This Week in Pulmonary — Aug 18, 2026

Welcome to This Week in Pulmonary. This week we're covering 10 notable papers spanning COPD — from beta-blocker safety to CT-based case finding and risk prediction — sleep apnoea epidemiology, critical care monitoring and ventilator liberation, and finally interstitial and paediatric airway disease. Let's dive in.

We'll start with the question that comes up in almost every clinic: should a patient with COPD and cardiovascular disease be on a beta-blocker? In Chest, Suissa and colleagues used the United Kingdom Clinical Practice Research Datalink to build a cohort of nearly twelve thousand beta-blocker initiators with both COPD and cardiovascular disease, each matched to a non-user on time, propensity score, and cardiac indication, in a prevalent new-user design intended to emulate a randomised trial [1]. Overall, starting a beta-blocker was associated with roughly a twelve percent lower risk of death. But the benefit was entirely concentrated where you would expect it from cardiology: about a twenty-three percent mortality reduction in heart failure and a sixteen percent reduction in ischaemic heart disease, with no signal at all in atrial fibrillation and, if anything, a neutral-to-unfavourable point estimate in hypertension. The more provocative finding is the interaction with inhaled therapy — among patients concurrently taking a long-acting beta-agonist the mortality benefit essentially disappeared, while among those not on a long-acting beta-agonist mortality was about twenty percent lower, and that interaction was statistically significant and driven by the ischaemic heart disease group. This is observational and residual confounding cannot be excluded, but the practical message is clear enough: do not withhold beta-blockers from COPD patients with heart failure or ischaemic disease, be much less enthusiastic when the only indication is hypertension or rate control in atrial fibrillation, and recognise that pharmacological antagonism between inhaled beta-agonists and systemic beta-blockade may be clinically real rather than theoretical.

Staying with COPD, two papers tackle detection and risk stratification. Also in Chest, Abdo and colleagues asked whether the low-dose CT scans already being done for lung cancer screening can be mined to find undiagnosed COPD [3]. Among just over five thousand screening participants who underwent spirometry, more than one in five had previously undiagnosed airflow obstruction — a prevalence of twenty-two percent, which is a striking amount of missed disease. Using emphysema extent alone at an optimised threshold of about five percent performed only moderately, with an area under the curve of 0.69 and a positive predictive value of forty-four percent, and it would have sent forty-one percent of the cohort for confirmatory spirometry. Adding artificial-intelligence-derived airway wall thickness and airway branch count together with smoking history and dyspnoea lifted the area under the curve to 0.83 and the positive predictive value to fifty-nine percent, while reducing referrals for spirometry to about a third of participants. So if your institution runs a screening programme, the structured CT report plus two simple clinical questions is a more efficient triage tool than emphysema score alone. Then, in Respiratory Medicine, Joo and colleagues examined post-bronchodilator mid-expiratory flow — FEF twenty-five to seventy-five — in eight hundred and thirty patients from the Korean KOCOSS cohort followed for three years [4]. Patients in the lowest quartile had a far heavier exacerbation burden, with eighty-one percent having at least one moderate-to-severe exacerbation versus fifty-four percent in the highest quartile, and mean cumulative exacerbations rising from two to nearly six. Crucially, though, that association survived adjustment for forced vital capacity but vanished completely once forced expiratory volume in one second was added to the model. The honest conclusion is that this index is a marker of overall physiological impairment, not an independent small-airway-specific predictor — so there is no case for reporting it as a separate risk flag beyond what the FEV1 already tells you.

Rounding out COPD, in the International Journal of Chronic Obstructive Pulmonary Disease, Fırat and colleagues developed and internally validated a prognostic model for in-hospital death among three hundred and ninety-two patients admitted to intensive care with an acute COPD exacerbation, of whom roughly a quarter died [5]. Six routinely available admission variables — age, male sex, urea, albumin, C-reactive protein, and the Charlson Comorbidity Index — produced an area under the curve of about 0.90, outperforming APACHE Two at 0.81, with good calibration and stability across a thousand bootstrap resamples. It is retrospective, single-centre, and lacks external validation or comparison with established COPD-specific scores, so treat this as promising rather than deployable.

Turning to sleep, two papers this week land in Respiratory Medicine and both should be read with your epidemiology hat firmly on. Azarian and colleagues used natural language processing to extract apnoea-hypopnoea index values from Veterans Health Administration records across more than one hundred and forty-six thousand veterans referred for sleep studies [7]. Unadjusted, severe sleep apnoea was associated with higher mortality in both sexes — but after full adjustment the direction reversed, with severe apnoea associated with roughly forty percent lower adjusted odds of death in women and forty-four percent lower in men. The apparent protective effect was largest in older adults, where adjusted odds of death were around seventy percent lower in women. Nobody should conclude that severe apnoea prolongs life. This is much more plausibly a signal of referral, diagnosis, and treatment bias — patients who get diagnosed are engaged with the health system and likely receive therapy — plus the collider effects of adjusting for obesity and cardiometabolic comorbidity in a referred population. It is a useful reminder of how badly clinically diagnosed cohorts can mislead. Alongside it, Huang and colleagues followed more than two thousand six hundred adults aged sixty and over from the Guangzhou Heart Study, all free of sleep apnoea at baseline, and identified three hundred and seventy-nine new cases over a median of about thirty-one months [9]. Of six inflammation-related biomarkers tested, only serum bilirubin was associated with incident apnoea — roughly a forty percent increase in risk comparing the highest with the lowest quartile, somewhat stronger in women and in those over sixty-eight — while C-reactive protein was associated with risk in men only. The four other markers, including neutrophil-to-lymphocyte ratio and the systemic immune-inflammation index, showed no association. Since outcomes were ascertained by Berlin Questionnaire rather than polysomnography, this is hypothesis-generating and should not change any testing pathway.

On the critical care side, Baptistella and colleagues offer a narrative synthesis in Chest proposing what they call a 3S framework for ventilator liberation — screening, separating, and securing [2]. Screening covers sedation management, preservation of limb and respiratory muscle strength and cardiac function, and the transition from controlled to partial support. Separating covers readiness assessment, the spontaneous breathing trial itself, and the extubation risk decision. Securing covers post-extubation management to prevent recurrent respiratory failure. There is no new outcome data here, but as a way to structure a fragmented and highly variable process — and as a teaching scaffold for trainees — it is worth reading. Complementing that, Kächele and colleagues in Respiratory Medicine tested transcutaneous capnometry against arterial blood gas in one hundred patients with acute respiratory failure, sampling at baseline and again up to twelve hours later [8]. Correlation was excellent at baseline but degraded meaningfully by the later timepoint, and just under a quarter of paired values fell outside the clinically acceptable range — though the device underestimated arterial carbon dioxide in only five cases. Deviations clustered significantly in patients with circulatory impairment, and every clinically relevant discrepancy traced back to a technical error that a trained operator would have caught. Use transcutaneous carbon dioxide for trends, not absolute decisions, and distrust it in the shocked or poorly perfused patient.

Finally, two papers on chronic parenchymal and airway disease. Patrone and colleagues report a multicentre retrospective cohort in Respiratory Medicine of fifty-eight patients with connective tissue disease-associated interstitial lung disease who had objectively progressed — a mean forced vital capacity decline of nearly sixteen percent over a standardised six-month window — despite mycophenolate or cyclophosphamide [6]. After rituximab, mean forced vital capacity rose by close to five percent early and by nearly seven percent at twelve months, with a concordant improvement in diffusing capacity and no toxicity leading to permanent discontinuation. Improvement was descriptively more frequent in inflammatory-predominant phenotypes and in those with better preserved baseline lung volumes. There is no control group and regression to the mean after a steep decline is a genuine concern, but this supports rituximab as rescue therapy chosen earlier rather than in end-stage fibrosis. And in a whole-available-population cohort from four remote Western Australian communities, Laird and colleagues re-screened one hundred and ninety-one Indigenous children six years after initial respiratory screening [10]. Seventeen percent had a chronic respiratory condition, with significantly lower forced expiratory volume and forced vital capacity z-scores than healthy peers. Among children who had chronic wet cough at baseline, thirty-five percent had chronic suppurative lung disease or bronchiectasis six years later — a risk close to thirty times higher than in children without wet cough — and prior hospitalisation for acute lower respiratory infection also carried a markedly elevated risk. Chronic wet cough in a young child is not a benign symptom to observe.

If you only have time for one paper this week, make it the Chest beta-blocker study from Suissa and colleagues [1]. It directly answers a decision you make repeatedly, it sharpens the answer by cardiac indication, and the interaction with long-acting beta-agonist therapy is a genuinely new consideration for how you co-prescribe.

Here are the key takeaways from this week in Pulmonary. First, in COPD with cardiovascular disease, beta-blockers reduce mortality in heart failure and ischaemic heart disease but not in atrial fibrillation or hypertension, and concurrent long-acting beta-agonist use may blunt that benefit. Second, if you have a lung cancer screening programme, combining CT airway biomarkers with smoking history and dyspnoea is a far more efficient way to select who needs spirometry than emphysema extent alone — and remember that more than one in five screening participants have undiagnosed COPD. Third, mid-expiratory flow adds nothing to FEV1 for predicting exacerbations. Fourth, the apparent mortality benefit of severe sleep apnoea in the veterans cohort is a lesson in referral and adjustment bias, not a clinical finding. Fifth, use transcutaneous capnometry for trends only, and be sceptical in circulatory failure. And finally, chronic wet cough in early childhood is a strong predictor of later bronchiectasis and warrants active management, not watchful waiting.

That's your roundup for This Week in Pulmonary. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And a quick rating on Apple Podcasts or Spotify helps other physicians discover the show.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.