Skip to main content
. Author manuscript; available in PMC: 2018 Dec 1.
Published in final edited form as: Ann Surg. 2017 Dec;266(6):952–961. doi: 10.1097/SLA.0000000000002286

Table 4.

Summary of Findings for Beta-Blockers after TBI

QUALITY ASSESSMENT SUMMARY OF FINDINGS
Participants
(Studies)		 Study
Design		 Inconsistency Indirectness Imprecision Publication
Bias		 Overall Quality
of Evidence		 Study Event Rates (%) Relative Effect Anticipated Absolute
Effects
With β-Blocker
after TBI		 No β-Blocker
after TBI		 Risk Difference using
β-Blocker after TBI
CRITICAL OUTCOME: In-hospital Mortality
8,245 (9 studies) Cohorts Very Serious1 Very Serious2 Very Serious3 Serious4 ++, low5 338/2005
(16.9%)6 		1103/6240
(17.7%)6 		0.39 (0.27 to 0.56) 99 fewer per 1000
(from 69 to 122 fewer)
CRITICAL OUTCOME: Function
0 (0 studies) N/A Unable to Assess7 Unable to Assess7 Unable to Assess7 Unable to Assess7 +, very low N/A N/A N/A N/A
CRITICAL OUTCOME: Quality of Life
0 (0 studies) N/A Unable to Assess7 Unable to Assess7 Unable to Assess7 Unable to Assess7 +, very low N/A N/A N/A N/A
IMPORTANT OUTCOME: Cardiac Morbidity by Biomarker/Arrhythmia
114 (1 study) Randomized Clinical Trial Unable to Assess7 Unable to Assess7 Unable to Assess7 Unable to Assess7 +, very low 2/27 (CK-MB) 6/46 (SVT) 9/30(CK-MB) 28/49 (SVT) N/A N/A
IMPORTANT OUTCOME: Hypotension
152 (2 studies)8 Randomized Clinical Trial Unable to Assess3 Unable to Assess3 Unable to Assess3 Unable to Assess3 +, very low 5/568 2/588 N/A N/A
IMPORTANT OUTCOME: Bradycardia
152 (2 studies)8 Randomized Clinical Trial Unable to Assess3 Unable to Assess3 Unable to Assess3 Unable to Assess3 +, very low 6/568 6/568 N/A N/A
IMPORTANT OUTCOME: Congestive Heart Failure
114 (1 study) Randomized Clinical Trial Unable to Assess3 Unable to Assess3 Unable to Assess3 Unable to Assess3 +, very low 0/56 0/58 N/A N/A
IMPORTANT OUTCOME: Bronchospasm
114 (1 study) Randomized Clinical Trial Unable to Assess3 Unable to Assess3 Unable to Assess3 Unable to Assess3 +, very low 1/56 0/58 N/A N/A
1

Inconsistency is very serious due to wide and unassessed baseline risk factors such as pre-injury cardiopulmonary comorbidities and pre-injury β-blocker use; substantial heterogeneity (I2 = 65%) indicating serious statistical inconsistency

2

Indirectness is very serious due to differences in population (e.g. TBI severity, non-TBI severity, age), differences across intervention (e.g., type dose, length, target of β-blocker), and differences across comparator (i.e., reasons for control or non-exposure)

3

Imprecision is also very seriously compromised with inability of the pooled sample to achieve optimal information size, as using a Type I error of 5%, power of 80%, over 35,000 subjects per arm would be required to enroll in a clinical trial

4

Known publication bias, as reviewed not but included RCT for this outcome states, “in-hospital deaths will be fully reported elsewhere”, but this critical outcome is not found elsewhere in the literature, despite being unable to quantify publication bias by funnel plot

5

Upgraded quality of evidence from very low to low quality given the consistent large magnitude of effect, 61% lower odds of mortality or 2.6 lower odds of mortality)

6

Number of deaths were derived from population rates but not directly reported for 1 of the pooled cohort studies

7

No published studies or No studies available for comparison

8

We only report the Randomized Clinical Trial; the other study is a prospective cohort that reports group statistical characteristics but not patient level data for these outcomes