• pulmonary alveolar proteinosis
• granulocyte-macrophage colony stimulating factor

The use of recombinant GM-CSF in the treatment of pulmonary alveolar proteinosis (PAP) has been reported recently in Thorax^1,2 in cases with anti-GM-CSF antibodies. We report the successful use of GM-CSF in a patient with no GM-CSF antibodies, which provides further evidence for considering GM-CSF as a possible keystone to the treatment of PAP.

A 35 year old man with a past history of idiopathic marrow aplasia treated with antilymphocytic serum in 1983 and Mycobacterium avium systemic infection presented in February 1999 with a 4 month history of dyspnoea. A diagnosis of PAP was suspected from the HRCT scan which showed fine or coarse irregular lines of attenuation (reticular pattern) involving predominantly the subpleural lung regions and diffuse ground glass opacities, honeycombing, and crazy paving in the lower lung zones. Pao₂ on room air was 6.9 kPa at rest. An open lung biopsy performed in May 1999 confirmed PAP. Between June 1999 and May 2001 the disease worsened, despite four whole lung lavages. At this time spirometric tests showed FEV₁ 1.93 l (55%) and FVC 2.94 l (72%); Pao₂ on room air was 7.3 kPa at rest. The HRCT scan showed an increase in pulmonary infiltrates and an HRCT scan severity score³ of 18. A blood cell count showed lymphopenia at 400/mm and monocytopenia at 70/mm. Anti-GM-CSF auto-antibodies were undetectable in the serum or in the bronchoalveolar lavage fluid.

The past history of Mycobacterium avium infection was related to the patient’s profound T cell lymphopenia that was already present at the time the aplasia was first diagnosed. In contrast, the profound monocytopenia, not present at that time, could have been induced by antilymphocytic globulins which corrected the other vital parameters such as neutropenia, thrombopenia, and anaemia. Indeed, this treatment may indirectly have revealed a defect within the monocyte lineage through deficient GM-CSF secretion by T cells. Based on this hypothesis, daily treatment with subcutaneous recombinant GM-CSF (Leucomax) in a dose of 5 μg/kg/day was started in September 2001. After 3 months of treatment the oxygen could be stopped; Pao₂ on room air was 8.6 kPa at rest. An exercise test showed oxygen saturation of 92% (compared with 86% before treatment). The HRCT scan showed a decrease in ground glass opacities and persistence of the fixed crazy paving lesions; the HRCT severity score was 4. No improvement was seen in the monocytopenia or lymphopenia after treatment. The patient remains stable after more than 1 year of follow up since stopping GM-CSF treatment.

Improvement in pulmonary symptoms with GM-CSF treatment, as occurred in our patient, reinforces the hypothesis of a defect in GM-CSF production in the pathogenesis of PAP. The progenitor cells of our patient did respond normally to GM-CSF in vitro and no anti-GM-CSF antibodies were detected. Later lesions, shown by radiological imaging not to have improved with this treatment, suggest that GM-CSF has to be introduced earlier in PAP to achieve definitive remission.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Laryngoscopy showing (A) glottal stenosis due to spastic adducting muscles while breathing and (B) the absence of paralysis of the abducting muscle.