Background
Main text
Methods
Literature search
Complexity of prediction scores
Examples score validation
Standard protocol approvals, and patient consents
Results
Literature review
Study (First Author, publication year) | Score name | Geographic location of derivation cohort (N of score derivation cohort) | Score components | Outcome measure(s) | Timing of outcome measures (in original score publication) | Score performance measures (in original score publication) |
---|---|---|---|---|---|---|
Tuhrim et al., 1988 [11] | No name | USA (82) | GCS ICH volume Pulse pressure | Mortality | 30 days | Expected-observed classification |
Tuhrim et al., 1991 [12] | No name | USA (191) | GCS ICH volume IVH Pulse pressure | Mortality | 30 days | Expected-observed classification |
Broderick et al., 1993 [13] | No name | USA (188) | GCS ICH volume | Mortality | 30 days | Sensitivity, specificity, PPV |
Masé et al., 1995 [14] | No name | Italy (138) | GCS ICH volume IVH | Mortality | 30 days | Expected-observed classification |
Hemphill et al., 2001 [15] | ICH score | USA (152) | Age GCS ICH volume Infratentorial origin IVH | Mortality | 30 days | Descriptive |
Cheung et al., 2003 [16] | New ICH score | Hong Kong (142) | IVH NIHSS Pulse pressure Subarachnoid extension Temperature | Mortality Favorable outcome (mRS < 3) | 30 days | Sensitivity, specificity, PPV, NPV, Youden index |
Godoy et al., 2006 [17] | Modified ICH Scores (mICH-A, −B) | Argentina (153) | Age Comorbidity GCS ICH volume IVH Infratentorial origin | Mortality Favorable outcome (GOS 4–5) | 30 days (mort.) 6 months (GOS) | Sensitivity, specificity, PPV, NPV, AUC, Youden index |
Weimar et al., 2006 [9] | Essen ICH score | Germany (260) | Age Level of consciousness NIHSS | Functional recovery (BI > 90) Favorable outcome (GOS 4–5, or BI > 50) | 100 days (functional recovery) 6 and 12 months (favorable outcome) | Sensitivity, specificity, AUC, external validation (independent cohort, n = 173) |
Ruiz-Sandoval et al., 2007 [18] | ICH grading scale | Mexico (378) | Age GCS ICH volume (supratentorial or infratentorial) IVH Location | Mortality | In-hospital 30 days | AUC, R2 |
Cho et al., 2008 [19] | Modified ICH score (mICH score) | China (226) | GCS ICH volume IVH or hydrocephalus | Mortality Favorable outcome (GOS 4–5 OR BI > 50) | 6 months 12 months (both endpoints at both time points) | AUC, Youden index |
Rost et al., 2008 [20] | FUNC score | USA (418) | Age GCS ICH location ICH volume Pre-ICH cognitive impairment | Functional independence (GOS 4–5) | 90 days | AUC, External validation (in independent patient cohort from same institution, n = 211) |
Chuang et al., 2009 [21] | Simplified ICH score | Taiwan (293) | Age Dialysis dependence GCS History of hypertension Serum glucose | Mortality | 30 days | Sensitivity, specificity, PPV, NPV, positive/negative likelihood ratios, AUC |
Li et al., 2012 [22] | ICH Index (ICHI) | China (227) | Age GCS Glucose WBC | Mortality | In-hospital | AUC |
Ji et al., 2013 [23] | ICH functional outcome score (ICH-FOS) | China (1953) | Age GCS Glucose ICH location ICH volume (supratentorial or infratentorial) IVH NIHSS | Mortality Unfavorable outcome (mRS 3–6) | 30 days 3, 6, and 12 months | Hosmer-Lemeshow test, AUC, external validation (in independent patient cohort from same institution, n = 1302) |
Romero et al., 2013 [24] | Spot sign score (SSSc) | USA (131) | CT characteristics Number of spot signs Maximum axial dimension Maximum attenuation | [ICH expansion] Mortality Unfavorable outcome | In-hospital (mortality) 3 months (mortality and unfavorable outcome) | Descriptive |
Zis et al., 2015 [25] | Emergency department ICH score (EDICH) | Greece (191) | GCS ICH location ICH volume INR IVH | Mortality | 30 days | Sensitivity, specificity, AUC |
Gupta et al., 2017 [26] | ICH outomes project (ICHOP) scores (ICHOP3, ICHOP12) | USA (365) | APACHE II GCS ICH volume NIHSS Pre-morbid mRS | Unfavorable outcome (mRS 4–6) | 3 and 12 months | AUC, McFadden R2, Cox&Snell R2, Nagelkerke R2 |
Sembill et al., 2017 [27] | Max ICH score | Germany (583) | Age IVH Lobar ICH volume NIHSS Non-lobar ICH volume Oral anticoagulation | Unfavorable outcome (mRS 4–6) | 12 months | AUC, Youden index |
Braksick et al., 2018 [28] | ICH scoreFS | USA (274) | Age FOUR score ICH volume Infratentorial origin IVH | Mortality | 30 days | AUC |
Scores and score complexities
Validation of the Essen-ICH-score
Discrimination
Calibration
Net benefit
Discussion
Clinical use of currently available scores
Problems of current prognostic scores
The withdrawal of care bias
Simplicity, accuracy, and timing of score assessment
Framework for future ICH prediction scores
Quality measures for score development and validation
Integration of new predictor variables
Physiologic variables | |
Serum hemoglobin [61] | |
Cerebral perfusion pressure and partial pressure of oxygen in interstitial brain tissue (PbtO2) [64] | |
Serum iron/ferritin/transferrin [65] | |
Chronic kidney disease [66] | |
Imaging Variables | |
Peak PHE [68] | |
Spot sign/island sign/black hole sign/blend sign [69] | |
EEG variables | Electrographic seizures [71] |
Periodic discharges [71] |
Beyond binary outcomes: modified Rankin scale and patient-reported outcomes
Pragmatic prognostication in clinical practice
First approach to the patient | Clarify code status. Be aware of your own biases. Especially in patients with large ICH volumes and/or IVH extension and/or hydrocephalus do not reflexively, consciously or subconsciously, provide sub-maximal care. Unless patients are at immediate risk of dying or fulfill criteria for brain death, provide maximal therapy, at least until contact with family is established and/or direct access to patient’s living will. |
First family communication | Establish a relationship of trust and try to speak with close family members in person rather than on the phone if possible. Inquire whether a documented living will exists. Provide objective information. Avoid choice of words or implicit communication elements that suggest a likely clinical outcome (it is reasonable however to say that the disease is “severe” or “potentially life-threatening” if that is the case). Assess the family’s overall understanding of the situation, explain the disease, leave room for questions. |
Score prognostication | Calculate the likelihood of unfavorable outcome using a recently validated prediction score. Be aware of the shortcomings of current prognostication tools, especially the lack of incorporation of worsening or improvement of the patient over time. Never make a definitive recommendation/decision solely based on the score results. |
Second family communication | Use the score information to give the family a sense how patients with a similar disease severity have done in the past. Avoid confronting patients/families with numbers. In rare cases, if the educational level of family allows it, explain biases and shortcomings of our current prediction tools. Explain that aggressive therapy may make a relative outcome difference even in situations where moderate to severe disability is very likely. |