Introduction

Trauma has become a growing epidemic over the years, one of the major causes of presentation at the emergency department (ED) that medical physicians have to encounter on a daily basis (1). This incessantly growing phenomena has been responsible for an increasing number of deaths throughout the years, causing an estimated number of five millions deaths every year, all over the world, being almost equal to the number of deaths caused by HIV, malaria and tuberculosis all three combined (1). Blunt trauma occurring during road traffic accidents is the predominant cause of death (23%). However, penetrating trauma whether by gun shots or by stab wounds (SWs) has a major impact on the healthcare system in so many regions of the globe (2,3). Stab wounds in particular are mostly prevalent in European countries, the United States and Australia (2,4).

By definition, SWs are considered as sharp force injuries and happen when a sharp-edged or pointed object comes into contact with the skin and whether or not the underlying tissues (5). SWs can be penetrating or perforating. Penetrating stab wounds (PSWs) refer to an injury occurring when the weapon goes through the body cavity or attend muscles depending on location (6). The circumstances during which SWs occur, are by order of frequency, homicidal, suicidal, or accidental (5,7). Studies agree that thoracic and abdominal injury are the most predominant sites (1,8). However, other studies agree that upper extremities injuries are to be more frequent (1,8,9). PSW can have dreadful consequences when located in the thoracic and abdominal area, causing shock, infection, and more dramatically death (2). A study conducted at Howard University Hospital has counted 11% of deaths caused SWs (10). To avoid such high risk of mortality, a rapid management of SWs must be ensured (9). The aim of this study was to assess the different epidemiological, and clinical features of SWs in the ED as well as prognostic factors determining short-term evolution (STE).

Methods

This is across-sectional single-center study conducted in the polyvalent ED room. Patients presenting to the ED for assault-induced SWs during a ten-month period, from March to December 2020, were enrolled. The inclusion criteria were: patients of both male and female genders, who are aged 14-years-old or more, patients must have at least one assault-induced SWs, and weapons used to inflict the injury can be stabbing or thrusting tools (knives, daggers, screwdrivers …). The non-inclusion criteria were: gunshot injuries, road accidents, and self-inflicted injuries. As for the exclusion criteria, they were: patients who left the ED without medical notice or who signed their release against medical advice before initial management, and patients who are already dead before their arrival to the ED.

The patients recruited in this study were identified during their presentation to the ED. They all went through the triage process and were all admitted to the ED. Clinical information was gathered using a well-established data sheet filled with information about each patient. This data sheet included the patient medical file’s number, their age and gender, as well as anamnestic data, physical examination data, laboratory blood tests, medical imaging results, as well as the prescribed treatments.

The STE is a composite variable included seven items: oxygen therapy, volume expansion, noradrenaline use, transfusion, insertion of chest tube, intubation, and urgent surgery. If the patient presented one of these items, he would be considered as having an unfavorable STE.

Statistical analysis

All data was recorded and analyzed using SPSS 22.0. The threshold for statistical significance ‘p’ was determined to be at least 0.05. The descriptive study analysis concerned the study population (SP) and the subgroups of patients who have penetrating stab wounds (PSWs) and those who have superficial stab wounds (SSWs).

The risk factors of STE search was carried out by an univariate analysis (Odds ratio, 95% confidence interval (CI). Binary logistic regression was performed to analyze the association between risk factors and STE. The threshold of the quantitative variable was established the Receiver Operating Characteristics curves (ROC). 

Results

We enrolled 130 patients admitted to the ED from March until December 2020 for having an assault-induced SW. Among these patients, 30 patients had penetrated SWs, and 100 had superficial SWs. Our study population included 122 males and 8 females. Gender-ratio was equal to 15.25. Median age was at 30 years, ranging from 15 to 66 years-old, with respective 25^th and 75^th percentiles of 24 and 37 years. Ninety-eight patients (75.4%) had a medium SES (n=98), and 30 patients (23.1%) had a good SES (n=30). A hundred and three patients (79.2%) were smokers, 60 patients (46.2%) were alcohol consumers, and 20 patients (15.4%) were drug users.

Twenty-seven patients (20.8%) were admitted on a Sunday, and 23 patients (17.7%) on a Monday. Thirty-one SWs occurred during June (23.9%), and 66 (50.8%) during the night. Stab wounds occurred in 64.6% of the cases (n=84) at home or nearby, and in the rest of the cases, on the streets. Weapons used to inflict the injury were cutting and thrusting tools: knives (n=108,) sabers (n=8), daggers (n=23), in all of the cases, except one patient who was stabbed by a screwdriver. Forty percent of the injuries were located on the torso (n=52). The rest of the locations are by order of frequency: the extremities (n=38, 29.2%), the face (n=23, 17.7%), the scalp (n=22, 16.9%) and the pelvic (n=2, 1.5%). As for wounds that were located on the torso, they are consecutively and by order of frequency abdominal (20%), thoracic (12.3%), and thoraco-abdominal (11.5%). One hundred and eight patients presented with only one SW (83.1%). (Table 1)

Mean RR was at 19±3cpm ranging from 16cpm to 28cpm. Seven patients (5.4%) were hypoxemic, three patients (2.3%) had signs of respiratory distress, and six patients had subcutaneous emphysema. Mean SBP was at 121±17 mmHg. Four patients (3.1%) had a SBP≤90mmHg and eight patients (6.15%) had a SBP<65mmHg. Thirty-two patients (24.6%) have had a HR>90 bpm on the initial physical examination. Two patients (1.5%) had an altered neurological status. Thirty-eight patients (29.2%) had SW located on the extremities. Among these patients, five had PSW (13.2%) of the extremities, and three among these five patients presented signs of acute limb ischemia. Fifty-three patients (40.8%) were performed a chest x-ray (CXR). Two CXRs (1.5%) revealed a PNO.

One patient was performed a bedside E-FAST (0.8%) while a hemodynamic resuscitation was initiated. This patient was hemodynamically unstable due to a SW located on the cardiac box. This E-FAST revealed a hemothorax as well as a heart wound complicated by a hemopericardium. Among the 79 patients (60.8%) who were admitted to the RsR, 58 were performed CT scans (44.6%) independently of the location. forty-four CT scans were thoracoabominal and/or pelvic and 14 were cerebral. Among the 58 performed, 23 CT scans (39.7%) showed abnormalities.

Thirty-eight injuries (29.2%) were located on the upper and/or lower limbs. Among these injuries, five were considered penetrating, and in which three had hard signs of vascular injury and underwent surgery, and two had soft signs and were performed an ACT. (Table 2)

In order to evaluate STE, and since in our population there was zero number of deaths in the ED, we created a composite variable indicating whether or not the patient received immediate life-saving therapeutic measures. Seventeen patients (13.1%) were considered as having unfavorable STE in the ED.

Gender and SES were not associated to an unfavorable STE. However, alcohol consumption and drug use were significantly associated with unfavorable STE with respectively crude OR of 3.25 and 3.85. Unfavorable STE were significantly associated with RR, HR, and BP. Respiratory rate was significantly higher among patients who had unfavorable STE compared to those who had favorable STE (p<0.001). (Table 3)

The ROC curve analysis showed an AUC of 0.813, CI 95% [0.699-0.927], indicating a high discrimination ability based on RR (Figure 1). The optimal cut-off value as an indicator of an unfavorable STE was at 19 cpm, with a sensibility of 82% and a specificity of 69%, p<0.001. Systolic blood pressure was significantly lower among patients who had unfavorable STE compared to those who had favorable STE (p<0.001). ROC curve analysis showed an AUC of 0.738, 95% CI [0.585–0.891] (figure 2). The optimal cut-off value of SBP as an indicator of an unfavorable STE was at 105mmHg, with a sensibility of 93% and a specificity of 53%, p=0.002. After logistic regression, we determined two risk factors associated with unfavorable STE. The first one was RR≥19cpm with p=0.018, OR=8.64, and the second one was SBP≤105mmHg, p<0.001, OR= 12.29.

Discussion

Our results had noted male predominance with a gender ratio of 15.25. Patients were young, mostly in their twenties (n=54, 41.5%), median age was 30 years-old. These results are in line with findings in literature. In a population-based study conducted over a 16-year period, by Johannesdottir U et al., a total of 88 patients who had PSW were included, and 73 patients were admitted to the ED. Age-standardized incidence was at 1.54/100,000 inhabitants and mean age was 32.6 years. In this study, a male predominance (90.4%) was noticed (8). Another study conducted by KA.Alenazi et al., over a period of 4 months in 2018 concerning 106 patients, showed that 72.6% of them were aged less than 40 years, and 87.7% were males (7). This male predominance founded in our study as well as in numerous studies raises the idea of a study attempting to compare males and females in terms of demographic and circumstantial factors, clinical features, and outcomes. In our study, 46.2% of the patients were alcohol consumers (n=60) and 15.4% were drug users (n=20). This was commensurate with literature. In a single-center study conducted by Jacob AO et al., 31% of the patients (n=481) were under the influence of alcohol (8). This could be explained by the fact that drug and alcohol use increased aggressiveness according to an article published in Alcohol Alert in 1997 (11). As a matter of fact, S. Manojkumar et al., mentioned in his study published in 2011, that 94% of stabbing victims who required surgical admission, 74% of them were admitted over the weekend. The same research mentioned that alcohol has ‘been identified as one of the most pressing problems. Furthermore, in November 2005, alcohol was allowed to be consumed beyond 11.00 PM and during the weekends, aiming at reducing alcohol induced violence. After that law was applied, a gradual decline between 2006 and 2008 in the general rate of PSWs was noticed (11).

The locations of injuries, in our study, were by order of frequency: the torso, (40%), the extremities, (29.2%), the face (17.7%), and the scalp (16.9%). Torso SW were predominantly abdominal (20%). This is consistent with literature. U. Johannesdottir et al. reported in their study that injuries to the chest, the abdomen, and the upper limbs were more common (26.4%, 21.5%, and 21.5%, respectively) (8). KA.Alenazi et al. reported that SW located on the head and the upper limbs (21.7% each) were more frequent, followed by the abdomen (17%) and the thorax (10.4%) (7). The predominance of SW inflicted to the extremities may be explained by the fact that victims attempted to protect the torso by exposing the upper extremities.

Most of the patients were hemodynamically stable. Four patients (3.1%) had a SBP<90mmHg and eight had a MAP≤65mmHg (6.2%). Mean heart rate was at 87±14 bpm. Seven patients (5.4%) were hypoxemic, and six (4.6%) presented abolished breath sounds on auscultation. Only two patients (1.5%) had an altered neurological status on the initial physical examination. Cervical ACT and cerebral CT were eliminated a traumatic and ischemic cause in particular carotid injury. Our results were similar to findings in a study conducted by C. Heus et al. in a level I trauma center over 159 patients. Mean HR in this study was 94±22 bpm, and 15.8% of the patients had a SBP below 90 mmHg. However, the median GCS was 14, and in 6.2% of the cases the GCS was below 9 (12).

In our study, 38 injuries (29.2%) were located on the upper (76.3%) and lower limbs (n=9, 23.7%). Among these injuries, five were penetrated, and among which three (13.2%) had hard signs of a vascular injury, and two had soft signs (5.3%). This was in line with what was found in a study conducted by H. Warwick et al. and which showed that 71% of patients had a normal vascular exam, 22% had soft signs, and 6% had hard signs (13).

Chest x-ray was performed for 53 patients (60.9%), and showed a PNO in two cases. However, 58 CT-scans were performed. Forty-four were thoracoabdominal and/or pelvic and 14 were cerebral. Five PNO, two hemothoraces, two hemopneumothoraces, and one pneumomediastinum were diagnosed. This led to conclude that CXR sensitivity was insufficient, that its interpretation should be cautious, its indication limited and optimized to include patients with respiratory or hemodynamic instability in case of penetrating trauma. A single-center study conducted by BM. Nguyen et al. over a 10-year period in California, found that CXR was associated with a high incidence of missed injury that was subsequently detected on CCT. It therefore concluded that in hemodynamically stable, asymptomatic patients, CCT was able to detect a large number of injuries in patients with thoracic SWs, allowing rapid diagnosis and disposition (14). Ct-scan is in consequence the primary imaging modality in the evaluation of hemodynamically stable patients (15). Extended focused assessment with sonography for trauma has also proved its accuracy in detecting injures caused by PSW and could even help save time and expenses since it could be a more reasonable alternative to detect PNOs and hemothoraces than CXR. In fact, according to an article published by  K. Charles et al., CXR has a low sensitivity in detecting PNO and even lower sensitivity in detecting hemothorax, compared to E-FAST (16). Ultra-sonography has owned its place in the management of trauma patients especially those in critical status since its bed-side availability and feasibility. Actually, it figures in multiple trauma guidelines (17).

We opted to evaluate STE based on therapeutic measures which were oxygen therapy, airway management, need for chest tube insertion, volume expansion, catecholamine and/or tranexamic acid use, urgent surgery, and death. In our study, 13.1% of the patients (n=17) had an unfavorable STE with no case of mortality occurring in the emergency setting. U. Johannesdottir et al. reported a 30-day mortality at 4.1% compared to 2% in a small center in Norway and 15% in a major urban trauma center in Australia. They explained the divergent rate of death by the differences in severity, body parts, and types of injuries that were included in the studies (8).

After running logistic regression, we determined two risk factors associated with unfavorable STE, which were RR with a cut-off of 19cpm (p=0.018, OR=8.64) and a SBP cut-off of 105mmHg (p<0.001, OR=12.29). RM. Hasler et al. reported in their multicenter cohort study that mortality was doubled among patients with SBP between 90-109 mmHg compared to those with a SBP between 110-129 mmHg and therefore determined SBP <110mmHg as a new threshold —instead of the old SBP <90mmHg—to triage patients with penetrating trauma to resuscitation areas within dedicated high-level care trauma centers (18,19). However, literature’s findings concerning RR as a risk factor related to unfavorable STE was noticed in many prognostic scoring systems such as RTS and qSOFA with respectively a normal interval of (10,17), and a threshold 22cpm. It figures in the recommendations related to thoracic trauma management which fixed a RR≥25cpm as factor predicting the severity of a penetrating trauma (18).

Some weak points have to be noted; it was a single-center study with a small sample size. We did not include the patients’ follow-up after hospital discharge which can underestimate unfavorable evolution.

Conclusion

Unfavorable short-term evolution could be predicted through clinical parameters such as RR≥19cpm and/or SBP≤105 mmHg in PSWs’ victims. These factors’ thresholds should probably be integrated in a new scoring system or assessment tools. A multicenter study is necessary to validate the usefulness of the composite indicator of the STE.