Table 3

Extract of the complete standardised form for the retained externally validated tools for SAH

Initial paper (author, date),
score name
Initial study designInitial outcomesInitial predictorsPaper (external validations),
date
Design,
sample size
Outcomes,
nb of events
DiscriminationCalibrationGlobal performanceStrengthsLimitations
De Oliveira Manoel et al
201535
VASOGRADE
3 SAHIT trials (CONSCIOUS-1, EPO trial, statin trial)+1 Canadian centreDCIWFNS, modified Fisher ScaleDengler et al 201736German hosp. registry, monocentric,
2009–2015,
n=423
  • Unfav. outcome (mRS 3–6) at 12 m (53.1%)

  • Unfav. outcome (mRS 4–6) at 12 m

AUC
  • mRS 3–6: 0.711

  • mRS 4–6: 0.709

Histogram pred. vs obs.No
  • Good discrimination

  • Choice of a functional outcome for EV

  • Weak calibration for 2/3 of patients due to a three-level ordinal scale with two predictors

Hijdra et al
198837
Hijdra score
NA
  • Unfav. outcome (GOS 1 to 2) at 1 m

  • DCI

  • Rebleeding

SAH volume, GCSClaassen et al 200438Prospective US monocentric cohort,
1996–2002,
n=413
Unfav. Outcome (mRS 4–6) at 3 m (40.4%)AUC
0.67 (0.61–0.73)
NoNo
  • Weaker discrimination

  • No calibration

Risselada et al
201043
ISAT
ISAT European cohortsMortality at 2 mAge, lumen size, Fisher grade, WFNSDijkland et al 201644Dutch hosp. registry,
Monocentric,
2007–2011,
n=307
Mortality at 2 m (30.6%)AUC
  • WFNS at ttt: 0.89

  • WFNS on adm: 0.82

Calibration curve
  • Intercept at trt: 2.248 on adm: 1.502

  • Slope at trt: 1.417on adm: 1.959

No
  • Good discrimination

  • Calibration curve

  • Low calibration for high-risk SAH

  • Choice of the outcome (short-term mortality)

Jaja et al
201841
SAHIT
SAHIT
(nine international trials and registries)
  • Mortality at 3 m

  • Unfav. outcome (GOS 1 to 3) at 3 m

  • Core: age, premorbid history of HT, WFNS

  • NI: core +CT vol of SAH, aneurysm size, aneurysm location

  • Full: NI +ttt modality

Mascitelli et al 201842US trial cohort,
Monocentric,
2003–2007,
n=338
  • Unfav. outcome (mRS 3–6) at 6 m (29.6%)

  • Mortality at 6 m (10.1%)

AUC
  • Unfav. outcome: core: 72.8 (66.8–78.9) NI: 73.2 (67.1–79.2) Full: 73.4 (67.5–79.4)

  • Mortality: core: 72.1 (62.1–82.2) NI: 73.9 (64.4–83.5) Full: 74.4 (65.1–83.8)

Calibration curve
intercept/slope
  • Brier score

  • Brier scale

  • Extended choice of predictors

  • Good discrimination

  • Good calibration curves

  • Reporting of global performances

Lee et al
201439
HAIR
American monocentric cohortIn-hosp. mortalityHunt & Hess, age, IVH, rebleeding within 24 hoursWitsch et al 201611
  • Cohort SHOP: Prospective US monocentric cohort, 1996–2014,n=1526

  • Cohort CONSCIOUS −1: multicentric Israel, Europe, North America, n=413

Unfav. outcome
(mRS 4–6) at 12 m
(79% SHOP)
(NA CONSCIOUS-1)
AUC
  • SHOP: 88.3 (86.4–90.2)

  • CONSCIOUS-1: 71.8 (66.0–77.5)

No
  • SHOP: N-R²: 0.45 C/S-R²: 0.32

  • CONSCIOUS-1: N-R²: 0.17 C/S-R²: 0.11

  • Long-term functional outcome

  • Good discrimination

  • Weaker perf. on CONSCIOUS-1

  • No calibration

Dengler et al. 201736Hosp. registry
Germany, monocentric
2009–2015,
n=423
  • Unfav. outcome (mRS 3–6) at 12 m (53.1%)

  • Unfav. outcome (mRS 4–6) at 12 months

AUC
  • mRS 3–6: 0.739

  • mRS 4–6: 0.737

Histogram pred. vs obs.No
  • Long-term functional outcome

  • Good discrimination

  • Weak calibration for high-risk SAH

Abulhasan et al 201740Canadian retrospective
monocentric cohort, 2010–2016,
n=434
Mortality at discharge (14.1%)AUC: 0.89Calibration curve
Intercept: −0.05
Slope: 0.77
No
  • Excellent discrimination

  • Calibration curve

Choice of outcome (short-term mortality)
  • adm, admission; AUC, Area Under the Receiving Operative Curve; C/S-R², Cox/Snell-R²; DCI, Delayed Cerebral Ischaemia; GCS, Glasgow Coma Scale; GOS, Glasgow Outcome Scale; Hosp, hospital; HT, HyperTension; ISAT, international subarachnoid aneurysm trial; IVH, intraventricular haemorrhage; m, months; mRS, modified Rankin-Scale; NA, Not available; NI, NeuroImaging; N-R², Nagelkerke-R²; obs, observed; pred, predicted; SAH, SubArachnoid Haemorrhage; SAHIT, SubArachnoid Haemorrhage International Trialists; trt, treatment; unfav., unfavourable; vol, volume; WFNS, World Federation of NeuroSurgeons.