The burden of autoimmune pulmonary alveolar proteinosis: a systematic review

Abstract

Background

Autoimmune pulmonary alveolar proteinosis (aPAP) is characterised by abnormal alveolar surfactant accumulation and reduced pulmonary gas transfer. Disease severity and progression depend on pulmonary surfactant accumulation, the rate of which varies widely among patients. Currently, whole-lung lavage (WLL) is the most widely accepted therapy. This review addresses the burden of aPAP on patients, caregivers and society.

Methods

MEDLINE and Embase databases were systematically searched for reports on the manifestations, treatment burden, caregiver impact and healthcare costs of aPAP published after 2000.

Results

Out of 1023 publications identified, 50 reported relevant data (for 2855 aPAP patients), including 43 observational studies and seven phase 2/3 trials. Commonly reported symptoms included dyspnoea, cough and sputum production. Clinical manifestations included progressive hypoxaemia, reduced exercise capacity, reduced quality of life, and an increased rate of serious infections. Low prevalence and nonspecific signs and symptoms contributed to delayed diagnosis of aPAP, frequent misdiagnoses, use of multiple tests with nondiagnostic results, and therapies that were inappropriate or exacerbated the disease. WLL was the most frequently administered therapy, and many patients required repeat procedures. Medical care costs were higher for PAP patients than for non-PAP control patients.

Conclusions

The results highlight the multifactorial and substantial burden of aPAP on patients. Significant unmet needs remain, particularly in achieving timely and accurate diagnosis and in providing effective, well-tolerated therapies that address the underlying pathophysiology of the disease.

Shareable abstract

This report outlines the healthcare burden on autoimmune pulmonary alveolar proteinosis patients including inaccurate/delayed diagnosis, progressive lung impairment, infection, lack of approved physiologically corrective therapy, and high healthcare costs https://bit.ly/4qt9vIb

Introduction

Autoimmune pulmonary alveolar proteinosis (aPAP) is a rare lung disease characterised by abnormal surfactant accumulation in the alveoli resulting in progressive impairment of pulmonary oxygen delivery and dyspnoea, an increased risk of serious secondary infections by a broad range of pathogens and, in some patients, pulmonary fibrosis, respiratory failure, and death [1–4]. The prevalence of aPAP has been estimated at 6–27 per million in the general population [5–7].

The pathogenesis of aPAP is mediated by an increase in the level of autoantibodies directed against granulocyte–macrophage-colony stimulating factor (GM-CSF), which alveolar macrophages require to clear excess surfactant from alveoli [8, 9]. Without GM-CSF, alveolar macrophages are unable to export surfactant-derived cholesterol [10, 11] that they internalise from the alveolar surface, resulting in abnormal accumulation of surfactant. Clinical disease course is variable and reflects the degree of surfactant accumulation.

Whole-lung lavage (WLL), the most widely accepted treatment option currently available for aPAP, involves repeatedly washing excess surfactant out of the lungs using warmed saline. While WLL improves gas transfer, it is invasive and requires hospital admission, general anaesthesia, and a specialised medical team to administer. Furthermore, it does not address the underlying pathophysiological defect or stop surfactant accumulation, and it is not available at most medical centres [12–14]. One form of inhaled recombinant human GM-CSF (rhGM-CSF) is in late-stage clinical development as therapy for aPAP and another form is approved for the treatment of aPAP in Japan. GM-CSF therapy is widely believed to be efficacious based on case studies/series, open-label trials and several randomised, blinded, controlled trials [13, 15–27]. Less well-studied treatment approaches include rituximab and plasmapheresis [28–37]; however, there are insufficient data to support their use in aPAP. Recently published clinical guidelines have recommended WLL and inhaled GM-CSF for treatment of aPAP patients [33, 38].

Because few published reports have addressed the healthcare burden of aPAP, we conducted a systematic literature review to determine how aPAP affects patients, caregivers and society. Specifically, we address the burden of disease manifestations, comorbidities, natural history, diagnosis, management and healthcare costs. The effectiveness of aPAP therapy is not addressed specifically in this report, but has been reviewed elsewhere [2, 3, 5, 23, 33, 39–42].

Materials and methods

Search methodology

This review (International Prospective Register of Systematic Reviews (PROSPERO) registration number CRD42024527178) [43] was conducted according to the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [44]. Publications of interest included reports of clinical trials, observational studies, and case series with at least five patients, including information on natural history, signs, symptoms, quality of life, treatment burden, impact on family members and caregivers, and impact on society (financial burden of healthcare) for aPAP patients (table 1).

TABLE 1.

Publication eligibility criteria

	Inclusion criteria
Population	Patients with aPAP#
Interventions	Any/none
Comparator	Any/none
Outcomes	aPAP natural history: presentation, comorbidities, pathway to diagnosis, secondary infections, pulmonary fibrosis Signs and symptoms: pulmonary function, gas transfer, hypoxia, dyspnoea, cough, sputum production, chest pain, fatigue, fever, haemoptysis Quality of life Medical burden of treatment Impact on family members/caregivers Financial burden: treatment costs and healthcare resource utilisation by country, loss of productivity
Study designs	Include clinical trials, observational studies and case series (n≥5) SLRs/meta-analyses to be screened for primary data Exclude letters, editorials, reviews and books
Date restrictions	2000 onwards for full-text articles 2019 onwards for case series 2019 onwards for SLRs 2021 onwards for congress abstracts from: American Thoracic Society international conference European Respiratory Society congress Chest annual meeting
Country restrictions	Unrestricted
Language	English-language publications only

aPAP: autoimmune pulmonary alveolar proteinosis; SLR: systematic literature review. ^#: aPAP rather than all forms of pulmonary alveolar proteinosis determined at screening.

Systematic published literature searches were conducted on 22 March 2024, using MEDLINE and Epub ahead-of-print, in-process, in-data-review and other nonindexed citations, daily and versions, and Embase. Searches included research reports (published from 2000 onwards, to include all research published over the past two decades), recent case series (published from 2019 onwards) and recent systematic reviews (to screen for any reporting primary data; published from 2019 onwards) that included PAP disease terms in their titles or abstracts. Search strings are shown in supplementary table S1. In addition, Embase was searched for abstracts from pre-defined congresses since 2021 (data in abstracts from earlier congresses were presumed to have been published in full-text articles); manual searches were conducted of the abstract book of the European Respiratory Society 2023 congress, which was not indexed in Embase. Publication eligibility was determined by title and abstract screening followed by full-text review, based on the pre-specified eligibility criteria detailed in table 1 [43].

Screening was completed by a single reviewer (Jennifer Y-T. Hung, Oxford PharmaGenesis, Melbourne, Australia), with uncertainties resolved by an independent adjudicating senior reviewer (Ruth Gandolfo, Oxford PharmaGenesis, Oxford, UK). Titles and abstracts of publications identified by the case report searches were screened to select those reporting case series of at least five patients with aPAP. Full-text articles and congress abstracts were screened to select publications that specifically reported on aPAP (including idiopathic PAP, a previous term used for aPAP), rather than other forms of PAP. Relevant congress abstracts were excluded if they were superseded by full-text publication of the data. In order to capture data on the natural history of aPAP rather than on treatment response, publications were selected that reported relevant data from patients with aPAP who had received supportive treatments such as WLL, and not systemic pharmacological treatment. Data were extracted manually by one researcher (Ruth Gandolfo) and recorded in a data extraction table according to pre-agreed data extraction fields (Microsoft Excel) [43].

Analysis

Numbers and proportions of patients with each outcome were calculated. For patient cohorts reported in more than one publication, outcomes data were extracted fully from one publication and only unique data were extracted from additional publications to avoid double-counting patients for individual outcomes. Ranges of means and medians were collated and weighted means were calculated (means of multiple means accounting for patient numbers in each cohort). No statistical tests were undertaken, and no meta-analyses were planned or conducted due to small sample sizes and the heterogeneity of available data.

Quality-of-life outcomes (e.g. St George's Respiratory Questionnaire (SGRQ), 36-Item Short Form Survey (SF-36)) were extracted when reported in eligible aPAP cohorts, because such data provide the only available information on the magnitude of quality-of-life impairment in aPAP. This review did not aim to evaluate or compare the effectiveness of specific therapies or the utility of specific instruments.

Results

Literature search results

Systematic searches identified 1023 unique publications (figure 1). Following review, 50 publications reporting data on the burden of aPAP were selected according to the pre-defined eligibility criteria (table 1) reporting data from 2855 aPAP patients. Additionally, healthcare costs from ∼500 hospital admissions for PAP syndrome (not only aPAP) from an inpatient database study in the USA and 164 patients with PAP from a claims database study in the USA were included because of the lack of published aPAP-specific cost data [6, 45].

FIGURE 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram of the systematic literature review process. ERS: European Respiratory Society; aPAP: autoimmune pulmonary alveolar proteinosis. ^#: hand searches of ERS 2023 congress abstract book were performed to identify relevant congress abstracts; ^¶: including publications that were duplicates across the three searches (Embase, MEDLINE and hand searches).

Characteristics of the included publications

Characteristics of the reports included are summarised in supplementary table S2. Most (43 out of 50) were observational studies and seven were treatment trials: two randomised treatment trials (n=64 and n=138) and five open-label treatment trials (10–50 patients each).

Comorbidities

Comorbidities reported for 428 aPAP patients included hypertension (7.7%), diabetes (6.1%), hyperlipidaemia (4.7%), infection (4.0%), liver disease (3.3%) and asthma (2.8%) (table 2, supplementary table S3).

TABLE 2.

Impact of autoimmune pulmonary alveolar proteinosis: common comorbidities

	Patients		Studies [references]
	Distribution	Total
Hypertension	7.7 (33)	428	7 [5, 46–51]
Diabetes	6.1 (26)	428	7 [5, 46–51]
Hyperlipidaemia	4.7 (20)	428	7 [5, 46–51]
Infection	4.0 (17)	428	7 [5, 46–51]
Liver disease	3.3 (14)	428	7 [5, 46–51]
Asthma	2.8 (12)	428	7 [5, 46–51]

Data are presented as % (n) or n.

Clinical measures

Clinical measures in aPAP patients, including symptoms, surfactant burden (seen in computed tomography (CT) scans), pulmonary gas transfer, disease severity and clinical course, exercise capacity, infections, respiratory health-related quality of life (HRQoL) and survival are summarised in tables 3 and 4 (supplementary tables S4 and S5).

TABLE 3.

Impact of autoimmune pulmonary alveolar proteinosis (PAP): symptoms and clinical measures

	Patients		Studies [references]
	Distribution	Total
Symptoms
None (range)	22.3 (169) (5.5–60)	757	12 [5, 7, 21, 46, 47, 49, 52–57]
Dyspnoea	61.2 (929)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Cough	43.8 (665)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Sputum production	18.7 (284)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Chest pain/discomfort	9.8 (149)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Fatigue	6.3 (96)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Pulmonary rales/crackles	4.9 (74)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Fever	3.8 (58)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Hypoxaemia	3.6 (54)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Weight loss	1.9 (29)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Cyanosis	1.5 (22)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Digital clubbing	1.2 (18)	1517	26 [5, 13, 17, 18, 21, 26, 27, 46, 47, 49, 50, 52–66]
Symptoms per patient median (range)	2 (0–4)	234	1 [64]
Surfactant burden (CT findings)
Ground-glass opacities	90.0 (162)	180	4 [49, 55, 58, 63]
Crazy-paving pattern	78.4 (149)	190	5 [58, 60, 63, 67, 68]
Pulmonary gas transfer
PaO2 mmHg
Normal range	75–100		[69]
Weighted mean	67.7±4.4	1211	25 [5, 7, 13, 17, 18, 21, 26, 27, 31, 32, 50, 52, 54–56, 60, 62, 66–68, 70–73]
Means	54.7–77.5	1211	25 [5, 7, 13, 17, 18, 21, 26, 27, 31, 32, 50, 52, 54–56, 60, 62, 66–68, 70–73]
Medians	57–78.8	591	8 [7, 31, 47, 56, 64, 74–76]
Alveolar–arterial oxygen difference (A–aDO2 or PA–aO2) mmHg
Normal value for a 50-year-old	15		[77]
Weighted mean (not weighted for age)	36.0±5.3	939	19 [5, 7, 13, 17, 18, 21, 26, 27, 31, 32, 50, 54–56, 62, 68, 70–72]
Means	28.3–50	939	19 [5, 7, 13, 17, 18, 21, 26, 27, 31, 32, 50, 54–56, 62, 68, 70–72]
Medians	30.1–49	233	5 [7, 31, 58, 74, 76]
DLCO % predicted
Normal values (depending on age and sex)	>80		[78]
Weighted mean	59.7±7.8	1314	26 [5, 7, 13, 17, 18, 21, 26, 27, 31, 32, 53–56, 58, 60, 62, 64, 67, 68, 70–73, 79, 80]
Means	41–75.8	1314	26 [5, 7, 13, 17, 18, 21, 26, 27, 31, 32, 53–56, 58, 60, 62, 64, 67, 68, 70–73, 79, 80]
Medians	44–91	401	7 [7, 47, 66, 75, 76, 81, 82]
Disease severity
PAP DSS 1 (asymptomatic, PaO2 ≥70 mmHg)	16.7 (181)	1087	15 [5, 7, 21, 26, 27, 54, 56, 57, 65, 71, 75, 79–81, 83]
PAP DSS 2 (symptomatic, PaO2 ≥70 mmHg)	34.4 (374)	1087	15 [5, 7, 21, 26, 27, 54, 56, 57, 65, 71, 75, 79–81, 83]
PAP DSS 3 (symptomatic, PaO2 ≥60 mmHg and <70 mmHg)	27.8 (302)	1087	15 [5, 7, 21, 26, 27, 54, 56, 57, 65, 71, 75, 79–81, 83]
PAP DSS 4 (symptomatic, PaO2 ≥50 mmHg and <60 mmHg)	14.5 (158)	1087	15 [5, 7, 21, 26, 27, 54, 56, 57, 65, 71, 75, 79–81, 83]
PAP DSS 5 (symptomatic, PaO2 <50 mmHg)	6.6 (72)	1087	15 [5, 7, 21, 26, 27, 54, 56, 57, 65, 71, 75, 79–81, 83]
Mean PAP DSS	2.5–2.7	93	3 [32, 56, 70]
Median PAP DSS	4	18	1 [13]
Disease course (in 12 weeks–∼9 years)
Spontaneous resolution (range)	8.4 (30) (2.9–13.2)	356	12 [13, 17, 47, 53, 55, 57, 58, 60, 62, 70, 72, 82]
Spontaneous improvement (range)	16.9 (60) (17.3–73.7)	356	12 [13, 17, 47, 53, 55, 57, 58, 60, 62, 70, 72, 82]
Stable without treatment (range)	15.4 (55) (20.0–50.0)	356	12 [13, 17, 47, 53, 55, 57, 58, 60, 62, 70, 72, 82]
Disease progression (range)	59.3 (211) (14.3–81.6)	356	12 [13, 17, 47, 53, 55, 57, 58, 60, 62, 70, 72, 82]

Data are presented as % (n), (range), n or mean±sd, unless otherwise stated. CT: computed tomography; P_aO2: partial pressure of arterial oxygen; D_LCO: diffusing capacity of the lung for carbon monoxide; DSS: disease severity score.

TABLE 4.

Impact of autoimmune pulmonary alveolar proteinosis (aPAP) on the patient

	Patients		Studies [references]
	Distribution	Total
Exercise capacity: 6MWD m
Mean (range) for a reference group of healthy individuals	571 (380–782)		1 [87]
Weighted mean	448.8±49.6
Means	239–501.9	326	8 [17, 18, 21, 27, 31, 58, 62, 68]
Median	498	104	1 [47]
Infections
Retrospective (at/before enrolment/diagnosis):		621	7 [5, 46, 47, 53, 54, 58, 59]
Any infection	11.4 (71)
Retrospective (up to 8.6 years’ follow-up)		224	2 [47, 58]
Infection leading to hospitalisation	15.6 (35)
Fatal pulmonary infection	0.9 (2)
Prospective in 24–42 weeks		78	2 [26, 27]
Any infection	32.1 (25)
Serious infection	1.3 (1)
Respiratory pathogens reported: Nocardia species, nontuberculous mycobacteria, Staphylococcus aureus, Escherichia coli, Chlamydia pneumoniae, Haemophilus influenzae, Pseudomonas aeruginosa, Mycobacterium tuberculosis, Mycoplasma pneumoniae, Klebsiella, pneumoniae, Pneumocystis jirocevii, Cryptococcus gattii, Streptococcus, pneumoniae, Aspergillus species, Serratia marcescens, Candida albicans		NA	5 [5, 53, 54, 59, 63]
Extrapulmonary pathogens reported: Helicobacter pylori (gastric), hepatitis C (liver), Cryptococcus gattii (CNS), Flavobacterium species (CNS), Nocardia species (CNS)		NA	5 [5, 53, 54, 59, 63]
Patient-reported quality of life
SGRQ functional health status: 0 (best health)–100 (worst health)#
Mild-to-moderate aPAP		36	1 [21]
Weighted mean total score   Means	27.0±0.5 26.5–27.4
Weighted mean symptom score	26.8±2.6
Weighted mean activity score	30.5±0.2
Weighted mean impact score	22.9±1.2
Mild-to-severe aPAP		138	1 [27]
Weighted mean total score	45.2±1.4
Means	44.1–47.2
Weighted mean symptom score	46.8±0.4
Weighted mean activity score	57.8±2.1
Weighted mean impact score	36.7±1.6
SF-36 quality of life: 0 (worst health)–100 (best health) (50: normative)		42 (25, 17, respectively)	2 [13, 74]
Median scores
Role limitations due to physical health	0, 45
Physical function	25, 0
Bodily pain	62, 40
General health	37, 35
Vitality	30, 56
Social function	50, 38
Role limitations due to emotional problems	33, 13
Mental health	52, 75
Physical component summary	33.4, NA
Mental component summary	40.2, NA
COPD Assessment Test: impact of COPD on health status: 0 (no effect)–  40 (severe effect)		64	1 [26]
Weighted mean total score	14.0±0.5
Means	13.5–14.5
Survival (in 24 weeks–9 years)		453	9 [7, 26, 27, 49, 53, 55, 58, 60, 67]
Death	3.5 (16)
5-year survival rate	97.7–100%
Mean survival time	57.8 months–16.1 years
Reported causes of 16 deaths		16
aPAP	11
Respiratory failure
Severe pulmonary infection and respiratory failure
Cerebrovascular accident	1
Fungal infection	1
Cancer	2
Unknown	1

Data are presented as mean±sd, n, range or % (n), unless otherwise stated. 6MWD: 6-min walk distance; CNS: central nervous system; SGRQ: St George's Respiratory Questionnaire; SF-36: 36-Item Short Form Survey; NA: not applicable/available. ^#: SGRQ component scores: symptom score, frequency of respiratory symptoms; activity score, effect on daily physical activity; impact score, impact on psychosocial function.

Symptoms

Symptoms were reported for 1517 patients (26 studies) (table 3, supplementary table S4). The proportions of patients reporting symptoms ranged from 40% to 94.5% across studies. In these 1517 patients, the most commonly reported symptoms were dyspnoea (61.2%), cough (43.8%), sputum production (18.7%), chest pain/discomfort (9.8%), fatigue (6.3%), fever (3.8%) and hypoxaemia (3.6%).

Surfactant burden

Surfactant burden (accumulation of surfactant within alveoli), as evaluated by lung CT scan before treatment, was reported (as ground-glass opacification (GGO) status, GGO score, GGO extent, crazy-paving status or CT grade/density) for 722 aPAP patients in 13 studies (table 3, supplementary table S4). Among 180 patients for whom GGO status was reported (in four studies), most (90.0%) were observed to have GGO, indicating abnormal surfactant accumulation. The mean baseline GGO scores in two studies were 4.9 and 10.8–10.9, respectively (score range 0–15 (higher score indicates GGO in a greater proportion of the lungs)) [13, 27]. Two studies utilising a different radiological assessment (extent of GGO and superimposed septal thickening (known as crazy paving [84])) reported a mean score of 19.1 (n=21) based on GGO extent (0 (none)–20 (extensive)) [52], and a mean score of 19.3 (n=13) based on GGO and CT reticulation extent (6 (limited)–30 (extensive)) [32]. Crazy paving was observed in 78.4% of 190 patients for whom crazy-paving status was reported (in five studies). In one study, patients were evenly distributed across severity grades determined by visual assessment of chest CTs: grade 1 (lowest) 23.4%; grade 2 28.6%; grade 3 25.5%; grade 4 (highest) 22.5% [79]. In aggregate, CT assessments using different visual scoring systems consistently demonstrated a high prevalence of GGO and crazy paving in aPAP.

Pulmonary gas transfer and clinical course

Pulmonary gas transfer is impaired in aPAP patients (table 3, supplementary table S4). The weighted mean±sd partial pressure of oxygen (P_aO2) for 1211 patients was 67.7±4.4 mmHg (normal: 75–100 mmHg [69]). Diffusing capacity of the lungs for carbon monoxide (D_LCO) is a direct measure of alveolar gas transfer that correlates well with oxygen transfer [85] and with aPAP severity [2, 5] (normal: >80% predicted, depending on age and sex [78]). Weighted mean±sd D_LCO was 59.7±7.8% pred (n=1314).

The Inoue disease severity score (DSS) is used to measure PAP severity based on the presence of symptoms and P_aO2, with scores ranging from 1 (asymptomatic, P_aO2 ≥70 mmHg) to 5 (P_aO2 <50 mmHg) [5, 86]. DSS was reported for 1087 patients (table 3, supplementary table S4): 16.7% had a DSS of 1; 34.4% had a DSS of 2; 27.8% had a DSS of 3; 14.5% had a DSS of 4; and 6.6% had a DSS of 5. Mean PAP DSS of 2.5–2.7 were reported in three observational studies, and median DSS was 4 in a phase 2 clinical trial.

Data on the clinical course were reported for 356 untreated patients enrolled in 12 studies with follow-up periods ranging from 12 weeks to 9 years (unspecified in two reports) (table 3, supplementary table S4). Within the time periods studied, it was reported that aPAP progressed in 59.3% of 356 patients, was stable in 15.4%, improved in 16.9%, and resolved in 8.4% (regardless of initial symptom status), based on WLL requirement, symptoms, CT evaluation, alveolar–arterial oxygen difference increase or disease worsening reported as an adverse event.

Data on pulmonary fibrosis were limited: only one study reported fibrosis data (two out of 31 patients had pulmonary fibrosis) [46].

Exercise capacity

The burden of aPAP can be demonstrated by patients’ reduced exercise capacity and inability to perform daily activities. For 326 aPAP patients in eight reports, the weighted mean±sd 6-min walk distance was 448.8±49.6 m compared with a mean of 571 m (range 380–782 m) in 444 healthy individuals [87] (table 4, supplementary table S5).

Infections

Patients with aPAP have increased susceptibility to a variety of microbial pathogens [88]. Infection rates differed according to study type. In seven retrospective studies, 71 (11.4%) out of 621 aPAP patients experienced serious infections at/before study enrolment or diagnosis. The severity of these infections was not described in most studies. However, in one study (median follow-up 4.3 years), 33.6% of 104 patients experienced infections leading to hospitalisation (nocardiosis n=10, pneumococcal pneumonia n=6, mycobacteriosis n=2), and one (1.0%) patient died from a pulmonary infection (Pseudomonas aeruginosa pneumonia) [47]. In another study (follow-up 8.6 years), one (0.8%) out of 120 patients died from a pulmonary infection (unspecified pathogen) [58]. In the placebo groups of two randomised controlled trials, a serious lower respiratory tract infection was reported in one (2.1%) out of 47 patients in 24 weeks, and no serious infections were reported in 31 patients in 42 weeks (table 4 and supplementary table S5). Reported pathogens are listed in table 4.

Respiratory HRQoL

The diverse manifestations of aPAP profoundly affect patients’ quality of life. Here, we look at the SGRQ and SF-36 as quality-of-life measures. Most available SGRQ and SF-36 data are reported from treated cohorts in clinical trials or open-label studies, and thus primarily reflect quality-of-life status in patients receiving GM-CSF or WLL rather than the untreated natural course. In this review these results are presented to illustrate the magnitude and domains of quality-of-life impairment in aPAP, not to assess the effectiveness of specific therapies or the utility of specific instruments.

There are no specific and validated patient-reported outcome measures for PAP. The SGRQ was designed to evaluate respiratory HRQoL in patients with COPD or asthma, with SGRQ total scores ranging from 0 (best health) to 100 (worst health), reflecting impact on respiratory health status. A 4-point change is the minimum clinically important difference in SGRQ total score in COPD [89]. Baseline SGRQ scores were available from two clinical trials (table 4, supplementary table S5). In a trial of 36 patients with mild/moderate aPAP (inhaled GM-CSF treatment (molgramostim) and control groups), the weighted mean±sd baseline SGRQ total score was 27.0±0.5 [21]. After 6 months of treatment, the mean SGRQ total score was 13.9 in the GM-CSF group (n=17), a 13.5-point reduction from 27.4 at baseline. Similarly, in the IMPALA trial (138 patients with mild/moderate aPAP), the weighted mean±sd baseline SGRQ total score was 45.2±1.4. After 24 weeks of daily GM-CSF treatment (inhaled molgramostim, n=46), the mean SGRQ total score was 35.1, a 12.1-point reduction from 47.2 at baseline [27].

The SF-36 is a general HRQoL questionnaire comprised of eight components measuring physical and mental health [90]. The scores for each component range from 0 (worst health) to 100 (best health) [90]. Median SF-36 scores in one open-label aPAP study (n=25) ranged from 0.0 to 62.0 across components at baseline and increased to 43.8–100.0 after 6 months of subcutaneous GM-CSF (sargramostim) [74]. In another open-label aPAP study (n=17), median SF-36 scores across components ranged from 0 to 75 before all patients received a baseline WLL and increased to 73–100 after WLL [13] (table 4, supplementary table S5).

Diagnosis

Data on aPAP diagnosis were included in 25 publications (table 5, supplementary table S3). The weighted mean±sd time between symptom onset and aPAP diagnosis was 15.9±9.9 months for 302 patients in five studies [53, 58–61].

TABLE 5.

Impact of autoimmune pulmonary alveolar proteinosis (aPAP): burden of the diagnostic pathway

	Patients		Studies [references]
	Distribution	Total
Time from symptom onset to aPAP diagnosis months		302	5 [53, 58–61]
Weighted mean	15.9±9.9
Means	4–34
Range	0–92
Misdiagnoses		148	4 [46, 53, 60, 61]
Pneumonia	42.6 (63)
Other interstitial lung diseases	40.5 (60)
Tuberculosis	20.9 (31)
Asthma/asthmatic syndrome	9.5 (14)
Other	8.9 (13)
Tests		1269	17 [5, 17, 31, 47, 49, 53–55, 58, 60–63, 67, 74, 75, 79]
Bronchoalveolar lavage	68.2 (866)
Lung biopsy	47.1 (598)
CT scan	46.0 (584)
GM-CSF antibody assay	7.2 (91)
Two tests	22.0 (279)
Three tests	21.3 (270)

Data are presented as n, mean±sd, range or % (n). CT: computed tomography; GM-CSF: granulocyte–macrophage colony-stimulating factor.

Most patients across six studies were diagnosed with a variety of disorders before the aPAP diagnosis. Some patients had multiple reported diagnoses by multiple physicians before aPAP was confirmed. Four studies [46, 53, 60, 61] reported 148 patients who were initially misdiagnosed; the most common misdiagnoses were pneumonia (42.6% of 148 patients), other interstitial lung diseases (40.5%) and tuberculosis (20.9%) (table 5, supplementary table S3). In a study of data from 1977 to 2013, 20% of 68 patients had at least three misdiagnoses before aPAP diagnosis [53]. Misdiagnosis may lead to inappropriate treatments that may be ineffective or even exacerbate aPAP. For example, most aPAP patients receiving corticosteroids in a Japanese cohort study were given them because of an initial diagnosis of idiopathic interstitial pneumonia, drug-induced interstitial lung disease, chronic hypersensitivity pneumonitis, chronic eosinophilic pneumonia, or alveolar cell carcinoma; PAP DSS worsened in ∼75% of these patients during corticosteroid therapy [46]. Reported “misdiagnoses” therefore include both clearly incorrect diagnoses (such as pneumonia or tuberculosis) and interstitial lung diseases that can coexist with aPAP (for example, chronic hypersensitivity pneumonitis).

Most patients undergo multiple nondiagnostic tests (e.g. transbronchial lung biopsy and open lung biopsy) before the diagnosis of aPAP is made. Of 1269 aPAP patients for whom data were available, 22.0% and 21.3% underwent two or three nondiagnostic tests, respectively, prior to ultimately undergoing a diagnostic test. Tests were most commonly bronchoalveolar lavage (68.2%), lung biopsy (open, transbronchial or thoracoscopic lung biopsy; 47.1% of patients) and CT scanning (46.0%) (table 5, supplementary table S3). It is important to note that while bronchoalveolar lavage and lung biopsy may support the presence of PAP, neither are diagnostic for aPAP. The only currently available method capable of diagnosing aPAP specifically is the presence of high-titre anti-GM-CSF antibodies [91, 92].

Survival

Survival rates were reported for 453 aPAP patients (table 4, supplementary table S5). In two clinical trials, none of the 78 patients who received placebo for 24 weeks died [26, 27]. In seven observational studies with follow-up periods of 2–9 years (unspecified in one study) and a total of 375 patients who did not receive pharmacological treatment, 16 (4.3%) died. Reported causes of death were aPAP (11 out of 16: pulmonary infection, respiratory failure), cerebrovascular accident (one out of 16), fungal infection (one out of 16), lung cancer (one out of 16), gastric cancer (one out of 16) and unknown (one out of 16). 5-year survival rates of 97.7% and 100% were reported in two retrospective studies (n=103 (1999–2017) and n=68 (1977–2013), respectively) [7, 53]. These rates may reflect the effect of WLL treatment on survival: it has previously been reported to improve 5-year survival from 85% to 94% [88].

Treatment burden

Data on the burden of treatment were reported in 36 publications (table 6, supplementary table S6). While treatment response is outside the scope of this review, data on the impact of undergoing aPAP-related therapies (WLL, lung transplantation, plasmapheresis and infection prophylaxis) are described herein.

TABLE 6.

Impact of autoimmune pulmonary alveolar proteinosis: treatment burden of whole-lung lavage (WLL) procedures (during study periods)

	Patients		Studies [references]
	Distribution	Total
WLL in 24 weeks to 8 years	49.6 (715)	1441	22 [12, 13, 27, 50, 53–55, 57–63, 66, 70–73, 79, 81, 83]
>1 WLL procedure	28.2 (407)	1441	22 [12, 13, 27, 50, 53–55, 57–63, 66, 70–73, 79, 81, 83]
>5 WLL procedures	4.8 (69)	1441	22 [12, 13, 27, 50, 53–55, 57–63, 66, 70–73, 79, 81, 83]
Weighted mean WLL procedures per patient	3.0±1.2	541	3 [12, 53, 61]
Means	1.9–5.7	541	3 [12, 53, 61]
Weighted mean time between WLL procedures months	9.3±2.9	271	2 [12, 54]
Means	8–15.7	271	2 [12, 54]
Median time between WLL procedures months	18	9	1 [13]

Data are presented as % (n), n or mean±sd.

Whole-lung lavage

Approximately half (49.6%) of 1441 patients (described in 22 reports) received WLL during follow-up periods of between 24 weeks and 8 years (unspecified in six studies). Only three studies (a phase 2 clinical trial, a long-term prospective observational study, and a retrospective study; 43 patients in total) had an observation period (12–36 months) starting with a baseline WLL [13, 59, 62] (table 6, supplementary table S6). Nearly 90% of patients (1256 out of 1441) were not receiving pharmacological therapies, while 10.8% (156 out of 1441) received GM-CSF and <2% received corticosteroids or rituximab (21 out of 1441 and eight out of 1441, respectively).

The data showed that 56.9% (407 out of 715) of patients requiring WLL underwent more than one procedure and 9.4% (67 out of 715) underwent more than five. The weighted mean±sd number of WLL procedures per patient (time periods unspecified) was 3.0±1.2 for 541 patients, and one study (n=11) reported a median of four WLLs per patient [81]. The highest numbers of WLL procedures performed on a single patient were 23 (over ∼5 years) [81] and 16 (period unspecified) [53]. Time between WLL procedures was a median 18 months for a control group of nine patients [13] and weighted mean±sd 9.3±2.9 months for 271 patients [12, 54].

In a survey of 1110 WLL procedures in 368 patients, complications (per procedure) included transient fever (18%), hypoxaemia (14%), wheezing (6%), pneumonia (5%), fluid leakage (4%), hydrothorax (3%), pneumothorax (1%) and cardiac arrest (1%) [12]. Another two studies reported that a single patient in each experienced pneumothorax following WLL, and one study also reported two cases of severe hypoxaemia after WLL [60, 81]. However, two studies reported no severe complications related to WLL in a total of 141 patients [58, 62].

Hospitalisation

Data on hospitalisations were reported in two studies (supplementary table S6). In a case series of five aPAP patients with infections, two were hospitalised for infections and one was hospitalised for respiratory failure requiring mechanical ventilation [63]. One aPAP patient in a GM-CSF treatment cohort (n=14) was hospitalised for acute respiratory distress and viral infection [93]. In addition, two claims database studies captured hospitalisations. In a United States claims database (OptumInsight) study, the mean±sem length of hospital stay was 15.96±20.71 versus 5.40±5.07 days (p=0.027) for 164 patients with PAP compared to age- and sex-matched patients without PAP (follow-up period 2008–2012) [6]. In a United States inpatient database (Nationwide Inpatient Database) study of 500 hospital admissions of PAP patients, mean (95% CI) hospital stay duration was 6.24 (3.92–8.55) days and mean (95% CI) hospital stay duration with WLL was 3.21 (1.56–4.85) days [45].

Lung transplantation and plasmapheresis

Lung transplantation was reported in two studies: one patient (GM-CSF treatment cohort; n=14) received transplantation and another patient (rituximab cohort; n=13) awaited transplantation [32, 93] (supplementary table S6). Although lung transplantation is conditionally recommended for eligible patients with PAP progression despite WLL or pharmacological treatment [33], continued GM-CSF autoantibody production in aPAP patients may lead to disease recurrence in transplanted lungs [94, 95]. While waiting for transplantation, these two patients were the only patients reported to have received plasmapheresis in the included studies [32, 93].

Burden on family and caregivers

No reports included data on the impact of aPAP on caregivers/families.

Financial burden of PAP

No reports included data on the healthcare costs of aPAP; data on healthcare costs for patients with any form of PAP (∼90% of whom are expected to have aPAP [3]) were therefore included in this systematic literature review. Two studies from the USA described costs associated with healthcare for PAP patients [6, 45].

In a United States claims database (OptumInsight) study, numbers of hospital visits were higher for 164 patients with PAP than for age- and sex-matched patients without PAP (follow-up period 2008–2012) [6]. The mean±sem number of outpatient visits was 17.30±13.77 versus 10.40±11.38 (p<0.01), the mean±sem number of emergency room visits was 1.49±1.17 versus 1.08±0.27 (p=0.014), and the mean annual per-patient healthcare costs (2008–2012) were USD 54 865±95 524 for PAP patients versus USD 10 214±20 333 (p<0.001) for patients without PAP [6]. Inpatient visit costs were 2.7 times higher for patients with PAP versus without (p=0.04). Outpatient visit costs were 3.8 times higher for patients with PAP versus without (p<0.001). Prescription costs were 4.75 times higher for patients with PAP versus without (p<0.001). Emergency room visit costs were similar between patients with and without PAP (p=0.0563). In a United States inpatient database (Nationwide Inpatient Database) study of PAP admissions, the mean hospitalisation cost per admission was USD 29 932.29 (versus a mean hospital admission cost of approximately USD 10 900 for the general US population in 2014), and the mean annual hospitalisation cost for PAP patients was USD 5 071 401.36 [45, 96].

Since the completion of this systematic literature review, data from a large United States claims database study has confirmed the high healthcare costs associated with PAP: mean annual nonpharmacy charges were USD 71 673 and USD 14 656 for patients with and without PAP, respectively [97].

There were no reports on the broader financial impact of aPAP, for example related to employment rates or work productivity among patients/caregivers.

Discussion

This systematic review summarises the medical literature related to the burden of aPAP on patients, caregivers/families and society. Here, we highlight the substantial burden of aPAP on affected individuals, shaped by progressive worsening clinical manifestations (breathlessness, cough and sputum production), diagnostic delays, and limitations in treatment access. Patients face escalating respiratory symptoms and physiological compromise due to the accumulation of surfactant in the alveoli, frequently complicated by severe infections (both intra- and extra-thoracic), progressive pulmonary fibrosis, and, in advanced stages, respiratory failure. Timely and accurate diagnosis remains a challenge due to limited awareness of GM-CSF antibody testing. Furthermore, the lack of an approved effective pharmacotherapy and the limited access to WLL contribute to both medical and psychosocial distress. Notably, no studies to date have systematically assessed the burden of aPAP on patients, caregivers, providers or society, representing a significant gap in the literature.

The diagnostic process for aPAP is frequently protracted, invasive and fraught with missteps. Invasive procedures such as bronchoscopy and surgical lung biopsy are often performed as part of the diagnostic workup, but pose risks such as pneumothorax, bleeding and cardiac complications and fail to specifically identify aPAP. The underutilisation and, until recently, limited availability of highly specific, blood-based diagnostic tests, such as GM-CSF autoantibody assays, contribute to diagnostic uncertainty and delay. This may lead to misdiagnosis and inappropriate treatment, some of which may be ineffective or even harmful. Misdiagnosed patients often undergo therapeutic approaches that offer no benefit, postponing effective intervention and potentially exacerbating disease progression.

Serum GM-CSF autoantibody measurement using a standardised, highly specific ELISA is now the cornerstone of aPAP diagnosis and is performed in a limited number of specialised reference laboratories worldwide. Routine clinical GM-CSF autoantibody testing is currently available through specialised laboratories in the United States (including Cincinnati Children's Hospital Medical Center and National Jewish Health), Japan, Germany and China, with information on additional reference centres curated by the PAP Foundation and the European Pulmonary Alveolar Proteinosis Network [98, 99]. These assays are best performed in such reference laboratories with validated methods and expertise in assay interpretation or via the recent introduction of commercially available testing in the United States and Japan and a free GM-CSF autoantibody testing programme offered by Trillium BioPharma in the USA, which expand access to standardised blood testing [100–104]. Because assay performance and interpretation depend on validated methodology and experienced centres, these reference-laboratory and sponsored testing approaches help reduce misdiagnosis and support timely, appropriate management for patients with suspected aPAP.

Currently available treatment of aPAP is similarly limited and burdensome. Access to WLL is restricted to specialised centres, and the procedure itself is resource-intensive, requiring hospitalisation, anaesthesia, and specialised personnel. From the patient perspective, WLL is often perceived as invasive, uncomfortable and disruptive to daily life. Furthermore, WLL is a symptomatic treatment that does not address the underlying disease mechanism. Other therapeutic options include plasmapheresis and lung transplantation. Plasmapheresis to remove anti-GM-CSF antibodies requires frequent procedures and is therefore impractical. Lung transplantation is reserved for more end-stage patients who have developed fibrosis; however, aPAP returns in the transplanted lung. Critically, no approved pharmacotherapy currently exists to restore alveolar macrophage function, underscoring the urgent need for effective disease-specific treatments.

The clinical manifestations of aPAP impose a continuous and unpredictable burden on patients. The risk of serious infections remains high due to impaired host defences, and these infections can occur without warning, contributing to morbidity and hospitalisation. In addition, some of the most serious infections are extrathoracic (e.g. Nocardia brain abscesses). In some patients, chronic inflammation and alveolar damage lead to the development of progressive pulmonary fibrosis, which may culminate in respiratory failure and premature death. The physical, emotional and logistical demands of living with aPAP can be profound, further compounding patient burden.

An important observation from the findings of this review is that most aPAP studies have limited follow-up time periods that do not allow for the full natural history of the disease to be observed. Given that 1) aPAP lung disease is driven by surfactant accumulation; 2) the rate of accumulation varies widely between patients; 3) patients often present clinically only after surfactant has accumulated to levels sufficient to cause hypoxaemia, reduced lung compliance, and exertional dyspnoea; and 4) some patients are identified by mandatory health screening programmes or incidentally, it is possible that some aPAP patients in the included reports were evaluated at a point in their disease prior to onset of clinically significant symptoms. It is also possible that patients categorised as “spontaneously improved” or “resolved” experienced a fluctuation in disease severity or experienced surfactant reaccumulation following a single WLL so that they were deemed to have disease resolution when in fact their alveoli were simply accumulating surfactant very slowly. Longer observational follow-up times would help answer these questions.

Several limitations affect the interpretation of the current literature. The study populations included in this review comprise patients treated at academic centres, who may have received more aggressive disease management than other patients with aPAP. Furthermore, emerging data on transbronchial lung cryobiopsy in PAP were published after our search cut-off and were not systematically captured in this review, but they support the growing use of cryobiopsy to improve diagnostic yield and to characterise coexisting fibrotic lung disease [105, 106]. Many existing reports address PAP as a syndrome and do not specifically delineate findings for the autoimmune subtype. Many studies were observational or retrospective in design and were not intended to assess the patient-centred burden of disease. In addition, the natural history of HRQoL in untreated aPAP remains poorly characterised because most available patient-reported outcome data derive from interventional or otherwise treated cohorts rather than from longitudinal, treatment-naïve populations. The lack of meta-analytic approach and the wide heterogeneity of study designs, including short follow-up durations, limit the strength of conclusions. Additionally, the potential for overlapping patient inclusion across studies cannot be excluded. These gaps highlight the need for future investigations to directly assess the clinical, psychosocial, and economic burden of aPAP using standardised and prospective methodologies.

Future research should prioritise studies specifically designed to characterise the burden of aPAP on patients, caregivers, healthcare providers and broader society. These studies should incorporate validated patient-reported outcome measures, economic impact assessments, and longitudinal data to document the natural history of the disease. There is also a need to understand the safety, tolerability, and long-term impact of WLL, including standardised criteria for its initiation and frequency. Most critically, the development and approval of effective pharmacotherapies that target the underlying pathophysiology of aPAP remain essential. Establishing global consensus treatment guidelines, alongside expanding access to accurate diagnostic testing and specialised care centres, would significantly improve the quality of care for patients living with this rare and burdensome disease.

Points for clinical practice

• Timely and accurate aPAP diagnosis using the specific GM-CSF autoantibody assay is important to avoid inappropriate diagnostic testing and symptom management, and to accelerate initiation of appropriate treatment.

• Many patients with aPAP require repeated whole-lung lavage procedures to improve pulmonary gas transfer; however, as this is an invasive, resource-intensive procedure that does not address the underlying pathophysiology of the disease, there is an unmet need for approved effective, well-tolerated pharmacotherapies to relieve the clinical and psychosocial burden of aPAP.

Questions for future research

• Will the availability of an accurate blood test for diagnosis of aPAP (serum GM-CSF autoantibody testing) change clinical practice by replacing the use of lung biopsies, which are unable to identify any PAP-causing diseases?

• Will the availability of an effective pharmacotherapy for aPAP relieve the burden on patients, caregivers, healthcare providers, and society?

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material

ERR-0237-2025.SUPPLEMENT