Atrial strain and occult atrial fibrillation in cryptogenic stroke patients: a systematic review and meta-analysis.

ABSTRACT

BACKGROUND: Cryptogenic stroke (CS) remains a significant cause of morbidity. Failure to identify the underlying pathology increases the rate of recurrence. Atrial fibrillation (AF) seems to be responsible for a substantial proportion of CS. Thus, there is an unmet need to identify and properly treat those with silent AF. PURPOSE: To investigate the association between left atrial strain and newly diagnosed AF in CS patients. OBJECTIVES: We searched major electronic databases for articles assessing the relationship between either peak left atrial longitudinal (PALS) or peak contractile (PACS) strain-quantified using speckle tracking echocardiography-and the incidence of occult AF during the diagnostic work-up of CS patients. RESULTS: Eleven studies (two thousand and eighty-one patients) were analyzed. Incidence of occult AF was 19%. Both PALS and PACS were significantly lower in patients with newly diagnosed AF (MD - 8.6%, 95%CI - 10.7 to - 6.4, I2 86.4% and MD - 5.5, 95%CI - 6.8 to - 4.2, I2 80.8%). According to the diagnostic accuracy meta-analysis, PALS < 20% present 71% (95%CI 47-87%) sensitivity and 71% (95%CI 60-81%) specificity for the diagnosis of occult AF, assuming a prevalence of 20%. The corresponding values for PACS < 11% are 83% (95%CI 57-94%) and 78% (95%CI 56-91%). CONCLUSION: Both PALS and PACS are significantly lower in patients with CS and silent AF. It seems that the cut-off values mentioned above could help physicians in identifying patients who may benefit more from prolonged rhythm monitoring. More studies are needed to confirm these findings.