This article is mentioned in:

Introduction

The liver is the second most commonly injured organ following blunt abdominal trauma and associated injuries contribute significantly to mortality and morbidity (1). Early diagnosis of the nature and extent of intra-abdominal organ injuries may result in significant reduction of morbidity and mortality (1). Focused abdominal sonography for trauma (FAST) is able to sensitively detect free fluid in the abdomen and pelvis, but its numerous limitations have been recognized (2–4). The overall sensitivity of emergency FAST for detection of blunt liver injury was reported to be as low as 64% (5). When there are parenchymal injuries of the liver only, with no free fluid, the sensitivity is even lower (6). Computed tomography (CT) is the standard diagnostic modality for stable trauma patients with a suspected abdominal injury (7,8). However, accurate diagnosis of significant injuries could be delayed as not all health institutions worldwide have ready access to CT scans. In addition, a CT scan suite, at times, may not provide a safe environment for resuscitation and additionally has limitations for patients who are too unstable for transportation. The high cost of a CT scan does not permit its widespread use in screening all patients with blunt abdominal trauma.

In view of the above issues, small-scale studies carried out in pediatric patients with trauma have revealed varied threshold admission levels of liver enzymes below which no clinically significant liver injury was observed (9–13). Should an association between laboratory tests and liver injuries exist, early identification of patients with liver injuries could be achieved. The usefulness of these tests in predicting intra-abdominal and liver injury in adults has not been intensively investigated. The present study was undertaken to determine the accuracy of selected laboratory tests in predicting the presence of liver injury and its severity following blunt abdominal trauma.

Materials and methods

Study population

The 101st Hospital of the Chinese People’s Liberation Army is a 1,000-bed hospital in Southeastern China that provides medical care to approximately 1 million individuals within the Wuxi metropolitan area and has more than 10,000 visits to the emergency department annually. It admits an average of 600 serious trauma cases yearly, of which 96% are blunt injuries.

A review of a retrospective electronic database of all patients with traumatic blunt abdominal injuries to our institution over a three-year period (January 2008 to December 2010) was performed. Patients were excluded if they suffered penetrating injuries, died in the emergency department or if the required laboratory tests were not performed within 24 h of the trauma.

Patients were then subdivided into two groups: patients with and without liver injuries. Liver injury grade was determined using the organ injury scale (1994 revision) described by the American Association of Surgery for Trauma (AAST). In this study, minor liver injuries were classified as AAST Grades I–III, while major liver injuries were classified as AAST Grades IV–VI. The information was obtained from either surgery or CT scans. The results of the CT scans were abstracted from attending radiology reports if the CT scan was performed at our hospital or from review of progress notes if the scan was obtained at another institution prior to transfer.

Test results for aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), lactate dehydrogenase (LDH), bilirubin, γ-glutamyl transpeptidase (GGT), C-reactive protein (CRP) and white blood cell (WBC) levels were collected. The values were compared with reference ranges for our institution (Table I). Data collection forms also included age, gender, trauma mechanisms, injury severity score (ISS), the AAST grade of liver injury, length of stay (LoS) in an intensive care unit (ICU), total inpatient LoS and the eventual outcome.

Table I. Reference ranges for laboratory tests.

Table I.

Reference ranges for laboratory tests.

Laboratory test	Reference range
Aspartate aminotransferase (AST)	0–50 U/l
Alanine aminotransferase (ALT)	0–50 U/l
Alkaline phosphatase (ALP)	32–135 U/l
Bilirubin	4–25 μmol/l
Lactate dehydrogenase (LDH)	50–250 U/l
γ-glutamyl transpeptidase (GGT)	0–50 U/l
White blood cell (WBC)	4–10×109/l
C-reactive protein (CRP)	0–5 mg/l

Median and range were calculated for continuous variables. The selected laboratory test levels in patients with liver injuries were compared with patients with abdominal non-liver injury by plotting the receiver operating characteristic (ROC) curves for threshold in the presence or absence of liver injury. Data analysis was performed by comparing the selected laboratory tests to the threshold using the Chi-square test. All P-values were 2-sided and considered significant at P<0.05. All statistical operations were performed using SPSS Statistics 13 for Windows (SPSS Inc, Chicago, IL, USA).

Results

Characteristics of the study sample

During the three-year study period, 182 patients with blunt abdominal trauma had the relevant laboratory tests performed and were included in our study. Table II summarizes the general characteristics of the two groups. In the two groups there was a preponderance of males. Patients with abdominal non-liver injury were, on average, older than those with liver injury. The majority of the patients underwent CT scans. The road traffic accident was the most common mechanism of injury. Patients with liver injury tended to be significantly more severely injured than those with non-liver injury in terms of their median ISS, LoS in ICU and total. However, more patients (67.4%) with abdominal non-liver injury underwent exploratory laparotomy than those (41.1%) with liver injury.

Table II. Characteristics of the 182 patients with blunt abdominal trauma.

Table II.

Characteristics of the 182 patients with blunt abdominal trauma.

	Liver injury (n=90)	Abdominal non-liver injury (n=92)
Age, median (range)	39.0 (8–73)	45.7 (10–83)
Gender, n (%)
Male	75 (83.3)	72 (78.3)
Female	15 (16.7)	20 (21.7)
Mechanism, n (%)
Road traffic accident	57 (63.3)	58 (63.0)
Fall	21 (23.3)	16 (17.4)
Assault	7 (7.8)	12 (13.0)
Crush	5 (5.6)	6 (6.5)
CT scans, n (%)
Performed	89 (98.9)	84 (91.3)
Not performed	1 (1.1)	8 (8.7)
Surgical intervention, n (%)
Performed	37 (41.1)	62 (67.4)
Not performed	53 (58.9)	30 (32.6)
ISS, median (range)	32.0 (9–75)	24.6 (9–75)
ICU LoS, median (range)	6.9 (0–30)	5.0 (1–30)
Total LoS, median (range)	30.3 (2–215)	22.1 (2–93)
Outcome, n (%)
Alive	85 (94.4)	83 (90.2)
Dead	5 (5.6)	9 (9.8)

[i] CT, computed tomography; ISS, injury severity score; ICU, intensive care unit; LoS, length of stay.

Patients with liver injuries

The grading of the 90 patients with liver injuries is shown in Table III. In those patients with liver injuries, patients were graded according to the severity of their liver injury as follows: grade I, 23 patients; grade II, 27 patients; grade III, 23 patients; grade IV, 6 patients; and grade V, 11 patients. There were no patients with grade VI liver injuries. Grouped according to severity, there were 73 patients (81.1%) with minor (grades I–III) injuries and 17 patients (18.9%) with major (grades IV–V) injuries.

Table III. Grading of liver injury for the 90 patients.

Table III.

Grading of liver injury for the 90 patients.

Grading of liver injury	n (%)
Minor liver injuries	73 (81.1)
I	23 (25.6)
II	27 (30.0)
III	23 (25.6)
Major liver injuries	17 (18.9)
IV	6 (6.7)
V	11 (12.2)
VI	0 (0.0)

Main results

Patients with non-liver injury and grades of liver injury from I to V were significantly different in regards to levels of ALT (P<0.001), AST (P<0.001), LDH (P<0.001) and GGT (P=0.001) (Table IV). Fig. 1 shows the ROC curves generated for the sensitivity and specificity comparison for the association between ALT, AST, LDH, GGT levels and the presence of liver injury. The area under the curve demonstrated that the test was a good discriminator for identification of liver injury. Using ROC curve assessment, optimum ALT, AST, LDH, GGT thresholds were determined to be >57 U/l, 113 U/l, 595 U/l and 50 U/l.

Figure 1. ROC curves showing sensitivity and specificity for (A) alanine aminotransferase, (B) aspartate aminotransferase, (C) lactate dehydrogenase and (D) γ-glutamyl transpeptidase threshold values. ROC, receiver operating characteristic. Diagonal segments are produced by ties.

Table IV. Relationship between the various laboratory tests and liver injuries.

Table IV.

Relationship between the various laboratory tests and liver injuries.

Laboratory test	No liver injury	Grade of liver injury					F-value	P-value
		I	II	III	IV	V
ALT	35± 23	96±64	306±274	263±206	628±448	681±440	39.01	<0.001
AST	59±42	126±90	309±261	292±230	777±482	695±368	43.40	<0.001
LDH	483±379	637±509	879±1084	775±652	2023±1715	1581±1011	10.27	<0.001
ALP	57±39	57±36	65±27	60±28	52±15	60±31	0.28	0.923
GGT	25±37	28±35	83±138	32±36	73±62	29±14	4.12	0.001
Bilirubin	17±10	18±13	17±11	14±7	18±6	25±24	1.20	0.310
WBC	14±6	13±7	14±6	13±5	17±6	11±2	0.97	0.439
CRP	54±80	52±84	53±87	31±32	16±29	40±67	0.63	0.681

[i] ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDH, lactate dehydrogenase; ALP, alkaline phosphatase; GGT, γ-glutamyl transpeptidase; WBC, white blood cell count; CRP, C-reactive protein.

ALT >57 U/l (OR, 66.1; P<0.001) and AST >113 U/l (OR, 30.6; P<0.001) were significantly associated with the presence of liver injuries. This was also observed in patients with LDH >595 U/l (OR, 3.8; P<0.001) and GGT >50 U/l (OR, 3.0; P<0.05) (Table V).

Table V. Relationship between selected laboratory tests and the presence of liver injuries (part I).

Table V.

Relationship between selected laboratory tests and the presence of liver injuries (part I).

Laboratory test	Liver injury (n)	No liver injury (n)	OR	95% CI	P-value
Median ALT (range)	309 (13–1500)	35 (8–118)	66.1	25.33–172.27	<0.001
ALT ≤57	7	78
ALT >57	83	14
Median AST (range)	336 (17–1637)	59 (15–228)	30.6	12.86–72.73	<0.001
AST ≤113	23	84
AST >113	67	8
Median LDH (range)	953 (173–5787)	482 (49–2875)	3.8	2.02–7.17	<0.001
LDH ≤595	41	70
LDH >595	49	22
Median GGT (range)	49 (4–551)	25 (3–319)	3.0	1.25–7.22	0.014
GGT ≤50	70	84
GGT >50	20	8

[i] ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDH, lactate dehydrogenase; ALP, alkaline phosphatase; GGT, γ-glutamyl transpeptidase; OR, odds ratio; CI, confidence interval.

Further analysis of selected laboratory tests also revealed that ALT >57 U/l is perhaps most suitable for detecting hepatic injuries. Its sensitivity (92.2%), specificity (84.8%), positive predictive value (85.6%) and negative predictive value (91.8%) are all favorable for its role as a screening tool compared to the other markers (Table VI). However, in patients with liver injuries, none of the selected laboratory tests were related to the severity of the liver injuries (Table VII).

Table VI. Relationship between selected laboratory tests to presence of liver injuries (part II).

Table VI.

Relationship between selected laboratory tests to presence of liver injuries (part II).

Laboratory test	Liver injury (n)	No liver injury (n)
ALT >57	83	14	Positive predictive value = 85.6%
ALT ≤57	7	78	Negative predictive value = 91.8%
	Sensitivity = 92.2%	Specificity = 84.8%
AST >113	67	8	Positive predictive value = 89.3%
AST ≤113	23	84	Negative predictive value = 78.5%
	Sensitivity = 74.4%	Specificity = 91.3%
LDH >595	49	22	Positive predictive value = 69.0%
LDH ≤595	41	70	Negative predictive value = 63.1%
	Sensitivity = 54.4%	Specificity = 76.1%
GGT >50	20	8	Positive predictive value = 71.4%
GGT ≤50	70	84	Negative predictive value = 54.5%
	Sensitivity = 22.2%	Specificity = 91.3%

[i] ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDH, lactate dehydrogenase; GGT, γ-glutamyl transpeptidase.

Table VII. Relationship between selected laboratory tests and severity of liver injuries.

Table VII.

Relationship between selected laboratory tests and severity of liver injuries.

Laboratory test	Minor liver injury	Major liver injury	OR	95% CI	P-value
Median ALT (range)	227 (13–1123)	663 (13–1500)	0.1	0.01–0.90	0.040
ALT ≤100	26	1
ALT >100	47	16
Median AST (range)	246 (17–1131)	724 (28–1636)	0.2	0.02–1.16	0.069
AST ≤113	22	1
AST >113	51	16
Median LDH (range)	770 (173–5787)	1737 (240–5189)	0.1	0.01–0.41	0.005
LDH ≤595	40	1
LDH >595	33	16
Median GGT (range)	50 (4–551)	45 (9–148)	0.8	0.15–4.22	0.788
GGT ≤100	66	15
GGT >100	7	2

[i] ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDH, lactate dehydrogenase; ALP, alkaline phosphatase; GGT, γ-glutamyl transpeptidase; OR, odds ratio; CI, confidence interval.

Discussion

The liver continues to be the second most commonly injured organ in blunt abdominal trauma (1). Physicians dealing with blunt abdominal trauma often use biochemical tests and radiographic imaging to aid in clinical assessment. Outcome of liver trauma has been shown to be related to several important factors: increased ISS, worse grading of hepatic injury, advanced age, operative blood loss and hemodynamic instability on admission (14,15).

Patients with blunt abdominal trauma resulting in liver injury may present with unstable hemodynamics and obvious hemoperitoneum. These patients usually do not represent a diagnostic challenge since the strategy is clear. They generally receive either prompt abdominal imaging (ultrasound or CT scan) or laparotomy or both. Usually the more difficult diagnosis is that of lesser, but still significant, liver injury in the stable patient with minimal physical findings following blunt abdominal trauma. Abdominal CT scan has proved to be a valuable diagnostic tool for evaluating intra-abdominal injury (8). However, it may only be carried out if available and in patients who are hemodynamically stable. Furthermore, it is costly, requires radiation exposure, and removes patients from direct clinical care. Another useful tool for initial survey is FAST. Yet, FAST is not always available in all healthcare institutions and one of the major limitations is the technical expertise of the operator (3,4). Furthermore, FAST has limited value if the blunt abdominal trauma does not result in hemoperitoneum, and it may miss nearly one third of the cases of abdominal injuries when used as the sole diagnostic tool in evaluating victims of blunt abdominal trauma.

Elevation of the serum liver enzymes AST and ALT is known to be associated with blunt traumatic liver injury. Presumably, as these transaminases are present in high concentrations in hepatocytes, they are released into the circulation in large quantities following acute traumatic hepatocellular injury. AST and ALT have been previously reported to indicate liver injury (12,16). One previous observational cohort study has reported serum ALT to be a sensitive diagnostic marker when evaluating harm caused by blunt hepatic injuries (17).

From our study, the authors preferred ALT to AST as a screening tool for hepatic injuries due to its associated high sensitivity, specificity, positive predictive value and negative predictive value. Although 7 (7.8%) and 23 (25.6%) of the patients with liver injuries had ALT and AST levels less than their thresholds in our study, most of them only had grades I and II injuries. We also found a trend that the more severe the liver injury of the patients, the higher the liver enzyme levels, but our study further demonstrates that patients with elevated ALT >100 U/l, AST >113 U/l, LDH >595 U/l and GGT >100 U/l did not necessarily have major liver injury. Therefore, even where patients whose liver enzyme levels are at lower levels and liver injury cannot be completely ruled out, they may still provide clues concerning liver injury, particularly in patients with high-grade liver injury.

There are several limitations to this study. Firstly, it is a retrospective chart review; therefore, data may not be present or properly recorded on the medical record. Secondly, the time interval between injury and the procurement of the blood test could not be standardized. This was predominantly due to the highly variable timing and location of the accident. However, we tried to confine our patients to the group having their blood drawn within 24 h to reduce study bias. Thirdly, as the threshold for undertaking CT scans varies greatly from one institution to another, our findings may not be applicable to the entire population of patients with blunt abdominal trauma.

In conclusion, this study suggests that, in patients with blunt abdominal trauma, abnormal transaminase levels are associated with liver injury. Patients with ALT >57 U/l and AST >113 U/l are strongly associated with liver injury and require further imaging studies and close management.

Acknowledgements

This research was supported by the Army Medicine Technology Fund (10MA008).

References