We present a case of an 8-month-old infant presenting with respiratory failure due to LPI-associated PAP, whom we successfully weaned from daytime respiratory support while receiving initially subcutaneous then inhaled GM-CSF therapy after two courses of WLL. Follow-up at two years of age revealed normal cognitive function and speech development, and, not unexpectedly, a one-year delay in motor development.

Whole lung lavage is the gold standard therapy for PAP since the early 1960s [17], with positive response rates of 60–84% [18]. The procedure removes the lipoproteinaceous material from the alveoli [17]. However, the procedure is difficult to perform, is expensive and is not well tolerated by children [19]. Our patient’s course confirmed the challenges WLL therapy presents. Initially the patient was unstable and did not tolerate WLL while on mechanical ventilation. Therefore, the decision was made to place her on VV ECMO to increase the chance of success. In addition, her vessels were relatively gracile and difficult to cannulate. Veno-venous ECMO modality was chosen as her cardiac function was normal and thus only pulmonary bypass was needed. Also, VV ECMO has been associated with lower rate of complications compared to venoarterial ECMO [20]. Finally, even though WLL is classically performed through gravity with the patient being positioned in the reverse Trendelenburg and Trendelenburg positions, we were unable to drain the injected lavage fluid from the lungs by positioning alone. Therefore, we applied gentle airway suctioning using a syringe. Importantly, no complications occurred during the procedure and following the second round of WLL which was tolerated off VV ECMO, the patient could be weaned to non-invasive ventilation one day after the procedure.

A number of studies have evaluated the effect of the use of GM-CSF in autoimmune PAP. In cases of autoimmune PAP, lipids and proteins accumulate within the alveoli since alveolar macrophages are unable to catabolize surfactants due to disrupted GM-CSF signaling by neutralizing GM-CSF autoantibodies [9]. Granulocyte–macrophage colony-stimulating factor enhances the ability of macrophages to perform this function; therefore, GM-CSF replacement improves PAP pathology [21,22,23]. However, in non-autoimmune PAP, there is no rationale based on current science for the use of GM-CSF. Specifically, in LPI Barelli et al. found an unaltered GM-CSF signaling pathway, and since LPI patients did not have high levels of ant-GM-CSF antibodies, the impaired function and phagocytic activity of the macrophages was attributed to be intrinsic to the SCL7A7 phenotype [12]. Surprisingly, recent case reports observed the resolution of PAP using GM-CSF in patients with LPI-associated PAP as well [13,14,15]. The rationale of this phenomenon is unknown, but the increase in the number and the activity of alveolar macrophages in the alveolar fluid by GM-CSF might be a reasonable explanation [13]. Additional hypotheses include that GM-CSF therapy might play a role in increasing in vivo gene expression involved in surfactant catabolism, and also that GM-CSF-dependent increase of SLC7A7 mRNA might increase y + LAT1 protein that ameliorates macrophage function [15].

The deterioration of our patient to maximal conventional therapy after an empirical course of high dose corticosteroid, azithromycin and hydroxychloroquine, and the PAP reoccurrence after the first WLL prompted us to consider GM-CSF administration after the second WLL, based on the beneficial effect of GM-CSF in LPI-associated PAP available in recent case reports [13,14,15]. Initially, GM-CSF was administered subcutaneously and then the drug was switched to the inhaled route of administration. Although subcutaneously administered GM-CSF was found to be effective in patients with autoimmune PAP, minor complications such as fever, fatigue, headache, and injection site complications have been reported [24]. Our patient experienced fever and leukocytosis without proven infectious origin during the administration of subcutaneous GM-CSF. Meta-analyses suggest that, compared to subcutaneous GM-CSF therapy, inhaled GM-CSF is associated with a higher response rate, may act more peripherally and improves alveolar partial pressure of oxygen to a greater extent [24]. In agreement with these observations [24], our findings also suggest that compared to subcutaneous GM-CSF, inhaled GM-CSF might have been more beneficial in mucus mobilization and had fewer side effects. Importantly, it appears to have been more effective since we were able to completely wean the patient off respiratory support during daytime.

Of note is that, although rare, spontaneous remission of PAP might occur. Nevertheless, spontaneous remission of PAP is described mainly in autoimmune PAP, accounting for 5–10% of cases [9]. On the other hand, the early onset of pulmonary disease in LPI is associated with a severe prognosis [25], and specifically the early development of PAP was found to be significantly associated with death in a retrospective study investigating the long-term management of LPI [6]. It is therefore logical to assume that the marked improvement in our patient’s clinical course is attributed to the completion of the two WLLs and to the introduction of the GM-CSF therapy. This case alongside with the recently published case reports [13,14,15], gives further support to the notion that GM-CSF therapy might be of benefit in patients with LPI-associated PAP. Additional studies are needed to clarify the exact mechanism of resolution of PAP using GM-CSF in patients with LPI-associated PAP.

In conclusion, our case gives support to the recently published case reports that GM-CSF therapy might be of benefit in the treatment of LPI-associated PAP, and might represent a long-term therapeutic option following initial induction of remission with WLL. Further research is warranted to elucidate the exact therapeutic mechanism of GM-CSF therapy in patients with LPI-associated PAP.