Pneumothorax Trends from 2010-2020 from a Large-Volume Pleural Unit

Previous studies in 2018 and 2022 have suggested increasing inpatient burden of pneumothorax and widespread variation in management. Local trends have never been elucidated. Northumbria Healthcare NHS Foundation Trust (NHCT) has a well-established pleural service, serving just over 600,000. Thus, we set up a local retrospective study to look at trends in pneumothorax presentation, management strategies, length of stay, and recurrence.

A coding search for 'pneumothorax' was performed for all patients attending NHCT between 2010 and 2020 was performed with local Caldicott approval. A total of 1840 notes were analysed to exclude iatrogenic, traumatic, and paediatric events. After excluding those cases, 580 remained for further analysis, consisting of 183 primary pneumothoraces (PSP) and 397 secondary pneumothoraces (SSP).

Median age for PSP was 26.5 years (IQR 17) with 69% male, and for SSP 68 years (IQR 11.5), 62% male; 23.5% of PSP and 8.6% of SSP were never smokers. The proportion of smokers and ex-smokers has not really changed over time: > 65% every year have been smokers or ex-smokers. Yearly pneumothorax incidence shows a downward trend for PSP but upwards for SSP. Median length of stay (LoS) for PSP was 2 days (IQR 2), and SSP 5 days (IQR 8), with a clear downward trend. From 2010 to 2015 > 50% PSP were managed with drain, but in 2019-2020 at least 50% managed conservatively, with a significant reduction in aspiration. Trends of recurrence for PSP are increasing, whereas for SSP is decreasing. Seventy-six (20 PSP, 56 SSP) went for surgery at the index time with 5.3% recurrence (20% recurrence in those without surgery).

This is the first known analysis of pneumothorax trends in a large trust in the northeast of England. The data in this study have certain limitations, including the lack of information on the size of pneumothorax and frailty indicators that may influence the decision for conservative management. Additionally, there is a reliance on clinical coding, which can introduce potential inaccuracies, and not all patient notes were accessible for analysis. Updated larger datasets should help elucidate trends better.