Last updated Aug 6, 2025
A 66-year-old male presents to the ED after a trip-and-fall with a 2 cm abrasion to his forehead. He admits hitting his head but denies LOC and isn’t on any anticoagulants or antiplatelets. After your exam, you place an order for a head CT… but it’ll be a 3 hour wait. He insists he feels great, jumps up to do a few squats, and asks to be discharged. You think to yourself: "What are the chances this guy has a bleed?"
The Paper
Dubucs X, Gingras V, Boucher V, et al. Risk factors for traumatic intracranial hemorrhage in older adults sustaining a head injury in ground-level falls: a systematic review and meta-analysis. Ann Emerg Med. 2025;xx(xx):1–11. doi:10.1016/j.annemergmed.2025.05.021. PMID: 40699169
Emergency medicine physicians increasingly encounter older adults presenting after ground-level falls. While most CTs in these cases are negative, a small subset have traumatic intracranial hemorrhage (ICH)—often without overt signs. The challenge? No validated rule exists for this population, and imaging overuse can strain ED resources. This systematic review and meta-analysis aimed to identify clinical predictors of traumatic ICH in patients ≥65 years old with GCS ≥13 following a ground level fall.
Methods
Clinical Question
What clinical factors are associated with traumatic intracranial hemorrhage (ICH) in older adults (≥65 years) presenting to the emergency department with head trauma following a ground-level fall?
Study Design
- Design: Systematic review + meta-analysis
- Registration: PROSPERO (CRD42023478239)
- Databases Searched: Medline (Ovid), Embase, Cochrane, CINAHL, Web of Science (updated to Dec 9, 2024)
- Inclusion Criteria:
- Age ≥65
- Ground-level fall with head trauma (including facial trauma)
- GCS ≥13
- Underwent head CT or follow-up to rule out ICH
- Exclusion: High-energy trauma (e.g., MVC, stairs, bike)
- Primary Outcome: Presence of traumatic ICH
- Secondary Outcomes: Neurosurgical intervention, hospital admission, mortality
Analyzing Risk Factors & Predictors of ICH
Unadjusted Odds Ratios (ORs):
The paper first looked at the association between each predictor and the presence of traumatic intracranial hemorrhage (ICH). This answers: “Among patients with this sign or symptom (e.g., vomiting), how many ended up having a bleed compared to those who didn’t?” It tells you how much more likely (or less likely) something is to happen in one group compared to another. It’s a measure of association, not causation.
These unadjusted results are easier to calculate, but they don’t control for confounding — for example, someone with vomiting might also have had a skull fracture, which could be the real reason they bled. They used a random-effects model to pool data from different studies while accounting for differences in sample sizes and study quality.
Adjusted Odds Ratios (AORs):
Next, they looked at a smaller group of studies that reported adjusted results — this means those studies used statistical methods (like logistic regression) to control for other factors that could be influencing the outcome.
Example: A study might adjust for age, sex, GCS score, or whether the patient was on anticoagulation — and then ask: “Even after controlling for all of that, is loss of consciousness still independently associated with ICH?”
Results
Study Characteristics
- 17 observational studies
- 22,520 patients
- ICH prevalence: 6.8%
- Secondary outcomes
- Urgent neurosurgical intervention: 8.0% of patients with ICH
- Hospital admission: 36.8% (only in 7 studies
- In-hospital mortality: 1.5% (only in 3 studies)
- High-quality studies: 6 of 17 (≥7 NOS score)
Significant Risk Factors for ICH
Factors significant in both adjusted and unadjusted analyses:
Risk Factor | AOR [95% CI] |
Focal neurologic signs | 4.4 [3.0–6.5] |
External sign of head trauma | 2.7 [2.1–3.5] |
Loss of consciousness | 1.6 [1.2–2.1] |
Male sex | 1.4 [1.2–1.6] |
Factors not predictive in adjusted or unadjusted analysis
Risk Factor | AOR [95% CI] | OR |
Vomiting | 1.2 [0.7–2.1] | 2.7 [2.1–3.5] |
Headache | 1.5 [0.9–2.4] | 2.1 [1.6–2.9] |
Reduced GCS from baseline | 1.0 [0.9–1.1] | 4.0 [3.4–4.7] |
Single antiplatelet | 1.0 [0.6–1.6] | N/A |
Preinjury anticoagulants
DOACs, VKAs (Warfarin) | N/A | 0.8-1.0 [0.6-1.4] |
Why were vomiting, headache, and reduced GCS from baseline predictive in unadjusted analysis but not in adjusted analysis? Explanation below.
Term | Meaning | Example | Why It Matters |
Unadjusted Odds Ratio (OR) | The relationship between a predictor (e.g., vomiting) and an outcome (e.g., traumatic ICH), without accounting for other variables. | "Patients with vomiting had 2.7x higher odds of ICH." | Can be misleading if other factors (like age, anticoagulants, or GCS) are influencing both vomiting and the risk of ICH. |
Adjusted Odds Ratio (AOR) | The association between a predictor and an outcome after controlling for confounding variables using statistical models (usually logistic regression). | "After adjusting for age, sex, and baseline GCS, vomiting was no longer significantly associated with ICH (AOR 1.2, 95% CI 0.7–2.1)." | Gives a more accurate and isolated estimate of the true effect of a variable. Helps rule out confounding. |
Additional unadjusted-only significant risk factors
Not all clinical predictors had adjusted odds ratios available because only a subset of studies performed multivariable analyses. For many risk factors, there were too few studies reporting adjusted data, or each study adjusted for different variables, making it impossible to pool consistent AORs across all predictors.
Risk Factor | OR [95% CI] |
Open/depressed skull fracture | 10.9 [6.4–18.7] |
Signs of basal skull fracture | 4.7 [3.4–6.5] |
Seizure | 3.2 [1.5–7.0] |
Amnesia | 2.4 [2.0–3.0] |
Chronic kidney disease | 1.4 [1.0–1.9] |
Dual antiplatelet therapy | 2.3 [1.5–3.5] |
Author Conclusions
“This study identified risk factors for traumatic ICH that can be recognized in older patients presenting in the emergency department for a ground-level fall-related head injury. Based on these ndings, there is a need for future prospective studies to evaluate potentially avoidable head computed tomography scans in this population.”
Appraisal Checklist
RAMMBO
BEEM Quality Checklist (Systematic review)
Strengths and Weaknesses
Strengths | Weaknesses |
Large pooled sample size (22,520 patients) across 17 studies | High heterogeneity across many predictors (I² > 50%) |
Focused on an older adult population often underrepresented in decision making rules | Inconsistent definitions of clinical predictors across studies |
Used CT-confirmed ICH as a gold standard outcome | Only 7 of 17 studies provided adjusted odds ratios (AORs) |
Identified key independent predictors of ICH (focal deficits, LOC, etc.) | Adjustment models varied — hard to standardize across studies |
Followed PRISMA guidelines and registered on PROSPERO | Some included patients had no head trauma (just falls), risking dilution bias |
Addressed antithrombotic use | Could not assess predictor performance |
Bottom Line
This systematic review identifies potential clinical predictors of traumatic ICH in older adults after ground-level falls — notably focal neurologic deficits, external signs of head trauma, loss of consciousness, and male sex. However, due to heterogeneity, limited adjusted data, and observational nature of the data, these findings are not yet ready for clinical decision-making. Prospective studies are needed to develop and validate a reliable risk calculator or decision tool to safely reduce head CTs in this growing patient population.