Non‐operative management versus operative management in high‐grade blunt hepatic injury
Abstract
Background
Surgery used to be the treatment of choice in cases of blunt hepatic injury, but this approach gradually changed over the last two decades as increasing non‐operative management (NOM) of splenic injury led to its use for hepatic injury. The improvement in critical care monitoring and computed tomographic scanning, as well as the more frequent use of interventional radiology techniques, has helped to bring about this change to non‐operative management. Liver trauma ranges from a small capsular tear, without parenchymal laceration, to massive parenchymal injury with major hepatic vein/retrohepatic vena cava lesions. In 1994, the Organ Injury Scaling Committee of the American Association for the Surgery of Trauma (AAST) revised the Hepatic Injury Scale to have a range from grade I to VI. Minor injuries (grade I or II) are the most frequent liver injuries (80% to 90% of all cases); severe injuries are grade III‐V lesions; grade VI lesions are frequently incompatible with survival. In the medical literature, the majority of patients who have undergone NOM have low‐grade liver injuries. The safety of NOM in high‐grade liver lesions, AAST grade IV and V, remains a subject of debate as a high incidence of liver and collateral extra‐abdominal complications are still described.
Objectives
To assess the effects of non‐operative management compared to operative management in high‐grade (grade III‐V) blunt hepatic injury.
Search methods
The search for studies was run on 14 April 2014. We searched the Cochrane Injuries Group's Specialised Register, The Cochrane Library, Ovid MEDLINE(R), Ovid MEDLINE(R) In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily and Ovid OLDMEDLINE(R), Embase Classic+Embase (Ovid), PubMed, ISI WOS (SCI‐EXPANDED, SSCI, CPCI‐S & CPSI‐SSH), clinical trials registries, conference proceedings, and we screened reference lists.
Selection criteria
All randomised trials that compare non‐operative management versus operative management in high‐grade blunt hepatic injury.
Data collection and analysis
Two authors independently applied the selection criteria to relevant study reports. We used standard methodological procedures as defined by the Cochrane Collaboration.
Main results
We were unable to find any randomised controlled trials of non‐operative management versus operative management in high‐grade blunt hepatic injury.
Authors' conclusions
In order to further explore the preliminary findings provided by animal models and observational clinical studies that suggests there may be a beneficial effect of non‐operative management versus operative management in high‐grade blunt hepatic injury, large, high quality randomised trials are needed.
PICOs
Plain language summary
Is surgery or observation better for people who have a severe blunt injury to the liver?
Background
The liver is the most commonly affected organ when a person is injured in the abdomen. Abdominal injury is usually caused by motor vehicle crashes, falling, being punched in the stomach, or from other causes. When a person is badly injured in the abdomen, they have a 10% to 15% chance of death. According to previous research, the chance of death following a liver injury has not reduced over the past 30 years.
Liver injury is classified on a scale from 1 to 6. A grade 1 injury is least severe, whereas a grade 6 injury is most severe. The majority of people with a grade 6 injury die. Usually, people with grade 1 and 2 liver injuries receive observation as their treatment; so their body can heal naturally. People with higher grade injuries may need surgery. During surgery doctors may stitch the liver together to help it heal.
Review question
We wanted to find out whether surgery or observation is better for people who have a severe blunt liver injury. Studies were included if people had a liver injury of grade 3, 4 or 5. We were interested in finding out if there is a difference in death, illness, or quality of life.
We searched for every randomised controlled trial undertaken worldwide, of surgery or observation for people with grade 3, 4 or 5 liver injury. We searched for trials on 14 April 2014.
Results
We found no randomised controlled trials on this topic. No studies are included in this review.
Conclusion
Trials are needed so that doctors and patients have research to use when making treatment decisions.
Authors' conclusions
Background
The liver is the most commonly injured organ following blunt abdominal trauma (Ahmed 2011). The mechanism of injury often includes motor vehicle and pedestrian crashes, falls or other causes. Hepatic injury is the primary cause of death after severe abdominal trauma, with related mortality of 10% to 15% (Parks 1999). In the last three decades the total number of hepatic injury cases diagnosed has increased due to a rise in urban road traffic crashes together with more accurate diagnosis (Matthes 2003).
Surgery used to be the treatment of choice in cases of blunt hepatic injury, but this approach gradually changed over the last two decades (Christmas 2005; Durham 1992) as increasing non‐operative management (NOM) of splenic injury led to its use for hepatic injury (Richardson 2000). The improvement in critical care monitoring and computed tomographic scanning, as well as the more frequent use of interventional radiology techniques, has helped to bring about this change to non‐operative management (Richardson 2005).
Non‐operative management for blunt hepatic injuries is safe and considered the treatment of choice in haemodynamically stable patients with only low‐grade injuries. In the study described by Christmas 2005, 561 patients with hepatic injury from blunt trauma between 1993 and 2003 underwent NOM: 59% were low‐grade injuries and 41% were high‐grade injuries (grade III‐IV‐V). Of those with grade IV or V injuries, 54.7% underwent NOM, whereas 70% with grade I‐II‐III injuries were managed without surgery.
Failure of NOM is defined as the need for a laparotomy to be performed more than six hours after hospital admission (Norrman 2009). The most common cause of failure of NOM is delayed bleeding and missed associated injuries. Predictive factors of the failure of NOM are: transfusion requirements of more than four units prior to surgery, injury severity score (ISS) of more than 34, injuries of grade IV‐V, large haemoperitoneum, and pooling of contrast (Norrman 2009). In patients with failure of NOM the incidence of postoperative complications is very high and differences have been found between small‐volume and specialised centres (Norrman 2009). In general, in hospitals with low trauma incidence NOM is feasible and safe (Giannopoulos 2009) but surgery may still be required. A study reported by Velmahos 2003 found that 206 patients with blunt injuries to the liver, spleen and kidney underwent NOM. NOM failed in 22%; the rate of failure for spleen injury (34%) was higher than for liver (17%) and kidney (18%) injury.
In the medical literature, the majority of patients who have undergone NOM have low‐grade liver injuries. The safety of NOM in high‐grade liver lesions (American Association for the Surgery of Trauma (AAST) grade IV and V) remains a subject of debate as a high incidence of liver (11%) and collateral extra‐abdominal (17%) complications are still described. Liver‐related and extra‐abdominal complication rates in NOM patients were 11% and 17% respectively in one study (Schnüriger 2009). The length of recovery is a consequence of the grade of liver lesion: mild hepatic injuries heal within seven days whereas high‐grade injuries need up to nine months (Bulas 1993).
Description of the condition
Liver trauma ranges from a small capsular tear, without parenchymal laceration, to massive parenchymal injury with major hepatic vein/retrohepatic vena cava lesions. In 1994, the Organ Injury Scaling Committee of the AAST revised the Hepatic Injury Scale (Table 1). Minor injuries (grade I or II) are the most frequent liver injuries (80% to 90% of all cases); severe injuries are grade III‐V lesions; grade VI lesions are frequently incompatible with survival (Clancy 2001). In this review we analysed only high‐grade blunt hepatic injury (grade III‐V).
| Grade* | Type of injury | Description of injury |
| I | Haematoma | Subcapsular, < 10% surface area |
| Laceration | Capsular tear, < 1 cm parenchymal depth | |
| II | Haematoma | Subcapsular, 10% to 50% surface area or intraparenchymal, < 10 cm in diameter |
| Laceration | 1 to 3 cm parenchymal depth, less than 10 cm in length | |
| III | Haematoma | Subcapsular, > 50% surface area, expanding ruptured subcapsular or parenchymal haematoma with active bleeding, or intraparenchymal haematoma > 10 cm or expanding |
| Laceration | > 3 cm depth | |
| IV | Laceration | Parenchymal disruption involving 25% to 75% of hepatic lobe or 1 to 3 Couinaud segments within a single lobe |
| V | Laceration | Parenchymal disruption involving > 75% of hepatic lobe or more than 3 Couinaud segments within a single lobe |
| Vascular | Juxtahepatic venous injuries, i.e. retrohepatic vena cava/major hepatic veins | |
| VI | Vascular | Hepatic avulsion |
| *Advance 1 grade for multiple injuries *Quoted from Moore (Moore 1995) and others | ||
Description of the intervention
Non‐operative management for blunt hepatic trauma involves monitoring and radiological intervention, including angio‐embolisation (Asensio 2003). In critical care units, patients undergo a protocol for intensive monitoring and an experienced surgical team must follow the patient closely (Buccoliero 2010). With the operative approach, patients immediately undergo hepatic resection (Polanco 2008) and conservative liver techniques (hepatorrhaphy, packing and wrapping) (Cirocchi 1999).
The goal of NOM is to avoid unnecessary laparotomy in selected patients with blunt abdominal injury. For patients with haemodynamic instability (Eastridge 2007), signs of peritonitis or associated extra‐abdominal injuries requiring surgery, it may be impossible to perform emergency surgery (Demetriades 2005). Within an hour of admission to the hospital CT scanning is recommended, because it enables classification of the severity of the abdominal injury and the presence of associated injuries (4.2%) (Schnuriger 2011); missed injuries are more frequently associated with hepatic rather than splenic injuries (Miller 2002). Patients are monitored in the critical care unit by an experienced team. Monitoring of abdominal trauma may include measuring vital signs, abdominal echography, CT scan, CBC (complete blood count), clotting profile, SGOT (aspartate aminotransferase), SGPT (alanine aminotransferase), and urine output. NOM is discontinued in the presence of active bleeding or when there is unstable haemodynamic status (Buccoliero 2010). This approach also applies in high‐grade hepatic injury cases provided that there are no other abdominal injuries that require surgical intervention (Norrman 2009). NOM can also be considered in patients with multiple solid organ injuries, even if it is associated with a higher failure rate (Kozar 2005).
How the intervention might work
The advantage of non‐operative management compared to operative management is a decrease in non‐therapeutic laparotomy with its related mortality, morbility, length of hospitalisation and blood transfusion. The disadvantages of non‐operative management include no immediate control of bleeding, with a risk of delayed bleeding, and missed associated abdominal injuries that require surgical treatment.
Why it is important to do this review
In current clinical practice non‐operative management is best for haemodynamically stable adult and paediatric low‐grade liver injuries. However, uncertainty still exists about its efficacy for high‐grade blunt hepatic injury.
Objectives
To assess the effectiveness of non‐operative management compared to operative management in high‐grade (grade III‐V) blunt hepatic injury.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs).
Cluster‐RCTs are excluded.
Types of participants
Any person who is haemodynamically stable and has high‐grade (grade III‐V) blunt abdominal trauma.
Types of interventions
Non‐operative management compared with operative management.
Types of outcome measures
Primary outcomes
-
Mortality in the first 30 days following trauma.
Secondary outcomes
-
30‐day all‐cause morbidity. All‐cause morbidity is defined as having documentation of serious morbidity or at least one of the following American College of Surgeons, National Surgical Quality Improvement Program (ACS‐NSQIP) complications (American College of Surgeons 2008): "superficial surgical site infection, deep surgical site infection, pneumonia, unplanned intubation (without preoperative ventilator dependence), peripheral neurological deficit, urinary tract infection, and deep vein thrombosis" (Ingraham 2011). Morbidity is a binary outcome (presence or absence of complication in each patient).
-
30‐day serious morbidity. Serious morbidity is defined as having documentation of at least one of the following ACS‐NSQIP complications (American College of Surgeons 2008): "organ space surgical site infection, wound dehiscence, neurological event (cerebrovascular accident or coma lasting 24 hours), cardiac arrest, myocardial infarction, bleeding requiring transfusion, pulmonary embolism, ventilator dependence more than 48 hours, progressive or acute renal insufficiency, and sepsis or septic shock" (Ingraham 2011). We treated morbidity as a binary outcome (presence or absence of complication in each patient).
-
Quality of life at final follow‐up. Quality of life assessed using the Gastrointestinal Quality of Life Index (GLQI). The GLQI was developed by a board of experts to measure quality of life in patients with gastrointestinal diseases, particularly those undergoing an operation. The questionnaire contains up to 36 items in five main categories: gastrointestinal symptoms (19 questions), physical condition (seven questions), emotions (five questions), social function (four questions) and effect of medical treatment (one question) (Eypasch 1995). Each question is scored on a five‐point Likert scale (range 0 to 144, where a higher score indicates a better quality of life). In the French version of the questionnaire a healthy control population scored 126 points out of a total of 144 (Slim 1999).
Search methods for identification of studies
In order to reduce publication and retrieval bias we did not restrict our search by language, date or publication status.
Electronic searches
The Cochrane Injuries Group's Trials Search Co‐ordinator searched the following:
-
Cochrane Injuries Group specialised register (14th April 2014);
-
Cochrane Central Register of Controlled Trials Online (CRSO) (14th April 2014);
-
Ovid MEDLINE(R), Ovid MEDLINE(R) In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily and Ovid OLDMEDLINE(R) (1946 to 14th April 2014);
-
Embase Classic + Embase (OvidSP) (1947 to 14th April 2014);
-
PubMed (15th May 2014);
-
ISI Web of Science: Science Citation Index Expanded (SCI‐EXPANDED) (1970 to April 2014);
-
ISI Web of Science: Conference Proceedings Citation Index‐Science (CPCI‐S) (1990 to April 2014);
-
Clinicaltrials.gov (www.clinicaltrials.gov) (accessed 14th April 2014);
-
International Clinical Trials Registry Platform http://apps.who.int/trialsearch/ (accessed 14th April 2014).
We adapted the MEDLINE search strategy illustrated in Appendix 1 as necessary for each of the other databases. We used search filters, a modified version of the 'Cochrane Highly Sensitive Search Strategies, for identifying randomized trials in MEDLINE and Embase (Lefebvre 2011).
Searching other resources
We handsearched abstracts presented at the following international scientific society conferences:
-
American Association for the Surgery of Trauma from 2003 to 2014
-
Eastern Association for the Surgery of Trauma from 2012 to 2014
-
Western Trauma Association from 2010 to 2014
We checked the reference lists of all relevant studies retrieved from our search and from relevant, published systematic reviews in order to identify other possible studies for inclusion. On 30 April 2014 we conducted an Internet search for grey literature and other information related to our topic on the following websites:
-
http://www.trauma.org/
-
http://traumamon.com/
-
http://archtrauma.com/
-
http://www.omicsgroup.org/journals/trauma‐treatment.php
-
http://tra.sagepub.com/
Data collection and analysis
We conducted the review according to the recommendations of the Cochrane Collaboration (Higgins 2011a; MECIR 2011). We used Review Manager software (RevMan 2012) to conduct the review.
Selection of studies
Two authors (RC, SA) assessed titles or abstracts of all records identified by the initial search and excluded clearly irrelevant studies. We obtained the full text of potentially relevant studies, including any studies with unclear methodologies. Two authors independently assessed the full‐text articles to determine whether they met the inclusion criteria for this review and to evaluate the method of randomisation. There were no disagreements about the inclusion of studies.
Data extraction and management
Two authors independently extracted the following information for each included trial: year of the study, study design, number of participants in each arm, number of participants who received the treatment they were allocated in each arm, number of participants who were allocated to receive one treatment but receive a different treatment, characteristics of non‐operative or operative management, method of generating the randomisation sequence, method of concealing the randomisation sequence, outcomes, drop‐outs, blinding of outcome evaluators and balance of prognostic factors.
Assessment of risk of bias in included studies
In the future if studies are included in the review the risk of bias regarding items of methodological quality in the included studies will be assessed independently by two review authors (RC, ST) using a pilot‐tested data extraction form. Disagreements will be resolved by a consensus meeting with a third review author (EP). In the case of discrepancies in the data extracted, we will contact the authors of the paper for clarification or missing information. We will assess and summarise the potential risk of bias for each trial by using the 'Risk of bias' tool described in theCochrane Handbook (Higgins 2011b).
Measures of treatment effect
In the future if studies are included in the review and the data extracted by the two authors (RC, ST) are different, a third author (EP) will resolve differences. We will express dichotomous outcomes as risk ratios (RR) with 95% confidence intervals (CI). We will express continuous outcomes, if possible, as mean differences (MD) with 95% CI. We will express time‐to‐event outcomes as hazard ratios (HR) with 95% CI. An intention‐to‐treat analysis will be performed in the first instance by extracting the number of patients originally allocated to each treatment group, irrespective of compliance.
Unit of analysis issues
A patient is the unit of analysis. In abdominal surgery, unit of analysis issues are usually not a problem, because cross‐over trials or self controlled trials are not feasible. We do not anticipate unit of analysis issues.
Dealing with missing data
In the future if studies are included in the review we will contact trial investigators to request missing data.
Assessment of heterogeneity
In the future if studies are included in the review RC and ST will assess heterogeneity according to the approach provided in section 9.5 of the Cochrane Handbook (Deeks 2011); 15 studies must be included in the review in order to assess statistical heterogeneity (Thorlund 2012). We will use the Chi2 test with a P value of <0.1 to indicate statistical significance. An I2 value of more than 50% will also be used as an indicator of moderate statistical heterogeneity (Higgins 2003).
Assessment of reporting biases
In the future if studies are included in the review we will use funnel plots to assess small study effects if there are 10 or more studies for a given outcome (Sterne 2011). Due to several explanations for funnel plot asymmetry, RC and ST will carefully interpret results (Egger 1997).
Data synthesis
In the future if studies are included in the review, when calculating risk ratios we will use a random‐effect Mantel‐Haenszel model for meta‐analysis. We will interpret random‐effects meta‐analyses with due consideration of the whole distribution of effects, ideally by presenting a prediction interval (Higgins 2011a). For hazard ratios we will use the generic inverse variance method, and for mean differences we will use the inverse variance method. We will present results on a forest plot. We will perform statistical analyses according to the statistical guidelines referenced in the newest version of theCochrane Handbook (Higgins 2011a).
Subgroup analysis and investigation of heterogeneity
In the future if studies are included in the review we will conduct the following subgroup analysis: hospitals with low‐volume trauma (less than 240 versus 240 or more trauma admissions per year involving patients with an injury severity score greater than 15) (Demetriades 2005).
Sensitivity analysis
In the future if studies are included in the review we will perform sensitivity analyses in order to explore the quality of allocation concealment (low risk of bias versus unclear or high risk of bias).
Results
Description of studies
Results of the search
The search identified 1305 references (Figure 1). After removal of duplicates, 816 references remained. We excluded 733 irrelevant references based on the title or abstract, or both. After reading 83 abstracts, we excluded them all because they were not randomised clinical trials. We retrieved the full‐text of 11 studies reporting on high‐grade blunt hepatic injury, but they did not meet the inclusion criteria. No cluster‐randomised trials were identified by the search for studies.
Study flow diagram.
Included studies
We found no randomised controlled trials comparing non‐operative management with operative management for high‐grade blunt hepatic injury.
Excluded studies
We excluded all the trials identified, as they were not randomised, or not relevant to the topic of this review.
Risk of bias in included studies
We included no trials in this review.
Effects of interventions
We included no trials in this review.
Discussion
We were unable to identify any randomised controlled trials of non‐operative management versus operative management in high‐grade blunt hepatic injury.
Summary of main results
We found no published or ongoing randomised controlled trials that compared non‐operative management versus operative management in high‐grade blunt hepatic injury.
Agreements and disagreements with other studies or reviews
We have not identified any other systematic reviews on this topic.
| Grade* | Type of injury | Description of injury |
| I | Haematoma | Subcapsular, < 10% surface area |
| Laceration | Capsular tear, < 1 cm parenchymal depth | |
| II | Haematoma | Subcapsular, 10% to 50% surface area or intraparenchymal, < 10 cm in diameter |
| Laceration | 1 to 3 cm parenchymal depth, less than 10 cm in length | |
| III | Haematoma | Subcapsular, > 50% surface area, expanding ruptured subcapsular or parenchymal haematoma with active bleeding, or intraparenchymal haematoma > 10 cm or expanding |
| Laceration | > 3 cm depth | |
| IV | Laceration | Parenchymal disruption involving 25% to 75% of hepatic lobe or 1 to 3 Couinaud segments within a single lobe |
| V | Laceration | Parenchymal disruption involving > 75% of hepatic lobe or more than 3 Couinaud segments within a single lobe |
| Vascular | Juxtahepatic venous injuries, i.e. retrohepatic vena cava/major hepatic veins | |
| VI | Vascular | Hepatic avulsion |
| *Advance 1 grade for multiple injuries *Quoted from Moore (Moore 1995) and others | ||
