies were heterogeneous in the diagnostic methods and were limited by referral bias and lack of control subjects. Importantly, they did not account for the presence and severity of underlying lung disease.
Major strengths of our study include the large sample size of participants enrolled from multiple centers and the extensive phenotyping of participants with quality-controlled CT imaging and spirometry. We standardized the visual height/width display of flow–volume curves per the American Thoracic Society standards (13). The study had a few limitations. We analyzed a single curve for each participant, which may have resulted in an incomplete capture of all flow–volume curve patterns. However, we analyzed the curve with the best sum of FEV1 and forced vital capacity per American Thoracic Society guidelines (13). The prevalence of radiologic ECAC was lower in the COPDGene cohort than prior estimates, which have ranged widely from 5% to 70%; these studies, however, used a range of thresholds and methods to detect ECAC and were performed mostly in symptomatic adults and hence had a referral bias (17). We acknowledge that radiologic ECAC is likely far more common than ECAC that meets criteria for clinical intervention. Another possible limitation was the use of paired static inspiratory/expiratory rather than dynamic CT imaging for the diagnosis of ECAC. Although dynamic imaging with forced exhalation is more sensitive, CT acquired with static breath holds may have greater specificity for ECAC and is more widely available clinically (2). Thirteen percent of expiratory scans were acquired at residual volume instead of functional residual capacity, but we have previously shown that the clinical associations with radiologic ECAC defined using scans acquired at either expiratory level are similar (2). We did not adjust for multiple comparisons, and it is possible that the association between the notching pattern and radiologic ECAC may be a spurious finding.
Conclusions
In conclusion, although the notching pattern was associated with higher odds of detecting radiologic ECAC and should raise clinical suspicion for ECAC, none of the flow–volume curve patterns were associated with diagnostic characteristics that would help inform clinical decisions in the diagnostic work-up of ECAC. Practitioners must maintain a high index of clinical suspicion when evaluating patients with symptoms compatible with ECAC.
Footnotes
Supported by National Institutes of Health grants R01 HL151421 (S.P.B.), R21EB027891 (S.P.B.), UH3HL155806 (S.P.B.), U01 HL089897, and U01 HL089856. COPDGene is also supported by the COPD Foundation through contributions made to an industry advisory board composed of AstraZeneca, Boehringer-Ingelheim, Genentech, GlaxoSmithKline, Novartis, Pfizer, Siemens, and Sunovion.
Author disclosures are available with the text of this letter at www.atsjournals.org.
References
- 1. Murgu S, Colt H. Tracheobronchomalacia and excessive dynamic airway collapse. Clin Chest Med . 2013;34:527–555. doi: 10.1016/j.ccm.2013.05.003. [DOI] [PubMed] [Google Scholar]
- 2. Bhatt SP, Terry NL, Nath H, Zach JA, Tschirren J, Bolding MS, et al. Genetic Epidemiology of COPD (COPDGene) Investigators Association between expiratory central airway collapse and respiratory outcomes among smokers. JAMA . 2016;315:498–505. doi: 10.1001/jama.2015.19431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Abia-Trujillo D, Majid A, Johnson MM, Mira-Avendano I, Patel NM, Makey IA, et al. Central airway collapse, an underappreciated cause of respiratory morbidity. Mayo Clin Proc . 2020;95:2747–2754. doi: 10.1016/j.mayocp.2020.03.004. [DOI] [PubMed] [Google Scholar]
- 4. Tanabe N, Terada K, Shima H, Hamakawa Y, Shiraishi Y, Shimizu K, et al. Expiratory central airway collapse and symptoms in smokers. Respir Investig . 2021;59:522–529. doi: 10.1016/j.resinv.2021.03.005. [DOI] [PubMed] [Google Scholar]
- 5. Majid A, Sosa AF, Ernst A, Feller-Kopman D, Folch E, Singh AK, et al. Pulmonary function and flow-volume loop patterns in patients with tracheobronchomalacia. Respir Care . 2013;58:1521–1526. doi: 10.4187/respcare.02277. [DOI] [PubMed] [Google Scholar]
- 6. Campbell AH, Faulks LW. Expiratory air-flow pattern in tracheobronchial collapse. Am Rev Respir Dis . 1965;92:781–791. doi: 10.1164/arrd.1965.92.5.781. [DOI] [PubMed] [Google Scholar]
- 7. Grandevia B. The spirogram of gross expiratory tracheobronchial collapse in emphysema. Q J Med . 1963;32:23–31. [PubMed] [Google Scholar]
- 8. Nuutinen J. Acquired tracheobronchomalacia. A clinical study with bronchological correlations. Ann Clin Res . 1977;9:350–355. [PubMed] [Google Scholar]
- 9. Koblet H, Wyss F. Clinical and functional picture of genuine bronchial collapse with lung emphysema [in German] Helv Med Acta . 1956;23:553–560. [PubMed] [Google Scholar]
- 10. Garcia-Pachon E. Tracheobronchomalacia: a cause of flow oscillations on the flow-volume loop. Chest . 2000;118:1519. doi: 10.1378/chest.118.5.1519. [DOI] [PubMed] [Google Scholar]
- 11. Bates CA, Rosenwasser LJA. A 49-year-old male with intractable dyspnea, wheeze, and cough. Ann Allergy Asthma Immunol . 2003;91:20–25. doi: 10.1016/S1081-1206(10)62053-4. [DOI] [PubMed] [Google Scholar]
- 12. Regan EA, Hokanson JE, Murphy JR, Make B, Lynch DA, Beaty TH, et al. Genetic Epidemiology of COPD (COPDGene) study design. COPD . 2010;7:32–43. doi: 10.3109/15412550903499522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Graham BL, Steenbruggen I, Miller MR, Barjaktarevic IZ, Cooper BG, Hall GL, et al. Standardization of spirometry 2019 update. An official American Thoracic Society and European Respiratory Society technical statement. Am J Respir Crit Care Med . 2019;200:e70–e88. doi: 10.1164/rccm.201908-1590ST. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sahni S, Blood A, Paulus S, Talwar A. “Saw-tooth sign” in upper airway disorders—a case report. Pneumonol Alergol Pol . 2015;83:140–143. doi: 10.5603/PiAP.2015.0022. [DOI] [PubMed] [Google Scholar]
- 15. Modrykamien AM, Gudavalli R, McCarthy K, Liu X, Stoller JK. Detection of upper airway obstruction with spirometry results and the flow-volume loop: a comparison of quantitative and visual inspection criteria. Respir Care . 2009;54:474–479. [PubMed] [Google Scholar]
- 16. Vincken WG, Cosio MG. Flow oscillations on the flow-volume loop: clinical and physiological implications. Eur Respir J . 1989;2:543–549. [PubMed] [Google Scholar]
- 17. Mitropoulos A, Song W-J, Almaghlouth F, Kemp S, Polkey M, Hull JH. Detection and diagnosis of large airway collapse: a systematic review. ERJ Open Res . 2021;7:00055-02021. doi: 10.1183/23120541.00055-2021. [DOI] [PMC free article] [PubMed] [Google Scholar]