Predictive markers related to local and systemic inflammation in severe COVID-19-associated ARDS: a prospective single-center analysis

Table 2 Correlations of biomarkers vs. clinical and neurological short- and long-time outcome

Biomarker	Outcome	r-value	p-value	Multivariate analysis	n =
1st Leukos	Leukos max	0.5288	0.0001	N/A	47
	Leukos mean	0.4566	0.0013	N/A	47
	1st PCT	0.3991	0.0066	N/A	45
	Horovitz min	− 0.3483	0.0164	ns	47
	NO supplementation	0.3544	0.0145	N/A	47
	ECMO/ECLS	0.2938	0.0450	ns	47
	RASS mean	− 0.3713	0.0102	0.0450	47
1st PCT	Leukos max	0.3471	0.0195	N/A	45
	ECMO/ECLS	0.4080	0.0054	0.0162	45
	mRS admission	0.5480	< 0.0001	N/A	45
	EQ-5D-5L index	− 0.4647	0.0338	0.0138, Intercept* (0.0203)	21
TLR3 blood	Age	0.3261	0.0308	N/A	44
	1st PCT	− 0.3129	0.0436	N/A	42
	Ppeak	− 0.3386	0.0376	N/A	38
	Horovitz min	0.3940	0.0081	ns	44
	ECMO/ECLS	− 0.3810	0.0107	ns	44
	RASS mean	0.4474	0.0023	0.0159	44
	Sedation	− 0.4059	0.0063	0.0055	44
	mRS 14d	− 0.3080	0.0445	N/A	43
	mRS discharge	− 0.3044	0.0445	N/A	44
	mRS 6 m	− 0.3184	0.0352	0.0114	44
	EQ-5D-5L index	0.6066	0.0059	N/A	19
IL-6 blood	RASS mean	− 0.3813	0.0152	ns	40
IL-6 blood	mRS 14d	0.3560	0.0261	N/A	39
IL-8 blood	1st PCT	0.4993	0.0012	N/A	39
	Horovitz min	− 0.4755	0.0019	ns	40
	NO supplementation	0.3794	0.0158	N/A	40
	ECMO/ECLS	0.3784	0.0161	ns	40
	RASS mean	− 0.5132	0.0007	ns	40
	mRS admission	0.4089	0.0088	N/A	40
	mRS 14d	0.4416	0.0049	N/A	39
	EQ-5D-5L index	− 0.5267	0.0142	Intercept* (0.0276)	21
IL-6 BAL	LOS ICU	0.7774	0.0104	N/A	10
	Ppeak	0.5740	0.0220	N/A	16
	Horovitz mean	− 0.5224	0.0399	ns	16
	Sedation	0.6924	0.0010	ns	19
	mRS admission	0.6411	0.0031	N/A	19
IL-8 BAL	Horovitz mean	− 0.5882	0.0185	ns	16
	Sedation	0.5895	0.0079	ns	19
	mRS admission	0.6408	0.0031	N/A	19

Shown are those correlations which reached significance. Depending on whether data showed normal distribution, Spearman nonparametrical or Pearson parametrical test was used for correlation testing. Akaike’s Information Criterion, Tjur’s R squared and Hosmer–Lemeshow hypothesis test were used for multivariate analyses. Multivariate analyses were composed of the respective biomarker, CCI and sex. Outcome parameters with a more descriptive than predictive value (when assessed at a similar time to sample acquisition, for example mRS at admission) were not tested in these models to avoid unnecessary tests. Similarly, markers with less clinical importance than other markers (for example Ppeak vs. Horovitz index) were not included into multivariate analyses. N/A, not tested or could not compute model. *p < 0.05. LOS ICU length of stay in intensive care unit, Ppeak mean peak ventilation pressure, NO nitric oxide, ECMO/ECLS extracorporeal membrane oxygenation/extracorporeal life support, RASS Richmond Agitation-Sedation Scale, mRS modified Rankin scale, BAL bronchoalveolar lavage

ISSN: 1471-2334