Chronic Subdural Hematoma Management
Chronic Subdural Hematoma Management
We undertook a systematic review and meta-analysis and reported our findings in accordance with recommendations made by the Meta-analysis Of Observational Studies in Epidemiology (MOOSE), the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA), the Cochrane handbook, and the Grading of Recommendation, Assessment, Development, and Evaluation (GRADE) working groups' guidelines.
Four independent reviewers (S.A.A., C.H., B.Y., and S.S.) performed a detailed electronic search of MEDLINE (PubMed and Ovid), EMBASE, CINAHL, Google scholar, and the Cochrane library for studies published between January 1970 and February 2013 written in the management of CSDH without language restriction. Key words and MeSH (Medical Subjective Headings) terms pertaining the condition (ie, CSDH, chronic subdural, hematoma, haematoma, hemorrhage, haemorrhage, or bleed) were cross-referenced with terms pertinent to treatment (ie, burr hole, twist drill, percutaneous, craniotomy, steroid, management, treatment, surgery, evacuation, irrigation, drainage, or conservative) in relevant combinations. The full-text versions of all studies that the reviewers considered relevant were obtained. References of studies with potential relevance and relevant reviews were screened manually to identify any applicable studies that were not previously identified.
Studies were included if they stated and evaluated a treatment modality for clinically symptomatic patients with radiological findings of CSDH and reported one or more of the management outcomes on follow-up. Articles with a sample size below 15 consecutive patients were excluded. In addition, we excluded duplicate references, studies that did not specify the treatment of choice, and resources that mixed the management outcomes of acute with chronic subdural hematomas. Papers with more than 1 treatment were considered eligible if outcomes of each management were separated. We excluded reports that examined only infantile CSDH. Abstracts presented in conferences and letters to editors were excluded if studies were not published as full reports. We included non-English articles if they presented the previously discussed selection criteria in a translated English abstract. Moreover, the current authors included their own unpublished observational study as it is one of the largest series in the literature with multiple cohorts and comparisons (for description of the single center series see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452). Any disagreement between the 4 reviewers concerning the decision to include or exclude a study was solved by discussion and consensus with a fifth reviewer (N.M.).
Data were abstracted by 6 investigators (S.A.A., C.H., B.B., B.M., P.A., and N.M.). We developed a system to ensure accuracy by having each reference abstracted and reviewed by 3 investigators independently, including the principal reviewer (S.A.A.). Discrepancies were settled by consensus and discussion. We gathered information from eligible articles using data abstraction forms to include title, first author, year of publication, study design, number of cohorts, total number of managed patients, total number of CSDHs (including bilateral cases), number of patients with previous use of antiplatelet or anticoagulant therapy, main treatment modality (percutaneous twist-drill drainage, burr holes, craniotomy, and corticosteroids), use of drains, duration of drainage, irrigation of the hematoma cavity, adjuvant use of steroids, the value of bed rest after surgical management, bed rest duration, single or multiple holes, recurrence rates, mortality, morbidity, types of complications, cure rates, treatment of recurrence, and recurrences following the second intervention (see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452).
Although the definition of percutaneous drainage is a relatively smaller skull puncture (usually <0.5 cm) made by a handheld twist-drill done at the bedside under local anesthetic, burr hole craniostomies are wider openings through the skull created using larger operating room drills. Craniotomy mostly requires intubation under general anesthetic and is defined as the creation of a cranial flap to access the hematoma cavity. The use of corticosteroids included dexamethasone and prednisone. We examined steroid usage as a nonsurgical conservative therapy and its use as adjuvant treatment. The definition of recurrence (primary outcome of success) was variable among the included articles. It was generally defined as reoccurrence of CSDH radiological and clinical findings warranting further treatment. Mortality, morbidity (secondary outcomes of safety), and improvement rates were reported at the time of discharge. The documented timing of recurrence rates was variable among included studies (up to 1, 3, or 6 months) based on reoccurrence of patients' symptomology that required management. Cure rates (tertiary outcome of efficacy) were assessed using multiple scales. The most commonly used was Markwalder classification system. We adopted the improved neurological outcomes as reported by the authors of included papers.
Methodological quality assessment was done by 2 reviewers (S.A.A. and W.A.). We used the Cochrane collaborations tool to assess the risk of selection, performance, detection, attrition, and reporting biases of randomized trials. The Newcastle-Ottawa Scale (NOS) was used to measure comparability, selection of cohorts, and assessment of outcomes (for quality assessment results see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452).
We performed separate analyses of randomized, comparative, and single cohort studies of prospective and retrospective designs. For each treatment option, we examined the outcomes of recurrence rate, mortality, morbidity, and cure rates. We reported the overall outcome estimates following CSDH management per 100 patients per year of the included studies separately for each of the previously discussed study designs. A comparison was made between percutaneous drainage, burr holes, craniotomy, and corticosteroids as main first-line treatment options. The adjuvant use of steroids to surgical management was evaluated. For burr holes, the roles of drainage, irrigation, and number of holes were analyzed. We did not perform similar analyses of percutaneous twist-drill drainage, because in all randomized and 80% of observational studies this procedure was described as the creation of a single skull puncture and insertion of a drain with no irrigation. The effect of bed rest following surgical management was examined. We compared percutaneous drainage, burr holes, and craniotomy as treatment options for recurrent cases. Randomized trials with similar comparisons were analyzed separately (Table 1). When appropriate, we pooled data from cohort studies reporting similar outcomes (Table 2). Moreover, we undertook separate meta-analyses of comparative cohort studies with similar arms of intervention (Fig. 1).
(Enlarge Image)
Figure 1.
Meta-analyses of high-quality observational studies comparing percutaneous twist-drill bedside drainage versus operative theater burr hole craniostomy. Outcomes evaluated are mortality (A), morbidity (B), cure (C), and recurrence rates (D).
We undertook sensitivity analyses of high-quality observational studies, defined as prospective study designs with high scores using The Newcastle-Ottawa Scale assessment (>4 points). Heterogeneity among studies was investigated using the Q test and quantified by the I statistic, which represents the percentage of total variation across studies with a predefined I > 50% as the cutoff point of statistical heterogeneity. We used the DerSimonian and Laird random-effects model to calculate weighted pooled proportions due to a priori that variable studies with different designs are prone to inherent heterogeneity. Weights were calculated using the Mantel-Haenszel method. Publication bias was evaluated visually by funnel plots, and quantified by Egger regression and Begg-Mazumdar tests. We used GRADE profiler 3.6 to stratify the level of evidence of variable outcomes (see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452). We calculated relative risks (RRs) and 95% confidence intervals (CIs) and generated forest plots for all outcomes. An alpha error of <0.05 was considered a criterion for statistical significance. We used RevMan (version 5.1) and StatsDirect (version 2.7.9) for statistical analyses.
Methods
We undertook a systematic review and meta-analysis and reported our findings in accordance with recommendations made by the Meta-analysis Of Observational Studies in Epidemiology (MOOSE), the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA), the Cochrane handbook, and the Grading of Recommendation, Assessment, Development, and Evaluation (GRADE) working groups' guidelines.
Search Strategy
Four independent reviewers (S.A.A., C.H., B.Y., and S.S.) performed a detailed electronic search of MEDLINE (PubMed and Ovid), EMBASE, CINAHL, Google scholar, and the Cochrane library for studies published between January 1970 and February 2013 written in the management of CSDH without language restriction. Key words and MeSH (Medical Subjective Headings) terms pertaining the condition (ie, CSDH, chronic subdural, hematoma, haematoma, hemorrhage, haemorrhage, or bleed) were cross-referenced with terms pertinent to treatment (ie, burr hole, twist drill, percutaneous, craniotomy, steroid, management, treatment, surgery, evacuation, irrigation, drainage, or conservative) in relevant combinations. The full-text versions of all studies that the reviewers considered relevant were obtained. References of studies with potential relevance and relevant reviews were screened manually to identify any applicable studies that were not previously identified.
Selection Criteria
Studies were included if they stated and evaluated a treatment modality for clinically symptomatic patients with radiological findings of CSDH and reported one or more of the management outcomes on follow-up. Articles with a sample size below 15 consecutive patients were excluded. In addition, we excluded duplicate references, studies that did not specify the treatment of choice, and resources that mixed the management outcomes of acute with chronic subdural hematomas. Papers with more than 1 treatment were considered eligible if outcomes of each management were separated. We excluded reports that examined only infantile CSDH. Abstracts presented in conferences and letters to editors were excluded if studies were not published as full reports. We included non-English articles if they presented the previously discussed selection criteria in a translated English abstract. Moreover, the current authors included their own unpublished observational study as it is one of the largest series in the literature with multiple cohorts and comparisons (for description of the single center series see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452). Any disagreement between the 4 reviewers concerning the decision to include or exclude a study was solved by discussion and consensus with a fifth reviewer (N.M.).
Data Extraction
Data were abstracted by 6 investigators (S.A.A., C.H., B.B., B.M., P.A., and N.M.). We developed a system to ensure accuracy by having each reference abstracted and reviewed by 3 investigators independently, including the principal reviewer (S.A.A.). Discrepancies were settled by consensus and discussion. We gathered information from eligible articles using data abstraction forms to include title, first author, year of publication, study design, number of cohorts, total number of managed patients, total number of CSDHs (including bilateral cases), number of patients with previous use of antiplatelet or anticoagulant therapy, main treatment modality (percutaneous twist-drill drainage, burr holes, craniotomy, and corticosteroids), use of drains, duration of drainage, irrigation of the hematoma cavity, adjuvant use of steroids, the value of bed rest after surgical management, bed rest duration, single or multiple holes, recurrence rates, mortality, morbidity, types of complications, cure rates, treatment of recurrence, and recurrences following the second intervention (see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452).
Definition of Variables
Although the definition of percutaneous drainage is a relatively smaller skull puncture (usually <0.5 cm) made by a handheld twist-drill done at the bedside under local anesthetic, burr hole craniostomies are wider openings through the skull created using larger operating room drills. Craniotomy mostly requires intubation under general anesthetic and is defined as the creation of a cranial flap to access the hematoma cavity. The use of corticosteroids included dexamethasone and prednisone. We examined steroid usage as a nonsurgical conservative therapy and its use as adjuvant treatment. The definition of recurrence (primary outcome of success) was variable among the included articles. It was generally defined as reoccurrence of CSDH radiological and clinical findings warranting further treatment. Mortality, morbidity (secondary outcomes of safety), and improvement rates were reported at the time of discharge. The documented timing of recurrence rates was variable among included studies (up to 1, 3, or 6 months) based on reoccurrence of patients' symptomology that required management. Cure rates (tertiary outcome of efficacy) were assessed using multiple scales. The most commonly used was Markwalder classification system. We adopted the improved neurological outcomes as reported by the authors of included papers.
Quality Assessment
Methodological quality assessment was done by 2 reviewers (S.A.A. and W.A.). We used the Cochrane collaborations tool to assess the risk of selection, performance, detection, attrition, and reporting biases of randomized trials. The Newcastle-Ottawa Scale (NOS) was used to measure comparability, selection of cohorts, and assessment of outcomes (for quality assessment results see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452).
Statistical Analyses
We performed separate analyses of randomized, comparative, and single cohort studies of prospective and retrospective designs. For each treatment option, we examined the outcomes of recurrence rate, mortality, morbidity, and cure rates. We reported the overall outcome estimates following CSDH management per 100 patients per year of the included studies separately for each of the previously discussed study designs. A comparison was made between percutaneous drainage, burr holes, craniotomy, and corticosteroids as main first-line treatment options. The adjuvant use of steroids to surgical management was evaluated. For burr holes, the roles of drainage, irrigation, and number of holes were analyzed. We did not perform similar analyses of percutaneous twist-drill drainage, because in all randomized and 80% of observational studies this procedure was described as the creation of a single skull puncture and insertion of a drain with no irrigation. The effect of bed rest following surgical management was examined. We compared percutaneous drainage, burr holes, and craniotomy as treatment options for recurrent cases. Randomized trials with similar comparisons were analyzed separately (Table 1). When appropriate, we pooled data from cohort studies reporting similar outcomes (Table 2). Moreover, we undertook separate meta-analyses of comparative cohort studies with similar arms of intervention (Fig. 1).
(Enlarge Image)
Figure 1.
Meta-analyses of high-quality observational studies comparing percutaneous twist-drill bedside drainage versus operative theater burr hole craniostomy. Outcomes evaluated are mortality (A), morbidity (B), cure (C), and recurrence rates (D).
We undertook sensitivity analyses of high-quality observational studies, defined as prospective study designs with high scores using The Newcastle-Ottawa Scale assessment (>4 points). Heterogeneity among studies was investigated using the Q test and quantified by the I statistic, which represents the percentage of total variation across studies with a predefined I > 50% as the cutoff point of statistical heterogeneity. We used the DerSimonian and Laird random-effects model to calculate weighted pooled proportions due to a priori that variable studies with different designs are prone to inherent heterogeneity. Weights were calculated using the Mantel-Haenszel method. Publication bias was evaluated visually by funnel plots, and quantified by Egger regression and Begg-Mazumdar tests. We used GRADE profiler 3.6 to stratify the level of evidence of variable outcomes (see the Supplemental Digital Content, available at: http://links.lww.com/SLA/A452). We calculated relative risks (RRs) and 95% confidence intervals (CIs) and generated forest plots for all outcomes. An alpha error of <0.05 was considered a criterion for statistical significance. We used RevMan (version 5.1) and StatsDirect (version 2.7.9) for statistical analyses.
Source...