Insertable cardiac monitors for detection of atrial fibrillation after cryptogenic stroke: a meta-analysis

Background In recent years, the implantable cardiac monitors (ICM) have enhanced the recognition ability of atrial fibrillation (AF), which makes ICM have a new application in AF detection. We conducted a meta-analysis to determine the total incidence of newly found AF detected by ICM after cryptogenic stroke and to evaluate the factors related to the detection of AF. Methods A literature search was conducted in the PubMed, EMBASE, Web of Science, and Cochrane library databases until March 1, 2020. Studies that reported the detection rate of AF using ICM in cryptogenic stroke patients with negative initial AF screening were analyzed. Results A total of 23 studies were included. The overall proportion of AF detected by ICM in cryptogenic stroke patients was 25% (95% confidence interval [CI], 22–29%). The rate of AF detected by ICM was independently related to both cardiac monitoring time (coefficient = 0.0003; 95% CI, 0.0001–0.0005; P = 0.0001) and CHA 2 DS 2 -VASc score (coefficient = 0.0834; 95% CI, 0.0339–0.1329; P = 0.001). In subgroup analysis, we found a significant difference in the detection rate of AF for monitoring duration (< 6 months: 9.6% [95% CI, 4.4–16.4%]; ≥ 6 and ≤ 12 months: 19.3% [95% CI, 15.9–23.0%]; > 12 and ≤ 24 months: 23.6% [95% CI, 19.9–27.5%]; > 24 months and ≤ 36 months: 36.5% [95% CI, 24.2–49.9%]; P < 0.001), and continent (Europe: 26.5% [95% CI, 22.2–31.0%]; North America: 16.0% [95% CI, 10.3–22.6%]; Asia: 17.4% [95% CI, 12.4–23.0%]; P = 0.005). Conclusions The longer the time of ICM monitoring after cryptogenic stroke, the higher the detection rate of AF. Further research is still needed to determine the optimal duration of long-term cardiac monitoring.