Few studies have addressed the BeLPT in patients with manifest disease.^17,18,21,22 To our knowledge, this is the largest analysis of the BeLPT in patients with granulomatous disease. A peak in proliferative response is observed early in the course of sensitization,²³ possibly linked to a higher proportion of poorly proliferating beryllium-specific effector memory T cells in more advanced stages.²⁴ In addition, proliferating T cells are detected in sarcoidosis,²⁵ and nonspecific or bystander activation of T cells in cell culture might interfere with specific BeLPTs.²⁶ Therefore, it seemed important to assess the performance of the BeLPT in this particular setting. We showed that the BeLPT can detect sensitization in manifest disease with high specificity. The main drawback of the BeLPT as a screening tool is its lack of sensitivity.^14,16 We confirmed this finding in a diagnostic setting.

This finding may have been influenced by our testing strategy: negative BeLPTs were usually repeated until either positive results were obtained or CBD was considered ruled out. Once sensitization criteria were met, testing was usually discontinued. This led to accumulation of negative tests in some patients but not of positive tests, magnifying the effect of false-negatives on sensitivity. Also, to limit false-positives, we chose a conservative threshold of 3 SDs. Lowering the cutoff to 2 SDs increased sensitivity, but at the cost of unacceptably low specificity for a clinical setting. Results between 2 and 3 SDs neither increased nor decreased the likelihood of sensitization, which supports interpreting these results as borderline, which requires retesting. The lack of sensitivity can be countered by repeated testing, as shown by the merging of split-sample tests. In every fourth patient, ≥ 5 tests were required to prove sensitization (e-Table 5). On average, sensitized patients received 1.3 more tests than nonsensitized patients, which is explained by the need for confirmation after the first positive test.

Despite similar test accuracies, PBMC and BAL BeLPTs appeared complementary. When both were performed, 64.3% of sensitized patients had positive results in only 1 modality. This might be explained by the compartmentalization of beryllium-specific T cells.²⁷ The addition of BAL proved beneficial because it identified sensitization in 27 cases where multiple PBMC BeLPTs had failed to demonstrate sensitization (45.0% of sensitized patients in whom BAL was performed). The high number of uninterpretable BAL BeLPTs should be noted. This was mostly because of low proliferation in positive control wells and can be attributed to inhibition of in vitro lymphocyte proliferation by alveolar macrophages²⁸ and poor cell recovery in patients with obstructive ventilatory defect.

Assessing the accuracy of the BeLPT is limited by the lack of a reference standard other than the BeLPT itself. Therefore, true sensitization can only be confirmed by repeated BeLPT testing, as previously done by Stange et al.¹⁴ This self-confirmatory approach is inherently biased because the BeLPT serves as its own reference standard and, given the low sensitivity, likely underestimates the rate of sensitization. Some cases may have been misclassified as sarcoidosis simply because the BeLPT was not repeated often enough. Therefore, the proposed numbers can only be taken as approximations of the test accuracy.

We could not show differences in false-negative testing between patients with CBD receiving and not receiving immunosuppressive treatment. However, none received high-dose corticosteroids and only 5 received second-line treatment because we hold or taper immunosuppression before testing whenever possible, which we would generally advocate for (e-Table 10).

It has been demonstrated that cases of CBD could be detected when performing the BeLPT in metal-exposed patients formerly diagnosed with sarcoidosis^17,18; however, this finding is inconsistently reported.²² In this study, the BeLPT revealed that 41.4% of metal-exposed patients who would otherwise have been misdiagnosed with sarcoidosis in fact had CBD. Only in 17 cases (8.1%) were we unable to make a clear distinction between CBD and sarcoidosis based on the BeLPT.

The BeLPT was performed in patients with suspected beryllium exposure, but because beryllium exposure is often occult, we broadly offer the BeLPT to patients with metal exposure. Except for electricians, who were underrepresented in CBD, occupations were similar across groups, with none of them emerging as more common in CBD than in sarcoidosis. Frequent occupations included those notoriously associated with high levels exposure (eg, machining) but also occupations with less conspicuous exposure, notably a high proportion of dental technicians, as previously observed,^17,29 even though beryllium is no longer used in dental alloys in Germany.^17,30 Thus, CBD should be considered as a differential diagnosis to sarcoidosis in a variety of work fields, not only in the beryllium industry itself, but also in downstream use of beryllium components, where awareness for beryllium exposure may not be given.³¹

Most patients in this analysis evidenced abnormal PFT results, which is striking when compared with other sarcoidosis cohorts.^32,33 This might be related to their specific, potentially ongoing inhalational exposure. Patients with CBD exhibited even lower lung function and more restrictive patterns compared with sarcoidosis. Except for higher age in CBD, confounding factors (eg, smoking, comorbidities, obesity) did not differ between both groups. Therefore, we may postulate that CBD causes more severe lung function impairment than sarcoidosis. A recent study comparing CBD and beryllium-exposed sarcoidosis found lower lung function in sarcoidosis.³⁴ However, patients with sarcoidosis were referred because of abnormal imaging, and patients with CBD were referred after positive screening on the BeLPT and generally had normal imaging. Conversely, in the present population, the cause for referral was almost universally disease manifestation, with the only difference between groups being the outcome of the BeLPT. Additionally, this demonstrates that in the absence of routine surveillance, BeS is identified late in its natural history. Earlier exposure cessation may have attenuated the clinical course of the patients.

Given the diversity and nonspecificity of extrapulmonary manifestations, and the difficulty in confirming most of them, an etiologic connection to the disease often remained presumptive. As such, we cannot accurately deduce the true prevalence of extrapulmonary manifestations in this cohort, especially because they were not evaluated systematically. Nonetheless, we think it is noteworthy that several patients with CBD displayed nonrespiratory symptoms, with a similar proportion as those with sarcoidosis. These mainly included eye and skin manifestations, well-recognized manifestations of CBD,⁷ but on some occasions we also found involvement of internal organs (e-Table 9).

The study is limited by its retrospective design, which might imply inconsistencies in data acquisition. Because our center performs the BeLPT for external referrals, we had access to blood samples but lacked clinical and lung functional data in some cases. Similarly, follow-up data were too scarce to make longitudinal assumptions. The included patients may not be representative of the general sarcoidosis population because they share specific exposures which occasioned the BeLPT, whereas other patients with sarcoidosis likely have various exposure backgrounds. Given these selection biases, we cannot infer the true prevalence of CBD in larger populations of patients with apparent sarcoidosis.