Recommendations from the Medical Education Editor

Chronic thromboembolic pulmonary hypertension (CTEPH) is a condition of pulmonary microvascular arteriopathy secondary to thrombotic occlusion of pulmonary vessels, with associated right ventricular dysfunction and elevated pulmonary arterial pressures.1 After assessment by a multidisciplinary team at an expert centre, pulmonary endarterectomy (PEA) is the treatment of choice for patients suitable for surgery. Unfortunately, not all patients are considered operable due to the distribution of thromboses or medical comorbidities.2 In this group of patients, balloon pulmonary angioplasty (BPA) and medical therapies are alternative strategies.3 Riociguat has an established role in non-operable CTEPH or those patients with persistent pulmonary hypertension after PEA but a role for agents more commonly used for pulmonary arterial hypertension (PAH), such as phosphodiesterase-5 inhibitors (PDE5i), prostacyclin analogues and endothelin receptor antagonists (ERA), is less clear.4 In the July 2024 issue of Respirology, Xia et al. report their evaluation of off-label use of PDE5i, ERA and prostacyclin analogues between 2009 and 2018 in a registry of patients from 18 centres in China. Patients who did not undergo PEA or BPA and were not receiving riociguat or treprostinil were included.5 Three hundred and forty-seven patients were enrolled in the study, of whom 212 received off-label treatment, while 135 did not (control). Targeted medications used included sildenafil (41.5%), tadalafil (24.5%), vardenafil (3.8%), bosentan (11.8%), ambrisentan (4.2%), beraprost (2.8%) and iloprost (2.4%). Nineteen patients were treated with dual agents. The median age was 56, with a marginal male predominance (51.6%). Mean pulmonary artery pressure was 50.7 ± 13.3 mm Hg, and the treatment group had more severe haemodynamic dysfunction at baseline. Most patients were also treated with warfarin, with only 16 receiving direct oral anticoagulants. In this study, the treatment group had significantly better survival than the control group despite having worse haemodynamic parameters at baseline. At 5 years, survival was 71% in the treatment group versus 55% in the control (Figure 1). Multivariate analysis demonstrated a significant survival benefit from receiving off-label medications (hazard ratio of 0.512, 95% CI 0.337–0.779, p = 0.002). Patients treated within the first 3 months of enrolment had significantly improved survival in the propensity-matched cohort after multivariate analysis compared with the control group (HR 0.453, 95% CI: 0.258–0.794, p = 0.006). The low rate of riociguat use in this cohort makes the role of alternative agents in countries where riociguat is available less clear, and the authors recommend using riociguat where possible. Other studies evaluating the use of PAH agents in CTEPH must clarify their role in patients receiving guideline-directed therapy. Clinicians caring for patients with community-acquired pneumonia (CAP) likely will be very familiar with the pros and cons of various prediction tools to aid risk stratification.6 My selection from the August 2024 issue seeks to answer a different question: in an ageing population, can we predict long-term prognosis to identify patients with community-acquired pneumonia who might benefit from advanced care planning discussions?7 The study includes patients over 65 years old who were admitted to 10 Japanese hospitals with CAP between 2018 and 2021. Patients with a diagnosis of COVID-19 infection were excluded, as were those already receiving palliative care. The chosen endpoint was 180-day mortality. Four hundred fifty-two patients were enrolled, of whom 21 were lost to follow-up, and several had missing data. The analysed cohort was 399 patients, of whom 68 (17%) had died at 180 days. The median patient age was 82 (IQR 76–88), with a 60:40 male-to-female ratio. Three hundred twenty-two cases were used to develop a prognostic score, and the remaining 77 cases were used for validation. After multiple logistic regression analysis, three factors were significantly associated with 180-day mortality: reduced Eastern Cooperative Oncology Group performance status (Odds Ratio [OR] 11.78), hypoalbuminaemia ≤2.5 g/dL (OR 5.28) and dementia (OR 3.15). Using these three factors, a scoring model was developed using two points for low performance status and one each for hypoalbuminaemia and dementia. In the 322-patient derivation cohort, the area under the receiver operating characteristic curve (AUC) was 0.8504, with a sensitivity of 95% and a negative predictive value (NPV) of 98% for a cut-off of 2 points out of 4. In the validation cohort, the AUC was 0.7925, sensitivity 92% and NPV 98%. The appropriate institution of symptom-driven rather than aggressive pneumonia-specific care in elderly patients with significant comorbidities is a worthwhile goal, and tools like this one might augment clinical judgement after further evaluation in larger datasets and other populations. Although the relationship between obstructive sleep apnoea (OSA) and cardiovascular risks appears well-established, OSA severity assessment methods like the apnoea–hypopnoea index (AHI) do not fully explain the relationship, and it has been challenging to prove that treatment reduces cardiovascular (CV) risk.8 It does appear that indices of hypoxia might be more useful for CV risk stratification, and the selected article from the September 2024 issue of Respirology explores the value of several novel polysomnographic indices.9 These include a reconstructed hypoxic burden (rHB), the Sleep Breathing Impairment Index (SBII) and the proportion of sleep time with the duration of respiratory events causing oxygen desaturation of more than 3% (pRED_3p). The patient data were derived from the Sleep Heart Health Study, and negative control groups were identified for both the CV mortality and morbidity groups. OSA severity indices included AHI, oxygen desaturation index and the three aforementioned novel hypoxic indices. Evaluated covariates included age, race, gender, body mass index, smoking status, hypertension, diabetes, the use of lipid-lowering drugs, coexisting chronic obstructive pulmonary disease and a previous history of cardiovascular events. A series of models incorporating the OSA severity indices and covariates were created and compared. In the CV mortality dataset, the SBII demonstrated statistical significance across its second to fifth quintiles and the pRED_3 across the third to fifth quintiles. SBII performed better than AHI, ODI, rHB, and pRED_3p in the adjusted model. For the CV morbidity dataset, the adjusted model using pRED_3p demonstrated the best performance. The authors suggest that the SBII model, which incorporates both sleep event duration and oxygen desaturation area, is likely particularly relevant for CV mortality. The pRED_3p model, which might capture events with a lower desaturation area, better predicted CV morbidity. Finally, the authors have made it possible to calculate these indices using an online tool. The ongoing research into and identification of polysomnographic indices that more accurately predict cardiovascular outcomes is critical to allow clinicians to provide the best evaluation of an individual patient's risk profile and guide treatment strategy. Perhaps with more modern indices of sleep disruption, we might clarify for which patients the treatment of OSA will reduce CV risk. None declared.