. 2020 Nov 3;133:109393. doi: 10.1016/j.ejrad.2020.109393

Table 3.

Subgroup meta-regression analysis of COVID-19 related neuroimaging findings.

Covariate	Cerebral microhemorrhages			Acute/subacute infarct			Spontaneous acute ICH			Encephalitis/encephalopathy
Imaging modality	Pooled incidences (%) [95 % CI]	I² (%)	P	Pooled incidences (%) [95 % CI]	I² (%)	P	Pooled incidences (%) [95 % CI]	I² (%)	P	Pooled incidences (%) [95 % CI]	I² (%)	P
MRI	13.8 [10.5–17.2]	96	<0.001	20.0 [7.9–32.2]	89	0.413	3.9 [0.6–7.3]	83	0.458	6.9 [4.2–9.7]	95	<0.001
CT or MRI	3.1 [1.0–5.2]	86		27.1 [16.4–37.7]	98		6.1 [3.3–8.9]	86		2.1 [0.4–3.7]	88
Region^a
Europe	14.0 [10.6–17.4]	96	<0.001	22.5 [14.7–30.3]	89	0.389	7.1 [3.4–10.8]	89	0.149	12.9 [9.3–16.4]	96	<0.001
USA	4.1 [1.1–7.1]	88		17.0 [9.1–24.8]	92		3.5 [0.1–7.0]	87		1.8 [0–4.7]	64
Asia	3.0 [0–8.8]	63		13.9 [1.7–26.0]	82		2.4 [0–7.8]	0		2.6 [0–7.4]	91
Mean/median age^b
>65 years	10.1 [5.6–14.7]	98	0.326	33.4 [17.2–49.6]	99	0.27	10.6 [4.8–16.4]	76	0.25	1.7 [0–5.2]	8	<0.001
≤65 years	6.9 [3.4–10.3]	88		18.0 [5.6–30.4]	81		6.6 [2.9–10.2]	89		10.5 [7.6–13.4]	96
Multicenter
Yes	5.7 [1.8–9.6]	87	0.119	30.4 [16.8–44.0]	99	0.257	6.1 [2.6–9.5]	92	0.525	3.3 [1.1–5.5]	95	0.545
No	9.9 [6.7–13.1]	96		20.0 [9.0–31.1]	88		5.0 [1.8–8.1]	78		4.3 [2.1–6.4]	85
Indication for imaging^c
Neurologic symptoms	6.1 [3.8–8.3]	95	0.053	23.2 [12.4–33.9]	98	0.624	7.8 [4.4–11.1]	88	0.574	5.0 [2.9–7.1]	93	0.054
Not reported or other clinical indication	11.8 [5.6–17.9]	78		29.2 [8.6–49.7]	78		4.3 [0–10.2]	85		9.7 [4.7–14.7]	95
Patients in ICU
Reported	11.8 [8.8–14.8]	96	<0.001	17.2 [5.3–29.1]	79	0.118	6.2 [2.8–9.5]	87	0.544	11.1 [8.2–14.0]	96	<0.001
Not reported	3.2 [0.6–5.7]	84		30.9 [19.1–42.7]	98		4.8 [1.6–7.9]	88		0.7 [0–2.7]	63

CI = confidence interval; ICH = intracranial hemorrhage; I²=Higgin’s inconsistency index.

A multinational study by Shahjouei et al involving all three continents was not included.

Three studies (Sawlani et al, Freeman et al, and Agarwal et al) not reporting mean/median ages were excluded for cerebral microhemorrhages; four studies (Xiong et al, Sawlani et al, Freeman et al, and Agarwal et al) were excluded in other three categories for the same reason.

Three studies (Lin et al, Freeman et al, and Agarwal et al) not reporting indication for neuroimaging were excluded.