Sarcoidosis: Epidemiology and clinical insights

Sarcoidosis is characterized by noncaseating granulomas which form in almost any part of the body, primarily in the lungs and/or thoracic lymph nodes. Environmental exposures in genetically susceptible individuals are believed to cause sarcoidosis. There is variation in incidence and prevalence by region and race. Males and females are almost equally affected, although disease peaks at a later age in females than in males.

Sarcoidosis is characterized by noncaseating granulomas which form in almost any part of the body, primarily in the lungs and/or thoracic lymph nodes. Environmental exposures in genetically susceptible individuals are believed to cause sarcoidosis. There is variation in incidence and prevalence by region and race. Males and females are almost equally affected, although disease peaks at a later age in females than in males.
The heterogeneity of presentation and disease course can make diagnosis and treatment challenging. Diagnosis is suggestive in a patient if one or more of the following is present: radiologic signs of sarcoidosis, evidence of systemic involvement, histologically confirmed noncaseating granulomas, sarcoidosis signs in bronchoalveolar lavage fluid (BALF), and low probability or exclusion of other causes of granulomatous inflammation. No sensitive or specific biomarkers for diagnosis and prognosis exist, but there are several that can be used to support clinical decisions, such as serum angiotensin-converting enzyme levels, human leukocyte antigen types, and CD4 Vα2.3+ T cells in BALF. Corticosteroids remain the mainstay of treatment for symptomatic patients with severely affected or declining organ function. Sarcoidosis is associated with a range of adverse long-term outcomes and complications, and with great variation in prognosis between populations.
New data and technologies have moved sarcoidosis research forward, increasing our understanding of the disease. However, there is still much left to be discovered. The pervading challenge is how to account for patient variability. Future studies should focus on how to optimize current tools and develop new approaches so that treatment and follow-up can be targeted to individuals with more precision.

Introduction
Sarcoidosis is a multisystem inflammatory disease characterized by the formation of noncaseating granulomas in affected organs, and the lungs and/or thoracic lymph nodes are engaged in more than 90% of all cases [1]. The cause or causes of sarcoidosis are still unknown; an interaction between several genetic variants and environmental exposures is thought to trigger and sustain granulomatous inflammation and clinical disease. The heterogeneity of onset and disease course is a major challenge to understanding the etiology of sarcoidosis. Variation in the incidence and prevalence by age, sex, race, and geographical region provides clues about the drivers of disease occurrence.

Epidemiology of sarcoidosis
The annual incidence of sarcoidosis varies between 1 and 15 per 100,000 depending on the region. The rates are lowest in Eastern Asian countries (0.5-1 per 100,000) [2,3], higher in North America and Australia (5-10) [4][5][6], and highest in Northern European (Scandinavian) countries (11)(12)(13)(14)(15) [7,8]. In Southern Europe, estimates are lower than in the North (Fig. 1) [9]. Incidence and prevalence vary even within countries, likely owing to differences in genetics, environmental exposures, or in the way sarcoidosis is detected and diagnosed. To this day, the burden of sarcoidosis is still unknown in most parts of the world because of the presence of mimicking diseases (e.g., tuberculosis), lack of diagnostic technology, expertise, and limited case registration.
The average age at diagnosis is around 50 years. Several studies, but not all, show a bimodal distribution of age primarily related to sex [9]. Even though many studies report a female to male ratio of 1:1, more males than females are diagnosed at 20-45 years old, whereas incidence peaks in females later (at 50-65 years old) [7].
Studies from the United States show that Black Americans have the highest incidence of sarcoidosis (17.8 per 100,000) compared to White (8.1), Hispanic (4.3), and Asian Americans (3.2) [5]. Similarly, in France, the highest incidence is in individuals of Afro-Caribbean and North African ancestry (16.9 and 9.7 per 100,000 per year, respectively) [10]. We should emphasize that ancestry and race are inadequate proxies for genetic risk and can reflect social inequities in health [11,12]. Most studies have focused on nondiverse popu-lations, limiting our understanding of the role of race.

Etiology of sarcoidosis
What causes sarcoidosis is likely multifactorial. It is thought that several different pathways, involving both genetic and environmental exposures, cause a dysregulated immune system to form granulomas and fail to resolve them. This complex process is not completely understood, but it is believed that environmental exposures may influence the disease risk depending on genetics through direct or indirect pathways (Fig. 2).

Genetics
Several human leukocyte antigen (HLA) alleles are associated with sarcoidosis susceptibility, especially in the class II region [13,14]. Of note, the strong linkage disequilibrium in the HLA region hampers our efforts to identify specific alterations involved in sarcoidosis occurrence and progression [13]. There are several associations with non-HLA genes as well [13]. Genetic factors influence not only the susceptibility to develop sarcoidosis, but also some distinct clinical phenotypes [15,16]. One such phenotype, Löfgren syndrome (LS), is seen in about 30% of cases in Sweden [17] but is less common in other parts of the world. It presents acutely but has a benign course, with disease resolution observed usually within 2 years. Because of its distinct clinical features and its strong association with HLA-DRB1*03 [14], we tend to divide sarcoidosis phenotypes into LS and non-LS.

Antigens
Several causative agents of sarcoidosis have been suggested, including self-antigens (such as vimentin [18]), as well as organic and inorganic particles [19][20][21]. Recently, pathogenic fungi, such as Aspergillus nidulans, were proposed as a causative agent of LS [22]. Among bacterial antigens, Cutibacterium acnes was identified in cultures of sarcoid lesions, and remnants of Mycobacterium tuberculosis were found in granulomas [23,24]. Epidemiological studies on the role of these and other infections show conflicting results [25,26]. It is important to acknowledge that bias due to reverse causation is likely because latent sarcoid inflammation increases the risk of infections before diagnosis [26]. In addition, several immunotherapeutic agents are associated with sarcoid-like reactions, including immune checkpoint inhibitors, antiretroviral therapy, interferons, and tumor necrosis factor-α antagonists [27]. These associations may provide useful hints on the immunologic mechanisms guiding the initiation and persistence of granulomatous inflammation [27].

Risk factors based on epidemiological studies
Familial predisposition is one of the strongest risk factors for sarcoidosis [28,29]. A nationwide study in Sweden showed that individuals with a firstdegree relative with sarcoidosis have an almost fourfold increased risk of sarcoidosis [28]. Thirty nine percent of the risk in the Swedish population could be attributed to genetic factors shared among relatives, whereas the remaining risk could be explained by non-shared environmental factors [28]. Interactions between the two have been observed (e.g., with smoking [30] or insecticide exposure [31]).
The role of tobacco smoking in sarcoidosis remains controversial. Smoking is associated with a 35%-66% lower risk of sarcoidosis in most [20,[32][33][34], but not all, epidemiological studies [35]. This may be due to the downregulation of sarcoid inflammation seen in molecular studies [36]. In terms of disease progression, older data from Sweden suggest that smoking is not associated with better prognosis [37], whereas a recent pilot study indicated a favorable effect in forced vital capacity within 6 months of initiation of transdermal nicotine compared to placebo [38]. Occupational exposures to inorganic and organic agents, including silica, wood dust, pesticides, mold, and metals, have been implicated in the etiology of sarcoidosis [20,[39][40][41][42]. Most studies relied on recall of exposure status, potentially overestimating the associations.
Obesity generates a proinflammatory environment that is thought to trigger granulomatous inflammation [43]. Several epidemiological studies support this notion [33,[44][45][46]. Obesity at age 18 in participants of two US female cohorts was associated with a 40%-50% higher risk of sarcoidosis compared to normal body mass index [44,46]. Obese pregnant women were also more likely to develop sarcoidosis several years after delivery [45]. The risk of sarcoidosis associated with obesity in men has not been extensively studied, and neither has the role of physical activity or diet. Female hormones (e.g., estrogens) appear to delay the onset of sarcoidosis or even protect against sarcoid inflammation. Two studies reported a decreased risk associated with later age at menopause, indicative of prolonged exposure to endogenous estrogens [47,48].

Clinical presentation
Sarcoidosis manifestations can have an insidious as well as an acute onset. Of note, 10%-15% of patients are asymptomatic [49], and diagnosis is incidental, for example, following a chest radiograph. In symptomatic patients, signs of systemic inflammation may be accompanied by symptoms from the affected organ(s). Lung involvement and lymph node enlargement are seen in >90% of patients; however, virtually every organ can be involved. The eyes, skin, and liver are affected in 15%-30% of cases. Sarcoidosis of the nervous system and heart is less common (2%-10%) [1]. Sarcoidosis mimics several diseases, making the diagnosis challenging. Cancer and inflammatory and infectious diseases share similar signs or symptoms and should be considered during diagnosis.
The pathophysiological mechanisms of various clinical phenotypes and the reasons for the varying frequency of clinical features across populations are topics of intense research. LS is an example of a more homogenous phenotype associated with a specific genetic background (HLA-DRB1*03). HLA types are also associated with organ manifestations. For example, HLA-DRB1*04 is associated with ocular involvement and hypercalcemia in Swedish patients [50,51]. The HLA-DRB1 alleles are also linked to various HLA-DQB1 alleles [14]. The DRB1*0401 and DQB1*0301 haplotypes were associated with uveitis in the United Kingdom [52]. Like LS, HLA-DRB1*04-particularly the *0401 haplotype-is overrepresented in patients with manifestations of another, less common syndrome, called Heerfordt syndrome (uveitis with salivary gland involvement), as well as cranial nerve palsy [50].
The clinical importance of sarcoidosis manifestations varies. For instance, liver and spleen sarcoidosis is relatively common but seldom causes symptoms or functional impairment [53]. Sarcoidosis of the skin can range from barely visible to devastating lesions. Lupus pernio, an aggressive form of skin sarcoidosis, is more common in African Americans [54]. Enlargement of intrathoracic and peripheral lymph nodes seldom causes symptoms in sarcoidosis. However, expansive lymph node complexes compressing near structures should prompt exclusion of lymphoma, a common diagnostic challenge in this patient group.
Some sarcoidosis manifestations need special attention as they can be devastating or even lifethreatening, calling for urgent treatment initiation. Sarcoidosis of the eyes, often presented as uveitis, can lead to blindness if not treated [55]. Severe hypercalcemia presenting with polyuria and thirst, cognitive dysfunction, and muscle weakness should prompt urgent treatment. Cardiac sarcoidosis can cause life-threatening arrhythmias and cardiac failure but can also remain silent. In some patients, involvement of the heart may precede other manifestations, and thus, the diagnosis of sarcoidosis can be easily missed. In older autopsy studies, the highest frequency of cardiac sarcoidosis was found in Japanese patients [56]. Patients with severe cardiac symptoms seem to have fewer extra-cardiac sarcoidosis manifestations than patients with less severe disease, further complicating diagnosis [57,58]. Sarcoid inflammation in the central and peripheral nervous systems is rarer than heart involvement. Sarcoidosis of the central nervous system most commonly presents with cranial neuropathies but can cause a wide range of symptoms-for instance, seizures, hydrocephalus, and motor and sensory disturbances [59].

Diagnosis
The diagnosis of sarcoidosis can be set with some confidence when the clinical presentation is suggestive of granulomatous inflammation in two or more systems or organs, non-necrotizing granulomas are identified in a biopsied lesion, and other granulomatous diseases are eliminated from the differential diagnosis (Table 1) [60,61].
Computerized tomography (CT) is now the cornerstone of diagnosis and follow-up of pulmonary sarcoidosis. In many guidelines, high-resolution CT is recommended [53], but some usually prefer conventional CT with intravenous contrast to visualize enlarged lymph nodes (biopsy targets) and the pulmonary arteries (a marker of pulmonary hypertension). Common findings in sarcoidosis include bilateral and symmetrical lymphadenopathy localized in the hila and mediastinum, as well as micronodules (≤4 mm in diameter) with bilateral perilymphatic distribution in the upper and middle zones that correspond to active sites of granulomatous inflammation, ground-glass opacities, consolidations, and interlobular septal thickening [62,63]. Fibrotic progression is marked by loss of lung volume, honeycombing with traction bronchiectasis, bullae, and coarse septal bands [64]. Magnetic resonance imaging (MRI) and fluorine-18 fluorodeoxyglucose positron emission tomography ( 18 F-FDG-PET) play a greater role in the diagnosis and follow-up of cardiac sarcoidosis [65], and MRI in neurosarcoidosis [66]. In cases that are difficult to diagnose, 18 F-FDG-PET can provide useful information for identifying biopsy targets.
Bronchoscopy adds information about bronchial mucosal involvement that is not seen on a CT scan. Bronchoalveolar lavage fluid (BALF) is useful for differential diagnosis of other interstitial lung diseases and infections. In BALF, a ratio of CD4/CD8 T cells >3.5 is indicative of sarcoidosis, and T cells expressing the receptor Vα2.3 >10.5% strengthens the diagnosis [67]. Bronchoscopy with BALF, mucosal and/or transbronchial biopsies, and blind or ultrasound-guided endobronchial or esophageal fine-needle aspiration of lymph nodes are commonly used in the diagnostic process [1,68]. Transbronchial cryobiopsies may be favored in some patients with complex differential diagnostics as the method provides larger tissue for pathology examination, although the need for anesthesia and the higher risk of bleeding and pneumothorax should be considered.
There is unfortunately no diagnostic biomarker for sarcoidosis, although many have been tested and suggested [69]. An increased level of angiotensinconverting enzyme (ACE) can support the diagnosis but is not specific for sarcoidosis. Because diagnosis is based on the clinical presentation, granulomas on histology, and excluding other causes, it can be delayed for some patients for months or even years. For other patients, sarcoidosis can be diagnosed and then later discovered to be another disease (see Table 1).

Prognosis
The disease course in sarcoidosis varies significantly, from complete resolution to pulmonary fibrosis with respiratory failure. Predicting prognosis is a challenging task. Historically, chest radiographic stages proposed by Scadding in the 1960s were used to assess pulmonary involvement [70], but it has been suggested that they are poorly correlated with disease severity and pulmonary function and not useful for predicting sarcoidosis progression [71].
Several epidemiological factors appear to be associated with prognosis. Older age is an important predictor for disease progression and a prolonged disease course [71,72]. African Americans present more often with extrapulmonary manifestations and advanced radiographic stages [73], and they are more likely to have a worse clinical course [74]. Some studies suggest that there are differences between females and males in terms of organ involvement and ultimately disease prognosis [73][74][75], but more research is needed. Some molecular factors have been identified as associated with disease severity and prognosis, but few are used in clinical practice [69]. The most widely used is the serum level of ACE which reflects the burden of sarcoid granulomas [76]. There are several other serum biomarkers that correlate with pulmonary involvement, extrapulmonary involvement, therapeutic response, or disease activity such as lysozyme [77], serum amyloid A [78], KL-6 [77], chitotriosidase [79,80]), and soluble interleukin-2 receptor (sIL-2R) [81]. Some studies showed no correlation between these biomarkers and disease activity/progression, mak-ing it hard to conclude which would be most reliable and useful [69].
The HLA-DRB1 alleles can be useful markers of increased risk of extrapulmonary manifestations (e.g., the *04 allele) [82]. Extra-thoracic disease (excluding LS manifestations) usually associates with a prolonged disease course [51,83,84] and is more common in patients with non-LS sarcoidosis [82]. On the other hand, LS is associated with spontaneous resolution, with most DRB1*03 positive patients recovering within 2 years of diagnosis compared to only half of the HLA-DRB1*03 negative patients [85]. Table 2 displays selected HLA types and their associations with phenotypes and disease course. BALF cell counts may also correlate with disease prognosis. For example, expression of CD4+ Vα2.3+ T cells in BALF >10.5% was shown to be specific for the more benign LS phenotype [67]. Other BALF markers have been associated with more severe disease and/or chronic disease-for instance, higher levels of lymphocytes, neutrophils, eosinophils, and mast cells [86].

Mortality and comorbidities
Sarcoidosis is associated with an increased risk of premature death, especially in individuals with non-resolving disease. In several cohorts, the risk of death was twice as high as the general population [87][88][89]. In a large representative cohort in Sweden, the 5-year probability of death was 7% in patients who were treated with systemic corticosteroids around the time of diagnosis, whereas patients who did not require treatment had a 2% risk of death, similar to that of the age-and sexmatched general population group [87]. Advanced age at diagnosis, non-LS disease phenotype, the extent of lung fibrosis, cardiac and neurologic involvement, and concomitant pulmonary hypertension are all predictors of mortality in sarcoidosis [87,90,91]. The risk of death, however, is similar between men and women despite differences in the age of onset in some cohorts [87,90,91].
Respiratory failure owing to extensive pulmonary disease is a common cause of death in sarcoidosis and accounts for up to 60% of deaths [87,92,93]. Sarcoidosis-associated comorbidities and complications account for the excess risk of mortality, especially for patients with non-LS disease. Nonfatal cardiac sarcoidosis and ischemic heart disease lead to heart failure that accounts for about 20% of deaths in sarcoidosis [87]. The risk of heart failure
is more than twice as high in sarcoidosis than in the general population. Sarcoidosis-specific factors like arrhythmias explain a considerable proportion of the excess of heart failure cases in this group [94][95][96], highlighting the suboptimal identification and treatment of these complications.
Risk of infection is also high in sarcoidosis, owing to the immune disturbance associated with granulomatous inflammation and the immunosuppressant treatment in some patients [97,98]. Pneumonia and infections of the urinary tract and skinwhich are common in the general populationare observed in twofold higher rates in patients with sarcoidosis and are sometimes recurrent [97]. Opportunistic infections are rarely seen in these patients [97,99] but can cause serious complications. One should keep in mind that differential diagnosis of infection and new pulmonary sarcoid infiltrates is an extremely challenging task.
Sarcoidosis-associated pulmonary hypertensiondiagnosed in up to 3%-20% of patients with sarcoidosis-is associated with high morbidity and mortality [100][101][102]. Its pathophysiology is multifactorial, with granulomatous inflammation of the pulmonary vasculature, fibrosis of the parenchyma, left ventricular disease due to cardiac sarcoidosis or ischemia, and chronic thromboembolic disease all being potential contributing factors [102]. The 2021 WASOG consensus statement calls for screening with transthoracic echocardiography in patients with clinical, imaging, or electrocardiographic biomarkers raising suspicion for pulmonary hypertension, which should be followed by right heart catheterization in highly probable cases [103]. Identifying the underlying mechanisms of pulmonary hypertension in sarcoidosis has notable implications for the choice of treatment.
Increased occurrence of cancer has been described in sarcoidosis compared to the general population, particularly for hematologic and skin cancers [104][105][106][107]. A cancer diagnosis is more likely during the first couple of years after sarcoidosis diagnosis [105][106][107]. Higher incidence of hematologic and skin cancers throughout the disease course points toward a possible influence of granulomatous inflammation on cancer development. The increase in incidence of solid tumors closer to sarcoidosis diagnosis [107] suggests that the extensive diagnostic screening of patients likely also contributes to the co-occurrence of cancer and sarcoidosis. Furthermore, screening of cancer patients, malignancy in itself, as well as some cancer treatments can give rise to sarcoidosis/granulomatous reactions.
Severe fatigue, anxiety, and depression are commonly seen in patients with sarcoidosis [108,109]. Besides physical factors (e.g., decline in lung function), they are some of the most important contributing factors to patients' decline in quality of life after diagnosis [110]. The causes are multifactorial-with, for instance, inflammation, organ dysfunction, side effects from treatment, the burden of having a chronic disease, psychosocial factors, and coping strategies all likely implicated [111]. In patients with troublesome fatigue, pulmonary rehabilitation programs and neurostimulants may be useful [112], although further studies are needed to disentangle the role of these comorbidities in disease prognosis and identify potential mechanisms to target with new therapies.

Treatment
In sarcoidosis, initiation of pharmacologic treatment is based on symptom severity and whether organ function is affected [112]. Despite the heterogeneity of sarcoidosis, similar treatment regimens are used for most intra-or extra-thoracic manifestations. Research on manifestation-specific treatment is lacking, and a precision medicine approach to treatment is currently not possible. However, knowledge is constantly evolving, and the hope is to be able to predict who will respond to which treatment and who will experience adverse effects due to treatment.

Current recommendations
For LS, nonsteroidal anti-inflammatory treatment is usually enough to reduce inflammation and its associated symptoms. Systemic corticosteroids remain the mainstay of non-LS sarcoidosis treatment. This treatment can relieve symptoms but can be accompanied by side effects and therefore mostly reserved for patients with symptoms and/or threatened organ function [112]. Treatment is typically initiated with prednisolone 30-40 mg/day and progressively tapered to reach a maintenance dose of <5-10 mg/day in most patients within the first year. Disease flares during dose reduction are common, and clinicians should carefully titrate corticosteroid dose to the lowest needed for symptom control.
In patients with treatment failure, the inability to adequately taper corticosteroid dose and the pres-ence of severe adverse effects should prompt initiation of a second-line steroid-sparing treatment, preferably methotrexate [112,113]. If contraindications exist (e.g., pregnancy), azathioprine is an alternative treatment, although it has been linked to a higher risk of side effects such as infections [114,115]. Empirical data suggest that leflunomide and mycophenolate mofetil may be efficacious in sarcoidosis and are therefore used in some countries [116,117]. In the absence of evidence, some experts debate when to introduce steroidsparing treatments [118]. The results of an ongoing randomized-controlled trial (PREDMETH [119]) aiming to examine the efficacy of methotrexate as first-line treatment compared to prednisolone are therefore much anticipated.
In steroid-and methotrexate-resistant disease, treatment strategy includes administration of a biologic-typically infliximab, a tumor necrosis factor α antibody [112]. Dosing schemes may vary among centers, and antibody-mediated resistance should be considered in patients with declining response to infliximab. Acthar Gel (swine adrenocorticotropic hormone) is licensed in the United States for use in sarcoidosis but is not available in the European market [112]. Other therapies for sarcoidosis (efzofitimod, anakinra, tocilizumab, tofacitinib, rituximab, sirolimus, or CLEAR [a combination of antibiotics]) have been or are being tested, but data are currently limited to support widespread use [120]. A major opportunity for future research is to develop new treatments and optimize current treatments for sarcoidosis.
Whether antifibrotic treatment delays progression and reduces the mortality associated with endstage disease remains unclear [64,121,122].
Scarce data indicate that patients with fibrosis may benefit from antifibrotic treatment [123,124]. Lung transplantation remains the last treatment option in eligible patients with progressive disease [125,126].

Future perspective: biomarkers and personalized medicine
New treatments are needed in sarcoidosis, and it is critical that they are targeted toward the upstream causes of granuloma formation and persistence. The identification, validation, and testing of specific markers driving sarcoidosis developmentwhether they be antigens, inflammatory cytokines, genetic variants, or a combination of thesewill lead to therapeutic targets. There is a great interest in discovering novel molecular markers of the underlying biological disease process in sarcoidosis. Biomarkers could be used to clarify many of the questions remaining on the topics discussed in this review, including etiology, phenotype identification, diagnosis, monitoring disease activity, predicting prognosis, and treatment response. The search for one or a combination of biomarkers obtained via various methods (e.g., serum, BALF, and imaging modalities) continues [69,127,128]. Currently, the validity and clinical utility of most potential biomarkers that have been reported in the literature have not been rigorously studied.
One of the major challenges in treating sarcoidosis is that we do not know which patients are at risk for sarcoidosis-related comorbidities and mortality. Thus, the decision on treatment is based only on whether the patient has symptoms, or whether an organ is currently threatened. Treatment decisions using information on predicted risks, as well as a patient's biology and preferences, will improve patient outcomes.
We do not only need more data but better data and better methods. Large studies linking highdimensional individual data (environmental, occupational, societal, genetic, transcriptomic, and proteomic) to detailed clinical information with long follow-up are critical for moving toward personalized care. To capture the variability between individuals, an omics approach (e.g., genomics, epigenomics, transcriptomics, proteomics, and metabolomics) can be used to form more homogenous and biologically based groups. Omics methods are not only more costly to perform but also more demanding in terms of data processing power and biostatistical expertise to deal with high-dimensional and highly correlated data. However, costs have decreased for many omics technologies, and experience and expertise have advanced. Several recent studies demonstrate the use of whole genome sequencing [129], differential gene expression [130], and RNA sequencing [131]. There remains the hope that the triggering antigen will be identified, and a more personalized medicine approach can be used, such as a sarcoidosis-omics assay to aid the diagnosis and choice of treatment strategy for each patient.

Acknowledgments
Thank you to our colleagues and patient partner who gave feedback on a draft of this manuscript.

Conflict of interest statement
Marios Rossides reports non-promotional speaker fees from Teva Pharmaceutical Industries, outside this work. Elizabeth Arkema received honoraria from the Milken Institute, outside this work. Pernilla Darlington and Susanna Kullberg disclose no conflict of interests.