Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2022 Mar 31;17(3):e0265932. doi: 10.1371/journal.pone.0265932

Comparative efficacy of prophylactic anticonvulsant drugs following traumatic brain injury: A systematic review and network meta-analysis of randomized controlled trials

Bo-Cyuan Wang 1,2,#, Hsiao-Yean Chiu 1,#, Hui-Tzung Luh 3,4,5, Chia-Jou Lin 1, Shu-Hua Hsieh 1,6, Ting-Jhen Chen 1, Chia-Rung Wu 1,6, Pin-Yuan Chen 7,8,9,*
Editor: Giuseppe Biagini10
PMCID: PMC8970384  PMID: 35358219

Abstract

We systematically compared the effects of prophylactic anticonvulsant drug use in patients with traumatic brain injury. We searched four electronic databases from their inception until July 13, 2021. Two researchers independently screened, appraised, and extracted the included studies. Network meta-analysis using multivariate random effects and a frequentist framework was adopted for data analysis. The risk of bias of each study was assessed using the Cochrane risk of bias tool, and confidence in evidence was assessed through confidence in network meta-analysis (CINeMA). A total of 11 randomized controlled trials involving 2,450 participants and six different treatments (i.e., placebo, carbamazepine, phenytoin, levetiracetam, valproate, and magnesium sulfate) were included. We found that anticonvulsant drugs as a whole significantly reduced early posttraumatic seizures (PTS) but not late PTS compared with placebo (odd ratios [ORs] = 0.42 and 0.82, 95% confidence intervals [CIs] = 0.21–0.82 and 0.47–1.43). For the findings of network meta-analysis, we observed that phenytoin (ORs = 0.43 and 0.71; 95% CIs = 0.18–1.01 and 0.23–2.20), levetiracetam (ORs = 0.56 and 1.58; 95% CIs = 0.12–2.55 and 0.03–84.42), and carbamazepine (ORs = 0.29 and 0.64; 95% CIs = 0.07–1.18 and 0.08–5.28) were more likely to reduce early and late PTS compared with placebo; however, the treatment effects were not significant. Sensitivity analysis, after excluding a study enrolling only children, revealed that phenytoin had a significant effect in preventing early PTS (OR = 0.33; 95% CI = 0.14–0.78). Our findings indicate that no antiepileptic drug had an effect on early or late PTS superior to that of another; however, the sensitivity analysis revealed that phenytoin might prevent early PTS. Additional studies with large sample sizes and a rigorous design are required to obtain high-quality evidence on prophylactic anticonvulsant drug use in patients with traumatic brain injury.

Introduction

Traumatic brain injury (TBI) is a serious public health problem worldwide. Approximately 2.8 million TBI-related emergency department visits, hospitalizations, and deaths occurred in the United States in 2013 [1]. Patients with head injuries can develop post-traumatic seizures (PTS), which vary in severity and affect up to 56% of cases [2–4], and these seizures may adversely affect the prognosis of functional outcomes [5, 6].

The Brain Trauma Foundation Guidelines for the Management of Severe Traumatic Brain Injury (Fourth Edition) recommend pharmacological therapy, particularly phenytoin, to reduce the incidence of early but not late PTS [7]. A pairwise meta-analysis in 2012 reported that phenytoin was not superior to levetiracetam in preventing early posttraumatic attacks [8]. Another meta-analysis published in 2015 indicated that traditional antiepileptic drugs (AEDs), namely phenytoin and carbamazepine, may reduce the risk of early PTS compared with placebo or standard care, and that AEDs were not more effective than placebo or standard care in relieving late PTS [9]. Because previous studies have produced inconsistent findings, the clinical application of these results has thus been restricted. In addition, one more randomized controlled trial (RCT) has been published since the two meta-analyses [10]. Conducting an updated and sophisticated meta-analysis to address relevant research questions is therefore of clinical relevance.

Network meta-analysis (NMA) is an advanced method that can simultaneously compare multiple treatments by estimating direct and indirect treatment effects, thereby providing physicians with more useful clinical information regarding treatment selection. In this systematic review, we used NMA to compare the efficacy of different AEDs in treating early and late PTS following TBI.

Materials and methods

We performed a systematic review and NMA in accordance with the Preferred Reporting Items for Systematic Review and Network Meta-analysis checklist [11], which was developed as a guideline for reporting research outcomes. The study protocol was registered at PROSPERO (CRD42020172968).

Search strategy and study selection

We searched PubMed, Embase, Scopus, and the Cochrane Central Register of Controlled Trials from database inception to July 13, 2021. The keyword combinations used were as follows: “traumatic brain injury” AND (“seizure” OR “post trauma seizure”) AND “anticonvulsant drugs.” Search strategies for each database are provided in S1 Table [12]. We included RCTs enrolling adults or children with TBI and comparing the efficacy of an AED with another AED or placebo in preventing early and late PTS. “Early PTS” refers to the occurrence of seizures within 1 week of head injury, and “late PTS” refers to seizures occurring 1 week or later after head trauma. We also searched Clinical Trials (www.clinicaltrials.gov) for ongoing trials, and reviewed reference lists from included and other related studies. No limitations in language were applied. We searched relevant articles from inception until July 13, 2021. AEDs of different doses (flexible or fixed) were considered identical treatments. Two reviewers (B.C.W. and C.J.L.) independently screened all titles and abstracts according to the inclusion criteria. Any disagreements were resolved through discussion with a third reviewer (P.Y.C.).

Data extraction and methodological assessment of study quality

Two authors (B.C.W and C.J.L.) independently extracted the data from each included study by using a predesigned form. The extracted data included the name(s) of the author(s), year of publication, study country, age, male percentage, sample size, drug use details (type, dosage, and frequency), AED prophylaxis, duration of follow-up, and outcomes (early and late PTS). Disagreements were resolved through consensus.

Two reviewers (B.C.W. and C.J.L) independently evaluated the risk of bias for each study by using version 2 of the Cochrane tool for assessing risk of bias in randomized trials (RoB 2.0) [13]. Five key domains of bias were evaluated: (1) bias arising from the randomization process; (2) bias due to deviations from intended interventions; (3) bias due to missing outcome data; (4) bias in measurement of the outcome; and (5) bias in selection of the reported result [13]. Responses to each domain resulted in an overall risk of bias for each study, which was judged to be low, high, or of some concern. Any discrepancies were resolved through discussion leading to consensus.

Data analysis

All statistical analyses were performed using the STATA statistical software package (version 14). Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for early and late PTS. DerSimonian and Laird random-effects models [14] were used to perform conventional pairwise meta-analyses to directly compare any two AEDs. The Cochrane Q test, with statistical significance set at p < 0.1, was used to test between-study heterogeneity. Heterogeneity across head-to-head trials was assessed using the I2 statistic, with values > 50% roughly indicating substantial heterogeneity [15]. Publication bias was assessed using Egger’s test [16], with the statistical significance level set at 5%. The NMA was conducted under a frequentist framework with a random-effects model (using the suite of network commands written by Ian White) [17]. We assessed statistical inconsistency by using the loop-specific approach, side-splitting model, and design-by-treatment interaction model [17]. We ranked probabilities for each treatment, and the surface under the cumulative ranking (SUCRA) curve was employed to facilitate the evaluation of relative treatment efficacy [16, 18]. A higher SUCRA score (range, 0%–100%) indicated a greater likelihood of therapy in the top rank or one of the top ranks [19]. Sensitivity analysis was performed by excluding the study by Young et al. [20], which enrolled only patients younger than 10 years old.

We evaluated the proportion of direct and indirect evidence contributing to each comparison by using the direct evidence plot. According to König et al. [21], minimal parallelism and mean path length are used to estimate the degree of indirectness in the reported pooled outcome, with lower values for minimal parallelism and mean path length > 2 indicating that results for a specific comparison should be interpreted with caution because additional similarity assumptions must be made when each direct comparison is serially combined. Transitivity, which refers to the absence of systematic differences between the comparisons of interest other than the treatments being compared, is a crucial assumption in NMA [22]. Transitivity was qualitatively assessed by comparing the distribution of potential effect modifiers, including age and male percentage, across various comparisons in the network; if substantial variation between comparisons on effect modifiers was identified, intransitivity was suggested [22].

We used the confidence in network meta-analysis (CINeMA) web application, which is based on the Grading of Recommendations, Assessment, Development and Evaluation method, to assess the certainty of evidence produced by the synthesis for outcomes of early and late PTS. The CINeMA framework accounts for six domains that affect confidence levels in NMA results: within-study bias, reporting bias, indirectness, imprecision, heterogeneity, and incoherence. Each domain is rated as “no concerns,” “some concerns”,” or “major concerns,” except for reporting bias, which is graded as “suspect” or “undetected.” Judgments are then summarized across the domains as “high,” “moderate,” “low,” and “very low” for each treatment comparison [23, 24]. For imprecision, the threshold was set at an OR of 1.05 for outcome comparisons after discussion.

Results

Study selection and characteristics

A total of 749 articles were identified and retrieved using the described search methodology, with two more articles retrieved manually after the reference lists of other relevant review articles had been reviewed. All articles were then imported into Endnote X7. Using the Endnote X7 “Find duplicate” function, we excluded 206 duplicate articles. The titles and abstracts of the remaining 543 articles were examined and then screened against the predetermined inclusion and exclusion criteria. After all ineligible articles were excluded, 10 articles were reviewed in full text. One of the 10 articles was excluded because it compared phenobarbital use with usual care, which could not connect with the network map [25]. Two additional articles were identified through a reference list search of the included studies. Eventually, 11 articles with 2,450 head injury cases were included for the analyses (Fig 1). Of these, seven articles were from the United States [20, 25–30], one was from the United Kingdom [31], one was from Germany [26], one was from France [27], and one was from Pakistan [32]. Among the included articles, five studies enrolled participants with severe TBI and four studies recruited patients with moderate to severe TBI. Most of the studies (n = 8) included more male patients. Most studies enrolled patients with a wide age range (toddlers/adolescents to older patients), except for the study of Young et al. [20], which enrolled only patients younger than 10 years old.

Fig 1. Flow diagram of included studies.

Fig 1

Open in a new tab

Among the included articles, six articles were related to the prevention of both early and late PTS [26–28, 33–35], three articles were related to the prevention of only early PTS [20, 29, 32], and the remaining two articles were related to the prevention of only late PTS [30, 31]. Nine articles compared phenytoin to either placebo [20, 27, 29–31, 34] or other AEDs [32, 33, 35], and one article each compared carbamazepine [26] and magnesium sulfate (MgSO4) [28] to placebo (S2 Table).

The transitivity of potential effect modifiers (age and male percentage) is presented in S1 Fig. Their distributions were not significantly different across treatments, except for Young’s (2004) study [20] for young age and Khan’s (2016) study [32] for low male percentage, indicating potential threats to the transitivity assumption.

Network plots

Regarding early PTS, head-to-head comparisons across nine RCTs comprising six treatments (i.e., placebo, carbamazepine phenytoin, levetiracetam, valproate, and MgSO4) and 2,071 patients were performed using NMA (Fig 2A). The network plot of head-to-head comparisons of AEDs for late PTS included eight RCTs comprising six treatments (i.e., placebo, carbamazepine, phenytoin, levetiracetam, valproate, and MgSO4) and 1,788 patients (Fig 2B). In the two network maps, the comparison between placebo and PTH was the most compared treatment arm. Because the two network plots did not form a closed loop, inconsistency was not evaluated.

Fig 2.

Fig 2

Open in a new tab

(A). Network maps for comparing different AEDs on preventing early post-traumatic seizures. (B). Network maps for comparing different AEDs on preventing late post-traumatic seizures.

Effects of AEDs on early and late PTS

S3 Table presents the results of the pairwise meta-analysis and heterogeneity estimates. In terms of early PTS, only two pairwise interventions compared with placebo groups had two or more studies: phenytoin vs. placebo and levetiracetam vs. phenytoin. The remaining pairwise interventions included single studies. No significant difference was observed between treatment comparisons, except for carbamazepine vs. placebo (OR = 0.29, 95% CI = 0.12–0.71, n = 1), indicating a greater reduction in attenuated early PTS. For late PTS, only the pairwise comparison between phenytoin and placebo had four studies. No AEDs had any significant treatment effects. Next, we compared the combined available AEDs with placebo for early and late PTS; the pooled ORs were 0.42 and 0.82, respectively, estimated using a random-effects model (95% CI = 0.21–0.82 and 0.47–1.43, S3 Table). This indicates that AEDs as a whole significantly reduced early PTS but not late PTS compared with placebo.

Regarding the results of NMA on early PTS (S4 Table), carbamazepine (OR = 0.29, 95% CI = 0.07–1.18), phenytoin (OR = 0.43, 95% CI = 0.18–1.01), and levetiracetam (OR = 0.56, 95% CI = 0.12–2.55) were more likely to reduce early PTS compared with placebo; however, the effects were not significant (S4 Table). Furthermore, no AED was superior to another. Heterogeneity was assessed using the I2 test (29.5%, p = 0.16) and tau-squared test (0.30). The SUCRA analysis (S2 Fig) indicated that carbamazepine ranked first (82.3%), followed by phenytoin (74.4%) and levetiracetam (57.2%). The Egger’s test for studies with small sample sizes returned a p-value of 0.457, indicating no publication bias.

Compared with the placebo, carbamazepine (OR = 0.64, 95% CI = 0.08–5.28) and phenytoin (OR = 0.71, 95% CI = 0.23–2.20) were more likely to reduce the risk of late PTS, but the treatment effects were nonsignificant (S4 Table). No one AED was superior to any other. Heterogeneity between studies was observed (I2 = 72.7%, p = 0.02; tau squared = 0.60). The SUCRA analysis (S2 Fig) indicated that phenytoin ranked the highest (62.7%), followed by carbamazepine (60.0%) and MgSO4 (48.0%). The Egger’s test for studies with small sample sizes returned a p-value of 0.24, indicating no publication bias.

Of the 15 unique treatment comparisons, only 5 contained entire direct evidence (proportion = 100%), whereas the remaining 10 were based entirely on indirect evidence (S3 Fig). For all estimates, the minimum number of independent paths contributing to the effect size estimate on an aggregated level (minimal parallelism) was 1, implying less robust estimates. We observed mean path lengths >2 in 4 of the 15 estimates; this implies that the estimate in question should be interpreted with caution.

Sensitivity analysis

After the study of Young et al., which involved only patients younger than 10 years, was excluded [20], phenytoin significantly reduced the occurrence of early PTS compared with placebo (OR = 0.33; 95% CI = 0.14–0.78; S4 Table).

Adverse events

Three of the included studies reported the occurrence of any adverse events [31, 33, 34] and two of them reported skin rash after the use of AEDs [31, 34]. The occurrence of any adverse events was 13.9% with phenytoin, 9.42% with placebo, and 6.48% with valproate, and that of skin rash was 10.3% with phenytoin and 6.5% with placebo. Mortality was the most frequently documented adverse event (8 of 10 studies). Levetiracetam was associated with the highest mortality rate (41.18%), followed by carbamazepine (36.0%), MgSO4 (18.03%), placebo (16.69%), phenytoin (13.94%), and valproate (12.96%). Because of the limited number of studies reporting any adverse events and skin rash, comparative NMA was performed only for mortality. As presented in S4 Fig, the higher mortality rate was comparable between AEDs.

Risk of bias assessment and confidence in evidence

We identified four studies with a high risk of overall bias [26, 27, 29, 30]. Regarding the randomization process of the included studies, one study was graded as being at high risk of bias because its allocation was based on participant admission on either odd or even days [27], and four were graded as being at unknown risk because their methods of allocation concealment were not clearly described [26, 29, 30, 34]. Regarding bias due to deviation from intended interventions, one study was rated as high risk because research personnel may have been aware of the group assignment [27] (S5 Table).

The grading of the comparisons with CINeMA revealed mainly low to very low confidence ratings. This was due to concerns about within-study bias due to poor reporting of the randomization and blinding procedures. Evidence of imprecision was noted, likely because of the low number of trials available for comparison. Owing to the star-shaped network, the evidence quality in the domain of incoherence bias was downgraded (S6 Table).

Discussion

In this NMA, we included 11 studies enrolling 2,450 patients, which compared various prophylactic AEDs for preventing early and late PTS in patients with TBI. Our NMA findings indicated that no AED was superior to any other in preventing early or late PTS. However, our sensitivity analysis revealed that phenytoin might reduce early PTS compared with placebo, which is consistent with the Brain Trauma Foundation Guidelines for the Management of Severe Traumatic Brain Injury (Fourth Edition) [7]. Notably, the results of our pairwise meta-analysis are partially similar to those of a pairwise meta-analysis by Thompson et al. (2015), who compared the efficacy of traditional AEDs in preventing early PTS following head injury with that of a placebo. Their results indicated that AEDs, namely carbamazepine, phenobarbital, phenytoin, levetiracetam, and valproate, effectively reduced the occurrence of early PTS following head injury compared with placebo. However, no statistically significant differences were observed when phenytoin and other AEDs, including levetiracetam and valproate, were compared [9]. The results of a meta-analysis by Zafar and colleagues in 2012 demonstrated no statistically significant differences between the prophylactic use of phenytoin and that of levetiracetam in reducing the occurrence of early PTS following head injury, further supporting our study results [8]. The inconsistency in the results of the two studies may be because the AEDs were pooled together using pairwise meta-analyses in both studies. The two pairwise meta-analyses [8, 9] pooled relevant studies together, thereby substantially increasing the statistical power. By contrast, NMA, which our study adopted, may have led to few treatment comparison studies, causing reduced statistical power [36]. Our study results should therefore be interpreted with caution, with future studies required to verify our findings.

According to the results of the sensitivity analysis and transitivity assessment, age potentially affects the treatment effect of anticonvulsant drugs on PTS. We observed that after the study of Young et al. [20], phenytoin exhibited higher efficacy in preventing early PTS. The influence of age on the development of PTS has been debated, with some studies suggesting that young children (<2 years old) are more likely to develop PTS than older children [37–39] and some rejecting the association [40]. In fact, pediatric PTS is still not well understood, and studies on the effects of AEDs on pediatric PTS are still scarce. Further studies are warranted to explore how age affects the efficacy of AEDs in patients with PTS.

Similar to the results of previous conventional meta-analyses [8, 9], this study revealed that treatment with each AED compared to other AEDs or placebo could not effectively reduce the occurrence of late PTS following head injury. In clinical practice, patients often stop taking drugs or forget to take drugs because of remission of their symptoms, which leads to insufficient drug concentrations in their blood [41]. However, most studies included in our analysis did not clearly describe how patients were followed for medication use after their discharge from the hospital, including confirming patients’ medication compliance, monitoring the drug concentration in blood, and adjusting drug dosages. As a result, the AEDs could have failed to effectively prevent the occurrence of late PTS following head injury.

Several limitations should be considered when interpreting our data. First, the study included patients with a wide age range (from children to older patients), which may have reduced the internal validity of the study. Second, only a few studies were included for single study arm comparison in the NMA, thus limiting the statistical power. Third, the CINeMA results and direct evidence proportion plots suggested that our evidence has low or very low quality and less reliable estimates; therefore, our data should be interpreted with caution. Finally, not all included studies provided clear details of medication administration, including loading dose, maintenance dose, and monitoring and target ranges of drug concentrations in blood. As a result, the therapeutically effective dose of the drugs could not be correctly estimated in the included studies, which could have caused either persistent epilepsy due to poor efficacy resulting from insufficient drug concentrations in blood or adverse reactions due to drug overdose.

Conclusion

Our pairwise MA suggests that AEDs as a whole are superior to placebo in preventing early PTS; however, there is currently no strong evidence to support any specific medication in our NMA. Phenytoin has the largest body of evidence for its efficacy and performed well in this meta-analysis despite the lack of statistical significance. Our results support the treatment guidelines regarding the prophylactic use of phenytoin in preventing early PTS, but the optimal medication is unclear. PTS is a possible complication that affects patients with head injuries. Clinicians should be alerted to our findings, and future high-quality RCTs are warranted to examine the effects of AEDs on early and late PTS, especially in children.

Supporting information

S1 Fig. The transitivity analysis of potential effect modifiers.

(A) age and (B) male percentage. 1 indicates Glotzner et al., (1983) study; 2 indicates Pechadre et al., (1991) study; 3 indicates Szaflarski et al., (2010) study; 4 indicates Temkin et al., (1990) study; 4 indicates Temkin et al., (1999) study; 6 indicates Temkin et al., (2007) study; 7 indicates Young et al., (1983) study; 8 indicates Young et al., 2004) study; 9 indicates Khan et al., (2016) study. CBZ = carbamazepine, LEV = levetiracetam, VPA = valproate, PBO = placebo, PTH = phenytoin, MgSO = Magnesium Sulfate (MgSO4).

(TIF)

Click here for additional data file. (943.3KB, tif)
S2 Fig. The surface under the cumulative ranking of anticonvulsant drugs according to early and late post-traumatic seizure.

The surface under the cumulative ranking curve value is the probability each intervention has of being among the best in the network, with larger values representing higher ranking probabilities.

(TIF)

Click here for additional data file. (1.1MB, tif)
S3 Fig

Direct evidence plots for (A) early post-traumatic seizure and (B) late post-traumatic seizure. This figure shows the proportion of direct evidence (orange shade) in overall reporting of network value (blue shade). Minimal parallelism: Bar chart displaying the minimum number of independent paths contributing to the effect estimate on an aggregated level. All interventions included in our analysis have values equal to one, highlighting that estimates had contributions that extended beyond single pair-wise comparisons. Mean path length. Bar chart displaying the mean path length, which characterizes the degree of indirectness of an estimate. Higher mean path lengths indicate less reliable estimates. Mean path length for indirect comparison was equal to or above 2, which points to a less reliable estimate.

(TIF)

Click here for additional data file. (1.3MB, tif)
S4 Fig. The forest plots of AEDs comparisons for mortality.

(TIF)

Click here for additional data file. (348.7KB, tif)
S1 Table. An example of search strategies.

(DOCX)

Click here for additional data file. (18.9KB, docx)
S2 Table. Characteristics of studies.

(DOCX)

Click here for additional data file. (20.6KB, docx)
S3 Table. Pairwise meta-analytic results for early and late posttraumatic seizures.

(DOCX)

Click here for additional data file. (18KB, docx)
S4 Table. League tables for comparing anticonvulsant drugs for early and late post-traumatic seizures.

(DOCX)

Click here for additional data file. (16.9KB, docx)
S5 Table. Risk of methodological bias score of the studies.

(DOCX)

Click here for additional data file. (17.9KB, docx)
S6 Table. CINeMA summary tables.

(DOCX)

Click here for additional data file. (21.2KB, docx)

Data Availability

All raw data files are available in the supplemental materials.

Funding Statement

This meta-analysis was supported by grants from the Chang Gung Medical Foundation (CMRPVVK0071) and Chang Gung Memorial Hospital, Keelung branch (CRRPG2K0041).” The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.Taylor CA, Bell JM, Breiding MJ, Xu L. Traumatic Brain Injury-Related Emergency Department Visits, Hospitalizations, and Deaths—United States, 2007 and 2013. MMWR Surveill Summ. 2017;66(9):1–16. Epub 2017/03/17. doi: 10.15585/mmwr.ss6609a1 ; PubMed Central PMCID: PMC5829835. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Frey LC. Epidemiology of posttraumatic epilepsy: a critical review. Epilepsia. 2003;44(s10):11–7. Epub 2003/09/27. doi: 10.1046/j.1528-1157.44.s10.4.x . [DOI] [PubMed] [Google Scholar]
  • 3.Mahler B, Carlsson S, Andersson T, Adelöw C, Ahlbom A, Tomson T. Unprovoked seizures after traumatic brain injury: A population-based case-control study. Epilepsia. 2015;56(9):1438–44. Epub 2015/09/04. doi: 10.1111/epi.13096 . [DOI] [PubMed] [Google Scholar]
  • 4.Tomkins O, Feintuch A, Benifla M, Cohen A, Friedman A, Shelef I. Blood-brain barrier breakdown following traumatic brain injury: a possible role in posttraumatic epilepsy. Cardiovasc Psychiatry Neurol. 2011;2011:765923. Epub 2011/03/26. doi: 10.1155/2011/765923 ; PubMed Central PMCID: PMC3056210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Ahmed S, Venigalla H, Mekala HM, Dar S, Hassan M, Ayub S. Traumatic Brain Injury and Neuropsychiatric Complications. Indian J Psychol Med. 2017;39(2):114–21. Epub 2017/05/19. doi: 10.4103/0253-7176.203129 ; PubMed Central PMCID: PMC5385737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Majidi S, Makke Y, Ewida A, Sianati B, Qureshi AI, Koubeissi MZ. Prevalence and Risk Factors for Early Seizure in Patients with Traumatic Brain Injury: Analysis from National Trauma Data Bank. Neurocrit Care. 2017;27(1):90–5. Epub 2016/12/22. doi: 10.1007/s12028-016-0363-6 . [DOI] [PubMed] [Google Scholar]
  • 7.Carney N, Totten AM, O’Reilly C, Ullman JS, Hawryluk GW, Bell MJ, et al. Guidelines for the Management of Severe Traumatic Brain Injury, Fourth Edition. Neurosurgery. 2017;80(1):6–15. Epub 2016/09/23. doi: 10.1227/NEU.0000000000001432 . [DOI] [PubMed] [Google Scholar]
  • 8.Zafar SN, Khan AA, Ghauri AA, Shamim MS. Phenytoin versus Leviteracetam for seizure prophylaxis after brain injury—a meta analysis. BMC Neurol. 2012;12:30. Epub 2012/05/31. doi: 10.1186/1471-2377-12-30 ; PubMed Central PMCID: PMC3406949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Thompson K, Pohlmann-Eden B, Campbell LA, Abel H. Pharmacological treatments for preventing epilepsy following traumatic head injury. Cochrane Database Syst Rev. 2015;(8):Cd009900. Epub 2015/08/11. doi: 10.1002/14651858.CD009900.pub2 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Younus SM, Basar S, Gauri SA, Khan AA, Imran M, Abubakar S, et al. Comparison of Phenytoin versus Levetiracetam in Early Seizure Prophylaxis after Traumatic Brain Injury, at a Tertiary Care Hospital in Karachi, Pakistan. Asian J Neurosurg. 2018;13(4):1096–100. Epub 2018/11/22. doi: 10.4103/ajns.AJNS_125_17 ; PubMed Central PMCID: PMC6208194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann Intern Med. 2015;162(11):777–84. Epub 2015/06/02. doi: 10.7326/M14-2385 . [DOI] [PubMed] [Google Scholar]
  • 12.Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. Epub 2021/03/31. doi: 10.1136/bmj.n71 ; PubMed Central PMCID: PMC8005924 http://www.icmje.org/conflicts-of-interest/ [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Sterne JAC, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898. Epub 2019/08/30. doi: 10.1136/bmj.l4898 . [DOI] [PubMed] [Google Scholar]
  • 14.DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177–88. Epub 1986/09/01. doi: 10.1016/0197-2456(86)90046-2 . [DOI] [PubMed] [Google Scholar]
  • 15.Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, et al. Cochrane handbook for systematic reviews of interventions: John Wiley & Sons; 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G. Graphical tools for network meta-analysis in STATA. PLoS One. 2013;8(10):e76654. Epub 2013/10/08. doi: 10.1371/journal.pone.0076654 ; PubMed Central PMCID: PMC3789683. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.White IR, Barrett JK, Jackson D, Higgins JP. Consistency and inconsistency in network meta-analysis: model estimation using multivariate meta-regression. Res Synth Methods. 2012;3(2):111–25. Epub 2012/06/01. doi: 10.1002/jrsm.1045 ; PubMed Central PMCID: PMC4433771. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J Clin Epidemiol. 2011;64(2):163–71. Epub 2010/08/07. doi: 10.1016/j.jclinepi.2010.03.016 . [DOI] [PubMed] [Google Scholar]
  • 19.Mbuagbaw L, Rochwerg B, Jaeschke R, Heels-Andsell D, Alhazzani W, Thabane L, et al. Approaches to interpreting and choosing the best treatments in network meta-analyses. Syst Rev. 2017;6(1):79. Epub 2017/04/14. doi: 10.1186/s13643-017-0473-z ; PubMed Central PMCID: PMC5389085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Young KD, Okada PJ, Sokolove PE, Palchak MJ, Panacek EA, Baren JM, et al. A randomized, double-blinded, placebo-controlled trial of phenytoin for the prevention of early posttraumatic seizures in children with moderate to severe blunt head injury. Ann Emerg Med. 2004;43(4):435–46. Epub 2004/03/25. doi: 10.1016/j.annemergmed.2003.09.016 . [DOI] [PubMed] [Google Scholar]
  • 21.König J, Krahn U, Binder H. Visualizing the flow of evidence in network meta-analysis and characterizing mixed treatment comparisons. Stat Med. 2013;32(30):5414–29. Epub 2013/10/15. doi: 10.1002/sim.6001 . [DOI] [PubMed] [Google Scholar]
  • 22.Rouse B, Chaimani A, Li T. Network meta-analysis: an introduction for clinicians. Intern Emerg Med. 2017;12(1):103–11. Epub 2016/12/04. doi: 10.1007/s11739-016-1583-7 ; PubMed Central PMCID: PMC5247317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Nikolakopoulou A, Higgins JPT, Papakonstantinou T, Chaimani A, Del Giovane C, Egger M, et al. CINeMA: An approach for assessing confidence in the results of a network meta-analysis. PLoS Med. 2020;17(4):e1003082. Epub 2020/04/04. doi: 10.1371/journal.pmed.1003082 ; PubMed Central PMCID: PMC7122720 following competing interests: ME is a member of the Editorial Board of PLOS Medicine. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Papakonstantinou T, Nikolakopoulou A, Higgins JPT, Egger M, Salanti G. CINeMA: Software for semiautomated assessment of the confidence in the results of network meta-analysis. Campbell Systematic Reviews. 2020;16(1):e1080. doi: 10.1002/cl2.1080 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Manaka S. Cooperative prospective study on posttraumatic epilepsy: risk factors and the effect of prophylactic anticonvulsant. Jpn J Psychiatry Neurol. 1992;46(2):311–5. Epub 1992/06/01. doi: 10.1111/j.1440-1819.1992.tb00865.x . [DOI] [PubMed] [Google Scholar]
  • 26.Glötzner FL, Haubitz I, Miltner F, Kapp G, Pflughaupt KW. [Seizure prevention using carbamazepine following severe brain injuries]. Neurochirurgia (Stuttg). 1983;26(3):66–79. Epub 1983/05/01. doi: 10.1055/s-2008-1053615 . [DOI] [PubMed] [Google Scholar]
  • 27.Pechadre JC, Lauxerois M, Colnet G, Commun C, Dimicoli C, Bonnard M, et al. [Prevention of late post-traumatic epilepsy by phenytoin in severe brain injuries. 2 years’ follow-up]. Presse Med. 1991;20(18):841–5. Epub 1991/05/11. . [PubMed] [Google Scholar]
  • 28.Temkin NR, Anderson GD, Winn HR, Ellenbogen RG, Britz GW, Schuster J, et al. Magnesium sulfate for neuroprotection after traumatic brain injury: a randomised controlled trial. Lancet Neurol. 2007;6(1):29–38. Epub 2006/12/15. doi: 10.1016/S1474-4422(06)70630-5 . [DOI] [PubMed] [Google Scholar]
  • 29.Young B, Rapp RP, Norton JA, Haack D, Tibbs PA, Bean JR. Failure of prophylactically administered phenytoin to prevent early posttraumatic seizures. J Neurosurg. 1983;58(2):231–5. Epub 1983/02/01. doi: 10.3171/jns.1983.58.2.0231 . [DOI] [PubMed] [Google Scholar]
  • 30.Young B, Rapp RP, Norton JA, Haack D, Tibbs PA, Bean JR. Failure of prophylactically administered phenytoin to prevent late posttraumatic seizures. J Neurosurg. 1983;58(2):236–41. Epub 1983/02/01. doi: 10.3171/jns.1983.58.2.0236 . [DOI] [PubMed] [Google Scholar]
  • 31.McQueen JK, Blackwood DH, Harris P, Kalbag RM, Johnson AL. Low risk of late post-traumatic seizures following severe head injury: implications for clinical trials of prophylaxis. J Neurol Neurosurg Psychiatry. 1983;46(10):899–904. Epub 1983/10/01. doi: 10.1136/jnnp.46.10.899 PubMed Central PMCID: PMC1027602. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Khan SA, Bhatti SN, Khan AA, Khan Afridi EA, Muhammad G, Gul N, et al. Comparison Of Efficacy Of Phenytoin And Levetiracetam For Prevention Of Early Post Traumatic Seizures. J Ayub Med Coll Abbottabad. 2016;28(3):455–60. Epub 2017/07/18. . [PubMed] [Google Scholar]
  • 33.Temkin NR, Dikmen SS, Anderson GD, Wilensky AJ, Holmes MD, Cohen W, et al. Valproate therapy for prevention of posttraumatic seizures: a randomized trial. J Neurosurg. 1999;91(4):593–600. Epub 1999/10/03. doi: 10.3171/jns.1999.91.4.0593 . [DOI] [PubMed] [Google Scholar]
  • 34.Temkin NR, Dikmen SS, Wilensky AJ, Keihm J, Chabal S, Winn HR. A randomized, double-blind study of phenytoin for the prevention of post-traumatic seizures. N Engl J Med. 1990;323(8):497–502. Epub 1990/08/23. doi: 10.1056/NEJM199008233230801 . [DOI] [PubMed] [Google Scholar]
  • 35.Szaflarski JP, Sangha KS, Lindsell CJ, Shutter LA. Prospective, randomized, single-blinded comparative trial of intravenous levetiracetam versus phenytoin for seizure prophylaxis. Neurocrit Care. 2010;12(2):165–72. Epub 2009/11/10. doi: 10.1007/s12028-009-9304-y . [DOI] [PubMed] [Google Scholar]
  • 36.Thorlund K, Mills EJ. Sample size and power considerations in network meta-analysis. Syst Rev. 2012;1:41. Epub 2012/09/21. doi: 10.1186/2046-4053-1-41 ; PubMed Central PMCID: PMC3514119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Bennett KS, DeWitt PE, Harlaar N, Bennett TD. Seizures in Children With Severe Traumatic Brain Injury. Pediatr Crit Care Med. 2017;18(1):54–63. Epub 2016/09/23. doi: 10.1097/PCC.0000000000000948 ; PubMed Central PMCID: PMC5218863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Ong LC, Dhillon MK, Selladurai BM, Maimunah A, Lye MS. Early post-traumatic seizures in children: clinical and radiological aspects of injury. J Paediatr Child Health. 1996;32(2):173–6. Epub 1996/04/01. doi: 10.1111/j.1440-1754.1996.tb00917.x . [DOI] [PubMed] [Google Scholar]
  • 39.Elsamadicy AA, Koo AB, David WB, Lee V, Zogg CK, Kundishora AJ, et al. Post-traumatic seizures following pediatric traumatic brain injury. Clin Neurol Neurosurg. 2021;203:106556. Epub 2021/02/27. doi: 10.1016/j.clineuro.2021.106556 . [DOI] [PubMed] [Google Scholar]
  • 40.Keret A, Bennett-Back O, Rosenthal G, Gilboa T, Shweiki M, Shoshan Y, et al. Posttraumatic epilepsy: long-term follow-up of children with mild traumatic brain injury. J Neurosurg Pediatr. 2017;20(1):64–70. Epub 2017/05/06. doi: 10.3171/2017.2.PEDS16585 . [DOI] [PubMed] [Google Scholar]
  • 41.Faught E. Adherence to antiepilepsy drug therapy. Epilepsy Behav. 2012;25(3):297–302. Epub 2012/10/27. doi: 10.1016/j.yebeh.2012.08.027 . [DOI] [PubMed] [Google Scholar]

Decision Letter 0

Giuseppe Biagini

4 May 2021

PONE-D-21-09280

Efficacy and safety of prophylactic anticonvulsant drug following traumatic brain injury: a systematic review and network meta-analysis of randomized controlled trials

PLOS ONE

Dear Dr. Chen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Jun 18 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Giuseppe Biagini, MD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include your tables as part of your main manuscript and remove the individual files. Please note that supplementary tables should remain uploaded as separate "supporting information" files.

3. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This manuscript is a network meta-analysis (NMA) of randomised controlled trials performed to evaluate the effect of prophylactic anticonvulsant therapy in patients with traumatic brain injury. The topic is of interest and certainly an area of equipoise in current practice. Current guidelines suggest the use of empiric anticonvulsant therapy in patients presenting with severe traumatic brain injury for up to one week, but practice has been demonstrated in surveys to vary widely among neurosurgeons. The authors’ analysis of the twelve included studies suggests no benefit to any individual antiepileptic drug.

Overall, this is a useful and important topic with interesting implications that might be somewhat answered using an NMA. However, there are methodological flaws which should be addressed before publication can be considered.

Statistical analysis: Lacking in detail, see comments below. However, the treatment ranking described by the authors is reproducible from the data provided.

English language/grammar: Requires further proof-reading.

Major comments:

1. Firstly, the authors should consider whether an NMA is appropriate. I note that, of the 5 drugs studied for the prevention of early seizures for example, 4 were compared in only a single study arm which limits power somewhat. An NMA, if performed, should be very carefully evaluated for integrity in such an instance.

2. How were early and late seizures defined? This is important for interpretation and also ensuring transitivity in the NMA and as such should be clearly stated in the methods section.

3. Transitivity is perhaps the most crucial assumption of the NMA and should be defined and assessed by the authors. This should be explained in the methods section.

4. The exact search strategy used should be provided (as supplementary) for each database in such a way that it is fully reproducible. This should ideally comply with the PRISMA-P statement.

5. The authors report to have assessed risk of bias in individual studies using the Cochrane Handbook, providing a reference from 2011. Risk of bias in randomised trials is more appropriately assessed using an updated, validated tool such as the RoB 2.0.

6. The authors report to have performed pairwise meta-analysis in addition to NMA for each pairing, but do not appear to have reported the results. Pairwise analyses should be reported separately from and clearly delineated from the network estimates derived from the NMA, perhaps in league table form.

7. The authors mention that random effects regression was applied to the pairwise analyses. Was a random effects model also used for the network estimates? This should be clarified.

8. The authors mention the calculation of a SUCRA treatment ranking and SUCRA ranking positions are alluded to in the manuscript but the treatment ranking for each category studied is not overtly provided. These rankings should be included in tabular form for each analysis with the exact SUCRA scores provided. A cumulative ranking plot may also be of benefit to illustrate.

9. SUCRA is a useful metric in the NMA but NMA is a relatively novel concept with which many readers will be unfamiliar and its calculation, implications and interpretations should therefore be explained in the methods.

10. On page 8, the authors mention that Egger’s p < 0.05 indicated no significant publication bias. This is incorrect - p < 0.05 indicates significant asymmetry and suggests the presence of publication bias.

11. The full results of the NMA including effect size estimates with associated confidence intervals and SUCRA scores should be reported for each analysis separately in the text and in the abstract.

12. The study by Manaka (1992) refers to a control group consisting of “usual care”, but the authors mention that some patients received anticonvulsants. There is no node relating to “usual care” in the network plots – was this group considered equivalent to placebo control? If so, this should be overtly stated as this may violate the transitivity assumption of the NMA.

13. The authors mention performing meta-regression in the methods (page 6) but no further detail was provided. If this was performed, please provide details of the statistical methods and software used for this in the methods along with results including regression coefficients, residual heterogeneity and significance levels in the results section.

14. The exploration of inconsistency is mentioned. Where available, the results of the node-splitting analysis performed should be provided for each comparison in the network. Heterogeneity within designs and between designs should be clearly quantitatively reported for each analysis.

15. The forest plots are currently rather illegible. The axes should be decompressed as the numbers are currently unreadable and the treatment identifiers (A, B, etc) should probably just be substituted for the full treatment names or three-letter identifiers for each treatment eg. PHT for phenytoin. In addition, the legend is absent from figure 5.

16. The authors conclude by stating that clinicians should incorporating their findings, that no anticonvulsant drug significantly decreases the likelihood of experiencing post-traumatic epilepsy, into clinical practice. I find this excessively assured. The authors did not attempt to pool all anticonvulsants into a pairwise anticonvulsants versus placebo meta-analysis, and a Cochrane review on the topic has demonstrated a lower likelihood of seizures when all anticonvulsants are considered together. It is possible, and even likely, that studies of individual treatments are underpowered to demonstrate an effect but when pooled demonstrate that prophylactic anticonvulsant therapy may in fact decrease the likelihood of post-traumatic seizures. The authors refer to this on page 9-10, stating that the study has better understanding of the individual treatments as a result, but consideration should also be given to the increased statistical power achieved by pooling the individual therapies.

Minor comments:

1. The authors should be commended for reporting raw data in Table 1, but this should be reported separately also as supplementary, ideally in spreadsheet format to allow easier reproduction of the results.

2. I find the choice of skin rash as the only adverse effect to be investigated unconvincing. While it may be the most common, its clinical relevance in patients with severe traumatic brain injury is questionable and perhaps more serious adverse effects should also be considered despite being less common? If this is retained, it should be defined – skin rash associated with anticonvulsants ranges from mild cutaneous irritation to life-threatening dermatological syndromes.

3. There are some references missing in the discussion wherein authors are mentioned but no superscript reference is provided eg. by Zafar et al. on page 9.

4. The authors should consider evaluating mean path length(1) given that many treatments are represented by a single study arm.

References:

1. König J, Krahn U, Binder H. Visualizing the flow of evidence in network meta-analysis and characterizing mixed treatment comparisons. Stat Med. 2013 Dec 30;32(30):5414-29. doi: 10.1002/sim.6001. Epub 2013 Oct 4. PMID: 24123165.

Reviewer #2: In this interesting systematic review the authors aimed was to systematically compare the effects and safety of prophylactic anticonvulsant drug use in patient with traumatic brain injury. Network meta-analysis using multivariate random effects and frequentist framework was applied. A total of 12 randomized controlled trials involving 1,431 participants were included. The authors underline that this study suggests no beneficial effect of anticonvulsant drug on early and late post-traumatic seizure, which does not support the recommendations of the clinical guideline of using Phenytoin as the first-line therapy in treatment of early post-traumatic seizure.

The manuscript is well written and it could have a relevant impact on the readership. However, there are few minor points that need to be clarified.

Specific comments:

- The sentence reported in the abstract “In comparison with placebo, phenytoin, valproate, levetiracetam, carbamazepine, and MgSO4 could significantly reduce the early post-traumatic seizure; and phenytoin, valproate, levetiracetam, carbamazepine, MgSO4, and phenobarbital were not significantly superior to placebo.” is not clear; please reformulate it.

- page 9: please change “seizure” with “seizures”

- Please check the different values included in the flow diagram of the study selection process.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Jack Henry

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2022 Mar 31;17(3):e0265932. doi: 10.1371/journal.pone.0265932.r002

Author response to Decision Letter 0

16 Sep 2021

Reviewer #1: This manuscript is a network meta-analysis (NMA) of randomised controlled trials performed to evaluate the effect of prophylactic anticonvulsant therapy in patients with traumatic brain injury. The topic is of interest and certainly an area of equipoise in current practice. Current guidelines suggest the use of empiric anticonvulsant therapy in patients presenting with severe traumatic brain injury for up to one week, but practice has been demonstrated in surveys to vary widely among neurosurgeons. The authors’ analysis of the twelve included studies suggests no benefit to any individual antiepileptic drug. Overall, this is a useful and important topic with interesting implications that might be somewhat answered using an NMA. However, there are methodological flaws which should be addressed before publication can be considered.

Response: Thank you for your thoughtful review. We have carefully studied your comments and suggestions and revised our paper accordingly. The following are our point-by-point responses to your specific comments. We hope our revisions are acceptable and our responses adequately address the comments. Thank you for your consideration.

Statistical analysis: Lacking in detail, see comments below. However, the treatment ranking described by the authors is reproducible from the data provided.

Response: Thank you for the comment. We have revised the statistical analysis according to your suggestion.

English language/grammar: Requires further proof-reading.

Response: Thank you for the comment. We have had the manuscript reviewed by a native English speaker with scientific expertise to ensure proper English grammar and usage.

Major comments:

1. Firstly, the authors should consider whether an NMA is appropriate. I note that, of the 5 drugs studied for the prevention of early seizures for example, 4 were compared in only a single study arm which limits power somewhat. An NMA, if performed, should be very carefully evaluated for integrity in such an instance.

Response: Thank you for the thoughtful comment. We agree with your concerns that the NMA may have caused reduced statistical power, thus threatening the internal validity of our findings. However, compared with traditional pairwise meta-analysis, the strength of the NMA lies in comparing different AEDs simultaneously. Even so, we have emphasized this concern in the section of Discussion and as one of the study limitations as follows: It is worth noting that the two pairwise meta-analyses [8, 9] that pooled relevant studies together may have substantially increased statistical power by doing so. By contrast, NMA, which our study adopted, may have led to few treatment comparison studies, causing reduced statistical power [33]. Our study results should therefore be interpreted with caution, with future studies required to verify our findings (Discussion section); Several limitations should be considered when interpreting our data. First,…Second, only a few studies were included for single study arm comparison, thus limiting the statistical power.…(Limitations)

2. How were early and late seizures defined? This is important for interpretation and also ensuring transitivity in the NMA and as such should be clearly stated in the methods section.

Response: We agree with your comment that clear definitions of early and late seizures are critical for interpretation and ensuring transitivity in the NMA. We have added the definition in the Material and Methods section as follows (P.4): “Early PTS” refers to the occurrence of seizures within 1 week of head injury, and “late PTS” refers to seizures occurring 1 week or later after head trauma.

3. Transitivity is perhaps the most crucial assumption of the NMA and should be defined and assessed by the authors. This should be explained in the methods section.

Response: We have defined and assessed transitivity, the most crucial assumption of the NMA in the methods section as follow (p 7): Transitivity, which refers to an absence of systematic differences between the available comparisons other than the treatments being compared, is a crucial assumption for an NMA [20]. Transitivity was qualitatively assessed by the authors; if a substantial variation between comparisons on effect modifiers was identified, intransitivity was suggested [20]. We have also qualitatively described the details of the included studies to confirm the issue of transitivity. We observed that age variations may cause intransitivity; hence, sensitivity analysis was further performed. All details can be found in the Results section (page 8 to 9).

4. The exact search strategy used should be provided (as supplementary) for each database in such a way that it is fully reproducible. This should ideally comply with the PRISMA-P statement.

Response: Thank you for the comments. We have provided the exact search strategy (as Supplementary table 1) for each database, which ideally complies with the PRISMA-P statement.

5. The authors report to have assessed risk of bias in individual studies using the Cochrane Handbook, providing a reference from 2011. Risk of bias in randomised trials is more appropriately assessed using an updated, validated tool such as the RoB 2.0.

Response: Thank you for the comments. We have reassessed the study quality using RoB 2.0. All changes are marked in red in both the Material and Methods and the Results sections.

6. The authors report to have performed pairwise meta-analysis in addition to NMA for each pairing, but do not appear to have reported the results. Pairwise analyses should be reported separately from and clearly delineated from the network estimates derived from the NMA, perhaps in league table form.

Response: Thank you for the comments. We did perform the pairwise meta-analysis before the NMA. We have added the pairwise analyses in a league table format.

7. The authors mention that random effects regression was applied to the pairwise analyses. Was a random effects model also used for the network estimates? This should be clarified.

Response: Thank you for the comments. This was a typo. We have deleted the sentence describing random effects regression and noted the random-effects model in the Data analysis section as follows: The network meta-analysis was conducted under a frequentist framework with a random-effects model. (Page 7)

8. The authors mention the calculation of a SUCRA treatment ranking and SUCRA ranking positions are alluded to in the manuscript but the treatment ranking for each category studied is not overtly provided. These rankings should be included in tabular form for each analysis with the exact SUCRA scores provided. A cumulative ranking plot may also be of benefit to illustrate.

Response: Thank you for your suggestions. We have added SUCRA plots with exact SCURA scores for each outcome in Supplementary Figure S1.

9. SUCRA is a useful metric in the NMA but NMA is a relatively novel concept with which many readers will be unfamiliar and its calculation, implications and interpretations should therefore be explained in the methods.

Response: Thank you for your suggestions. We have added the implications and interpretations of the SUCRA score in the section of Data Analysis as follows: We ranked probabilities for each treatment, and the surface under the cumulative ranking (SUCRA) curve was employed to facilitate the evaluation of relative treatment efficacy [16, 18]. A higher SUCRA score (range, 0%–100%) indicated a greater likelihood of therapy in the top rank or one of the top ranks [19].

10. On page 8, the authors mention that Egger’s p < 0.05 indicated no significant publication bias. This is incorrect - p < 0.05 indicates significant asymmetry and suggests the presence of publication bias.

Response: It was a typo. We have provided the exact p-value indicative of no significant publication bias.

11. The full results of the NMA including effect size estimates with associated confidence intervals and SUCRA scores should be reported for each analysis separately in the text and in the abstract.

Response: Thank you for your suggestion. We have added the effect size estimates with associated CIs and SUCRA scores for each outcome separately in the main text and abstract.

12. The study by Manaka (1992) refers to a control group consisting of “usual care”, but the authors mention that some patients received anticonvulsants. There is no node relating to “usual care” in the network plots – was this group considered equivalent to placebo control? If so, this should be overtly stated as this may violate the transitivity assumption of the NMA.

Response: Thank you for your suggestion. After reviewing the article, we found that usual care arm is not equivalent to placebo, causing the study (Manaka [1992]) could bot connect with NMA map. We therefore removed it from final analysis.

13. The authors mention performing meta-regression in the methods (page 6) but no further detail was provided. If this was performed, please provide details of the statistical methods and software used for this in the methods along with results including regression coefficients, residual heterogeneity and significance levels in the results section.

Response: Thank you for your suggestions. We have deleted this sentence because we did not perform meta-regression due to the inclusion of small-size number of studies.

14. The exploration of inconsistency is mentioned. Where available, the results of the node-splitting analysis performed should be provided for each comparison in the network. Heterogeneity within designs and between designs should be clearly quantitatively reported for each analysis.

Response: Thank you for your suggestions. Because the network plots did not form a closed loop, the pairwise comparisons were based on direct or indirect evidence. That is, inconsistency was not evaluated for the networks. We have mentioned this point in the Network Plots section.

15. The forest plots are currently rather illegible. The axes should be decompressed as the numbers are currently unreadable and the treatment identifiers (A, B, etc) should probably just be substituted for the full treatment names or three-letter identifiers for each treatment eg. PHT for phenytoin. In addition, the legend is absent from figure 5.

Response: Thank you for the comments. To enhance the readability of our data, we have replaced the forest plots with league tables to present the results of treatment comparisons.

16. The authors conclude by stating that clinicians should incorporating their findings, that no anticonvulsant drug significantly decreases the likelihood of experiencing post-traumatic epilepsy, into clinical practice. I find this excessively assured. The authors did not attempt to pool all anticonvulsants into a pairwise anticonvulsants versus placebo meta-analysis, and a Cochrane review on the topic has demonstrated a lower likelihood of seizures when all anticonvulsants are considered together. It is possible, and even likely, that studies of individual treatments are underpowered to demonstrate an effect but when pooled demonstrate that prophylactic anticonvulsant therapy may in fact decrease the likelihood of post-traumatic seizures. The authors refer to this on page 9-10, stating that the study has better understanding of the individual treatments as a result, but consideration should also be given to the increased statistical power achieved by pooling the individual therapies.

Response: We agree with your comment; hence, we have attributed increased statistical power to the pairwise meta-analyses that pooled the individual therapies, and stated that our network meta-analysis involved only a few studies for each comparison in the Discussion section as follows: It is worth noting that the two pairwise meta-analyses [8, 9] that pooled relevant studies may have substantially increased statistical power by doing so. By contrast, NMA, which our study adopted, may have led to few treatment comparison studies, causing reduced statistical power [33]. Our study results therefore should be interpreted with caution, with future studies required to verify our findings.

Minor comments:

1. The authors should be commended for reporting raw data in Table 1, but this should be reported separately also as supplementary, ideally in spreadsheet format to allow easier reproduction of the results.

Response: Thank you for the suggestion. We have provided the raw data in spreadsheet format as a supplementary Table S3.

2. I find the choice of skin rash as the only adverse effect to be investigated unconvincing. While it may be the most common, its clinical relevance in patients with severe traumatic brain injury is questionable and perhaps more serious adverse effects should also be considered despite being less common? If this is retained, it should be defined – skin rash associated with anticonvulsants ranges from mild cutaneous irritation to life-threatening dermatological syndromes.

Response: Thank you for the suggestion. After rechecking the included studies, original authors did not define skin rash in the papers; therefore, we could not identify the effects of skin rash on TBI survivors. We have deleted all relevant context regarding the adverse effects of AEDs on skin rash from the manuscript.

3. There are some references missing in the discussion wherein authors are mentioned but no superscript reference is provided eg. by Zafar et al. on page 9.

Response: We have added citation numbers for the missing references.

4. The authors should consider evaluating mean path length (1) given that many treatments are represented by a single study arm.

Response: We have added the average path length to depict the numbers of treatments in a single study arm (Figure 2).

 

Reviewer #2: In this interesting systematic review the authors aimed was to systematically compare the effects and safety of prophylactic anticonvulsant drug use in patient with traumatic brain injury. Network meta-analysis using multivariate random effects and frequentist framework was applied. A total of 12 randomized controlled trials involving 1,431 participants were included. The authors underline that this study suggests no beneficial effect of anticonvulsant drug on early and late post-traumatic seizure, which does not support the recommendations of the clinical guideline of using Phenytoin as the first-line therapy in treatment of early post-traumatic seizure. The manuscript is well written and it could have a relevant impact on the readership. However, there are few minor points that need to be clarified.

Response: Thank you for your encouraging words. We have revised our manuscript according to your suggestions, and the following are our point-by-point responses to your specific comments. We hope that our revisions are acceptable and that our responses adequately address your comments. Thank you for your consideration.

Specific comments:

- The sentence reported in the abstract “In comparison with placebo, phenytoin, valproate, levetiracetam, carbamazepine, and MgSO4 could significantly reduce the early post-traumatic seizure; and phenytoin, valproate, levetiracetam, carbamazepine, MgSO4, and phenobarbital were not significantly superior to placebo.” is not clear; please reformulate it.

Response: We have revised the abstract as follows: We found that phenytoin (odd ratios [ORs] = 0.43 and 0.71; 95% confidence intervals [CIs] = 0.18–1.01 and 0.23–2.20), levetiracetam (ORs = 0.56 and 1.58; 95% CIs = 0.12–2.55 and 0.03–84.42), and carbamazepine (OR= 0.29 and 0.64; 95% CIs = 0.07–1.18 and 0.08–5.28) were more likely to reduce early and late PTS compared with placebo; however, the treatment effects were not significant.

- page 9: please change “seizure” with “seizures”

Response: Because the abbreviation of posttraumatic seizures (i.e., PTS) was defined upon first usage, we have replaced “posttraumatic seizures” with “PTS” throughout the manuscript.

- Please check the different values included in the flow diagram of the study selection process.

Response: We have rechecked and updated the values included in the flow diagram. All changes are marked in red.

Attachment

Submitted filename: response to reviewers comment-0905 HYC.docx

Click here for additional data file. (32.5KB, docx)

Decision Letter 1

Giuseppe Biagini

8 Oct 2021

PONE-D-21-09280R1Comparative efficacy of prophylactic anticonvulsant drugs following traumatic brain injury: a systematic review and network meta-analysis of randomized controlled trialsPLOS ONE

Dear Dr. Chen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Nov 22 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Giuseppe Biagini, MD

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors should be commended for a substantial review, which has significantly improved the manuscript. The plots are significantly improved and more legible, and the writing has significantly improved. I have the following comments:

Major comments:

1. The authors now provide a definition of transitivity as suggested. However, they mention that transitivity was assessed qualitatively only (methods, page 7-8). They do not appear to have defined specific potential effect modifiers, or collected arm-level data on these modifiers. This insufficient to assess the plausibility of transitivity, especially in the context of a star-shaped network, which limits the assessment of its statistical manifestation (inconsistency). Ideally, the authors should define potential effect modifiers and compare their distribution across treatment nodes in the network.

2. On this note, the network heterogeneity, quantified by tau^2, should also be reported. I am aware that, for the late seizure analysis in particular, this will be a skewed estimate given that most treatments are comprised of only a single study. However, some effort to quantify the heterogeneity is needed.

3. The authors now report to have included the odds ratios, including direct pairwise estimates, in league table format as Table 2. This does not appear to have been provided.

4. The authors have now removed the skin rash adverse event from the NMA assessment, which seems appropriate. However, there is now no assessment of adverse events or mortality, which is a significant limitation for any clinician reading, because this informs treatment choice as much as, or even more than efficacy. Ideally, the authors should record data on the number of adverse events recorded and the mortality rate between the various treatments evaluated. This may not be appropriate for network meta-analysis, but should at least be recorded, displayed and discussed.

5. The authors claim to now have provided the mean path length for comparisons. What is provided in the plots is in fact the number of studies forming the comparison, as far as I can see. The mean path length I refer to is as described by Konig et al. (2013), below.

6. It is not emphasised enough that anticonvulsants as a whole are superior to placebo for the prevention of early seizures. This has been strongly shown by the Cochrane review. In addition, if studies comparing an antiepileptic drug to placebo in the authors' provided data are meta-analysed (any AED vs. placebo) the summary result from a random effects model is OR 0.42, 95%CI 0.42, p = 0.01, I^2 = 37.5%, tau^2 = 0.25. To me, this is a central point that should be better emphasised by the authors. When this meta-analysis is also performed, the findings suggest that anticonvulsants as a whole are superior to placebo, but the various drugs appear to have equivalent efficacy.

7. In my opinion, the conclusion that PHT may be a beneficial treatment, without reference to the aggregate of anticonvulsant drugs, is rather unsupported as a result. This is especially given the case that phenytoin did not rank highest in the NMA, did not have superior efficacy to any of the individual drugs in the point estimates and was only superior to placebo in a seemingly arbitrary sensitivity analysis.

8. In line with recent updates to the PRISMA guidelines, the authors confidence in their findings should be discussed. In network meta-analysis, the CINeMA tool provides a framework and associated web application for this (referenced below).

König J, Krahn U, Binder H. Visualizing the flow of evidence in network meta-analysis and characterizing mixed treatment comparisons. Stat Med. 2013 Dec 30;32(30):5414-29. doi: 10.1002/sim.6001. Epub 2013 Oct 4. PMID: 24123165.

Nikolakopoulou A, Higgins JPT, Papakonstantinou T, Chaimani A, Del Giovane C, Egger M, Salanti G. CINeMA: An approach for assessing confidence in the results of a network meta-analysis. PLoS Med. 2020 Apr 3;17(4):e1003082. doi: 10.1371/journal.pmed.1003082. PMID: 32243458; PMCID: PMC7122720.

Reviewer #2: The Authors have edited the manuscript as suggested, furthermore I have no further comments to add .

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Jack Henry

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2022 Mar 31;17(3):e0265932. doi: 10.1371/journal.pone.0265932.r004

Author response to Decision Letter 1

10 Feb 2022

Thank you for your valuable comments that have helped improve our manuscript. We have carefully studied your comments and suggestions and revised our paper accordingly. The following are our point-by-point responses to your specific comments. We hope that our revisions are acceptable and that our responses adequately address the comments.

Attachment

Submitted filename: response to reviewers comment-0121.docx

Click here for additional data file. (25.3KB, docx)

Decision Letter 2

Giuseppe Biagini

14 Feb 2022

PONE-D-21-09280R2Comparative efficacy of prophylactic anticonvulsant drugs following traumatic brain injury: a systematic review and network meta-analysis of randomized controlled trialsPLOS ONE

Dear Dr. Chen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Mar 31 2022 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Giuseppe Biagini, MD

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors should be congratulated for providing an excellent revision, which has substantially improved the rigour of this meta-analysis. My concerns have been largely addressed by this revision. This is an interesting paper that provides useful data regarding the comparative efficacy of various antiepileptic medications as seizure prophylaxis in traumatic brain injury.

Major issues:

1. The authors have now added to the conclusion that they do not believe their findings support guidelines advocating the use of prophylactic phenytoin in traumatic brain injury. This is not supported by the data. Both the referenced Cochrane review, and the authors' own results, show that antiepileptics reduce the likelihood of early seizures. While there is currently no strong evidence to support any specific medication, phenytoin has the largest body of evidence for its efficacy and performed well in this meta-analysis despite the lack of statistical significance.

In addition, the efficacy of phenytoin relative to placebo IS arguably significant (OR 0.43, 95%CI 0.18-1.01). I would encourage the authors not to over-interpret p-values and the concept of 'statistical significance' in the context of network meta-analysis, where comparisons inevitably suffer from multiplicity (see doi: 10.1002/jrsm.1377). Therefore, the level IIA recommendation in the Brain Trauma Foundation guidelines to give one week seizure prophylaxis with phenytoin remains supported by the data in this study.

This is a good study but the conclusion in its current form is misleading and even potentially harmful because it implies that prophylaxis should not be given. The conclusion should be modified to make it clear that antiepileptic drugs reduce the risk of early seizures and the findings of this meta-analysis support current guidelines, but that the optimal medication is unclear. This can be considered a minor or major issue at the discretion of the editor, but should be rectified prior to publication.

Minor issues:

1. In the text, (page 9), the effect of phenytoin relative to placebo in reducing early seizures is described as OR 0.43, 95%CI 0.18-0.99. In the corresponding table S4, this is listed as OR 0.43, 95%CI 0.18-1.01. Please check these values and ensure they are correct and matching.

2. The results of the pairwise analysis comparing all AEDs versus placebo should also be included in the abstract, as the results of the paper cannot be properly interpreted without this knowledge.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Jack Henry

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2022 Mar 31;17(3):e0265932. doi: 10.1371/journal.pone.0265932.r006

Author response to Decision Letter 2

23 Feb 2022

Thank you for your valuable comments that have significantly improved our manuscript. We have revised our manuscript based on your suggestions. The point-by-point responses to your specific comments are listed as below. We hope that our revisions are acceptable and that our responses adequately address the comments.

Attachment

Submitted filename: response to reviewers comment-0217 HYC.docx

Click here for additional data file. (23.4KB, docx)

Decision Letter 3

Giuseppe Biagini

11 Mar 2022

Comparative efficacy of prophylactic anticonvulsant drugs following traumatic brain injury: a systematic review and network meta-analysis of randomized controlled trials

PONE-D-21-09280R3

Dear Dr. Chen,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Giuseppe Biagini, MD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have addressed all of my concerns about this manuscript. This is an interesting meta-analysis that warrants publication.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Acceptance letter

Giuseppe Biagini

22 Mar 2022

PONE-D-21-09280R3

Comparative efficacy of prophylactic anticonvulsant drugs following traumatic brain injury: a systematic review and network meta-analysis of randomized controlled trials

Dear Dr. Chen:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Giuseppe Biagini

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Fig. The transitivity analysis of potential effect modifiers.

    (A) age and (B) male percentage. 1 indicates Glotzner et al., (1983) study; 2 indicates Pechadre et al., (1991) study; 3 indicates Szaflarski et al., (2010) study; 4 indicates Temkin et al., (1990) study; 4 indicates Temkin et al., (1999) study; 6 indicates Temkin et al., (2007) study; 7 indicates Young et al., (1983) study; 8 indicates Young et al., 2004) study; 9 indicates Khan et al., (2016) study. CBZ = carbamazepine, LEV = levetiracetam, VPA = valproate, PBO = placebo, PTH = phenytoin, MgSO = Magnesium Sulfate (MgSO4).

    (TIF)

    Click here for additional data file. (943.3KB, tif)
    S2 Fig. The surface under the cumulative ranking of anticonvulsant drugs according to early and late post-traumatic seizure.

    The surface under the cumulative ranking curve value is the probability each intervention has of being among the best in the network, with larger values representing higher ranking probabilities.

    (TIF)

    Click here for additional data file. (1.1MB, tif)
    S3 Fig

    Direct evidence plots for (A) early post-traumatic seizure and (B) late post-traumatic seizure. This figure shows the proportion of direct evidence (orange shade) in overall reporting of network value (blue shade). Minimal parallelism: Bar chart displaying the minimum number of independent paths contributing to the effect estimate on an aggregated level. All interventions included in our analysis have values equal to one, highlighting that estimates had contributions that extended beyond single pair-wise comparisons. Mean path length. Bar chart displaying the mean path length, which characterizes the degree of indirectness of an estimate. Higher mean path lengths indicate less reliable estimates. Mean path length for indirect comparison was equal to or above 2, which points to a less reliable estimate.

    (TIF)

    Click here for additional data file. (1.3MB, tif)
    S4 Fig. The forest plots of AEDs comparisons for mortality.

    (TIF)

    Click here for additional data file. (348.7KB, tif)
    S1 Table. An example of search strategies.

    (DOCX)

    Click here for additional data file. (18.9KB, docx)
    S2 Table. Characteristics of studies.

    (DOCX)

    Click here for additional data file. (20.6KB, docx)
    S3 Table. Pairwise meta-analytic results for early and late posttraumatic seizures.

    (DOCX)

    Click here for additional data file. (18KB, docx)
    S4 Table. League tables for comparing anticonvulsant drugs for early and late post-traumatic seizures.

    (DOCX)

    Click here for additional data file. (16.9KB, docx)
    S5 Table. Risk of methodological bias score of the studies.

    (DOCX)

    Click here for additional data file. (17.9KB, docx)
    S6 Table. CINeMA summary tables.

    (DOCX)

    Click here for additional data file. (21.2KB, docx)
    Attachment

    Submitted filename: response to reviewers comment-0905 HYC.docx

    Click here for additional data file. (32.5KB, docx)
    Attachment

    Submitted filename: response to reviewers comment-0121.docx

    Click here for additional data file. (25.3KB, docx)
    Attachment

    Submitted filename: response to reviewers comment-0217 HYC.docx

    Click here for additional data file. (23.4KB, docx)

    Data Availability Statement

    All raw data files are available in the supplemental materials.

    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES