3.2.3. Low Quality of Evidence

There may be minimal effect on the frequency of early seizures (7 days) for neuroprotective agents compared to placebo, (moderate-to-severe TBI, 1 study, *n* = 499) [25]. Antiepileptic interventions compared with placebo may reduce the frequency of early seizures (moderate-to-severe, 5 studies, *n*=987) [25]. Neuroprotective agents versus other antiepileptic drugs may have minimal effect on adverse events (moderate-to-severe TBI, 2 studies, *n* = 431) [25].

#### 3.2.4. Very Low Quality of Evidence

A review comparing baclofen 50 μg versus saline placebo included one study (*n* = 11) and examined the effects on spasticity (6 h), and adverse events [28]. The findings could not be extracted since they were not reported in the randomized control trial. The efficacy and safety of the intervention remain thereby unclear.

A review evaluated the efficacy of botulinum toxin A × 1 dose (500/1000 U) or botulinum toxin A × 1 dose of 200 U + serial casting versus placebo on spasticity (4–12 weeks post treatment), and adverse events (2 studies; *n* = 47) [28]. No statistically significant differences were detected between groups and the quality of evidence was rated very low. This hindered the ability to ascertain the true treatment effects of either intervention.

Evaluating 1029 participants and six studies, one CSR examined the difference in effects on late seizure occurrence (3 to 24 months after the start of the treatment) comparing between antiepileptic medications and placebo [25]. No significant differences were found for either outcome. The comparison was judged to provide very low quality of evidence, which indicates that the effects of antiepileptic interventions on these two outcomes remain uncertain.

In a total sample of 67 participants and one study, the reviewers found a significant difference in depression level between the repetitive transcranial magnetic stimulation and repetitive transcranial magnetic stimulation plus antidepressant groups (TBI severity unspecified) [24]. While the treatment effect was in favor of the comparator, repetitive transcranial magnetic stimulation plus tricyclic antidepressants, the true treatment effect remains uncertain due to the very low quality of evidence.

#### **4. Discussion**

This overview summarizes evidence on the effects of non-pharmacological and pharmacological interventions for any level of TBI severity, and reports the challenges identified in TBI research that are critical for further developing the integration and augmentation of rehabilitation services.

Amongst the options for non-pharmacological interventions, hospital-based cognitive rehabilitation and cognitive didactic therapy likely produce minimal or no changes in the return-to-work rate (moderate certainty evidence). These findings agree with published reports in the literature on neurocognitive status and the return-to-work rates, ref. [29–31] which maintain that favorable outcomes are facilitated by the inclusion of multidisciplinary/interdisciplinary rehabilitation services, and not by a monotherapy approach, such as cognitive rehabilitation or cognitive training alone [32,33]. Executive functions, especially sequencing and inhibitory control, are necessary to perform well at work and their status predicts the return-to-work rate following TBI [29]. Ensuring that available cognitive interventions and cognitive strategy training lead to improvements in cognitive functioning and are properly integrated in the rehabilitation management are crucial for increasing return-to-work rates, as well as improving life satisfaction and the wellbeing of individuals with TBI and their families.

The low-certainty of evidence found in acupuncture, splint therapy, and exercise of large muscle groups prevented us from ascertaining the role of these interventions on Glasgow Coma Scale scores, spasticity, and cardiorespiratory fitness, respectively. With respect to acupuncture, the lack of information on the etiology of the TBI from three of the four RCTs prevented us from determining whether the results are equally applicable to acquired brain injury, traumatic brain injury, and non-traumatic brain injury cases. Likewise, there is insufficient quality of evidence to support the roles of cognitive therapeutic approaches as monotherapy in improving community integration, depression, and activities of daily living (very low certainty evidence).

Amongst the pharmacological interventions used to reduce the number and frequency of late-seizure episodes (i.e., 6 months after the start of treatment; high-quality evidence), neuroprotective agents produced little to no difference on the frequency of late-seizures (high-quality evidence) and minimal differences on early seizures (low-quality evidence). The anti-convulsant drug, phenytoin, for example, appeared to have little effect on the number and frequency of late seizures (moderate quality evidence) and little to no effect on early-seizure events (low quality evidence). This finding aligns with current guidelines that support the use of phenytoin to treat early seizures or active seizures, but not late seizures [34].

Our evidence mapping shows that other antiepileptic drugs do not reduce the number and frequency of late seizure events. The literature primarily focuses on early seizures, and data on late seizures after TBI are limited. Discussions of study results typically note that no evidence supports the use of neuroprotective agents and antiepileptic drugs for late seizures, mainly due to the differences observed in studies on pathogenesis of early seizures in post-traumatic epilepsy [34,35]. This feature of post-TBI care warrants further attention since late seizure episodes may impair otherwise positive neurological and rehabilitation outcomes [36].

For the remaining two pharmacological interventions (botulinum toxin A × 1 dose (500/1000 U) or botulinum toxin A × 1 dose of 200 U + serial casting; intrathecal baclofen 50 μg), uncertainty of their effects on spasticity and adverse events remain, as the quality of evidence for these two therapies has been assessed as very low [28].

The absence and/or low quality of evidence for pharmacological interventions to reduce early- and late-seizure frequency, and improve spasticity, may be associated in part with the following situations: (1) research challenges exacerbated by the narrow window for effective intervention; (2) the inability of candidate medications to cross the blood–brain barrier; and (3) possible delays and ethical issues encountered when patients are unable to provide consent [37]. These difficulties are exacerbated among pediatric groups [38], which may explain the limited results for pediatric patients with TBI among the CSRs that analyzed pharmacological interventions.

The low to very- low quality evidence found is in accordance with past reviews that focused on clinical practice guidelines for TBI [39,40], which stressed the persistent paucity of quality evidence and the major gaps between the bench and the bedside in the context of rehabilitation interventions associated with both methodological issues and clinical complexity. The reviewers stated that few published trials examined rehabilitation outcomes, such as cognitive and physical function, with the majority of studies targeting symptom management or reduction [39,40].

For non-pharmacological trials, the primary issues concerned the number of studies and the small sample sizes (cumulative <500 participants), which affected the estimated effect sizes, heterogeneity among the respondents, and the imprecision of the results (i.e., wide 95% confidence intervals). Similar to pharmacological trials, some studies showed a lack of clarity regarding random sequence generation, blinding, and allocation concealment.

Overall, our evidence map shows that no prominent advances were reported in any of the CSRs, confirming the concerns expressed a decade ago by Maas et al. [41], who observed that randomized control trials (RCTs) fail to showcase significant recovery trajectories when assessing the effectiveness of interventions on TBI populations. Other study designs (e.g., observational) could provide additional insights when conducting systematic reviews for patients with TBI.

The landscape displayed by this evidence map places strong emphasis on the need to prioritize and augment rehabilitation research efforts for patients with TBI. Hence, we reiterate four priorities for bolstering the quality of evidence associated with rehabilitation outcomes: (1) revisit the recruitment and consent process and preserve ethical standards; (2) increase efforts and funding to support trials that examine functioning (i.e., cognitive, physical, and emotional); (3) consider multi-site recruitment options to increase participant diversity and sample sizes; (4) clearly identify the etiology of brain injury or offer disaggregate data in studies with mixed brain injury populations; and (5) promote the transparent reporting of adverse events, if applicable.
