Article 1
Statistical Power of Randomized Controlled Trials in Trauma Surgery. Berg A, Lyons NB, Badami A, Reynolds J, Pizano L, Pust G, Meizoso J, Namias N, Yeh DD. J. American College of Surgeons. 2023 Nov 1;237(5):731-736.
Development of contemporary practice management guidelines in trauma is fueled by the evidence provided through quality clinical investigation. Among the stratum of study type that yields this evidence, randomized controlled trials (RCT) provide highly valued data due to their presumed minimization of bias and rigor of study design. Critical to the reliability of this data is the attainment of statistical power based on a priori power analysis as to minimize the potential effect of both Type I and Type II error. Recent studies have discovered a startling proportion of underpowered published trials within many realms of medical literature. To analyze this phenomenon further, this bibliometric study sought to determine the prevalence of underpowered RCTs within the field of trauma surgery.
The authors begin by describing the importance of attaining statistical power in determining treatment effect and elaborating on the number of insufficiently powered studies in other fields, setting adequate background and warrant for examination of trauma-specific literature. In addition, they define the evolution of study reporting by describing the rise of both CONSORT guidelines and the concept of fragility index.
Query of the Ovid Medline database identified 187 trauma-focused RCTs published between 2001-2021 across 60 different journals and documented study subtype (superiority, noninferiority, or equivalence trial), presence of an a priori sample size calculation, power analysis, and primary outcome. In addition, it was noted whether studies achieved target sample size and if the conclusion was positive or negative. Each RCT was secondarily analyzed by the publishing journal’s impact factor at time of publication, CONSORT adherence, number of subsequent citations, country of origin, data registry affiliation, and whether the trial was industry-sponsored.
Through appropriately powered post hoc analysis, authors found that of the 187 RCTs included, only 46%, 56%, and 65% were adequately powered to detect small, medium, and large effect sizes, respectively. Only 91 studies (49%) documented performing a priori sample size calculation, and that of those who did, only 62% met their target sample size. Nearly half (46%) of studies used means to describe primary outcome, with standard deviation portraying effect size. These studies had significantly smaller sample sizes on average compared to studies reporting primary outcome measures as a proportion (n=94 vs. n=198; p < 0.001). Regarding geographical discrepancies, North American studies were least likely to be powered to detect medium effect sizes (46%) compared to studies from Asia who were most likely to be appropriately powered for this effect (79%). In addition, European studies were most likely to perform a priori power calculation (65%) and had the highest proportion of studies affiliated with a clinical trial registry (39%). Lastly, analysis of RCTs published after 2010 shockingly demonstrated that only 11% strictly adhered to CONSORT guidelines. Further, those within the upper and lower quartiles of CONSORT adherence were published in journals with a median impact factor of 3.59 and 1.55, respectively.
This study sheds light on the previously unreported prevalence of RCTs in trauma that do not report a priori sample size calculations, fall short in meeting target enrollment, and fail to be adequately powered to detect even large effect sizes. The authors recognize certain limitations to conducting RCTs in trauma such as obtainment of consent in the acute care setting, funding, and time to accumulate appropriate sample size. Based on these findings, areas for substantial improvement are identified within future study design, utility of CONSORT adherence, and completeness of data reporting.
Article 2
Association Between Emergency Medical Service Agency Volume and Mortality in Trauma Patients.
Silver D, Sperry J, Beiriger J, Lu L, Guyette F, Wisniewski S, Moore E, Schreiber M, Joseph B, Wilson C, Cotton B, Ostermayer D, Fox E, Harbrecht B, Patel M, Brown J. Ann Surgery. 2023 Aug 28.
Over the past two decades, evidence has consistently demonstrated the relationship between trauma center volume and patient outcomes. This has led to minimum patient volume requirements for center accreditation and verification and an emphasis on the regionalization of care. Within trauma systems, emergency medical services (EMS) provide the critical link between severely injured patients and definitive care. While the association of volume and outcomes has been heavily studied at the trauma center level, the impact of this variable on prehospital care had yet to be investigated. This retrospective multicenter study aimed to determine whether a similar increase in trauma patient volume was associated with mortality benefits at the EMS agency level.
Using the prospectively collected Linking Investigations in Trauma and Emergency Services (LITES) database, 33,511 patients aged 18-90 with and injury severity score (ISS) >9 transported by the 20 participating air and ground EMS agencies between 2017-2021 were identified and included in analysis, with 6-hour mortality being the primary outcome. Risk-adjusted multi-level logistic regression models were implemented to determine the association of EMS trauma volume with mortality. Generalized additive mixed model (GAMM) and mixed effect random forest model (MERF) analyses were additionally performed to examine potential non-linear associations between volume and mortality.
Median annual volume of trauma transports per EMS agency was 374 (IQR 90-580). For every incremental increase of 50 trauma patients transported annually per agency, a 5% adjusted odds reduction in 6-hour mortality (aOR 0.954; 95% CI 0.916-0.994, p=0.025) and 3% adjusted odds reduction in 24-hour mortality (aOR 0.973; 95% CI 0.949-0.998, p=0.038) was observed. GAMM analysis of non-linear effects exhibited a notable decrease in mortality when more than 250 trauma patients were transported by an agency per annum. MERF analysis of factors impacting 6-hour mortality revealed that EMS volume was the most influential, with necessitation of pRBCs in the first 4 hours being second. Additionally, risk-standardized mortality ratio (RSMR) for EMS agencies with transport volumes above the median were significantly lower than those below the median, 1.85% vs. 4.82%; p=0.003, respectively.
This study uniquely illustrates how the association between patient volume and mortality extends into the prehospital setting. In doing so, areas for evidence-backed quality improvement are identified for both EMS agencies and trauma systems as a whole. By using this data to supplement standards for EMS provider preparedness and define best practices from high-volume agencies, reductions in mortality can be made through further optimizing the prehospital care of critically ill patients.
Article 3
Early venous thromboembolism prophylaxis in patients with trauma intracranial hemorrhage: Analysis of the prospective multicenter Consortium of Leaders in Traumatic Thromboembolism study. Wu YT, Chien CY, Matsushima K, Schellenberg M, Inaba K, Moore EE, Sauaia A, Knudson MM, Martin MJ; CLOTT Study Group. J Trauma Acute Care Surg. 2023 Nov 1;95(5):649-656.
Timing of venous thromboembolism (VTE) prophylaxis following traumatic brain injury remains a controversial topic in the care of traumatically injured patients. TBI patients are thought to be at elevated risk for VTE given their reduced mobility and resulting venous stasis, as well as the systemic inflammatory effects of their injuries. This concern for VTE is countered by apprehension that pharmacologic VTE prophylaxis could result in bleeding progression and worsened neurologic outcome. This results in large variations in practice patterns related to the timing of VTE prophylaxis initiation. The majority of previous literature on this subject has been limited to single site investigation or the repurposing of larger datasets not specifically designed for the study of VTE. For these reasons, the authors decided to use a large multi-center prospective dataset designed for VTE investigation to better examine the safety and timing of pharmacologic VTE prophylaxis in the TBI population.
The authors performed their secondary analysis using data procured by the Consortium of Leaders in Traumatic Thromboembolism (CLOTT) study group, which comprised almost 8,000 patients from 17 level-1 trauma centers over a 3-year period, who were studied prospectively in order to answer questions regarding post-traumatic thromboembolism. From this dataset, Head AIS was used to identify patients with TBI. Inclusion criteria was an AIS score of 3-5, hospital stay >72 hours, and either no pharmacologic VTE prophylaxis in the first 24 hours of their admission due to intra-cranial bleeding, or emergent neurosurgical intervention (Craniotomy / Craniectomy / ICP monitoring / Ventricular drain) at admission. Patients with a non-survivable head injury (AIS 6) were excluded, as were those who never received VTE prophylaxis.
A total of 881 patients were identified and included, with 378 (43%) receiving VTE prophylaxis within 48 hours of admission, and the remaining 503 (57%) receiving it at greater than 48 hours. In the late VTE group, median time until receiving pharmacologic prophylaxis was 4 days. Groups were similar in demographics and mechanisms of injury, though the late VTE group had a larger percentage of patients with a Head AIS of 5, and those undergoing neurosurgical intervention. The early VTE group demonstrated a lower rate of VTE (7.2% vs 12.4%, P=0.01) with an equivalent rate of intracranial bleeding (1.9% vs 1.8%, P=0.95). Backwards stepwise multivariate logistic regression was then performed using the above data, where the authors found that an increased VTE rate was independently associated with VTE prophylaxis >48hours after admission, ventilator use >3 days, and Risk Assessment Profile (RAP) score >4. Additionally, enoxaparin use was associated with lower VTE rate as compared to unfractionated heparin (OR, 0.54; 95% CI, 0.33–0.88). Multivariate modeling did not show any independent association between the early initiation of VTE prophylaxis and bleeding events.
This study has some limitations, which are readily acknowledged by the authors. VTE surveillance and the prophylactic agent were at the discretion of physicians at each site and not subject to universal protocol, and details regarding specific head injury type, operative findings, and timing of intracranial hemorrhage progression were absent. Additionally, the CLOTT dataset includes only patients aged 18-40, as it was funded by the Department of Defense, who specified the patient population of interest reflect the ages of active military members. Despite these limitations, this study offers some of the most compelling research thus far in the ongoing investigation of timing and safety of VTE prophylaxis in traumatic brain injury, providing a prospective multi-center perspective to the issue. The authors propose additional prospective investigation regarding specific high-risk subgroups, which would aid in clinical decision making moving forward.
