Table 3.

Sensitivity and subgroup analyses for studies evaluating flow-mediated dilation (FMD) in convalescent coronavirus disease 2019 (COVID-19) patients and controls.

	N of studies	N of datasets	N of patients	Effect size
Sensitivity analyses	A. ‘High quality’ studies
	7	9	517 COVID-19 489 controls	MD: −2.30 (95% CI: −3.45, −1.15); p < 0.0001 I2 = 88.4%; p < 0.0001 PI: −6.25, 1.66
	B. Exclusion of participants with cardiovascular risk factors or history of events
	5	6	202 COVID-19 188 controls	MD: −1.73 (95% CI: −3.04, −0.41); p = 0.010 I2 = 92.5%; p < 0.0001 PI: −6.38, 2.93
	C. Exclusion of paediatric populations
	11	16	606 COVID-19 624 controls	MD: −2.28 (95% CI: −3.19, −1.37); p < 0.0001 I2 = 91.8%; p < 0.0001 PI: −6.04, 1.48
	D. Use of an automatic edge detection software
	4	5	227 COVID-19 213 controls	MD: −1.53 (95% CI: −2.98, −0.07); p = 0.040 I2 = 93.9%; p < 0.0001 PI: −7.09, 4.04
Subgroup analyses	E. Follow-up ≤ 3 months
	9	11	408 COVID-19 383 controls	MD: −2.20 (95% CI: −3.35, −1.05); p < 0.0001 I2 = 91.2%; p < 0.0001 PI: −6.37, 1.97
	F. Follow-up > 3 months
	3	5	211 COVID-19 253 controls	MD: −2.53 (95% CI: −4.19, −0.86); p = 0.003 I2 = 94.1%; p < 0.0001 PI: −8.98, 3.93

Panel A: ‘high quality’ studies (Newcastle–Ottawa Scale ≥ 7); Panel B: studies specifically excluding participants with any cardiovascular risk factor or history of coronary artery disease; Panel C: exclusion of studies on paediatric populations; Panel D: studies using an automatic edge detection software for FMD assessment; Panel E: studies evaluating convalescent patients within the first 3 months of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection; Panel F: studies evaluating COVID-19 participants more than 3 months after recovery.

N: number; MD: mean difference; 95% CI: 95% confidence interval; PI: prediction interval.