September 2026 Research Update

This month added one real-world comparison of 40 patients suggesting that, for some people undergoing whole-lung lavage, having ECMO support during the procedure may be linked to bigger gains in breathing tests and oxygen levels than lavage without ECMO. [1]

In this 40-patient retrospective study, the group that received ECMO-assisted whole-lung lavage plus nebulized GM-CSF improved more on lung function measures and oxygenation (for example, higher blood oxygen levels) than the group that had whole-lung lavage plus GM-CSF without ECMO. The ECMO group also showed larger drops in inflammation blood tests and blood lipid levels, and did better on walking distance and daily activity scores. [1]

Because the comparison was based on past medical records rather than a randomized trial, the study cannot fully separate the effect of ECMO itself from differences between the groups. Even so, it makes the picture clearer that some centers are combining ECMO with lavage and inhaled GM-CSF, and that this approach may be linked to broader improvement across oxygen, function, and inflammation measures. [1]

For patients, this mainly adds context about how whole-lung lavage may be delivered in more severe cases, and what kinds of improvements have been reported when ECMO is used to support the procedure. It may matter now for understanding care discussions at specialized hospitals, but the evidence is not strong enough to say how much ECMO itself changes outcomes for any one person. What remains uncertain is who benefits most and what the risks and tradeoffs are, since safety details and reasons for choosing ECMO were not described. [1]

This month, the case reports and small case series broadened the picture of pulmonary alveolar proteinosis by showing how different the underlying situations can be. Across these reports, PAP appeared alongside long-term immune changes in autoimmune PAP, serious lung infections (nocardiosis), and a medication-linked secondary form that improved after changing immune therapy. [2] [3] [4]

In three people with autoimmune PAP followed for 11 to 24 years, anti-GM-CSF autoantibodies gradually fell in an exponential pattern and eventually became undetectable after about 7 to 16 years, even though treatments varied. This supports the idea that autoimmune activity in aPAP may quiet down over long periods in some individuals, although the report did not show that antibody changes always matched day-to-day symptoms. [2]

A four-patient series described pulmonary nocardiosis occurring in adults with PAP, with diagnosis supported by lung fluid cultures and treatment guided by antibiotic susceptibility testing. Outcomes varied from improvement and microbiological cure to death, underscoring that infections can be an important competing problem when PAP is present and that correct identification of the infection can shape treatment choices. [3]

One report described secondary PAP developing in a person with chronic graft-versus-host disease while taking ruxolitinib, with rapid clearing of lung imaging changes after switching to belumosudil. Another report described a child with Still's disease-related lung involvement that included PAP and pulmonary hypertension, where baricitinib plus cyclosporine was followed by long-lasting clinical improvement and eventual stopping of corticosteroids. Both reports point to immune pathways as potential drivers in some non-autoimmune settings, but they are single-patient stories. [4] [5]

For patients and caregivers, these reports mostly help explain why PAP can look and behave differently from person to person, depending on the cause and the immune system context. Some findings may matter now in awareness terms, such as the possibility of serious lung infection in someone with PAP, or the possibility of a medication-related secondary form in people on strong immune therapies, but none of these reports can predict what will happen for others. What remains uncertain is how common these situations are and whether the same outcomes would occur in larger groups, since the evidence here comes from small numbers and single cases. [2] [3] [4] [5]