A Systematic Review and Meta-Analysis on the Therapeutic Efficacy of Heparin and Low Molecular Weight Heparins in Animal Studies of Traumatic Brain Injury

Abstract

The identification of effective pharmacotherapies for traumatic brain injury (TBI) remains a major challenge. Treatment with heparin and its derivatives is associated with neuroprotective effects after experimental TBI; however, the optimal dosage and method of administration, modes of action, and effects on hemorrhage remain unclear. Therefore, this review aimed to systematically evaluate, analyze, and summarize the available literature on the use of heparin and low molecular weight heparins (LMWHs) as treatment options for experimental TBI. We searched two online databases (PubMed and ISI Web of Science) to identify relevant studies. Data pertaining to TBI paradigm, animal subjects, drug administration, and all pathological and behavior outcomes were extracted. Eleven studies met our pre-specified inclusion criteria, and for outcomes with sufficient numbers, data from seven publications were analyzed in a weighted mean difference meta-analysis using a random-effects model. Study quality and risk of bias were also determined. Meta-analysis revealed that heparin and its derivatives decreased brain edema, leukocyte rolling, and vascular permeability, and improved neurological function. Further, treatment did not aggravate hemorrhage. These findings must be interpreted with caution, however, because they were determined from a limited number of studies with substantial heterogeneity. Also, overall study quality was low based on absences of data reporting, and potential publication bias was identified. Importantly, we found that there are insufficient data to evaluate the variables we had hoped to investigate. The beneficial effects of heparin and LMWHs, however, suggest that further pre-clinical studies are warranted.

Introduction

Traumatic brain injury (TBI) is one of the leading causes of disability and death among the active population aged less than 45 years. It is caused by an external mechanical force to the head, and it is associated with incidents such as falls, motor accidents, bullet injury, industrial accidents, or sports-related injuries. With the increasing global incidence of TBI, it has been estimated that 69 million persons worldwide will have a TBI each year.¹

A TBI triggers a cascade of events that are classified as primary and secondary injury mechanisms. Primary injury, which includes axonal shearing, tissue destruction, and hemorrhage, results from the initial impact at the time of trauma. Secondary injury occurs within minutes to days and is characterized by a number of molecular, metabolic, and cellular processes, including disruption of the blood–brain barrier, brain edema, neuroinflammation, ischemia, apoptosis, oxidative stress, and lipid degradation, that may eventually result in additional brain dysfunction and chronic neurodegeneration.^2
–4

There is currently no effective drug treatment available for the ongoing brain damage associated with TBI despite many promising pre-clinical studies that have identified potential therapeutic agents over the years.⁵ Moreover, drug development for TBI continues to face significant challenges because of the presentation of very diverse and varying pathophysiologies among patients with TBI.⁶

One group of drugs that is of interest as a potential pharmacotherapy for TBI is heparins (including unfractionated heparin [UFH] and modified heparin 2,3-O-desulphated heparin, ODSH) and low molecular weight heparins (LMWHs). Heparins and LMWHs are well known for their use as anticoagulants. These glycosaminoglycans have a chemical structure consisting of recurring disaccharide units of uronic acid and d-glucosamine residues linked by 1,4 glycosidic bonds.⁷ LMWHs are produced through either chemical or enzymatic fractionation of UFH, and while the average molecular weight of heparin is 15 kDa, LMWHs lie in the range of 4–6 kDa.⁸

The highly negatively charged UFH has many unwanted characteristics, such as drug-induced adverse effects and an unpredictable dose-response after its administration, because it interacts with a wide range of proteins.^8,9 Conversely, LMWHs have fewer negative charges and exhibit low affinity for plasma proteins and, therefore, have a better safety profile, more predictable pharmacokinetic properties, and improved efficacy, with reduced risk of bleeding compared with heparin.¹⁰ Hence, LMWHs are more preferred for clinical use.

Heparins and enoxaparin, one of the most frequently used LMWHs, have demonstrated promising neuroprotective effects in pre-clinical TBI models, including reduced cerebral edema and lesion size and improved motor and cognitive recovery.¹⁰ Overall, the mechanisms underlying these pharmacological effects remain unknown; however, they are distinct from heparin and LMWHs' anticoagulant actions.

Importantly, despite the apparent benefits of heparin and LMWHs in TBI, these drugs are not used clinically, at least in the first 24 h after trauma, because of the potential risk of bleeding.¹¹ In the pre-clinical TBI literature, however, the incidence of hemorrhage after treatment with heparins/LMWHs remains unclear. It is also unclear whether UFH, modified heparins, or LMWHs have different protective actions, and what dosage and method of administration is optimal. As well, the overall efficacy of heparins/LMWHs to reduce specific secondary injury pathologies remains to be determined.

Therefore, this systematic review and meta-analysis aims to critically evaluate, summarize, and analyze data from the literature describing the use of heparins and LMWHs as treatment in animal models of TBI. In doing so, we will address the following questions: (1) do heparins and LMWHs (enoxaparin and dalteparin) have protective effects after TBI in animal models?; (2) what dose and route of administration produces maximal beneficial effects?; (3) what is the relative efficacy of heparins versus LMWHs?; and (4) does treatment increase brain hemorrhage after TBI?

Methods

A comprehensive systematic search of both PubMed and Web of Science was conducted based on the current guidelines of Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA)¹² (Supplementary Data S1). The protocol was registered with the international prospective register of systematic reviews (PROSPERO; CRD42020205574).

Sources of information and search strategy

Two electronic databases (Medline via PubMed and Web of Science) were searched since inception to August 2021 for relevant articles. Additional studies were sourced from the reference list of included articles.

Using Systematic Review Centre for Laboratory Animal Experimentation's (SYRCLE) guidelines,¹³ a detailed search strategy was developed in collaboration with a research librarian. Comprehensive synonyms and alternate words/terms were generated (Supplementary Data S2) for each of the main keywords (heparin/LMWHs, traumatic brain injury, and animal model) to search titles and abstracts in the databases mentioned above. More alternate terms were added from PubMed MeSH terms as searches were performed and from the references of the relevant articles that were included in the study.

The advanced search tools of PubMed and Web of Science were employed to search for both titles and abstracts. The Boolean operators “OR, AND, NOT, or AND NOT” were used to combine or exclude keywords in the searches for this review. Once the search was completed, duplicate references were removed using the EndNote citation management software (EndNote version X9, Thomson Reuters, New York). Detailed information on the search strategy can be found in Supplementary Data S3.

Inclusion criteria and exclusion criteria

We included peer-reviewed primary research articles describing the use of heparin (including modified heparins) and/or LMWHs as potential treatment for TBI in any animal model of TBI, including TBI transgenic models. We also included randomized controlled studies with control and experimental groups that clearly stated the number of animals in each group. Studies that administered heparin and/or LMWHs irrespective of the dose, route, frequency, duration, and time after brain injury were included.

We included studies that assessed functional and/or neuropathological outcomes of TBI, such as: motor, behavioural, or cognitive function, lesion volume, blood–brain barrier permeability, brain edema, neuroinflammation, hemorrhage, and the recruitment of immune cells from the blood. Finally, to be included, studies were required to report data as group means and variance (standard error of the mean or standard deviation). (No studies were excluded for violating this criterion.)

We excluded review articles and studies published in languages other than English. Studies involving other drugs as co-treatment were also excluded.

Study selection and data extraction

Two authors (MA and XYL) independently screened the titles and/or abstracts to identify publications that met the eligibility criteria, and differences were resolved in cases of no consensus by discussion with a third reviewer (NB).

The studies were exported to Rayyan QCRI (Qatar Computing Research Institute) software after the removal of duplicates.

All articles approved after the initial title and abstract screening underwent a further full-text screening step by two independent reviewers (MA and XYL), and disagreements between the independent reviewers were resolved by a third reviewer (NB).

Data were extracted from the manuscripts by text or by extrapolating numerical values from graphs in eligible studies. We extracted data for study identification elements, including the title, name of authors, journal, and publication date of articles; components of study design including animal details, injury characteristics, time, route, dose, frequency and duration of drug administration; and TBI pathological and behavioral outcomes assessed and the method of assessment.

Data were extracted and treated as independent experiments when a single study reported more than one experiment, and, as recommended by the Cochrane collaboration, the groups were combined to generate a single pair-wise comparison against the control group.¹⁴ Authors were contacted directly for further clarification or more information in cases of missing data.

Risk of bias/quality assessment

The risk of bias (RoB) of the included studies was assessed with SYRCLE's RoB tool,¹⁵ and the quality assessment of all studies was evaluated according to a modified checklist from the Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES) with slight modification.^16,17 The checklist comprised the following 10 criteria: (1) peer-review publication; (2) random group allocation; (3) blinded induction of injury; (4) blinded assessment of outcome measures; (5) use of an anesthetic/analgesic agent without marked neuroprotective activity; (6) statement of compliance with animal welfare regulatory requirements; (7) statement of potential conflicts of interest; (8) use of appropriate animal model; (9) calculation of sample size; and (10) explanation of excluded animals. One point was given to indicate the presence of each criterion.

Data analysis

In this review, a meta-analysis and a narrative synthesis were both employed. The effect of heparins and LMWHs on brain edema, leukocyte rolling, microvascular permeability, hemorrhage/hemorrhagic contusion, neurological function, and body weight loss was estimated by calculating the standardized mean difference (SMD; mean difference between groups divided by the standard deviation of outcome effect) for each comparison and combining these values in a weighted mean difference meta-analysis using a random-effects model.¹⁸ This method was chosen because it permits aggregation of data from studies measuring the same outcome in different ways.^19,20 For instance, evaluating neurological function using different neurological assessment tools. Results are presented as overall SMD and its 95% confidence intervals (CIs).

The presence of heterogeneity was assessed using the I² statistic, the percentage of the variability in effect estimates that is because of heterogeneity rather than sampling error (chance).¹⁴ Higher values indicate a greater degree of heterogeneity.

A priori examination of which experimental factors contribute to outcome heterogeneity was performed primarily by subgroup analysis. Subgroup analysis for drug types and their doses was conducted only for the brain edema outcome where data were most abundant. In addition, funnel plots and Egger tests were conducted to investigate the presence of small sample effects that might result from problems such as publication bias.

For all UFH studies,^21,22 the dose given was recorded as U/kg, and so was converted to mg/kg according to standard dosing guidelines (100 U/kg = 1 mg/kg).²³ After conversion, doses of both enoxaparin and heparin were either categorized as low dose (≤1 mg/kg) or high dose (≥2 mg/kg).

In this review, the controls were untreated injured animals while the treatment groups received either heparin or enoxaparin at varied times and doses post-injury.

For the narrative synthesis, all 26 outcomes that were extracted from the included studies were summarized in tables based on, but not limited to, the drug dosage, drug type, and overall efficacy of the different drugs (heparin, ODSH, and enoxaparin), with their major findings discussed. Meanwhile, meta-analyses were conducted for outcomes where data were available from at least three publications. (See Supplementary Data S4 for a flowchart illustrating the relationship between the narrative and quantitative aspects of this review.)

All meta-analyses were performed using Review Manager (RevMan, Version 5.4, the Cochrane Collaboration, 2020) and/or StataSE software (version 16, StataCorp LP, Texas) with a significance level set at p < 0.05. The effect estimates of all the outcomes in the summarized table (Table 3) were also performed using RevMan. For data represented as forest plots, the central lines and lateral tips of the diamond signify pooled point estimates and the corresponding 95% CIs, respectively. The solid black vertical line indicates the line of no effect while the red vertical line is the overall effect size line.

Results

Study selection

Our search found 339 potentially relevant publications. Sixty-three were replicates because of database overlap and were removed. Of the remaining 276 publications subject to title and abstract screening, 12 were selected for full-text review. One of these was excluded for not meeting our pre-specified inclusion criteria. Eleven publications ^{21,22,24

–32} were included in the review (Fig. 1). From the included studies, meta-analysis was conducted on the data from seven publications^{21,22,24,25,28,29,32} for brain edema, neurological function, hemorrhage/hemorrhagic contusion, microvascular permeability, leukocyte rolling, and body weight (see Supplementary Data S5 for number of comparisons of each outcome).

FIG. 1.

Flowchart of study process. A total of 11 studies were included in this review after the literature search process, and meta-analysis was conducted on seven studies. *One study was excluded during full-text screening because enoxaparin was administered in combination with another drug.

Study characteristics

The characteristics of the included studies, published between 2000 and 2020, are shown in Table 1. Of the 11 studies, seven administered enoxaparin,^24

–30 two studies administered unfractionated heparin,^21,22 and one study each administered ODSH,³² and heparin oligosaccharides,³¹ respectively. All the animals studied were male. The CD-1 mouse strain was the most commonly used animal,^21,22,24,32 followed by Sprague-Dawley^25,28,29 and Wistar^30,31 rats. One study each used the C57BL/6²⁶ and BALB/c²⁷ mouse strains, respectively. Mice ranged in weight from 20 to 32 g and rats ranged from 100 to 450 g. Six studies reported using adult animals but did not specify an exact age,^{21,22,24,28,30,32} four studies reported using mouse models aged between 4 to 12 weeks,^25
–27,31 and one study failed to report the age of animals used.²⁹

Table 1.

Characteristics of Included Studies

First author, Year	Animal	Strain	Sex	Age (weeks)	Weight (g)	TBI model	Injury severity	Anesthetic	Treatment	Mode of administration	QA score
Dhir, 2018 ²⁵	Rat	Sprague-Dawley	Male	4	100 - 200	Weight drop	Mild	Isoflurane	Enoxaparin	SC injections	5
Hayashi, 2004 ³¹	Rat	Wistar	Male	8 - 10	NR	Cryo-cortical injury	NR	Pentobarbital	Heparin oligosaccharides	applied directly to wound cavity after injury	4
Kerr, 2020 ²⁶	Mouse	C57BL/6	Male	8 - 12	24 - 32	CCI	Severe	Ketamine/Xylazine	Enoxaparin	IV injections	6
Keskin, 2017 ²⁷	Mouse	BALB/c	Male	8 - 10	20 - 25	Cold induced brain injury	NR	Ketamine hydrochloride / Xylazine hydrochloride	Enoxaparin	IP injections	5
Li, 2015 ²⁴	Mouse	CD-1	Male	Adult^*	25 - 30	CCI	Severe	Ketamine/Xylazine / Acepromazine	Enoxaparin	SC injections	4
Nagata, 2017 ²²	Mouse	CD-1	Male	Adult^*	25 - 30	CCI	Severe	Ketamine/Xylazine / Acepromazine	Unfractionated heparin	Repeated SC injections	6
Nagata, 2016 ²¹	Mouse	CD-1	Male	Adult^*	25 - 30	CCI	Severe	Ketamine/Xylazine / Acepromazine	Unfractionated heparin	Repeated SC injections	4
Nagata, 2018 ³²	Mouse	CD-1	Male	Adult^*	25 - 30	CCI	Severe	Ketamine/Xylazine / Acepromazine	O-desulphated heparin	SC injections	5
Sen, 2011 ²⁸	Rat	Sprague-Dawley	Male	Adult^*	200 - 290	Weight drop injury	Severe	Ketamine	Enoxaparin	IV bolus injection+SC injections	4
Wahl, 2000 ²⁹	Rat	Sprague-Dawley	Male	NR	260 - 300	FPI	Moderate	Halothane	Enoxaparin	IV bolus injection+SC injections	4
Župan, 2011 ³⁰	Rat	Wistar	Male	Adult^*	350 -450	LFPI	Moderate	Isoflurane	Enoxaparin	SC injections	7

QA, quality assessment; NR, not reported, CCI, controlled cortical injury, IV, intravenous; IP, intraperitoneal; SC, subcutaneous; FPI, fluid percussion injury; LFPI, lateral fluid percussion injury, ^*Adult as specified by the authors of the different studies.

Controlled cortical impact was the most frequently used model of TBI,^{21,22,24,26,32} while two studies used the weight-drop impact model.^25,28 In most studies, the injury induced was severe,^{21,22,24,26,28,32} while two studies reported moderate injury^29,30 and one study reported a mild injury.²⁵ Two studies failed to report injury severity.^27,31

Drugs were either administered immediately after trauma induction or between 15 min and 2 h after injury as a single dose^26,27,31 or multiple doses.^{21,22,24, 28
–30,32} The subcutaneous (SC) injection was the most used mode of drug delivery as reported by six studies^{21,22,24,25,30,32}; two studies used both intravenous (IV) and SC injections.^28,29 Direct administration into the wound cavity,³¹ intravenous,²⁶ and intraperitoneal (IP)²⁷ injection were each reported in one study, respectively. Drugs, dosing, time to first administration, and frequency of administration are detailed in Table 2.

Table 2.

Details of Treatment Intervention from the Included Studies

Study	Drug	Time to first drug delivery post-TBI	Treatment dose	Treatment frequency
Dhir, 2018 ²⁵	Enoxaparin	Unclear	1 mg/kg	Unclear
Hayashi, 2004 ³¹	Heparin oligosaccharides	Immediately onto wound cavity	10 mg/ml	Single
Kerr, 2020 ²⁶	Enoxaparin	30 min	1 mg/kg	Single
Keskin, 2017 ²⁷	Enoxaparin	20 min	3 mg/kg, 10 mg/kg	Single
Li, 2015 ²⁴	Enoxaparin	2 h	1 mg/kg	Multiple
Nagata, 2017 ²²	Unfractionated heparin	2 h	75 U/kg, 225 U/kg	Multiple
Nagata, 2016 ²¹	Unfractionated heparin	2 h	75 U/kg, 225 U/kg	Multiple
Nagata, 2018 ³²	ODSH	2 h	25 mg/kg, 50 mg/kg	Multiple
Sen, 2011 ²⁸	Enoxaparin	15 min	0.5 + 1 mg/kg	Multiple
Wahl, 2000 ²⁹	Enoxaparin	1) 15 min; 2) 2 h	1) 0.5 + 2 mg/kg, 2) 0.5 + 0.5, 0.5 + 1, 0.5 + 2 mg/kg	Multiple
Župan, 2011 ³⁰	Enoxaparin	1 h	1 mg/kg	Multiple

TBI, traumatic brain injury; ODSH, 2,3-O-desulphated heparin.

As detailed in Table 3, of the 26 outcomes extracted, brain edema and neurological function were the outcomes most commonly assessed, being reported in five studies,^{21,22,24,29,32} followed by leukocyte rolling,^21,24,32 body weight loss,^21,22,24 microvascular permeability,^21,24,32 hemorrhage,^24,25,28 and apoptosis,^27,28,30 all with three studies each. Brain infarct volume^27,29 and leukocyte adhesion^21,22 were reported by two studies each, and the remaining outcomes including oxidative stress,³⁰ cognitive function,²⁹ learning,³² hemoglobin levels,²² motor function,³² inflammation,²⁸ glial scar formation,³¹ interleukin (IL)-1β,²⁶ pro-IL-1β,³⁰ caspase-1,²⁶ cyclooxygenase (COX) expression,³⁰ reactive gliosis,³⁰ lesion surfaces,²⁹ polymorphonuclear (PMN) sequestration,²² neurodegeneration,³⁰ as well as total antioxidant status and total oxidant status,²⁷ were reported by one study each.

Table 3.

Summary and Findings of All Outcomes from Included Studies

Outcomes assessed	Treatment	Vehicle	Animal species	Model and time points assessed	Method of assessment	Key laboratory findings post-injury	Brain region assessed	Effect size % [95% confidence interval]
Brain oedema ^{21,22,24,29,32}	Enoxaparin, UFH, ODSH	Saline	Mouse, rat	48 and 96 h post-injury (CCI, FPI)	Calculation of water content	Enoxaparin, high dose ODSH, low and high doses of UFH decreased brain edema at 48 h. Low dose UFH reduced edema at 96 h	Cortex, whole brain	0.95 [0.35, 1.55]
Leukocyte rolling ^21,24,32	Enoxaparin, UFH, ODSH	Saline	Mouse	48 h post-TBI (CCI)	Quantification of fluorescently labelled intravascular cells as LEU rolling	Enoxaparin, low and high dose UFH and ODSH reduced LEU rolling	NR	1.93 [1.42, 2.44]
Microvascular permeability ^21,24,32	Enoxaparin, UFH, ODSH	Saline	Mouse	48 h post-TBI (CCI)	Quantified the ratio of FITC-labeled albumin fluorescence	Enoxaparin, both doses of UFH and ODSH decreased albumin leakage	NR	2.23 [1.50, 2.96]
Oxidative stress ³⁰	Enoxaparin	NR	Rat	48 h post-TBI (LFPI)	Measured the levels of TBARS, SOD and GSH-Px activity, evaluated immunochemical detection of protein carbonyl groups	Reduced the levels of TBARS and hippocampal oxidized proteins. No effect on hippocampal SOD and GSH-Px activities	Hippocampus	0.11 [-0.38, 0.60]
Body weight loss ^21,22,24	Enoxaparin, UFH	Saline	Mouse	24, 48, 72 and 96 h post-TBI (CCI)	Measured body weight	Enoxaparin reduced body weight loss at 24 h and not at 48 h. Both doses of UFH had no differences on body weight	NR	0.25 [-0.34, 0.85]
Inflammation ²⁸	Enoxaparin	NR	Rat	48 h post-TBI (WDI)	Quantified H&E stained mild inflammatory infiltrates	Reduced inflammation	Frontoparietal cortex, hippocampus, corpus callosum	0.61 [-0.29, 1.51]
Glial scar formation ³¹	Heparin oligosaccharides	Sterile PBS	Rat	6 - 9 days post-TBI (WDI)	Quantified CS-A and GFAP immunoreactivity	Reduced the formation of glial scar in cortical regions closest to injury. No change in cortical region further away from injury.	Cerebral cortex	0.34 [-0.12, 0.81]
Hemorrhage ^24,25,28	Enoxaparin	Saline	Rat, mouse	24, 48 and 72 h post-TBI (WDI, CCI)	Histologically assessed using H&E staining, measured hemoglobin levels and hemorrhagic contusion.	No differences observed	Frontoparietal cortex, hippocampus, corpus callosum, dorsal surface of the injured hemisphere	0.95 [0.44, 1.46]
IL-1β ²⁶	Enoxaparin	NR	Mouse	24 h post-TBI (CCI)	Measured levels of IL-1β using Western blot	Reduced expression of IL-1β	Cortex	1.56 [0.42, 2.70]
Neurological function ^{21,22,24,29,32}	Enoxaparin, UFH, ODSH	Saline	Mouse, rat	24, 48, 72 and 96 h post-TBI, 1-4 weeks post-TBI (CCI, FPI)	A validated modified Neurological Severity Scale (NSS) or Garcia Neurological Test (GNT)	Low dose, but not high dose UFH and enoxaparin improved neurological function at all time points. Both low and high doses of ODSH improved outcome.	NA	-1.11 [-1.75, -0.46]
Brain infarct volume ^27,29	Enoxaparin	Saline	Mouse, rat	24 h and 7 days post-TBI (FPI, CIBI)	Histology assessment with H&E and Cresyl Violet staining	Low dose did not reduce infarct volume at 24 h post-TBI but reduced lesion volume at 7 days post-TBI. High dose reduced infarct volume at 24 h post-TBI but did not reduce lesion volume at 7 days after injury.	Throughout the brain	0.65 [0.20, 1.10]
Apoptosis ^27,28,30	Enoxaparin	Saline	Mouse, rat	24 and 48 h post-TBI (WDI, LFPI, CIBI)	Quantified TUNEL positive cells, H&E evaluation of neuronal cell death and assessing the expression of active caspase-3 using Western blot	Decreased apoptosis, however, high dose and not low dose decreased outcome in one study	Frontoparietal cortex, hippocampus, corpus callosum	0.62 [0.07, 1.16]
Caspase – 1 ²⁶	Enoxaparin	NR	Mouse	24 h post-TBI (LFPI)	Measured levels of caspase-1 using Western blot	Reduced the expression of inflammasome protein caspase-1	Cortex	1.12 [0.06, 2.18]
Total anti-oxidant status (TAS) ²⁷	Enoxaparin	Isotonic saline solution	Mouse	24 h post-TBI (CIBI)	Measured serum levels of TAS	Low dose had no effect, but high dose increased serum levels	Blood sample	-0.41 [-1.16, 0.34]
Total oxidant status (TOS) ²⁷	Enoxaparin	Isotonic saline solution	Mouse	24 h post-TBI (CIBI)	Measured serum levels of TOS	No significant difference observed	Blood sample	-0.60 [-1.36, 0.16]
PMN sequestration ²²	UFH	Saline	Mouse	96 h post-TBI (CCI)	Counted immunostained positive PMNs cells	Both low and high doses of UFH decreased PMN sequestration	Cortex	1.89 [1.02, 2.75]
Haemoglobin levels ²²	UFH	Saline	Mouse	96 h post-TBI (CCI)	Measured serum levels of hemoglobin	Low dose had no difference on serum levels. High dose lowered serum hemoglobin levels	Mean arterial blood pressure	0.25 [-0.47, 0.97]
Leukocyte adhesion ^21,32	UFH, ODSH	Saline	Mouse	48 h post-TBI (CCI)	Quantification of fluorescently labelled intravascular cells as LEU rolling.	No significant effect	NR	0.82 [0.33, 1.31]
Neurodegeneration ³⁰	Enoxaparin	NR	Rat	48 h post-TBI (LFPI)	Quantification of Fluoro-Jade-B stained hippocampal neurons	No significant decrease	Hippocampus	0.09 [-0.55, 0.72]
Motor function ³²	ODSH	Saline	Mouse	14-17 days post-TBI (CCI)	Measured swim velocity on Morris water maze	High dose improved swimming velocity on days 16 and 17	NA	-0.89 [-1.39, -0.38]
Lesion surfaces ²⁹	Enoxaparin	NR	Rat	7 days post-TBI (FPI)	H&E histological evaluation of coronal sections	Low dose reduced lesion volume. High dose had no effect	NR	0.69 [0.54, 0.84]
Pro IL-1β ³⁰	Enoxaparin	NR	Rat	48 h post-TBI (LFPI)	Measured pro IL-1β using Western blot	No change in hippocampal pro IL-1β expression	Hippocampus	0.00 [-1.24, 1.24]
COX-2 expression ³⁰	Enoxaparin	NR	Rat	48 h post-TBI (LFPI)	Measured COX-2 expression using western blot	Lowered COX-2 expression	Hippocampus	0.99 [-0.37, 2.35]
Reactive gliosis ³⁰	Enoxaparin	NR	Rat	48 h post-TBI (LFPI)	Evaluated the area of astrocytes' perikarya and GFAP staining intensity by analyzing immunostained GFAP cells	No effect on the area of astrocytes' perikarya. decreased GFAP staining intensity	Hippocampus	0.14 [-0.79, 1.06]
Cognitive function ²⁹	Enoxaparin^*	NR	Rat	2, 6, 13, 20, 27 h post-TBI (FPI)	Measured latency time on the complex maze at days 2, 6,13, 20 and 27.	Decreased latency time at days 2, 6 and 13 but had no difference on days 20 and 27	NA	0.26 [0.06, 0.45]
Learning ³²	ODSH	Saline	Mouse	14-18 post-TBI (CCI)	Measured latency to the platform (learning) on Morris water maze	Low dose but not high dose improved learning	NA	-2.13 [-3.28, -0.99]

UFH, unfractionated heparin; ODSH, 2,3-O-desulfated heparin; TBARS, thiobarbituric acid-reactive substances; SOD, Superoxide dismutase; GSH-Px, glutathione peroxidase; H&E, Haematoxylin and Eosin; PBS, phosphate-buffered saline; CS-A, chondroitin sulfate-A; GFAP, glial fibrillary acidic protein; PMN, Polymorphonuclear; IL-1, Interleukin-1; COX-2, Cyclooxygenase-2; CIBI, cold-induced brain injury; CCI, controlled cortical injury; WDI, weight drop injury; FPI, fluid percussion injury; LFPI, lateral fluid percussion injury.

(Only low dose was assessed)

The wet/dry weight ratio method was used to assess brain edema, while the validated, modified Garcia Neurological Test and the Neurological Severity Score were the most frequent methods used to assess neurological function. The summary of the method of assessment of the other outcome measures used is included in Table 3.

RoB and quality assessment

The RoB analysis, as shown in Figure 2, highlights the possible sources of bias in the included studies. Baseline characteristics were generally well reported (91% of studies) with only one study failing to report these key details.²⁹ Most other facets of study reporting, however, suggest the data are at significant risk of bias. No studies described random housing, random outcome assessment, or blinding of assessors. Of all the included studies, 70% had evidence of selective outcome reporting^{21,22,24
–26,28,31,32} because the sample size planned or analyzed was not reported. Only three studies did not have selective outcome reporting because they reported all outcomes and sample sizes evaluated and analyzed.^27,29,30

FIG. 2.

Risk of bias (RoB) graph of included studies evaluated by SYRCLE's RoB tool. Baseline characteristics was the most reported in the included studies (91%) while more than 70% of studies demonstrated selective outcome reporting and no study described random housing, random outcome assessment, and blinding of assessors (detection bias). Each risk of bias item is presented as percentages.

In three studies, attrition bias was not evident^27,29,30 but was present in seven studies (more than 60%) where the number of sample size did not equate to the final number that was analyzed and an explanation was not provided for the missing data.^{21,22,24
–26,28,32} One study made no mention of the sample size used or analyzed.³¹ Nine studies reported randomization into experimental groups, two studies did not report whether randomization was performed, and only one study (about 9%) actually described sequence generation achieved by a double coin toss.²²

Of the included studies, only one reported allocation concealment, achieved by using an investigator who was not involved in treatment delivery or analysis.²⁶ The same study was the only one that reported that blinding was achieved by masking treatment bottles with opaque tape.²⁶ No studies indicated whether outcome assessment was blinded.

While perhaps less critical than the factors discussed above, another potential source of bias identified was potential conflicts of interest from sources of funding or materials; these are important to report because the financial interests of authors could lead to biased interpretation of the data. A total of six studies did not report on whether a potential conflict of interest from sources of funding existed,^{21,22,24,27
–29} although two of the six did state that they had no potential conflict of interest but without specifically referring to their source of funding.^22,27 The remaining five studies did declare their sources of funding: two of the five stated that this did not pose a conflict of interest,^25,30 one did not indicate whether the funding created a potential conflict of interest,³¹ and two did appear to have a potential conflict of interest because of their funding and material source.^26,32

A modified CAMARADES checklist gave a median quality score of six with scores ranging from four to seven out of 10 (Supplementary Data S6). In this assessment, by definition, all included studies came from peer-reviewed articles and used the appropriate animal models. Nine of the 11 included publications (81%) reported compliance with animal welfare regulations, and six studies (55%) reported whether a potential conflict of interest existed.^{22,25
–27,30,32} No study, however, reported blinded induction of injury by the investigator or blinded assessment of outcome measures.

Random group allocation was reported in nine of the 11 included publications.^{21,22,24

–28,30,32} Most studies used ketamine alone or in combination with other agents such as xylazine and acepromazine for anesthesia, only three used gaseous anesthetics,^25,29,30 and one used pentobarbital.³¹ Two publications described the sample size calculation,^22,26 and two others provided an explanation for excluded animals.^29,30 One study did not have to provide an explanation because there were no excluded animals.²⁷

Meta-analysis

Overall, there was significant heterogeneity between studies investigating drug efficacy on both brain oedema and neurologic function outcomes, which subgroup analyses accounted for.

Overall effect of treatment on brain edema

Five of the publications used in the meta-analysis assessed the effect of treatment on brain edema.^{21,22,24,29,32} The data on therapeutic efficacy of heparins/LMWHs and its derivatives on brain edema was contained within the 27 experimental comparisons (Supplementary Data 5). We limited our analysis, however, to data from the ipsilateral cortex, because it best answers our research questions. The efficacy of treatment on brain edema was evaluated based on 17 comparisons involving a total of 407 animals. Taken together, drug treatment reduced brain edema with an overall estimated effect of -1.69 (95% CI: -2.55, -0.82, p < 0.001, Fig. 3) but with substantial heterogeneity, I² of 90.46% among studies (Fig. 3).

FIG. 3.

Effect of all treatments on brain edema after trauma. The forest plots represent the pooled analysis of data on the use of heparin and enoxaparin on brain edema. The overall effect estimates favor treatment compared with untreated control (-1.69 (95% cofidence interval [CI]: -2.55, -0.82, p < 0.001). Heterogeneity was significant (I² = 90.46%), however. The square symbols and their error bars (95% CI) are the individual studies contributing to the pooled estimates. The central lines and lateral tips of the diamonds indicate the pooled point estimates and the corresponding 95% CIs, respectively. The solid black vertical line indicates the line of no effect while the gray vertical line is the overall effect size line. This annotation is used in all subsequent forest plots. SD, standard deviation.

Effect of heparin and enoxaparin treatment on brain edema

Comparing the two studies that examined the effect of UFH^21,22 with the two that examined enoxaparin^24,29 on brain edema revealed similar point estimates for both (test of group difference, brain edema, p = 0.96, Fig. 4) but the degree of heterogeneity differed markedly (enoxaparin I² = 94%, UFH I² = 34.38%).

FIG. 4.

No difference in drug subgroup analysis on brain edema. For enoxaparin and unfractionated heparin (UFH), the overall effect favored treatment compared with the untreated control for brain edema (-1.67 (95% confidence interval [CI]: -2.78, -0.56; p < 0.001). There was no difference, however, between the effect estimates for the two drugs although heterogeneity differed substantially (I² enoxaparin = 93%, I² UFH = 34.38%). SD, standard deviation.

Effect of different doses of heparin and enoxaparin on brain edema

Segregating the data for enoxaparin and UFH into low and high dose regimes on brain edema suggested a trend toward greater effect for low doses (low dose: -2.40 (95% CI: -3.01, -1.80; high dose: -1.40 (95% CI: -2.42, -0.38); however, no statistical difference between the doses was detected (test of group difference, Q_b, 2.76, p = 0.10; Fig. 5). While heterogeneity was not significant in the low dose group (I² = 6.55%), it was moderate in the high dose group (I² = 60.81%).

FIG. 5.

Effect of low and high doses on brain edema after trauma. The drug doses were grouped into either low or high dose regardless of the mode of administration. Overall, the pooled estimate of both doses favored treatment compared with control with a substantial heterogeneity (I² = 61.44%). There was no significant difference between the effect of low and high dose on brain edema (p = 0.10). Heterogeneity was moderate in the high dose group (I² = 60.81%) and not significant in the low dose group (I² = 6.55%). SD, standard deviation.

The study from Wahl and associates, 2000²⁹ was excluded from this analysis; we could not categorically group the dose administered as either a low or high dose because the authors used a combination of drug doses at various times. For example, they started with an IV administration of a low dose a few minutes or hours post-injury and afterward with a SC injection of a high dose. Overall, the pooled estimate of both low and high drug doses favored treatment compared with the untreated controls, and the heterogeneity (I² = 61.44%) was reduced slightly but remained substantial.

Effect of treatment on neurological function

Four studies assessed neurological function at 24 h and 48 h post-TBI,^21,22,24,32 with one study also including 72 h and 96 h time points,²² while a fifth study assessed this outcome at weekly intervals for four weeks post-TBI,²⁹ We selected the 48-h time point data for meta-analysis because it was the longest post-injury interval with the most data. The four studies assessing neurological function at 48 hours post-TBI^21,22,24,32 were evaluated after treatment with heparin and its derivatives based on seven comparisons involving 129 animals. Overall, the pooled adjusted estimated effect of 1.51 (95% CI: 0.31, 2.72, p = 0.01, Fig. 6) favored treatment. There was significant heterogeneity (I² = 88.58%), however.

FIG. 6.

Effect of all treatment on neurological function after trauma. The forest plots represent the pooled analysis of data on the use of heparin and its derivatives on neurological function. The overall effect estimates favor treatment compared with control (1.51 (95% confidence interval [CI]: 0.31, 2.72, p = 0.01). Drug treatment improved neurological function after trauma. Heterogeneity was significant (I² = 88.58%), however. The central lines and lateral tips of the diamonds indicate the pooled point estimates and the corresponding 95% CIs, respectively. The solid black vertical line indicates the line of no effect while the gray vertical line is the overall effect size line. SD, standard deviation.

Effect of treatment on hemorrhage/hemorrhagic contusion

Only three studies examined the effect of treatment on bleeding using two different methodologies—detection of red cells/clot on hematoxylin and eosin stain^25,28 and visualization of hemorrhage on the surface of the brain.²⁴ Based on four comparisons involving 70 animals, together, these studies with a pooled estimate of 0.95 (95% CI: 0.44, 1.46, p = 0.0003, Fig. 7A) found that treatment did not increase bleeding. Heterogeneity was not significant (I² = 0%).

FIG. 7.

Effect of all treatments on hemorrhage, inflammation, and vascular permeability. Leukocyte rolling and microvascular permeability were decreased after treatment with unfractionated heparin, enoxaparin, and 2,3-O-desulphated heparin. Varying effects of both enoxaparin and unfractionated heparin were observed on body weight loss. Enoxaparin treatment did not increase hemorrhage and treatment. SD, standard deviation; CI, confidence interval.

Effect of treatment on microvascular permeability and inflammation

Three studies,^21,24,32 all in mice involving five comparisons in 95 animals, quantified FITC-labelled albumin fluorescence at 48 h post-TBI. The authors found that enoxaparin, UFH, and ODSH, respectively, improved blood–brain barrier permeability (2.23 (95% CI: 1.50, 2.96, p = 0.00001; I² = 44%, Fig. 7B). The same three studies (five comparisons involving 95 animals) also reported decreased leukocyte rolling with an overall pooled estimate of 1.93 (95% CI: 1.42, 2.44, p = 0.00001, Fig. 7C) and an insignificant heterogeneity (I² = 0%).

Effect of treatment on body weight

When body weight was examined in the same three studies as above,^21,24,32 albeit with 14 comparisons involving 260 animals, treatment with enoxaparin was associated with reduced body weight loss while treatment with either UFH or ODSH was not. Overall, there was no effect of treatment on body weight loss (0.25 (95% CI: -0.34, 0.85), p = 0.40, I² = 79%, Fig. 7D).

Other variables we had hoped to investigate, including the optimal time at which the drug produced the most effect and the best route of administration, could not be evaluated because of insufficient data.

Potential publication bias

Publication bias was assessed using funnel plot on the overall therapeutic effect of treatment on brain edema and neurological function after trauma. Four studies were evaluated, and we found that there was funnel plot asymmetry for both outcomes, which was confirmed with Egger regression test (brain edema, z = -4.86, p < 0.001; neurological function, z = 3.25, p = 0.0012). Contour-enhanced funnel plot explored funnel plot assymetry at 1%, 5% and 10% (Fig. 8A, 8B) for each outcome, suggesting the possibility of publication bias. This finding, however, should be interpreted with caution because of the low number of studies evaluated.

FIG. 8.

Funnel plot asymmetry in the studies evaluated for brain edema and neurological function after trauma. Contour-enhanced funnel plot assessed at 1%, 5%, and 10% show marked asymmetry in reporting of brain edema (A) and neurological function (B) after treatment of traumatic brain injury with heparins.

Discussion

In this review, we have evaluated the pre-clinical data supporting the potential use of heparins/LMWHs as treatments for TBI. To our knowledge, this review is the first to systematically collect and analyze all available evidence in the literature regarding the therapeutic efficacy of heparins and LMWHs in animal models of TBI.

There is a significant body of work examining this potential. We were surprised, however, at the paucity of data suitable for inclusion in a meta-analysis. Indeed, of the 276 publications that appeared relevant on screening, only 11 met our overall eligibility criteria and only seven contained sufficient data for inclusion in meta-analysis. Moreover, within the studies, the injury models and outcome assessments were variable. This is not surprising, because the presentation of TBI with respect to age, sex, site, and nature of injury, severity, and pathophysiology is complex and diverse.³³

The paucity of data is in part because of the strict criteria that are applied to facilitate valid meta-comparisons across sometimes disparate datasets. For example, many articles were the only reports for a particular outcome and could not sensibly be grouped with others for analysis. Also, we were only able to use data from experiments that provided controlled studies (in this case, untreated injured animals) where cohort sizes and outcome data and its variability were provided or could be reliably imputed.

This speaks to the need for coordination across the field, particularly with respect to core standardized analyses^34,37 and better application of current reporting standards such as ARRIVE (Animal Research: Reporting of In Vivo Experiments)^35,36 and the pre-clinical TBI Common Data Elements (CDEs) developed by the National Institute of Neurological Disorders and Stroke, describing experimental parameters and outcome measures.^38,39 These guidelines facilitate reproducible reporting of key variables, which enables the comparison of results between studies.⁴⁰

As mentioned above, we were limited in our ability to utilize the analytical power of systematic review^41,42 by the breadth of different analyses performed and the overall paucity of suitable numerical data; this is an important observation for the field and common in pre-clinical analyses^43,44 where exploration of pathobiology is usually combined with evaluation of potential therapeutic efficacy. Nevertheless, these limitations and the meta-analysis itself provide some important additional insights.

Because all the animals used were male (Table 1), we were unable to explore the role of sex on the efficacy of heparins/LMWHs for TBI. This indicates that researchers are ignoring this important facet of human TBI where sex clearly affects outcome,^45,46 the female hormones estrogen and progesterone are known to be neuroprotective,^47
–49 and published guidelines recommend use of both sexes in studies of TBI.⁵⁰

We did find studies that analyzed multiple outcomes using multiple model systems as detailed in Tables 1–3. The main themes were that heparin and its derivatives modify lesion size,^27,29 neurological function,^21,22,29,32 brain edema,^21,22,24,32 blood–brain barrier dysfunction,^21,24,32 neuroinflammation,^{21,22,24,26,28,30
–32} a variety of mechanisms of cell death,^26
–28,30 the pathobiology of bleeding in the face of heparin-based drug treatments,^22,24,25,28 and a range of other markers of injury such as oxidative stress.³⁰ Here, we first summarize the main findings from the individual articles and report the meta-analysis of components of these data.

Increased lesion volume after TBI, whether because of ischemic or other secondary changes, is associated with unfavorable outcomes.^51,52 Of the studies included in this review, only two studying the effects of enoxaparin explicitly examined lesion volume. One reported that high dose enoxaparin (10 mg/kg IP) decreased lesion volume,²⁷ but the other found no effect at high doses (0.5 mg/kg IV + 4 × 2 mg/kg SC) and detected a decrease at low doses (0.5 mg/kg IV + 4 × 1 mg/kg SC).²⁹ These differences might be a consequence of the different models and species used—cold-induced TBI in mice for the former²⁷ and fluid percussion in rats for the later.²⁹ Unfortunately, there were insufficient data to explore this by meta-analysis.

Poor neurological function, as a consequence of both the initial lesion and subsequent damaging changes, is perhaps the most critical consequence of TBI.^53,54 The studies reported here that examined enoxaparin, UFH, and ODSH all improved neurological function and enhanced recovery up to four weeks post-TBI.^21,22,29,32 Motor and cognitive function were also reported to be improved by ODSH³² and enoxaparin.²⁹

Brain edema, which is a common pathology in TBI, could worsen outcomes by increasing intracranial pressure causing secondary ischemia and reducing cerebral perfusion and tissue oxygenation.^55,56 The studies we assessed reported that enoxaparin, UFH, and ODSH all decreased brain edema after experimental TBI.^21,22,24,29 The effect of dosing in these studies is unclear. Low dose enoxaparin²⁴ and both low and high dose UFH ^21,22 but only high dose ODSH³² were reported to reduce edema. In related studies in photothrombotic stroke, only higher doses of enoxaparin are reported to improve cerebral edema.¹⁰ These edema outcomes are discussed further in the meta-analysis section below.

One contributor to brain edema and worse TBI outcome is damage to the blood–brain barrier, the complex structure that controls the movement of nutrients, proteins, and immune cells among others in and out of the brain.⁵⁷ Assessed here as microvascular permeability, enoxaparin, UFH, and ODSH all reduced blood–brain barrier leakage in the controlled cortical impact model at 48 h post-TBI irrespective of the drug dose, mode of delivery, or drug type.^21,24,32

Overall, our meta-analysis suggests that UFH and enoxaparin do appear to reduce brain edema (Fig. 3) and improve neurological function (Fig. 6) in animal models of TBI. Moreover, it appears that enoxaparin and UFH provided very similar effects, albeit with considerable differences in experimental heterogeneity, which we cannot currently explain (Fig. 4). Intriguingly, there was a trend suggesting that lower doses (0.75 mg/kg minimum, not greater than 1 mg/kg) might be more effective than higher doses (≥2 mg/kg) (Fig. 5).

As an additional note of caution, however, heterogeneity in the two high dose studies was high despite coming from the same research team and employing very similar methodologies including animal species, mode of drug delivery, and dose of drug.^21,22 Conversely, heterogeneity was very low between the low dose studies.^21,22,24

The studies included in this review did not report an increase in hemorrhage after the administration of enoxaparin post-trauma^22,24,25,28; indeed the meta-analysis (Fig. 7A) would appear to suggest that the treatment improved hemorrhagic outcome, suggesting that drugs like enoxaparin may be relatively safe. This suggestion, however, has to be weighed against the risk of clinically significant bleeding complications because early use of drugs with anti-coagulant properties^11,58 might be expected to worsen outcome by increasing intracranial/cerebral hemorrhage, which is commonly observed after TBI.⁵⁹ It should be noted, however, that the early administration of enoxaparin in patients with severe TBI did not show significant increase in cerebral hemorrhage in a human clinical trial.⁶⁰

Neuroinflammation is a critical secondary injury mechanism that interacts with blood–brain barrier damage after TBI, contributing to ongoing neuronal damage and death.^61,62 It encompasses infiltration of peripheral immune cells into the brain, increased circulating and brain concentrations of inflammatory mediators, and consequent astrogliosis and microgliosis.^61,63 In related models of subarachnoid hemorrhage, heparin is reported to reduce microglial activation and expression of the inflammatory cytokines, tumor necrosis factor-α and IL-1β.⁶⁴

Eight of our included studies together investigated the effect of UFH, ODSH, and enoxaparin on leukocyte rolling, leukocyte adhesion, PMN cell sequestration, pro-IL-1β, COX-2 expression, reactive gliosis, and glial scar formation after experimental TBI.^{21,22,24,26,28,30
–32} Overall, these inflammatory outcomes were improved by heparin,^21,22,31 enoxaparin,^24,26,28,30 and ODSH³² regardless of dose. There were no effects reported on leukocyte adhesion,^21,32 pro-IL-1β concentration,³⁰ or astrogliosis, however.³⁰ While most of these parameters provided insufficient data for meta-analysis, three studies that examined both vascular permeability and leukocyte rolling did provide evidence for a consistent beneficial effect (Fig. 7B, 7C).

In the studies evaluated here, reduced apoptosis was reported in response to enoxaparin.^27,28,30 One study compared between doses and actually demonstrated that the decrease observed was dose-dependent.²⁷ Similarly, enoxaparin also decreased the expression of inflammasome proteins IL-1β and caspase-1.²⁶ There were insufficient data to examine this by meta-analysis.

Oxidative stress is one of many other important contributors to the pathophysiology of TBI that were investigated in the studies we reviewed. In TBI, markers of oxidative stress such as thiobarbituric acid-reactive substances (TBARS) are increased in the brain while antioxidant defence enzymes such as superoxide dismutase and glutathione are reduced.^65
–67

The studies we report that examined the effects of enoxaparin on oxidative stress found decreased concentrations of TBARS and oxidized proteins in the hippocampus in the lateral fluid percussion injury rat model of TBI at 48 h after injury,³⁰ while serum levels of “total antioxidant status” were increased after administration of high dose enoxaparin in a cold-induced TBI model.²⁷ It remains to be seen whether antioxidant actions are important contributors to the effects of heparins/LMWHs overall.

We had originally pre-planned for the possibility of a range of subgroup analyses (PROSPERO; CRD42020205574; e.g., drug type, mode of drug administration, drug dose, sex of animal, animal species, timing, and frequency of treatment) to explore the factors that contributed to experimental heterogeneity. Most proved impractical to analyze effectively, and some could not be analyzed at all. For example, many injury models were used but none sufficiently often to allow comparison.

Controlled cortical impact producing a severe TBI was the most common way to induce TBI.^{21,22,24,26,32} Similarly, multiple routes of drug delivery were used, including within studies,^28,29 all with apparent benefit. No study, however, provided data indicating that any route of delivery was to be preferred, and we found that insufficient data were available for meta-analysis to provide guidance for this important practical question.^68,69 Similarly, the timing of treatment was also considered important but not possible to analyze.

The potential for bias is another reason for caution. While systematic review and meta-analysis are important tools for finding, aggregating, and helping interpret data,^70
–72 it must be remembered that if the input data are biased, the output data will also be influenced by this bias. While we have endeavored to not do so, it is also possible for introduction of bias during the process of systematic review and meta-analysis.^73
–75

Overall, the studies we were able to find were generally of low quality (median four of 10) based on the modified checklist from CAMARADES.^16,17 Despite having been subjected to peer review, there were notable absences of data reporting: for example, whether measures such as blinding had been taken to help avoid the introduction of bias. These absences are particularly well highlighted by use of the SYRCLE Risk of Bias tool (Fig. 2). Sadly, this finding is consistent with the analyses of the use of these measures across in vivo pre-clinical research⁷⁶ where such absences are associated with inflated effect sizes.^77,78

It was reassuring to discover, however, that random allocation into groups was very well reported by virtually all the studies included in this review. Funnel plots asymmetry and Egger regression, albeit based on a small sample size (Fig. 8), also indicates the possibility of publication bias or some other source of asymmetry in these data.^79,80 Debray and colleagues,⁸¹ however, have suggested that funnel plot should be avoided for reviews with few studies because it could be unreliable.

The search for a treatment in TBI is a truly long and challenging one. Despite demonstrating efficacy in pre-clinical studies, several potential therapeutic agents such as progesterone⁸² failed to achieve clinically meaningful primary end points in large scale phase III clinical trials. Improved clinical methodology, such as selection of patients and adaptive Bayesian phase II trials, have been suggested as some of the key factors to be considered for future trials.⁸³

In addition to optimizing clinical trial designs, robust pre-clinical data on pharmacokinetics-pharmacodynamic (PKPD) of a therapeutic candidate, in this case, heparins/LMWHs, to support the design of later clinical trials are needed. Although being used clinically in many disorders, the understanding of the PKPD relationship of the heparins/LMWHs in the context of neuroinflammation and neurological injuries is still insufficient.⁸⁴ Future investigations to capture these data will provide better insights on the function and effects of heparins/LMWHs, including where and how they act, whether they cross the blood–brain barrier, dose-response relationship, risk/benefit ratio at different doses, as well as optimization of time and route of administration.

Limitations

To our knowledge, this is the first systematic review and meta-analysis investigating the therapeutic effects of heparins and LMWHs in animal models of TBI. The study has important limitations, especially with respect to the small sample size available and presence of substantial heterogeneity, and more than 50% of the studies included for meta-analysis coming from one research group.^21,22,24,32 Further, we found that there are insufficient data to evaluate the experimental variables we had hoped to investigate, including dose, timing, and route of administration, as well as the relative efficacy of heparins and LMWHs on different TBI outcomes.

Despite these shortcomings, we have identified that across the 11 studies included in our review, treatment with UFH, enoxaparin, and/or ODSH shows consistent beneficial effects, including decreased cerebral edema, inflammation, microvascular permeability, recruitment and accumulation of leukocytes, apoptosis, and brain contusion volume, and improved neurological function. These pharmacological outcomes were elicited using different dosing regimens and different rodent models of TBI, demonstrating a robustness to these findings that suggests further pre-clinical studies are warranted.

Our review highlights areas where more work needs to be done, such as comparing alternate dosing regimens within single studies, because the effect of dosing remains unclear. Also, while assessing a variety of outcome measures will be important to elucidate the range of therapeutic effects of heparins and LMWHs, it would be valuable for future studies to also include analyses of key TBI outcomes, such as those identified by the Operation Brain Trauma Therapy consortium³⁷ to facilitate the comparison of therapeutic efficacy across different studies. Finally, the possibility of publication bias that we've identified speaks to the need for publishing negative results to prevent the exaggeration of beneficial findings.

Conclusion

Overall, in this review, we found that heparins and LMWHs, including UFH, ODSH, and enoxaparin, appeared to improve outcome after TBI with no increased hemorrhage. With enoxaparin's better bioavailability profile, more predictable dose-response relationship, longer half-life, less anti-IIa (thrombin) activity, and half the anti-Xa (prothrombinase) activity than heparin,^10,85 it may have advantages for clinical use. The inferences made above, however, are based on a very limited number of animal studies, and therefore these findings must be interpreted with caution. Nevertheless, there is a clear need for more detailed pre-clinical studies to evaluate the therapeutic potential of heparins and LMWHs as a treatment for TBI.

Footnotes

Acknowledgments

We would like to thank the University of Tasmania research librarian, Michaela Venn for her assistance in developing the final search strategy. Mimieveshiofuo Aiyede was supported by an Australian Government Research Training Program (RTP) Stipend and RTP Fee-offset Scholarship through the University of Tasmania.

Authors' Contributions

MA and NB devised and coordinated the study. MA, DWH, and NB developed the search strategy and registered the protocol. MA and XYL reviewed the studies and extracted data. MA and AAMR designed the statistical analysis, and MA conducted the risk of bias assessments and statistical analysis. MA, DWH and NB performed the data interpretation. MA drafted the manuscript. MA, XYL, AMR, RPP, NG, DWH and NB helped to revise the manuscript. All authors read and approved the final manuscript.

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary Data S1

Supplementary Data S2

Supplementary Data S3

Supplementary Data S4

Supplementary Data S5

Supplementary Data S6

References

Dewan

, Rattani

, Gupta

, et al. Estimating the global incidence of traumatic brain injury. J Neurosurg, 2018; 1-18; doi: 10.3171/2017.10.jns17352

Prasetyo

The primary, secondary, and tertiary brain injury. Crit Care Shock, 2020; 23:4-13

Jarrahi

, Braun

, Ahluwalia

, et al. Revisiting traumatic brain injury: from molecular mechanisms to therapeutic interventions. Biomedicines, 2020; 8(10):389; doi: 10.3390/biomedicines8100389

, Lee

AYW

. Traumatic brain injuries: pathophysiology and potential therapeutic targets. Front Cell Neurosci, 2019; 13:528; doi: 10.3389/fncel.2019.00528

Xiong

, Mahmood

, Chopp

. Emerging treatments for traumatic brain injury. Expert Opin Emerg Drugs, 2009; 14(1):67-84; doi: 10.1517/14728210902769601

Vink

, Nimmo

. Multifunctional drugs for head injury. Neurotherapeutics, 2009; 6(1):28-42; doi: 10.1016/j.nurt.2008.10.036

Oduah

, Linhardt

, Sharfstein

. Heparin: past, present, and future. Pharmaceuticals (Basel), 2016; 9(3):38; doi: 10.3390/ph9030038

Paluck

, Nguyen

, Maynard

. Heparin-mimicking polymers: synthesis and biological applications. Biomacromolecules, 2016; 17(11):3417-3440; doi: 10.1021/acs.biomac.6b01147

Weiss

, Esko

, Tor

. Targeting heparin–and heparan sulfate–protein interactions. Org Biomol Chem, 2017; 15(27):5656-5668; doi: 10.1039/c7ob01058c

10.

Stutzmann

, Mary

, Wahl

, et al. Neuroprotective profile of enoxaparin, a low molecular weight heparin, in in vivo models of cerebral ischemia or traumatic brain injury in rats: a review. CNS Drug Rev, 2002; 8(1):1-30; doi: 10.1111/j.1527-3458.2002.tb00213.x

11.

Glassner

, Srivastava

, Cofnas

, et al. Prevention of venous thrombotic events in brain injury: review of current practices. Rambam Maimonides Med J, 2013; 4(1):e0001; doi: 10.5041/RMMJ.10101

12.

Page

, McKenzie

, Bossuyt

, et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. PloS Med, 2021; 18(3):e1003583; doi: 10.1371/journal.pmed.1003583

13.

Leenaars

, Hooijmans

, van Veggel

, et al. A step-by-step guide to systematically identify all relevant animal studies. Lab Anim, 2012; 46(1):24-31; doi: 10.1258/la.2011.011087

14.

Higgins JPT

, Chandler

, Cumpston

, Li

, Page

, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.2 (updated February 2021) Available from www.training.cochrane.org/handbook. [Last accessed: 5/20/2021]

15.

Hooijmans

, Rovers

, de Vries

RBM

, et al. SYRCLE's risk of bias tool for animal studies. BMC Med Res Methodol, 2014; 14:43; doi: 10.1186/1471-2288-14-43

16.

Macleod

, O'Collins

, Howells

, et al. Pooling of animal experimental data reveals influence of study design and publication bias. Stroke, 2004; 35(5):1203-1208; doi: 10.1161/01.STR.0000125719.25853.20

17.

Hirst

, Watzlawick

, Rhodes

, et al. Study protocol - A systematic review and meta-analysis of hypothermia in experimental traumatic brain injury: why have promising animal studies not been replicated in pragmatic clinical trials?. Evid Based Preclin Med, 2016; 3(2):e00020; doi: 10.1002/ebm2.20

18.

DerSimonian

, Laird

. Meta-analysis in clinical trials. Control Clin Trials, 1986; 7(3):177-188; doi: 10.1016/0197-2456(86)90046-2

19.

Haidich

AB.

Meta-analysis in medical research. Hippokratia, 2010; 14(Suppl 1):29-37.

20.

Sedgwick

, Marston

. Meta-analyses: standardised mean differences. BMJ, 2013; 347:2; doi: 10.1136/bmj.f7257

21.

Nagata

, Kumasaka

, Browne

, et al. Unfractionated heparin after TBI reduces in vivo cerebrovascular inflammation, brain edema and accelerates cognitive recovery. J Trauma Acute Care Surg, 2016; 81(6):1088-1094; doi: 10.1097/TA.0000000000001215

22.

Nagata

, Browne

, Suto

, et al. Early heparin administration after traumatic brain injury: prolonged cognitive recovery associated with reduced cerebral edema and neutrophil sequestration. J Trauma Acute Care Surg, 2017; 83(3):406-412; doi: 10.1097/TA.0000000000001590

23.

Busti

AJ.

Dosing differences and rationale among low molecular weight heparins (LMWH; enoxaparin, Lovenox, dalteparin, Fragmin, etc). Available from: https://www.ebmconsult.com/articles/lmwh-enoxaparin-lovenox-dalteparin-fragmin-dosing-differences. [Last accessed:12/20/2021]

24.

, Marks

, Eisenstadt

, et al. Enoxaparin ameliorates post-traumatic brain injury edema and neurologic recovery, reducing cerebral leukocyte endothelial interactions and vessel permeability in vivo. J Trauma Acute Care Surg, 2015; 79(1):78-84; doi: 10.1097/TA.0000000000000697

25.

Dhir

, Weiss

, Wolanin

, et al. The impact of enoxaparin administration in relationship to hemorrhage in mild traumatic brain injury. Injury, 2018; 49(12):2174-2177; doi: 10.1016/j.injury.2018.09.027

26.

Kerr

, Vaccari

JPD

, Weaver

, et al. Enoxaparin attenuates acute lung injury and inflammasome activation after traumatic brain injury. J Neurotrauma, 2021; 38(5):646-654; doi: 10.1089/neu.2020.7257

27.

Keskin

, Gunal

, Ayturk

, et al. Dose-dependent neuroprotective effect of enoxaparin on cold-induced traumatic brain injury. Neural Regen Res, 2017; 12(5):761-764; doi: 10.4103/1673-5374.206646

28.

Sen

, Sonmez

, Cekinmez

, et al. Antithrombin III and enoxaparin treatment inhibit contusion-triggered cell death, inflammation, hemorrhage and apoptosis after severe traumatic brain injury in rats. Turk Neurosurg, 2011; 21(2):203-209; doi: 10.5137/1019-5149.JTN.3646-10.1

29.

Wahl

, Grosjean-Piot

, Bareyre

, et al. Enoxaparin reduces brain edema, cerebral lesions, and improves motor and cognitive impairments induced by a traumatic brain injury in rats. J Neurotrauma, 2000; 17(11):1055-1065; doi: 10.1089/neu.2000.17.1055

30.

Zupan

, Pilipovic

, Dangubic

, et al. Effects of enoxaparin in the rat hippocampus following traumatic brain injury. Prog Neuropsychopharmacol Biol Psychiatry, 2011; 35(8):1846-1856; doi: 10.1016/j.pnpbp.2011.08.005

31.

Hayashi

, Miyata

, Kariya

, et al. Attenuation of glial scar formation in the injured rat brain by heparin oligosaccharides. Neurosci Res, 2004; 49(1):19-27; doi: 10.1016/j.neures.2004.01.007

32.

Nagata

, Suto

, Cognetti

, et al. Early low-anticoagulant desulfated heparin after traumatic brain injury: reduced brain edema and leukocyte mobilization is associated with improved watermaze learning ability weeks after injury. J Trauma Acute Care Surg, 2018; 84(5):727-735; doi: 10.1097/ta.0000000000001819

33.

Prins

, Greco

, Alexander

, et al. The pathophysiology of traumatic brain injury at a glance. Dis Model Mech, 2013; 6(6):1307-1315; doi: 10.1242/dmm.011585

34.

Karp

NA.

Reproducible preclinical research—Is embracing variability the answer?. PloS Biol, 2018; 16(3):e20055413; doi: 10.1371/journal.pbio.2005413

35.

Percie du Sert

, Hurst

, Ahluwalia

, et al. The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research. PLoS Biol, 2020; 18(7):e3000410; doi: 10.1371/journal.pbio.3000410

36.

Percie du Sert

, Ahluwalia

, Alam

, et al. Reporting animal research: Explanation and elaboration for the ARRIVE guidelines 2.0. PLoS Biol, 2020; 18(7):e3000411; doi: 10.1371/journal.pbio.3000411

37.

Kochanek

, Dixon

, Mondello

, et al. Multi-center pre-clinical consortia to enhance translation of therapies and biomarkers for traumatic brain injury: operation brain trauma therapy and beyond. Front Neurol, 2018; 9:640; doi: 10.3389/fneur.2018.00640

38.

Smith

, Hicks

, Johnson

, et al. Pre-clinical traumatic brain injury common data elements: toward a common language across laboratories. J Neurotrauma, 2015; 32(22):1725-1735; doi: 10.1089/neu.2014.3861

39.

LaPlaca

, Huie

, Alam

, et al. Pre-clinical common data elements for traumatic brain injury research: progress and use cases. J Neurotrauma, 2021; 38(10):1399-1410; doi: 10.1089/neu.2020.7328

40.

Smith

, Kochanek

, Rosi

, et al. Roadmap for advancing pre-clinical science in traumatic brain injury. J Neurotrauma, 2021; 38(23):3204-3221; doi: 10.1089/neu.2021.0094

41.

Egger

, Davey Smith

, O'Rourke

Rationale, potentials, and promise of systematic reviews. BMJ Publishing: London; 2001

42.

Petticrew

Why certain systematic reviews reach uncertain conclusions. BMJ, 2003; 326(7392):756-758; doi: 10.1136/bmj.326.7392.756

43.

Voelkl

, Altman

, Forsman

, et al. Reproducibility of animal research in light of biological variation. Nat Rev Neurosci, 2020; 21(7):384-393; doi: 10.1038/s41583-020-0313-3

44.

Rodriguez

, Herzog

, Gee

. Variability in human-animal interaction research. Front Vet Sci, 2021; 7:619600; doi: 10.3389/fvets.2020.619600

45.

Duncan

, Garijo-Garde

. Sex, genes, and traumatic brain injury (TBI): a call for a gender inclusive approach to the study of TBI in the lab. Front Neurosci, 2021; 15:681599; doi: 10.3389/fnins.2021.681599

46.

Farace

, Alves

. Do women fare worse: a metaanalysis of gender differences in traumatic brain injury outcome. J Neurosurg, 2000; 93(4):539-545; doi: 10.3171/jns.2000.93.4.0539

47.

Brann

, Dhandapani

, Wakade

, et al. Neurotrophic and neuroprotective actions of estrogen: basic mechanisms and clinical implications. Steroids, 2007; 72(5):381-405; doi: 10.1016/j.steroids.2007.02.003

48.

Webster

, Wright

, Sun

, et al. Progesterone treatment reduces neuroinflammation, oxidative stress and brain damage and improves long-term outcomes in a rat model of repeated mild traumatic brain injury. J Neuroinflammation, 2015; 12:238; doi: 10.1186/s12974-015-0457-7

49.

Brotfain

, Gruenbaum

, Boyko

, et al. Neuroprotection by estrogen and progesterone in traumatic brain injury and spinal cord injury. Curr Neuropharmacol, 2016; 14(6):641-653; doi: 10.2174/1570159x14666160309123554

50.

DeWitt

, Hawkins

, Dixon

, et al. Pre-clinical testing of therapies for traumatic brain injury. J Neurotrauma, 2018; 35(23):2737-2754; doi: 10.1089/neu.2018.5778

51.

Niesman

, Schilling

, Shapiro

, et al. Traumatic brain injury enhances neuroinflammation and lesion volume in caveolin deficient mice. J Neuroinflammation, 2014; 11:39; doi: 10.1186/1742-2094-11-39.

52.

Eriksen

, Pakkenberg

, Rostrup

, et al. Neurostereologic lesion volumes and spreading depolarizations in severe traumatic brain injury patients: a pilot study. Neurocrit Care, 2019; 30(3):557-568; doi: 10.1007/s12028-019-00692-w.

53.

Corrigan

, Vink

, Blumbergs

, et al. Evaluation of the effects of treatment with sAPPα on functional and histological outcome following controlled cortical impact injury in mice. Neurosci Lett, 2012; 515(1):50-54; doi: 10.1016/j.neulet.2012.03.017

54.

Zhou

, Li

, Zheng

, et al. Erythropoietin regulates immune/inflammatory reaction and improves neurological function outcomes in traumatic brain injury. Brain Behav, 2017; 7(11):e00827; doi: 10.1002/brb3.827

55.

Unterberg

, Stover

, Kress

, et al. Edema and brain trauma. Neuroscience, 2004; 129(4):1021-1029; doi: 10.1016/j.neuroscience.2004.06.046

56.

Winkler

, Minter

, Yue

, et al. Cerebral edema in traumatic brain injury: pathophysiology and prospective therapeutic targets. Neurosurg Clin N Am, 2016; 27(4):473-488; doi: 10.1016/j.nec.2016.05.008

57.

Jha

, Kochanek

, Simard

. Pathophysiology and treatment of cerebral edema in traumatic brain injury. Neuropharmacology, 2019; 145(Pt B):230-246; doi: 10.1016/j.neuropharm.2018.08.004

58.

Norwood

, Berne

, Rowe

, et al. Early venous thromboembolism prophylaxis with enoxaparin in patients with blunt traumatic brain injury. J Trauma, 2008; 65(5):1021-1027; doi: 10.1097/TA.0b013e31818a0e74

59.

Kaur

, Sharma

. Recent advances in pathophysiology of traumatic brain injury. Curr Neuropharmacol, 2018; 16(8):1224-1238; doi: 10.2174/1570159X15666170613083606

60.

Baharvahdat

, Ganjeifar

, Etemadrezaie

, et al. Enoxaparin in the treatment of severe traumatic brain injury: a randomized clinical trial. Surg Neurol Int, 2019; 10:10; doi: 10.4103/sni.sni_112_18

61.

Sulhan

, Lyon

, Shapiro

, et al. Neuroinflammation and blood-brain barrier disruption following traumatic brain injury: pathophysiology and potential therapeutic targets. J Neurosci Res, 2020; 98(1):19-28; doi: 10.1002/jnr.24331

62.

Stoll

, Jander

, Schroeter

. Detrimental and beneficial effects of injury-induced inflammation and cytokine expression in the nervous system. Adv Exp Med Biol, 2002; 513:87-113; doi: 10.1007/978-1-4615-0123-7_3

63.

Morganti-Kossmann

, Semple

, Hellewell

, et al. The complexity of neuroinflammation consequent to traumatic brain injury: from research evidence to potential treatments. Acta Neuropathol, 2019; 137(5):731-755; doi: 10.1007/s00401-018-1944-6

64.

Simard

, Tosun

, Ivanova

, et al. Heparin reduces neuroinflammation and transsynaptic neuronal apoptosis in a model of subarachnoid hemorrhage. Transl Stroke Res, 2012; 3(Suppl 1):S155-S165; doi: 10.1007/s12975-012-0166-9

65.

Rodriguez-Rodriguez

, Egea-Guerrero

, Murillo-Cabezas

, et al. Oxidative stress in traumatic brain injury. Curr Med Chem, 2014; 21(10):1201-1211; doi: 10.2174/0929867321666131217153310

66.

Lin

, Chen

, Wang

, et al. Association between peripheral oxidative stress and white matter damage in acute traumatic brain injury. Biomed Res Int, 2014; 2014:340936; doi: 10.1155/2014/340936

67.

Shohami

, Beit-Yannai

, Horowitz

, et al. Oxidative stress in closed-head injury: brain antioxidant capacity as an indicator of functional outcome. J Cereb Blood Flow Metab, 1997; 17(10):1007-1019; doi: 10.1097/00004647-199710000-00002

68.

Kim

, De Jesus

Medication Routes of Administration. StatPearls Publishing: Treasure Island, FL; 2022.

69.

Wen

, Jung

, Li

. Drug delivery approaches in addressing clinical pharmacology-related issues: opportunities and challenges. AAPS J, 2015; 17(6):1327-1340; doi: 10.1208/s12248-015-9814-9

70.

Sandercock

, Roberts

. Systematic reviews of animal experiments. Lancet, 2002; 360(9333):586-586; doi: 10.1016/s0140-6736(02)09812-4

71.

Roberts

, Kwan

, Evans

, et al. Does animal experimentation inform human healthcare? Observations from a systematic review of international animal experiments on fluid resuscitation. BMJ, 2002; 324(7335):474-476; doi: 10.1136/bmj.324.7335.474

72.

Pound

, Ritskes-Hoitinga

. Can prospective systematic reviews of animal studies improve clinical translation?. J Transl Med, 2020; 18:15; doi: 10.1186/s12967-019-02205-x

73.

Pannucci

, Wilkins

. Identifying and avoiding bias in research. Plast Reconstr Surg, 2010; 126(2):619-625; doi: 10.1097/PRS.0b013e3181de24bc

74.

McDonagh

, Peterson

, Raina

, et al. Avoiding bias in selecting studies. 2013 Feb 20. In: Methods Guide for Effectiveness and Comparative Effectiveness Reviews [Internet]. Agency for Healthcare Research and Quality (US): Rockville, MD; 2008

75.

Kirkham

, Altman

, Williamson

. Bias due to changes in specified outcomes during the systematic review process. PloS One, 2010; 5(3):e9810; doi: 10.1371/journal.pone.0009810

76.

The NPQIP Collaborative group. Did a change in Nature journals' editorial policy for life sciences research improve reporting?. BMJ Open Sci, 2019; 3(1):e000035; doi: 10.1136/bmjos-2017-000035

77.

Hirst

, Howick

, Aronson

, et al. The need for randomization in animal trials: an overview of systematic reviews. PloS One, 2014; 9(6): e98856; doi: 10.1371/journal.pone.0098856

78.

van der Worp

, Macleod

. Preclinical studies of human disease: time to take methodological quality seriously. J Mol Cell Cardiol, 2011; 51(4):449-450; doi: 10.1016/j.yjmcc.2011.04.008

79.

Egger

, Smith

, Schneider

, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ, 1997; 315(7109):629-634; doi: 10.1136/bmj.315.7109.629

80.

Sterne

JAC

, Sutton

, Ioannidis

JPA

, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ, 2011; 343:d4002; doi: 10.1136/bmj.d4002

81.

Debray

TPA

, Moons

KGM

, Riley

. Detecting small-study effects and funnel plot asymmetry in meta-analysis of survival data: a comparison of new and existing tests. Res Synth Methods, 2018; 9(1):41-50; doi: 10.1002/jrsm.1266

82.

Skolnick

, Maas

, Narayan

, et al. A clinical trial of progesterone for severe traumatic brain injury. N Engl J Med, 2014; 371(26):2467-2476; doi: 10.1056/NEJMoa1411090

83.

Stein

DG.

Lost in translation: understanding the failure of the progesterone/traumatic brain injury Phase III trials. Future Neurol, 2016; 11(1):9-13; doi: 10.2217/fnl.15.49

84.

Qiu

, Huang

, Luo

, et al. Pharmacological and clinical application of heparin progress: an essential drug for modern medicine. Biomed Pharmacother, 2021; 139:111561; doi: 10.1016/j.biopha.2021.111561

85.

Nutescu

, Burnett

, Fanikos

, et al. Pharmacology of anticoagulants used in the treatment of venous thromboembolism. J Thromb Thrombolysis, 2016; 41(1):15-31; doi: 10.1007/s11239-015-1314-3

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.22 MB

0.16 MB

0.14 MB

0.09 MB

0.19 MB

0.18 MB