Open Access
Open Peer Review

This article has Open Peer Review reports available.

How does Open Peer Review work?

Emergency department spirometric volume and base deficit delineate risk for torso injury in stable patients

BMC Surgery20044:3

DOI: 10.1186/1471-2482-4-3

Received: 29 April 2003

Accepted: 19 January 2004

Published: 19 January 2004

Abstract

Background

We sought to determine torso injury rates and sensitivities associated with fluid-positive abdominal ultrasound, metabolic acidosis (increased base deficit and lactate), and impaired pulmonary physiology (decreased spirometric volume and PaO2/FiO2).

Methods

Level I trauma center prospective pilot and post-pilot study (2000–2001) of stable patients. Increased base deficit was < 0.0 in ethanol-negative and ≤ -3.0 in ethanol-positive patients. Increased lactate was > 2.5 mmol/L in ethanol-negative and ≥ 3.0 mmol/L in ethanol-positive patients. Decreased PaO2/FiO2 was < 350 and decreased spirometric volume was < 1.8 L.

Results

Of 215 patients, 66 (30.7%) had a torso injury (abdominal/pelvic injury n = 35 and/or thoracic injury n = 43). Glasgow Coma Scale score was 14.8 ± 0.5 (13–15). Torso injury rates and sensitivities were: abdominal ultrasound negative and normal base deficit, lactate, PaO2/FiO2, and spirometric volume – 0.0% & 0.0%; normal base deficit and normal spirometric volume – 4.2% & 4.5%; chest/abdominal soft tissue injury – 37.8% & 47.0%; increased lactate – 39.7% & 47.0%; increased base deficit – 41.3% & 75.8%; increased base deficit and/or decreased spirometric volume – 43.8% & 95.5%; decreased PaO2/FiO2 – 48.9% & 33.3%; positive abdominal ultrasound – 62.5% & 7.6%; decreased spirometric volume – 73.4% & 71.2%; increased base deficit and decreased spirometric volume – 82.9% & 51.5%.

Conclusions

Trauma patients with normal base deficit and spirometric volume are unlikely to have a torso injury. Patients with increased base deficit or lactate, decreased spirometric volume, decreased PaO2/FiO2, or positive FAST have substantial risk for torso injury. Increased base deficit and/or decreased spirometric volume are highly sensitive for torso injury. Base deficit and spirometric volume values are readily available and increase or decrease the suspicion for torso injury.

Background

The American College of Surgeons Committee on Trauma developed the Advanced Trauma Life Support (ATLS) guidelines for evaluating and managing acutely injured patients.[1] The purpose of the ATLS primary survey is to detect vital function instability and enhance resuscitation. Patients with instability need an aggressive evaluation to detect torso injuries. Patients with vital function stability undergo the ATLS secondary survey, a comprehensive history and head-to-toe examination. An objective of the secondary survey is to detect clinically significant torso injuries not identified during the primary survey.

There are numerous publications describing missed injuries in patients incurring blunt and penetrating trauma [24]. Of particular concern, missed torso injuries have been documented (chest, [24] abdomen, [24] and pelvis[2, 3]). There is no simple method to readily detect all clinically important torso injuries. The primary problem with identifying torso injuries is related to the diversity of organ and skeletal injuries and the variability of clinical manifestations[1, 5]. Patients with a torso injury may have clinical findings that are suggestive, non-specific, or occult. A tertiary survey has been recommended to complement ATLS guidelines so that missed injuries will be decreased[4]

Depending on the clinical findings, there are numerous diagnostic recommendations for detecting chest, abdominal, and pelvic injuries [1, 5]. The large number of diagnostic procedures for identification of torso injuries underscores the complexity associated with their detection. Chest computed tomography (CT) scan is recommended for select trauma patients following blunt[6, 7] or penetrating [8, 9] mechanisms. Several other diagnostic procedures, e.g., aortography, echocardiography, electrocardiography, and bronchoscopy, may be needed to identify a variety of chest injuries[1, 5]. Following blunt trauma, there are variable recommendations for detecting abdominal injuries: diagnostic peritoneal lavage (DPL), [10] CT scan, [10] ultrasonography, [11] laparoscopy, [12] and celiotomy[13] There are divergent recommendations for detecting abdominal injuries following a stab wound to the abdomen: observation, [14] DPL, [15] CT scan, [16] laparoscopy, [17] and celiotomy[14]. Following a gunshot wound to the abdomen, there are variable recommendations for detecting abdominal injuries: observation, [18] DPL, [19] CT scan, [20] ultrasonography, [21] laparoscopy, [22] and celiotomy[23]. Additionally, there are variable recommendations for detecting pelvic ring disruption: antero-posterior, inlet, outlet, and Judet x-rays, and CT scan [2427]. The multiplicity of diagnostic recommendations to evaluate the chest, abdomen, and pelvis underlie the notion that early torso injury identification is difficult.

There is a need to develop an objective and simple complement to the ATLS secondary survey to indicate the probability for torso injury in stable patients. An assessment of metabolic acidosis status, pulmonary physiology, and abdominal ultrasound findings may be useful to determine the presence of a torso injury. Severity of injury has been shown to correlate with base deficit[28] and serum lactate[29]. Abdominal ultrasound has evolved as a method for initial screening for the detection of abdominal injuries; however, the sensitivity of this procedure is variable[11]. A reduction in PaO2/FiO2 has been described in patients with pulmonary contusion[30] and acute chest trauma, [31] while a decrease in spirometric volume has been found in patients with rib fractures[32] or operative torso trauma [33, 34]. The purpose of this study was to determine the relationship of torso injury with base deficit, lactate, presence of fluid on abdominal ultrasound, PaO2/FiO2, and spirometric volume. This investigation was performed at the St. Elizabeth Health Center in Youngstown, OH, a Level I trauma center.

Methods

The Institutional Review Board for Human Investigations approved the study.

Inclusion and exclusion criteria

Inclusion criteria were blunt or penetrating trauma patients evaluated by the trauma service (emergency medical service trauma team alert or emergency department physician consultation). Exclusion criteria were patients who: failed to have spirometric volume, lactate, focused abdominal sonography for trauma (FAST), or arterial blood gas (ABG) within 4 hours of injury; were under 18 years or over 65 years; fell from a standing height; had persistent hemodynamic instability; smoked two or more packs of cigarettes per day; have COPD and require bronchodilator or home oxygen therapy; needed urgent tracheal intubation; or were unable to understand the proper technique of incentive spirometry. Routine evaluation for patients evaluated by the trauma service include spirometric volume, lactate, FAST exam, and ABG evaluation.

Clinically significant torso injuries

Clinically significant torso injuries included: chest wall or abdominal contusion (moderate to severe pain and tenderness with impaired chest wall excursion and decreased cough sound intensity); sternal fracture; three or more rib fractures; splenic injury; liver injury; gastrointestinal tract injury; pelvic ring disruption; thoracic esophageal injury; pneumothorax; hemothorax; lung contusion; diaphragmatic injury; great vessel injury; cardiac contusion or tamponade; pancreatic injury; renal injury; ureter or bladder injury; hip fracture or dislocation; thoracic spinal injury; lumbar spinal injury; and abdominal vascular injury.

Patient assessment and identification of torso injuries

An initial pilot study was performed. Prospective documentation was accomplished by completing data collection forms at the time of patient evaluation. The chief resident, under the supervision of the emergency department or trauma surgical attending, performed a standard FAST exam. The FAST exam was used to document the presence of fluid in the pericardial space, right upper quadrant, left upper quadrant, and pelvis. Base deficit and PaO2/FiO2 were routinely obtained from the ABG. Lactate was obtained from a venous blood specimen. Incentive spirometry was taught to the patient by a respiratory therapist. The patient had three practice spirometry attempts. The patient was instructed to take in a deep breath, maximally expire, and then perform a maximal inspiratory effort. The goal was to generate the highest volume possible.

A torso physical examination was performed by the chief surgical resident who documented the presence of torso tenderness (none, mild, moderate, or severe), ecchymosis, lacerations, and abrasions. Ethanol status was positive when the toxicology screen revealed a positive ethanol level and was negative when the level was not detected. When there was no toxicology screen, the ethanol status was positive when there was a history of ethanol consumption or the physical examination indicated that the patient was intoxicated or smelled of alcohol. Otherwise, the ethanol status was negative.

There was a routine assessment of patients discharged within 72 hours of injury. Patients were discharged only when their vital signs, chest wall excursion, and cognition were stable. Patients were considered stable when there was normotension, heart rate ≤ 100 beats per minute, respiratory rate ≤ 24 breaths per minute, incentive spirometric volume ≥ 30 mL/kg weight, and normal cognition. The patient's weight was recorded during hospitalization. Routine clinic visit and/or home telephone call was performed within 3 to 7 days for patients discharged within 72 hours of injury.

Diagnostic imaging included routine chest x-ray and pelvic x-rays. Chest CT and abdominal/pelvic CT scans were performed at the discretion of the attending trauma surgeon.

There were 88 patients in the prospective observational pilot study. There was one missed torso injury found during the routine follow-up. Total patients with torso injuries in the pilot study were 26 (29.5%).

A post-pilot study was conducted that had the same study design as the pilot study. However, there was no obligatory clinic visit or telephone follow-up for patients discharged within 72 hours of injury. The patients seen in the trauma clinic were assessed for missed torso injury. The complete study included data from the pilot and post-pilot studies. This data were combined for statistical assessment.

Demographics

Mechanism of injury was studied for both blunt (motor vehicular crash, motorcycle crash, fall, assault, other) and penetrating (gunshot wound, stab wound) injuries. Age and weight were documented as well as Glasgow Coma Score and Injury Severity Score (ISS). The non-torso injury with the highest Abbreviated Injury Scale (AIS) score and its body region were documented.

Torso injury relationships with metabolic acidosis status and pulmonary physiology

The base deficit for patients with and without torso injury was determined. A previous study by the first author showed a univariate relationship between base deficit and ethanol status and between base deficit and ISS[28]. Multivariate regression analysis also showed that base deficit was independently associated with ethanol status and ISS. The relationship between base deficit and ethanol status and ISS was determined in the current study. The levels selected for increased base deficit were the base deficit values where the odds ratio and sensitivity for torso injury were the greatest for ethanol-negative and ethanol-positive patients.

The lactate for patients with and without torso injury was determined. The relationship between lactate and ethanol status and ISS was determined. The levels selected for increased lactate were the lactate values where the odds ratio and sensitivity for torso injury were the greatest for ethanol-negative and ethanol-positive patients.

The PaO2/FiO2 for patients with and without torso injury was determined. The level selected for decreased PaO2/FiO2 was the PaO2/FiO2 value where the odds ratio and sensitivity for torso injury were the greatest. The spirometric volume for patients with and without torso injury was determined. The level selected for decreased spirometric volume was the volume where the odds ratio and sensitivity for torso injury were the greatest.

FAST and torso injury relationship

The right upper quadrant (RUQ), left upper quadrant (LUQ), and pelvis were categorized relative to the presence of fluid: present, absent, or poor visualization. A positive abdominal FAST was the presence of fluid in the RUQ, LUQ, or pelvis. The pericardial space was categorized relative to the presence of fluid: present, absent, or poor visualization. A positive pericardial FAST was the presence of fluid in the pericardial space. The relationship between a positive pericardial FAST and chest injury was determined. The relationship between a positive abdominal FAST and abdominal injury was determined. The relationship between fluid-positive thoracic or abdominal/pelvic ultrasound and torso injury was determined.

Truncal soft tissue injury and torso injury relationship

Truncal soft tissue injury was the presence of chest or abdominal wall abrasion, ecchymosis, or laceration. The relationship between truncal soft tissue injury and torso injury was determined.

Statistical analysis

Statistical analysis was performed on the SAS System for Windows, release 6.11, SAS Institute Inc. (Cary, NC). A Fisher's exact test was used to compare the torso injury rate with other dichotomous variables. Correlation coefficient analysis was utilized to assess the relationship between two continuous variables. A T-test was used to compare the mean value of two groups. Multiple regression analysis was utilized to assess the influence of ethanol status and base deficit on base deficit and lactate. Multivariate logistic regression analysis was used to determine the effect of two more independent variables on the presence or absence of torso injury. A P value < 0.05 was considered statistically significant.

Results

Study patients

There were 88 trauma patients who met the criteria for inclusion in the pilot study and 26 (29.5%) had a torso injury. Sixty-three patients (71.6%) were discharged within 72 hours of emergency department presentation. One of these patients had a torso injury undiagnosed during hospitalization (rib fractures), but was found during the routine follow-up. The post-pilot study group included another 127 trauma patients who met the criteria for study inclusion.

Of the 215 patients in the total study group, 66 patients (30.7%) had a torso injury (abdominal/pelvic and/or thoracic injury). Of the 66 patients with torso injury, 35 had abdominal or pelvic injury and 43 had thoracic injury. Specific abdominal/pelvic and thoracic injuries are depicted in Table 1. Of the total study group, 135 patients (62.8%) were discharged within 72 hours. All 215 patients had a chest x-ray taken. A pelvic x-ray was taken in 177 patients (82.3%). An abdominal CT scan was performed in 59 patients (27.4%) and a chest CT scan was done in 38 patients (17.7%).
Table 1

Torso injuries in study patients

Abdominal injuries:

#

Thoracic injuries:

#

gastrointestinal

1

cardiac

2

kidney

2

hemothorax

7

liver

8

lung contusion

17

spleen

6

pneumothorax

22

urinary bladder

1

rib fractures

23

lumbar spine

3

sternal fracture

2

pelvic ring

10

thoracic spine

1

hip fracture/dislocation

8

severe chest contusion

3

Patients with abdominal injury: 35; Patients with thoracic injury: 43; Patients with torso injury: 66.

Demographics

Blunt trauma was present in 194 patients (90.2%) (motor vehicle crash, motorcycle crash, fall, assault, or crush injury). Penetrating injury was found in 21 patients (9.8%) (gunshot or stab wound). The torso injury rate was similar for the blunt trauma (30.4%) and the penetrating trauma patients (33.3%; P = 0.78). There were no significant differences (P ≥ 0.05) between the patients with torso injury and without torso injury relative to age (36.7 and 32.9 years), weight (172 and 176 pounds), and Glasgow Coma Scale score (14.8 and 14.8). The ISS was higher in the patients with torso injury (16.2 ± 9.2) when compared to the patients without torso injury (5.6 ± 4.3, P = 0.0001). The hospital length of stay for patients with a torso injury was 5.3 days (62.8% had length of stay > 72 hours). Body regions with the highest non-torso injury AIS score were head – 80 (37.2%), neck – 18 (8.4%), upper extremity – 38 (17.7%), lower extremity – 60 (27.9%), and none – 19 (8.8%). The distribution of the highest non-torso injury AIS score was 1 – 59 (27.4%), 2 – 80 (37.2%), 3 – 53 (24.7%), 4 – 4 (1.9%), and 0 – 19 (8.8%). Ethanol status was positive in 26.5% (57/215) and negative in 73.5% (158/215). An ethanol level was measured in 80.5% (173/215) of the patients. Of the 42 without an ethanol level, four (9.5%) had clinical evidence of ethanol consumption and were considered to be ethanol-positive.

Torso injury relationships with metabolic acidosis status and pulmonary physiology

The base deficit was greater in the patients with torso injury (-2.6 ± 3.0) when compared to the patients without torso injury (-1.0 ± 2.0, P = 0.0001). The base deficit was inversely related to the ISS (r = -0.44, P = 0.0001). The base deficit was greater in the ethanol-positive patients (-3.4 ± 2.2) when compared to the ethanol-negative patients (-0.8 ± 2.2, P = 0.0001). Multivariate regression analysis showed that base deficit was independently associated with ethanol status and ISS (r = 0.61, P = 0.0001). The torso injury rate was similar in the ethanol-positive patients (27.3%) and the ethanol-negative patients (26.2%, P = 0.87). Base deficit was greater for patients with highest non-torso injury AIS score 3–4 (-2.6 ± 2.9) when compared to those with highest non-torso injury AIS score 0–2 (-1.1 ± 2.2; P = 0.0005). Base deficit was independently associated with torso injury, ethanol status, and the highest non-torso injury AIS score (r = 0.61; P = 0.0001). The threshold levels selected for increased base deficit were the base deficit values where the odds ratio and sensitivity for torso injury were the greatest for ethanol-negative and ethanol-positive patients. The threshold base deficit values were < 0.0 for ethanol-negative patients and ≤ -3.0 for ethanol-positive patients. The torso injury positive predictive value, sensitivity, specificity, negative predictive value, and risk of increased base deficit are displayed in Tables 2 and 3. Increased base deficit was greater with highest non-torso injury AIS score 3–4 (68.4% [39/57]) when compared to AIS score 0–2 (51.9% [82/158]; OR 2.0, P = 0.03). Increased base deficit was independently associated with torso injury and the highest non-torso injury AIS score (P = 0.006). Torso injury was not associated with the highest non-torso injury AIS score (P = 0.22).
Table 2

Risk assessment for torso injuries in 215 stable trauma patients

Risk Factor

#

TI Rate

Sensitivity

Specificity

NPV

OR

P-value

increased BD

121

41.3%

75.8%

52.3%

83.0%

3.4

.0001

increased lactate

78

39.7%

47.0%

68.5%

74.5%

1.9

.03

decreased PaO2/FiO2

45

48.9%

33.3%

84.6%

74.1%

2.7

.003

decreased SV

64

73.4%

71.2%

88.6%

87.4%

19.2

<<.0001

positive FAST

8

62.5%

7.6%

98.0%

70.5%

4.0

.06

chest/abdominal STI

82

37.8%

47.0%

65.8%

73.7%

1.7

.08

increased BD and/or decreased SV

144

43.8%

95.5%

45.6%

95.8%

17.6

<<.0001

TI, torso injury; NPV, negative predictive value for a negative test; OR, odds ratio; BD, base deficit; SV, spirometric volume; FAST, focused-abdominal sonography for trauma; STI, soft tissue injury (abrasions, ecchymosis, laceration)

Table 3

Rates of torso injuries in 215 stable trauma patients

Risk Factor

Number

TI Rate

95% CI

normal BD and normal SV

71

4.2%

0.0–8.9%

increased BD and normal SV

80

20.0%

12.7–30.0%

normal BD and decreased SV

23

56.5%

36.8–74.4%

increased BD and decreased SV

41

82.9%

71.4–94.4%

TI, torso injury; CI, confidence intervals; BD, base deficit; SV, spirometric volume

Lactate was greater in the patients with torso injury (2.9 ± 2.2 mmol/L) when compared to the patients without torso injury (2.3 ± 1.1 mmol/L, P = 0.04). Lactate was directly related to ISS (r = 0.26, P = 0.0001). The lactate was greater in the ethanol-positive patients (3.1 ± 1.2 mmol/L) when compared to the ethanol-negative patients (2.3 ± 1.6 mmol/L, P = 0.0001). Multivariate regression analysis showed that lactate was independently associated with ethanol status and ISS (r = 0.33, P = 0.001). The threshold levels selected for increased lactate were the lactate values where the odds ratio and sensitivity for torso injury were the greatest for ethanol-negative and ethanol-positive patients. The threshold lactate values were > 2.5 mmol/L for ethanol-negative patients and ≥ 3.0 mmol/L for ethanol-positive patients. The positive predictive value, sensitivity, specificity, negative predictive value, and risk ratio of increased lactate for torso injury are displayed in Table 2.

The PaO2/FiO2 was lower in the patients with torso injury (394 ± 139) when compared to the patients without torso injury (445 ± 102, P = 0.01). The threshold level selected for decreased PaO2/FiO2 was the PaO2/FiO2 value where the odds ratio and sensitivity for torso injury were the greatest. The threshold PaO2/FiO2 value was < 350. The positive predictive value, sensitivity, specificity, negative predictive value, and risk ratio of decreased PaO2/FiO2 for torso injury are displayed in Table 2.

The spirometric volume was lower in the patients with torso injury (1,543 ± 620 mL) when compared to the patients without torso injury (2,287 ± 344, P = 0.0001). The threshold level selected for decreased spirometric volume was the volume where the odds ratio and sensitivity for torso injury were the greatest. The threshold spirometric volume was < 1,800 mL (< 25 mL/kg). The positive predictive value, sensitivity, specificity, negative predictive value, and risk ratio of decreased spirometric volume for torso injury are displayed in Table 2.

FAST and torso injury relationships

A positive abdominal/pelvic ultrasound result was found in seven patients (3.3%). All patients with fluid-positive FAST had an abdominal/pelvic CT scan. The positive predictive value for abdominal/pelvic injury was 57.1% and the sensitivity was 11.4%. The risk ratio for abdominal/pelvic injury with a positive ultrasound result was 7.6 (P = 0.01). A positive pericardial ultrasound result was present in 1 patient (0.5%). The positive predictive value for thoracic injury was 100.0% and the sensitivity was 2.3%. The risk ratio for thoracic injury with a positive ultrasound result was 4.1 (P = 0.2). The positive predictive value, sensitivity, specificity, negative predictive value, and risk ratio of a positive thoracic/abdominal/pelvic ultrasound result for torso injury are displayed in Table 2.

Truncal soft tissue injury and torso injury relationship

The positive predictive value, sensitivity, specificity, negative predictive value, and risk ratio of chest or abdominal soft tissue injury for torso injury are displayed in Table 2.

Torso injury relationships with decreased spirometric volume or increased base deficit

Multivariate logistic regression analysis indicated that torso injury was independently and inversely related to base deficit and spirometric volume (P = 0.0001). Regression analysis also showed that torso injury was independently related to increased base deficit status and decreased spirometric volume status (P = 0.004). Increased lactate, decreased PaO2/FiO2, and positive FAST each had an insignificant relationship with torso injury (P >> 0.05) when increased base deficit and decreased spirometric volume status were included in the logistic regression analysis. Base deficit was increased and/or spirometric volume was decreased in 144 patients (67.0%) (see Tables 2 and 3). The positive predictive value of increased base deficit and/or decreased spirometric volume for torso injury was 43.8% (95% CI, 35.7–51.9%) and the sensitivity for torso injury was 95.5% (95% CI, 92.1–98.9%). The risk ratio for torso injury was 17.6 (P << 0.0001). A normal base deficit and spirometric volume was found in 71 patients (33.0%) (see Table 3). Three of these patients had a torso injury (negative predictive value 95.8% [95% CI, 91.1–100.0%]). The sensitivity of increased base deficit and/or decreased spirometric volume for abdominal/pelvic injury was 94.3%; the sensitivity for thoracic injury was 97.7%.

The torso injury rates for the four combinations of base deficit and spirometric volume are described in Table 3 (Chi-square 87.5; P < 0.0001). The rates for the four combinations are significantly different (P < 0.05) from each other. Of the 46 patients with a negative FAST and a normal base deficit, lactate, PaO2/FiO2, and spirometric volume, there were no torso injuries identified (positive predictive value 95% CI, 0.0–3.5%).

The three patients with a torso injury and a normal base deficit and normal spirometric volume also had a normal FAST exam. The lactate was increased in two patients, and the PaO2/FiO2 was decreased in the third patient. One patient had an isolated sternal fracture and was discharged within 3 days with a spirometric volume of 2,000 mL at discharge. The only torso injury in the other two patients was a hip fracture.

Discussion

Spirometric volume and base deficit provide a probability for the presence of torso injury in stable patients. The literature is miniscule relative to chest wall mechanical assessment in acutely injured patients. Patients with increased base deficit or lactate, decreased spirometric volume, impaired PaO2/FiO2, or a positive FAST were found to have substantial risk for torso injury. Trauma patients with normal base deficit and spirometric volume are unlikely to have a torso injury. Conversely, increased base deficit and/or decreased spirometric volume indicate that there is a markedly enhanced risk for torso injury. An increased base deficit and/or decreased spirometric volume were highly sensitive for detecting patients with torso injury.

The trauma surgeon is frequently confronted with a plethora of injuries in a diverse cohort of patients. However, our aim was to evaluate the typical 30–40 year old stable, trauma patient with a significant mechanism of injury and determine the risk for torso injury using readily available, objective tests. We excluded extremes of age to optimize cohort homogeneity. Also, patients with severe pre-existing pulmonary conditions were excluded because we were concerned that it would alter the relationship between spirometric volume and torso injury. Additionally, these patients commonly have a metabolic alkalosis, which may modify the association of base deficit with torso injury. Unconscious patients would not be able to perform a spirometric volume. Patients undergoing immediate tracheal intubation or with persistent hemodynamic instability typically need a comprehensive diagnostic evaluation for torso injury. Patients falling from a standing height have variable risk for serious injury, are often elderly, and frequently have pre-existing medical conditions.

The association between base deficit and torso injury was highly significant. Base deficit was independently associated with ethanol status, torso injury, and highest non-torso injury AIS score. Patients with an increased base deficit had a substantial increase in torso injury risk. Three-quarters of patients with torso injury had an increased base deficit.

The literature indicates that there are several relationships that have been previously established between injury severity and base deficit. Base deficit has been found to be associated with trauma patient mortality [35, 36] and ISS[28, 35, 37]. Base deficit has also been associated with the trauma score[35, 38]. Additionally, base deficit has been linked to hypotension and resuscitation [28, 36] and the presence of an abdominal injury[28, 39]. Similar to the current study, base deficit level has been found in other investigations to be increased in patients with ethanol [28, 37, 40].

Similar to base deficit, there was an association between lactate and torso injury. Patients with torso injury had a greater lactate when compared to those without torso injury. Lactate was directly related to ISS and was greater in ethanol-positive patients when compared to ethanol-negative patients. Lactate was also found to be independently associated with ethanol status and ISS. The patients with increased lactate had a substantial risk for torso injury when compared to those with normal lactate. Several other studies provide evidence that lactate is associated with injury severity. Lactate has been shown to have a positive relationship with mortality[29, 41] and injury severity[29, 37, 41]. Additionally, lactate has been associated with the need for resuscitation[38]. Similar to base deficit, others have found lactate to be increased in patients who have consumed ethanol[37].

Patients with torso injury had a lower PaO2/FiO2 than those without torso injury. When the PaO2/FiO2 was < 350, the torso injury risk was substantially increased. One-third of the patients with torso injury had a decreased PaO2/FiO2. Other investigators have noted that trauma patients with a decreased PaO2/FiO2 had increased mortality [30, 31] and a need for prolonged mechanical ventilation[30, 42]. Other investigators have also noted a reduction in PaO2/FiO2 in patients with pulmonary contusion [30, 43, 44] and acute chest trauma[31, 45].

Torso injury patients had a lower spirometric volume than those with no torso injury. A spirometric volume < 1,800 mL (< 25 mL/kg) was associated with increased torso injury risk. Three-quarters of the patients with a decreased spirometric volume had a torso injury and three-quarters of the patients with torso injury had a decreased spirometric volume. The only literature identified regarding spirometric volume assessment in trauma patients described patients with rib fractures[32]. This study showed a decrease in spirometric volume prior to the institution of a regimen to mitigate chest wall pain. Rib fractures are associated with torso pain and are likely to cause a decrement in chest wall mechanics [32, 4648].

A reduction in vital capacity following operative torso trauma has been documented in other studies. Upper abdominal surgery has been associated with a postoperative vital capacity 45–55% that of the preoperative value[33, 49, 50]. Postoperative vital capacity has been found to have a greater reduction with open cholecystectomy (48%) when compared to laparoscopic cholecystectomy (26%)[34]. Other studies have shown that stable patients undergoing median sternotomy had a marked reduction in postoperative vital capacity[51]. These studies suggest that significant trauma to the chest or abdomen may cause a clinically significant reduction in spirometric volume.

The patients with a positive FAST had a substantial torso injury risk. However, the 95% confidence band was quite large due to the small number of patients. Less than 10% of the patients with torso injury had a positive FAST. The latter is due to the fact that FAST is not likely to detect most retroperitoneal, pelvic, or chest injuries[11, 52]. The variable sensitivity of FAST for detecting abdominal injuries is a concern. Amoroso found in 13 studies that the abdominal injury sensitivity for FAST ranged from 81–99%[52]. In six of these studies, the abdominal injury sensitivity was ≤ 90%. Pearl has also described variation in FAST sensitivity depending on the clinical endpoint[11]. An examination of four studies revealed a sensitivity of 87–98% for detecting intraperitoneal fluid. The sensitivity for identifying organ injury was 69–96% in six studies. A review of four studies showed that the sensitivity for therapeutic laparotomy ranged from 84–93%. The current study and the literature suggest that FAST has substantial limitations for identifying torso injuries.

Patients with chest or abdominal abrasions, ecchymosis, or lacerations had a substantial torso injury risk. However, the odds ratio was insignificant. Only one-half of the patients with torso injury had a truncal soft tissue injury. The relationship between spirometric volume and torso injury was superior. Others have shown an important association between intrathoracic injuries and chest ecchymosis [53, 54] and abrasions[55]. Similarly, intra-abdominal injuries have been linked to abdominal ecchymosis, [54, 5658] abrasions, [55, 58] and lacerations[59]. Data from the current study suggest that spirometric volume assessment may be more reliable for detecting torso injuries. Others have shown that physical examination has a modest sensitivity for detecting abdominal injuries (60–87%) [6062]. The above literature suggests that an objective appraisal for torso injury may be needed to complement the physical exam.

Torso soft tissue injury can be considered a clinical standard for which to compare the other torso injury risk factors. Sensitivity of increased base deficit and sensitivity of decreased spirometric volume for torso injury was greater than the other risk factors. Sensitivity for increased base deficit and/or decreased spirometric volume for torso injury approached 100%. Decreased spirometric volume had a superior specificity when compared to torso soft tissue injury. The specificity of increased base deficit and the specificity of increased base deficit and/or decreased spirometric volume were inferior to torso soft tissue injury. In other words, the false-positive rates were increased. The decreased specificity of increased base deficit and the decreased specificity of increased base deficit and/or decreased spirometric volume are, in part, explained by the association between base deficit and the highest non-torso injury AIS score value. Specifically, increased base deficit is present with either torso injury or complex non-torso injuries. The negative predictive value for normal base deficit and the negative predictive value for normal spirometric volume were greater than the rate for no torso soft tissue injury. The negative predictive value for normal base deficit and normal spirometric volume was nearly 100%. Torso soft tissue injury was not a significant risk for torso injury, however, increased base deficit was. Decreased spirometric volume and increased base deficit and/or decreased spirometric volume performed well as risk factors for torso injury. Decreased spirometric volume had the highest positive predictive value for torso injury.

There were no torso injuries in the patients with a negative FAST and a normal base deficit, lactate, PaO2/FiO2, and spirometric volume. When base deficit and spirometric volume were normal, the probability for torso injury was low. When base deficit was increased and spirometric volume was normal, there was a 20% torso injury rate. The relatively low torso injury rate is likely due to the influence of complex non-torso injuries on base deficit. When spirometric volume is normal and base deficit is increased, the clinician should evaluate the patient for evidence of torso and non-torso injuries and determine the need for a torso CT scan. When spirometric volume is decreased, a comprehensive evaluation of the torso is indicated. Because the clinical objective is to minimize risk for missing potentially life threatening injuries, the high sensitivity and relatively low specificity of increased base deficit and/or decreased spirometric volume are reasonable.

There were study weaknesses that need to be addressed in future investigations. Patients falling from a standing height need to be studied in the future to determine the potential relevance of base deficit and spirometric volume in assessing risk for torso injury. The study focus was to detect clinically significant torso injuries and describe risk relationships using readily available, objective tests. The standard for determining the presence or absence of a torso injury was a routine chest x-ray on admission, select chest or abdominal CT scans on admission, and daily evaluation with additional appropriate diagnostic studies until hospital discharge. A torso injury was determined to be present or absent at hospital discharge. During the pilot-study, patients discharged within 72 hours of injury were also evaluated by post-discharge telephone call and/or clinic visit. Eighty-eight patients were in the pilot-study and 26 torso injuries were detected (29.5%). One injury was detected after discharge. If this injury had not been detected or two more had been identified the torso injury rate would have been 28.4–31.8%. The 30.7% torso injury rate for the 215 patients was strikingly similar to that in the pilot-study. Routine chest, abdominal, and pelvic CT scans may have been more elucidating. However, CT scans may divulge clinically insignificant injuries. Our study should provide an impetus to perform a future investigation with routine CT scans, as well as a description of the clinical import of each identified torso injury. Such a study would help to better define the rate of clinically significant torso injuries and the ability to assess risk during the patient's initial evaluation. A cost analysis comparing routine CT scans with potential discharge home to the cost and ability of arterial blood gas analysis, spirometric volume, lactate, and FAST to indicate the need for admission or discharge or torso CT scans is important and needs to be studied. Future studies should routinely obtain an ethanol test.

Conclusions

This is the only study to describe the use of bedside spirometric volume assessment in a wide spectrum of stable trauma patients during the early postinjury period. Patients with decreased spirometric volume have a high torso injury rate and the majority of patients with a torso injury have a decreased spirometric volume. Spirometric volume and base deficit are two simple objective measures that indicate the probability of torso injury in stable patients. Trauma patients with normal base deficit and spirometric volume are unlikely to have a torso injury. An increased base deficit and/or a decreased spirometric volume are highly sensitive for torso injury. An increased base deficit is associated with torso and complex non-torso injuries. Base deficit and spirometric volume may be useful complements to the ATLS secondary survey to risk stratify the likelihood for torso injury. The physician may find that the objective and clinical information can be used to determine the need for a more comprehensive evaluation of the trauma patient.

Abbreviations

ABG: 

arterial blood gas

ATLS: 

Advanced Trauma Life Support

BD: 

base deficit

CT: 

computed tomography

DPL: 

diagnostic peritoneal lavage

FAST: 

focused abdominal sonography for trauma

ISS: 

Injury Severity Score

LUQ: 

left upper quadrant

RUQ: 

right upper quadrant

SV: 

spirometric volume

Declarations

Authors’ Affiliations

(1)
Department of Surgery, St. Elizabeth Health Center
(2)
Department of Surgery, Ferguson Clinic

References

  1. American College of Surgeons Committee on Trauma: Advanced Trauma Life Support, Instructor Manual. 1997, Chicago: American College of Surgeons, SixthGoogle Scholar
  2. Buduhan G, McRitchie DI: Missed injuries in patients with multiple trauma. J Trauma. 2000, 49: 600-605.View ArticlePubMedGoogle Scholar
  3. Robertson R, Mattox R, Collins T, Parks-Miller C, Eidt J, Cone J: Missed injuries in a rural area trauma center. Am J Surg. 1996, 172: 564-567. 10.1016/S0002-9610(96)00247-4.View ArticlePubMedGoogle Scholar
  4. Enderson BL, Reath DB, Meadors J, Dallas W, DeBoo JM, Maull KI: The tertiary trauma survey: a prospective study of missed injury. J Trauma. 1990, 30: 666-669.View ArticlePubMedGoogle Scholar
  5. Mattox KL, Feliciano DV, Moore EE: Trauma. 2000, New York: McGraw-Hill, FourthGoogle Scholar
  6. Zinck SE, Primack SL: Radiographic and CT findings in blunt chest trauma. J Thorac Imaging. 2000, 15: 87-96. 10.1097/00005382-200004000-00003.View ArticlePubMedGoogle Scholar
  7. Kuhlman JE, Pozniak MA, Collins J, Knisely BL: Radiographic and CT findings of blunt chest trauma: aortic injuries and looking beyond them. Radiographics. 1998, 18: 1085-1106.View ArticlePubMedGoogle Scholar
  8. LeBlang SD, Dolich MO: Imaging of penetrating thoracic trauma. J Thorac Imaging. 2000, 15: 128-135. 10.1097/00005382-200004000-00008.View ArticlePubMedGoogle Scholar
  9. Nagy KK, Gilkey SH, Roberts RR, Fildes JJ, Barrett J: Computed tomography screens stable patients at risk for penetrating cardiac injury. Acad Emerg Med. 1996, 3: 1024-1027.View ArticlePubMedGoogle Scholar
  10. Catre MG: Diagnostic peritoneal lavage versus abdominal computed tomography in blunt abdominal trauma: a review of prospective studies. Can J Surg. 1995, 38: 117-122.PubMedGoogle Scholar
  11. Pearl WS, Todd KH: Ultrasonography for the initial evaluation of blunt abdominal trauma: A review of prospective trials. Ann Emerg Med. 1996, 27: 353-361.View ArticlePubMedGoogle Scholar
  12. Taner AS, Topgul K, Kucukel F, Demir A, Sari S: Diagnostic laparoscopy decreases the rate of unnecessary laparotomies and reduces hospital costs in trauma patients. J Laparoendosc Adv Surg Tech A. 2001, 11: 207-211. 10.1089/109264201750539718.View ArticlePubMedGoogle Scholar
  13. Fakhry SM, Brownstein M, Watts DD, Baker CC, Oller D: Relatively short diagnostic delays (<8 hours) produce morbidity and mortality in blunt small bowel injury: an analysis of time to operative intervention in 198 patients from a multicenter experience. J Trauma. 2000, 48: 408-414.View ArticlePubMedGoogle Scholar
  14. Demetriades D, Rabinowitz B: Indications for operation in abdominal stab wounds. A prospective study of 651 patients. Ann Surg. 1987, 205: 129-132.View ArticlePubMedPubMed CentralGoogle Scholar
  15. Boyle EM, Maier RV, Salazar JD, Kovacich JC, O'Keefe G, Mann FA: Diagnosis of injuries after stab wounds to the back and flank. J Trauma. 1997, 42: 260-265.View ArticlePubMedGoogle Scholar
  16. Albrecht RM, Vigil A, Schermer CR, Demarest GB, Davis VH, Fry DE: Stab wounds to the back/flank in hemodynamically stable patients: evaluation using triple-contrast computed tomography. Am Surg. 1999, 65: 683-687.PubMedGoogle Scholar
  17. Ivatury RR, Simon RJ, Stahl WM: A critical evaluation of laparoscopy in penetrating abdominal trauma. J Trauma. 1993, 34: 822-827.View ArticlePubMedGoogle Scholar
  18. Velmahos GC, Demetriades D, Foianini E, Tatevossian R, Cornwell EE, Asensio J: A selective approach to the management of gunshot wounds to the back. Am J Surg. 1997, 174: 342-346. 10.1016/S0002-9610(97)00098-6.View ArticlePubMedGoogle Scholar
  19. Nagy KK, Krosner SM, Joseph KT, Roberts RR, Smith RF, Barrett J: A method of determining peritoneal penetration in gunshot wounds to the abdomen. J Trauma. 1997, 43: 242-245.View ArticlePubMedGoogle Scholar
  20. Ginzburg E, Carrillo EH, Kopelman T, McKenney MG, Kirton OC, Shatz DV: The role of computed tomography in selective management of gunshot wounds to the abdomen and flank. J Trauma. 1998, 45: 1005-1009.View ArticlePubMedGoogle Scholar
  21. Boulanger BR, Kearney PA, Tsuei B, Ochoa JB: The routine use of sonography in penetrating torso injury is beneficial. J Trauma. 2001, 51: 320-325.View ArticlePubMedGoogle Scholar
  22. Fabian TC, Croce MA, Stewart RM, Pritchard FE, Minard G, Kudsk KA: A prospective analysis of diagnostic laparoscopy in trauma. Ann Surg. 1993, 217: 557-564.View ArticlePubMedPubMed CentralGoogle Scholar
  23. Salim A, Velmahos GC: When to operate on abdominal gunshot wounds. Scand J Surg. 2002, 91: 62-66.PubMedGoogle Scholar
  24. Mostafavi HR, Tornetta P: Radiologic evaluation of the pelvis. Clin Orthop. 1996, 329: 6-14. 10.1097/00003086-199608000-00003.View ArticlePubMedGoogle Scholar
  25. Duane TM, Cole FJ, Weireter LJ, Britt LD: Blunt trauma and the role of routine pelvic radiographs. Am Surg. 2001, 67: 849-852.PubMedGoogle Scholar
  26. Noh HM, Scott WW, Fishman EK: Imaging of pelvic trauma: the role of CT with multiplanar and three-dimensional reconstruction. J South Orthop Assoc. 1996, 5: 111-125.PubMedGoogle Scholar
  27. Resnik CS, Stackhouse DJ, Shanmuganathan K, Young JW: Diagnosis of pelvic fractures in patients with acute pelvic trauma: efficacy of plain radiographs. AJR Am J Roentgenol. 1992, 158: 109-112.View ArticlePubMedGoogle Scholar
  28. Dunham CM, Watson LA, Cooper C: Base deficit level indicating major injury is increased with ethanol. J Emerg Med. 2000, 18: 165-171. 10.1016/S0736-4679(99)00188-2.View ArticlePubMedGoogle Scholar
  29. Lavery RF, Livingston DH, Tortella BJ, Sambol JT, Slomovitz BM, Siegel JH: The utility of venous lactate to triage injured patients in the trauma center. J Am Coll Surg. 2000, 190: 656-664. 10.1016/S1072-7515(00)00271-4.View ArticlePubMedGoogle Scholar
  30. Tyburski JG, Collinge JD, Wilson RF, Eachempati SR: Pulmonary contusions: quantifying the lesions on chest X-ray films and the factors affecting prognosis. J Trauma. 1999, 46: 833-838.View ArticlePubMedGoogle Scholar
  31. Croce MA, Fabian TC, Davis KA, Gavin TJ: Early and late acute respiratory distress syndrome: two distinct clinical entities. J Trauma. 1999, 46: 361-366.View ArticlePubMedGoogle Scholar
  32. Haenel JB, Moore FA, Moore EE, Sauaia A, Read RA, Burch JM: Extrapleural bupivacaine for amelioration of multiple rib fracture pain. J Trauma. 1995, 38: 22-27.View ArticlePubMedGoogle Scholar
  33. Stock MC, Downs JB, Gauer PK, Alster JM, Imrey PB: Prevention of postoperative pulmonary complications with CPAP, incentive spirometry, and conservative therapy. Chest. 1985, 87: 151-157.View ArticlePubMedGoogle Scholar
  34. Frazee RC, Roberts JW, Okeson GC, Symmonds RE, Snyder SK, Hendricks JC: Open versus laparoscopic cholecystectomy. A comparison of postoperative pulmonary function. Ann Surg. 1991, 213: 651-653.View ArticlePubMedPubMed CentralGoogle Scholar
  35. Kincaid EH, Chang MC, Letton RW, Chen JG, Meredith JW: Admission base deficit in pediatric trauma: a study using the National Trauma Data Bank. J Trauma. 2001, 51: 332-335.View ArticlePubMedGoogle Scholar
  36. Rixen D, Raum M, Bouillon B, Lefering R, Neugebauer E: Base deficit development and its prognostic significance in posttrauma critical illness: an analysis by the trauma registry of the Deutsche Gesellschaft fur unfallchirurgie. Shock. 2001, 15: 83-89.View ArticlePubMedGoogle Scholar
  37. Dunne J, Napolitano LM, Tracy MA, Scalea TM: Proceedings of the Sixty-First Annual Meeting of the American Association for the Surgery of Trauma, 2002: Lactate and base deficit in trauma: does alcohol impair their predictive accuracy?. www aast org/02 abstracts. 2002Google Scholar
  38. Bannon MP, O'Neill CM, Martin M, Ilstrup DM, Fish NM, Barrett J: Central venous oxygen saturation, arterial base deficit, and lactate concentration in trauma patients. Am Surg. 1995, 61: 738-745.PubMedGoogle Scholar
  39. Davis JW, Mackersie RC, Holbrook TL, Hoyt DB: Base deficit as an indicator of significant abdominal injury. Ann Emerg Med. 1991, 20: 842-844.View ArticlePubMedGoogle Scholar
  40. Davis JW, Kaups KL, Parks SN: Effect of alcohol on the utility of base deficit in trauma. J Trauma. 1997, 43: 507-510.View ArticlePubMedGoogle Scholar
  41. Milzman DP, Boulanger BR, Wiles C, Hinson D: Admission lactate predicts injury severity and outcome in trauma patients. Crit Care Med. 1992, 20 (suppl): S94-Google Scholar
  42. Velmahos GC, Belzberg H, Chan L, Avari S, Cornwell EE, Berne TV: Factors predicting prolonged mechanical ventilation in critically injured patients: introducing a simplified quantitative risk score. Am Surg. 1997, 63: 811-817.PubMedGoogle Scholar
  43. Kollmorgen DR, Murray KA, Sullivan JJ, Mone MC, Barton RG: Predictors of mortality in pulmonary contusion. Am J Surg. 1994, 168: 659-663.View ArticlePubMedGoogle Scholar
  44. Hoff SJ, Shotts SD, Eddy VA, Morris JA: Outcome of isolated pulmonary contusion in blunt trauma patients. Am Surg. 1994, 60: 138-142.PubMedGoogle Scholar
  45. Tanaka H, Yukioka T, Yamaguti Y, Shimizu S, Goto H, Matsuda H: Surgical stabilization of internal pneumatic stabilization? A prospective randomized study of management of severe flail chest patients. J Trauma. 2002, 52: 727-732.View ArticlePubMedGoogle Scholar
  46. Thompson BM, Finger W, Tonsfeldt D, Aprahamian C, Troiano P, Hendley G: Rib radiographs for trauma: useful or wasteful?. Ann Emerg Med. 1986, 15: 261-265.View ArticlePubMedGoogle Scholar
  47. Wu CL, Jani ND, Perkins FM, Barquist E: Thoracic epidural analgesia versus intravenous patient-controlled analgesia for the treatment of rib fracture pain after motor vehicle crash. J Trauma. 1999, 47: 564-567.View ArticlePubMedGoogle Scholar
  48. DeLuca SA, Rhea JT, O'Malley TO: Radiographic evaluation of rib fractures. AJR Am J Roentgenol. 1982, 138: 91-92.View ArticlePubMedGoogle Scholar
  49. Minschaert M, Vincent JL, Ros AM, Kahn RJ: Influence of incentive spirometry on pulmonary volumes after laparotomy. Acta Anaesthesiol Belg. 1982, 33: 203-209.PubMedGoogle Scholar
  50. Dohi S, Gold MI: Comparison of two methods of postoperative respiratory care. Chest. 1978, 73: 592-595.View ArticlePubMedGoogle Scholar
  51. Stock MC, Downs JB, Cooper RB, Lebenson IM, Cleveland J, Weaver DE: Comparison of continuous positive airway pressure, incentive spirometry, and conservative therapy after cardiac operations. Crit Care Med. 1984, 12: 969-972.View ArticlePubMedGoogle Scholar
  52. Amoroso TA: Evaluation of the patient with blunt abdominal trauma: an evidence based approach. Emerg Med Clin North Am. 1999, 17: 63-75. viiiView ArticlePubMedGoogle Scholar
  53. Rozycki GS, Tremblay L, Feliciano DV, Tchorz K, Hattaway A, Fountain J: A prospective study for the detection of vascular injury in adult and pediatric patients with cervicothoracic seat belt signs. J Trauma. 2002, 52: 618-623.View ArticlePubMedGoogle Scholar
  54. Velmahos GC, Tatevossian R, Demetriades D: The "seat belt mark" sign: a call for increased vigilance among physicians treating victims of motor vehicle accidents. Am Surg. 1999, 65: 181-185.PubMedGoogle Scholar
  55. Hayes CW, Conway WF, Walsh JW, Coppage L, Gervin AS: Seat belt injuries: radiologic findings and clinical correlation. Radiographics. 1991, 11: 23-36.View ArticlePubMedGoogle Scholar
  56. Wotherspoon S, Chu K, Brown AF: Abdominal injury and the seat-belt sign. Emerg Med (Fremantle). 2001, 13: 61-65. 10.1046/j.1442-2026.2001.00180.x.View ArticleGoogle Scholar
  57. Sivit CJ, Taylor GA, Newman KD, Bulas DI, Gotschall CS, Wright CJ: Safety-belt injuries in children with lap-belt ecchymosis: CT findings in 61 patients. AJR Am J Roentgenol. 1991, 157: 111-114.View ArticlePubMedGoogle Scholar
  58. Chiu WC, Cushing BM, Rodriguez A, Ho SM, Mirvis SE, Shanmuganathan K: Abdominal injuries without hemoperitoneum: a potential limitation of focused abdominal sonography for trauma (FAST). J Trauma. 1997, 42: 617-623.View ArticlePubMedGoogle Scholar
  59. Johnstone BR, Waxman BP: Transverse disruption of the abdominal wall – a tell-tale sign of seat belt related hollow viscus injury. Aust N Z J Surg. 1987, 57: 455-460.View ArticlePubMedGoogle Scholar
  60. Rodriguez A, DuPriest RW, Shatney CH: Recognition of intra-abdominal injury in blunt trauma victims. A prospective study comparing physical examination with peritoneal lavage. Am Surg. 1982, 48: 457-459.PubMedGoogle Scholar
  61. Bivins BA, Sachatello CR, Daugherty ME, Ernst CB, Griffen WO: Diagnostic peritoneal lavage is superior to clinical evaluation in blunt abdominal trauma. Am Surg. 1978, 44: 637-641.PubMedGoogle Scholar
  62. Olsen WR, Redman HC, Hildreth DH: Quantitative peritoneal lavage in blunt abdominal trauma. Arch Surg. 1972, 104: 536-543.View ArticlePubMedGoogle Scholar
  63. Pre-publication history

    1. The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2482/4/3/prepub

Copyright

© Dunham et al; licensee BioMed Central Ltd. 2004

This article is published under license to BioMed Central Ltd. This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose, provided this notice is preserved along with the article's original URL.

Advertisement