Getting the invite list right: a discussion of sepsis severity scoring systems in severe complicated intra-abdominal sepsis and randomized trial inclusion criteria

Table 3 Predictive capabilities of potential COOL study sepsis and critical illness scoring systems using Helsinki outcomes data

System	Identified	Outcome mortality	Sensitivity (%)	Specificity (%)	AUC^b	95% CI
qSOFA ≥ 2	34 (36.6%)	13 (32%)	37	95	0.723	[0.653–0.792]
SOFA ≥ 2	72 (77.4%)	25 (27%)	77	83	0.825	[0.766–0.885]
SOFA ≥ 3	70 (75.3%)	24 (27%)	75	85
SOFA ≥ 4	60 (64.5%)	20 (28%)	65	91
Septic shock^a	36 (38.7%)	15 (37%)	39	96	0.82	[0.761–0.88]
MPI ≥ 30	48 (51.6%)	21 (28%)	51	79	0.774	[0.713–0.835]
MPI ≥ 32	42 (45.2%)	18 (32%)	45	89
MPI ≥ 34	22 (23.7%)	9 (33%)	24	96
APACHE II ≥ 14	64 (68.8%)	24 (26%)	69	78	0.828	[0.775–0.881]
APACHE II ≥ 16	52 (55.9%)	20 (30%)	56	89
APACHE II ≥ 18	42 (45.2%)	19 (39%)	45	95
WSESSSS ≥ 8	68 (73.1%)	27 (27%)	73	76	0.809	[0.752–0.866]
WSESSSS ≥ 9	58 (62.4%)	24 (29%)	62	82
WSESSSS ≥ 10	47 (50.5%)	20 (32%)	51	88
CPIRO ≥ 3	54 (58.1%)	21 (31%)	58	90	0.856	[0.806–0.905]
CPIRO ≥ 4	28 (30.1%)	13 (42%)	30	98

Ninety-three patients were selected out of the database based on 30-day mortality or ICU admission
^aAUC is for sepsis classification according to Sepsis-3 consensus definitions
^bOnly one area under the curve (AUC) was calculated for each scoring system without thresholds within

ISSN: 1749-7922