Pulmonary alveolar proteinosis (PAP) was first described in 1958 by Samuel H. Rosen et al. Since that time, clinicians' understanding of this rare lung disease has improved dramatically. Initial reports of this disease described it as respiratory failure secondary to over-production of surfactant proteins within the alveoli. It was believed to be a consequence of inhaled environmental irritants or infectious agents and was initially called acquired or idiopathic PAP. Practitioners now understand that there are 3 separate pathways to the development of surfactant accumulation within alveoli: congenital, secondary, and autoimmune. All 3 of these pathways result in decreased clearance of surfactant, rather than increased production. Autoimmune PAP is the most common pathophysiologic mechanism accounting for 90% of documented cases. Autoimmune PAP is initiated by immunoglobulin (Ig)-G anti-granulocyte macrophage colony stimulating factor (anti-GM-CSF) antibodies, which decrease functional alveolar macrophages. Secondary PAP lacks anti-GM-CSF antibodies but has decreased functional macrophages secondary to hematological malignancies (myelodysplastic syndrome, chronic myelogenous leukemia, among others) or primary immunodeficiency diseases (common variable immunodeficiency, DiGeorge syndrome, among others). Congenital PAP is the least common and results from genetic mutations in GM-CSF receptor proteins or surfactant proteins.
Autoimmune PAP is a direct consequence of antibodies against GM-CSF. There has been speculation that cigarette smoke or infectious diseases stimulate the development of these autoantibodies due to a high prevalence of smoking and infections in patients with PAP. However, no causal link has been found between cigarette smoke and autoimmune PAP. Additionally, no causal link has been found between infections and autoimmune PAP.
Secondary PAP is caused by any disease that reduces the functionally effective alveolar macrophage population. In one retrospective analysis, 34.1% of secondary PAP was associated with myelodysplastic syndrome, and 15.2% was associated with chronic myelogenous leukemia. Acute myeloid leukemia also has a well-established association with PAP. Less frequent are associations between PAP and acute lymphoid leukemia, lymphoma, and myeloma. Few case reports have associated PAP with non-hematological cancers such as lung cancer, mesothelioma, and glioblastoma. Rare reports have associated PAP with immunodeficiencies such as severe combined immunodeficiency, agammaglobulinemia, adenosine deaminase deficiency, common variable immunodeficiency, DiGeorge syndrome, dermatomyositis, rheumatoid arthritis, Behcet’s disease, AIDS, GATA2 deficiency, and organ transplantation. Secondary PAP has also been associated with inhalation of several environmental exposures. These environmental exposures include silica, talc, cement, kaolin, aluminum, titanium, indium, and cellulose. Several animal studies have induced PAP following inhalation of aluminum, fiberglass, indium, nickel, quartz, silica, and titanium. French analysis of PAP patients found that 39% had an environmental exposure of cement, cereal dust, copper, epoxy, paint, polyvinyl chloride, silica, welding, wood dust, zirconium, or smoke. A report from Japan showed that 23% of patients with PAP had significant environmental exposures. A Korean study reported 53% of PAP patients were smokers and 48% had exposure to dust. These dust exposures included metal (26.5%), stone or sand (20.6%), chemical or paint (17.7%), farming dust (14.7%), diesel (14.7%), textile (2.9%), and wood (2.9%). One report suggests a link between indium inhalation and anti-GM-CSF antibodies, suggesting these toxic inhalations might induce autoantibodies.
Hereditary dysfunction in one of many proteins responsible for surfactant regulation causes Congenital PAP. This includes mutations in the GM-CSF receptor alpha-subunit or beta-subunit, surfactant protein B, surfactant protein C, ATP-binding cassette 3, NK2 homeobox 1, or the lysinuric protein intolerance disease.
PAP is a rare lung disease. Prevalence has been reported to be from 3.7 to 40 cases per million depending on the country. The incidence has been estimated to be 0.2 cases per million. Autoimmune PAP accounts for approximately 90% of cases, while 4% is secondary PAP, 1% is congenital PAP, and undetermined PAP-like disease represents the remaining 5%. Smoking is reported in 53% to 85% of PAP patients.
In autoimmune PAP, males are more commonly affected than females at a 2:1 ratio. Median age at time-of-diagnosis is reported from 39 years to 51 years, although ages range from newborn to 72 years old.
Secondary PAP has been reported with a slightly younger median age at time-of-diagnosis; 37 years to 45 years. Gender-ratio is also different in secondary PAP, with a 1.2:1 male to female ratio.
Surfactant is a key component within alveoli that prevents end-expiratory collapse by decreasing alveolar surface tension at the air-liquid interface. Surfactant is also crucial for host defense at the alveolar level by opsonizing microbes and signaling innate defense mechanisms. It is synthesized and secreted by alveolar type II epithelial cells. Both type II epithelial cells and alveolar macrophages are involved in the breakdown and clearance of surfactant. Surfactant is made of approximately 90% lipid, most of which is phosphatidylcholine. The next 9% of surfactant is protein, and the remaining 1% is carbohydrate. There are 4 main subtypes of surfactant known as surfactant proteins A, B, C, and D. Each of these molecules corresponds to the genes SFTPA, SFTPB, SFTPC, and SFPTD. Mutations in these genes lead to malformed surfactant which accumulates within type II alveolar epithelial cells, causing ineffective clearance of surfactant from the alveoli and cell death. When PAP was first described in 1958, patients were noted to have eosinophilic materials in their alveoli that were filled with lipids and some proteins and carbohydrates. Since that time, it has been confirmed that the eosinophilic material is indeed surfactant accumulation.
GM-CSF is a crucial cytokine involved in host defense of pulmonary bacterial, viral, and fungal infections. Antibodies that interfere with this cytokine can lead to increased opportunistic infections.
In 1994, researchers discovered that GM-CSF is important in the pathophysiology of PAP when GM-CSF knock-out mice demonstrated PAP-like disease. GM-CSF is a 23-kD growth factor cytokine. GM-CSF binds to various cells via cell-surface receptors to induce biological effects of differentiation, proliferation, and survival. The affected cells include monocytes, neutrophils, dendritic cells, macrophages, and type II alveolar epithelial cells. When anti-GM-CSF antibodies are formed, as with autoimmune PAP, they bind with high affinity to GM-CSF and block their biological potential. This leads to dysfunctional alveolar macrophages which are unable to clear surfactant. Instead, these macrophages accumulate large lysosomes filled with surfactant and become defective in normal anti-microbial activities such as phagocytosis, chemotaxis, superoxide production, pathogen-recognition receptor expression, cytokine release, and cellular adhesion. Furthermore, lymphocytes and neutrophils are less functional which leads to an increase in opportunistic infections. In GM-CSF knock out mice, messenger RNA of the surfactant proteins was not increased, indicating that surfactant overproduction is not involved in the development of PAP. This model instead demonstrated decreased clearance of surfactant as evidenced by the accumulation of lipoprotein in macrophages and eosinophils. Under ultra-microscopy, the lipoprotein material within the eosinophils and macrophages consists of tubular myelin, lamellar bodies, surfactant phospholipids, and surfactant proteins.
GM-CSF levels are normal in bronchoalveolar lavage fluid (BALF) in autoimmune PAP. However, elevated IgG anti-GM-CSF levels are observed in BALF and serum in autoimmune PAP patients. Elevated levels are not seen in other lung diseases.
In secondary PAP, it is believed that the underlying hematological disease causes either a reduction in the number of alveolar macrophages or a reduction in the functional quality of these macrophages. At least 3 patients with acute myeloid leukemia have been found to have a defect in the expression of GM-CSF receptor on alveolar macrophages. Following successful treatment of underlying leukemia, the receptor expression returned to normal in these individuals.
In congenital PAP, surfactant accumulation is a consequence of a genetic mutation resulting in dysfunctional GM-CSF receptor activation. In humans, the GM-CSF receptor is composed of an alpha and a beta subunit, each which corresponds to the genes CSF2RA and CSF2RB respectively. Mutations in CSF2RA have been reported in young children (newborn to 9 years) with an autosomal recessive inheritance pattern with incomplete penetrance. Mutations in CSF2RB are less common and have been reported from newborns up to 36 years old. Either mutation leads to an increased GM-CSF level in BALF and serum. Congenital PAP has also been described in lysinuric protein intolerance, which is an autosomal recessive disease caused by a mutation in SLC7A7 and leads to defective transport of amino acids at the epithelial membrane. This disease presents at an early age with gastrointestinal, renal, and less commonly pulmonary manifestations. SLC7A7 mutations lead to dysfunctional arginine transport and dysfunctional alveolar macrophages. Although SLC7A7 is a target of GM-CSF, which upregulates it, the GM-CSF pathway does not seem to be altered in this disease or lead to PAP. Mutations in surfactant protein B, surfactant protein C, ATP-binding cassette 3, or NK2 homeobox can all lead to dysfunctional surfactant release from type II epithelial cells and dysfunctional clearance from the alveoli.
Whatever the inciting factor, all of these pathways lead to the same end: accumulation of lipoproteinaceous material in the alveoli due to dysfunctional clearance by alveolar macrophages or type II epithelial cells.
The initial gross description of lungs from PAP patients during autopsy revealed multiple yellow-grey nodules measuring from a few millimeters to 2 centimeters throughout the lung. These nodules were firm, but upon dissection, they leaked a thick, milky substance with palpation. General light-microscopy with hematoxylin and eosin staining showed preserved interalveolar septa with lipoproteinaceous material filling the alveoli and some bronchioles. The epithelial cells lining the alveoli and bronchioles were damaged and sloughed from the wall, while large mononuclear cells were present. A more recent examination of surgical lung biopsies demonstrated persevered lung parenchyma with peribronchial lymphocytic infiltrations and alveoli filled with macrophages and amorphic eosinophilic material that stains periodic-acid-Schiff (PAS)-positive. Immunohistochemical staining of this material confirms surfactant protein. BALF cellularity is often increased with a predominance of lymphocytes. Cytological examination of the BALF shows foamy macrophages which contain eosinophilic granules and amorphic material that stains PAS-positive. High-powered microscopy of BALF, although not routinely done, shows lamellar bodies that resemblance myelin. The rare cases of congenital PAP may demonstrate varying degrees of interstitial remodeling and type-II epithelial cell hyperplasia on biopsy.
Clinical presentation of PAP varies from indolent to emergent and symptoms are often non-specific. Dyspnea is the most common presenting complaint; present in 39% of patients in one report. A cough was reported in 21% of patients. Hemoptysis, fever, and chest pain are rare complaints of autoimmune PAP and should prompt consideration of another diagnosis. Fever may be present in 24% of secondary PAP patients due to concomitant hematological malignancies or opportunistic infections. Thirty-three percent of patients may be asymptomatic at the time of presentation.
A majority of patients are smokers (53% to 85%). Many report occupational exposures (39% to 48%). Even though 90% of PAP cases are autoimmune, only 1.7% of patients have other identified autoimmune diseases. Physical examination is frequently normal; however, patients can present with cyanosis (25% to 30%), clubbing (30%), or inspiratory crackles (50%).
Given that the clinical presentation of PAP is nonspecific, diagnosis of this disease demands appropriate serological, radiological, and bronchoscopic evaluation.
Chest radiography may demonstrate bilateral alveolar opacities in a peri-hilar and basilar distribution without air-bronchograms. This is sometimes referred to as a “batwing distribution,” or may resemble pulmonary edema without cardiomegaly or pleural effusions. Computed tomography reveals intralobular thickening and diffuse ground-glass opacities. This pattern is often referred to as “crazy paving.” "Crazy paving" is highly suggestive of PAP although not specific or sensitive enough to confirm the diagnosis. Lower-lobe predominance has been reported in 22% of cases. Pulmonary nodules, mediastinal adenopathy, and focal parenchymal consolidations are absent in PAP and should prompt consideration of a different diagnosis. Secondary PAP is less likely to have interlobular septal thickening (only 33.3% of cases). Congenital PAP is also less likely to present with interlobular septal thickening and additionally may present with lung cysts.
Several biomarkers have been studied in PAP, including surfactant protein A, B, and D levels, cytokeratin 19, serum carcinoembryonic antigen, serum lactate dehydrogenase, GM-CSF levels, anti-GM-CSF antibodies, and KL-6. Serum lactate dehydrogenase is elevated in 50% of PAP patients and serum carcinoembryonic antigen is also often elevated. Serum levels of the surfactant protein A, B, and D are all increased in PAP and seem to be associated with disease severity. Overall, these biomarkers have not proven sensitive or specific for diagnosis. Testing for IgG anti-GM-CSF antibodies is the only clinically relevant biomarker to date. Detection of these antibodies is done via enzyme-linked immunosorbent assay (ELISA) which is the gold-standard and has been validated. Serum antibody levels can also be detected via latex-agglutination test with 100% sensitivity and 98% specificity. An anti-GM-CSF antibody level of 2.8 micrograms/ml or greater is abnormal and is consistent with PAP.
Pulmonary Function Testing
Pulmonary function testing is not necessary for diagnosis, nor is it specific for PAP; however, the most common pattern on spirometry is a restrictive pattern. Spirometry may be normal in 30% of patients with PAP and may have a mixed obstruction and restrictive pattern in those who are smokers. A significant reduction in diffusion capacity and an increase in the alveolar-arterial gradient are the most common findings on pulmonary function testing in patients with PAP.
When PAP is suspected, bronchoscopy with bronchoalveolar lavage is the gold standard for diagnosis. In approximately 75% of suspected cases, BALF examination will confirm a diagnosis. The lavage return will often appear milky and opaque. BALF cytological examination will reveal large foamy macrophages with amorphous material that stains PAS-positive. BALF cellularity is typically lymphocyte-predominant. Secondary PAP may be associated with opportunistic infections which can lead to mixed cellularity, microbial growth, and potential for a missed diagnosis of PAP (only 61.9% of secondary PAP cases diagnosed with bronchoscopy).
Biopsies are not necessary for a diagnosis of PAP but can be helpful. Open-lung biopsy or video-assisted thoracic surgery is rarely done but will demonstrate PAS-positive lipoproteinaceous material. Transbronchial biopsies were found to positively diagnose PAP in 42% of cases in one study. There is at least one report of a bronchoscopic cryobiopsy resulting in a diagnosis of PAP in a patient with otherwise normal BALF and negative for anti-GM-CSF antibodies.
Whole Lung Lavage
The current standard of care for autoimmune PAP is whole lung lavage (WLL). The decision to treat is based upon signs of poor gas exchange and symptoms of respiratory distress. Patients with mild dyspnea or no symptoms can do well with supportive care and monitoring of pulmonary function tests and chest imaging. Whole lung lavage should be considered in patients with dyspnea at rest, resting PaO2 less than 65 mm Hg, resting alveolar-arterial gradient greater than 40 mm Hg or oxygen desaturations on 6-minute walk test. WLL was first performed in 1961. Today, the procedure is done under general anesthesia with a double lumen endotracheal tube. One lung is ventilated while the contralateral lung is lavaged with warm saline (37 C) and then allowed to drain to gravity. This process is continued until the returning fluid is clear. An average of 15 liters of saline is required. The procedure is repeated on the contralateral lung after 48 hours or more. A global analysis of WLL therapy showed that centers differ in the time between lavaging the contralateral lung (50% wait 1 to 2 weeks), the choice of lung to begin with, the position of the patient (50% supine, 50% lateral decubitus), chest percussion use during the procedure, and amount of saline used (5 to 40 L). Retrospective studies of WLL have shown several parameters improve after therapy. Symptomatic improvement occurs in 85% of cases. Radiographic findings also improve. Forced expiratory volume in 1 second (FEV1) improves by 0.26 L, forced vital capacity (FVC) improves by 0.5 L, diffusion capacity improves by 4.4 mL, and alveolar-arterial gradient improves by 30 mm Hg. Retrospective data also suggests a mortality benefit (94% at 5 years with WLL compared to 85% at 5 years without WLL). WLL has not been verified in randomized prospective trials. Approximately 50% of patients will require a second WLL procedure. The average duration of benefits following WLL is 15 months. A review found that 70% of patients underwent WLL within 5 years of diagnosis and that on average patients required two whole lung lavages.
GM-CSF Replacement Therapy
Clinical trials for GM-CSF replacement therapy were performed in the late 1990s. Subcutaneous injections of GM-CSF showed a positive response in 48% of a small cohort of 25 patients. Improvement following GM-CSF injections was much slower than the standard WLL, and therefore, this therapy has fallen out of favor. In a small trial of 12 patients of inhaled GM-CSF therapy, 11 of patients (91%) showed improvement. However, a larger subsequent trial of 35 patients resulted in only 68% of patients showing improvement. These therapies are safe and show some improvement in autoimmune PAP; they were ineffective in congenital PAP and inferior to WLL. These are currently considered alternative therapies to WLL.
Systemic corticosteroids were trialed in autoimmune PAP but were ineffective, and they were found to increase the risk of pulmonary infections. Plasmapheresis has been reported to decrease circulating anti-GM-CSF in 2 cases successfully. One of these patients demonstrated significant clinical improvements while the other did not. Rituximab, an anti-CD20 therapy, was studied in 2010 in 10 patients. Eight of these patients had significant clinical improvement despite no change in serum anti-GM-CSF levels. Both Rituximab and plasmapheresis are considered alternative therapies for PAP refractory to WLL.
The only proven therapy for secondary PAP is the treatment of the underlying disease. Case reports have demonstrated return of functional alveolar macrophages following the appropriate treatment of leukemia. WLL does not seem to benefit this population as only two of 14 patients studied had clinical improvement. Allogeneic hematopoietic stem cell transplant is a potential treatment option with two reported cases of complete resolution of secondary PAP after transplant.
WLL may benefit patients with congenital PAP, but is not curative. GM-CSF replacement therapies have not shown effectiveness. Corticosteroids and steroid-sparing agents such as azathioprine have been used at some centers with mixed results. Animal models show promise with gene therapy and direct pulmonary macrophage transplantation resulting in complete resolution of PAP. At least one patient has had successful hematopoietic stem cell transplant, which cured their PAP, but this therapy carries potential risks of graft-versus-host disease and opportunistic infections.
The differential diagnosis of pulmonary alveolar proteinosis includes diseases that share restrictive physiology and diffuse interstitial changes on computed tomography. Although a “crazy paving” pattern on computed tomography is suggestive of PAP, it can also be associated with cardiogenic pulmonary edema, acute respiratory distress syndrome, alveolar hemorrhage, lipoid pneumonia, mycoplasma pneumonia, pneumocystis pneumonia, radiation pneumonitis, drug-induced pneumonitis, sarcoidosis, non-specific interstitial pneumonitis, or bronchioloalveolar carcinoma. Acute silicosis can present with the same clinical and pathologic findings as PAP. Acute silicosis can display radiographic “crazy paving,” milky BALF with foamy macrophages, and PAS-positive amorphous material within the alveolus on biopsy. The similarities between acute silicosis and PAP emphasize the importance of a thorough occupational history for recent silica exposure.
Disease course varies, and the prognosis is unpredictable. When treated with whole lung lavage, the 5-year survival of autoimmune PAP is 95%. Previously spontaneous remission had been reported as frequently as 50% of the time. However, more recent and more extensive analyses have concluded that spontaneous remission occurs in less than 10% of patients. In one prospective cohort study of 39 asymptomatic PAP patients, 64% of patients remained stable while 7% progressed. In a retrospective analysis of mortality in PAP patients, 72% of deaths occurred as a direct result of respiratory failure while 20% were due to secondary infections. The prognosis for secondary PAP is worse with a median time for survival less than 15 months. The causes of death in secondary PAP patients include the underlying hematological disease (33%), infection (25%), respiratory failure (25%), and bleeding complications (13%).
Patients with PAP are at an increased risk of developing an opportunistic infection with approximately 5% of PAP patients developing one. The lung is the most common site of infection, although extra-pulmonary infections represent 32% of opportunistic infections in PAP patients. Typical bacterial cases of pneumonia are possible but infrequently reported in PAP. Nocardia and Mycobacterium tuberculosis are the 2 most commonly reported opportunistic infections in PAP. Fungal infections reported include Histoplasma, Aspergillus, Cryptococcus, and Blastomyces. Other reported infections include Acinetobacter, Coccidiodies, Mucorales, and Streptomyces.