Eur J Trauma Emerg Surg (2014) 40:309–313 DOI 10.1007/s00068-013-0341-z
ORIGINAL ARTICLE
Serum lipase for assessment of pancreatic trauma B. Mitra • M. Fitzgerald • M. Raoofi G. A. Tan • J. C. Spencer • C. Atkin
•
Received: 4 June 2013 / Accepted: 30 September 2013 / Published online: 17 October 2013 Ó Springer-Verlag Berlin Heidelberg 2013
Abstract Purpose Pancreatic enzymes are routinely measured during reception of trauma patients to assess for pancreatic injury despite conflicting evidence on their utility. The aim of this study was to investigate the utility of routine initial serum lipase measurement for the diagnosis of acute pancreatic trauma. Materials and methods Lipase measurements were introduced as part of the trauma pathology panel and requested on all patients who presented to an adult major trauma service and met trauma call-out criteria. Clinical records of these patients were extracted from the trauma registry and retrospectively reviewed. The performance of an initial serum lipase level measured on presentation to detect pancreatic trauma was determined. Results There were 2,580 patients included in the study, with 17 patients diagnosed with pancreatic trauma. An elevated lipase was recorded in 390 patients. Statistically significant associations were observed for B. Mitra (&) M. Fitzgerald G. A. Tan J. C. Spencer Emergency and Trauma Centre, The Alfred Hospital, Commercial Rd., Melbourne, VIC 3004, Australia e-mail: [email protected] B. Mitra Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia B. Mitra M. Fitzgerald National Trauma Research Institute, Melbourne, Australia M. Fitzgerald M. Raoofi C. Atkin Trauma Service, The Alfred Hospital, Melbourne, Australia M. Fitzgerald G. A. Tan Central Clinical School, Monash University, Melbourne, Australia
elevated lipase in patients with pancreatic trauma, head injury, acute alcohol ingestion and massive blood transfusion. As a test for pancreatic trauma, an abnormal serum lipase result had a specificity of 85.3 % (95 % CI 83.8–86.6), sensitivity of 76.5 % (95 % CI 49.8–92.2), positive predictive value of 3.3 % (95 % CI 1.8–5.8) and negative predictive value of 99.8 % (95 % CI 99.4–99.9). Higher cut-offs of serum lipase did not result in better performance. Conclusions A normal serum lipase result can be a useful adjunct to exclude pancreatic injury. A positive lipase result, regardless of the cut-off used, was not reliably associated with pancreatic trauma, and should not be used to guide further assessment. Keywords Wounds and injuries Pancreas Abdomen Resuscitation Hematologic tests Tomography X-Ray computed
Introduction Pre-determined laboratory tests are commonly requested during reception of major trauma patients. The rationale of this practice stems from the fear of missing vital investigation results during the hectic phase of initial resuscitation. However, routine panels have been previously shown to be neither useful, nor cost-effective [1–3]. Selective, patient specific investigations have been previously suggested, but translation of this evidence to practice has been slow with a range of laboratory panels continuing to be recommended for trauma resuscitation [4]. A measure of pancreatic enzymes, serum amylase, serum lipase or both, are commonly part of trauma laboratory panels. The utility of abnormal pancreatic
123
310
B. Mitra et al.
Table 1 The Alfred trauma call-out criteria Mechanism
MBA/cyclist impact [30 km/h
Table 2 Grading of pancreatic injury Grade
Criteria
I
Simple contusion of the pancreas
II
Major contusion or laceration without tissue loss or involvement of the main pancreatic duct
III
Complete transection of the pancreas or a parenchymal injury with involvement of the major duct to the left of the superior mesenteric vein
Explosion
IV
All significant blunt trauma assessed by ambulance
Ductal transection or a major parenchymal injury to the right of the superior mesenteric vein
V
Massive disruption of the head of the pancreas
Pedestrian impact [30 km/h Extrication [30 min Vehicle rollover Fatality in same vehicle Ejection from vehicle Fall [3 m Injuries (not exclusive)
All penetrating head, neck and truncal injuries Traumatic amputation of carpus/tarsus # pelvis/pulseless limb/#dislocations with vascular compromise Evisceration
lipase measurement for the diagnosis of acute pancreatic trauma following the introduction of routine lipase measurements on all trauma call-out patients.
Blast injuries Severe crush injury Severe burns ([20 % TBSA); All facial burns Signs
SBP \ 100 mmHg GCS \ 14 SpO2 \ 90 % RR \ 10 or [30 beats min-1
Treatment
Any airway manoeuvre including intubation Assisted ventilation Pleural decompression Haemostatic dressings/tourniquet application [1,000 mL IV fluid or blood administration Neuromuscular blockade
Other criteria
Mass casualty incident All inter-hospital trauma transfers Pregnancy Significant comorbidity Anticoagulant therapy
enzymes is presumed to aid the early diagnosis of pancreatic trauma. Among the paediatric population, evidence for the lack of utility of routine amylase or lipase measurements is relatively strong with multiple studies and cost-effectiveness analyses [5–7]. However, conflicting evidence persists for the adult population. The inclusion of serum amylase or lipase to the routine laboratory parameters after trauma in adult patients have been previously recommended as being indicative of the need for early surgical intervention [8, 9]. However, such findings have been refuted by suggestions that acute serum amylase and lipase levels are neither specific nor sensitive for acute pancreatic trauma [10–13]. The aim of this study was to determine the incidence of patients with pancreatic trauma presenting to a major trauma centre and investigate the utility of initial serum
123
Methods Setting The state of Victoria, Australia has one paediatric and two adult major trauma services (MTS) located within metropolitan Melbourne. Major trauma triage guidelines direct 85 % of major trauma patients to an MTS for definitive treatment. The Alfred Hospital is one of the two adult MTS, with a trauma database prospectively capturing prehospital and hospital data on all major trauma patients, defined as ISS [15, or requiring urgent surgery, or intensive care unit (ICU) admission or death or[72 h admission to hospital post trauma. The Alfred trauma call-out criteria are listed in Tables 1, 2. All patients meeting these criteria are received at the trauma centre with guidelines recommending trauma panel laboratory investigations that include serum lipase requested on presentation. Patient selection All patients included in the Alfred Hospital Trauma registry, presenting between August 2010 and December 2011 post blunt trauma, were included in this study. Cases that did not have trauma team activation or did not have an initial serum lipase result recorded as part of the trauma panel of tests were excluded, as trauma reception guidelines may not have been followed. An injury to the pancreas was recorded if confirmed by imaging or intraoperatively. Hyperlipasaemia as the sole criteria for pancreatic injury was not accepted. The serum lipase was measured using spectrophotometry. The normal reference range was 10–70 U/L. The shock index is defined by heart rate over systolic blood pressure and recorded to be abnormal for a value over 1.0 [14]. Massive blood
Serum lipase for assessment of pancreatic trauma
transfusion was defined as C5 units of packed red blood cells in 4 h [15]. Any positive blood alcohol result was considered abnormal. Pancreatic trauma was classified as proposed by the American Association for the Surgery of Trauma Committee on Organ Injury Scaling [16].
311 Table 3 Comparison of patients with elevated and non-elevated serum lipase Elevated lipase (n = 390) Age (years) ISS
43.0 (20.1)
10 (5–22)
Temperature (°C)
10 (5–18)
36.3 (1.0)
Heart rate (beats min-1) Systolic blood pressure (mmHg) Systolic blood pressure \100 mmHg
36.4 (0.8)
90.1 (21.4)
88.1 (20.3)
143.5 (25.0)
146.1 (27.4)
15 (3.8 %)
88 (4.0 %)
Shock index [1
20 (5.1 %)
90 (4.1 %)
GCS
15 (14–15)
15 (14–15)
Massive transfusion*
23 (5.9 %)
61 (2.8 %)
103 (26.4 %) 148 (37.9 %)
398 (18.2 %) 693 (31.6 %)
13 (3.3 %)
4 (0.2 %)
Blood alcohol positive* Head injury* Pancreatic injury*
0.50
Sensitivity
0.75
1.00
* Statistically significant
0.00
Medical records of all patients who had a pancreatic injury coded were extracted to determine the mode of diagnosis. Diagnostic accuracy of serum lipase in detecting pancreatic trauma was reported using sensitivity, specificity, positive and negative predictive values and 95 % confidence intervals. A two-sided p value of \0.05 was considered to be statistically significant. Normally distributed continuous data were reported as mean (standard deviation). Ordinal or skewed data were reported as median (interquartile range) and categorical data as counts (proportions). Statistical significance for differences between means were tested using the Student’s t test, differences between medians were tested using the Wilcoxon rank sum test while differences between proportions were tested the v2 test. Predictive accuracy of accumulative effect of these variables was assessed by measuring the area under a receiver operating characteristic (ROC) curve. Cases with data missing with regards to lipase measurement were handled by list-wise exclusion. All analyses were performed using Stata version 11.0 (Statacorp, College Station, TX, USA). This study was approved by The Alfred Hospital Research and Ethics Committee.
41.0 (18.1)
0.25
Statistical analysis
Non-elevated lipase (n = 2,190)
Results
0.00
0.25
0.50
0.75
1.00
1 - Specificity Area under ROC curve = 0.8941
A total of 4,808 patients were identified in the database for the study period. There were 1,936 patients excluded who did not have a trauma team activation and a further 287 patients excluded as not having an initial serum lipase result recorded. There were 22 patients who were coded as having pancreatic trauma. Following review of medical records, five more cases were excluded as there was no radiological or operative evidence of pancreatic trauma, with the diagnosis based on the lipase level alone. There were 2,580 patients included, of which 17 patients were diagnosed with pancreatic injury. This represented an incidence of 0.7 % of pancreatic trauma among patients presenting after blunt trauma. There were 390 patients with an initial abnormal serum lipase recorded. Demographic and injury characteristics of patients with elevated and non-elevated serum lipase are presented in Table 3. As a test for pancreatic trauma, an abnormal serum lipase result ([70 U/L) was found to have
Fig. 1 ROC for lipase and pancreatic injury
a specificity of 85.3 % (95 % CI 83.8–86.6) with a sensitivity of 76.5 % (95 % CI 49.8–92.2), a positive predictive value (PPV) of 3.3 % (95 % CI 1.8–5.8) and a negative predictive value (NPV) of 99.8 % (95 % CI 99.4–99.9). Figure 1 illustrates the area under the ROC of 0.89 (95 % CI 0.84–0.95). Using a cut-off of serum lipase of 500 U/L, specificity for pancreatic injury was 98.2 % (95 % CI 97.6–98.7) with a sensitivity of 35.3 % (95 % CI 15.2–61.4), a PPV of 11.7 % (95 % CI 4.9–24.5) and a NPV of 99.6 % (99.2–99.8 %). At a cut-off of 1,000 U/L, specificity for pancreatic injury was 99.5 % (95 % CI 99.1–99.7) with a sensitivity of 17.6 % (95 % CI 4.7–44.2), a PPV of 20.0 % (95 % CI 5.3–48.6) and a NPV of 99.4 % (95 % CI 99.1–99.7).
123
312
The initial diagnosis of pancreatic injury was made using abdominal CT in 16 patients with one patient diagnosed using ultrasound. Of the 17 patients diagnosed with pancreatic trauma, nine patients underwent an explorative laparotomy, with two patients undergoing distal pancreatectomies. These two patients were classified as having grade IV pancreatic trauma. A further four patients had drains inserted only with Grade I or II injuries. Patients who did not undergo surgery were all classified as having grade I or II injuries. There were no deaths among patients with pancreatic trauma. The average length of stay was 23.2 (14.4) days with 10 patients requiring admission to the ICU. Among the four patients with normal initial serum lipase result but diagnosed to have pancreatic trauma, the diagnoses were made on CT abdomen. None of these patients had repeat serum lipase measurements. All four patients had other injury to abdominal organs with two patients sustaining bowel injuries, one renal injury and the other sustaining liver lacerations. Among excluded patients, there were no patients diagnosed with pancreatic injury and 81 of the excluded patients had an abnormal lipase level of [70 U/L.
Discussion In our population of adult patients post blunt trauma presenting to a major trauma centre, the incidence of pancreatic injury was low. A serum lipase level recorded immediately on presentation to the ED was highly sensitive and may be useful in its ability to exclude pancreatic injury with a NPV of over 99 %. The PPV for an initial lipase measurement to detect pancreatic injury was low, suggesting that the serum lipase result should not be used to guide further investigations or management for pancreatic injury. The PPV did not improve when higher cut-offs of serum lipase were considered. Provided the limitations are considered, this study supports the inclusion of serum lipase among the panel of pathology tests during initial trauma reception. The incidence of pancreatic trauma been previously reported at \1 % is consistent with our population [8, 17]. The low incidence has resulted in a paucity of clinical studies related to the diagnosis of pancreatic trauma. As a result, retrospective case-series have used inclusion criteria of patients with diagnosed pancreatic trauma only, where the NPV could not be evaluated. The reported poor utility of serum lipase, therefore, had been previously based on the low PPVs of serum pancreatic enzymes to detect pancreatic trauma—reported to be 1.5 % and supported by this study [11, 12]. The low PPV of serum lipase for pancreatic trauma is likely due to other causes of hyperlipasaemia in major trauma patients. A high serum lipase has been previously
123
B. Mitra et al.
associated with head injury and confirmed in this study [18–21]. The serum lipase was also univariately associated with recent alcohol ingestion. An association with massive blood transfusion was also observed and this had also been previously reported along with an absence of association with mortality [22]. The blunt force required to injure the pancreas is substantial and a pancreatic injury is rarely a solitary injury. When the pancreas is injured in the setting of blunt trauma, other abdominal organs are also likely to be affected. Where clinical indications exist for further evaluation of other abdominal organs using imaging or surgery, the pancreas can be assessed as part of that procedure. In the setting of no such indication, an isolated elevated serum lipase result should not be used as an indication to further assess for pancreatic injury. Patients with a reliable history, lack of clinical signs and isolated elevated serum lipase results in the setting of trauma should be observed for clinical signs and undergo repeated measurements [23]. An arbitrary cut-off value of both amylase and lipase of three times normal is frequently used by clinicians to discriminate between those with acute pancreatitis from those with non-pancreatic abdominal pain and translated into the setting of pancreatic trauma [24]. We have shown that a higher cut-off does not improve performance of the test in this setting and therefore, should not be used. This study is limited in being a retrospective review of the data, but included a large number of patients who underwent serum lipase testing in the setting of blunt trauma immediately upon presentation to the ED. Patients who did not have a trauma call-out were excluded, but there were no cases of pancreatic trauma among excluded patients and as such, their inclusion regardless of the lipase result, would not have altered the NPV. On the contrary, it would have further lowered the PPV, strengthening our view that an elevated initial serum lipase result alone should not be used to guide further assessment for pancreatic trauma. Post-traumatic pancreatitis has been previously described, with an incidence of up to 17 % among trauma patients and may be secondary to undiagnosed pancreatic trauma [25]. No cases of delayed diagnosis were observed in our cohort, but follow up beyond hospital discharge was not conducted. The exact timing of injury and time from injury to presentation to the ED was not available. This could be important as delayed measurement of lipase has previously been suggested to have higher specificity for detecting pancreatic trauma [7]. The current state-wide Victorian Trauma System mandates direct transport of patients who are within 30 min of transport time and early transfer for other major trauma services [26]. Adding in time from injury to ambulance arrival and scene times, we estimate most patients to have had a serum lipase measurement after at least 1 h of injury. It is, therefore, unlikely that
Serum lipase for assessment of pancreatic trauma
specificity of the test would improve to the point of being clinically useful. While an initial serum lipase appears to be efficacious in excluding pancreatic trauma, further prospective studies are required to confirm the utility of a serum lipase together with clinical examination to determine clinical effectiveness and if established, cost-effectiveness studies will be required to determine efficiency. The low incidences of pancreatic trauma presenting to single centres make such studies difficult to design and confidently determine PPVs. In the absence of an appropriate level of evidence to guide efficiency, cost-effective studies will be premature. Collaborative multi-centre trauma registries may provide the answer to facilitating such clinical studies in the future.
313
7.
8.
9.
10.
11.
12.
Conclusions Despite discouraging previous reports, a negative initial serum lipase result appears to be an useful adjunct during the assessment of the major trauma patient in being able to exclude pancreatic injury. We support the retention of serum lipase as a screening test for all patients presenting with major trauma. A positive result, regardless of the cutoff used, was not diagnostic of pancreatic trauma, and should not be used to guide further assessment.
13. 14.
15.
16. Acknowledgments We would like to thank staff at The Alfred Hospital Trauma Registry and in particular, Ms. Louise Niggemeyer for extraction of data. Conflict of interest
17.
None to declare. 18.
References 1. Namias N, McKenney MG, Martin LC. Utility of admission chemistry and coagulation profiles in trauma patients: a reappraisal of traditional practice. J Trauma. 1996;41(1):21–5. 2. Tasse JL, Janzen ML, Ahmed NA, Chung RS. Screening laboratory and radiology panels for trauma patients have low utility and are not cost effective. J Trauma. 2008;65(5):1114–6. doi:10. 1097/TA.0b013e318184b4f2. 3. Tortella BJ, Lavery RF, Rekant M. Utility of routine admission serum chemistry panels in adult trauma patients. Acad Emerg Med. 1995;2(3):190–4. 4. Chu UB, Clevenger FW, Imami ER, Lampard SD, Frykberg ER, Tepas JJ 3rd. The impact of selective laboratory evaluation on utilization of laboratory resources and patient care in a level-I trauma center. Am J Surg. 1996;172(5):558–62. doi:10.1016/ s0002-9610(96)00234-6 (discussion 62-3). 5. Adamson WT, Hebra A, Thomas PB, Wagstaff P, Tagge EP, Othersen HB. Serum amylase and lipase alone are not costeffective screening methods for pediatric pancreatic trauma. J Pediatr Surg. 2003;38(3):354–7. doi:10.1053/jpsu.2003.50107 (discussion-7). 6. Herman R, Guire KE, Burd RS, Mooney DP, Ehlrich PF. Utility of amylase and lipase as predictors of grade of injury or outcomes
19.
20.
21. 22.
23.
24.
25.
26.
in pediatric patients with pancreatic trauma. J Pediatr Surg. 2011;46(5):923–6. doi:10.1016/j.jpedsurg.2011.02.033. Matsuno WC, Huang CJ, Garcia NM, Roy LC, Davis J. Amylase and lipase measurements in paediatric patients with traumatic pancreatic injuries. Injury. 2009;40(1):66–71. doi:10.1016/j. injury.2008.10.003. Mayer JM, Tomczak R, Rau B, Gebhard F, Beger HG. Pancreatic injury in severe trauma: early diagnosis and therapy improve the outcome. Dig Surg. 2002;19(4):291–7. doi:64576 (discussion 7-9). Kumar S, Sagar S, Subramanian A, Albert V, Pandey RM, Kapoor N. Evaluation of amylase and lipase levels in blunt trauma abdomen patients. J Emerg Trauma Shock. 2012;5(2):135–42. doi:10.4103/0974-2700.96482. Buechter KJ, Arnold M, Steele B, Martin L, Byers P, Gomez G, et al. The use of serum amylase and lipase in evaluating and managing blunt abdominal trauma. Am Surg. 1990;56(4):204–8. Boulanger BR, Milzman DP, Rosati C, Rodriguez A. The clinical significance of acute hyperamylasemia after blunt trauma. Can J Surg. 1993;36(1):63–9. Farkouh E, Wassef R, Atlas H, Allard M. Importance of the serum amylase level in patients with blunt abdominal trauma. Can J Surg. 1982;25(6):626–8. Olsen WR. The serum amylase in blunt abdominal trauma. J Trauma. 1973;13(3):200–4. Tourtier JP, Jost D, Domanski L. Shock index: a simple clinical parameter for mortality risk assessment in trauma? J Trauma Acute Care Surg. 2012;73(3):780–1. doi:10.1097/TA.0b013 e31825ff540 (author reply 1). Mitra B, Cameron PA, Gruen RL, Mori A, Fitzgerald M, Street A. The definition of massive transfusion in trauma: a critical variable in examining evidence for resuscitation. Eur J Emerg Med. 2011;18(3):137–42. doi:10.1097/MEJ.0b013e328342310e. Moore EE, Cogbill TH, Malangoni MA, Jurkovich GJ, Shackford SR, Champion HR, et al. Organ injury scaling. Surg Clin North Am. 1995;75(2):293–303. Lin BC, Chen RJ, Fang JF, Hsu YP, Kao YC, Kao JL. Management of blunt major pancreatic injury. J Trauma. 2004;56(4): 774–8. Pezzilli R, Billi P, Barakat B, Fiocchi M, Re G, Gullo L, et al. Serum pancreatic enzymes in patients with coma due to head injury or acute stroke. Int J Clin Lab Res. 1997;27(4):244–6. Justice AD, DiBenedetto RJ, Stanford E. Significance of elevated pancreatic enzymes in intracranial bleeding. South Med J. 1994;87(9):889–93. Vitale GC, Larson GM, Davidson PR, Bouwman DL, Weaver DW. Analysis of hyperamylasemia in patients with severe head injury. J Surg Res. 1987;43(3):226–33. Bouwman DL, Altshuler J, Weaver DW. Hyperamylasemia: a result of intracranial bleeding. Surgery. 1983;94(2):318–23. Malinoski DJ, Hadjizacharia P, Salim A, Kim H, Dolich MO, Cinat M, et al. Elevated serum pancreatic enzyme levels after hemorrhagic shock predict organ failure and death. J Trauma. 2009;67(3):445–9. doi:10.1097/TA.0b013e3181b5dc11. Degiannis E, Glapa M, Loukogeorgakis SP, Smith MD. Management of pancreatic trauma. Injury. 2008;39(1):21–9. doi:10. 1016/j.injury.2007.07.005. Smith RC, Southwell-Keely J, Chesher D. Should serum pancreatic lipase replace serum amylase as a biomarker of acute pancreatitis? ANZ J Surg. 2005;75(6):399–404. doi:10.1111/j. 1445-2197.2005.03391.x. Fleming WR, Collier NA, Banting SW. Pancreatic trauma: Universities of Melbourne HPB Group. Aust N Z J Surg. 1999;69(5):357–62. Cox S, Currell A, Harriss L, Barger B, Cameron P, Smith K. Evaluation of the Victorian state adult pre-hospital trauma triage criteria. Injury. 2012;43(5):573–81. doi:10.1016/j.injury.2010.10.003.
123
ORIGINAL ARTICLE
Serum lipase for assessment of pancreatic trauma B. Mitra • M. Fitzgerald • M. Raoofi G. A. Tan • J. C. Spencer • C. Atkin
•
Received: 4 June 2013 / Accepted: 30 September 2013 / Published online: 17 October 2013 Ó Springer-Verlag Berlin Heidelberg 2013
Abstract Purpose Pancreatic enzymes are routinely measured during reception of trauma patients to assess for pancreatic injury despite conflicting evidence on their utility. The aim of this study was to investigate the utility of routine initial serum lipase measurement for the diagnosis of acute pancreatic trauma. Materials and methods Lipase measurements were introduced as part of the trauma pathology panel and requested on all patients who presented to an adult major trauma service and met trauma call-out criteria. Clinical records of these patients were extracted from the trauma registry and retrospectively reviewed. The performance of an initial serum lipase level measured on presentation to detect pancreatic trauma was determined. Results There were 2,580 patients included in the study, with 17 patients diagnosed with pancreatic trauma. An elevated lipase was recorded in 390 patients. Statistically significant associations were observed for B. Mitra (&) M. Fitzgerald G. A. Tan J. C. Spencer Emergency and Trauma Centre, The Alfred Hospital, Commercial Rd., Melbourne, VIC 3004, Australia e-mail: [email protected] B. Mitra Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia B. Mitra M. Fitzgerald National Trauma Research Institute, Melbourne, Australia M. Fitzgerald M. Raoofi C. Atkin Trauma Service, The Alfred Hospital, Melbourne, Australia M. Fitzgerald G. A. Tan Central Clinical School, Monash University, Melbourne, Australia
elevated lipase in patients with pancreatic trauma, head injury, acute alcohol ingestion and massive blood transfusion. As a test for pancreatic trauma, an abnormal serum lipase result had a specificity of 85.3 % (95 % CI 83.8–86.6), sensitivity of 76.5 % (95 % CI 49.8–92.2), positive predictive value of 3.3 % (95 % CI 1.8–5.8) and negative predictive value of 99.8 % (95 % CI 99.4–99.9). Higher cut-offs of serum lipase did not result in better performance. Conclusions A normal serum lipase result can be a useful adjunct to exclude pancreatic injury. A positive lipase result, regardless of the cut-off used, was not reliably associated with pancreatic trauma, and should not be used to guide further assessment. Keywords Wounds and injuries Pancreas Abdomen Resuscitation Hematologic tests Tomography X-Ray computed
Introduction Pre-determined laboratory tests are commonly requested during reception of major trauma patients. The rationale of this practice stems from the fear of missing vital investigation results during the hectic phase of initial resuscitation. However, routine panels have been previously shown to be neither useful, nor cost-effective [1–3]. Selective, patient specific investigations have been previously suggested, but translation of this evidence to practice has been slow with a range of laboratory panels continuing to be recommended for trauma resuscitation [4]. A measure of pancreatic enzymes, serum amylase, serum lipase or both, are commonly part of trauma laboratory panels. The utility of abnormal pancreatic
123
310
B. Mitra et al.
Table 1 The Alfred trauma call-out criteria Mechanism
MBA/cyclist impact [30 km/h
Table 2 Grading of pancreatic injury Grade
Criteria
I
Simple contusion of the pancreas
II
Major contusion or laceration without tissue loss or involvement of the main pancreatic duct
III
Complete transection of the pancreas or a parenchymal injury with involvement of the major duct to the left of the superior mesenteric vein
Explosion
IV
All significant blunt trauma assessed by ambulance
Ductal transection or a major parenchymal injury to the right of the superior mesenteric vein
V
Massive disruption of the head of the pancreas
Pedestrian impact [30 km/h Extrication [30 min Vehicle rollover Fatality in same vehicle Ejection from vehicle Fall [3 m Injuries (not exclusive)
All penetrating head, neck and truncal injuries Traumatic amputation of carpus/tarsus # pelvis/pulseless limb/#dislocations with vascular compromise Evisceration
lipase measurement for the diagnosis of acute pancreatic trauma following the introduction of routine lipase measurements on all trauma call-out patients.
Blast injuries Severe crush injury Severe burns ([20 % TBSA); All facial burns Signs
SBP \ 100 mmHg GCS \ 14 SpO2 \ 90 % RR \ 10 or [30 beats min-1
Treatment
Any airway manoeuvre including intubation Assisted ventilation Pleural decompression Haemostatic dressings/tourniquet application [1,000 mL IV fluid or blood administration Neuromuscular blockade
Other criteria
Mass casualty incident All inter-hospital trauma transfers Pregnancy Significant comorbidity Anticoagulant therapy
enzymes is presumed to aid the early diagnosis of pancreatic trauma. Among the paediatric population, evidence for the lack of utility of routine amylase or lipase measurements is relatively strong with multiple studies and cost-effectiveness analyses [5–7]. However, conflicting evidence persists for the adult population. The inclusion of serum amylase or lipase to the routine laboratory parameters after trauma in adult patients have been previously recommended as being indicative of the need for early surgical intervention [8, 9]. However, such findings have been refuted by suggestions that acute serum amylase and lipase levels are neither specific nor sensitive for acute pancreatic trauma [10–13]. The aim of this study was to determine the incidence of patients with pancreatic trauma presenting to a major trauma centre and investigate the utility of initial serum
123
Methods Setting The state of Victoria, Australia has one paediatric and two adult major trauma services (MTS) located within metropolitan Melbourne. Major trauma triage guidelines direct 85 % of major trauma patients to an MTS for definitive treatment. The Alfred Hospital is one of the two adult MTS, with a trauma database prospectively capturing prehospital and hospital data on all major trauma patients, defined as ISS [15, or requiring urgent surgery, or intensive care unit (ICU) admission or death or[72 h admission to hospital post trauma. The Alfred trauma call-out criteria are listed in Tables 1, 2. All patients meeting these criteria are received at the trauma centre with guidelines recommending trauma panel laboratory investigations that include serum lipase requested on presentation. Patient selection All patients included in the Alfred Hospital Trauma registry, presenting between August 2010 and December 2011 post blunt trauma, were included in this study. Cases that did not have trauma team activation or did not have an initial serum lipase result recorded as part of the trauma panel of tests were excluded, as trauma reception guidelines may not have been followed. An injury to the pancreas was recorded if confirmed by imaging or intraoperatively. Hyperlipasaemia as the sole criteria for pancreatic injury was not accepted. The serum lipase was measured using spectrophotometry. The normal reference range was 10–70 U/L. The shock index is defined by heart rate over systolic blood pressure and recorded to be abnormal for a value over 1.0 [14]. Massive blood
Serum lipase for assessment of pancreatic trauma
transfusion was defined as C5 units of packed red blood cells in 4 h [15]. Any positive blood alcohol result was considered abnormal. Pancreatic trauma was classified as proposed by the American Association for the Surgery of Trauma Committee on Organ Injury Scaling [16].
311 Table 3 Comparison of patients with elevated and non-elevated serum lipase Elevated lipase (n = 390) Age (years) ISS
43.0 (20.1)
10 (5–22)
Temperature (°C)
10 (5–18)
36.3 (1.0)
Heart rate (beats min-1) Systolic blood pressure (mmHg) Systolic blood pressure \100 mmHg
36.4 (0.8)
90.1 (21.4)
88.1 (20.3)
143.5 (25.0)
146.1 (27.4)
15 (3.8 %)
88 (4.0 %)
Shock index [1
20 (5.1 %)
90 (4.1 %)
GCS
15 (14–15)
15 (14–15)
Massive transfusion*
23 (5.9 %)
61 (2.8 %)
103 (26.4 %) 148 (37.9 %)
398 (18.2 %) 693 (31.6 %)
13 (3.3 %)
4 (0.2 %)
Blood alcohol positive* Head injury* Pancreatic injury*
0.50
Sensitivity
0.75
1.00
* Statistically significant
0.00
Medical records of all patients who had a pancreatic injury coded were extracted to determine the mode of diagnosis. Diagnostic accuracy of serum lipase in detecting pancreatic trauma was reported using sensitivity, specificity, positive and negative predictive values and 95 % confidence intervals. A two-sided p value of \0.05 was considered to be statistically significant. Normally distributed continuous data were reported as mean (standard deviation). Ordinal or skewed data were reported as median (interquartile range) and categorical data as counts (proportions). Statistical significance for differences between means were tested using the Student’s t test, differences between medians were tested using the Wilcoxon rank sum test while differences between proportions were tested the v2 test. Predictive accuracy of accumulative effect of these variables was assessed by measuring the area under a receiver operating characteristic (ROC) curve. Cases with data missing with regards to lipase measurement were handled by list-wise exclusion. All analyses were performed using Stata version 11.0 (Statacorp, College Station, TX, USA). This study was approved by The Alfred Hospital Research and Ethics Committee.
41.0 (18.1)
0.25
Statistical analysis
Non-elevated lipase (n = 2,190)
Results
0.00
0.25
0.50
0.75
1.00
1 - Specificity Area under ROC curve = 0.8941
A total of 4,808 patients were identified in the database for the study period. There were 1,936 patients excluded who did not have a trauma team activation and a further 287 patients excluded as not having an initial serum lipase result recorded. There were 22 patients who were coded as having pancreatic trauma. Following review of medical records, five more cases were excluded as there was no radiological or operative evidence of pancreatic trauma, with the diagnosis based on the lipase level alone. There were 2,580 patients included, of which 17 patients were diagnosed with pancreatic injury. This represented an incidence of 0.7 % of pancreatic trauma among patients presenting after blunt trauma. There were 390 patients with an initial abnormal serum lipase recorded. Demographic and injury characteristics of patients with elevated and non-elevated serum lipase are presented in Table 3. As a test for pancreatic trauma, an abnormal serum lipase result ([70 U/L) was found to have
Fig. 1 ROC for lipase and pancreatic injury
a specificity of 85.3 % (95 % CI 83.8–86.6) with a sensitivity of 76.5 % (95 % CI 49.8–92.2), a positive predictive value (PPV) of 3.3 % (95 % CI 1.8–5.8) and a negative predictive value (NPV) of 99.8 % (95 % CI 99.4–99.9). Figure 1 illustrates the area under the ROC of 0.89 (95 % CI 0.84–0.95). Using a cut-off of serum lipase of 500 U/L, specificity for pancreatic injury was 98.2 % (95 % CI 97.6–98.7) with a sensitivity of 35.3 % (95 % CI 15.2–61.4), a PPV of 11.7 % (95 % CI 4.9–24.5) and a NPV of 99.6 % (99.2–99.8 %). At a cut-off of 1,000 U/L, specificity for pancreatic injury was 99.5 % (95 % CI 99.1–99.7) with a sensitivity of 17.6 % (95 % CI 4.7–44.2), a PPV of 20.0 % (95 % CI 5.3–48.6) and a NPV of 99.4 % (95 % CI 99.1–99.7).
123
312
The initial diagnosis of pancreatic injury was made using abdominal CT in 16 patients with one patient diagnosed using ultrasound. Of the 17 patients diagnosed with pancreatic trauma, nine patients underwent an explorative laparotomy, with two patients undergoing distal pancreatectomies. These two patients were classified as having grade IV pancreatic trauma. A further four patients had drains inserted only with Grade I or II injuries. Patients who did not undergo surgery were all classified as having grade I or II injuries. There were no deaths among patients with pancreatic trauma. The average length of stay was 23.2 (14.4) days with 10 patients requiring admission to the ICU. Among the four patients with normal initial serum lipase result but diagnosed to have pancreatic trauma, the diagnoses were made on CT abdomen. None of these patients had repeat serum lipase measurements. All four patients had other injury to abdominal organs with two patients sustaining bowel injuries, one renal injury and the other sustaining liver lacerations. Among excluded patients, there were no patients diagnosed with pancreatic injury and 81 of the excluded patients had an abnormal lipase level of [70 U/L.
Discussion In our population of adult patients post blunt trauma presenting to a major trauma centre, the incidence of pancreatic injury was low. A serum lipase level recorded immediately on presentation to the ED was highly sensitive and may be useful in its ability to exclude pancreatic injury with a NPV of over 99 %. The PPV for an initial lipase measurement to detect pancreatic injury was low, suggesting that the serum lipase result should not be used to guide further investigations or management for pancreatic injury. The PPV did not improve when higher cut-offs of serum lipase were considered. Provided the limitations are considered, this study supports the inclusion of serum lipase among the panel of pathology tests during initial trauma reception. The incidence of pancreatic trauma been previously reported at \1 % is consistent with our population [8, 17]. The low incidence has resulted in a paucity of clinical studies related to the diagnosis of pancreatic trauma. As a result, retrospective case-series have used inclusion criteria of patients with diagnosed pancreatic trauma only, where the NPV could not be evaluated. The reported poor utility of serum lipase, therefore, had been previously based on the low PPVs of serum pancreatic enzymes to detect pancreatic trauma—reported to be 1.5 % and supported by this study [11, 12]. The low PPV of serum lipase for pancreatic trauma is likely due to other causes of hyperlipasaemia in major trauma patients. A high serum lipase has been previously
123
B. Mitra et al.
associated with head injury and confirmed in this study [18–21]. The serum lipase was also univariately associated with recent alcohol ingestion. An association with massive blood transfusion was also observed and this had also been previously reported along with an absence of association with mortality [22]. The blunt force required to injure the pancreas is substantial and a pancreatic injury is rarely a solitary injury. When the pancreas is injured in the setting of blunt trauma, other abdominal organs are also likely to be affected. Where clinical indications exist for further evaluation of other abdominal organs using imaging or surgery, the pancreas can be assessed as part of that procedure. In the setting of no such indication, an isolated elevated serum lipase result should not be used as an indication to further assess for pancreatic injury. Patients with a reliable history, lack of clinical signs and isolated elevated serum lipase results in the setting of trauma should be observed for clinical signs and undergo repeated measurements [23]. An arbitrary cut-off value of both amylase and lipase of three times normal is frequently used by clinicians to discriminate between those with acute pancreatitis from those with non-pancreatic abdominal pain and translated into the setting of pancreatic trauma [24]. We have shown that a higher cut-off does not improve performance of the test in this setting and therefore, should not be used. This study is limited in being a retrospective review of the data, but included a large number of patients who underwent serum lipase testing in the setting of blunt trauma immediately upon presentation to the ED. Patients who did not have a trauma call-out were excluded, but there were no cases of pancreatic trauma among excluded patients and as such, their inclusion regardless of the lipase result, would not have altered the NPV. On the contrary, it would have further lowered the PPV, strengthening our view that an elevated initial serum lipase result alone should not be used to guide further assessment for pancreatic trauma. Post-traumatic pancreatitis has been previously described, with an incidence of up to 17 % among trauma patients and may be secondary to undiagnosed pancreatic trauma [25]. No cases of delayed diagnosis were observed in our cohort, but follow up beyond hospital discharge was not conducted. The exact timing of injury and time from injury to presentation to the ED was not available. This could be important as delayed measurement of lipase has previously been suggested to have higher specificity for detecting pancreatic trauma [7]. The current state-wide Victorian Trauma System mandates direct transport of patients who are within 30 min of transport time and early transfer for other major trauma services [26]. Adding in time from injury to ambulance arrival and scene times, we estimate most patients to have had a serum lipase measurement after at least 1 h of injury. It is, therefore, unlikely that
Serum lipase for assessment of pancreatic trauma
specificity of the test would improve to the point of being clinically useful. While an initial serum lipase appears to be efficacious in excluding pancreatic trauma, further prospective studies are required to confirm the utility of a serum lipase together with clinical examination to determine clinical effectiveness and if established, cost-effectiveness studies will be required to determine efficiency. The low incidences of pancreatic trauma presenting to single centres make such studies difficult to design and confidently determine PPVs. In the absence of an appropriate level of evidence to guide efficiency, cost-effective studies will be premature. Collaborative multi-centre trauma registries may provide the answer to facilitating such clinical studies in the future.
313
7.
8.
9.
10.
11.
12.
Conclusions Despite discouraging previous reports, a negative initial serum lipase result appears to be an useful adjunct during the assessment of the major trauma patient in being able to exclude pancreatic injury. We support the retention of serum lipase as a screening test for all patients presenting with major trauma. A positive result, regardless of the cutoff used, was not diagnostic of pancreatic trauma, and should not be used to guide further assessment.
13. 14.
15.
16. Acknowledgments We would like to thank staff at The Alfred Hospital Trauma Registry and in particular, Ms. Louise Niggemeyer for extraction of data. Conflict of interest
17.
None to declare. 18.
References 1. Namias N, McKenney MG, Martin LC. Utility of admission chemistry and coagulation profiles in trauma patients: a reappraisal of traditional practice. J Trauma. 1996;41(1):21–5. 2. Tasse JL, Janzen ML, Ahmed NA, Chung RS. Screening laboratory and radiology panels for trauma patients have low utility and are not cost effective. J Trauma. 2008;65(5):1114–6. doi:10. 1097/TA.0b013e318184b4f2. 3. Tortella BJ, Lavery RF, Rekant M. Utility of routine admission serum chemistry panels in adult trauma patients. Acad Emerg Med. 1995;2(3):190–4. 4. Chu UB, Clevenger FW, Imami ER, Lampard SD, Frykberg ER, Tepas JJ 3rd. The impact of selective laboratory evaluation on utilization of laboratory resources and patient care in a level-I trauma center. Am J Surg. 1996;172(5):558–62. doi:10.1016/ s0002-9610(96)00234-6 (discussion 62-3). 5. Adamson WT, Hebra A, Thomas PB, Wagstaff P, Tagge EP, Othersen HB. Serum amylase and lipase alone are not costeffective screening methods for pediatric pancreatic trauma. J Pediatr Surg. 2003;38(3):354–7. doi:10.1053/jpsu.2003.50107 (discussion-7). 6. Herman R, Guire KE, Burd RS, Mooney DP, Ehlrich PF. Utility of amylase and lipase as predictors of grade of injury or outcomes
19.
20.
21. 22.
23.
24.
25.
26.
in pediatric patients with pancreatic trauma. J Pediatr Surg. 2011;46(5):923–6. doi:10.1016/j.jpedsurg.2011.02.033. Matsuno WC, Huang CJ, Garcia NM, Roy LC, Davis J. Amylase and lipase measurements in paediatric patients with traumatic pancreatic injuries. Injury. 2009;40(1):66–71. doi:10.1016/j. injury.2008.10.003. Mayer JM, Tomczak R, Rau B, Gebhard F, Beger HG. Pancreatic injury in severe trauma: early diagnosis and therapy improve the outcome. Dig Surg. 2002;19(4):291–7. doi:64576 (discussion 7-9). Kumar S, Sagar S, Subramanian A, Albert V, Pandey RM, Kapoor N. Evaluation of amylase and lipase levels in blunt trauma abdomen patients. J Emerg Trauma Shock. 2012;5(2):135–42. doi:10.4103/0974-2700.96482. Buechter KJ, Arnold M, Steele B, Martin L, Byers P, Gomez G, et al. The use of serum amylase and lipase in evaluating and managing blunt abdominal trauma. Am Surg. 1990;56(4):204–8. Boulanger BR, Milzman DP, Rosati C, Rodriguez A. The clinical significance of acute hyperamylasemia after blunt trauma. Can J Surg. 1993;36(1):63–9. Farkouh E, Wassef R, Atlas H, Allard M. Importance of the serum amylase level in patients with blunt abdominal trauma. Can J Surg. 1982;25(6):626–8. Olsen WR. The serum amylase in blunt abdominal trauma. J Trauma. 1973;13(3):200–4. Tourtier JP, Jost D, Domanski L. Shock index: a simple clinical parameter for mortality risk assessment in trauma? J Trauma Acute Care Surg. 2012;73(3):780–1. doi:10.1097/TA.0b013 e31825ff540 (author reply 1). Mitra B, Cameron PA, Gruen RL, Mori A, Fitzgerald M, Street A. The definition of massive transfusion in trauma: a critical variable in examining evidence for resuscitation. Eur J Emerg Med. 2011;18(3):137–42. doi:10.1097/MEJ.0b013e328342310e. Moore EE, Cogbill TH, Malangoni MA, Jurkovich GJ, Shackford SR, Champion HR, et al. Organ injury scaling. Surg Clin North Am. 1995;75(2):293–303. Lin BC, Chen RJ, Fang JF, Hsu YP, Kao YC, Kao JL. Management of blunt major pancreatic injury. J Trauma. 2004;56(4): 774–8. Pezzilli R, Billi P, Barakat B, Fiocchi M, Re G, Gullo L, et al. Serum pancreatic enzymes in patients with coma due to head injury or acute stroke. Int J Clin Lab Res. 1997;27(4):244–6. Justice AD, DiBenedetto RJ, Stanford E. Significance of elevated pancreatic enzymes in intracranial bleeding. South Med J. 1994;87(9):889–93. Vitale GC, Larson GM, Davidson PR, Bouwman DL, Weaver DW. Analysis of hyperamylasemia in patients with severe head injury. J Surg Res. 1987;43(3):226–33. Bouwman DL, Altshuler J, Weaver DW. Hyperamylasemia: a result of intracranial bleeding. Surgery. 1983;94(2):318–23. Malinoski DJ, Hadjizacharia P, Salim A, Kim H, Dolich MO, Cinat M, et al. Elevated serum pancreatic enzyme levels after hemorrhagic shock predict organ failure and death. J Trauma. 2009;67(3):445–9. doi:10.1097/TA.0b013e3181b5dc11. Degiannis E, Glapa M, Loukogeorgakis SP, Smith MD. Management of pancreatic trauma. Injury. 2008;39(1):21–9. doi:10. 1016/j.injury.2007.07.005. Smith RC, Southwell-Keely J, Chesher D. Should serum pancreatic lipase replace serum amylase as a biomarker of acute pancreatitis? ANZ J Surg. 2005;75(6):399–404. doi:10.1111/j. 1445-2197.2005.03391.x. Fleming WR, Collier NA, Banting SW. Pancreatic trauma: Universities of Melbourne HPB Group. Aust N Z J Surg. 1999;69(5):357–62. Cox S, Currell A, Harriss L, Barger B, Cameron P, Smith K. Evaluation of the Victorian state adult pre-hospital trauma triage criteria. Injury. 2012;43(5):573–81. doi:10.1016/j.injury.2010.10.003.
123
