Abstract
Rationale
Historically, sarcoidosis was described as a restrictive lung disease, but several alternative phenotypes of pulmonary function have been observed. Pulmonary function phenotypes in sarcoidosis may represent different clinical and/or molecular phenotypes.
Objectives
To characterize the prevalence of different pulmonary function phenotypes in a large and diverse sarcoidosis cohort from a tertiary care referral center.
Methods
We identified individuals seen between 2005–2015 with a confirmed diagnosis of sarcoidosis. Data were collected from the first pulmonary function test (PFT) performed at our institution which included spirometry and diffusing capacity of the lung for carbon monoxide (DlCO). Demographics and clinical data were collected. Chi-squared analyses and multiple linear regressions were done to assess statistical differences and associations. Global Lung Function Initiative equations were used to calculate percent predicted measurements for spirometry and DlCO.
Results
Of 602 individuals with sarcoidosis, 93% (562) had pulmonary involvement, 64% (385) were female, and 57% (341) were Black. Of those with pulmonary involvement, 56% had abnormal pulmonary function. Lung function impairment phenotypes included: 47% restriction, 22% obstruction, 15% isolated reduction in DlCO, and 16% combined obstructive restrictive phenotype. Restriction was the most common PFT phenotype among Black individuals (41%), while no lung impairment was most common among White individuals (66%) (P < 0.001). Males more frequently had obstruction (19%) compared with females (9%) P = 0.001, and females had more restriction (30%) compared with males (21%) P = 0.031.
Conclusions
Among individuals with sarcoidosis and pulmonary function impairment, less than half demonstrated a restrictive phenotype. There were significant differences in pulmonary function phenotypes by race and sex.
Keywords: sarcoidosis, respiratory function tests, pulmonary disease
Sarcoidosis is a heterogeneous disease which affects the lungs in approximately 90% of individuals (1). Pulmonary function impairment has been described as occurring in up to 55% of individuals with sarcoidosis (2–4). Sarcoidosis is frequently described as a predominantly restrictive lung disease (2) resulting in the majority of clinical trials in sarcoidosis using forced vital capacity (FVC) % predicted as the primary pulmonary function outcome measure (5–7). However, several additional phenotypes of pulmonary impairment have been described in sarcoidosis including an obstructive phenotype (2, 4, 8), a combination of obstructive and restrictive phenotypes (9), and an isolated reduced diffusing capacity of the lungs for carbon monoxide (DlCO) phenotypes (10). In fact, an international cohort of predominately White individuals of European ancestry with sarcoidosis reported only 7% of 830 cases having restriction as the pulmonary function impairment (11). Pulmonary function phenotypes in sarcoidosis may represent different clinical and/or molecular phenotypes of this heterogenous disease and warrant further exploration.
There has been heterogeneity in clinical characteristics of sarcoidosis and outcomes across geographic regions and countries (12–14). An international cohort of 164 individuals with pulmonary sarcoidosis found more than 20% of individuals had a phenotype of obstruction and noted racial differences with White and Asian individuals being more likely to have an obstructive phenotype as compared with Black individuals (15). Pulmonary function among individuals with sarcoidosis has been described in the United States with racial differences in the degree of impairment noted (8, 16, 17), but pulmonary function phenotypic patterns have not been similarly explored. Given the heterogeneity of sarcoidosis on both a molecular and epidemiologic level, there is a need for further studies that comprehensively characterize pulmonary function phenotypes among diverse populations in the United States. The aims of this study were to characterize the prevalence of different pulmonary function phenotypes in a diverse sarcoidosis cohort from a tertiary care referral center.
Methods
Procedures
The study was reviewed and approved by the Johns Hopkins University Institutional Review Board (IRB00195158). This study includes the establishment of a retrospective cohort to include patients seen within the Johns Hopkins Sarcoidosis Clinic between 2005 and 2015. The inclusion criteria were having a pulmonary function test (PFT) performed between 2005 and 2015 at Johns Hopkins institutions, being seen as an outpatient by a physician that routinely staffed the Johns Hopkins Sarcoidosis Clinic (sarcoidosis physician), and having the diagnosis of sarcoidosis. Our pulmonary function lab follows the American Thoracic Society guidelines for reproducibility and reliability of pulmonary function tests (18). Pulmonary function data including spirometry and DlCO were obtained from the Johns Hopkins legacy database for individuals with tests including with an ICD-10 (International Classification of Diseases, Tenth Revision) code of D86.0 or an ICD-9 code of 135 (sarcoidosis) in the diagnosis. The diagnosis of sarcoidosis was made using criteria from the American Thoracic Society Statement in 2020 (19). Individuals were excluded if they did not have a PFT measured within a year of a clinic visit with the sarcoidosis physician. Once individuals were identified from the pulmonary function database, chart abstraction was used to assess sarcoidosis organ involvement. The organ assessment tool (OAT) utilized in the Genomic Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS) (20), which was adapted from prior tools (2, 21) was used to assess the probability of sarcoidosis involvement across 16 organ systems. The tool utilizes radiologic findings, bronchoscopy findings including alveolar lavage, PFT findings, and histological findings to determine probability of pulmonary involvement (2, 20, 21). Organ involvement among the 16 organs was included as present if noted to be highly probable or probable by the OAT. For analyses organ involvement was grouped into one organ, two to four organs, and five or more organs.
We defined four pulmonary function impairment phenotypes: 1) restriction as forced vital capacity (FVC) < the lower limit of normal (LLN) and forced expiratory volume in one second (FEV1)/FVC ⩾ LLN; 2) obstruction as FEV1/FVC < LLN; 3) combined restrictive and obstructive (combined) as FVC < LLN and FEV1/FVC < LLN; and 4) isolated gas transfer defect as DlCO < LLN with no restrictive or obstructive defect. A normal phenotype was defined as not having any impairment. The percentage predicted measurements for FVC, FEV1, and DlCO were calculated using Global Lung Function Initiative reference equations calculated by assigning the race to a race-composite set of equations with the designation “other” (22, 23). In a sensitivity analysis, predicted values were also calculated using reference equations specifically derived for White or Black populations.
Chart review was performed abstracting data from the clinic visit closest to the time of first PFT (baseline). Demographic information including race were obtained from medical chart review. Ethnicity was not consistently documented in the chart and therefore is not described. The first clinic visit note was reviewed and the following was abstracted: 1) tobacco use: past, current, or no use; 2) year of diagnosis; and 3) comorbidities using Charlson Comorbidity Index. Disease duration was defined as the time between the year of diagnosis and the baseline PFT. Laboratory data including hemoglobin was included if performed within 1 year of the first PFT. Finally, chest imaging was abstracted by the lead author (M.S.) with a preference for computed tomography (CT) of the chest over chest radiographs. The chest imaging was described as hilar or mediastinal lymph node involvement (yes/no), parenchymal involvement (yes/no), and fibrocystic disease (yes/no). Imaging information was included if done within 1 year of the date of first PFT.
Statistical Analysis
Descriptive statistics were used to characterize the cohort and were compared between White and Black participants using Mann Whitney U and Chi square and Fisher exact tests, for continuous and categorical variables, respectively. Similarly, distributions of pulmonary function phenotypes were compared across race, sex, tobacco use, age, disease duration, and organ involvement using Kruskal-Wallis and Chi square or Fisher exact tests. After grouping the pulmonary function phenotypes as noted above, we performed chi-squared analyses to assess differences between pulmonary function phenotypes by race, sex, tobacco use, age, disease duration, and organ involvement. We conducted multiple linear regression models with each pulmonary function parameters at baseline (FVC% predicted, FEV1% predicted, and DlCO % predicted). All regression models included these a priori identified covariates: sex, tobacco use, and organ involvement. Results of models are presented as the difference in parameter estimates with corresponding 95% confidence interval (CI). We performed an analysis using DlCO corrected for hemoglobin in the subset of individuals who had hemoglobin values. A P value < 0.05 was considered statistically significant. All analyses were performed using Stata Version 15 (STATA Corp).
Results
Participants
Six hundred and seventy-eight individuals were eligible after reviewing the Johns Hopkins Pulmonary Function database. Of the 678 individuals eligible, 24 were excluded for either not having a diagnosis of sarcoidosis or not being seen by a sarcoidosis clinic physician at the time of the initial PFT. We abstracted the medical charts of 654 individuals, but excluded 22 from the analysis due to having a clinic visit more than 1 year from the time of the initial PFT. 30 individuals who had race documented as “other” in the medical chart were excluded since we were unable to adequately categorize into an appropriate racial category. See Figure 1 for study population diagram.
Figure 1.
Inclusion of participants for analysis.
Clinical Characteristics
Of the 602 individuals in the cohort, 93% (562) had pulmonary involvement defined by the GRADS OAT. Individuals were predominantly female 64% (385). The median age was 51 years and median disease duration from time of diagnosis was 4 years. Most individuals within the cohort (64%) had more than one organ involved (Table 1). There were significant racial differences in baseline clinical characteristics, with Black individuals more frequently having greater than one organ involved compared with White individuals (70% versus 56%) (P < 0.001). Black individuals also had a longer disease duration from initial diagnosis compared with White individuals, Table 1. There were no racial differences seen in age or comorbidities (Table 1). Females had a longer disease duration from initial diagnosis compared with males (median of 5 years versus 3 years) (P = 0.030). There were no sex differences seen in age, comorbidities, tobacco use, or pulmonary involvement. A total of 77% (461) of the cohort had chest imaging done within a year of pulmonary function and 40% (184) of the images were chest radiographs. There were racial differences in chest imaging (P = 0.006). Black individuals were more likely to have fibrocystic disease compared with White individuals (P = 0.002).
Table 1.
Participant characteristics
| All N = 602 |
Black N = 341 |
White N = 261 |
|
|---|---|---|---|
| Age in years, median (IQR) | 51 (44–58) | 49 (44–57) | 53 (44–60) |
| Sex, Female n (%) | 385 (64) | 253 (74) | 132 (51) |
| Pulmonary Involvement, yes n (%) | 562 (93) | 312 (91) | 250 (95) |
| Organ Involvement n (%) | |||
| 1 organ | 215 (36) | 102 (30) | 117 (44) |
| 2–4 organs | 364 (60) | 226 (66) | 138 (53) |
| >5 organs | 23 (4) | 16 (4) | 7 (3) |
| Disease duration in years, median (IQR) | 4 (1–12) | 6 (1–17) | 2 (1–8) |
| Charlson Comorbidity Index, median (IQR) | 2 (1–4) | 2 (1–3) | 2 (1–4) |
| Tobacco use, n (%) | |||
| No use | 346 (59) | 181 (54) | 165 (64) |
| Past use | 196 (33) | 115 (35) | 81 (32) |
| Current use | 47 (8) | 37 (11) | 10 (4) |
| Chest imaging, n (%) | |||
| Normal | 96 (21) | 59 (24) | 37 (18) |
| Lymph node involvement only | 62 (13) | 33 (13) | 29 (14) |
| Lymph node and parenchymal involvement | 154 (33) | 81 (32) | 73 (35) |
| Parenchymal involvement only | 127 (28) | 59 (24) | 68 (32) |
| Fibrocystic disease | 22 (5) | 19 (7) | 3 (1) |
Definition of abbreviation: IQR = interquartile range.
Pulmonary Function Phenotype
Of those with pulmonary involvement defined by the OAT, 56% had abnormal pulmonary function. Restriction was the most common impairment with 47% of individuals with abnormal lung function (27% of cohort with pulmonary involvement defined by the GRADS OAT). Of the remaining 53% with phenotypes other than restriction, obstruction was the second most common phenotype with 22% of individuals with abnormal lung function having obstruction (13% of cohort with pulmonary involvement). Isolated reduction in DlCO was seen in 15% of individuals with abnormal lung function (8% of the cohort with pulmonary involvement) and combined restriction/obstruction were also identified in 16% of individuals with abnormal lung function (9% of the cohort with pulmonary involvement).
Of the cohort, 316 had a hemoglobin value within 1 year of the PFTs. The mean for the corrected DlCO % predicted was higher compared with using the uncorrected DlCO % predicted (85.90 versus 83.85, P < 0.01). However, using DlCO corrected did not significantly change our findings among pulmonary function phenotypes. Therefore, we used DlCO uncorrected to include the 562 individuals.
Differences in Pulmonary Function Phenotypes
There were racial differences seen among the phenotypes. Normal pulmonary function tests were observed in 66% of White individuals compared with only 26% of Black individuals. Restriction was the most common pulmonary function phenotype identified among Black individuals (41%). The most common abnormal pulmonary function phenotype among White individuals was obstruction (17%), whereas only 9% of Black individuals were found to have obstruction (Figure 2, Table 2). There were sex differences seen among the pulmonary function phenotypes with males having more obstruction (19%) compared with females (9%) P = 0.008 and females having more restriction (30%) compared with males (21%) P = 0.031. There were differences in chest imaging characteristics among the phenotypes with fibrocystic disease more commonly being seen among those with the restriction or combined phenotypes (P < 0.001). Additionally, there were differences in disease duration with the longest median duration (10 years) among those with the combined phenotype (P = 0.037). There were also differences by tobacco use (P = 0.001) with a higher number of current smokers among individuals with an isolated reduction in DlCO phenotype and combined phenotype compared with the other phenotypes; however, less than 20% of individuals in each phenotype were current smokers (Table 2). Of note, among those with the isolated reduction in DlCO, there were no racial differences seen among tobacco use (P = 0.067).
Figure 2.
Prevalence of pulmonary function phenotypes by A) Race and B) Sex. Restriction was defined as FVC < the lower limit of normal (LLN) with FEV1/FVC ⩾ (LLN), obstruction as FEV1/FVC ⩽ LLN with FVC ⩾ LLN, combined (obstruction/restriction) as FVC < LLN and FEV1/FVC < LLN, and isolated gas transfer defect as DlCO < LLN with no restrictive or obstructive defect. Normal was defined as not having any impairment. DlCO = diffusing capacity of the lung for carbon monoxide; FEV1 = forced expiratory volume in one second; FVC = forced vital capacity; LLN = lower limit of normal.
Table 2.
Pulmonary function phenotypes among individuals with pulmonary involvement
| Restriction (n = 149) | Obstruction (n = 71) |
Combined (n = 50) |
Isolated DlCO (n = 46) |
Normal (n = 246) |
P Value | |
|---|---|---|---|---|---|---|
| Race n (%) | ||||||
| Black (n = 312) | 127 (41) | 28 (9) | 38 (12) | 37 (12) | 82 (26) | <0.001 |
| White (n = 250) | 22 (9) | 43 (17) | 12 (5) | 9 (4) | 164 (66) | |
| Sex n (%) | ||||||
| Female (n = 355) | 105 (30) | 32 (9) | 30 (8) | 30 (8) | 158 (45) | 0.008 |
| Male (n = 207) | 44 (21) | 39 (19) | 20 (10) | 16 (8) | 88 (43) | |
| Tobacco use n (%) | ||||||
| No use (n = 320) | 94 (29) | 34 (11) | 22 (7) | 19 (6) | 151 (47) | 0.001 |
| Past use (n = 184) | 39 (21) | 29 (16) | 18 (10) | 18 (10) | 80 (44) | |
| Current use (n = 45) | 10 (22) | 8 (18) | 9 (20) | 8 (18) | 10 (22) | |
| Organ Involvement n (%) | ||||||
| 1 organ (n = 197) | 52 (26) | 25 (13) | 19 (10) | 9 (5) | 92 (47) | 0.077 |
| 2–4 organs (n = 342) | 88 (26) | 44 (13) | 31 (9) | 31 (9) | 147 (43) | |
| >4 organs (n = 23) | 9 (39) | 2 (9) | 0 (0) | 5 (22) | 7 (30) | |
| Chest imaging n (%) | ||||||
| Normal (n = 76) | 12 (16) | 5 (7) | 6 (8) | 11 (15) | 42 (55) | <0.001 |
| Lymph node involvement only (n = 59) | 11 (19) | 4 (7) | 1 (2) | 5 (8) | 38 (64) | |
| Lymph node and parenchymal involvement (n = 153) | 46 (30) | 22 (14) | 7 (5) | 12 (8) | 66 (43) | |
| Parenchymal involvement only (n = 121) | 30 (25) | 21 (17) | 12 (10) | 8 (7) | 50 (41) | |
| Fibrocystic disease (n = 22) | 11 (50) | 3 (14) | 5 (23) | 0 (0) | 3 (14) | |
| Age at testing (yr), median (IQR) | 50 (44–57) | 49 (42–57) | 48 (43–55) | 52 (44–61) | 52 (46–61) | 0.596 |
| Disease duration (yr from diagnosis), median (IQR) | 6 (2–15) | 4 (1–12) | 10 (2–20) | 6 (1–18) | 3 (1–9) | 0.037 |
Definition of abbreviations: DlCO = diffusing capacity of the lung for carbon monoxide; IQR = interquartile. range.
Pulmonary Function Severity
Overall Black individuals had significantly worse pulmonary function compared with White individuals including FVC% predicted, FEV1% predicted and DlCO % predicted indicating worse severity, Table 3. Analysis of parameters stratified by phenotype showed that this difference remained in all except the combined restriction/obstruction phenotype, Table 3. Females overall had slightly worse pulmonary function compared with males across all the lung function parameters, Table 4. However, the analysis stratified by phenotype showed that this difference was more prominent among the restriction and obstruction phenotypes. Males had worse pulmonary function in the combined restriction/obstruction phenotype, Table 4.
Table 3.
Lung function parameters by across lung function phenotypes by race
| Overall N = 562 |
Restriction N = 149 |
Obstruction N = 71 |
Combined N = 50 |
Isolated DlCO N = 46 |
Normal N = 246 |
|
|---|---|---|---|---|---|---|
| FVC % predicted | 84.72 (19.23) | 65.42 (10.90) | 95.45 (13.18) | 62.05 (12.30) | 90.24 (7.95) | 96.89 (12.82) |
| Black | 77.06 (16.51) | 65.30 (11.28) | 91.59 (9.45) | 63.00 (11.24) | 89.90 (7.39) | 91.04 (9.01) |
| White | 94.27 (18.10) | 66.12 (8.61) | 97.96 (14.69) | 59.05 (15.36) | 91.63 (10.35) | 99.81 (13.46) |
| FEV1 % predicted | 79.35 (20.15) | 64.41 (11.16) | 75.65 (13.18) | 47.19(11.73) | 89.25 (9.21) | 94.16 (13.13) |
| Black | 72.79 (17.78) | 64.67 (11.66) | 72.94 (9.60) | 47.80 (10.80) | 89.08 (9.30) | 89.54 (8.85) |
| White | 87.55 (19.96) | 62.93 (7.72) | 77.41 (14.90) | 45.27 (14.67) | 89.96 (9.36) | 96.47 (14.28) |
| DlCO % predicted | 84.92 (23.11) | 71.27 (22.16) | 89.92 (18.98) | 64.19 (22.86) | 66.26 (10.16) | 99.44 (14.66) |
| Black | 76.40 (22.47) | 69.80 (21.89) | 83.60 (23.07) | 63.73 (22.72) | 64.66 (10.66) | 95.32 (11.97) |
| White | 95.55 (21.36) | 79.74 (22.34) | 94.04 (14.64) | 65.65 (24.27) | 72.84 (3.00) | 101.50 (15.46) |
| FEV1/FVC | 0.75 (0.09) | 0.80 (0.06) | 0.64 (0.07) | 0.61 (0.09) | 0.79 (0.05) | 0.78 (0.05) |
| Black | 0.76 (0.09) | 0.80 (0.06) | 0.64 (0.06) | 0.62 (0.09) | 0.79 (0.05) | 0.80 (0.05) |
| White | 0.74 (0.08) | 0.76 (0.05) | 0.64 (0.07) | 0.61 (0.09) | 0.80 (0.05) | 0.78 (0.05) |
Definition of abbreviations: DlCO = diffusing capacity of the lung for carbon monoxide; FEV1 = forced expiratory volume in one second; FVC = forced vital capacity.
Table 4.
Lung function parameters by across lung function phenotypes by sex
| Overall N = 562 |
Restriction N = 149 |
Obstruction N = 71 |
Combined N = 50 |
Isolated DlCO N = 46 |
Normal N = 246 |
|
|---|---|---|---|---|---|---|
| FVC % predicted | 84.72 (19.23) | 65.42 (10.90) | 95.45 (13.18) | 62.05 (12.30) | 90.24 (7.95) | 96.89 (12.82) |
| Female | 83.42 (19.02) | 64.25 (10.44) | 92.19 (10.68) | 63.79 (12.06) | 90.50 (8.24) | 96.77 (12.54) |
| Male | 86.94 (19.43) | 68.21 (11.58) | 98.12 (14.51) | 59.43 (12.49) | 89.75 (7.61) | 97.09 (13.38) |
| FEV1 % predicted | 79.35 (20.15) | 64.41 (11.16) | 75.65 (13.18) | 47.19 (11.73) | 89.25 (9.21) | 94.16 (13.13) |
| Female | 78.97 (19.62) | 63.93 (10.81) | 72.47 (11.72) | 49.61 (10.83) | 89.45 (8.17) | 93.86 (12.67) |
| Male | 80.02 (21.06) | 65.58 (12.02) | 78.25 (13.87) | 43.57 (12.35) | 88.87 (11.19) | 94.70 (13.97) |
| DlCO % predicted | 84.92 (23.11) | 71.27 (22.16) | 89.92 (18.98) | 64.19 (22.86) | 66.26 (10.16) | 99.44 (14.66) |
| Female | 84.72 (22.23) | 72.04 (22.17) | 86.90 (20.76) | 67.95 (24.92) | 65.90 (10.25) | 100.17 (15.15) |
| Male | 85.03 (23.64) | 69.43 (20.94) | 92.40 (17.27) | 58.55 (18.56) | 66.95 (10.29) | 98.13 (13.71) |
| FEV1/FVC | 0.75 (0.09) | 0.80 (0.06) | 0.64 (0.07) | 0.61 (0.09) | 0.79 (0.05) | 0.78 (0.05) |
| Female | 0.73 (0.10) | 0.81 (0.06) | 0.64 (0.08) | 0.63 (0.06) | 0.80 (0.05) | 0.79 (0.05) |
| Male | 0.77 (0.08) | 0.77 (0.06) | 0.64 (0.06) | 0.59 (0.12) | 0.79 (0.05) | 0.78 (0.05) |
Definition of abbreviations: DlCO = diffusing capacity of the lung for carbon monoxide; FEV1 = forced expiratory volume in one second; FVC = forced vital capacity.
Values shown are mean (standard deviation).
In our regression models, race was significantly associated with all three pulmonary function parameters with Black individuals having significantly worse pulmonary function compared with White individuals after controlling for sex, tobacco use, disease duration, and organ involvement (FVC% predicted: 16.39, FEV1% predicted: 13.41, DlCO % predicted: 16.20) (Table 5). DlCO differed by sex with females having significantly higher DlCO % predicted (4.45) compared with males after including other covariates. Disease duration was significantly associated with worse pulmonary function in all three models. In models that included FVC% predicted and FEV1% predicted calculated using race-specific reference equations, the association between race and FVC % was attenuated with FVC % predicted 3.7% lower among Black individuals compared with White and the difference in FEV1% predicted was 2.55% lower and no longer statistically significant (Table E1 in the online supplement).
Table 5.
Patient characteristics and the association with lung function
| β | 95% CI | P Value | |
|---|---|---|---|
| FVC % predicted | |||
| Black | −16.39 | (−19.49 to −13.29) | <0.001 |
| Female | 0.78 | (−2.28 to 3.84) | 0.617 |
| Disease duration | −0.21 | (−0.36 to −0.65) | 0.005 |
| Tobacco use | |||
| Former | 3.44 | (0.34 to 6.54) | 0.030 |
| Current | 1.58 | (−3.84 to 7.00) | 0.567 |
| FEV1% predicted | |||
| Black | −13.41 | (−16.77 to −10.04) | <0.001 |
| Female | 2.70 | (−0.62 to 6.02) | 0.111 |
| Disease duration | −0.28 | (−0.44 to −0.12) | 0.001 |
| Tobacco use | |||
| Former | 1.71 | (−1.81 to 4.93) | 0.364 |
| Current | −2.21 | (−8.10 to 3.68) | 0.461 |
| DlCO % predicted | |||
| Black | −16.20 | (−19.80 to −12.60) | <0.001 |
| Female | 4.45 | (0.89 to 8.00) | 0.014 |
| Disease duration | −0.42 | (−0.59 to −0.24) | <0.001 |
| Tobacco use | |||
| Former | −2.56 | (−6.16 to 1.05) | 0.164 |
| Current | −7.51 | (−13.81 to −1.21) | 0.019 |
Definition of abbreviations: CI = confidence interval; DlCO = diffusing capacity of the lung for carbon monoxide; FEV1 = forced expiratory volume in one second; FVC = forced vital capacity.
Table displays three distinct models describing the association between patient characteristics and lung function. All regression models include organ involvement.
Discussion
In a large and diverse cohort of individuals with sarcoidosis from an urban tertiary referral center, restriction represented less than half of the pulmonary function impairment phenotype. We observed different pulmonary function phenotypes by race, sex, disease duration, and tobacco use. We also found that Black individuals had worse lung function compared with White individuals, and these differences were seen in all pulmonary function phenotypes except the combined phenotype. In our regression analyses, race was significantly associated with all three pulmonary function parameters (FVC, FEV1, and DlCO) with Black individuals having worse pulmonary function when controlling for organ involvement, sex, and tobacco use.
To our knowledge, our cohort is the first large diverse U.S. cohort to describe the prevalence of pulmonary function impairment phenotypes. Our findings have implications in both clinical care and research in sarcoidosis. The etiology of sarcoidosis remains elusive including what genetic or environmental factors trigger the disease and drive the development of unique phenotypes. The clinical heterogeneity of sarcoidosis contributes to the challenges in advancing disease knowledge. Identification of clinical pulmonary function phenotypes have the potential to improve understanding in sarcoidosis and allow for further scientific understanding of the disease. Future studies evaluating therapeutic outcomes linked to pulmonary function phenotypes can lead to the identification high-risk endotypes and facilitate personalized approaches to clinical care.
The historical acceptance of sarcoidosis as a restrictive lung disease has led clinical practice and research to focus on spirometry alone to assess pulmonary sarcoidosis disease course (2, 5, 7). We have demonstrated, however, that such a strategy could miss clinical progression in a substantial proportion of individuals whose pulmonary impairment was isolated reduction in DlCO; we noted this phenotype in 15% of our cohort with pulmonary impairment representing a group of individuals who may have pulmonary vascular disease necessitating further evaluation. Additionally, we found that those individuals with a mixed phenotype (both obstructive and restrictive phenotypes) had a longer disease duration and worse lung function in all parameters compared with the individuals with other phenotypes, suggesting higher disease severity among this group, warranting further investigation. The dominant pulmonary impairment phenotype is likely due to the distribution of granulomatous inflammation with obstruction often being the result of airways granulomas, restriction resulting from interstitial granulomas, and isolated reduction of diffusion capacity from pulmonary vascular involvement. Further clinical phenotyping of patients has the potential to assist in disease understanding, clinical prognostication, and to uncover potential therapeutic targets.
Our cohort is more diverse than many of the sarcoidosis cohorts published in the literature with Black individuals comprising 57% of our cohort, enabling a robust comparison of racial differences. We found that Black individuals were significantly more likely to have abnormal pulmonary function compared with White individuals. There are several potential explanations for our observed racial differences. First, the racial differences may be explained by differences in environmental exposures. There is evidence that sarcoidosis is associated with environmental occupational exposures (24) and in the literature, there are significant differences in environmental exposures based on race (25). Second, the differences may be caused by genetic and environmental exposure interactions. There have been studies suggesting a gene-environment interaction with specific gene-exposures varying by race and being associated with a different sarcoidosis phenotype (26, 27). However, race is an imperfect proxy for genetics (28). The racial differences we observed may be explained by health disparities in medicine, including barriers to care and possible delays in care. How algorithms account for race in medicine are an increasingly recognized potential source of bias (23). Our findings of worse lung function among Black patients (defined using universal reference equations that do not account for race) would be less apparent if race-specific approaches are applied to interpret lung function, an approach that could result in delays in diagnosis and treatment in clinical practice. This could potentially explain previous studies which have demonstrated that Black individuals are found to have more advanced disease at diagnosis and are less likely to have clinical improvement compared with non-Hispanic Whites (29, 30). Additionally, it has been shown that Black individuals reported lower medication adherence compared with non-Hispanic White individuals (31). The reasons for the differences in medication adherence in sarcoidosis are unknown, but may contribute to the differences in lung function observed in our study. Future research is needed to understand the contribution of societal, environmental, and possible genetic factors to the racial differences that exist in sarcoidosis to inform development of potential solutions to reduce these differences.
In addition to racial differences found in our study, we also found differences in pulmonary function phenotype by sex. We found that males were more likely to have an obstructive phenotype, while females were more likely to have a restrictive phenotype. Although the cause of this difference is unknown, it is plausible that environmental exposures may contribute to a different distribution of granulomatous inflammation in sarcoidosis. Previous literature has found that females were more likely to have reduced lung compliance compared with males, but there was no evaluation of pulmonary function phenotype (11). Of note, previous research has found that females with sarcoidosis have worse lung function compared with males with sarcoidosis (17); however, our study found no difference in FVC% predicted or FEV1% predicted and higher DlCO % predicted among females, even when including tobacco use in the model. Further research is needed to understand the sex differences seen in sarcoidosis.
Our cohort includes individuals at different stages of diagnosis and treatment. Our findings were that 56% of individuals in our cohort had pulmonary function impairment, which is higher than some of the larger incident cohorts in sarcoidosis (2). Our findings are consistent with other prevalence cohorts in the literature, with a cohort from Israel recently describing an impairment rate of almost 56% (3). Our finding of multiple pulmonary function phenotypes is consistent with international cohorts of individuals with sarcoidosis. However, the international cohorts have reported even lower rates of restriction, although are often less diverse with predominantly White individuals (11, 15). These findings however raise the question of whether using FVC% predicted as the sole primary outcome measure of pulmonary function in clinical studies assessing treatment response in sarcoidosis is appropriate. Given the known heterogeneity of sarcoidosis (32), it may be helpful to identify the pulmonary function phenotype of an individual within a clinical trial context to ensure that the appropriate pulmonary function outcome is measured to ensure an appropriate interpretation of results.
There is significant clinical heterogeneity in sarcoidosis, and differences in organ involvement are well described in the sarcoidosis literature (2, 33). Pulmonary function phenotypes may add to the further understanding of the clinical heterogeneity in sarcoidosis by representing different specific phenotypes of one organ system. More recent studies have worked to identify different molecular and immunologic phenotypes and endotypes of disease to inform not only the potential pathobiological underpinnings of this clinical heterogeneity, but to explore novel molecular signatures that may assist in clinical management and treatments (33, 34). In this context, further investigation is needed to understand if pulmonary function phenotype and outcomes are associated with different molecular and immunologic profiles that can be used to guide clinical management in sarcoidosis.
Limitations
Our study has several limitations given the retrospective nature of our cohort. We were not able to assess other patient-level measures such as socioeconomic status, education, or exposures, given the retrospective chart review. Future studies should include these important measures to further understand potential race differences in lung function severity and if there is an intersection with socioeconomic status. We chose to exclude individuals with race identified as other by the medical chart given, which was a small percentage. Additionally, we used DlCO uncorrected for hemoglobin given that we only had hemoglobin values on 336 individuals within our cohort, although the results did not differ in the subsample when we conducted the sensitivity analyses. We did, however, analyze the pulmonary function phenotype using DlCO corrected and it did not significantly change our findings. It is important to note that DlCO predicted values were created based on non-Hispanic White populations that lack diversity. In addition, our cohort is from a tertiary, urban referral center and may not represent the entire population of individuals in the United States with sarcoidosis. In summary, our cohort is a large, diverse cohort and therefore poised to make potentially relevant observations of race and sex differences. Future prospective studies should allow for more inclusiveness with race and ethnicity to detail the differences and therefore, the opportunities to improve clinical outcomes.
Conclusions
Among patients from a tertiary referral center with sarcoidosis and pulmonary function tests, there were significant race and sex differences across pulmonary function phenotypes. There were also significant differences in pulmonary function severity with Black individuals having significantly worse pulmonary function compared with White individuals. Less than half of patients demonstrated restriction, the phenotype classically described in sarcoidosis. Longitudinal studies are required to identify pulmonary function changes over time and the clinical implications of different pulmonary function phenotypes on management strategies and outcomes.
Footnotes
Supported by the National Heart, Lung, and Blood Institute grants K23HL148527 and T32HL007534.
Author Contributions: M.S. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. M.S., K.J.P., A.B., A.V.P., E.S.C., S.-A.W.B., S.C.M., N.A.G., J.C., R. Bascom, R. Berstein, M.N.E., R.A.W., D.R.M., and M.C.M. contributed substantially to the study design, data analysis and interpretation, and/or the writing of the manuscript. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1. Costabel U, Hunninghake GW, Sarcoidosis Statement Committee. American Thoracic Society. European Respiratory Society. World Association for Sarcoidosis and Other Granulomatous Disorders ATS/ERS/WASOG statement on sarcoidosis. Eur Respir J . 1999;14:735–737. doi: 10.1034/j.1399-3003.1999.14d02.x. [DOI] [PubMed] [Google Scholar]
- 2. Baughman RP, Teirstein AS, Judson MA, Rossman MD, Yeager H, Jr, Bresnitz EA, et al. Case Control Etiologic Study of Sarcoidosis (ACCESS) research group Clinical characteristics of patients in a case control study of sarcoidosis. Am J Respir Crit Care Med . 2001;164:1885–1889. doi: 10.1164/ajrccm.164.10.2104046. [DOI] [PubMed] [Google Scholar]
- 3. Markevitz N, Epstein Shochet G, Levi Y, Israeli-Shani L, Shitrit D. Sarcoidosis in Israel: clinical outcome status, organ involvement, and long-term follow-up. Lung . 2017;195:419–424. doi: 10.1007/s00408-017-0015-4. [DOI] [PubMed] [Google Scholar]
- 4. Te HS, Perlman DM, Shenoy C, Steinberger DJ, Cogswell RJ, Roukoz H, et al. Clinical characteristics and organ system involvement in sarcoidosis: comparison of the University of Minnesota Cohort with other cohorts. BMC Pulm Med . 2020;20:155. doi: 10.1186/s12890-020-01191-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Baughman RP, Drent M, Kavuru M, Judson MA, Costabel U, du Bois R, et al. Sarcoidosis Investigators Infliximab therapy in patients with chronic sarcoidosis and pulmonary involvement. Am J Respir Crit Care Med . 2006;174:795–802. doi: 10.1164/rccm.200603-402OC. [DOI] [PubMed] [Google Scholar]
- 6. Baughman RP, Nunes H, Sweiss NJ, Lower EE. Established and experimental medical therapy of pulmonary sarcoidosis. Eur Respir J . 2013;41:1424–1438. doi: 10.1183/09031936.00060612. [DOI] [PubMed] [Google Scholar]
- 7. Judson MA, Baughman RP, Costabel U, Drent M, Gibson KF, Raghu G, et al. Safety and efficacy of ustekinumab or golimumab in patients with chronic sarcoidosis. Eur Respir J . 2014;44:1296–1307. doi: 10.1183/09031936.00000914. [DOI] [PubMed] [Google Scholar]
- 8. Sharma OP, Johnson R. Airway obstruction in sarcoidosis. A study of 123 nonsmoking black American patients with sarcoidosis. Chest . 1988;94:343–346. doi: 10.1378/chest.94.2.343. [DOI] [PubMed] [Google Scholar]
- 9. Kouranos V, Ward S, Kokosi MA, Castillo D, Chua F, Judge EP, et al. Mixed ventilatory defects in pulmonary sarcoidosis: prevalence and clinical features. Chest . 2020;158:2007–2014. doi: 10.1016/j.chest.2020.04.074. [DOI] [PubMed] [Google Scholar]
- 10. Aleksonienė R, Zeleckienė I, Matačiūnas M, Puronaitė R, Jurgauskienė L, Malickaitė R, et al. Relationship between radiologic patterns, pulmonary function values and bronchoalveolar lavage fluid cells in newly diagnosed sarcoidosis. J Thorac Dis . 2017;9:88–95. doi: 10.21037/jtd.2017.01.17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Boros PW, Enright PL, Quanjer PH, Borsboom GJ, Wesolowski SP, Hyatt RE. Impaired lung compliance and DL,CO but no restrictive ventilatory defect in sarcoidosis. Eur Respir J . 2010;36:1315–1322. doi: 10.1183/09031936.00166809. [DOI] [PubMed] [Google Scholar]
- 12. Rossides M, Kullberg S, Eklund A, Grunewald J, Arkema EV. Sarcoidosis diagnosis and treatment in Sweden: a register-based assessment of variations by region and calendar period. Respir Med . 2020;161:105846. doi: 10.1016/j.rmed.2019.105846. [DOI] [PubMed] [Google Scholar]
- 13. Jamilloux Y, Maucort-Boulch D, Kerever S, Gerfaud-Valentin M, Broussolle C, Eb M, et al. Sarcoidosis-related mortality in France: a multiple-cause-of-death analysis. Eur Respir J . 2016;48:1700–1709. doi: 10.1183/13993003.00457-2016. [DOI] [PubMed] [Google Scholar]
- 14. Sharma OP. Sarcoidosis around the world. Clin Chest Med . 2008;29:357–363. doi: 10.1016/j.ccm.2008.03.013. [DOI] [PubMed] [Google Scholar]
- 15. Thillai M, Potiphar L, Eberhardt C, Pareek M, Dhawan R, Kon OM, et al. Obstructive lung function in sarcoidosis may be missed, especially in older white patients. Eur Respir J . 2012;39:775–777. doi: 10.1183/09031936.00103811. [DOI] [PubMed] [Google Scholar]
- 16. Sharp M, Eakin MN, Drent M. Socioeconomic determinants and disparities in sarcoidosis. Curr Opin Pulm Med . 2020;26:568–573. doi: 10.1097/MCP.0000000000000704. [DOI] [PubMed] [Google Scholar]
- 17. Rabin DL, Thompson B, Brown KM, Judson MA, Huang X, Lackland DT, et al. Sarcoidosis: social predictors of severity at presentation. Eur Respir J . 2004;24:601–608. doi: 10.1183/09031936.04.00070503. [DOI] [PubMed] [Google Scholar]
- 18. Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, et al. ATS/ERS Task Force Standardisation of spirometry. Eur Respir J . 2005;26:319–338. doi: 10.1183/09031936.05.00034805. [DOI] [PubMed] [Google Scholar]
- 19. Crouser ED, Maier LA, Wilson KC, Bonham CA, Morgenthau AS, Patterson KC, et al. Diagnosis and detection of sarcoidosis. An Official American Thoracic Society Clinical Practice Guideline. Am J Respir Crit Care Med . 2020;201:e26–e51. doi: 10.1164/rccm.202002-0251ST. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Moller DR, Koth LL, Maier LA, Morris A, Drake W, Rossman M, et al. GRADS Sarcoidosis Study Group Rationale and design of the Genomic Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS) Study. Sarcoidosis protocol. Ann Am Thorac Soc . 2015;12:1561–1571. doi: 10.1513/AnnalsATS.201503-172OT. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Judson MA, Costabel U, Drent M, Wells A, Maier L, Koth L, et al. The WASOG sarcoidosis organ assessment instrument: an update of a previous clinical tool. Sarcoidosis Vasc Diffuse Lung Dis . 2014;31:19–27. [PubMed] [Google Scholar]
- 22. Cooper BG, Stocks J, Hall GL, Culver B, Steenbruggen I, Carter KW, et al. The Global Lung Function Initiative (GLI) Network: bringing the world’s respiratory reference values together. Breathe (Sheff) . 2017;13:e56–e64. doi: 10.1183/20734735.012717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Bhakta NR, Kaminsky DA, Bime C, Thakur N, Hall GL, McCormack MC, et al. Addressing race in pulmonary function testing by aligning intent and evidence with practice and perception. Chest . 2022;161:288–297. doi: 10.1016/j.chest.2021.08.053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Newman LS, Rose CS, Bresnitz EA, Rossman MD, Barnard J, Frederick M, et al. ACCESS Research Group A case control etiologic study of sarcoidosis: environmental and occupational risk factors. Am J Respir Crit Care Med . 2004;170:1324–1330. doi: 10.1164/rccm.200402-249OC. [DOI] [PubMed] [Google Scholar]
- 25. Ringquist EJ. Assessing evidence of environmental inequities: a meta-analysis. Journal of Policy Analysis and Management . 2005;24:223–247. [Google Scholar]
- 26. Rossman MD, Thompson B, Frederick M, Iannuzzi MC, Rybicki BA, Pander JP, et al. HLA and environmental interactions in sarcoidosis. Sarcoidosis Vasc Diffuse Lung Dis . 2008;25:125–132. [PubMed] [Google Scholar]
- 27. Culver DA, Newman LS, Kavuru MS. Gene-environment interactions in sarcoidosis: challenge and opportunity. Clin Dermatol . 2007;25:267–275. doi: 10.1016/j.clindermatol.2007.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Fine MJ, Ibrahim SA, Thomas SB. The role of race and genetics in health disparities research. Am J Public Health . 2005;95:2125–2128. doi: 10.2105/AJPH.2005.076588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Judson MA, Boan AD, Lackland DT. The clinical course of sarcoidosis: presentation, diagnosis, and treatment in a large white and black cohort in the United States. Sarcoidosis Vasc Diffuse Lung Dis . 2012;29:119–127. [PubMed] [Google Scholar]
- 30. Israel HL, Karlin P, Menduke H, DeLisser OG. Factors affecting outcome of sarcoidosis. Influence of race, extrathoracic involvement, and initial radiologic lung lesions. Ann N Y Acad Sci . 1986;465:609–618. doi: 10.1111/j.1749-6632.1986.tb18537.x. [DOI] [PubMed] [Google Scholar]
- 31. Sharp M, Brown T, Chen ES, Rand CS, Moller DR, Eakin MN. Association of medication adherence and clinical outcomes in sarcoidosis. Chest . 2020;158:226–233. doi: 10.1016/j.chest.2020.01.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Gerke AK, Judson MA, Cozier YC, Culver DA, Koth LL. Disease burden and variability in sarcoidosis. Ann Am Thorac Soc . 2017;14:S421–S428. doi: 10.1513/AnnalsATS.201707-564OT. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Schupp JC, Freitag-Wolf S, Bargagli E, Mihailović-Vučinić V, Rottoli P, Grubanovic A, et al. Phenotypes of organ involvement in sarcoidosis. Eur Respir J . 2018;51:1700991. doi: 10.1183/13993003.00991-2017. [DOI] [PubMed] [Google Scholar]
- 34. Vukmirovic M, Yan X, Gibson KF, Gulati M, Schupp JC, DeIuliis G, et al. GRADS Investigators Transcriptomics of bronchoalveolar lavage cells identifies new molecular endotypes of sarcoidosis. Eur Respir J . 2021;58:2002950. doi: 10.1183/13993003.02950-2020. [DOI] [PMC free article] [PubMed] [Google Scholar]