Surgical Versus Non-surgical Treatment of Unstable Lateral Compression Type I (LC1) Injuries of the Pelvis With Complete Sacral Fractures in Non-fragility Fracture Patients: A Systematic Review

Lateral compression type 1 (LC1) injuries comprise two-thirds of pelvic fractures. Approximately one-third of LC1 fractures are unstable and may benefit from surgical fixation to improve stability but it is not clear if this leads to better clinical or cost-effectiveness outcomes. This study explores differences in patient-reported outcomes, complications, time-to-mobilisation, cost-effectiveness, and length of hospital stay between surgically and non-surgically treated unstable LC1 non-fragility fractures. We performed a systematic review to determine whether surgical or non-surgical treatment yielded better clinical and cost-effectiveness outcomes for the treatment of unstable LC1 pelvic injuries with complete sacral fractures, excluding fragility fractures. We searched Medline, Embase and Cochrane databases from inception to June 2022, as well as clinical trial registries. A formal meta-analysis was not possible due to available study designs and heterogeneity. Therefore, a narrative review of the findings has been provided. Five observational studies met the inclusion criteria. A total of 183 patients were treated surgically, and 314 patients were treated non-surgically. Patients treated surgically had lower pain levels (Visual Analogue Scale) and fewer days to mobilisation. Quality of life (EuroQol-5 domains and 36-Item Short Form questionnaires) was better in the surgical group, but not statistically significant. No statistical differences in the length of hospital stay or complication rates were found. This review highlights the low quantity and quality of existing data on patients with unstable LC1 pelvic fractures and the need for a definitive randomised controlled trial to determine whether surgical or non-surgical care should be the preferred treatment concerning clinical and cost-effective care.


Introduction And Background
Lateral compression (LC) fractures are the most common type of pelvic fracture, with lateral compression type 1 (LC1) injuries accounting for two-thirds of all pelvic fractures [1,2]. Unstable LC1 fractures are synonymous with the 61B2.1 pelvic ring fractures of the Orthopaedic Trauma Association (OTA)/AO Foundation (AO) classification [3]. LC1 fractures typically result from a lateral force, which frequently leads to pubic rami fractures of the anterior pelvic ring and an associated sacral impaction fracture [4]. The mortality rate of LC1 fractures during index hospital admission is reported to be between 5% and 9% [2]. LC1 fractures may be stable or unstable, which has caused controversy over the optimal management strategy for this injury pattern [5][6][7]. There is currently no high-quality evidence or agreed surgical strategy for their management. Sagi et al. reported that up to 37% of LC1 fractures are unstable, highlighting the potential role of surgical stabilisation [8]. Surgical fixation may improve time to mobilisation, reduce complications associated with prolonged immobility, and improve long-term functional outcomes [9][10][11]. Despite this, the risk of surgical site infection, neurovascular injury, reoperation, and the risks of general anaesthesia are potential risks of surgical intervention [10].
Historically, LC1 injuries have been managed non-surgically with restricted weight-bearing, physiotherapy, and walking aids [4]. The contemporary rationale for surgical management of LC1 injuries is to achieve stability, but this depends on the independent clinical judgment of the stability of the injury, which has been demonstrated to be inconsistent among surgeons [9,12]. Uncertainty concerning the definition of 'instability' remains; some surgeons state it depends on the tendency of the fracture to displace over time, while others define it in terms of excessive pain on mobilisation regardless of the bone position (adopting the definition of 'functional' instability) [6,12].
This study aims to systematically review the existing literature to explore if there are any differences in patient-reported outcomes, complications, time-to-mobilisation, cost-effectiveness, and length of hospital stay between surgically and non-surgically treated unstable LC1 non-fragility fractures.

Data Sources and Search Strategy
We performed a systematic review per the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines based on a pre-defined methodology registered with the International Prospective Register of Systematic Reviews (PROSPERO ID: CRD42021256075). We systematically searched Medline, Embase and Cochrane (Central) databases from inception to 5 June 2022 for studies reporting on patient-reported outcome measures following surgical versus non-surgical management of unstable LC1 pelvic fractures. The International Clinical Trial Registry Platform was also searched for completed, ongoing, and registered trials. The search terms included a combination of Medical Subject Heading (MeSH) terms, synonyms, and related terms for LC fracture, surgical fixation, and non-surgical management. The detailed search strategies are outlined in Appendix 1. The results of the searches were then integrated into Rayyan [13] online bibliographic software.
One reviewer (JV) reviewed the titles and abstracts of retrieved studies for inclusion suitability and then obtained the full texts of potentially relevant articles for further evaluation. Two reviewers (JV, MFS) independently evaluated the full texts of the articles identified, and any disputes regarding eligibility criteria were discussed with and resolved by a third author (RLD).
Studies that (i) included patients with fragility fractures resulting from low-energy trauma (fall from standing height or less); (ii) included children aged 16 years or younger; or (iii) included animals or cadavers were excluded. Two authors (JV, MFS) independently extracted the relevant data from eligible studies. Any disputes were explored and resolved by a third author (RLD).

Risk of Bias and Data Synthesis
The Cochrane Risk of Bias in Non-randomised Studies -of Interventions (ROBINS-I) tool [14], a validated tool for assessing the quality of non-randomised studies, was selected to assess the risk of bias in observational studies. Continuous variables were represented as mean differences (MDs) or standardised mean differences (SMDs) with 95% confidence intervals (CIs). Risk estimates (risk ratios) were used as the standard measure of association across studies for dichotomous variables with 95% CIs. A two-sided statistical significance of 0.05 was used. Due to a paucity of data, we were unable to conduct a formal metaanalysis and have, therefore, provided a narrative review of the findings.

Study Identification
The PRISMA flowchart in Figure 1 summarises the study identification and selection process in detail. After removing 1,853 duplicates from the initial searches, 2,393 studies were screened; 2,373 studies were removed after screening titles and abstracts, and an additional 15 studies were excluded after full-text evaluation. Finally, five studies [15][16][17][18][19] were included in this systematic review.  Table 1 summarises the characteristics of the studies included in the review. The included papers were published between 2016 and 2021. A total of 490 LC1 (or equivalent) injuries were included, of which 183 (37%) were treated surgically. Four studies were retrospective observational studies [15][16][17][18], and one was a prospective study [19]. All studies compared surgical versus non-surgical treatment of LC1 (or equivalent) injuries. Most patients were female and in their third or fourth decades of life. Two studies were performed in the United States and three in Europe (Germany and Lithuania). Three studies used shorter-term followup time points ranging from 24-48 hours [15] to 1-12 weeks [16,19] post-injury, one study had a mean follow-up period of 47 months (standard deviation (SD): 13-82 months) [17], and the largest follow-up duration was a mean of 52 months (SD: 48-84 months) [18]. Primary outcome data from the five individual papers are listed in Table 2. Post-injury -At least at 1 year, mean 47 months

Risk of Bias
Two studies had a serious risk of bias due to confounding, while the remaining three had a moderate risk of bias, according to the ROBINS-I tool. Appendix 2 contains the detailed risk of bias assessments for the included studies.
The EQ-5D questionnaire is a standardised scoring system that assesses a patient's QoL across five domains, including physical and mental health (scale: -0.594 to 1.000, where 0 represents death, and less than 0 represents a QoL worse than death). After a follow-up of at least one year, Höch et al. observed no significant difference in EQ-5D between surgical (mean: 91.8, SD: 9.3) and non-surgical (mean: 89.5, SD: 18.4; p > 0.05) treatment groups (EQ-5D scores were multiplied by 100). Similarly, after a mean follow-up of 4.3 years, Zwingmann et al. reported no significant differences in EQ-5D scores between the surgical (mean: 0.88, SD: 0.14) and non-surgical groups (mean: 0.85, SD: 0.14; p = 0.965).
The SF-36 questionnaire (scale: 0-100, where the lower limit represents extreme disability/symptoms, and the upper limit represents no disability/symptoms) is another standardised scoring system that assesses a patient's QoL. Petryla  Pain: Four studies assessed pain using the VAS (scale: 0-10, where 0 represents no pain, and 10 represents the worst pain) [15,16,18,19]. Some results were presented as an overall surgical treatment effect, a single figure denoting the difference between the mean score in the non-surgical cohort versus the surgical cohort. Associated complications: Complications after treatment of LC1 injuries were discussed in two studies [16,17]. Höch et al. were the only researchers to detail complication rates in both treatment groups. The surgical cohort had a complication rate of 6% versus 4% in the non-surgical cohort, with no statistical significance (p > 0.05). Bladder injury, pneumonia, acute respiratory distress syndrome, surgical malpositioning, and infected metalwork were among the reported complications. Petryla et al. only reported the complications in the surgical cohort, including wound infection (n = 1), screw migration (n = 2) and S1 neuropathy (n = 4), yielding a total surgical complication rate of 18.9%.
Days to mobilisation: Hagen et al. measured the number of days required for the patient to transfer independently from the bed to the chair. The surgical treatment group had a mean reduction of 1.7 days to independent mobilisation compared to the non-surgical group (mean difference (surgical minus nonsurgical): -1.7 days, SD: 0.08, 95% CI: -3.3 to -0.01; p = 0.034).

Discussion
This is the first systematic review to investigate the surgical versus non-surgical treatment of unstable LC1 injuries in the non-fragility fracture population. Five observational studies were eligible for inclusion in this study. Two of the included studies displayed a serious risk of bias, with the remainder deemed to display moderate risk.
Three studies used the SF-36 or EQ-5D to assess QoL, with no statistical differences between treatment groups. Two studies also demonstrated no correlations between functional outcomes at a mean follow-up of ten weeks and four months, respectively.
Surgical fixation was positively correlated with reduced pain in the short term, particularly from the postoperative period to discharge from the hospital and up to 12 weeks post-injury. Surgical fixation was also found to reduce the time to mobilisation by 1.7 days, measured in one study. However, in two studies investigating longer-term outcomes, there was no statistical difference in pain between treatment groups after four to five years. Despite these results, there is ongoing debate regarding the minimal clinically important difference (MCID) for VAS in patients with these injuries. The VAS MCID for pain due to osteoarthritis of the knee, for example, has been demonstrated to be a 2.5 score reduction for surgical or rehabilitative therapy [20]. In the aforementioned study, the differences in VAS were 1.7 and 1.1 at six weeks and 12 weeks follow-up, respectively. Nonetheless, the VAS has proven reliable for pain measurement among patients of all ages and educational levels [21].
Few studies investigated post-treatment complications but found no statistical difference in their results, albeit they are likely to be underpowered for determining true statistical significance. In Höch et al., although the overall complication rate was negligible between the non-surgical and surgical cohorts (11.5 vs. 16.7%, p > 0.5), 8% of the complications in the surgical group were directly related to the surgery.
Although LC1 injuries include both unstable and stable fractures, this review has focused on those with unstable injury patterns. Currently, there are conflicting views on what constitutes an unstable LC1 injury, with some authors basing their criteria on the degree of fracture displacement [22] and others on the patient's pain, for example [12]. A recent cadaveric study also revealed that the displacement of LC1 injuries may be related to pubic rami fracture morphology with oblique pubic rami fractures being most susceptible to displacement over time [23].
Treatment of LC1 injuries varies with no agreed algorithm for determining whether patients would be better off receiving surgical or non-surgical care [24]. Beckmann et al. used the expert opinion of 111 orthopaedic surgeons to develop a validated radiographic scoring system for LC1 fractures [24]. Specific parameters, such as ramus displacement, Denis classification, and sacral displacement, were included in the system, with a score greater than nine out of 14 indicating the need for surgery [25]. Similar efforts have been proposed by Sagi et al. suggesting >1 cm and Giannoudis et al. advising ≥2 cm of pubic rami displacement as cut-offs for surgical intervention [8,26]. Despite these efforts, Hadeed et al. outlined that fellowship-trained orthopaedic surgeons often exhibit inconsistent measurement of sacral fracture completeness in LC1 injuries [27].
Although there have been attempts to study the optimum management algorithm, only one clinical trial has been published to date. In patients with minimally displaced LC fractures of the pelvis, Slobogean et al. discovered that surgical fixation results in a minor but sustained improvement in pain and function for up to 12 months. Patients with a 5 mm displacement were the best surgical candidates, with the most short-term pain relief after surgery [28]. However, because the cohort contained both LC1 and LC2 fractures, this study was excluded from our review.
Lykomitros et al. demonstrated that patients with sacral fractures treated non-surgically had better SF-36 scores in all domains than those treated surgically. The authors highlighted that patients treated surgically had more concomitant injuries, resulting in a higher Injury Severity Score (ISS), a possible confounder for the lack of improvement in QoL following surgical management [29]. Similarly, Tosounidis et al. reported that surgical stabilisation of LC1 injuries reduced pain and analgesic requirements significantly in the immediate post-injury period [26]. However, this study was excluded from our review due to the absence of complete sacral fractures in the non-surgical comparison group.
The existing literature is very limited in the reporting of function and postoperative complication outcomes. A feasibility study has been conducted [30], demonstrating that it would be feasible to recruit into a definitive RCT comparing the surgical and non-surgical treatment of LC1 fractures. A trial could further investigate PROMs, function, pain, QoL, complications, and economic evaluations.
Our review has several limitations. All of the included studies were observational, inherently increasing the risk of confounding and selection bias; this was emphasised in the risk of bias assessment, as all studies were of moderate-to-serious risk of bias. The severity of concomitant injuries was not controlled for; for example, a polytraumatised patient may be more likely to undergo surgical intervention than a patient with an isolated unstable LC1 fracture. The surgical group tended to have more severe injuries, perhaps more significant pain, and reduced QoL. We were unable to perform a formal meta-analysis due to heterogeneity of the time points at which outcomes were measured in individual studies and a lack of consistency in reporting. There were variations in surgical technique, which further increased the risk of confounding. Furthermore, there is no widely accepted protocol for LC1 fracture management, and each study had variable eligibility criteria for surgical intervention: some placing greater emphasis on pain and mobility, and others on initial displacement on static radiographs to aid decision-making.

Conclusions
Unstable LC1 fractures are a heterogenous injury pattern, with no current agreed protocol for surgical intervention. The studies that met our eligibility criteria demonstrated small improvements in short-term pain and days-to-mobilisation with surgical fixation, but no statistically significant differences in functional outcomes or QoL; however, all included studies were of moderate-to-serious risk of bias. This review highlights the low quantity and quality of existing data on patients with unstable LC1 pelvic non-fragility fractures and the need for a definitive RCT to determine whether surgical or non-surgical care should be the preferred treatment in terms of clinical and cost-effective care.