Cystic echinococcosis (CE) is caused by infection with the larval stage of Echinococcus granulosus sensu lato
. This develops as fluid-filled cysts, most frequently in the liver, after accidental ingestion of parasite eggs [1
]. The parasite life cycle develops mostly between domestic dogs and livestock ungulates, and the infection is prevalent worldwide where livestock breeding is practiced [2
]. The spectrum of clinical manifestations of human CE ranges from being asymptomatic to causing serious, disabling pathology; when present, symptoms are nonspecific [4
CE cysts pass through different stages, from unilocular, fluid-filled CE1, to more “complex” stages (CE2, CE3a, CE3b), and “solid” inactive stages (CE4, CE5) [5
]. Ultrasound is at the basis of the diagnosis and staging of abdominal CE, which are pivotal to guide clinical decision-making [5
]. Pathognomonic features of CE can be visualized with ultrasound; however, in their absence or when clinicians are not experienced in recognizing them, the diagnosis of CE may be difficult. The wide spectrum of differential diagnoses ranges from biliary cysts to malignancies. Serology is applied to support imaging in doubtful cases; however, the interpretation of serological test results may be challenging. Tests are not standardized, often lack appropriate validation, and their performances vary widely [7
]. Furthermore, serology should be applied only after a lesion is visualized on imaging; however, most seroassay performance studies were not designed to fulfill this criterion [8
]. False positive reactions may occur, with frequent cross-reactivity in case of infection with E. multilocularis
, causing alveolar echinococcosis (AE) [7
]. Misdiagnosis of CE and AE [11
], and likewise of other etiologies, may have dramatic consequences. Finally, sensitivity is influenced by a number of factors, resulting in false negative results [7
]. Therefore, a negative serology may occur even in the presence of active CE.
Serology results must be interpreted in the light of imaging characteristics and pre-test probabilities of a lesion being CE or another etiology [13
]. At present, no evidence-based consensus algorithm exists to guide the use of serological tests in the diagnostic process of CE-evocative lesions. In a retrospective study, it has been suggested that the application of a Western blot test, as a single-test approach or as a confirmation after two first-level tests, might have the best performance [14
Here, we aim to evaluate the performances of nine serological tests commercialized in Europe for the diagnosis of “echinococcosis”, using a panel of well-characterized sera from patients with hepatic CE and with non-CE lesions potentially in differential diagnosis with CE. The work is presented according to STARD (Standards for Reporting Diagnostic accuracy studies) recommendations [15
The diagnosis of CE may be challenging due to the different morphology of CE cysts, their wide spectrum of differential diagnoses, and the often little familiarity of physicians with this infection. Serology is applied when the diagnosis cannot be based on imaging alone. The current WHO-IWGE Expert Consensus [5
] recommends the application of “a high-sensitivity serological test, confirmed by a separate high-specificity serological test”; however, no consensus algorithm is available to guide the application and interpretation of serology. Based on a retrospective analysis of serology results, it was suggested that WB, alone or to confirm the result of two first-line tests, might have the best performance [14
]. Here, we examined the diagnostic performances of nine serological tests commercialized in Europe for the diagnosis of “echinococcosis”, and their combination, using a panel of well-characterized, clinically appropriate sera.
We did not pre-set performance thresholds to define an assay as acceptable; however, we excluded three tests from the combination analysis due to their unsatisfactory performances. For all other tests, specificity was very high. This is not surprising because control sera were from selected patients with hepatic focal lesions, coming from areas not endemic for CE, and did not include patients with AE. Therefore, the results of our study may apply to the clinical settings in low/nonendemic areas for CE, but should be translated with caution to high-CE or CE/AE co-endemic areas.
Several papers reported the diagnostic performance of some of the commercial seroassays evaluated in this work [8
]. As diagnostic performances substantially depend on the characteristic of CE cysts and controls forming the examined sample cohort [13
], it is difficult to directly compare the results of studies, even more so when cyst stages and other relevant factors are not described and the aims of the studies differ. However, diagnostic accuracy parameters found in this study were, overall, within the reported ranges for those tests previously evaluated in the published literature [8
In accordance with previous results of a study with comparable design [14
], the LDBIO WB assay resulted the best single test to apply in the presence of suggestive focal liver lesions (Se 83%, 95% CI 72–91%; Sp 98%, 95% CI 91–100%). Of note, the visible band identifying the only false positive result of this test was faint and irregular; therefore, the specificity of the assay was probably underestimated in this study. When considering other assays and combinations, higher sensitivities were, logically, achieved when borderline results were considered as positive and using the “one-positive-only” approach. However, this approach should probably not be recommended in all settings, especially in areas of intense E. granulosus
circulation and co-endemicity with potentially cross-reactive parasitoses, where false positive results are expected to be much more frequent. Since a diagnosis of CE cannot be excluded by a negative serology and prompts further investigations, a false negative result would be a more “cautionary” result. We therefore privileged the interpretation of borderline results as negative and of test combinations using the “concordant-positive” approach. When applying these stringent criteria, the application of two first-level tests followed by WB in case of discordant results (Se 67–73%, 95% CI 55–86%; Sp 100%, 95% CI 94–100%), or in case of discordant and concordant negative results (Se 73–86%, 95% CI 61–93%; Sp 98–100%, 95% CI 92–100%), had the best performances.
Deserving a final mention is the accuracy of Echinococcus
species identification based on WB banding pattern. It is well known that WB can discriminate CE and AE only in a percentage of cases [10
]; furthermore, our results show that a variable proportion of sera from patients with CE and from uninfected patients can be identified as AE based on WB banding pattern. This is of great concern because CE and AE (and other causes of focal liver lesions) have very different clinical management and prognoses. Therefore, a very cautious interpretation of WB banding pattern for Echinococcus
species identification is absolutely needed, which in no way can constitute the sole basis of species identification.
This study had several limitations. Sera from patients with AE were unavailable; furthermore, controls did not come from the same areas of patients with CE. This may have been responsible for the very high specificity of all tests, and highlights the need to perform a similar study in CE-endemic and CE/AE co-endemic areas. We could not exclude pulmonary CE in control patients. This possible source of false-positive results, however, seems of marginal concern, because very few false-positives were obtained and because anyway lung CE is associated with a low rate of seropositivity [31
]. Finally, half of sera from patients with CE came from patients who received albendazole treatment, one third of which recently (File S1
), that may have increased tests sensitivities [13
]. On the other hand, the strength of this study is the use of a relatively large sample size of sera, from patients with well-characterized lesions, and selected using stringent criteria, to minimize the possible influence of previous treatment on serology results. This allowed evaluating assays performance in the conditions most similar to a first diagnosis of the focal liver lesion.