Confirmatory Tissue Sampling in Clinical Stage I Sarcoidosis

To the Editor:

A policy of tissue verification of stage I sarcoidosis (S1S) in subjects presenting with asymptomatic bilateral hilar lymphadenopathy (ABHL) to identify an alternative diagnosis (AD) simulating S1S that might be materially benefited by earlier diagnosis (lymphoma or tuberculosis [TB]) appears to be a self-evident, categorical good. This view was challenged by Winterbauer and colleagues on grounds that ABHL is such a stereotypical feature of sarcoidosis that one can make a confident provisional clinical diagnosis (1) (with confirmation by mandated follow-up), thus sparing individuals with S1S, who comprise the vast majority of subjects presenting with ABHL, from an invasive procedure. In the 50+ years since its publication, to my knowledge, not a single verified exception has been published. The decision to proceed thus rests on a quantitative assessment: What proportion of ABHL is caused by AD? What is the benefit of their earlier ascertainment? How many persons with ABHL must undergo an invasive procedure to confirm an AD? What are the harms and costs of a confirmatory invasive procedure?

I reviewed the abstract or text of the case series cited by the authors of the guideline policy (2) and found that most confounded their analysis by conflating BHL with cases demonstrating radiographically evident mediastinal lymphadenopathy, a feature characteristic of lymphoma. Collectively, they reported 1.96% with an AD, more than 100-fold a prior estimate in which ADs presented as ABHL (3). None furnished a documented instance. Some guideline authors commented on rare reports of metastatic hypernephroma presenting in this fashion. Because earlier diagnosis confers no material benefit, we did not consider it an AD. Based on the product of the incidence and radiographic presentation of AD, we estimated that they comprise ≤0.05% of persons presenting with ABHL (3). We reported that a delay in diagnosis of lymphoma would have, at most, a trivial effect on its course, that primary TB was typically unilateral and self-limited in 95% of cases, and that both lymphoma and progressive primary TB would become evident during mandated follow-up.

Major complications of transbronchial needle aspiration (TBNA) appear to be infrequent, but, absent a reporting requirement or incentive, they cannot be quantified.

Under the assumption that our estimate of the positive predictive value of a clinical diagnosis of S1S based on ABHL underestimated AD by a twofold order of magnitude (i.e., ≤0.05% vs. 1.96%), that TBNA is 100% sensitive for AD and sarcoidosis, that it is complication free, and that the procedural (99–232) charge is $5,000, if 102 persons with ABHL were submitted to TBNA, the net charge would be a half-million dollars, 100 would receive no offsetting, tangible benefit, and 2 would be found to have a lymphoma or TB. Under our estimate of ≤0.05% AD, under the same assumptions, if 10,005 persons with ABHL underwent TBNA, the net charge would be 50 million dollars, 10,000 would receive no tangible benefit, and, at most, 5 persons with an AD might be marginally benefitted by an earlier diagnosis (vs. mandated clinical follow-up).

Tissue confirmation of S1S is indicated in the occasional patient with high-risk history (e.g., lymphoma, renal cell carcinoma, or TB exposure), but, because of the extreme rarity with which an AD (whose earlier diagnosis would be materially beneficial) presents in this fashion, tissue confirmation is not required; it imposes numerous individuals destined to spontaneous resolution to an invasive and costly procedure not offset by commensurate benefit. In the exquisitely rare individual (none have been reported!) with an AD simulating S1S, the diagnosis will become evident with routine, mandated follow-up with trivial to no harm imposed by the diagnostic delay.

The disproportion between potential benefits versus collective harms (including the procedure as a harm) and costs strongly favor a provisional clinical diagnosis of S1S over tissue confirmation.