Recent studies have reported that PEF plays an important role in haemostasis through stabilisation of the pelvic ring in pelvic fracture with shock to reduce additional damage and the reduction effect of the pelvic cavity (11-13). However, a recent study that analysed 10-year data on external emergent stabilisation using the German pelvic trauma registry showed a decreasing trend in the use of PEF in patients with pelvic ring fracture. In contrast, the use of PB has increased rapidly, and it was used in almost 40% patients (10). Moreover, in a recent multi-centre study conducted in a level I trauma centre in the United States, PB was performed in 50% patients with pelvic fracture and shock, and PEF was performed in only 4% patients (12). These results show that PB has been increasingly used instead of PEF, and its use has been continuously increasing due to its simplicity and speed in application (11, 15). It is difficult to compare the effects of PEF and PB in the treatment of patients with haemorrhage due to pelvic fracture compared to haemorrhage due to other injuries because a combination of various modalities is possible. In 2007, in a comparative study between the PEF and PB groups, Croce et al. showed that the mortality rates were similar, but the requirement for packed RBC transfusion at 24 and 48 h was significantly lower in the PB group than in the PEF group. However, there was a difference in characteristics between the two groups. Since the recently used procedures such as PPP or REBOA were not analysed together, it is difficult to accept the results in the current scenario (11). In our study, to minimise the effect of other haemostatic procedures and compare the differences between the effects of PB and PEF, the proportions of patients who had undergone PPP, PA, and laparotomy were corrected using PSM. The results showed that PEF was superior to PB in terms of haemorrhage-induced mortality and 7-day mortality rates. There are several reasons for this difference. First, if PB is not removed quickly or over-tighten, complications such as skin necrosis and pressure ulceration may occur; therefore, it is recommended that PB be maintained for >24–48 h (2, 16). In the trauma centres included in this study, PB was removed within 48 hours when the patient was haemodynamically stabilised, but the definitive fixation of the pelvis was determined considering the patient’s condition and was performed after an average of 6 days after injury. Therefore, it is thought that the effect of pelvic stabilisation in PEF, but not in PB, lasted sufficiently until definitive fixation. Second, PEF maintains constant pelvic stabilisation during additional procedures such as PA, REBOA, and femoral vessel access after application, whereas PB may increase the risk of bleed because of the possibility of additional damage when reapplied after removal for additional procedures. Although not statistically significant in this study, the amount of packed RBCs transfused after 4 h in the PEF group tended to be lower than that in the PB group. This result seems to be consistent with that reported by the Denver group, showing that patients with pelvic fracture and haemodynamic instability undergoing PEF with PPP had a very good overall mortality rate (5, 13). Third, the tile classification of the PEF and PB groups was equally corrected by PSM, but there was a difference in the pattern of the type of pelvic ring fracture according to the Burges and Young classification between the two groups. In other words, the PB group tended to have more vertical shear-type pelvic fractures than the PEF group (60% vs 35%, p = 0.205). Unlike APC-and LC-type pelvic fractures, vertical shear-type injuries are best stabilised by posterior C-clamp (17, 18), and C-clamp is not commonly used in trauma centres in Korea due to the lack of orthopaedic trauma specialists (19).
Therefore, it appears that in patients with vertical shear-type pelvic fracture, PB was first applied rather than PEF, and then definitive fixation was performed when their condition was haemodynamically stabilised. The difference in the mortality rate might be due to the difference in the incidence of acute haemorrhage between the two groups. The pelvic trauma management algorithm of the World Society of Emergency Surgery was used to defined severe lesions (WSES grade IV) regardless of mechanical instability in cases of haemodynamic instability. After application, haemostatic procedures such as PPP, mechanical fixation, REBOA, and PA should be performed complementarily (2). In our study, before PSM, PPP with PEF was the most commonly performed in the PEF group (71.4%), whereas PA was most commonly performed in the PB group (52.1%). This result shows that PEF was mainly applied in the operating room while performing surgery such as PPP or laparotomy. When moving to the angiography suite for PA, it is thought that PB that can be easily and quickly removed and reapplied is combined.
REBOA has recently been increasingly used in patients with haemodynamic instability instead of emergent resuscitative thoracotomy (20-23). In Korea, REBOA was first used in regional trauma centres in 2016. It is being used as a bridge procedure before other haemostasis in patients with pelvic fracture accompanied by severe shock (23). In our study, 12 patients were treated with REBOA, of which 11 (91.7%) patients underwent PPP and nine (75%) patients underwent PA. The haemorrhage-induced mortality and total mortality rates in these patients were 33.3% and 66.7%, respectively, and PEF was performed in only three (25%) patients. These results are thought to be because REBOA was used in patients with a clinically critical condition, and PB, which can be easily applied, was preferred over PEF when it was necessary to move to the operating room. Although patients were not matched according to REBOA application by PSM, the application rate between the two groups was the same after PSM; therefore, it is judged that the effect of REBOA application did not affect the clinical outcome.
Since this was a retrospective study, this study has limitations, such as a selection bias between the two groups. Further, the statistical power was low because the number of patients who underwent PEF was very small. Nevertheless, this study is a rare study on the effectiveness of pelvic stabilisation procedures performed with various haemostatic procedures in patients with haemodynamic instability and pelvic fractures. The advantages of this study are that PSM was performed to correct various confounding factors and that patients from three institutions were included in the study. However, a larger prospective study is needed to confirm the results of our study.