| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Medical University of South Carolina (Dr. Judson), Charleston, SC; George Washington University (Dr. Hirst and Ms. Ghormli), Washington, DC; Case Western Reserve University (Drs. Iyengar and Elston), Cleveland, OH; Henry Ford Health System (Dr. Rybicki), Detroit, MI; University of Cincinnati Medical Center (Dr. Baughman), Cincinnati OH; University of North Carolina School of Medicine (Dr. Donohue), Chapel Hill, NC; Cleveland Clinic Foundation (Dr. Kavuru), Cleveland OH; Johns Hopkins University School of Medicine (Dr. Moller), Baltimore, MD; National Jewish Research and Medical Center (Dr. Newman), Denver, CO; Georgetown University (Dr. Rabin), Washington, DC; University of Pennsylvania Medical Center (Dr. Rossman), Philadelphia PA; and Mt. Sinai Medical Center (Drs. Teirstein and Iannuzzi), New York, NY.
For a complete list of SAGA Study Consortium members, please see Appendix 1.
Correspondence to: Marc A. Judson, MD, FCCP, Division of Pulmonary and Critical Care Medicine CSB-812, Medical University of South Carolina, Charleston, SC 29425; e-mail: judsonma{at}musc.edu
Abstract
Study objective: To test the hypothesis that sibling pairs, who share genes and environmental exposures, might have similar phenotypic expressions of sarcoidosis beyond what would be expected by chance alone.
Design: Multicenter family study with study subjects recruited from 11 clinical centers.
Subjects: Subjects were African-American sibling pairs with sarcoidosis. Sarcoidosis and organ pattern involvement were defined according to specific criteria. Fifteen different organ systems were evaluated.
Results: For full-sibling pairs, ocular involvement was found in both siblings more often than expected by chance alone (p < 0.05), but the concordance was weak (
= 0.18). When analyzing full-sibling and half-sibling pairs, ocular and liver involvement showed a significant concordance between sibling pairs (p < 0.05), but again the agreement was poor ( = 0.16 for both). Concordance in pulmonary function change over time was also weak. Clinical outcomes of sibling pairs were not significantly correlated except for whether treatment was prescribed, and this level of agreement was poor ( = 0.14 for full-sibling and half-sibling pairs; = 0.15 for full-sibling pairs only). Modeling phenotypic expression in sibling pairs using logistic regression did show that the presence of ocular and liver sarcoidosis in the first affected sibling conferred a statistically significant increased risk to the second affected sibling for having those organs involved (odds ratio [OR], 3; 95% confidence interval [CI], 1.7 to 5.4 for ocular; OR, 3.3; 95% CI, 1.5 to 7.4 for liver).
Conclusions: The phenotypic features and clinical outcomes of sarcoidosis in sibling pairs show minimal concordance, with the possible exception that the presence of ocular or liver involvement in the first sibling with a diagnosis of sarcoidosis makes involvement of these organs more likely in other affected siblings.
Key Words: genetics • phenotype • sarcoidosis • sibling
Sarcoidosis is a multisystem granulomatous disease of unknown cause.1 Although the lung is most commonly affected, sarcoidosis may occur in any organ.1 The severity of pulmonary sarcoidosis may range from an abnormal screening chest radiograph finding in a patient free of symptoms, to end-stage lung disease with respiratory failure.23 Some patients have clinical disease that is most prominent in extrapulmonary organs.4 The prognosis of sarcoidosis is highly variable with a tendency to wax and wane, and may resolve either spontaneously or in response to therapy.1 The clinical course is chronic or progressive in 10 to 30% of patients.1
Sarcoidosis is more common in first-degree relatives of an affected patient than in the general population.5 In the A Case Controlled Etiology of Sarcoidosis Study (ACCESS), the odds ratios (ORs) for a sarcoidosis patient having an affected parent or sibling were 16.6 (95% confidence interval [CI], 2.2 to 126.1) and 3.1 (95% CI, 1.4 to 7.1) for white and African-American patients, respectively.5 It has been estimated that approximately 19% of African Americans with sarcoidosis have an affected first-degree relative with the disease, compared to 6% in whites.6 It is not known if affected relatives have similar phenotypic expression of the disease. It is thought that sarcoidosis is caused by an abnormal host response to an unknown exposure in genetically susceptible individuals.1 Conceivably, first-degree relatives who are likely to share genes and environmental exposures might also have similar phenotypic disease expression.
The Sarcoidosis Genetic Analysis (SAGA) study was a multicenter study7 that enrolled African-American sibling pairs with sarcoidosis plus unaffected siblings to perform a genome scan to identify chromosomally linked regions. Affected sibling pairs were phenotypically characterized, gave detailed exposure histories, and were subjected to genetic analysis. The purpose of this article is to describe the phenotypic expression of sarcoidosis in SAGA sibling pairs, and to determine if there was significant similarity in phenotypic expression between sibling pairs.
Materials and Methods
Enrollment and Data Collection
Patients from 11 clinical centers were enrolled in this study (Appendix 1). All clinical centers had this study approved by their institutional review boards. Subjects were required to sign an informed consent statement for participation. Subjects could be enrolled only if they self-designated their race as "black or African American." Only African Americans were selected for the SAGA study6 because they have a higher occurrence of first-degree relatives affected with sarcoidosis than whites. Study subjects were enrolled as families with the minimal family configuration consisting of an initial case (proband) and at least one sibling with a diagnosis of sarcoidosis. Full siblings were defined as siblings who had the same two parents. Half siblings were defined as sharing one of their two parents. Table 1
gives the criteria for the diagnosis of sarcoidosis of each member of the sibling pair. For the purposes of this study, an affected sibling pair was defined as two siblings having "definite" or "highly probable" sarcoidosis on the basis of Table 1. In addition, unaffected siblings for each sibling pair were recruited for enrollment, as were parents. The strategies for enrollment varied at the different clinical centers and included enrollment directly from clinics, contact from local physicians practicing outside the clinical centers, and various forms of advertising. Details about the study design and enrollment experience have been published elsewhere.8
|
An attempt was made to obtain two sets of chest radiographs and spirometry, one at or close to date of diagnosis, and one close to time of enrollment. The chest radiographs were categorized by Scadding stage (stages 0 to 4),9 and improvement or worsening was determined by a change of 
1 Scadding stage. FVC and FEV1 were considered to have improved or worsened if they increased or decreased by > 10% over time. Percentage of predicted FVC and FEV1 were computed for men and women from age and height according to the methods of Hankinson et al.10 The principal investigator (PI) at each clinical center was also asked to make a subjective assessment whether sarcoidosis had gotten better, worse, or stayed the same.
Analysis
To determine the degree that concordance occurred for a particular organ or for other clinical outcomes, we computed statistics for affected sibling pairs using a technique developed by Fleiss.11 While this technique was initially developed for measuring agreement between raters, it has also been applied to sibling analysis in multiple sclerosis12 and rheumatoid arthritis.13 Each variable tested for concordance was dichotomized by the presence or absence of the condition. Proportions of siblings with the condition present were computed to determine if the concordance differed from what would be expected by chance. The resulting statistic ranged from + 1 (indicating perfect concordance) to – 1 (indicating perfect discordance). A separate analysis was performed for full siblings only.
Because is unreliable in cases of extreme lack of variation in the distribution of a response variable, concordance results were not computed for organ systems in which there was almost universal involvement (97.6% of the affected siblings had lung involved) or rare involvement. Fewer than 10% of the affected siblings had neurologic, cardiac, spleen, calcium metabolism, renal, parotid/salivary gland, muscle, or bone marrow involved.
Observed and expected numbers of sibling pairs concordant or discordant for a specific phenotype (or clinical outcome) were computed. Observed and expected numbers counted all permutations of sibling pairs from each family. For continuous variables, intraclass correlations coefficients (ICCs) were used as measures of sibling concordance.
ORs and 95% CIs from general estimating equation logistic regression models were used to estimate the risk of the condition in the sibling with a later date of diagnosis. The presence of the condition in the sibling with an earlier date of diagnosis was used as the predictor in the models. All permutations of sibling pairs within a family were included in the analysis. Each general estimating equation model was adjusted for full-sibling and half-sibling pair status and accounted for the clustering of observations within a family (tendency of members within the same family to be more alike than members between two different families).
The primary study hypothesis in the SAGA study concerned detecting one or more disease predisposing genetic loci in a full genome scan, and the study sample size was configured for that purpose. The present phenotype analysis deals with secondary hypotheses and was considered exploratory in nature. Therefore, no formal corrections of p values for multiple comparisons were made. Results were considered significant if the p < 0.05 level was attained.
Results
The final study sample consisted of 231 families containing 2 affected siblings with sarcoidosis comprising a total of 509 affected siblings. Women represented the majority of the affected siblings in the sample (75.1%). Approximately 15% of the enrolled families were comprised of three or more affected siblings, in which all possible sibling pair combinations were counted; there were 197 pairs, 23 trios, 9 quartets, and 2 quintets of affected siblings with sarcoidosis representing a total of 340 affected sibling pairs. Of these, 25% were half-sibling pairs.
Some phenotype data were missing due to incomplete or unavailable medical records. The degree of missing data for the various organs involved with sarcoidosis ranged from 2% (n = 10) to 11.6% (n = 59). Missing data concerning Scadding radiographic stage and most recent spirometry ranged from 3.7% (n = 19) to 12.2% (n = 62). The only clinical parameters that had > 12.2% of missing data were for FEV1 and FVC closest to the time of the diagnosis of sarcoidosis (28.1% [n = 143] and 28.55% [n = 145], respectively). When the patients with missing spirometry were compared with those with spirometry data available, there were no significant differences in any of the other clinical parameters (data not shown).
Time between diagnosis and most recent spirometry measurements averaged 7.3 years (median, 5.0 years). The mean and SD of FEV1 and FVC among affected siblings at time closest to diagnosis were 2.3 ± 1.1 L and 2.8 ± 0.9 L, respectively. More than 50% of the affected siblings had a change in FVC and FEV1 > 10%, but no significant concordance was found among siblings. Sibling concordance for percentage of predicted FVC and percentage of predicted FEV1 was low (ICCs of 0.27 and 0.23, respectively).
The mean age of affected siblings was 47.0 years (range, 25 to 71 years), and the mean age at diagnosis was 35 years (range, 13 to 68 years). Among sibling pairs, the younger sibling was on average 3 years younger than the older sibling at the time of diagnosis. Sibling concordance for age at diagnosis was weak (ICC = 0.30). Full-sibling pairs and half-sibling pairs were not more likely to be of the same sex than expected by chance alone.
Table 2
shows the sibling concordance by organ system. Overall values were low, ranging from – 0.04 to 0.18, demonstrating poor concordance for these phenotypes. There was significant but weak sibling concordance for ocular ( = 0.16; p < 0.05) and liver involvement ( = 0.16; p < 0.05). Concordance was also statistically significant, but weak, for ocular involvement when only full siblings were considered ( = 0.18; p < 0.05).
|
= 0.14 for full-sibling and half-sibling pairs, = 0.15 for full-sibling pairs alone). Seventy-six percent of the sibling pairs were concordant with respect to ever being treated.
|
|
It has been well established that sarcoidosis is more severe in African-American compared to white populations. African Americans more frequently have symptomatic disease, require treatment, have extrathoracic disease, and have a worse prognosis than whites.1415161718 However, no previous study has systematically examined if the phenotypic expression of sarcoidosis is more similar within families with multiple affected family members compared to unrelated cases. In this study of African-American sibling pairs affected with sarcoidosis, we identified only a few phenotypic similarities that were more common among the siblings, and the correlations were weak. The strongest correlations found were that the second sibling with sarcoidosis was three times more likely to have ocular or liver involvement if the first sibling had involvement of these organs. However, the concordance of these phenotypes was relatively weak. In addition, there was virtually no concordance between the members of sibling pairs in terms of clinical course.
Our approach in analyzing phenotypic expression of a disease in sibling pairs was not unique. A similar approach as been used to examine disease similarity in sibling pairs with multiple sclerosis12 and rheumatoid arthritis.13 Trojano and colleagues12 found that sibling pairs with multiple sclerosis were more likely than unrelated patients to have a similar age of onset, progression of disease, and sensory symptoms. However these investigators could not identify a genetic explanation for these phenotypic differences. Silman et al13 failed to find a greater concordance of rheumatoid arthritis phenotypes within families than between them.
We separated full-sibling pairs and half siblings in our analysis because we suspected that stronger associations would be found between full siblings, since full-sibling pairs share twice as many genes as do half-sibling pairs and could be more likely to share environmental exposures. In both full-sibling and half-sibling pairs, sarcoidosis was diagnosed in the younger sibling at an age approximately 3 years earlier than the older sibling. One possible explanation is that the younger sibling and his/her physician were attuned to the possibility of sarcoidosis after the older sibling had acquired the disease. However, this statistically significant difference in age at diagnosis showed an extremely weak correlation between the siblings (R2 = 0.09 and R2 = 0.11 for the full-sibling and half-sibling pairs and full-sibling pairs, respectively).
No significant differences in baseline spirometry, chest radiograph, or number of organs involved were found between members of sibling pairs. Of the 15 organ systems examined in full-sibling pairs, only eye involvement was statistically more likely in both siblings than by chance alone, but the level of agreement was poor. When the full-sibling and half-sibling pairs were analyzed together in a logistic regression model, we found a statistically significant threefold increased risk of liver or ocular involvement for the second sibling if the first member of the pair had involvement of one of these organs. In terms of degree of pulmonary involvement and clinical course of disease, almost no phenotypic similarities between the siblings were detected regardless whether the full-sibling and half-sibling pairs were analyzed together or only the full-sibling pairs were examined. Sibling pairs were not significantly concordant in baseline radiographic stage, changes in spirometry, number of organ systems involved, chest radiograph, or overall clinical assessment over time. The only statistically significant concordance found concerned the need for treatment. Although this concordance was detected in both the full-sibling and half-sibling pair and the full sibling pair groups, the level of agreement was poor ( = 0.14 and = 0.15, respectively).
There are very few previously published reports of phenotypic similarities among siblings, and these are all case reports. Sharma and colleagues19 reported four siblings with sarcoidosis in which three siblings presented with parotid enlargement and liver involvement. Two of the four siblings had erythema nodosum at presentation. Brennan et al20 identified 13 siblings of sarcoidosis cases in Ireland and noted no significant difference in the mode of presentation or sex distribution between the sibling pairs and the sarcoidosis patients without an affected sibling. Nassif and coauthors21 described 22 French families in which two or three members had sarcoidosis. Sixteen families were white, and 6 were mixed Caribbean. There was a tendency, especially in monozygotic twins, for the dates of revelation of the disease to be very close to each other, and for the clinical and radiologic features to be similar. Human leukocyte antigen polymorphisms were examined, but the study was vastly underpowered to detect any possible relationships.
The cause of sarcoidosis remains unknown, but it is thought to result from by an interplay between environmental exposures and abnormal host responses.1 Evidence suggests that the immunologic process that leads to sarcoidosis begins when an antigen is presented to a T-lymphocyte by an antigen-presenting cell via an human leukocyte antigen class II molecule. This induces a T-helper type 1 response, whereby cytokines are released that result in cellular recruitment and granuloma formation.1 Scandinavian investigators22 have demonstrated a strong association of a specific class of lung T-cells bearing V
ß2,3 T-cell antigen receptors in patients with clinically active sarcoidosis having Lofgren syndrome-type phenotype.
Because siblings share both a portion of their genome and some environmental exposures, it is logical to test the hypothesis that phenotypic expression of sarcoidosis between affected siblings should be more similar than between cases selected at random. The concordance of liver and ocular sarcoidosis between affected siblings supports, albeit weakly, this hypothesis, whereas the other organ systems and phenotypic parameters of sarcoidosis we examined failed to show concordance between siblings.
In addition, if sarcoidosis is a multigenetic disease associated with environmental exposure, then our inability to find close phenotypic similarities between siblings would not necessarily disprove a genetic basis for the disease. It may be that a specific combination of genes determines sarcoid phenotypes. In addition, it is possible that the genes responsible for the development of sarcoidosis are distinct from those that determine the phenotypic expression of the disease. Thus, while affected siblings may share disease predisposing genes, differences in genes and/or exposures important in the determination of sarcoidosis phenotypes may still exist between sibling pairs. The concept of multiple sarcoidosis susceptibility genes is supported by two studies723 that identified multiple linkage sites associated with sarcoidosis. It is also possible that few phenotypic similarities were seen in sibling pairs because environmental influences were different enough to override the genetic similarities. Because a genetic analysis of affected sibling pairs that included environmental histories was not performed in this study, these issues remain unresolved.
There are several potential limitations to our study. The principal investigators at all clinical centers were pulmonologists and were therefore more likely to recruit patients with pulmonary disease. We do not believe that our population was skewed in this regard, as it has been estimated that 90 to 95% of sarcoidosis patients have abnormal chest radiograph findings,24 and 96% (308 of 320 of our subjects) had pulmonary involvement.
Another potential limitation is that there were some inconsistencies in data collection. This was a retrospective multicenter study. The clinical data evaluated in this study were collected from medical record review that may have been incomplete. In addition, physicians caring for the sarcoidosis patients may have differed in their clinical evaluation, and this may have accounted for different frequencies of organ involvement across the clinical centers. However, we doubt that this potential bias significantly impacted our results, in that it is likely that clinically significant organ involvement was detected in most cases and the prevalence of organ involvement was within the range of that found in other studies of African-American sarcoidosis patients.2 It could be argued that organ involvement in sarcoidosis is often subclinical,1 and it may be important to detect both clinical and subclinical organ involvement when classifying sarcoidosis phenotypes. This argument will only be settled when a connection is made between the phenotypic expression of sarcoidosis and genetic patterns. Until this is connection is made, this study used reasonable clinical criteria to detect organ involvement.
The time period over which these data were collected was also not uniform. However, for most sibling pairs, data were collected over many years, so that it is likely that a true estimate of prevalence and phenotypic expression was determined. For example, spirometric data were compared over a median of 5 years, and 85% of the affected sibling pairs had 2 years of follow-up after diagnosis. This study demonstrated more involvement of every organ than the large ACCESS25 study of US sarcoidosis patients, but patients in ACCESS were enrolled within 6 months of diagnosis and extrathoracic manifestations may not be apparent near the time of diagnosis. Finally, our results could have been affected by missing data. However, for all clinical parameters except spirometry, the percentage of missing data was < 12.2%. In the case of spirometry, in which there was missing data in slightly more than one fourth of the cases, there were no significant differences in any other measured clinical parameter between those with and without spirometry data.
For most organ systems, we had medical record data on > 90% of patients, and our sample exceeded 200 pairs, which provided adequate statistical power even with missing data. Nevertheless, the prevalence of organ involvement for some organs was so low that it was difficult to reach meaningful statistical conclusions. Finally, sibling relationships were not ascertained randomly, and an examination of disease course in probands compared with their affected siblings showed a significant greater severity of disease in the former.7 This tendency toward ascertaining index cases with severe disease could bias the overall sibling sample in terms of how well it represents all sibling pairs affected with sarcoidosis.
In conclusion, the phenotypic expression of sarcoidosis sibling pairs shows little concordance in terms of organ involvement, severity of pulmonary involvement, or clinical course. One exception was a modest sibling concordance of liver and ocular involvement. The fact that the few associations found were weak may suggest that sibling phenotypic variation is related to the timing, duration, or dose of environmental exposures. Possibly, specific genes affect the phenotypic expression of sarcoidosis only in the presence of specific environmental exposures. Another possibility is that the genes for sarcoidosis development and phenotypic expression of the disease are distinct. Future studies of the SAGA study population will explore these possibilities.
Appendix
Appendix 1: The SAGA Study Consortium
Clinical Centers
Cleveland Clinic Foundation, Cleveland, OH (M. Kavuru [PI], C. Johnson); Georgetown University, Washington, DC (D. Rabin [PI], H. Yeager, K. Odegbile, L. Stewart); Henry Ford Health System, Detroit, MI (M. Iannuzzi [PI], W. Bibbs, M. Major, B. Rybicki); University of Cincinnati Medical Center, Cincinnati, OH (R.P. Baughman [PI], D. Winget); University of North Carolina, Chapel Hill, NC (J. Donohue [PI], J. Montenegro); University of Pennsylvania Medical Center, Philadelphia PA (M. Rossman [PI], J. Regovich); Johns Hopkins School of Medicine (D. Moller [PI], R. Robinson); Medical University of South Carolina, Charleston SC (M. Judson [PI], S. Forrest); Mt. Sinai School of Medicine, New York, NY (A. Teirstein [PI]); National Jewish Research and Medical Center (L.S. Newman [PI], J. Barnard, C. Rose, L. Silveira).
Coordinating Center
George Washington University, Washington, DC (K. Hirst [PI], L. El Ghormli, S. Fowler, S. Grau, R. Laney, L. Pyle).
DNA Core Laboratory
Henry Ford Health System (M. Maliarik [director]).
Genetics Core Laboratory
Case Western Reserve University, Cleveland OH (R.C. Elston [director], S.K. Iyengar, C. Gray-McGuire).
Project Office
National Heart, Lung, and Blood Institute (National Institutes of Health) [S. Hatch].
Appendix 2: Definition of Definite Organ Involvement
Other than biopsy of an organ showing granulomatous inflammation with no known cause, definite organ involvement is defined for the following organs if any one of the criteria listed is present and there is no other clinical explanation for these findings.
Lung
(1) Mediastinal or hilar lymph node biopsy showing granulomatous inflammation of no known cause
(2) Chest radiograph showing bilateral hilar adenopathy
(3) Chest radiograph showing diffuse infiltrates
(4) Chest radiograph showing upper lobe fibrosis
Neurologic
(1) Cranial nerve abnormality
(2) Culture-negative lymphocytic meningitis
(3) Positive magnetic resonance scan result with uptake in the meninges, brain, brainstem, or spinal cord
(4) Diabetes insipidus
Ocular
(1) Anterior or posterior uveitis
(2) Lacrimal gland enlargement
Cardiac
(1) Steroid-responsive cardiomyopathy
(2) ECG showing intraventricular conduction delay or nodal block
(3) Cardiac magnetic resonance scan showing cardiac uptake
(4) 67Ga scan showing cardiac uptake
Liver
(1) Serum alkaline phosphatase greater than three times the upper limit of normal
(2) Serum total bilirubin greater than three times the upper limit of normal
(3) Aspartate aminotransferase or alanine aminotransferase greater than three times the upper limit of normal
Calcium Metabolism
(1) Hypercalcuria
(2) Hypercalcemia
(3) Nephrolithiasis with stone analysis showing calcium
Renal
(1) Steroid-responsive renal failure
Skin
(1) Erythema nodosum
(2) Lupus pernio
(3) Waxy nodular lesions around nose or eyelids
Parotid/Salivary Gland
(1) 67Ga scan showing parotid/salivary gland uptake
(2) Parotiditis/gland enlargement
Muscle
(1) Elevated creatine phosphokinase/aldolase that responds to treatment
Bone Marrow
(1) Unexplained thrombocytopenia
Bone and Joint
(1) Cystic changes on radiographs of hands or feet phalanges
Footnotes
Abbreviations: ACCESS = A Case Controlled Etiology of Sarcoidosis Study; CI = confidence interval; ICC = intraclass correlations coefficient; OR = odds ratio; PI = principal investigator; SAGA = Sarcoidosis Genetic Analysis
This work was supported by National Institutes of Health Grant U01 HL060263.
None of the authors have any conflicts of interest.
Received for publication November 7, 2005. Accepted for publication February 21, 2006.
References
ß2,3 gene usage by CD4+ T lymphocytes in bronchoalveolar lavage fluid from sarcoidosis patients correlates with HLA-DR#. Eur J Immunol 1992;22,129-135[ISI][Medline]Read all eLetters
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
