Perianal Tuberculosis: Lessons Learned in 57 Patients From 743 Samples of Histopathology and Polymerase Chain Reaction and a Systematic Review of Literature.

BACKGROUND

Few data are published on perianal tuberculosis.

OBJECTIVE

This study aimed to determine the best method to diagnose tuberculosis in patients with fistula-in-ano and to conduct a systematic review to determine the incidence and characteristics of tuberculosis fistula-in-ano.

DATA SOURCES

The prospective study data and existing literature were derived from PubMed, Google scholar, and Scopus STUDY SELECTION:: Prospective analysis of patients with tuberculous fistula-in-ano treated between 2014 and 2018 was conducted, and a systematic review of studies describing ≥3 patients with tuberculosis fistula-in-ano was completed.

INTERVENTION

Testing of tuberculosis was performed by histopathology or polymerase chain reaction of tissue or pus from the fistula tract.

MAIN OUTCOME MEASURES

The primary outcomes measured were the detection rate of various tests to detect tuberculosis in fistula-in-ano and the prevalence rate of tuberculosis in simple versus complex fistulas.

RESULTS

In 637 samples (410 patients) tested, tuberculosis was detected in 49 samples (43 patients). Additional samples (n = 106) sent in patients with a high index of suspicion tested positive in 14 more patients. Thus, overall, 63 samples tested positive in 57 patients (total: 743 samples in 410 patients were tested). Tuberculosis was detected in 2 of 181 patients (1.1%) in tissue (histopathology), in 28 of 341 patients (8.2%) in tissue (polymerase chain reaction), and in 19 of 115 patients (16.5%) in pus (polymerase chain reaction) samples. To detect tuberculosis, tissue (polymerase chain reaction) was significantly better than tissue (histopathology) (28/341 vs 2/181, p < 0.00001) and pus (polymerase chain reaction) was significantly better than tissue (polymerase chain reaction) (19/115 vs 28/341, p < 0.0009). Tuberculosis was significantly more common in complex fistulas than in simple fistulas (20.3% vs 7.2%, p = 0.0002). The systematic review (n = 199) highlighted that tubercular fistulas are more common in recurrent and complex fistulas and in tuberculosis endemic regions.

LIMITATIONS

The true sensitivity and specificity of each testing modality could not be determined because not all patients with tuberculosis fistula-in-ano were tested by every diagnostic modality studied.

CONCLUSIONS

The tuberculosis detection rate of polymerase chain reaction was significantly higher than histopathology. Among polymerase chain reaction, pus had higher detection rate than tissue. Tuberculosis was associated with more complex and recurrent fistulas.