Dear Editor,
Pneumothorax and pneumomediastinum may complicate acute respiratory distress syndrome (ARDS). Early studies in ARDS caused by coronavirus disease 2019 (COVID-19) suggested increased pneumothorax incidence but lacked relevant controls [1, 2]. We investigated whether COVID-19 ARDS is associated with more radiographic pneumothorax and/or pneumomediastinum than pre-pandemic ARDS and whether pneumothorax/pneumomediastinum in COVID-19 ARDS is associated with worse outcomes or differing treatments.
This retrospective cohort study included adult ARDS patients admitted between 2017 and 2021 to a 23-hospital system in the Intermountain West. We abstracted data from the electronic health record and used natural language processing to identify radiographic pneumothorax and/or pneumomediastinum [3, 4]. We performed bivariate and adjusted analyses to compare patients with pre-pandemic ARDS (2017–2020) to patients with a positive SARS-CoV-2 polymerase chain reaction (PCR) result proximate to ARDS (2020–2021) (see also Supplemental Methods).
Comparing 2,211 patients with COVID-19 ARDS and 5522 with pre-pandemic ARDS (Table 1 and Supplemental Fig. 1), unadjusted incidence of pneumothorax/pneumomediastinum was similar (24% vs. 22.5%, p < 0.148). After adjustment, pneumothorax/pneumomediastinum risk was significantly higher in COVID-19 vs. pre-pandemic ARDS (adjusted odds ratio 1.31, 95% CI 1.13–1.52, p < 0.001). COVID-19 ARDS patients had significantly higher rates of pneumomediastinum but not pneumothorax in unadjusted and adjusted analyses (Table 1 and Supplemental Table 2). Compared to COVID-19 ARDS, chest tube placement for pre-pandemic pneumothorax patients was more frequent (52.1% vs. 38.2%, p < 0.001), occurred earlier (− 0.4 vs. 1.3 days, p < 0.001) and remained in place longer (9.9 days vs. 7 days, p < 0.001).
Table 1.
Summary demographic and outcome data presented as n (%) or median [IQR]
| Overall | p | No pneumothorax or pneumomediastinum | Pneumothorax and/or pneumomediastinum | p | ||||
|---|---|---|---|---|---|---|---|---|
| COVID-19 | Prepandemic | COVID-19 | Prepandemic | COVID-19 | Prepandemic | |||
| n | 2211 | 5522 | 1680 | 4282 | 531 | 1240 | ||
| Female | 810 (36.6) | 2377 (43) | < 0.001 | 641 (38.2) | 1913 (44.7) | 169 (31.8) | 464 (37.4) | < 0.001 |
| Age | 61 [49, 70] | 63 [51, 74] | < 0.001 | 61 [49, 71] | 64 [52, 74] | 61 [49, 69] | 61 [46, 71] | < 0.001 |
| ARDS Qualifying Diagnostic Group (more than one category possible) | ||||||||
| Trauma | 130 (5.9) | 1199 (21.7) | < 0.001 | 89 (5.3) | 742 (17.3) | 41 (7.7) | 457 (36.9) | < 0.001 |
| Pneumonia | 2132 (96.4) | 2689 (48.7) | < 0.001 | 1621 (96.5) | 2124 (49.6) | 511 (96.2) | 565 (45.6) | < 0.001 |
| Sepsis | 37 (1.7) | 1208 (21.9) | < 0.001 | 26 (1.5) | 985 (23) | 11 (2.1) | 223 (18) | < 0.001 |
| Aspiration | 95 (4.3) | 1390 (25.2) | < 0.001 | 70 (4.2) | 1123 (26.2) | 25 (4.7) | 267 (21.5) | < 0.001 |
| Shock | 228 (10.3) | 787 (14.3) | < 0.001 | 155 (9.2) | 505 (11.8) | 73 (13.7) | 282 (22.7) | < 0.001 |
| Acute pancreatitis | 16 (0.7) | 144 (2.6) | < 0.001 | 11 (0.7) | 115 (2.7) | 5 (0.9) | 29 (2.3) | < 0.001 |
| Overdose | 15 (0.7) | 336 (6.1) | < 0.001 | 14 (0.8) | 288 (6.7) | 1 (0.2) | 48 (3.9) | < 0.001 |
| Worst ARDS severity (first 7 days) | ||||||||
| Mild | 38 (1.7) | 955 (17.3) | < 0.001 | 32 (1.9) | 794 (18.5) | 6 (1.1) | 161 (13) | < 0.001 |
| Moderate | 232 (10.5) | 2581 (46.7) | 194 (11.5) | 2045 (47.8) | 38 (7.2) | 536 (43.2) | ||
| Severe | 1941 (87.8) | 1986 (36) | 1454 (86.5) | 1443 (33.7) | 487 (91.7) | 543 (43.8) | ||
| Hospital Day 1 Lowest P/F Ratio | 78.1 [61.7, 118.9] | 148.6 [96.7, 204.8] | < 0.001 | 79.4 [62.9, 124.1] | 148.4 [96.8, 203.5] | 74.4 [60.8, 101.5] | 149.2 [96.2, 212.9] | < 0.001 |
| Hospital day 1 SOFA Score | 6 [4, 9] | 8 [5, 11] | < 0.001 | 6 [4, 9] | 8 [5, 11] | 7 [4, 10] | 8 [5, 11] | < 0.001 |
| BMI | 32.8 [28.4, 38.9] | 28.9 [24.3, 35.4] | < 0.001 | 33.5 [28.5, 40.1] | 29.2 [24.4, 35.9] | 31.5 [27.5, 35.8] | 28 [23.9, 34] | < 0.001 |
| Weighted (von Walraven) Elixhauser comorbidity score | 15 [7, 25] | 23 [13, 32] | < 0.001 | 15 [6.5, 24] | 23 [14, 32] | 17 [10, 26] | 22 [12, 32] | < 0.001 |
| Days from admission until endotracheal Intubation | 0.8 [0, 3.9] | 0.1 [0, 1.3] | < 0.001 | 0.7 [0, 3.3] | 0.1 [0, 1.1] | 1.2 [0, 5.3] | 0.3 [0, 1.8] | < 0.001 |
| Maximum respiratory support on hospital day 1 | ||||||||
| FiO2 | 100 [66, 100] | 76.5 [50, 100] | < 0.001 | 100 [66, 100] | 74 [50, 100] | 100 [70, 100] | 81 [40, 100] | < 0.001 |
| PEEP | 14 [10, 18] | 8 [7, 10] | < 0.001 | 14 [10, 18] | 8 [7, 12] | 15 [12, 18] | 8 [7, 10] | < 0.001 |
| Plateau pressure | 29 [25, 32] | 22 [18, 26] | < 0.001 | 29 [25, 32] | 22 [18, 26] | 29 [26, 33] | 21 [17, 26] | < 0.001 |
| Peak inspiratory pressure | 31 [21, 36] | 26 [20, 32] | < 0.001 | 30 [21, 36] | 25 [19, 32] | 33 [23, 37] | 27 [21, 33] | < 0.001 |
| Positive pressure ventilation | 1281 (58) | 4189 (76) | < 0.001 | 970 (57.8) | 3306 (77.3) | 311 (58.6) | 883 (71.4) | < 0.001 |
| Invasive mechanical ventilation | 872 (39.5) | 2902 (52.6) | < 0.001 | 639 (38.1) | 2194 (51.3) | 233 (43.9) | 708 (57.3) | < 0.001 |
| Outcomes | ||||||||
| Pneumomediastinum | 288 (13) | 188 (3.4) | < 0.001 | 288 (54.2) | 188 (15.2) | < 0.001 | ||
| Pneumothorax | 448 (20.3) | 1201 (21.7) | 0.158 | 448 (84.4) | 1201 (96.9) | < 0.001 | ||
| Pneumothorax or pneumomediastinum | 531 (24) | 1240 (22.5) | 0.148 | |||||
| Days from admission until pneumothorax or pneumomediastinum | 7.3 [2.9, 12.6] | 1.3 [0.1, 5.1] | < 0.001 | 7.3 [2.9, 12.6] | 1.3 [0.1, 5.1] | < 0.001 | ||
| Hospital Length of Stay (days) | 14.5 [9.5, 23.7] | 9.2 [5.3, 15.4] | < 0.001 | 13.1 [8.9, 21] | 8.3 [4.9, 13.6] | 20.8 [12.8, 33.4] | 13.9 [8.4, 21] | < 0.001 |
| 30 Day Mortality | 871 (39.4) | 1572 (28.5) | < 0.001 | 608 (36.2) | 1265 (29.5) | 263 (49.5) | 307 (24.8) | < 0.001 |
| ICU Length of Stay | 10.4 [6, 18.4] | 4.9 [2.5, 9.9] | < 0.001 | 9 [5.2, 15.2] | 4.3 [2.2, 8.6] | 17.1 [10.1, 27.3] | 7.9 [4, 14.4] | < 0.001 |
| Management of pneumothorax and/or pneumomediastinum | ||||||||
| Chest tube placed | 203 (38.2) | 646 (52.1) | < 0.001 | |||||
| Days from admission until chest tube placement | 8.6 [3.9, 15.9] | 0.9 [0.2, 3.8] | < 0.001 | |||||
| Duration of chest tube (days) | 9.9 [4.9, 17] | 7 [4.1, 11.5] | < 0.001 | |||||
| Treatment occurring prior to pneumothorax/pneumomediastinum | ||||||||
| Nasal canula utilized | 228 (44) | 598 (56.5) | < 0.001 | |||||
| High-flow nasal canula utilized | 402 (77.6) | 179 (16.9) | < 0.001 | |||||
| Non-invasive ventilation utilized | 251 (48.5) | 368 (34.8) | < 0.001 | |||||
| Invasive ventilation utilized | 400 (77.2) | 783 (74) | < 0.001 | |||||
| Positive pressure ventilation | 481 (92.9) | 918 (86.8) | < 0.001 | |||||
| Nasal canula days | 0 [0, 0.9] | 0.2 [0, 2] | < 0.001 | |||||
| High-flow nasal canula | 0.6 [0, 3.5] | 0 [0, 0] | < 0.001 | |||||
| Non-invasive ventilation | 0 [0, 0.7] | 0 [0, 0.3] | < 0.001 | |||||
| Invasive ventilation days | 2 [0, 8.4] | 0.3 [0, 2.6] | < 0.001 | |||||
| Maximum FiO2 | 100 [100, 100] | 100 [65.5, 100] | < 0.001 | |||||
| Maximum PEEP | 16 [14, 20] | 10 [8, 12] | < 0.001 | |||||
| Maximum plateau pressure | 34 [30, 40] | 24 [19, 30] | < 0.001 | |||||
| Maximum peak inspiratory pressure | 38 [28, 45] | 29 [23, 36] | < 0.001 | |||||
Mortality rates in COVID-19 ARDS were higher than pre-pandemic ARDS (39.4% vs. 28.5% p < 0.001). Among COVID-19 ARDS patients, we observed higher 30-day mortality rates with pneumothorax/pneumomediastinum (49.5% vs. 36.2%, p < 0.001), while we observed a lower mortality in pre-pandemic ARDS patients with pneumothorax/pneumomediastinum (24.8% vs. 29.5%, p < 0.001). Adjusted analyses yielded similar results (Supplemental Table 3).
Prior to pneumothorax/pneumomediastinum, both COVID-19 and pre-pandemic ARDS cohorts had similar receipt of invasive mechanical ventilation (77% vs. 74%, p = 0.17). COVID-19 patients received higher maximum PEEP (16 vs. 10 mmHg, p < 0.001). The median duration of invasive ventilation prior to pneumothorax/pneumomediastinum was much longer in the COVID-19 patients (2 vs. 0.3 days, p < 0.001; Supplemental Fig. 2), as was time from admission until pneumothorax/pneumomediastinum (7.3 vs. 1.3 days, p < 0.001).
Study strengths include comparison of large, multi-hospital COVID-19 and control ARDS cohorts. Limitations include the possibility of unmeasured confounding and potentially counting radiographic pneumothorax/pneumomediastinum events that were “clinically insignificant” or not due to acute lung injury. We note a substantially higher rate of pneumothorax/pneumomediastinum compared with other published cohorts (Supplemental Table 5). Our detection is more sensitive than clinically reported as all events are included, not just pneumothorax/pneumomediastinum > 2 cm or presence in clinical notes, which may limit generalizability. The relationships between radiographic and clinically significant pneumothorax/pneumomediastinum, pneumothorax/pneumomediastinum risk factors (including use of guideline-endorsed “high positive end-expiratory pressure (PEEP)” ventilation [5]), and pneumothorax management warrant further study.
In conclusion, COVID-19 ARDS patients experienced similar rates of radiographic pneumothorax but more pneumomediastinum. Chest tubes were used less frequently and placed later in COVID-19 ARDS than in pre-pandemic ARDS. Radiographic pneumothorax/pneumomediastinum in COVID-19 ARDS patients is associated with an increased mortality.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We would like to acknowledge Kyle Henry MD for assistance with proofreading the manuscript, Colin Grissom MD and James Orme MD for assistance in design of the study and Jason R. Jacobs for assistance with the data queries. The study was determined to be exempt from review by the Intermountain Healthcare IRB #1051342.
Author contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Daniel Knox. The first draft of the manuscript was written by Alex Brunhoeber and Daniel Knox and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding
National Institutes of Health (K23GM129661).
Data availability
To protect patient privacy and comply with relevant regulations, identified data are unavailable. Requests for deidentified data from qualified researchers with appropriate ethics board approvals and relevant data use agreements will be processed by the Intermountain Office of Research, officeofresearch@imail.org.
Declarations
Conflicts of interest
SMB chairs a DSMB for Hamilton Ventilators and has received research funding from National Institutes of Health, Department of Defense, and Centers for Disease Control and Prevention. IDP reports grant support from NIH, Centers from Disease Control and Prevention, Intermountain Research and Medical Foundation, and Janssen Pharmaceuticals, and payments to his institution for trial enrollment from Asahi Kasei Pharma and Regeneron. Otherwise no conflicts for any other authors.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Belletti A, Todaro G, Valsecchi G, et al. Barotrauma in Coronavirus Disease 2019 patients undergoing invasive mechanical ventilation: a systematic literature review. Crit Care Med. 2021 doi: 10.1097/CCM.0000000000005283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Shrestha DB, Sedhai YR, Budhathoki P, et al. Pulmonary barotrauma in COVID-19: a systematic review and meta-analysis. Ann Med Surg. 2022;73:103221. doi: 10.1016/j.amsu.2021.103221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Afshar M, Joyce C, Oakey A, et al. A computable phenotype for acute respiratory distress syndrome using natural language processing and machine learning. AMIA Annu Symp Proc AMIA Symp. 2018;2018:157–165. [PMC free article] [PubMed] [Google Scholar]
- 4.Mayampurath A, Churpek MM, Su X, et al. External validation of an acute respiratory distress syndrome prediction model using radiology reports. Crit Care Med. 2020;48:e791–e798. doi: 10.1097/CCM.0000000000004468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fan E, Del Sorbo L, Goligher EC, et al. An Official American Thoracic Society/European Society of Intensive Care Medicine/Society of Critical Care Medicine Clinical Practice Guideline: mechanical ventilation in adult patients with acute respiratory distress syndrome. Am J Respir Crit Care Med. 2017;195:1253–1263. doi: 10.1164/rccm.201703-0548ST. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
To protect patient privacy and comply with relevant regulations, identified data are unavailable. Requests for deidentified data from qualified researchers with appropriate ethics board approvals and relevant data use agreements will be processed by the Intermountain Office of Research, officeofresearch@imail.org.