Clinical practice guideline for the diagnosis of granulomatous-lymphocytic interstitial lung disease in patients with common variable immunodeficiency disorders: an ERS Clinical Research Collaboration

Abstract

Background

Granulomatous-lymphocytic interstitial lung disease (GLILD) is a lung disease first described in people affected by common variable immunodeficiency disorders (CVID). Despite growing recognition of GLILD, there is no accepted diagnostic and management guideline, and current practice varies significantly across centres and countries. This clinical practice guideline provides evidence- and consensus-based recommendations on the screening and diagnosis of GLILD in patients with CVID.

Methods

A panel representing multiple interdisciplinary perspectives convened with methodologists to prioritise clinical questions, and review and assess the evidence using Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology. Evidence-to-decision frameworks were used to decide on the direction and strength of recommendations.

Results

Screening for GLILD is recommended in all adult patients with CVID, preferably using high-resolution computed tomography. Evaluation should be performed by a multidisciplinary team. Routine lung biopsies are not mandatory but necessary for atypical presentations. As part of the diagnostic process, pulmonary infections should be excluded and lymphocytic alveolitis sought for by means of bronchoalveolar lavage. The severity of lung function impairment should be evaluated using pulmonary function tests including gas transfer assessment. Most of the recommendations are graded as conditional because of low certainty in the evidence regarding health effects.

Conclusions

This guideline allows for international homogeneity in the diagnosis of GLILD, thereby paving the way for improved comparability between centres, improving equity in healthcare for those affected by GLILD and facilitating multicentre research collaborations for future studies.

Shareable abstract

This guideline provides an evidence-based framework for diagnosing GLILD in CVID, recommending HRCT screening, multidisciplinary evaluation and selective biopsy, to enhance diagnostic consistency, equity in patient care and global research collaboration https://bit.ly/437GHuE

Introduction

Granulomatous-lymphocytic interstitial lung disease (GLILD) is a serious noninfectious complication of common variable immunodeficiency disorders (CVID), the most frequent symptomatic primary immunodeficiency in adults. CVID affects approximately one in 25 000 individuals, and GLILD is estimated to develop in 15–20% of these patients, posing a significant burden of morbidity and mortality [1, 2]. The term GLILD was originally introduced by Bates et al. [1] in 2004 to refer to CVID patients showing histological patterns of granulomas and/or lymphocytic infiltrates. It was later defined more formally by the British Lung Foundation/UK Primary Immunodeficiency Network collaboration as “a distinct clinico-radio-pathological interstitial lung disease (ILD) occurring in patients with CVID, characterised by a lymphocytic infiltrate and/or granulomatous inflammation in the lung, in whom alternative diagnoses have been carefully considered and, where possible, excluded” [3].

GLILD should be suspected in CVID patients who present with chronic cough and/or dyspnoea on exertion, although symptoms can be subtle or patients asymptomatic. Furthermore, a restrictive pattern or loss of diffusion capacity on pulmonary function tests (PFTs) and small nodules, consolidations, interlobular septal thickening and/or ground-glass opacities on chest computed tomography (CT) in the absence of infectious agents should also raise suspicion of GLILD.

Despite growing recognition of GLILD in the literature, there is no accepted diagnostic and management guideline, and current practice varies significantly across centres and countries [4, 5]. The inconsistency stems in part from the lack of large-scale, prospective studies and the limited availability of multidisciplinary expertise in CVID-related lung disease. Furthermore, patient perspectives are rarely included despite the fact that one of the top-ranked questions in a GLILD research prioritisation exercise, as identified by patients themselves, was about the optimal screening and treatment approach [6]. This reflects patient uncertainty and concern about current diagnostic and management pathways.

In 2023, an international task force was established with the support of the European Society for Immunodeficiencies (ESID) in collaboration with European Reference Network for Rare Immunodeficiency, Autoinflammatory and Autoimmune Diseases (ERN-RITA) and with methodological support from the Evaluation Unit of the Canary Islands Health Service (SESCS), with the goal of reviewing the best available evidence and developing an evidence-based guideline for the diagnosis and management of GLILD in both children and adults. The network was originally formed from the European Respiratory Society (ERS) e-GLILDnet Clinical Research Collaboration [7]. The task force prioritised the needs of both clinicians and patients by identifying key clinical questions and systematically reviewing the supporting evidence [8]. The guideline is being developed in two phases: the first, presented in this paper, focuses on diagnosis, while the second phase will address management and treatment strategies. This guideline is meant to provide practice recommendations in clinical care for those treating people with CVID, related inborn errors of immunity and/or suspected GLILD, and can provide a consensus definition on the diagnosis of GLILD to guide further research.

Methods

Organisation, working group composition, planning, coordination and management of competing interests

The Guideline Development Group (GDG) was composed of members from various backgrounds and from different countries. It consisted of a core group of four clinicians with expertise in pulmonology and immunology (K. Warnatz, J.R. Hurst, A.A.J.M. van de Ven and H.M. Bintalib), one methodologist expert in guideline development (M.X. Rojas-Reyes) who drove the process forward, and various other clinical experts (paediatric and adult immunologists, paediatric and adult pulmonologists, a radiologist and pathologist), a social scientist (M. Calderon-Jaramillo), a patient representative (F. Strauss) active at the International Patient Organization for Primary Immunodeficiency (IPOPI) and a methodological team from the SESCS. Finally, the methodological team consisted of eight researchers with experience in evidence synthesis: one clinical epidemiologist and guideline expert (M.X. Rojas-Reyes), one documentalist (J. Tomas Ramos) and six researchers with expertise in evidence synthesis (A. de Armas Castellano, Y. González Hernández, Z. López Mújica, D. Ingante Vuntura, T. del Pino Sedeñox, I.C. Rada Ramirez). The evidence search was designed and conducted by a team member from the Epistemonikos Foundation (J. Tomas Ramos). The overall guideline development was guided by the methodological principles developed within the European Reference Network: Clinical Practice Guidelines and Clinical Decision Support Tools Programme [9] and was intended to align with the recommendations for trustworthy guidelines by the Institute of Medicine and the Guidelines International Network [10, 11]. The GDG developed and graded the recommendations based on the certainty of supporting evidence following the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. All participants declared any potential conflicts of interest at the beginning of the project and again prior to final publication. Because no relevant conflicts of interests were identified, all members were fully involved in the generation of recommendations.

Definition of questions and outcomes of interest

The GDG employed a structured process to prioritise clinical questions, beginning with the core group identifying key challenges in the diagnosis and management of GLILD in CVID patients. This led to the definition of three major topic areas: 1) screening and diagnosis, focused on selecting the most accurate, least invasive and most cost-effective diagnostic tests; 2) disease progression assessment, aimed at identifying optimal tools and timing for monitoring disease changes; and 3) treatment and maintenance, addressing first- and second-line therapies, steroid use, alternatives in case of treatment failure and maintenance strategies. An iterative, collaborative process involving electronic communication enabled the clinicians of the GDG, guided by a lead methodologist (M.X. Rojas-Reyes), to define and refine clinical questions for each topic. After collecting all proposals, comments and suggestions, the core group drafted an initial list of questions to be addressed in the guideline. This list was then presented at an online meeting, where it was validated by all panel members. At this stage, contributions from the social scientist and patient representative were incorporated, resulting in a final set of 21 questions for inclusion in the GLILD guidelines. Of these, seven questions related to screening and diagnosis were prioritised by the GDG for the first phase of recommendations and are presented in this manuscript, while the remaining 14 questions, addressing initial treatment, maintenance treatment and follow-up, will be covered in a second phase (supplement 1). Outcome measures of interest were initially identified based on expert input and literature review, which subsequently were ranked by the panel members independently using an online survey to determine their relative importance for decision-making following the GRADE approach (table 1).

TABLE 1.

Outcomes rated by the guideline panel as critical for decision-making

Screening and diagnosis	Assessment of disease activity and progression during follow-up	Treatment
Diagnostic accuracy Safety (adverse events) Changes in clinical management of the disease Exclusion of infection, malignancy and lymphoma	Progression (changes) over time Safety (adverse events) Changes in clinical management of the disease (length or frequency of follow-up)	Changes in disease activity parameters (including but not limited to symptoms, pulmonary function tests and radiological findings) Safety (adverse events) Treatment response Mortality

Evidence review

To address the anticipated paucity of evidence on GLILD in patients with CVID, a comprehensive and structured search strategy was developed (supplement 2). The Epistemonikos Foundation carried out a highly sensitive, multistep process to identify and classify studies relevant to each predefined Population, Intervention, Comparison and Outcome (PICO) question. Searches were conducted in MEDLINE, Embase, Cochrane, HTA, CINAHL and PsycINFO, with no restrictions on language or publication date, from database inception to 5 February 2024. An automated system enabled continuous updating, deduplication and incorporation of records into the dedicated CVID-GLILD Epistemonikos L·OVE collection.

Study selection proceeded in three stages: initial screening for relevance to CVID and confirmed or suspected GLILD or ILD; categorisation of eligible studies according to thematic areas aligned with the guideline's evidence-to-decision (EtD) framework, such as diagnostic interventions, treatment strategies and health equity considerations; and full-text review to assign studies to each guideline question. Title and abstract screening was performed by two technical team members (A. de Armas Castellano and I.C. Rada Ramirez). Full-text screening was carried out in pairs (A. de Armas Castellano plus one other technical team member), with a third member (M.X. Rojas-Reyes) resolving any discrepancies that arose.

The team organised the selected studies into evidence matrices and included observational studies when high-level evidence was lacking (supplement 3). The SESCS technical team extracted the data and evaluated the risk of bias using validated instruments, including the Newcastle–Ottawa Scale and Joanna Briggs Institute (JBI) tools [12, 13]. To evaluate potential undesirable effects of diagnostic tests, the guideline group also used indirect evidence from patients with other ILDs and respiratory conditions who frequently undergo the same procedures, such as bronchoscopy and lung biopsy (supplement 4). The last screening for diagnostic studies was performed on 24 February 2026. Identified studies were incorporated into the corresponding questions and the effect of the new evidence on the original recommendations was assessed in an online survey by the GDG. Results of the above-described process are summarised in a Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) flowchart (supplement 2).

A quantitative synthesis of the evidence was not performed because meta-analysis was not feasible in most cases. This was due to a lack of comparative assessment of the tests to estimate test accuracy and the substantial variability in study designs, which ranged from case reports to follow-up cohorts. Therefore, we conducted a qualitative (narrative) synthesis of the data from each included study.

Certainty of evidence was graded using the GRADE approach. The certainty of evidence was classified as high, moderate, low and very low, considering the criteria of limitations in study design (risk of bias), inconsistency, imprecision, indirect evidence and publication bias. Results of the assessment are included in summary of findings tables (supplement 5). For each guideline question, the methodological team prepared the summary of findings table and an EtD table using the GRADEpro software. Each EtD table includes information about the effects of interventions on outcomes of interest, patient values and preferences regarding the importance of outcomes, and information on contextual factors such as resource utilisation, health equity issues, acceptability of interventions to interest-holders, and the feasibility of implementation (supplement 6). There is limited evidence on health inequalities related to GLILD in CVID, likely due both to the rarity of the condition and a limited awareness of the sociodemographic characteristics of affected individuals. To address this, the social scientist involved in guideline development highlighted key aspects to consider when designing a clinical practice guideline for this condition, which are presented in supplement 7. All guideline panel and group members received these considerations prior to the validation survey, and they were incorporated into the development of the recommendations.

Good practice statement development

During question framing, the panel identified key aspects of clinical care in GLILD for which formal evidence synthesis was not considered appropriate. This was primarily due to the anticipated absence of direct comparative studies and the existence of well-established clinical practices supported by indirect evidence and expert consensus. In accordance with guidance from the GRADE Working Group, these topics were addressed by means of good practice statements, defined as recommendations for which the expected benefits clearly outweigh potential harms despite the lack of formally graded evidence [14, 15].

Recommendation development process

As mentioned earlier, in the first phase we developed recommendations for the prioritised questions related to screening and diagnosis. Each question was assigned to at least two clinical experts and the methodological lead, who collaboratively reviewed and validated the evidence in the EtD tables, suggesting further literature review when needed. Prior to the guideline recommendations meetings, all members of the GDG received the EtD and were invited to provide their judgements on each EtD aspect in advance of the meeting. Preliminary recommendations and justifications were developed during online meetings, along with implementation and subgroup considerations. These drafts were reviewed by all GDG members, who suggested corrections and clarifications. Three virtual meetings, led by an external GRADE expert methodologist, were held to finalise recommendations for the seven screening and diagnosis questions. The GDG discussed and voted on each EtD criterion until consensus was reached on the direction and strength of recommendations. After the virtual meetings, the methodological lead (M.X. Rojas-Reyes) drafted the final recommendations document, incorporating subgroup, research and implementation considerations. This draft was circulated to the GDG via a survey, with consensus defined as >75% agreement. Two recommendations were revised and voted on in a subsequent online round. Final approval of the recommendations was achieved at the ERS e-GLILDnet CRC Meeting in Lausanne, Switzerland, in May 2025. Following updated literature searches conducted prior to this publication, which identified five new studies, the final recommendations were revalidated in an online survey in March 2026.

Results

Seven questions were prioritised to address screening and diagnostic challenges in GLILD, resulting in nine recommendations and two good practice statements.

Question 1. Should high-resolution computed tomography (HRCT), positron emission tomography (PET) or magnetic resonance imaging (MRI) of the chest be used to screen for and diagnose GLILD in patients with CVID?

Recommendations

• The panel recommends screening for GLILD in all adults with CVID. (Strong recommendation; low-certainty evidence)

• The panel recommends that children with CVID who are at high risk for GLILD, or who have impaired lung function or respiratory symptoms, undergo regular screening for GLILD. (Strong recommendation; low-certainty evidence)

• In patients with CVID and in whom a decision to screen for GLILD has been made, the panel suggests the use of HRCT of the chest as the first line test for screening or early diagnosis of GLILD over the use of MRI or PET-CT. (Conditional recommendation; very low-certainty evidence)

Background

GLILD is a significant noninfectious complication of CVID. Because GLILD can be asymptomatic in its early stages, imaging is often performed when symptoms emerge. However, timely identification of GLILD is critical to avoid disease progression and enable early intervention when needed. Prior to defining the optimal screening strategy, it is necessary to determine which subgroups of patients with CVID should undergo screening for GLILD. Although direct evidence evaluating the consequences of not screening is lacking, the potential for delayed diagnosis and progression to irreversible lung damage supports a proactive approach to identifying patients at risk. Among screening and early diagnosis strategies, HRCT of the chest is the primary imaging modality used to evaluate patients with suspected ILD [16]. It provides detailed pulmonary imaging that can detect early parenchymal changes, which may be missed by conventional chest X-rays because it is not able to identify subtle early damage. MRI and PET imaging can provide functional information about tissue processes; however, both have limitations compared to CT. MRI lacks the spatial resolution needed to detect early lung damage, while PET-CT, although informative, also falls short in resolution and exposes patients to significantly higher radiation doses than CT [4, 17]. The taskforce assessed whether using HRCT compared to MRI and PET could help in the diagnosis of GLILD.

Evidence assessment

In total, 11 observational studies were included to inform this PICO question [18–28]. Due to the nature of the available studies, it was not possible to estimate the accuracy of the index test (nor PET or MRI of HRCT). The authors do not provide information about the use of a reference test (gold standard) for the diagnosis of GLILD, nor about an independent, blind comparison of the tests with a reference (gold) standard. Based on the identified studies, the summary of findings is provided descriptively, which reflects the limitations in assessing diagnostic accuracy based on the available data. The complete version of the EtD table developed for this question can be found in supplement 6 Q1-EtD.

HRCT

Four observational studies, comprising one prospective cohort and three nested case–control studies, reporting on HRCT use for diagnosis of GLILD were included [18–21]. Maarschalk-Ellerbroek et al. [18] investigated the prevalence of pulmonary abnormalities in adult patients with CVID using thin-slice CT imaging and examined how these findings relate to clinical, functional and immunological markers. Among the 47 patients studied, 51% exhibited significant CT abnormalities, with ILD detected in 34%. The ILD was evaluated using a scoring system based on the presence of opacities, ground-glass changes, septal thickening and pulmonary nodules. Although 34 patients showed some signs of ILD, only 16 reached an ILD score >5, the threshold considered clinically significant. Scarpa et al. [21], in a case–control study, evaluated chest CT features in 38 CVID patients with GLILD and compared them with 38 age- and sex-matched CVID patients without GLILD. They found that patients with GLILD more frequently exhibited bronchiectasis, ground-glass opacities, reticular infiltrate, fibrotic ILD and bands or scars.

Mannina et al. [19] conducted a case–control study to identify clinical predictors of GLILD in patients with CVID. They analysed data from 34 CVID patients with confirmed GLILD and compared them with 52 age- and sex-matched CVID controls without lung disease. They reported that HRCT was highly valuable in identifying characteristic radiographic patterns of GLILD. Notably, these HRCT patterns were consistent regardless of whether the diagnosis was confirmed by surgical lung biopsy, suggesting that in the appropriate clinical context, HRCT alone may support the diagnosis of GLILD. Additionally, Hartono et al. [20] conducted a case–control study to identify clinical predictors of GLILD. They included 26 patients with GLILD confirmed radiologically and/or histologically and compared them to 26 CVID controls without lung disease. In addition to developing a clinical prediction model, they highlighted that HRCT was helpful in distinguishing GLILD cases from controls.

PET

Four case reports were identified in the literature describing the radiological findings of both PET and HRCT [22–25]. No comparative test accuracy studies were identified. The patients included in the studies were young individuals and children between 12 and 32 years old, including four women. All patients were diagnosed with GLILD through lung or lymph node biopsy. The PET reports in all cases showed hypermetabolic pulmonary nodules and adenomegalies. The main findings on HRCT were pulmonary nodules, ground-glass opacities, and mediastinal and cervical adenomegalies. The involvement was mostly in the lower lobes.

MRI

Three small cohort studies investigated the use of MRI for evaluating lung abnormalities in patients with primary immunodeficiencies, including a total of 62 patients, 90% of whom had CVID [26–28]. In all studies, chest MRI was compared to chest CT by two independent observers using the Bhalla scoring system, with an additional item to assess the presence of nodules. The Arslan et al. [26] cohort included 23 patients with primary immunodeficiencies (18 with confirmed CVID), the Milito et al. [27] cohort included 16 CVID patients and two with X-linked agammaglobulinemia, and the Serra et al. [28] cohort included 21 patients with CVID. Across the studies, CT consistently overtook MRI in detecting milder and more peripheral bronchial alterations [28], as well as nodules smaller than 1 cm, which MRI failed to detect in some patients [25]. CT also proved superior in assessing the extent of bronchiectasis and peripheral airway abnormalities [27, 28]. However, for other pulmonary findings, such as larger nodules, bullae, mucus plugs, bronchial wall thickening, the severity of bronchiectasis, and the presence and extent of consolidation, MRI and CT showed comparable performance [26]. Overall, while MRI offered functional imaging benefits, its lower sensitivity limited its ability to detect early or subtle structural changes compared to CT.

Justification

The guideline panel considered HRCT the preferred screening test for GLILD because it is highly sensitive for detecting relevant pulmonary abnormalities and is more practical than alternative imaging modalities. Although the available studies in patients with GLILD do not allow a precise estimate of test accuracy, they consistently show that GLILD is associated with a higher frequency of small nodules, consolidation, interlobular septal thickening and ground-glass opacities than found in patients without GLILD [1, 18–21]. While these findings may appear nonspecific within the broader spectrum of ILDs, their diagnostic value increases when they occur in the appropriate clinical context, particularly in the presence of otherwise unexplained splenomegaly on CT [4]. Taken together, these features provide a more recognisable radiological pattern that can support screening and subsequent diagnostic assessment in patients with CVID. Although screening may reveal more interstitial lung abnormalities, not all findings require further evaluation, and some may only need monitoring [29]. The panel also considered the potential harms of HRCT, particularly radiation exposure [30, 31]. In the targeted population of patients with CVID, however, the panel judged this risk to be low and acceptable in view of the potential consequences of missed or delayed diagnosis, including progression to irreversible lung damage. This recommendation does not apply to patients with known radiosensitivity, in whom the balance of benefits and harms may differ.

When compared with other imaging modalities (MRI and PET), HRCT offered a more favourable balance of feasibility, diagnostic yield and acceptability. MRI avoids ionising radiation, but available evidence suggests that it is less sensitive for detecting pulmonary abnormalities and less practical because it requires patients to remain still for longer periods [26–28]. PET involves greater radiation exposure, has limited availability and is not suitable for routine screening [32, 33]. In addition, both MRI and PET are generally more costly than HRCT. Although the cost of HRCT may be a concern in some settings, especially when it is not covered by the health system or insurance, the panel considered it modest within the broader pathway of disease management. In this context, the potential benefits of reducing underdiagnosis and enabling timely treatment were judged to outweigh the associated costs.

Overall, the certainty of the evidence was very low. Nevertheless, the panel favoured HRCT because it placed a high value on avoiding missed diagnoses of GLILD and facilitating timely intervention when needed. For these reasons, the panel did not support the use of MRI or PET as screening or diagnostic tools for routine practice. Some representative images are shown in figure 1. These illustrate the CT findings described above and help define the radiological profile of GLILD, thereby supporting clinicians in the use of HRCT for screening and diagnostic evaluation in patients with CVID.

FIGURE 1.

Axial and coronal computed tomography (CT) images of a 36-year-old male patient diagnosed with granulomatous-lymphocytic interstitial lung disease (GLILD). a, b) Axial CT images demonstrate diffuse irregular nodules in both lungs which can have a bronchocentric distribution. Consolidation (arrow), as seen in the right middle lobe, is not infrequently present in GLILD. c) The pulmonary nodularity shows lower lobe predominance. Patchy regions of ground glass opacification (arrowheads) are also commonly seen in GLILD. d) Though mediastinal lymph node enlargement, commonly associated with GLILD, is a relatively non-discriminatory finding, the presence of splenomegaly (arrow) increases the confidence of a diagnosis of GLILD.

Subgroup considerations

In children, special consideration must be applied owing to a different risk–benefit ratio. Because of the potential side effects of radiation, HRCT screening is restricted to children with suspected increased risk of GLILD based on (a combination of) clinical, genetic, laboratory or PFT information. In patients with known radiosensitivity, MRI would be the preferred choice.

Implementation considerations

Implementation of HRCT for GLILD requires attention to several considerations. Although access to HRCT is generally not the main limitation, accurate interpretation of GLILD-related findings often depends on specialised expertise. Its effective use therefore requires trained radiologists and clinicians, multidisciplinary review when appropriate, and continued familiarity with evolving diagnostic criteria. Standardised acquisition protocols and quality assurance measures should also be adopted to minimise radiation exposure, particularly with the increasing availability of low-dose CT techniques. Inspiratory and expiratory CT imaging in the supine position is recommended to detect air trapping, whereas routine prone imaging is not recommended because it adds radiation without sufficient justification in this radiation-sensitive population. In this guideline, HRCT is considered only for initial screening and diagnostic evaluation; follow-up imaging is beyond its scope. Nevertheless, when repeated imaging is needed, clinicians should balance cumulative radiation exposure against the anticipated clinical benefit. This guideline also does not address the use of PET-CT in specific contexts, such as the differential diagnosis of malignancy.

Question 2: Should the gas transfer test, spirometry, lung volumes (plethysmography) or a combination of these be used to diagnose GLILD in patients with CVID?

Recommendations

• In patients with CVID in whom GLILD is suspected, the panel suggests performing PFTs including assessment of gas transfer to determine the severity of lung function impairment. (Conditional recommendation; very low-quality evidence)

Recommendation remarks

• No PFT confirms the presence of GLILD, and PFTs can be normal in the presence of GLILD.

Background

Spirometry is a widely used, noninvasive test for the assessment of lung function. Guidelines for its performance have been standardised by the ERS and the American Thoracic Society [34]. In patients with ILD, spirometry often shows a restrictive ventilatory pattern. Despite its value, spirometry may miss early changes in gas exchange. The diffusing capacity of the lungs for carbon monoxide (D_LCO) is considered more sensitive in detecting ILD, which is primarily attributed to a loss of alveolar surface area and thickening of the alveolar–capillary barrier due to inflammation and fibrosis [32, 35]. We assessed the role of PFTs in the diagnosis of GLILD.

Evidence assessment

The summary of findings is based on 11 studies identified to address this question, including six retrospective cohort studies [1, 17, 19, 36–43]. None of the papers provided data suitable for estimating diagnostic accuracy. The pulmonary function profiles of patients with GLILD were heterogeneous. Reported patterns include restrictive, obstructive, normal spirometry or a mixed pattern of both obstruction and restriction [3, 4]. The evaluated studies that directly compared GLILD with CVID patients without GLILD [1, 19, 36, 37, 40, 42, 43] consistently showed that patients with GLILD exhibit worse results in PFTs than CVID patients without GLILD.

Two studies incorporated PFT parameters into predictive models for identifying GLILD. Mannina et al. [19] included forced vital capacity <80% of the predicted value, along with splenomegaly, autoimmune cytopenia and low serum IgA levels, achieving an area under the curve (AUC) of 0.92. Similarly, Cinetto et al. [36] used the % predicted value of D_LCO alongside splenomegaly, autoimmune cytopenia and an increased percentage of CD21^low B-cells in their model, which demonstrated even higher performance, with an AUC of 0.98.

None of the studies reported on the outcomes of interest. As a result, the quality of evidence was rated as very low. The complete version of the EtD table developed for this question can be found in supplement 6 Q2-EtD.

Justification

The guideline panel considered PFTs useful for assessing pulmonary involvement and the severity of GLILD. The combined use of spirometry, gas transfer and lung volumes is preferred, because each test provides complementary information on different aspects of lung function. However, studies have shown that PFTs alone are insufficient to distinguish GLILD from other lung diseases, because GLILD can be present even when lung function tests are normal [1, 19, 21, 36, 40].

Current evidence indicates that gas transfer is the most sensitive test for monitoring the progression of ILDs [35]. When incorporated into noninvasive diagnostic scores, altered gas transfer has been shown to improve the predictive performance for GLILD diagnosis [36]. These findings highlight the potential utility of lung function parameters, particularly gas transfer measurements, in supporting the early identification of GLILD. The cost and feasibility of PFTs are not a concern, because these tests are routinely available in all pulmonary function laboratories.

Subgroup considerations

In paediatric patients with CVID, PFTs can generally be performed from the age of 6–7 years.

Implementation considerations

Implementation of PFTs should account for patient-specific characteristics and the healthcare context to minimise barriers and practical challenges, with particular attention to children and individuals with disabilities. The use of PFTs in managing GLILD, primarily for assessing disease progression, severity, treatment response and prognosis, falls outside the scope of this guideline, which focuses on initial diagnosis and classification.

Question 3: Should blood biomarkers such as B-cell-activating factor (BAFF) or IgM in serum or the soluble form of the interleukin-2 receptor (sIL-2R) be used to diagnose GLILD in patients with CVID?

Recommendation

• In patients with CVID in whom GLILD is suspected, the panel does not suggest routine assay of blood biomarkers as part of the diagnostic assessment. (Conditional recommendation; very low-certainty evidence)

Recommendation remarks

• If biomarker testing is considered necessary and available, measurement of sIL-2R is currently preferred over other biomarkers.

Background

The use of biomarkers as early warning systems in the evaluation of disease risk has increased markedly in the last decade. Biomarkers are indicators of typical biological processes, pathogenic processes or pharmacological reactions to therapy. The application and identification of biomarkers in the medical and clinical fields have an enormous impact on society; therefore, we assessed the role of biomarkers in the diagnosis of the GLILD.

Evidence assessment

A total of 12 studies addressed this question [19–21, 36, 40, 43–49]. Of these, nine case–control studies compared biomarkers between GLILD and non-GLILD groups in a total of 593 CVID patients with ILD. Three studies assessed CVID patients with varying severity of ILD, rather than specifically focusing on GLILD. None of the studies reported on the outcomes of interest. The available evidence did not allow for formal test accuracy estimates. Therefore, the findings are summarised descriptively. In the included studies, levels of BAFF, IgM and sIL-2R (also known as sCD25) were measured as part of patient diagnosis.

The included studies consistently reported lower serum levels of IgA, IgG or IgM in patients with GLILD than in CVID patients without GLILD. However, none of the studies identified significant differences in IgM levels specifically between GLILD and non-GLILD CVID patients. Evidence regarding serum BAFF was inconsistent: one study reported elevated BAFF levels in CVID patients with progressive ILD compared to CVID patients with stable ILD [45], while another found no elevation in GLILD patients compared with other CVID patients [44]. A similar pattern was observed with sIL-2R, which was found to be elevated in GLILD patients in two studies [47, 50] but not in a separate cohort study involving seven patients [46]. Overall, the certainty of evidence was very low. The complete version of the EtD table developed for this question can be found in supplement 6 Q3-EtD.

Justification

The guideline panel considered that the current literature does not provide sufficient evidence for the routine use of blood biomarkers in the diagnosis of GLILD [21, 36, 40, 45, 46]. Although the evidence is uncertain, the absence of elevated levels of sIL-2R may suggest the absence of GLILD, indicating its potential diagnostic value [47, 50]. By contrast, very low-certainty evidence indicates that IgM levels might be altered between GLILD cases and controls, and BAFF levels may not differ between patients with and without GLILD at diagnosis [44]. Therefore, BAFF measurement is unlikely to provide additional diagnostic value for GLILD in patients with CVID-ILD. The panel also considered that patient and healthcare professional acceptance is anticipated to be adequate; however, the feasibility of implementing these measurements in the diagnosis of GLILD would depend on the biomarker and may increase healthcare costs. Therefore, incorporating blood biomarker measurements into the diagnostic process may adversely affect health equity.

Subgroup considerations

Although data on paediatric patients are very limited, there were no relevant arguments to make different recommendations for children.

Implementation considerations

The panel has not suggested routinely measuring biomarkers as part of the diagnostic assessment. Because biomarker measurement is not common practice, implementation barriers are anticipated. This guideline has not evaluated the use of cellular phenotyping of blood cells or other serum markers in diagnosis of GLILD.

Research considerations

Several areas of uncertainty about use of biomarkers in GLILD patients are identified where further research is needed. The first of these is the role of biomarkers in diagnosis. Future research should focus on evaluating biomarkers at the time of diagnosis to establish a baseline for monitoring disease activity and progression in GLILD. Although beyond the scope of this guideline, tracking changes in these biomarkers over time could provide valuable insights into the disease's evolution and response to treatment.

The second area of future research is an investigation of additional biomarkers. Research should explore additional biomarkers not covered in current evaluations, such as soluble T-cell immunoglobulin and mucin-domain containing-3 (sTIM3), as well as those investigated by Fraz et al. [44] and Berbers et al. [51]. These biomarkers might enhance the diagnostic accuracy and understanding of GLILD.

A third area of future research is the role of sIL-2R and negative predictive value should be explored. Further work is needed to understand the role of biomarkers such as sIL-2R, which shows increased values in a very high percentage of patients with GLILD, suggesting a negative predictive value of normal sIL-2R for GLILD. This could refine diagnostic criteria and help rule out the disease more effectively.

Finally, diagnostic scores should be developed. Investigating how single biomarkers might contribute to a comprehensive diagnostic score could improve the accuracy and reliability of GLILD diagnosis. Integrating multiple biomarkers into a scoring system, as suggested by Cabanero-Navalon et al. [40], Scarpa et al. [21] and Cinetto et al. [36], could enhance diagnostic precision and guide treatment decisions.

Question 4a: Should bronchoalveolar lavage (BAL) be used to exclude infection as part of the diagnosis of GLILD in patients with CVID?

Good practice statement

• Healthcare professionals should exclude pulmonary infection in patients with CVID in whom GLILD is suspected. (Ungraded good practice statement)

Recommendation

• In patients with CVID in whom GLILD is suspected, the panel suggests the use of bronchoscopy and BAL to exclude infection as part of the diagnostic process. (Conditional recommendation; low-certainty evidence)

Background

Patients with CVID are highly susceptible to pulmonary infections, with respiratory tract infections often serving as both a primary presenting feature and a major cause of morbidity and mortality. The most commonly identified respiratory pathogens include Staphylococcus aureus, Haemophilus influenzae, Streptococcus pneumoniae, rhinovirus and cytomegalovirus [4]. ILD affects the lung parenchyma rather than the bronchial airways, so patients do not produce sputum because of this manifestation. The resulting inflammation and scarring cause a dry, nonproductive cough, because there is little or no mucus in the small airways or alveoli to be expectorated. To obtain adequate samples for infection confirmation, bronchoscopy with BAL is routinely used in these patients. BAL allows for the detection of bacteria, mycobacteria, fungi and respiratory viruses and is therefore the preferred method for excluding infection in any ILD. While sputum induction is possible, it is less accurate and frequently results in an inadequate sample.

Evidence assessment

As with the limitations noted in the previous PICO questions, the diagnostic accuracy of BAL tests for GLILD in CVID could not be assessed. Six studies performed in CVID patients were identified to inform this question, two case series [22, 52] and four case reports [53–56]. All of them were rated 7 or >7 points in the risk of bias assessment following the JIB instruments [12, 57]. Another five studies identified for this question were not included as part of this synthesis because of the high risk of bias in their reports [23, 58–61]. Studies on the use of BAL cellular analysis (e.g. cell counts and differentials, lymphocyte subsets or soluble components) or to identify cell patterns highly consistent with GLILD are evaluated as part of another question in this guideline.

None of the six included studies in GLILD patients (n=14) reported the outcomes of interest or identified evidence of infection (viral, bacterial, fungal or mycobacterial) based on culture results. In our previous systematic review on diagnostic testing for GLILD, we found that BAL was primarily used to exclude infection [4]. Indirect evidence also supports the use of BAL to confirm or rule out pulmonary infections, including ventilator-associated pneumonia [62], SARS-CoV-2 infection [63] and infections in patients with haematological malignancies [64]. The diagnostic accuracy of fluid testing for different pathogens depends on the specific method used, such as culture, PCR, ELISA or galactomannan.

The complete version of the EtD table developed for this question can be found in supplement 6 Q4a-EtD.

Justification

The panel considered lower respiratory tract infection an essential differential diagnosis in patients with CVID and suspected GLILD, given that immunosuppressive treatment may be harmful if infection is present. Flexible bronchoscopy with BAL is widely used in the evaluation of pulmonary disease and is generally considered safe. Although direct evidence in GLILD is limited, indirect evidence supports its role in excluding infection. The anticipated benefit lies in enabling the identification and treatment of infection before initiating immunosuppressive therapy for GLILD. In addition, BAL findings may inform both short- and long-term treatment decisions, including the need for antimicrobial therapy because airway disease and ILD frequently coexist in these patients. In the broader context of ILD, BAL also provides important clinical information by excluding infection and identifying cytological patterns associated with specific subtypes, such as alveolar haemorrhage, diffuse alveolar damage and organising pneumonia [21].

Subgroup considerations

While BAL is generally considered a safe procedure, there are subgroups of patients in whom the risk of adverse effects is increased. Although data on paediatric patients are very limited, there were no relevant arguments to make different recommendations for children.

Implementation considerations

The panel suggests excluding bacterial (culture), fungal (antigen detection, culture), mycobacterial (Ziehl-Neelsen staining, PCR and culture), Pneumocystis jirovecii (staining and PCR) and common respiratory viral (PCR) infections. Given the presence of immunodeficiency, additional pathogens might need to be considered. Screening for pathogens may also vary depending on the region and should be adapted accordingly. It is essential to ensure that the laboratory has well-defined protocols for handling BAL to achieve maximum accuracy in the diagnostic process.

Question 4b: Should BAL with differential cell count and flow cytometry in addition to the clinical and radiological assessment be used to diagnose CVID-related GLILD instead of biopsy?

Recommendations

• In patients with CVID in whom GLILD is suspected, the panel suggests performing a differential cell count on BAL fluid (BALF) to support the diagnosis of GLILD. (Conditional recommendation; very low-certainty evidence)

Background

While lung biopsy is considered the gold standard in unclassifiable ILDs, it carries procedural risks and may not be suitable for all patients. Aside from using BAL to exclude infection in GLILD, some studies suggest that significant lymphocyte expansion, especially increased CD21^low B-cells, may support a GLILD diagnosis [36, 65]. However, immunological analysis of BALF is not routinely performed owing to uncertainty about its diagnostic value. It has been proposed that broader use of BALF differential cell counts and lymphocyte phenotyping could improve diagnostic accuracy and potentially guide treatment, especially when biopsy is not feasible [4, 19, 65].

Evidence assessment

Three studies were identified to inform this question [19, 36, 65]; none of them provided data on the outcomes of interest nor information suitable for estimating diagnostic accuracy. Therefore, the findings are presented descriptively.

Mannina et al. [19] and Cinetto et al. [36] reported findings from BALF flow cytometry, highlighting alveolar lymphocytosis. The study by Friedmann et al. [65] involved the retrospective analysis of BALF samples from patients with GLILD, comparing them with samples from patients with sarcoidosis and healthy controls to characterise immune cell populations and cytokine profiles. The analysis showed that BALF from patients with GLILD contained expanded lymphocyte counts in 83%, including an increased abundance of CD21^low B-cells, T-follicular helper-1 and T-peripheral helper-like T-cells, suggesting active local immune interactions. Indirect evidence from studies on the harmful effects of BAL in other populations was used to inform the safety outcomes of the test. These studies, conducted on other cohorts of patients, suggest BAL may be safer than lung biopsy [66].

The complete version of the EtD table developed for this question can be found in supplement 6 Q4b-EtD.

Justification

The panel considered BAL differential cell count a useful component of the diagnostic evaluation for suspected GLILD. BAL lymphocytosis, reflecting lymphocytic alveolitis, is a frequent finding in GLILD [65, 67, 68]. Although not disease-specific, a BAL lymphocyte count of ≥15% can support the diagnosis in patients with a compatible clinical, radiological and functional profile, whereas its absence should prompt consideration of alternative diagnoses [65, 67–71]. Compared with lung biopsy, BAL differential cell count offers a less invasive, lower-risk, more acceptable and more widely available alternative at lower cost. This advantage is particularly relevant given the heterogeneity of GLILD, which may limit the accessibility and representativeness of biopsy specimens. However, the panel recognises that BAL lymphocytosis alone does not establish the diagnosis of GLILD and must be interpreted in the context of the overall assessment. The panel also concluded that the current evidence on additional BAL immunophenotyping markers, including expansion of CD21^low B-lymphocytes, reduction of class-switched memory B-cells and altered T-cell homeostasis, is insufficient to support a diagnostic recommendation. Although of research interest, their added diagnostic value remains uncertain, availability is limited, quality control requirements are greater, costs are higher and their results do not currently guide management.

Subgroup considerations

While BAL is generally considered a safe procedure, there are subgroups of patients in whom the risk of adverse effects is increased. Although data on paediatric patients are very limited, there were no relevant arguments to make different recommendations for children.

Implementation considerations

Given that BAL has been recommended for exclusion of infection, differential cell counts can be easily added concurrently, saving time and effort in the diagnostic evaluation of GLILD.

Research considerations

Future research should evaluate the incremental benefit of flow cytometric testing in the diagnostic evaluation of GLILD, particularly on whether it improves diagnostic confidence, accuracy or clinical decision-making sufficiently to justify the additional costs and technical requirements.

Question 5b: Should transbronchial biopsy (TBB) versus transbronchial lung cryobiopsy (TBLC) versus video-assisted thoracoscopic surgical biopsy (VATS) be used to diagnose CVID-related GLILD?

Good practice statement

• Healthcare professionals should ensure that CVID patients in whom GLILD is suspected are evaluated by a multidisciplinary team skilled in the diagnosis of GLILD. (Ungraded good practice statement)

Recommendations

• In patients with CVID in whom GLILD is suspected, the panel does not suggest routinely performing a lung biopsy. (Conditional recommendation; low-certainty evidence)

• In patients with CVID in whom GLILD is suspected, the panel suggests performing a lung biopsy to establish a definitive diagnosis only when radiological findings are unclear or alternative diagnoses are being considered. (Conditional recommendation; very low-certainty evidence)

Recommendation remarks

• The use of biopsy for diagnosis of GLILD needs to be weighed against the risks of the procedure and include factors such as the general and lung-specific condition of the patient, location and size of lesions and available technical expertise.

• Clinicians should consider the presence of other conditions when patients present with atypical manifestations or significant systemic symptoms.

• A negative biopsy cannot reliably exclude GLILD given the possibility of sampling error.

• The decision on which biopsy modality to use should be made collaboratively by a multidisciplinary team, weighing all factors to optimise diagnostic accuracy while minimising risks to the patient.

Background

Diagnosis of GLILD typically relies on clinical, functional and radiological findings, while the role of lung biopsy remains uncertain. Although biopsy provides a definitive diagnosis, it is unclear whether it is superior to diagnostic procedures without lung biopsy. An overview of histopathological features consistent and inconsistent with GLILD is provided in table 2.

TABLE 2.

Histopathological features consistent or inconsistent with granulomatous-lymphocytic interstitial lung disease

Consistent features	Inconsistent features (consider alternate diagnosis)
Lymphoid hyperplasia, typically around small airways (follicular bronchiolitis) and/or distributed throughout the interstitium; may also have foci of nodular lymphoid hyperplasia	Features suggestive of lymphoma (clonal expansion of lymphocytes); increased IgG4 expressing plasma cells Consider lymphocytic interstitial pneumonia; requires multidisciplinary team discussion
Granulomas: usually poorly formed and/or scattered giant cells	Necrotising granulomas (consider infection) Well-formed granulomas in a lymphangitic distribution (consider sarcoidosis) Airway-centred poorly formed granulomas with absence of lymphoid hyperplasia (consider hypersensitivity pneumonia)
Organising pneumonia, usually mild to moderate	Preponderance of organising pneumonia (consider infection, autoimmune disease, cryptogenic and others)
Interstitial fibrosis, may be mild to severe with honeycomb cysts	Requires multidisciplinary discussion to rule out other fibrosing lung disease (most commonly usual interstitial pneumonia, nonspecific interstitial pneumonia)

Biopsy methods vary in invasiveness and diagnostic yield: TBB is less invasive but offers limited tissue, while VATS yields more comprehensive samples at higher procedural risk. TBLC may offer a balance between the two. The impact of disease duration and histological findings on diagnostic decisions is not well defined, and the absence of biopsy may lead to missed or incorrect diagnoses.

Evidence assessment

Five studies that assess the use of biopsy in GLILD were identified to inform this question, two retrospective case–control studies [19, 20], two retrospective cohort studies [72, 73] and one cross-sectional case–control study [74]. The most frequently reported histological patterns were lymphoid interstitial pneumonitis, granulomatous inflammation, organising pneumonitis and lymphoid hyperplasia [4]. A recent study by van Stigt et al. [74] evaluated the histopathological and spatial proteomic characteristics of granulomas in patients with CVID using lymph node and skin biopsies. They report that CVID-associated granulomas are smaller, less well circumscribed, and show minimal fibrosis and multinucleated giant cells compared with sarcoidosis, tuberculosis and pseudo-sarcoidosis. They further showed that CVID granulomas display a distinct inflammatory profile, with increased infiltration of CD3⁺/CD4⁺ T-cells, CD20⁺ B-cells, CD163⁺ macrophages and myeloperoxidase⁺ neutrophils, supporting a unique and more inflammatory pathogenesis. Although no lung biopsies were included, the findings suggest that histological features may help differentiate CVID-related granulomatous disease from other granulomatous disorders, which may be of diagnostic relevance in patients when ILD is diagnosed before CVID.

No studies comparing the diagnostic accuracy of TBB versus VATS versus TBLC in patients with CVID and suspected GLILD were found. However, two systematic reviews provide indirect evidence from broader ILD populations [37, 75]. Sharp et al. [37] conducted a systematic review and meta-analysis comparing the diagnostic yield of TBLC, forceps TBB and VATS lung biopsy in adults with ILD. The pooled diagnostic yields were TBLC 84.4% (95% CI 75.9–91.4%), forceps TBB 64.3% (95% CI 52.6–75.1%) and VATS 91.1% (95% CI 84.9–95.7%). For forceps TBB, no deaths were reported. The mortality was 0.5% and 2.3% for TBLC and VATS, respectively. Similarly, Rodrigues et al. [75] reviewed 23 studies on TBLC and 10 on VATS, reporting diagnostic yields of 80.7% (95% CI 71.6–87.4%) for TBLC in centres that performed ≥70 procedures and 93.5% (95% CI 88.3–96.5%) for VATS. The 30-day mortality rate was 0.6% (range 0–3.2%) for TBLC and 1.7% (range 0–6.7%) for VATS. These findings suggest that VATS has the highest diagnostic yield, leading to a higher percentage of patients correctly diagnosed with ILD, but imposes a high risk of complications.

The complete version of the EtD tables developed for these questions can be found in supplement 6 Q5a-EtD and Q5b-EtD.

Justification

The panel considered that while biopsy has historically been regarded as the gold standard for a definitive diagnosis of GLILD, current experience and modern practice increasingly favour noninvasive approaches, particularly when HRCT findings strongly suggest GLILD and are interpreted by experienced radiologists [76]. Evidence from Mannina et al. [19] indicates that among 34 GLILD patients, 19 with histological confirmation via surgical biopsy and 15 without, there were no significant differences in HRCT patterns between the groups, and HRCT features and forced vital capacity remained similar over time. The panel noted that biopsies are most appropriately reserved for cases with diagnostic uncertainty, such as suspected alternative conditions including lymphoma. Sampling error means that even a negative biopsy does not definitively exclude GLILD. In most experienced centres, biopsy is performed in <50% of cases [6], reflecting increasing reliance on the combination of radiological and clinical assessments to guide diagnosis and management, thereby reducing the need for invasive procedures. Biopsies infrequently yield unexpected findings, and the literature offers limited guidance on when biopsy results meaningfully influence therapy [18–20, 72, 77]. The panel highlighted the potential risks and challenges associated with biopsy, including procedure-related complications, patient concerns, logistical limitations in resource-limited centres, increased costs and impacts on health equity, particularly for patients in remote areas. Ethical considerations reinforce the need to balance these risks against uncertain benefits, emphasising patient autonomy and prioritising noninvasive diagnostic strategies whenever appropriate.

Subgroup considerations

Although data on paediatric patients are very limited, there were no relevant arguments to make different recommendations for children.

Implementation considerations

As reflected in the good practice statement, the panel regards the implementation of a multidisciplinary team skilled in the diagnosis of GLILD, including pulmonologists, clinical immunologists and an experienced radiologist, as crucial in the decision-making process regarding the need for a lung biopsy. The choice of biopsy modality for diagnosing conditions such as GLILD in patients with CVID is highly case specific and should be tailored to individual patient needs and circumstances, because no comparative data in GLILD exist to support a strong recommendation in favour of any single modality [77]. Guidance on electing the optimal sampling method is provided in table 3. Key considerations include the location of the lesion within the lung, the accuracy and reliability of the biopsy method, the availability of the procedure at the medical centre and the expertise of the healthcare team. Additionally, ensuring that an adequate sample of lung tissue is obtained is crucial for making an accurate diagnosis. The potential harms associated with different biopsy procedures must also be considered, although specific risks related to CVID may not be well-differentiated.

TABLE 3.

Pros and cons of transbronchial biopsy (TBB) versus transbronchial cryobiopsy (TBCB) versus surgical lung biopsy (SLB) via video-assisted thoracoscopic surgery (VATS)

Procedure	Pros	Cons
TBB	Safe, minimally invasive	Small specimen (2–3 mm), often with crush artefact
	Most helpful in ruling out infections	Low diagnostic yield
	Better for peribronchial abnormalities	Sampling of periphery is difficult
		Heterogeneous diseases not adequately sampled
TBCB	Larger sample size, but still small (5–7 mm)	Sample size does not allow for assessment of all lung compartments
	Lower morbidity than surgical lung biopsy in experienced hands	May have higher complication rates than transbronchial biopsy
	Less crush artefact	Poor sampling of pleural/subpleural regions
SLB via VATS	Gold standard for heterogeneous diseases	May not be tolerated by some patients
	Much larger sample size (20–30 mm) and two to three lobes typically biopsied	Requires experienced thoracic surgeon
	Able to assess all lung compartments involved in parenchymal lung diseases	Mostly samples peripheral lung; not good for large airway disease assessment
	Combined with multidisciplinary discussion has a high diagnostic yield (90%+)

Research considerations

More research is needed to refine the criteria for biopsy, ensuring it is reserved for cases in which it will meaningfully impact GLILD diagnosis and treatment.

Conclusions

The diagnosis of GLILD in patients with CVID requires a thoughtful and systematic approach that balances accuracy with minimal invasiveness. The recommendations provided in this guideline are intended to standardise diagnostic practices across healthcare settings, thereby improving the identification and management of GLILD in the CVID population. Ongoing research and clinical feedback will be essential to further refine these guidelines and ensure they continue to meet the needs of this complex patient group.

Diagnostic pathway and implementation of recommendations

For clinical practice, table 4 provides an overview of the recommendations and good practice statements. Schematic flow charts of the screening and diagnosis of GLILD for adult and paediatric CVID patients are provided in figure 2.

TABLE 4.

Guideline recommendations on screening and diagnosis

PICO question	Recommendations and good practice statements
1: Should HRCT, PET or MRI of the chest be used to screen for and diagnose GLILD in patients with CVID?	• The panel recommends screening for GLILD in all adults with CVID. (Strong recommendation; low-certainty evidence) • The panel recommends that children with CVID who are at high risk for GLILD, or who have impaired lung function or respiratory symptoms, undergo regular screening for GLILD. (Strong recommendation; low-certainty evidence) • In patients with CVID and in whom a decision to screen for GLILD has been made, the panel suggests the use of HRCT of the chest as the first line test for screening or early diagnosis of GLILD over the use of MRI or PET-CT. (Conditional recommendation; very low-certainty evidence)
2: Should the gas transfer test, spirometry, lung volumes (plethysmography) or a combination of these be used to diagnose GLILD in patients with CVID?	• In patients with CVID in whom GLILD is suspected, the panel suggests performing PFTs including assessment of gas transfer to determine the severity of lung function impairment. (Conditional recommendation; very low-quality evidence)
3: Should blood biomarkers such as B cell-activating factor or IgM in serum or the soluble form of the interleukin-2 receptor be used to diagnose GLILD in patients with CVID?	• In patients with CVID in whom GLILD is suspected, the panel does not suggest routine assay of blood biomarkers as part of the diagnostic assessment. (Conditional recommendation; very low-certainty evidence)
4a: Should BAL be used to exclude infection as part of the diagnosis of GLILD in patients with CVID?	• Healthcare professionals should exclude pulmonary infection in patients with CVID in whom GLILD is suspected as part of the diagnosis process. (Ungraded good practice statement) • In patients with CVID in whom GLILD is suspected on HRCT, the panel suggests the use of bronchoscopy and BAL to exclude infection as part of the diagnostic process. (Conditional recommendation; low-certainty evidence)
4b: Should BAL with differential cell count and flow cytometry in addition to the clinical and radiological assessment be used to diagnose CVID-related GLILD instead of biopsy?	• In patients with CVID in whom GLILD is suspected, the panel suggests performing a differential cell count on BAL fluid to support the diagnosis of GLILD. (Conditional recommendation; very low-certainty evidence)
5a: Should lung biopsy be used to diagnose GLILD in patients with CVID? 5b: Should transbronchial biopsy versus transbronchial lung cryobiopsy versus video-assisted thoracoscopic surgical biopsy be used to diagnose CVID-related GLILD?	• Healthcare professionals should ensure that CVID patients in whom GLILD is suspected are evaluated by a multidisciplinary team skilled in the diagnosis of GLILD. (Ungraded good practice statement) • In patients with CVID in whom GLILD is suspected, the panel does not suggest routinely performing a lung biopsy. (Conditional recommendation; low-certainty evidence) • In patients with CVID in whom GLILD is suspected, the panel suggests performing a lung biopsy to establish a definitive diagnosis only when radiological findings are unclear or alternative diagnoses are being considered. (Conditional recommendation; very low-certainty evidence)

BAL: bronchoalveolar lavage; CVID: common variable immunodeficiency disorders; GLILD: granulomatous-lymphocytic interstitial lung disease; HRCT: high-resolution computed tomography; MRI: magnetic resonance imaging; PET: positron emission tomography; PFT: pulmonary function test.

FIGURE 2.

Screening and diagnostic approach for patients with common variable immunodeficiency disorders (CVID) and suspected granulomatous-lymphocytic interstitial lung disease (GLILD). Screening and diagnostic approach in a) adult and b) paediatric patients. BAL: bronchoalveolar lavage; HRCT: high-resolution computed tomography of the chest; IEI: inborn errors of immunity; PFT: pulmonary function tests. ^#: see figure 1 for typical GLILD HRCT characteristics. ^¶: BAL with microbiology and cytology should be used to rule out infection; the panel suggests excluding bacterial (culture), fungal (antigen detection, culture), mycobacterial (Ziehl-Neelsen staining, PCR and culture), Pneumocystis jirovecii (staining and PCR) and common respiratory viral (PCR) infections. Given the presence of immunodeficiency, additional pathogens might need to be considered. Screening for pathogens may also vary depending on the region and should be adapted accordingly. Spontaneous or induced sputum are an option to exclude infection. Cytology consists of leukocyte+differential and may additionally be used to assess alveolar haemorrhage and certain types of malignancies, including low-grade lymphoma. ⁺: see tables 2 and 3 for guidance on histopathological features and selecting diagnostic modality. ^§: a negative biopsy cannot reliably exclude GLILD given the possibility sampling error. ^ƒ: risk factors in the clinical, laboratory, genetic and pulmonary function domain, or a combination thereof. Examples include but are not limited to the presence of pulmonary symptoms and/or impaired pulmonary function, presence of other noninfectious disease manifestations such as autoimmune cytopenia, lymphoproliferation; laboratory risk profiles including low class-switched memory B-cells and/or high fractions of CD21^low B-cells; established molecular diagnosis associated with an increased risk of noninfectious disease manifestations such as variants affecting the function of CTLA4 and LRBA.

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

ERJ-02224-2025.Supplement