Chest ACCP Member Benefits
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (25)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Reich, J. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Reich, J. M.
(Chest. 2002;121:32-39.)
© 2002 American College of Chest Physicians

Mortality of Intrathoracic Sarcoidosis in Referral vs Population-Based Settings*

Influence of Stage, Ethnicity, and Corticosteroid Therapy

Jerome M. Reich, MD, FCCP

* From the Center for Health Research, Kaiser Permanente, Northwest Region, Portland, OR.

Correspondence to: Jerome M. Reich, MD, FCCP, 5051 SW Barnes Rd, Portland, OR 97221-1517; e-mail: Reichje{at}dnamail.com


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Meta-analysis and Review
 Discussion
 Conclusions
 References
 
Study objectives: To compare the sarcoidosis mortality in referral settings (RS) and population-based settings (PS), and to identify the contribution of stage, ethnicity, and corticosteroid therapy (CST) to their disparate outcomes.

Design: All observational studies identified in a MEDLINE search and bibliographic review published in the English language since 1960 dealing with the course and prognosis of sarcoidosis in large, unsorted, adult, ambulatory RS and PS providing long-term follow-up were reviewed and subjected to meta-analysis.

Measurements and results: Sarcoidosis mortality in RS (4.8%), in which 17% of patients had the most unfavorable prognosis as judged by stage (stage III), was 10-fold that reported in PS (0.5%), in which 11% of patients were identified at this stage. The magnitude of this disparity could not be accounted for solely by adverse selection, as indicated by stage or by ethnicity. Patients in RS received CST with sevenfold the frequency of PS, and its provision was highly correlated with stage-normalized mortality.

Conclusion: The prognosis of patients with intrathoracic sarcoidosis in PS is far more favorable than that obtained in RS. Sarcoidosis mortality is largely independent of ethnicity. The possibility cannot be excluded that excessive employment of CST may unfavorably influence the long-term course of the disease in some individuals.

Key Words: drug therapy • ethnology • meta-analysis • mortality • sarcoidosis


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Meta-analysis and Review
 Discussion
 Conclusions
 References
 
While some authors cite a sarcoidosis mortality of 10%,1 2 a range of 1 to 5% was provided in a recently published joint statement of the American Thoracic Society, European Respiratory Society, and World Association of Sarcoidosis and Other Granulomatous Diseases.3 The authors of the joint statement ascribed the fivefold range of sarcoidosis mortality to patient source (referral bias) and genetic factors, principally ethnicity: individuals of Scandinavian descent are believed to experience the most, and African Americans the least, favorable course.

The populations served by referral settings (RS) are a composite of unselected patients derived from their serviced community plus a variable proportion of patients who have been referred, most often because of the expectation or realization of an unfavorable course. Most reports on the course and prognosis of sarcoidosis originate in specialty clinics, university settings, and tertiary-care facilities in which high mortality is likely to mirror referral bias. Population-based settings (PS) such as health maintenance organizations and government clinics in Scandinavian countries, by contrast, serve a defined population, and are therefore free from referral bias. The ethnic composition of inner city residents served by universities in the United States, in which sarcoidosis mortality is high, is largely African American. Conversely, the reported prognosis of sarcoidosis in the ethnically white, Scandinavian PS have been highly favorable. Thus, the attribution of unfavorable outcomes to a combination of referral bias and ethnic composition appears highly plausible. Neither explanation of the outcome disparity, however, has been subjected to critical analysis.

The objectives of this study are to ascertain the sarcoidosis mortality in RS and PS, to identify the contribution of adverse selection, ethnicity, and corticosteroid therapy (CST) to this outcome disparity and, in light of these findings, to assess the implications of a policy of providing CST for individuals with pulmonary shadowing and trivial or no symptoms who do not demonstrate improvement or resolution within 6 to 12 months. This study is limited to patients with intrathoracic sarcoidosis reported from Western countries because cardiopulmonary failure due to pulmonary fibrosis is the predominant cause of sarcoidosis mortality in Western countries.

Stage was selected as the sole criterion of disease severity for comparison between studies because of its value as a prognostic guide,3 4 stage I being the most favorable and stage III being the least favorable, and because most compendia supply this information. Persons with stage III have the highest mortality, and this stage correlates with most other prognostic guides—chronicity, extrathoracic involvement, physiologic impairment, and progression—that are omitted in reports on course and prognosis. The proportion of patients with stage III was employed as a metric of adverse selection. This value need not equate to referral bias, ie, referral of patients with an adverse course or those judged to be at high risk. The proportion of patients with stage III will, in addition, be inversely related to the proportion identified by means of mass population radiographic screening (who are asymptomatic and preponderantly in stage I). In Scandinavian countries, which provide this service, about half of patients are identified by this means. Referral bias can be quantified if the author provides explicit information; it can be inferred if the studied population includes a substantial proportion of patients previously hospitalized for treatment. The contribution of African-American ethnicity to sarcoidosis mortality can be estimated both by comparing the mortality in series of differing ethnic composition, and the overall mortality in African-Americans adjusted for incidence.

The indications for CST in patients with pulmonary shadowing (stages II, III, and IV) remain unresolved3 ; it is not, however, currently recommended for patients with stage I sarcoidosis (isolated hilar adenopathy). CST is not considered a risk factor for mortality; it provides consistent short-term benefit (its long-term benefit is less certain), and its employment is confounded by its indication, disease severity. However, because few studies supplied specific indications for its employment, and because of the strength of the association between its provision and mortality in RS, the possibility of a causal association was explored. To minimize confounding by disease severity, the comparative strength of the association of CST vs stage III with mortality, the consistency between studies, and its correlation with stage-normalized mortality were evaluated. Cumulative sarcoidosis mortality—hereafter referred to as "mortality"—was selected as the sole outcome metric because other outcome measures—clinical, physiologic, or radiographic course—were not universally provided.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Meta-analysis and Review
 Discussion
 Conclusions
 References
 
All reports dealing with the long-term course and prognosis of sarcoidosis originating in large, outpatient adult practices in Western countries, unsorted by chronicity, classified according to stage, and identified in a MEDLINE search and bibliographic survey of publications in the English language from 1960 through 2000, were reviewed and allocated by origin in RS or PS. The search terms employed were sarcoidosis AND course, AND prognosis, AND outcome. Compendia based on hospitalized patients were excluded to minimize the effect of outcome disparity attributable to patient source. To ensure an adequate duration of follow-up, compendia based solely on newly diagnosed patients were excluded. Series not susceptible to meta-analysis because they lacked stage or CST information were reviewed and summarized to broaden geographic and ethnic representation. Japanese RS were omitted both because there were no comparable PS studies and because the biology of sarcoidosis in Japan differs from that in Western countries.

The proportion of patients identified by means of either population screening radiographs or with stage I disease (both categories have, almost invariably, a favorable outcome) were employed as markers of nonadverse selection. Conversely, the difference in the proportion of patients identified with stage III disease in RS vs PS was employed as a metric of adverse selection. Few reports distinguished between stage III (pulmonary shadowing without hilar adenopathy) and stage IV (radiographic evidence of pulmonary fibrosis); patients in these two categories were therefore pooled under the rubric "stage III."

Statistical Analysis
Reports suitable for meta-analysis supplied stage, CST (with two exceptions), and mortality data. Meta-analysis was employed to compare the outcomes in RS and PS by computing the ratios of the two predictive variables—stage III and CST—to mortality. The analysis was extended by correlating mortality in each report with the proportion with stage III and the proportion receiving CST. While stage III frequency distribution was approximately normal, CST and mortality were bimodal in distribution. For this reason, Spearman’s rank correlation coefficient (rs) was employed. Because the population—rather than a sample—of eligible studies was analyzed, p values are omitted. Attributable risk (this term is used in its statistical sense; it does not imply causation) was computed as the square of the correlation coefficient (rs2). Linear trend lines were calculated by the least-squares fit method.


   Meta-analysis and Review
TOP
 Abstract
 Introduction
 Materials and Methods
 Meta-analysis and Review
Discussion
 Conclusions
 References
 
Ten reports fulfilling the inclusion criteria were identified4 5 6 7 8 9 10 11 12 13 ; they are summarized in Tables 1 , 2 .


View this table:
[in this window]
[in a new window]

  		Table 1.. Sarcoidosis Mortality Reported From RS

 

View this table:
[in this window]
[in a new window]

  		Table 2.. Sarcoidosis Mortality Reported From Population-Based Settings

 
RS Patient Source
Two RS series provided information indicating referral bias. Scadding8 stated that, " ... many [all were stage IV patients] were referred to me after prolonged periods of observation by other physicians because of an unfavorable course." It is noteworthy that the proportion of patients with stage III (ie, III and IV) disease in this series (47%) was more than threefold that of the case-weighted mean of the remaining six tabulated RS series (15%). Reisner’s series14 of 69 patients, which was not tabulated because staging was not provided, comprised 11 patients (16%) referred following hospitalization for severe disease and 58 patients identified in the community. The six remaining compendia4 5 6 7 8 9 10 summarized in Table 1 lacked patient source information that would serve as a measure of referral bias, ie, the proportion referred because of an adverse course vs the proportion assembled from the ambulatory population served by the institution.

The proportion of patients in each setting identified by mass population screening chest radiographs is an indicator of nonadverse selection; asymptomatic persons identified by this means would be expected to have the most favorable prognosis. Where figures are available, the proportion identified by radiographic screening was higher in the PS than in the RS compendia. Among the former, the series by Huhti et al15 had the highest percentage (64%), reflecting the performance of triennial mass screening, followed by the series by Hillerdal et al11 (57%) and Romer12 (50%). In RS reporting this information, the proportions, though lower than in PS, were substantial: 34% (Siltzbach et al4 ), 46% (Smellie and Hoyle5 ), and 30% (Johnston10 ).

Stage
More than half the patients in the two tabulated Scandinavian PS11 12 studies were identified by mass population screening; this, rather than referral bias, may account for the higher proportion of patients with stage I in the PS (59%) than the RS (49%) and the lower proportion of least favorable stage (III) in PS (11%) than in RS (17%).

Ethnicity
The populations in the European PS and RS series5 7 8 9 10 11 12 15 16 17 were predominantly or exclusively white. Of four Western RS series summarized by Siltzbach et al,4 those originating in London and Paris were predominantly white, and those in New York and Los Angeles were predominantly African American. The US RS series reported by Sones and Israel6 was 87% African American. The US RS series by Reisner14 reported on 69 patients; 54 (79%) were African-American. The Mayo Clinic PS18 reported on 75 patients; most were of northern European extraction, and one was African American. This ethnicity was well represented in only one PS series13 in which 23 of 86 (27%) were African Americans.

Outliers
The proportion of stage III patients in the series by Scadding8 was far higher than that of the remaining six tabulated RS series. Stage III patients constituted 24% in the US PS study,13 nearly threefold that of the remaining two PS series (9%).11 12 Johnston’s10 experience at the University Chest Clinic in Dundee, Scotland, differed markedly from other RS with respect to the indication for CST, the frequency with which it was provided, and mortality. It is therefore discussed separately below.

CST
Both Hillerdal et al11 and Johnston10 provided CST only to patients exhibiting progressive pulmonary shadowing. The authors did not specify the indications for CST in the remaining tabulated reports. The proportion of patients receiving CST (41%) in RS was sevenfold that of PS (5.9%). Scadding8 treated 54 of 136 patients (40%); however, because he observed no sustained benefit in asymptomatic patients, only 29 of the 54 (21%) received CST for > 3 to 8 weeks. Despite evidence of adverse selection, indicated by both patient source and stage, the mortality (6.6%) in this series was within the range of the remaining RS (Table 1) .

Mortality in Untabulated Series
RS: Bunn and Johnston17 reported a zero mortality in 71 patients identified at the Dundee Chest Clinic or the University of Dundee, Scotland, 44% of whom had been identified by screening radiographs. Staging was not provided. Only six patients (7%) received CST for pulmonary indications. Reisner14 in a series of patients followed in the Chest Clinic, New York City Department of Health, reported a zero mortality in the nonhospital-referred group of 58, and two sarcoidosis deaths (total, 2.9%) in 11 patients referred following hospitalization for severe disease. Staging was not provided. None received CST.

PS: Huhti et al15 reported zero mortality in a study of 178 patients with intrathoracic sarcoidosis identified in Oulu, Finland, and followed up for 5 to 12 years. Sixty-three patients (36%) showed pulmonary shadowing at inception; 22 patients (12%) received CST. This study was not tabulated because the author’s classification merged patients with stages II and III. Two series supplied mortality information incident to their primary purpose, which was epidemiologic; neither provided CST data. Henke et al18 reported a mortality of 1.4 (1 of 75 patients) from The Mayo Clinic, Rochester, MN. Kolek16 reported a series of nearly 1,500 patients in Moravia and Silesia, 5% of whom had stage III, and stated that, "... advanced cases causing death were extremely rare." In summary, the mortality rate in the RS exhibiting restraint in the provision of CST was zero; in three, geographically diverse PS studies, it mirrored the tabulated PS studies that supplied more information.

Mortality in Tabulated Series
The case-weighted mortality ratio, RS:PS was 4.8:0.5, or 10:1. With the exception of the report by Johnston,10 which is discussed separately below, mortality clustered in two narrow ranges with no overlap: in RS,4 5 6 7 8 9 the case-weighted mean is 5.1% (range, 3.6); in PS,11 12 13 it is 0.5% (range, 0.8).

Correlation of Mortality With Predictive Variables
In data from the eight series4 6 7 8 9 10 12 13 (Table 3 ) providing specific information on mortality, stage, and CST, mortality, as expected, correlated with stage III (Fig 1 ; rs = 0.37; attributable risk, 14%). The correlation between CST and mortality was slightly stronger and by visual inspection, closer (Fig 2 ; rs = 0.39; attributable risk, 15%.) To minimize the confounding effect of disease severity, the correlation between CST and mortality was further evaluated by stage-normalizing mortality (dividing mortality by the proportion with stage III in each series). Figure 3 shows a close linear relationship between provision of CST and stage-normalized mortality (rs = 0.68; attributable risk of the product of stage III and CST, 46%). Solving this linear regression equation for mortality gives:

where M = mortality, S = stage III, and C = CST. Solving for predicted M by inputting the S and C case-weighted data from the last rows of Tables 1 , 2 gives a value of 4.9 (actual, 4.8) for RS, and of 0 (actual, 0.05) for PS.


View this table:
[in this window]
[in a new window]

  		Table 3.. Data From Studies With Predictive Variables

 


View larger version (59K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1.. Stage III vs mortality. See variables in Table 3 .

 


View larger version (60K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2.. CST vs mortality. See variables in Table 3 .

 


View larger version (59K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3.. CST vs stage-normalized mortality. See variables in Table 3 .

 

   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Meta-analysis and Review
 Discussion
Conclusions
 References
 
Imprecision of the available data, in combination with varying means of data presentation, constitute the major limitation of this study. Cumulative mortality is an imprecise estimate of mortality rate (which was not supplied), neither the dose or duration of CST was provided, no study supplied details of physiologic impairment, and only one study10 detailed the radiographic extent of disease. Patient age and duration of follow-up were supplied by different authors in noncomparable formats: range, mean, frequency distribution, modal decade, minimum or maximum (for follow-up). Because there were no obvious systematic biases within or between the RS and PS in any of these data fields, it seems reasonable to expect that pooling the available studies, in combination with the one-log magnitude of the mortality differences, will largely offset this limitation.

The variability in sarcoidosis mortality has been ascribed to a combination of referral bias and ethnicity.3 With the qualification that, stage for stage, patients in RS might have more severe disease than those in PS (stage-independent disease severity [SIDS]), adverse selection, as evidenced by advanced stage, is insufficient to account for a 10-fold difference in mortality.

While the frequency distribution of patients with stage III found by conflating the tabulated RS and PS is normal, and the proportions in the two settings overlap, the frequency distribution of mortality and CST in the RS and PS settings—with the exception of Johnston’s series, which is discussed below—are bimodal. Under the assumption that the proportion of patients adversely selected for referral would differ in each RS, one would expect the range in mortality between RS (4 to 7.6; Table 1 ) and PS (0 to 0.8; Table 2 ) to overlap, exhibiting a unimodal distribution skewed to the right. Instead, there is a fivefold difference between the highest PS and the lowest RS mortality. The bimodal mortality distribution must be the consequence of a marked and systematic divergence in the two settings in one or more of the following: duration of follow-up, referral late in the course of the disease, SIDS, or treatment policy. The first two are unlikely since the observations in most cited reports extended over one or more decades, and the ages of the patients, when provided, did not notably differ in the two settings.

Stage and Mortality
That stage III is a valid marker of adverse selection is evidenced by the experience of Scadding,8 in which the proportion of patients with this stage was more than threefold that of the case-weighted remaining RS series, and in which adverse selection, based on clinical assessment was explicitly detailed.

Under the assumption that persons with stage III disease approximate the total at risk of dying of sarcoidosis, one can compute a stage-corrected mortality in the PS by adjusting for the difference in proportion of this stage in the two settings. This would increase the mortality in PS by a factor of 17/11 (Tables 1 , 2) , from 0.5 to 0.8, and diminish the mortality ratio, RS:PS to 4.8/0.8 = 6. Conversely, if 100% of PS patients, under the same assumption, were identified with stage III disease, the mortality would increase by a factor of 100/11 x 0.5 = 4.5%, still less than the RS rate. Furthermore, four of the RS series,4 5 7 9 with similar mortalities (Table 1) , had a smaller proportion of stage III (12 to 22%) patients than the US PS study (24%),13 with zero mortality (Table 2) .

Ethnicity and Mortality
The adverse prognostic import of African-American ancestry, to which the unfavorable outcome in some RS has been attributed, has not been consistently observed. In the compilation by Siltzbach et al,4 the mortality in preponderantly African-American populations in Los Angeles (4%) and New York (5%) did not differ markedly from the largely white populations reported from Paris (1.8%) or London (5%). Absence of ethnic-specific mortality data in this compilation precludes analysis of the African-American component. Israel stated, "Studies in Philadelphia have shown no ethnic difference in character or stage of thoracic involvement or in frequency of clearing. Large scale studies by the Veterans Administration have shown no ethnic differences in the outcome of sarcoidosis."19 Reisner’s series14 is particularly noteworthy, not only because of clear evidence of referral bias and its ethnic composition, but also because of its favorable outcome in comparison with other RS. The author reported on 69 patients with intrathoracic sarcoidosis followed in a New York City health clinic. Although 79% of the patients were African American, they constituted only 66% of those categorized as experiencing unfavorable outcomes (defined as radiographic or clinical deterioration). None of the patients received CST. There were two deaths (2.9%) from sarcoidosis in this series, both from the group of 11 referred following hospitalization for severe disease, half the mortality of the tabulated RS series. There were no deaths in the US PS13 study and no difference in outcome by stage between 23 African Americans and 62 whites.

Gideon and Mannino20 reported a sarcoidosis mortality rate ratio for US African-Americans vs whites of 14:1; however, the incidence rate ratio has been estimated at 13:1.21 Thus, the differential mortality appears to be due largely or entirely to a difference in case rate—not to a difference in case-fatality rate. In summary, although African Americans evidence, overall, a more exuberant response, with greater clinical evidence of systemic involvement, their mortality does not appear to differ appreciably from that experienced by whites.

Similarly treated white ethnic groups in series originating in Scandinavian countries,11 12 15 Scotland,10 17 and the United States (white component),13 experienced a similar mortality, < 1%. The central European experience16 was highly favorable, but details are lacking.

CST and Mortality
Absent information on the proportion of patients referred to RS because of adverse prognostic features—as opposed to those seen because of indigence, health-care policy, or population radiographic screening—it is not possible to determine either to what extent the far greater utilization of CST in RS reflects treatment policy vs disease severity, or whether the differential distribution of SIDS between RS and PS is sufficient to account for a mortality difference of this magnitude. It was not possible, in most instances, to determine specifically, from the supplied data, the stage-adjusted rate of CST. Furthermore, there were no specific data on the mortality among CST recipients in the two settings.

Characteristics that suggest that an association is causal are its strength, consistency between studies, dose-response relationship, and biological plausibility. Confounding by disease severity does not appear to fully account for the strength of the association between CST and mortality: the sixfold stage-adjusted difference in mortality in RS vs PS is closely aligned with far higher proportion of patients in RS who received CST (Tables 1 , 2 : 41/5.9 = 7). Moreover, the association between restraint in the provision of CST and a favorable outcome was evidenced in all series, irrespective of ethnic composition or geographic origin. The close linear relationship of CST and stage-normalized mortality (Fig 3) suggests that the propensity to provide CST is a major outcome determinative. Mortality can be modeled as a function of stage III and CST; and because stage III was normally distributed, and overlapped in the two settings, the bimodal distribution of mortality in the tabulated series (with the exception of the Johnston10 report, as noted) must, of necessity, be attributed to the bimodal distribution of CST.

The possibility that the provision of CST in RS is confounded by measures of disease severity unreflected by stage (SIDS) requires consideration. This explanation appears implausible for the following reasons: (1) To achieve the close linear relationship evident in Figure 3 as a consequence of confounding by SIDS presupposes a nearly 1:1 correspondence between SIDS and the provision of CST in each RS that is unlikely; (2) The bimodal provision of CST presupposes a consistency in referral bias according to SIDS that is unlikely to be present in each RS; (3) Johnston10 reported zero mortality in his RS series which provided a 20-year follow-up; 17% were stage III, similar to the tabulated RS. CST for pulmonary disease was provided for only 4 of 159 patients (3%). If the higher mortality in RS were, in part, attributable to SIDS, one would have to hypothesize that this setting was unique in not receiving patients with SIDS. Similar considerations apply to confounding by duration of observation and disease chronicity.

Under the assumption that sarcoidosis is a genetically conditioned immunologic response to specific environmental agents,3 it is biologically plausible that CST, by attenuating this response, might exert, in susceptible individuals, a long-term adverse effect.

Other Potential Confounding Variables and Mortality
Neither late-age onset, disease chronicity, or duration of follow-up are plausible explanatory candidates for the mortality difference or the decision to employ CST in the RS and PS series, since both the age distribution and length of observation (which extended over 1 to 2 decades in most series) appeared roughly similar. For example, the modal age range in the RS series by Neville et al7 was 21 to 30 years, which suggests that the majority had disease of recent onset. Neither does symptom severity appear likely to account for the frequency with which CST was employed in this series, which had the highest percentage of stage I sarcoidosis (65%) of any tabulated, 45% of whom were treated, presumably to interdict potential progression. In the RS series by Chappel et al,9 11% of patients with stage I disease received CST either at presentation or for failure to resolve.

Outcomes in Series With Defined Treatment Indications
Scadding’s8 decision to provide sustained CST only to symptomatic patients exhibiting a favorable response was associated with a low mortality relative to stage. The conservative treatment policy adopted by Hillerdal et al,11 in a PS, was associated with a highly favorable outcome: CST was given only to persons exhibiting rapidly progressive pulmonary shadowing. Of 490 persons, 4 patients (0.8%) died of respiratory failure—at a mean age of 74 years—due to pulmonary fibrosis. The authors stated: "Treatment with corticosteroids has been recommended for patients in whom stage II or III persists for > 6 to 12 months, but, as illustrated by the present series, the value of this recommendation is questionable: restoration to normal can be expected in most cases even without treatment." Johnston,10 whose sole pulmonary indication for CST was progressive shadowing, reported that only 4 of his 159 Scottish RS patients required it. Although the stages of his patients were nearly identical with the case-weighted mean of the other RS (Table 1) —81 patients (51%) were stage I and 27 patients (17%), stage III—he reported no deaths attributed to sarcoidosis.

Number Needed To Treat
Approximately 50%11 (precise figures are unavailable) of patients with intrathoracic sarcoidosis will present with or evolve pulmonary shadowing that will be sustained for > 6 to 12 months. In a hypothetical cohort of 200 patients in a PS setting, a policy of treating those failing to improve within this interval, and in whom treatment is not compelled by symptom severity, would necessitate sustained CST in somewhat < 100 patients in hopes of preventing one death. In addition, the possibility that CST may exert a long-term adverse effect on the course of the disease in some individuals renders judicious the recommendation proposed by Hillerdal et al11 : that absent compelling symptoms, treatment be confined to patients experiencing rapidly progressive pulmonary shadowing.


   Conclusions
TOP
 Abstract
 Introduction
 Materials and Methods
 Meta-analysis and Review
 Discussion
 Conclusions
References
 
(1) Sarcoidosis mortality representative of patients who are likely to be encountered by community physicians in western countries is 0.5%, less than one tenth of that reported from RS; (2) although ethnic differences strongly influence the incidence and clinical manifestations of sarcoidosis, they have little appreciable influence on its mortality; and (3) the possibility that a lower threshold for administering CST is associated with an increased mortality risk, beyond the selection effect that distinguishes the RS and PS group, is open to question.


   Footnotes
 
Abbreviations: CST = corticosteroid therapy; PS = population settings; rs = Spearman’s rank correlation coefficient; RS = referral settings; rs2 = square of correlation coefficient; SIDS = stage-independent disease severity

Received for publication November 7, 2000. Accepted for publication August 8, 2001.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Meta-analysis and Review
 Discussion
 Conclusions
 References
 

  1. Crystal, RG (1998) Sarcoidosis. Fauci, AS Braunwald, E Isselbacher, KJ eds. Harrison’s principles of internal medicine 14th ed. ,1922-1928 McGraw Hill New York, NY.
  2. Fraser, RS, Müller, NL, Paré, PD (1999) Diagnosis of diseases of the chest 4th ed. ,1533-1583 W.B. Saunders Philadelphia, PA.
  3. . American Thoracic Society, European Respiratory Society, World Association of Sarcoidosis and Other Granulomatous Diseases (1999) Statement on sarcoidosis. Am J Respir Crit Care Med 160,736-755[Free Full Text]
  4. Siltzbach, LE, Geraint James, DG, Neville, E, et al (1974) Course and prognosis of sarcoidosis around the world. Am J Med 57,847-852[CrossRef][ISI][Medline]
  5. Smellie, H, Hoyle, C (1960) The natural history of pulmonary sarcoidosis. Q J Med 116,539-558
  6. Sones, M, Israel, HL (1960) Course and prognosis of sarcoidosis. Am J Med 29,84-93[CrossRef][ISI][Medline]
  7. Neville, E, Walker, AN, Geraint James, D (1983) Prognostic factors predicting the outcome of sarcoidosis: an analysis of 818 patients. Q J Med 208,525-533
  8. Scadding, JG (1961) Prognosis of intrathoracic sarcoidosis in England: a review of 136 cases after five years of observation. BMJ 4,1165-1172
  9. Chappel, AG, Cheung, W, Hutchings, HA (2000) Sarcoidosis: a long term follow up study. Sarcoidosis Vasc Diffuse Lung Dis 17,167-173[ISI][Medline]
  10. Johnston, RN (1986) Pulmonary sarcoidosis after ten to twenty years. Scott Med J 31,72-78[ISI][Medline]
  11. Hillerdal, G, Nou, E, Osterman, K, et al (1984) Sarcoidosis epidemiology and prognosis: a 15-year European study. Am Rev Respir Dis 130,29-32[ISI][Medline]
  12. Romer, FK (1982) Presentation of sarcoidosis and outcome of pulmonary changes. Dan Med Bull 29,27-32[ISI][Medline]
  13. Reich, JM, Johnson, RE (1985) Course and prognosis of sarcoidosis in a nonreferral setting: analysis of 86 patients observed for 10 years. Am J Med 78,61-67[CrossRef][ISI][Medline]
  14. Reisner, D (1967) Observations on the course and prognosis of sarcoidosis. Am Rev Respir Dis 96,361-368[ISI][Medline]
  15. Huhti, E, Poukkula, A, Lilja, M (1987) Prognosis for sarcoidosis in a defined geographical area. Br J Dis Chest 81,381-390[CrossRef][ISI][Medline]
  16. Kolek, V (1994) Epidemiologic study on sarcoidosis in Moravia and Silesia. Sarcoidosis 11,110-112[ISI][Medline]
  17. Bunn, D, Johnston, R (1972) A ten year study of sarcoidosis. Br J Dis Chest 66,45-52[Medline]
  18. Henke, CW, Henke, G, Elveback, LR, et al (1986) The epidemiology of sarcoidosis in Rochester, Minnesota: a population-based study of incidence and survival. Am J Epidemiol 123,840-845[Abstract/Free Full Text]
  19. Israel, HL (1970) Prognosis of sarcoidosis [letter]. Ann Intern Med 73,1038-1039[CrossRef]
  20. Gideon, NM, Mannino, DM (1996) Sarcoidosis mortality in the United States, 1979–1991. Am J Med 100,423-427[CrossRef][ISI][Medline]
  21. Reich, JM, Johnson, RE (1996) Estimated incidence of clinically identified sarcoidosis in a Northwest United States population. Sarcoidosis Vasc Diffuse Lung Dis 13,173-177[ISI][Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (25)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Reich, J. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Reich, J. M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS