BBMI, particularly mild to major mesenteric bleeding, may present with subclinical bleeding and remain undiagnosed. Once lethal exsanguination manifests, patients usually exhibit life-threatening symptoms, meaning that clinicians have no time to respond. Accordingly, in light of the dramatically deteriorating course, most existing scoring systems are not clinically practical because of the availability of time-consuming laboratory data, not to mention the initial vital signs, although they are all helpful adjuvant tools in clinical decision-making. Many scoring systems have been developed and validated retrospectively [14], reflecting their insufficient clinical value and the need to be refined. Joseph et al. [5] used SI > 1.0 and the presence of pelvic fracture instead of hypotension and tachycardia to refine the Assessment of Blood Consumption (ABC) score as the Revised Assessment of Bleeding and Transfusion (RABT) score due to concerns over triage of MT and increased product waste. Some authors [3] considered that more weighted variables could be better than the absence of weighted scoring systems in predicting MT in trauma patients, such as the trauma-associated severe hemorrhage (TASH) score and the Prince of Wales Hospital (PWH) score, both of which comprise seven parameters. In our opinion, we support the concept that different scoring systems should not be used alone or over-dependent due to the dynamic nature of trauma; re-assessment of these scores is clinically necessary [3].
The principal treatment for BBMI is based on early diagnosis, timely operation, and attempted resuscitation. However, apart from early therapeutic laparotomy, timely recognition of patients in need of rapid activation using the MT protocol could be another challenge before laparotomy. Although the need for MT to resuscitate patients with BBMI is well known, both clinically and academically, fewer lectures are associated with this issue. Niederee et al. conducted a 11-year retrospective study associated with outcome of delayed diagnosis (> 24 hours) in patients with BBMI, and reported that the patients with diagnosis within 24 hours had received mean 8.5 unites blood products and 26.7% mortality rate, whereas patients with diagnosis beyond 24 hours had received 11.7 units blood products and 36.4% mortality rate [11]. Our results show that there are approximately 31% patients in need of MT after BBMI, who present 32% overall mortality rate and 8% mortality rate within 24 hours, respectively. The median administered packed RBC within 24 hours in the MT(+) and MT(-) group were 16U and 2U, respectively. Accordingly, in addition to the early diagnosis and attempted laparotomy, the effectiveness and importance of administering blood to these patients in the trauma bay should not be ignored. To predict the MT for patients after abdominal trauma, El-Menyar et al. created FASILA scores using the acronym FAST, SI, and initial serum lactate to investigate the implications of MTP activation, exploratory laparotomy (ExpLap), and mortality and proposed that from FASILA scores of 4 to 6 points, the percentage of ExpLap increases from 28 to 85%, whereas the MTP activation increases from 17 to 68% [4]. They validated the novel FASILA score as superior to the RABT, SI, and ABC scores in the prediction of MTP and blood transfusion using receiver operating characteristic analysis. Although the study included all injured organs after abdominal trauma, either penetration or blunt mechanisms, it seemed impractical in the prediction of ExLap, as there were fewer BAT patients who still required ExLap in the era of NOM. El-Menyar et al. [12] also conducted another retrospective analysis with abdominal trauma patients sustaining solid organ injury, using SI ≥ 0.8 to predict the transfusion, and concluded that on multivariate analysis, although SI had significant prediction of transfusion and MTP activation, it failed to show significant prediction of ExLap. Therefore, we thought it would be interesting to explore the association between MT and hollow organ injuries after BAT in a different prediction model in terms of investigation of associated injury of hollow organ injury.
Our study is the first to explore the implications of MT and hollow organ injuries after BAT and shows that besides the involvement of mesenteric injury, TBI and pelvic fracture are significant risk factors for the prediction of MT in patients with surgical BBMI after multivariate analysis. As expected, pelvic fracture was found to be a significant risk factor. However, it is interesting to note that TBI remained a significant risk factor after controlling for other confounding factors. Our study is also the first to report a significant correlation between MT and TBI as an associated injury in patients with BBMI. Although it is well known that BBMI could occur with other associated injury concomitantly [10], the amount of intracerebral hemorrhage (ICH) does not lead to significant blood loss theoretically. Based on our observations, there are some viewpoints for this finding: (1) ICH needs to increase SBP to compensate the intracranial hypertension (ICP), which would speed more bleeding given the increased SBP; (2) the necessity of transfused RBC administrated for BBMI would worsen cerebral autoregulation [15], which would result in further deterioration of increased ICP; (3) coagulopathy can occur early after TBI with being 22%, contributing to more bowel mesenteric hemorrhage [16, 17]; there are some evidence reporting the implication between TBI and coagulopathy, which may associate with the release of injury-mediated thromboplastin or the activation of the protein C pathway in responsive of hypoperfusion [18, 19]; (4) splanchnic vasoconstriction would develop after TBI through some mechanisms, such as increased levels of circulating catecholamines or cytokines in response to neurotrauma, and the possibility of re-distribution of blood flow within the bowel system, which may contribute to accentuate the bowel mesenteric bleeding [20]; (5) patients would develop hypovolemia after sustaining subarachnoid hemorrhage and need more transfusions to replace the reduced circulating blood volume [21]; (6) increased bowel permeability was noted from some evidence [22], speculating from adjustment of protein expression after TBI via animal experiments [23], which maybe the reason that why patents sustaining concomitant BBMI and TBI would accelerate blood volume loss so that they needed more blood transfusions. Peiniger et al. [16] also supported the concept of presence of early coagulopathy in patients sustaining major trauma (ISS ≥ 16) involved with TBI (AIS ≥ 3) and receiving massive transfusion (≥ 10U pRBC),and described that high FFP: pack RBC transfusion ratio (> 1:2) was an independent predictor of survival compared with low FFP: pRBC transfusion ratio (≤ 1:2). They further analyzed patients with and without TBI and suggested that severely injured patients with TBI benefited significantly more than those without TBI, with an odds ratio of mortality (0.48 vs. 0.70). Some evidence also suggests that brain injuries can initiate systemic inflammatory response syndrome (SIRS) and multiple organ failure [22], which can deteriorate splanchnic tissue perfusion depending on the severity of the neurological injury. Although the abovementioned evidence has been described in associated lectures, the definite pathophysiological processes are yet to be clarified. On the other hand, our observations seemed to echo the above-mentioned concept regarding the mutual interaction between TBI and the gut-brain axis [24]. Studies involving TBI and gut pathophysiology are difficult to corroborate in human subjects or clinical practice. Future prospective studies or animal experiments focusing on the pathophysiology of TBI and the bowel are required to further clarify this correlation.
Our results showed that pelvic fracture is another significant predictor of MT in patients sustaining BBMI with concomitant pelvic fracture, which is similar to other scoring systems that use pelvic fracture as a variable, such as the TASH, PWH, and RABT scores [3, 5]. Pelvic disruption with major retroperitoneal hemorrhage has a high lethality rate ranging from 17 to 35%, 43% manifest with hemodynamic instability, and 30% require MT [25]. Patients with BAT who sustain pelvic fractures with concomitant intraperitoneal and retroperitoneal hemorrhage were thought to have a high possibility of requiring laparotomy and activation of the MTP. It was found that patients receiving laparotomy needed a mean blood transfusion of 914.5 ml, patients receiving laparotomy plus subsequent TAE required a mean blood transfusion of 1542.8 ml [26]. Additionally, patients with concomitant unstable hemodynamic and pelvic fractures were difficult to manage, in whom intra-abdominal injury was easily overlooked, particularly BBMI, with an 86% missed rate. These difficult-to-manage patients would require initial TAE for hemostasis and post-TAE laparotomy for intra-abdominal missed injuries, requiring a mean transfusion of 2016.7 ml [27]. If the intra-abdominal lesion requiring operation is not recognized in time in such patients and the required laparotomy is delayed by different times, they would require a mean transfusion of 1035.7 ml with a delay < 6 hours, 1625.0 ml with a delay of 6–24 hours, and 3291.7 ml with a delay > 24 hours [27]. It is well known that hemodynamic instability frequently occurs in patients with complex pelvic fractures, although in our present series, only 12.3% (20/163) of patients had combined pelvic fractures without stratifying stable or unstable fracture patterns, which had insufficient power to corroborate it as a predictor. Nevertheless, it had been reported that even patients with stable pelvic fractures may have 82% severe injuries in abdominal regions (AIS ≥ 3) with manifestation of unstable hemodynamic, whereas hemodynamic instability resulted from retroperitoneal hemorrhage in patients with pelvic fracture may also occur in 46% non-complex pelvic fractures [28]. The pattern, stability, and severity of pelvic fractures do not play significant roles in predicting mortality in patients with blunt pelvic fractures [29]. In prior scores, PWH and TASH used displaced or unstable pelvic fractures as variables to predict MT [3], which would decrease the sensitivity of missing the opportunity to activate the MTP [30]. On the other hand, a recent RABT score using any pelvic fracture as a variable was validated in Canadian level I centers, and it was reported that after sub-analysis for patients with only unstable pelvic fractures compared with any pelvic fractures, the RABT score had decreased sensitivity from 62–55% but improved specificity from 63–68% without a change in the positive predictive value (23%) and negative predictive value (90%) [2]. Accordingly, the presence of any pelvic fracture in patients with BBMI should alert clinicians to the high possibility of the need for MT.
In our study, we did not include the ISS in the multivariate logistic regression because the ISS interacts with other associated injuries. The ISS is an anatomical scoring system that can provide a good severity assessment; however, it can only select an injured organ within a body region and fails to predict prognosis. The MT(+) group had higher ISS and blood transfusion amounts and percentages, which could explain the early presence of morbidity and mortality in the MT(+) group, regardless of its inherently higher injury severity or complications of receiving MT [31]. Furthermore, a higher ISS resulted from a higher incidence of ICH in the MT(+) group. Although ICH can contribute to a high ISS, ICH usually does not require transfusion, which is another reason we did not include ISS in the logistic regression. Our results showed that ICH and pelvic fractures were risk factors for MT in patients with BBMI. Additionally, the association between ICH and hemorrhagic pelvic fractures has not been discussed in prior studies. Kido et al.[29] conducted a retrospective analysis including 102 patients with concomitant fatal bleeding pelvic fracture and severe associated injuries (AIS ≥ 3), and concluded that TBI and shock were significant independent predictors in mortality within 24 hours, indicating that TBI had higher risk of mortality than other body parts injury in patients with fatal bleeding pelvic fracture, whose pelvic artery active bleeding were identified by enhanced CT. In other words, this finding seemed to reflect that although TBI had a worse outcome inherently in trauma, as expected, it cannot exclude the fact that TBI may advance to hemorrhagic shock in the acute stage in patients with fatally bleeding pelvic fractures. However, the authors did not mention the relationship between blood transfusions in their study. Accordingly, future studies on TBI and bleeding after BAT are warranted.
4.1 Limitations
There are some limitations in our study. First, its retrospective design, institutional experience, and small sample size are weaknesses in terms of powerful statistical significance and generalization. Secondly, our study mainly focused on MT in patients, however, we did not have data on the quantity of hemoperitoneum on CT or FAST [7]. Third, we did not stratify the severity of every concomitant or associated injury, which introduces bias in the consistency of severity. Despite these weaknesses, our study provided important insights into the predictions of MT in BBMI and showed that the associated injury with TBI or pelvic fracture had a significant impact on BBMI patients requiring MT. We hoped we could innovatively seek easy, non-laboratory, and rapid assessment models to predict the requirement of MT during the expeditious diagnosis and resuscitation course for the purpose of dealing with inherent trauma nature with a dynamic change and aid in decision making for treating difficult-to-manage patients with surgical BBMI.