CHEST Roundup: Diabetes in RA Patients Ups Risk of Interstitial Lung Disease

BOSTON — Some of the latest research advancements in the field of pulmonology presented at the CHEST 2024 annual meeting, hosted by the American College of Chest Physicians, included dupilumab’s (Dupixent) positive effect on quality of life for patients with chronic obstructive pulmonary disease; SGLT2 inhibitors’ ability to lower mortality in pulmonary arterial hypertension; and an increased risk for pulmonary embolism with checkpoint inhibitors for treating metastatic lung cancer.

Below are some additional scientific highlights from this year’s meeting.

Diabetes Increases ILD Risk in Patients With RA

Type 2 diabetes (T2D) appeared to double the risk of interstitial lung disease (ILD) in patients with rheumatoid arthritis (RA), according to a retrospective cohort study.

In propensity score-matched cohorts of 121,046 hospitalized patients, those with RA and T2D had an ILD prevalence of 2.25%, compared with 1.11% among those with RA but no diabetes (OR 2.023, 95% CI 1.843-2.219, P

Moreover, diabetes was linked with a nearly fourfold increase in pulmonary fibrosis among patients with RA (OR 3.704, 95% CI 3.298-4.161, P

When Khattar and co-investigators looked at other lung conditions, they found that patients with RA who also had T2D had significantly elevated risks for pulmonary arterial hypertension (OR 1.4, PP

The study was the first to explore the association of T2D and pulmonary outcomes in the RA population on such a large scale, Khattar stated. He hypothesized that T2D in the setting of RA may compound chronic inflammation, fibroblast activation and collagen deposition, neurologic dysfunction, and endothelial damage to result in “diabetic pneumopathy.”

“Screening guidelines for pulmonology pathologies in the RA population are still not established,” Khattar said.

“Due to the high prevalence of T2D in the RA population, clinicians should consider having a low threshold for screening high-risk patients for ILD or pulmonary arterial hypertension,” he said. “More studies on the role of aggressive glycemic control in this population are needed.”

Study limitations included its retrospective design, potential for residual confounding, and that it only looked at data from hospitalized patients.

Nalbuphine Relieves Cough in IPF

The oral opioid receptor agonist/opioid receptor antagonist nalbuphine reduced cough bout duration, time spent coughing, and cough intensity in idiopathic pulmonary fibrosis (IPF), according to a post-hoc analysis of the phase IIa CANAL trial.

The small study (N=38) demonstrated that extended release nalbuphine increased relief-of-cough time by 21% from baseline, compared with 7.4% with placebo (P=0.0019), reported Alyn Morice, MD, of the University of Hull in Kingston-upon-Hull, England.

The novel opioid also reduced cough time by 66.5% versus 26.4% with placebo (P=0.0039). In addition, nalbuphine also reduced cough intensity by 26.5% compared with an increase of 19.3% with placebo, for a relative change of 38.4% (P=0.0049).

“I don’t need to tell this audience how dreadful the cough of IPF is,” Morice said, noting that 85% of patients with IPF experience chronic cough, for which there are no approved treatments.

In the initial primary analysis of the CANAL trial, researchers had looked at the effects of nalbuphine on daytime cough frequency and 24-hour cough frequency. The current analysis evaluated different cough parameters to further quantify the effects of nalbuphine. Researchers defined relief-of-cough duration as any cough-free period of at least 15 minutes. Cough time was defined as total observation time minus relief time, and cough intensity was defined as number of coughs during coughing time.

Adverse events were consistent with the known safety profile from previous trials, according to a press release from drugmaker Trevi Therapeutics. “Nalbuphine is mostly a κ-agonist and is actually a blocker of the µ-opiate receptor so we have a completely different safety and tolerability profile with nalbuphine compared with morphine,” Morice explained.

The most frequently reported adverse events were nausea, dizziness, anxiety, constipation, vomiting, dry mouth, headache, somnolence, dyspnea, decreased appetite, and fatigue.

Addressing Pediatric Lung Allograft Discards With AI

Three machine-learning algorithms were able to accurately predict whether a pediatric lung allograft would be discarded, according to a retrospective analysis.

Of five machine-learning models, the Logistic Regression, Random Forest, and XGBoost models had areas under the curve (AUC) above 0.9, outperforming the Naive Bayes and Decision-Tree Machine-Learning models which had AUCs of about 0.8, reported Tahir Malik, MD, of Mount Sinai Hospital in New York City.

“When making transplant decisions, clinicians can use machine-learning techniques to make robust quantitative predictions and gain valuable insight into which factors may contribute to discard,” Malik said.

The Logistic Regression method achieved an AUC of 0.908 (95% CI 0.895-0.920). In this model, the three most important variables were the donor’s history of cigarette smoking, donor hepatitis C antibody status, and donor chest x-ray.

The Random Forest method achieved an AUC of 0.911 (95% CI 0.900-0.923). In this method, the three most significant variables used to predict discard were the donor’s arterial oxygen partial pressure/fractional inspired oxygen (PO2:FiO2) ratio, donor age, and donor height.

The XGBoost model had an AUC of 0.901 (95% CI 0.888-0.914), with PO2:FiO2 ratio, donor height, and donor age being the most important variables.

From 2012 to 2023, 3,500 pediatric patients died while waiting for transplantation, Malik explained. Over those same 11 years, 865 pediatric patients became too sick to be transplanted. In addition, 7,500 pediatric lung allografts were discarded — 77% of all pediatric lung allografts that were donated. Decreasing donor allograft discards could potentially decrease waitlist mortality for lung transplant patients, he said.

Malik and colleagues retrospectively analyzed 9,726 donors from 2011 to 2023 with data available from United Network for Organ Sharing (UNOS), focusing on individuals ages 18 years or younger. They partitioned the data into a training set (70%, n=6,802) and a testing set (30%, n=2,918). Models were constructed using the training data and then evaluated for their predictive performance using the different machine-learning techniques for predicting pediatric lung allograft discard.

Next steps are to create a discard risk score to predict the relative risk of allograft discard, Malik stated.

Please enable JavaScript to view the comments powered by Disqus.