Diagnostic laparoscopy or selective non-operative management for stable patients with penetrating abdominal trauma: What to choose?

BACKGROUND: Selective non-operative management (NOM) and diagnostic laparoscopy (DL) are well-accepted approaches in the management of stable patients with penetrating abdominal trauma (PAT). The aim of this pilot study was to investigate the advantages and disadvantages of early DL in stable asymptomatic or minimally symptomatic patients with PAT as opposed to NOM, a standard of care in this scenario. The secondary aim was to suggest possible indications for DL.
METHODS: Patients managed with DL or NOM over a 12-month period were included in this study. The age, gender, mechanism and location of injuries, trauma scores, haemodynamic and metabolic parameters, intraoperative findings and length of hospital stay (LOS) were recorded and correlated with outcomes.
RESULTS: Thirty-six patients were in the NOM group and 35 in the DL group. Stab wounds were more common. The most common location of injury was the anterior abdominal wall in the NOM group and the lower chest in the DL group. Computed tomography (CT) scan was performed more often in the NOM group (75% vs. 17.1%). The injury severity score (ISS), New ISS and PAT Index were higher in the DL group. Nearly 23 (66%) patients in the DL group had a penetration of the peritoneum, but no significant abdominal injuries. LOS in the NOM group was 2 days versus 3.1 days in the DL group. There were no missed injuries, complications or mortality.
CONCLUSION: NOM is a preferred modality for minimally symptomatic stable patients. However, there is a risk of missed injuries and delayed treatment. DL accurately visualizes injuries, decreases unnecessary CT scans and avoids nontherapeutic laparotomies.

INTRODUCTION

Since the introduction of bladed weapons and the invention of firearms, the management of penetrating abdominal trauma (PAT) has been changed many times. Initially, PAT was managed mainly non-operatively with high mortality. After the first reported laparotomy for PAT in 1834, the mortality rate significantly decreased, and the concept of mandatory laparotomy was a standard of care.[1] Today, the concept of 'selective conservatism’ is widely accepted.[2] Recently, laparoscopy for trauma has become more popular and is proven to be feasible and safe.[34]

It is uniformly accepted that haemodynamically unstable patients should be managed with immediate laparotomy. However, the optimal management of stable patients with PAT is still under debate. Traditional laparotomy is sufficient, however, morbidity is reported to be high. In the presence of appropriate expertise, laparoscopy has more benefits than laparotomy with a minimal intraoperative cost increase and is further enhanced by fewer complications, a shorter length of hospital stay (LOS) and faster recovery. Modern imaging techniques, combined with thorough clinical examination, have decreased the rate of nontherapeutic laparotomy to 24%.[5] Diagnostic laparoscopy (DL) allows the detection of intra-abdominal injuries with a reported accuracy of 100% in recent studies.[346] DL is a minimal but still invasive technique requiring general anaesthesia. The concept of selective non-operative management (NOM) was reintroduced in the 1960s.[1] In South Africa, the 'selective conservatism’ philosophy became a necessity to provide care for an overwhelming number of patients with PAT.[2] Selective NOM has been well studied and currently is being practiced by many surgeons.[78] It is considered to be a safe approach in carefully selected patients with a failure rate from 17% to 20%.[79] On the other hand, early DL may accurately confirm the absence of injuries and the patient may be discharged as early as 6 h after the procedure.[10]

The aim of this pilot study was to investigate the advantages and disadvantages of early DL in stable asymptomatic or minimally symptomatic patients with PAT as opposed to NOM, a standard of care in this scenario. The secondary aim was to suggest possible indications for DL.

METHODS

The study was performed at a single institution providing trauma services for a population of 7.3 million people.

The prospectively collected database of all patients with PAT was analysed. All patients managed non-operatively or with DL from January to December 2015 were included in the study. Patients, who failed NOM or converted to laparotomy, were excluded from the comparison. The cases of patients younger than 12 years and patients with missing records were also excluded from the study. All included patients were allocated into either the DL or NOM group.

All patients in the DL group were asymptomatic or minimally symptomatic; however, operating surgeon had suspicion for possible injury and opted for laparoscopy. The two groups were similar regarding the absence of intra-abdominal injuries requiring any therapeutic options. Cases of failed NOM or converted to laparotomy were excluded to avoid heterogenicity of therapeutic procedures. Retrospectively, all patients could have been successfully managed non-operatively or, theoretically, had been even safely discharged without further investigations.

All patients were managed according to the Advanced Trauma Life Support guidelines. Informed consent for intervention was obtained in all cases.

Penetration of the abdominal cavity was defined as a violation of the parietal peritoneum or retroperitoneal haematoma signifying the possibility of injury to the intra- or retroperitoneal organs. DL was defined as a procedure when there were no injuries or the identified injuries did not require any repair. Evacuation of liquid blood or clots and mobilisation of any intra- or retroperitoneal organs for diagnostic purposes were not considered as therapeutic. DL was subsequently subdivided as follows: (1) diagnostic without penetration and (2) diagnostic (nontherapeutic) with penetration of the abdominal cavity but not requiring any therapeutic procedures.

NOM included serial clinical examinations of the patient, preferably by the same clinician, and laboratory testing over a minimum of 24 h. The decision to select the patient for NOM or for DL was made by the attending surgeon. All surgeons were proficient with advanced laparoscopy and able to perform trauma laparoscopy.

Patients were allocated into the DL or NOM group. The absence of significant injuries was confirmed by DL and clinical recovery in the DL group and by serial physical examination and clinical recovery in the NOM group. All patients were requested to come for a follow-up visit in 1 week.

The patients’ age, gender, mechanism and location of injuries, coexisting injuries in other anatomical regions, haemodynamic and metabolic parameters, type of laparoscopy, intraoperative findings, therapeutic and diagnostic procedures were recorded and correlated with outcomes. The Injury Severity Score (ISS), New ISS (NISS) and the PAT Index (PATI) were calculated. Significant complications (Clavien-Dindo Grades 3–5) were recorded and investigated.[11] LOS and reasons for a prolonged stay were analysed.

Data analysis

Demographic details of the patients were summarised descriptively by mean, median, minimum and maximum values for continuous variables and frequency counts and percentage calculations for categorical variables. Clinical outcomes after laparoscopic surgery were described descriptively in relation to the procedural outcomes. Mean values were compared by the two-sample t-test. Median values were compared by the nonparametric Wilcoxon rank-sum test. P < 0.05 were considered statistically significant. An analysis of variance was performed for each of trauma scores, for comparisons of mean values for outcome, followed by pairwise comparison of least square means by t-test.

The study was approved by the University Research Ethics Committee.

RESULTS

Over the period of 12 months, 38 patients were managed non-operatively and 36 patients underwent DL [Figure 1]. Two patients (5%) failed NOM and were excluded from the study. One patient developed signs of peritonitis and underwent DL (failed NOM). The second patient refused any further treatment and left the hospital (missed data). In the DL group, one (3%) patient was converted to laparotomy due to large retroperitoneal haematoma and was excluded from the study. In total, 36 patients in the NOM group and 35 patients in the DL group were analysed [Table 1]. Significantly more patients with stab wounds were managed with laparoscopy (85.7% vs. 52.8%, P = 0.004). The most common location of injury was the anterior abdominal wall in the NOM group versus the lower chest in the DL group. Twenty-seven patients in the NOM group underwent computed tomography (CT) scan as compared to only six patients in the DL group (75% vs. 17.1% respectively; P < 0.001). Inversely, only nine patients did not have any imaging in the NOM group versus 29 in DL group. All patients with gunshot wound (GSW) in the NOM group were imaged with CT scan but only one with ultrasound. GSW was located on the flank in seven patients, on the anterior abdominal wall in six patients and on the back in four patients. The ISS, NISS and PATI were significantly higher in the DL group (5.08 vs. 6.8; 5.08 vs. 7.26; 0.19 vs. 1.3; respectively, P < 0.001). Two patients in the NOM group had an intercostal drain (ICD) for pneumohaemothorax. Both had an additional injury to the chest caused by GSW in one patient and stab in another patient. Eleven patients in the DL group presented with pneumohaemothorax and eight of them underwent thoracoscopy. Thoracoscopy was performed for evacuation of clotted haemothorax documented by post-ICD chest X-ray. Approximately 23 (66%) patients in the DL group had a penetration of the peritoneum; however, no significant abdominal injuries. Two patients in the DL group had iatrogenic minor injuries during laparoscopic exploration. One patient had deserosation and pinpoint perforation of the small bowel and deserosation of the ascending colon during the right colon mobilisation. The injury was reinforced with seromuscular (Lembert) sutures. The second patient had a minor splenic injury during the diaphragm exploration and bleeding was easily controlled with a diathermy. Both patients recovered uneventfully.

Figure 1

Non-operative management group and diagnostic laparoscopy group

Table 1

Baseline characteristics of patients with penetrating abdominal trauma in non-operative management and diagnostic laparoscopy groups

Characteristics	NOM group (%)	DL group (%)	P
Total	36	35
Male	30 (83.3)	31 (88.6)	0.735
Female	6 (16.7)	4 (11.4)
Age	29.14 (14-51)	29.49 (17-59)	0.756
Mechanism
Stab	19 (52.8)	30 (85.7)	0.004
GSW	17 (47.2)	5 (14.3)
Location of injury
Anterior abdominal wall	14 (38.9)	12 (34.3)	0.002
Back	9 (25)	1 (2.9)
Flank	10 (27.8)	8 (22.9)
Lower chest	3 (8.3)	14 (40.0)
CT not done	9 (25)	29 (82.9)	<0.001
CT scan	27 (75)	6 (17.1)
Negative	24	3
Positive	3	3
ISS	5.08 (4-13)	6.8 (4-13)	0.007
NISS	5.08 (4-13)	7.26 (4-13)	<0.001
PATI	0.19 (0-3)	1.3 (0-6)	0.002
Pneumohaemothorax	2	11
Thoracoscopy	0	8
Evisceration
Bowel	0	3
Omentum	0	4
DL, peritoneum
No penetration		12 (34)
Penetrated		23 (66)

NOM: Non-operative management, DL: Diagnostic laparoscopy, GSW: Gunshot wound, ISS: Injury Severity Score, NISS: New Injury Severity Score, PATI: Penetrating Abdominal Trauma Index

LOS in the NOM group was 2 days (1–13) versus 3.1 days (1–8) in the DL group (P = 0.009). Patients without ICD stayed in hospital for a shorter period; 1.62 (1–4) days in the NOM group and 2.83 (1–8) days in the DL group (P < 0.001). The ICD was responsible for the longer LOS but the difference was not statistically significant; 8.5 (4–13) days for the NOM and 3.77 (2–8) for DL group (P = 0.483) [Table 2]. Three patients in the NOM group and ten patients in the DL group stayed in hospital more than 3 days [Table 3]. Seven patients had pneumohaemothorax and ICD. Four patients with prolonged LOS did not have peritoneal penetration on DL but all of them had pneumohaemothorax and ICD. The patient with iatrogenic minor injury to the bowel (deserosation) stayed in hospital for 6 days for observation. The patient with the iatrogenic minor splenic injury had concomitant clotted haemothorax and underwent thoracoscopy, he stayed in the hospital for 4 days. No missed injuries were recorded in all trauma patients managed with laparoscopy over the study. No significant post-operative complications or mortality were reported in the DL group. All patients in the NOM group recovered successfully, and there were no complications or mortality reported.

Table 2

Length of hospital stay in patients with penetrating abdominal trauma in non-operative management and diagnostic laparoscopy groups

	NOM group	DL group	P
Total	2 (1-13)	3.1 (1-8)	0.009
ICD
no ICD	1.63 (1-4)	2.83 (1-6)	<0.001
ICD	8.5 (4-13)	3.77 (2-8)	0.483
Mechanism
Stab	1.79	3.1	<0.001
GSW	2.24	3.6	0.360
Location of injury
Anterior abdominal wall	1.57	2.42	0.043
Back	2.11	6.00	0.008
Flank	2.50	3.22	0.579
Lower chest	2.00	3.57	0.148

NOM: Non-operative management, DL: Diagnostic laparoscopy, GSW: Gunshot wound, ICD: Intercostal drain

Table 3

Patients with prolonged length of hospital stay

Patient (year old)	LOS	Group	Mechanism	Comment
Female, 29	13	NOM	GSW	Bilateral pneumohaemothorax, ICD, right kidney injury Grade 2
Male, 28	4	NOM	Stab	Pneumohaemothorax, ICD
Male, 50	4	NOM	GSW	ICU, high care, not improving clinically
Female, 24	5	DL	Stab	Pregnant 8 weeks, Grade 1 Colon injury
Male, 42	9	DL	GSW	Bilateral pneumohaemothorax, thoracoscopy, ICD
Male, 31	5	DL	Stab	Small bowel evisceration, not improving clinically
Male, 30	4	DL	Stab	Pneumohaemothorax, ICD
Male, 23	4	DL	Stab	Not improving clinically
Male, 50	4	DL	GSW	Colon mobilised, not improving clinically
Male, 28	4	DL	Stab	Pneumohaemothorax, ICD
Female, 21	5	DL	Stab	Pneumohaemothorax, thoracoscopy, ICD
Male, 59	6	DL	Stab	Iatrogenic colon and small bowel injuries, seromuscular stitches
Male, 32	4	DL	Stab	Thoracoscopy, ICD, iatrogenic splenic injury, diathermy

LOS: Length of hospital stay, NOM: Non-operative management, DL: Diagnostic laparoscopy, GSW: Gunshot wound, ICD: Intercostal drain

Most patients in the NOM group failed to return for a follow-up visit. There were no complications and mortality recorded in patients presenting for a follow-up visit 1 week after discharge.

DISCUSSION

Selective NOM of PAT is well studied and practiced worldwide.[78] NOM reduces the rate of unnecessary nontherapeutic surgical exploration; however, at the same time, it may delay the necessary surgery. Such delay may increase morbidity and mortality. Careful patient selection and close observation are the critical components of safe NOM. Patients are admitted to high-care facilities for continuous haemodynamic monitoring. Frequent laboratory tests, including 4-hourly haemoglobin, white cell count and lactate, are done. Serial physical examinations are performed, preferably by the same, experienced team, in the absence of brain, spinal injuries, intoxication, narcotics, antibiotics or anaesthesia.[81213] Diffuse abdominal tenderness is associated with significant injuries. In contrast, the localised tenderness may be caused by soft-tissue damage or by insignificant intra-abdominal injuries.[12]

In most cases, 24-h observation is sufficient to discharge the patient.[131415] In unequivocal cases, oral feeding is introduced and the observation may be extended to 72 h.[816]

NOM may be unreliable in the presence of distracting injuries and unequivocal tenderness around the wound and in an environment where the same surgical team is not available.[10] In this situation, DL may be the investigation of choice. The accuracy of DL in detecting abdominal injuries is close to 100% in recent studies.[34] In this study, there were no missed injuries.

Overall, 25% of patients with PAT would be selected for NOM.[8] The failure rate of NOM is reported as high as 20%.[7] In these patients, the necessary surgical intervention will be delayed to the average of 30–40 h (range, 8 h to 5 days).[78] The delay up to 12 h is considered to be safe, whereas surgery performed after 12 h is associated with increased morbidity.[717] However, mortality is not affected by the delay.[818] In this study, the failure rate of NOM was 5% (2/38).

The single small, randomised study comparing DL to NOM was identified by Oyo-Ita in his recent Cochrane review.[19] The authors recommended that future randomised controlled studies are needed to make a conclusion.

The main role of imaging of the patients selected for NOM is to detect early who will fail NOM. Abdominopelvic CT is strongly recommended to facilitate initial management decisions. Focused assessment with sonography for trauma is noninvasive and reliable in diagnosis of free intraperitoneal fluid, but its utility is limited in PAT. It cannot distinguish fluid originating from a solid organ or a hollow viscus.

CT allows the tract mapping and grading of injuries.[12] CT tractography is reported to identify peritoneal violation with an accuracy of 100%.[20] The overall sensitivity of the CT in detecting bowel and mesenteric injuries was 63.6%; specificity was 79.6%, positive predictive value only 53.9% and accuracy 75.3%. The majority of patients with missed bowel injuries demonstrated only indirect signs of injury.[21] In patients selected for NOM, CT predicts the need for laparotomy with sensitivity (94.90%), specificity (95.38%), negative predictive value (98.62%), positive predictive value (84.51%) and accuracy (94.70%).[22] In nonoperatively managed patients, CT will assess the grade of solid organ injury and the need for angiography.[18] On the grounds that most patients who failed NOM had free fluid on CT and only 5.9% failed without it, Bennett et al. suggested DL in patients with free fluid on CT.[15]

Di Saverio et al. consider DL as a valid alternative to CT.[23] In the current study, the need for CT was significantly reduced in the DL group (17% vs. 75%). We support Navsaria et al. that DL cannot replace CT.[24] Pre-operative CT may guide the laparoscopic exploration and prevent unnecessary extensive mobilisation of abdominal viscera cleared by imaging.

CT findings may identify four possible groups: (1) obvious intra-abdominal pathology (immediate surgery); (2) fully extra-abdominal trajectory (potential discharge); (3) isolated solid organ injury (observation, potential angioembolisation or surgery); and (4) patients with concern for intra-abdominal trajectory but unclear injury (can be observed).[12] The last group may benefit from laparoscopy and early discharge. Although we do not recommend routine CT scan before DL, the pre-operative track mapping and grading of injuries may be of benefit.

Both NOM and DL significantly reduced LOS in comparison to laparotomy. There is little data published on a comparison of these two approaches. The earliest safe discharge after NOM was reported to be 24 h and only 6.4 h for DL.[1013] The average LOS is 2–3 days for NOM.[915] Moreover, in the event of failure, this time is added to post-operative stay. In this study, the LOS for the NOM group was 2 days (1–13) versus 3.1 days (1–8) in the DL group (P = 0.009). Patients with pneumohaemothorax and ICD management stayed in the hospital longer; however, the difference was not statistically significant. In this study, LOS for DL was significantly longer (3.1 vs. 2 days). However, the interpretation of LOS should be done with caution as we did not have strict discharge criteria and logistical issues unnecessarily prolonged LOS.

Minor iatrogenic injuries during laparoscopy are possible and we report two minor iatrogenic injuries. In these patients, the prolonged LOS was accounted for concomitant clotted haemothorax managed with thoracoscopy (4 days) in one patient and overcautious observation (6 days) in another patient.

To identify patients for early discharge, the laparoscopic assessment in surgical trauma (LAST) was developed. It is performed in the emergency room under local anaesthesia and sedation. Patients with negative findings can be discharged from the emergency room after a short period of recovery.[25] LAST is an accurate and safe diagnostic tool in the management of patients with an equivocal penetrating stab wound.[26]

Serial examination may be misleading in patients with PAT to the thoracoabdominal region.[7] da Silva et al. performed DL to rule out diaphragmatic injury after 24–36 h of successful NOM.[27] We prefer an early laparoscopic assessment of abdominal injuries including the diaphragm.

Despite the discouraging recommendations on the use of laparoscopy to identify hollow viscus injuries, we reported high accuracy in detecting hollow viscus injuries.[41328] In addition, our data are in agreement with recent reviews.[329]

The benefits of DL as compared with negative laparotomy for PAT are well documented. DL significantly lowers the risk of wound infection and pneumonia. It shortens LOS and avoids nontherapeutic laparotomies.[29]

The cost of successful NOM ($1580) is cheaper than surgery, although the cost of failed NOM is not reported. The cost of DL is similar to that of negative laparotomy ($8,318 vs. $8,640).[10] This study did not compare the cost. It is clear that NOM is preferable option due to its low cost and noninvasive nature.

Based on this study, we suggest that NOM should be opted in cases of:

Haemodynamically stable, asymptomatic or minimally symptomatic patients

Soft abdomen without any tenderness away from the wound

The absence of distracting injuries and the absence of brain, spinal injuries, intoxication, narcotics, antibiotics or anaesthesia

The availability of facilities for continuous haemodynamic monitoring and frequent laboratory tests

The availability of an experienced, preferably the same, surgical team to perform serial physical examinations.

DL should be beneficial in haemodynamically stable patients with:

The presence of unequivocal tenderness around the wound

The presence of distracting injuries, brain and spinal injuries, intoxication, narcotics, antibiotics or anaesthesia

The lower chest (thoracoabdominal) injuries to exclude diaphragmatic injury

Unequivocal findings on imaging (concerns for intra-abdominal trajectory but unclear hollow viscus injury)

Free fluid in the abdominal cavity

The failure to improve after 12 h of serial physical examination, as morbidity increasing afterward

The same surgical team is not available and in circumstances where serial physical examination is not feasible.

‘Limitations of the study

The study was not randomised and there were no strict criteria for discharge, which resulted in inaccurate LOS calculations. The decision to opt NOM or DL was dependent on the surgeon's judgement. The heterogenicity of groups prevents us from drawing any strong conclusion. Most of the successfully recovered patients were lost to follow-up (many patients without problems opted not to come to save on transport expenses). Finally, it is difficult to reproduce similar results in centres without the same.

Future research

Future studies are needed to validate our suggested indications for DL. Prospective evaluation of the feasibility and safety of the early discharge after nontherapeutic DL and its overall cost is required.

CONCLUSION

NOM still remains the well-accepted low-cost noninvasive modality in the management of asymptomatic or minimally symptomatic stable patients with PAT. The benefits of NOM should be weighed against the risks of missed abdominal viscus injuries and delayed treatment. In some circumstances, early DL may offer some benefits over NOM. It accurately visualises intra-abdominal injuries, decreases unnecessary CT scans and avoids nontherapeutic laparotomies. It removes the anxiety of serial examination in the setting of limited work hours and same-surgeon availability. However, the patient may stay in hospital for 1 day longer.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.