In this issue of CHEST (see page 376), Zvezdin and colleagues9 overcome this major limitation of previous studies by reporting on an autopsy series of 43 consecutive patients who were admitted to the hospital for an acute exacerbation and died within 24 h of hospitalization. Because the local jurisdiction mandates that all patients who die within this 24 h window undergo an autopsy, these investigators were able to evaluate all decedents, thus avoiding selection bias. Although the study was small, there were several notable findings. First, they found that the leading cause of death was, surprisingly, not respiratory failure but cardiac failure, accounting for 37% of all deaths, followed by pneumonia and thromboembolic events, each contributing 28% and 21%, respectively, to total mortality. Only 14% of the deaths could be primarily attributed to respiratory failure secondary to COPD. Interestingly, there were six patients, who were (incidentally) found to have carcinoma at autopsy, but all of them died from another cause. Second, because a majority of these patients were smokers and had multiple comorbid conditions, none of the routinely collected demographic or clinical data, including patient symptoms, and the findings of general blood tests, chest radiographs, and ECGs, could reliably predict the causes of death before the terminal event. It is plausible that the wider use of more sensitive instruments such as CT scanning or blood tests including serum troponin or brain natriuretic peptide levels may have increased the diagnostic yield of these terminal events. Third, a majority of patients in this study died of causes that were potentially modifiable but only if they had been diagnosed early and treated promptly. In this study, although all patients received appropriate therapies for their lung disease,1 only one-third of patients who eventually died from thromboembolic disease received anticoagulant therapy (likely due to delayed diagnosis). It is unclear how many of the patients who died from cardiac failure had received anti-heart failure medications, including diuretics, statins, angiotensin-converting enzyme inhibitors, or β-blockers, which are life preserving even in patients with COPD,10,11 but based on prior literature on COPD,10,11 it is likely that very few of these patients would have received these medications. This begs the following question: could some of these deaths have been averted had the diagnosis of thromboembolic disease or cardiac failure been made earlier in the course of their illness? Clinical intuition, along with prior clinical studies,12,13 suggest that this is likely.