abdominal ct scanning for trauma: how low can we go?

7
Abdominal CT scanning for trauma: how low can we go? David G. Jacobs*, Jennifer L. Sarafin, John A. Marx Department of Surgery and Emergency Medicine, Carolinas Medical Center, P.O. Box 32861, Charlotte, NC 28232-2861, USA Accepted 15 December 1999 Abstract Purpose: computed tomography (CT) of the abdomen is an established, albeit expensive and perhaps overused, diagnostic modality for the evaluation of the injured patient. We developed a practice management guideline for blunt abdominal trauma intended to reduce the percentage of negative CT scans, yet minimize delayed recognition of injury and non-therapeutic laparotomy. Procedures: between April 1996 and March 1997, 1147 adult patients at risk for blunt abdominal injury were admitted to our Level I trauma centre and underwent abdominal evaluation according to the practice management guideline. Main Findings: abdominal CT was performed in 522 patients (45%), and 441 scans were negative (85%). Delayed recognition of injury and non-therapeutic laparotomy rates were low, 4% and 1.6%, respectively. Principal Conclusion: abdominal CT scanning in trauma patients can achieve low non-therapeutic laparotomy and delayed recognition of injury rates but at the expense of high negative CT scan rates. Greater reliance on the physical examination and perhaps abdominal ultrasound may reduce negative CT scan rates and yet preserve low non-therapeutic laparotomy and delayed recognition of injury rates. 7 2000 Elsevier Science Ltd. All rights reserved. 1. Introduction Computed tomography (CT) of the abdomen has become the diagnostic modality of choice for the evaluation of hemodynamically stable trauma patients with the potential for abdominal injury. However, it is relatively expensive, and most abdominal CT examin- ations obtained under these circumstances are normal [1–5]. As a result, there is growing interest in more cost-eective diagnostic strategies, particularly abdomi- nal ultrasound and serial physical examinations in selected clinical situations. Simultaneously, the devel- opment of practice management guidelines for the diagnosis and treatment of injured patients has been advocated as a way to minimize provider variability in common clinical situations and to potentially reduce overall costs [6]. We developed a practice management algorithm for the evaluation of patients at risk for intra-abdominal injury, examined its clinical ecacy and safety, and established a benchmark for appropri- ate use of abdominal CT scanning against which we could compare subsequent modifications of the diag- nostic algorithm. 2. Patients and methods Our medical center is an 843-bed, acute care hospital and a Level I trauma center designated by the state of North Carolina, USA, and by the American College of Surgeons Committee on Trauma. Injured patients ful- filling specific trauma triage criteria are evaluated im- mediately upon arrival by a multi-disciplinary trauma resuscitation team supervised by attending physicians from emergency medicine and trauma surgery. De- cisions regarding the need for CT or diagnostic perito- neal lavage (DPL) for patients requiring admission to the trauma service are made by that service’s chief resi- dent, with input from the service attending and from emergency medicine. A comprehensive practice management guideline Injury, Int. J. Care Injured 31 (2000) 337–343 0020-1383/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved. PII: S0020-1383(99)00306-X www.elsevier.com/locate/injury * Corresponding author. Tel.: +1-704-355-3176; fax: +1-704-355- 5619. E-mail address: [email protected] (D.G. Jacobs).

Upload: david-g-jacobs

Post on 14-Sep-2016

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Abdominal CT scanning for trauma: how low can we go?

Abdominal CT scanning for trauma: how low can we go?

David G. Jacobs*, Jennifer L. Sara®n, John A. Marx

Department of Surgery and Emergency Medicine, Carolinas Medical Center, P.O. Box 32861, Charlotte, NC 28232-2861, USA

Accepted 15 December 1999

Abstract

Purpose: computed tomography (CT) of the abdomen is an established, albeit expensive and perhaps overused, diagnosticmodality for the evaluation of the injured patient. We developed a practice management guideline for blunt abdominal traumaintended to reduce the percentage of negative CT scans, yet minimize delayed recognition of injury and non-therapeutic

laparotomy. Procedures: between April 1996 and March 1997, 1147 adult patients at risk for blunt abdominal injury wereadmitted to our Level I trauma centre and underwent abdominal evaluation according to the practice management guideline.Main Findings: abdominal CT was performed in 522 patients (45%), and 441 scans were negative (85%). Delayed recognition of

injury and non-therapeutic laparotomy rates were low, 4% and 1.6%, respectively. Principal Conclusion: abdominal CT scanningin trauma patients can achieve low non-therapeutic laparotomy and delayed recognition of injury rates but at the expense ofhigh negative CT scan rates. Greater reliance on the physical examination and perhaps abdominal ultrasound may reducenegative CT scan rates and yet preserve low non-therapeutic laparotomy and delayed recognition of injury rates. 7 2000

Elsevier Science Ltd. All rights reserved.

1. Introduction

Computed tomography (CT) of the abdomen hasbecome the diagnostic modality of choice for theevaluation of hemodynamically stable trauma patientswith the potential for abdominal injury. However, it isrelatively expensive, and most abdominal CT examin-ations obtained under these circumstances are normal[1±5]. As a result, there is growing interest in morecost-e�ective diagnostic strategies, particularly abdomi-nal ultrasound and serial physical examinations inselected clinical situations. Simultaneously, the devel-opment of practice management guidelines for thediagnosis and treatment of injured patients has beenadvocated as a way to minimize provider variability incommon clinical situations and to potentially reduceoverall costs [6]. We developed a practice managementalgorithm for the evaluation of patients at risk for

intra-abdominal injury, examined its clinical e�cacyand safety, and established a benchmark for appropri-ate use of abdominal CT scanning against which wecould compare subsequent modi®cations of the diag-nostic algorithm.

2. Patients and methods

Our medical center is an 843-bed, acute care hospitaland a Level I trauma center designated by the state ofNorth Carolina, USA, and by the American College ofSurgeons Committee on Trauma. Injured patients ful-®lling speci®c trauma triage criteria are evaluated im-mediately upon arrival by a multi-disciplinary traumaresuscitation team supervised by attending physiciansfrom emergency medicine and trauma surgery. De-cisions regarding the need for CT or diagnostic perito-neal lavage (DPL) for patients requiring admission tothe trauma service are made by that service's chief resi-dent, with input from the service attending and fromemergency medicine.

A comprehensive practice management guideline

Injury, Int. J. Care Injured 31 (2000) 337±343

0020-1383/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved.

PII: S0020-1383(99 )00306 -X

www.elsevier.com/locate/injury

* Corresponding author. Tel.: +1-704-355-3176; fax: +1-704-355-

5619.

E-mail address: [email protected] (D.G. Jacobs).

Page 2: Abdominal CT scanning for trauma: how low can we go?

was developed that speci®cally de®ned the indicationsfor abdominal CT in patients with suspected bluntinjury to the abdomen. In an attempt to eliminate un-necessary CT examinations, the role of the abdominalphysical examination was expanded in the new algor-ithm. Instead of undergoing CT (the practice beforeguideline implementation), hemodynamically stablepatients with minimal decreases in levels of conscious-ness, as well as those requiring brief extra-abdominaloperative procedures, were evaluated with serial physi-cal examinations of the abdomen and hemoglobin de-

terminations (Fig. 1Fig. 1). Abdominal CT wasreserved for hemodynamically stable patients whosemental status precluded reliance on the abdominalexamination and for those with reliable examinationsbut localized abdominal tenderness. Abdominal ultra-sound was not incorporated into the diagnostic algor-ithm, since su�cient experience with this modality hadnot yet been achieved at the time of algorithm devel-opment.

CT scanning for acute abdominal injury was doneusing a General Electric CTI scanner (G.E. Medical

Fig. 1. Algorithm for blunt abdominal trauma. (1) Includes hypotension, tachycardia, base de®cit, unexplained fall in hemoglobin, and signi®cant

¯uid requirement to maintain stability. (2) DPT = diagnostic peritoneal tap; done using closed Seldinger technique (open technique if patient is

pregnant or has had prior abdominal surgery); positive de®ned as >10 cc gross blood; should be preceded by chest X-ray and pelvis ®lm when

possible. Immediate laparotomy (bypassing DPT) is acceptable alternative assuming other sources of hemorrhage have been ruled out. (3) DPL

= diagnostic peritoneal lavage; e�uent volume should be at least 50% of infused volume Ð if not, proceed to open DPL; positive DPL is

>100,000 RBC/mm3 or >500 WBC/mm3; should be performed while pursuing extra-abdominal sources of hemorrhage (i.e., angiography, etc.)

or other causes of shock (i.e., cardiogenic, neurogenic). Patients with negative DPLs should be admitted for serial abdominal examinations and

hemoglobin determinations. (4) GCS = glasgow coma scale score. (5) Refers to signi®cant epidural or subdural hematoma requiring emergent

craniotomy. (6) Abdominal examination may be confounded by alcohol, drugs, brain injury, spinal cord injury, or by contiguous (i.e., rib frac-

tures) or remote distracting (i.e., femur fracture) injuries. Assessment should be based on clinical ®ndings and not on lab studies (i.e., ETOH

level, toxicology screen, etc.). (7) To operating room for repair of extra-abdominal injuries under general anesthesia. (h) ``High risk'' refers to

patients who, as determined by the responsible physician, are at signi®cant risk for the development of intraoperative hemodynamic instability

(therefore, raising question of intra-abdominal injury). Determination of risk to be based on all available data including injury mechanism, antici-

pated length of surgical procedure, and anticipated blood loss. (9) For high risk patients, DPL may be substituted here at the discretion of the re-

sponsible physician. (10) Frequency of abdominal examinations and hemoglobin determinations dictated by clinical circumstances.

D.G. Jacobs et al. / Injury, Int. J. Care Injured 31 (2000) 337±343338

Page 3: Abdominal CT scanning for trauma: how low can we go?

Systems, Waukesha, WI) with images obtained at 1 cmintervals between diaphragm and the lower pole of thekidney, and at 2 cm intervals from this point to thepelvic outlet. Intravenous contrast enhancement wasacheived with a 2 cc/s injection of 125±150 cc ofOptiray-320 (Ioversal Injection 68%; Mallinkrodt,Inc., St. Louis, MO). Per trauma protocol, enteral con-trast material was not administered. Scans were inter-preted immediately on completion by in-house, board-certi®ed radiologists and later reviewed by radiologistswith particular expertise in CT interpretation. This®nal dictated interpretation was used in our analysis.

After adopting the practice management guideline asthe standard of care for the trauma service, we felt itwas critical to evaluate, concurrently, the e�cacy andsafety of the algorithm. Data were collected on alladult (ager 16 years) patients admitted to or seen inconsultation by the trauma service who were believed(due to mechanism of injury)to be at risk for havingabdominal injury. The presence or absence of abdomi-nal injury and the interval between admission and thediagnosis of abdominal injury (if present) were noted.The performance of negative or nontherapeutic laparo-tomies was determined from review of the dictated op-erative note and through discussions with theresponsible surgeons. For those patients who under-went abdominal CT as part of their initial traumaevaluation, scans were categorized as positive or nega-tive depending on the presence or absence, respect-ively, of injury-related abdominal ®ndings on the scan.From these data, the following rates were determined.

1. CT utilization (CTU) rate: percentage of patients atrisk for blunt abdominal injury who undergo ab-dominal CT scan as part of admission traumawork-up.

2. Negative abdominal CT scan (NCT) rate: percen-tage of scanned patients without evidence of ab-dominal injury on initial CT of abdomen.

3. Nontherapeutic laparotomy (NTL) rate: percentageof laparotomies performed in which no injuryrequiring repair was identi®ed.

4. Delayed recognition of abdominal injury (DRI)rate: percentage of patients with documented intra-abdominal injuries whose injuries were diagnosedmore than 12 h following admission.

3. Results

From April 1996 through March 1997, 2630 patientswith traumatic mechanisms were admitted to ourmedical center; 2014 of these were admitted to thetrauma service or had a trauma service consultation.Of these, 1181 had mechanisms of injury consistent

with possible blunt intra-abdominal injury. Twelve ofthese had no documentation of Glasgow coma scalescore (GCS); therefore, 1169 patients with completedata were evaluated in this study.

The practice management algorithm for the evalu-ation of blunt abdominal injury is shown in Fig. 1.Compliance with the algorithm was not speci®callydetermined, and seven patients had ultrasound examin-ation as the sole mode of abdominal evaluation in vio-lation of the algorithm. However, the majority oftreatment decisions were made with input by one ofthe ®ve full-time trauma faculty who developed the al-gorithm. In addition, adherence to the guideline by theresident sta� was monitored and enforced by one ofthree full-time trauma case managers.

Fig. 2 shows the primary means used to evaluate theabdomen in the 1169 patients in this study. CT scanand serial physical examinations with hemoglobin de-terminations were the most commonly employedmethods, with DPL, alone or in combination with CT,being used much less frequently. Twenty-two patientswere transferred to our institution having alreadyundergone either CT (20 patients) or DPL (2 patients).CT scans performed at outside institutions were notincluded in the determination of CTU or NCT rates.

CT utilization rates are shown in Table 1. Of the1147 patients who arrived at our trauma center with-out prior diagnostic studies, 522 underwent abdominalCT scanning as the sole means of initial abdominalevaluation (CTU = 45%). Generally, patients withlower GCS had higher CTU rates, although the rateswere actually higher for patients with GCS between 8and 10 than for patients with GCS 7 or less.

Negative abdominal CT scan rates are shown in

Table 1

CT utilization (CTU) rates for blunt abdominal trauma

Population CTU rate (%)

Entire group (n = 1147) 522/1147 (45.5)

GCSR 7 125/170 (73.5)

GCS 8±10 16/19 (84.2)

GCS 11±13 28/50 (56)

GCS 14±15 353/908 (38.9)

Table 2

Negative abdominal CT scan (NCT) rates for blunt abdominal

trauma

Population NCT rate (%)

All patients (n = 522) 441/522 (84.5)

GCSR 7 110/125 (88)

GCS 8±10 16/16 (100)

GCS 11±13 24/28 (85.7)

GCS 14±15 291/353 (82.5)

D.G. Jacobs et al. / Injury, Int. J. Care Injured 31 (2000) 337±343 339

Page 4: Abdominal CT scanning for trauma: how low can we go?

Table 2. Overall, 85% of abdominal CT scansobtained as part of the initial evaluation for abdominalinjury were negative. When the data were analyzedaccording to the patient's mental status at the time ofadmission, negative abdominal CT rates remained highin all GCS groups. With regard to the overall safety ofthe diagnostic algorithm, only 1 of 63 laparotomiesperformed for blunt abdominal injury was judged tobe nontherapeutic (NTL rate = 1.6%). Similarly, theDRI rate was only 4%, with only 5 of 138 patientswith documented intra-abdominal injuries followingblunt trauma being diagnosed more than 12 h follow-ing admission (Table 3).

4. Discussion

The role of CT scanning in the diagnosis of bluntabdominal trauma has evolved signi®cantly since itsintroduction into clinical practice almost 20 years ago[7]. Despite initial enthusiasm for this modality, early

reports focused on the low sensitivity of CT scanningcompared with peritoneal lavage [8±10]. Subsequentre®nements in CT technology have markedly reducedscanning time and increased scan resolution, resultingin excellent sensitivity and speci®city for most injuriesand making CT the test of choice for the hemodynami-cally stable patient with suspected blunt abdominalinjury [11]. However, since CT sensitivity is una�ectedand CT speci®city improved by large numbers of truenegative scans, there has previously been no impetusto focus on the negative abdominal CT scan fortrauma. Given that the patient charge associated withCT is approximately $1000 [5,12,13] and that the ma-jority of recent studies demonstrate NCT rates exceed-ing 60% [1±5], reports examining the role of CT inblunt abdominal trauma should consider CTU andNCT rates as well as sensitivity and speci®city.

After implementation of our algorithm, 45% ofpatients believed to be at risk for abdominal injuryunderwent CT scanning of the abdomen. CTU ratesdetermined from the literature vary from 19% to 56%

Fig. 2. Initial diagnostic modalities for evaluation of blunt abdominal trauma in 1169 patients. PTA = prior to admission; number in parenthesis

indicates number of patients.

Table 3

DRI: injuries diagnosed more than 12 after admission

Patient Primary method of abdominal

evaluation

Injuries Method of diagnosis

Patient 1 CT Grade I liver laceration; spleen laceration Review of original CT scan

Patient 2 DPL Grade II spleen and Grade III kidney injuries Laparotomy for hemodynamic instability

Patient 3 Physical examination Mesenteric laceration, small bowel necrosis Laparotomy for worsening exam

Patient 4 Physical examination Jejunal perforation, spleen laceration Delayed DPL for worsening exam;

laparotomy

Patient 5 CT Small bowel perforation Laparotomy for worsening exam

D.G. Jacobs et al. / Injury, Int. J. Care Injured 31 (2000) 337±343340

Page 5: Abdominal CT scanning for trauma: how low can we go?

[8,14,15,16], though none of these studies cite DRIrates for the entire population at-risk. Thus, althoughCTU rates lower than our reported 45% appear feas-ible, the bene®t of achieving this lower CTU rate (i.e.,reduced costs) may be o�set by the risk of harmfuldelays in the diagnosis and treatment of abdominalinjuries.

In spite of our attempt to de®ne the criteria for ab-dominal CT scanning in the trauma population, 85%of our abdominal CT scans showed no evidence ofintra-abdominal injury. Surprisingly, our NCT rate didnot vary with GCS (Table 2). Given the higher CTUrate in patients with GCS R 7, higher NCT rateswould have been an expected ®nding. Our data suggestthat low GCS, in and of itself, is a poor predictor ofinjury detection on abdominal CT, in agreement withthe ®ndings of Freid and colleagues [2]. Comparedwith CTU rates, NCT rates are more often reported inthe medical literature and range from 35% [17] to90% [16]. Most recent series are associated with NCTscan rates between 63% and 69% [1,2,4,5], althoughsome of these studies include chest injuries demon-strated on abdominal scanning among their positiveabdominal CT scans.

The true measure of any practice managementguideline is its safety. Any attempt to reduce the CTUrate or the NCT rate could be met with greater delaysin the diagnosis of abdominal injury. Such delays,however, are not necessarily evidence of algorithm fail-ure unless associated with delays in laparotomy oractual patient morbidity. Having said this, our DRIrate compares favorably with similar studies in the lit-erature. Sorkey and coworkers noted a 7% DRI ratebut de®ned this as operations undertaken more than24 following admission [18]. Watanabe and colleaguesdescribed three delayed laparotomies in a series of 155patients with abdominal injuries, for a delayed recog-nition rate of only 2% [19]. However, delayed laparot-omy rates would be expected to be lower than DRIrates in that the former do not include injuries whichwere recognized in a delayed fashion and subsequentlytreated non-operatively. Several authors cite low oreven absent false-negative CT scan rates[1,4,5,8,12,20,21,22]. Once again, however, DRI ratesmay not compare well with false-negative CT scanrates in that the latter include only patients whounderwent abdominal CT scanning, while our DRIrate includes every patient deemed at risk for blunt ab-dominal injury.

Compared with DRI rates, high NTL rates maymore closely re¯ect algorithm failure; speci®cally,excessive reliance on physical examination as a diag-nostic modality. Our NTL rate of 1.6% comparesquite favorably with other reported series. Unfortu-nately, most authors report their NTL rates only forthe subset of patients undergoing CT scanning

[4,14,16,19,20,22]. In an early population-based studysimilar to ours, Sorkey et al. noted a NTL rate of 5%[18], although NTL rates extracted from older studiessuch as this may not be truly comparable to ours fortwo reasons. First, operations deemed therapeutic asrecently as 10 years ago (e.g., drainage of a Grade I orII liver laceration) might not be considered therapeuticby today's standards. Second, given the recent trendtoward non-operative management of selected solidorgan injuries, operations undertaken 10 years ago anddeemed therapeutic (e.g., splenectomy or splenorrha-phy for Grade II splenic laceration) might well be trea-ted non-operatively today. Despite these di�erences, itis unlikely that a 1.6% NTL rate can be reduced sig-ni®cantly.

Our data suggest that, although safe as evidenced bylow NTL rates and DRI rates, our algorithm might beimproved by imposing even more strict criteria for theperformance of an abdominal CT scan. The high CTUrate in patients with GCS R 7 suggests that, in theabsence of more compelling physical ®ndings, thesepatients may not bene®t from routine abdominal CTscanning. Support for this notion can be found in areport by Freid and colleagues in which 150 adulttrauma patients underwent abdominal CT scanning fora variety of indications [2]. Of 112 scans performed forsolid clinical indications (abdominal pain or tender-ness, abdominal bruises or abrasions, nausea, vomit-ing, rib fractures, hematuria, dropping hemoglobin),45 demonstrated potentially signi®cant intra-abdomi-nal injury likely related to their present episode ofblunt trauma (NCT = 60%). However, of the 38scans for ``incidental'' indications (e.g., diminished sen-sorium, unreliable abdominal examination, need forgeneral anesthesia) only two scans were positive (NCT= 95%) [2]. Limiting CT scanning to patients withsolid indications will clearly reduce the CTU and NCTrates but will likely increase DRI rates, and perhapsplace undue reliance on the physical examination inthe bluntly injured patient. Richards et al., in a recentstudy of 444 patients who underwent abdominal CTfor blunt abdominal trauma, noted that the physicalexamination yielded a sensitivity of 63%, a speci®cityof 65%, a positive predictive value of 19%, and anegative predictive value of 93% [16].

Another suggested method of reducing reliance onabdominal CT is the use of DPL as a ``screening test''to determine the subset of patients who will bene®tfrom CT. Employing a protocol limiting CT topatients with a DPL red blood count more than 105/mm3, Schreiber and colleagues reduced CT use by 791scans over a 2-year period, and thus reduced patientcharges by $580,594 [13]. In addition, the non-thera-peutic laparotomy rate using this protocol was only5%, compared with 31% when the decision for lapar-otomy was based on DPL criteria alone.

D.G. Jacobs et al. / Injury, Int. J. Care Injured 31 (2000) 337±343 341

Page 6: Abdominal CT scanning for trauma: how low can we go?

Emergency department ultrasound performed by theemergency medicine physician or the trauma surgeonhas also been proposed as a means of reducing relianceon abdominal CT scanning for blunt abdominaltrauma. Algorithms for blunt abdominal trauma usingabdominal ultrasound have been proposed whichwould restrict CT scanning to hemodynamically stablepatients with either free ¯uid or equivocal ®ndings onthe ultrasound examination [23]. Recently, Branneyand colleagues demonstrated a reduction in the CTUrate from 56% to 26% by the addition of ultrasoundto their blunt abdominal trauma practice managementguideline [15]. Although this technology seems promis-ing, it is also very operator-dependent and requiresconsiderable experience before one can reliably baseclinical decisions on the ultrasound interpretation. Ad-ditional experience with greater numbers of patientswill be required before practice management guidelinesrelying on abdominal ultrasound can be accepted as astandard of care.

Our study su�ers from several methodological short-comings that may limit broad application of our ®nd-ings. First, we did not document actual compliancerates with the algorithm, and thus cannot be certainthat abdominal CT was used only as speci®ed withinthe algorithm. Because of the tendency to obtain CTscans in equivocal situations, our CTU and NCT ratesmay be somewhat higher than the minimum ratesobtainable with strict adherence to the algorithm.However, since the majority of CT scans were orderedwith the consent of one of the algorithm's developers,variance from the algorithm should have been mini-mal. Secondly, we do not have baseline CTU, NCT,DRI, or NTL rates obtained before algorithm im-plementation with which to compare our post-im-plementation rates. We, therefore, have no proof thatthe current approach to the diagnosis of intra-abdomi-nal injury o�ers any advantage over our previousmethods. However, our primary focus in this studywas not to demonstrate improved resource utilizationas a result of algorithm implementation, but rather toemphasize the high CTU and NCT rates whichoccurred despite our e�orts to reduce unnecessary CTscans. Finally, our analysis was limited to relativelycrude outcome measures (CTU, NCT, DRI, and NTLrates) and does not consider other equally important,though more di�cult-to-measure, outcomes such asthe true cost savings related to decreased CT use, mor-bidity/mortality rates associated with delayed recog-nition of abdominal injury and non-therapeuticlaparotomy, or the medico-legal bene®ts which mightresult from ``de®nitively'' excluding the presence of anabdominal injury. Neither are we able to take intoaccount the potential negative ®nancial impact of CTscans that, despite revealing abdominal injury, do notalter patient management (e.g., mild renal contusion).

Such detailed analyses were simply beyond the scopeof this study.

Before concluding, we must emphasize that,although attempts to reduce CTU and NCT rates areimportant, a negative abdominal CT scan may, undercertain circumstances, have considerable value. Arecent multi-institutional trial involving 2299 patientsfound that the negative predictive value of an abdomi-nal CT scan, as de®ned by the subsequent need forlaparotomy, was 99.6% and concluded that patientswith a CT scan after blunt abdominal trauma thatshows no organ injury or peritoneal ¯uid, could safelybe discharged from the emergency department withouta period of either inpatient or outpatient observation[24]. Certainly, many of the patients in our study whohad negative abdominal CT scans would still haverequired hospital admission for their extra-abdominalinjuries, but it may be precisely this patient populationin whom serial physical examination, peritoneal lavage,or abdominal ultrasound may be more cost-e�ectivealternatives to abdominal CT. For those patients with-out other indications for admission, abdominal CTscanning may prove more cost-e�ective than hospitaladmission in achieving abdominal clearance. Such anapproach could result in a considerable reduction inpatient charges if, as Livingston et al. estimates, 40%of patients who undergo CT scanning for blunt ab-dominal trauma could be sent home from the emer-gency department [24].

In summary, practice management guidelines for theevaluation of blunt abdominal trauma are valuable inthat they standardize an approach to care and prob-ably contain cost as well. Our current practice manage-ment guideline, however, although safe, has resulted inhigh CTU and NCT rates. There are probably mini-mum CTU and NCT rates below which unacceptableDRI and NTL rates will result. Neither the currentstudy nor the current literature suggests what theseminimum rates are. Only modi®cation of and sub-sequent surveillance of standardized practice manage-ment guidelines will lead us to the appropriate use ofthese costly resources.

References

[1] Ross jr P, Perkal MF, Degutis LC, Baker CC. Clari®cation of

the role of CT scan in the acute evaluation of blunt abdominal

trauma. Conn Med 1991;55:330±2.

[2] Fried AM, Humphries R, Scho®eld CN. Abdominal C.T. scans

in patients with blunt trauma: low yield in the absence of clini-

cal ®ndings. J CAT 1992;16:717±21.

[3] Udekwu PO, Gurkin B, Oller DW. The use of computed tom-

ography in blunt abdominal injuries. Am Surg 1996;62:56±9.

[4] Pevec WC, Peitzman AB, Udekwu AO, McCoy B, Straub W.

Computed tomography in the evaluation of blunt abdominal

trauma. S,G&O 1991;173:262±7.

[5] Rizzo AG, Steinberg SM, Flint LM. Prospective assessment of

D.G. Jacobs et al. / Injury, Int. J. Care Injured 31 (2000) 337±343342

Page 7: Abdominal CT scanning for trauma: how low can we go?

the value of computed tomography for trauma. J Trauma

1995;38:338±43.

[6] Rhodes M. Practice management guidelines for trauma care:

presidential address seventh scienti®c assembly of the Eastern

Association for the Surgery of Trauma. J Trauma 1994;37:635±

44.

[7] Federle MP, Goldberg HI, Kaiser JA, Moss AA, Je�rey Jr RB,

Mall JC. Evaluation of abdominal trauma by computed tom-

ography. Radiology 1981;138:637±44.

[8] Fabian TC, Mangiante EC, White TJ, Patterson CR,

Boldreghini S, Britt LG. A prospective study of 91 patients

undergoing both computed tomography and peritoneal lavage

following blunt abdominal trauma. J Trauma 1986;26:602±8.

[9] Davis RA, Shayne JP, Max MH, Wool®tt RA, Schwab W. The

use of computerized axial tomography versus peritoneal lavage

in the evaluation of blunt abdominal trauma: a prospective

study. Surgery 1985;98:845±9.

[10] Marx JA, Moore EE, Jorden RC, Eule Jr J. Limitations of

computed tomography in the evaluation of acute abdominal

trauma: a prospective comparison with diagnostic peritoneal

lavage. J Trauma 1985;25:933±7.

[11] Catre MG. Diagnostic peritoneal lavage versus abdominal com-

puted tomography in blunt abdominal trauma: a review of pro-

spective studies. CJS 1995;38:117±22.

[12] Blow O, Bassam D, Butler K, Cephas GA, Brady W, Young

JS. Speed and e�ciency in the resuscitation of blunt trauma

patients with multiple injuries: the advantage of diagnostic per-

itoneal lavage over abdominal computerized tomography. J

Trauma 1998;44:287±90.

[13] Schreiber MA, Gentilello LM, Rhee P, Jurkovich GJ, Maier

RV. Limiting computed tomography to patients with peritoneal

lavage-positive results reduces cost and unnecessary celiotomies

in blunt trauma. Arch Surg 1996;131:954±9.

[14] Matsubara TK, Fong HMT, Burns CM. Computed tomogra-

phy of abdomen (CTA) in management of blunt abdominal

trauma. J Trauma 1990;30:410±4.

[15] Branney SW, Moore EE, Cantrill SV, Burch JM, Terry SJ.

Ultrasound based key clinical pathway reduces the use of hospi-

tal resources for the evaluation of blunt abdominal trauma. J

Trauma 1997;42:1086±90.

[16] Richards JR, Derlet RW. Computed tomography and blunt ab-

dominal injury: patient selection based on examination, haema-

tocrit and haematuria. Injury 1997;28:181±5.

[17] Liu M, Lee C-H, Peng F-K. Prospective comparison of diagnos-

tic peritoneal lavage computed tomographic scanning, and

ultrasonography for the diagnosis of blunt abdominal trauma. J

Trauma 1993;35:267±70.

[18] Sorkey AJ, Farnell MB, Williams Jr HJ, Mucha Jr P, Ilstrup

DM. The complementary roles of diagnostic peritoneal lavage

and computed tomography in the evaluation of blunt abdomi-

nal trauma. Surgery 1989;106:794±801.

[19] Watanabe S, Ishi T, Kamachi M, Takahashi T. Computed tom-

ography and nonoperative treatment for blunt abdominal

trauma. Japanese J Surg 1990;20:56±63.

[20] Wing VW, Federle MP, Morris Jr JA, Je�rey RB, Bluth R. The

clinical impact of CT for blunt abdominal trauma. AJR

1985;145:1191±4.

[21] Nelson EW, Holliman CJ, Juell BE, Mintz S. Computerized

tomography in the evaluation of blunt abdominal trauma. Am J

Surg 1983;146:751±4.

[22] Peitzman AB, Makaroun MS, Slasky BS, Ritter P. Prospective

study of computed tomography in initial management of blunt

abdominal trauma. J Trauma 1986;26:585±92.

[23] Rozycki GS, Shackford SR. Ultrasound what every trauma sur-

geon should know. J Trauma 1996;40:1±4.

[24] Livingston DH, Lavery RF, Passannante MR, Skurnick JH,

Fabian TC, Fry DE, et al. Admission or observation is not

necessary after a negative abdominal computed tomographic

scan in patients with suspected blunt abdominal trauma: results

of a prospective, multi-institutional trial. J Trauma

1998;44:273±82.

D.G. Jacobs et al. / Injury, Int. J. Care Injured 31 (2000) 337±343 343