Abdominal Pain An Analysis of 1,000 Consecutive Room Richard J. Brewer, MD, Charlottesville, Gerald T. Golden, MD, Charlottesville,
Cases in a University Hospital Emergency
Virginia Virginia
David C. Hitch, MD,* Charlottesville,
Virginia
Leslie E. Rudolf, MD, Charlottesville,
Virginia
Stephen L. Wangensteen, MD, Charlottesville,
Virginia
Abdominal pain unrelated to trauma has been the presenting complaint in one of every twenty patients seen in the emergency clinic at the University of Virginia Medical Center. A retrospective study of 1,000 consecutive cases of abdominal pain has been performed to (1) delineate the causes of abdominal pain in a large number of patients, (2) determine the efficiency of an active emergency facility in the evaluation of abdominal pain, and (3) outline those features frequently associated with the acute surgical abdomen and assess their potential value as screening devices in the selection of patients who may require immediate operation. Design of Study The University of Virginia Medical Center, located in Charlottesville, Virginia, serves a local city-county population of ,approximately 90,000 and is the primary hospital for an additional 70,000 persons in seven surrounding counties. In this emergency service, all patients who complain of abdominal pain are examined by a junior surgical resident, who may consult with the senior resident at his discretion. The charts of 1,000 consecutive patients who presented with abdominal pain between July 1, 1971 and January 15, 1972 were reviewed; excluded were patients under fifteen years of age and thosewith a recent history of abdominal trauma. During this six month period, 19,876 patients were treated for all causes in this emergency clinic. To insure uniformity of data abstraction, a questionnaire containing eightynine items of information was constructed and each patient’s chart was reviewed in detail. An initial diagnosis was made during the first visit in the emergency room; the final diagnosis was established by confirmatory eviFrom the Department of Surgery, University of Virginia Medical Center. Charlottesville, Virginia. Reprint requests should be addressed to Gerald T. Golden, MD, Box 71, Department of Surgery, University of Virginia Hospital, Charlottesville, Virginia 22901. = Present address: Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada.
Vokmo 131, February 1076
dence obtained
at surgery, by subsequent
cultures, by
roentgenograms, or by laboratory data. There were three exceptions. First, in patients with cramping abdominal pain, diarrhea, and vomiting, whose symptoms subsided within 48 hours, a final diagnosis of gastroenteritis was assigned without further supporting evidence. Second, in female patients with a history, physical findings, and clinical course typical of pelvic inflammatory disease, this diagnosis was assigned without obtaining positive cervical cultures. Third, in patients in whom no diagnosis was established, a diagnosis of abdominal pain of unknown cause was assigned. All information was entered into a computer program to facilitate handling of data.
Results The age and race distributions of this group of patients are presented in Figures 1 and 2. More than 40 per cent of the patients were between fifteen and twenty-four years of age. This constitutes a disproportionate number of young patients compared with the general population distribution in this area, even though the University students are included in the distribution. The twenty most common diagnoses in this group of 1,000 patients with abdominal pain are as follows: abdominal pain of unknown cause, 41.3 per cent gastroenteritis, 6.9 per cent pelvic inflammatory disease, 6.7 per cent urinary tract infection, 5.2 per cent ureteral stone, 4.3 per cent appendicitis, 4.3 per cent acute cholecystitis, 2.5 per cent intestinal obstruction, 2.5 per cent constipation, 2.3 per cent duodenal ulcer, 2.0 per cent dysmenorrhea, 1.8 per cent simple pregnancy, 1.8 per cent pyelonephritis, 1.7 per cent gastritis, 1.4 per cent
219
Brewer
et al
AGE RANGE
Figure 1. Population distribution by age.
TABLE
I Surgery for a Nonsurgical Condition (20 patients)
Number of Patients 1
4 2 1
3 1 1
Acute Acute
appendicitis appendicitis
Acute Acute Acute
appendicitis abdomen abdomen
2
Acute abdomen Ectopic pregnancy Ectopic pregnancy
1
Ovarian
1
Diverticulitis perforation Tubo-ovarian abscess
3
220
Preoperative Diagnosis
cyst with
Postoperative Diagnosis Mesenteric lymphadenitis Abdominal pain of unknown cause Acute salpingitis Pyelonephritis Abdominal pain of unknown pause Acute salpingitis Acute salpingitis Abdominal pain of unknown cause Abdominal pain of unknown cause Acute salpingitis Acute
salpingitis
chronic cholecystitis, 1.2 per cent ovarian cyst, 1.0 per cent incomplete abortion, 1.0 per cent pancreatitis, 0.9 per cent abdominal aneurysm, 0.7 per cent epididymitis, 0.7 per cent In more than 40 per cent of the patients, the cause of abdominal pain was not determined. Although 12 per cent of this group were reevaluated in the hospital or outpatient department, the majority did not return after the initial evaluation in the emergency room. Gastroenteritis, pelvic inflammatory disease, urinary tract infections, and ureteral calculi accounted for the difficulty in the majority of patients in whom a diagnosis could be assigned whereas acute appendicitis was the most common disease for which immediate operation was required. Of the 1,000 patients in this series, 274 were ultimately admitted to the hospital. Of this group, 124 (45 per cent) were treated by observation, appropriate diagnostic tests, and nonoperative management. The additional 150 patients underwent operation during their hospitalization. In 20 patients, a condition that did not require immediate surgical treatment was found at operation. (Table I.) The remaining 130 patients all had conditions that required immediate surgical intervention. In eleven cases, an acute surgical condition was not recognized initially and the patients were discharged from the emergency room. Early acute appendicitis was present in eight of these patients. (Table II.) In only one patient did significant morbidity occur, presumably as a result of such a diagnostic error. This patient had perforation of the appendix and a prolonged hospitalization but survived without sequelae. Three patients had early acute intestinal obstruction. All had undergone intra-abdominal operations, and all had normal abdominal roentgenograms when first evaluated. Three hundred eighty-six patients had had abdominal pain longer than 48 hours when first exTABLE
No. of Patients 1 4 3 1 2
II
Diagnostic Errors in 11 Patients Initially Discharged from the Emergency Room Initial Diagnosis Endometriosis Gastroenteritis Abdominal pain of unknown cause Urinary tract infection Abdominal pain of unknown cause
Correct Diagnosis Acute Acute Acute
appendicitis appendicitis appendicitis
Acute intestinal obstruction Acute intestinal obstruction
The American Journal GI Surgery
Abdominal
amined in the emergency clinic. Of this group, 82 (21 per cent) were admitted to the hospital and 43 (11 per cent) were operated on. In 614 patients with symptoms less than 48 hours, 192 (31 per cent) were hospitalized and 107 (18 per cent) were operated on. Although patients who presented with abdominal pain of less than 48 hours’ duration were hospitalized more often and were operated on with greater frequency (chi square analysis, p < O.OOl), the differences were not striking. Of the patients sixty-five years of age or older, a total of 51 per cent were admitted to the hospital and 33 per cent required immediate operation. By contrast, only 16 per cent of patients less than sixty-five years of age underwent a surgical procedure while in. the hospital (p < 0.001). Moreover, during hospitalization the mortality for patients older than sixty-five years was 8.4 per cent compared with 0.9 per cent for patients less than sixtyfive years (p < 0.001). A hematocrit value and white blood cell count were obtained in 95 per cent of the patients. Forty per cent of the blood leukocyte counts were greater than 10,006 per mm3. Of patients who had a disease requiring immediate surgical intervention, only 39 per cent of those older than sixty-five years had a white blood cell count greater than 10,000 per mm3 compared with 71 per cent of those patients under sixty-five. Abdominal roentgenograms, obtained in 427 patients, were considered abnormal in 174 (38 per cent). In only 58 patients did roentgenographic findings provide positive information concerning a specific diagnosis such as appendolith, mechanical small bowel obstruction, or a visible ureteral calculus. In no case was the clinical diagnosis changed by examination of the roentgenograms. In the remaining cases, the findings were not diagnostic,
TABLE I II
consisting usually of a nonspecific ileus pattern. Chest roentgenograms were abnormal in 15 per cent of the 320 patients in whom they were obtained, but in only 3 per cent did this examination yield specific diagnostic information, such as free air under the diaphragm, that was helpful in the diagnosis of the abdominal problem. The symptoms, signs, and laboratory data in patients with abdominal pain of unknown cause and gastroenteritis were compared with those in patients with acute appendicitis, acute cholecystitis, and acute mechanical small bowel obstruction. Table III lists the frequency of associated complaints in patients with these conditions. Anorexia and vomiting were common in all diagnostic categories. In patients with abdominal pain of unknown cause and gastroenteritis, the temporal relation between pain and vomiting was erratic. In patients with surgical illnesses, however, the onset of vomiting invariably followed the onset of pain. The abdominal physical findings in patients with these conditions are listed in Table IV. Rigidity, defined as intense, involuntary muscle spasm, was an infrequent finding. Guarding, defined as a moderate resistance to palpation, and rebound tenderness were associated with surgical disease in a larger proportion of patients. Abnormal bowel sounds (hypoactive, hyperactive, or obstructive) were associated with a high percentage of illnesses in every category but, as would be expected, were most frequent in patients with acute intestinal obstruction (93 per cent). Tenderness on rectal examination was found in 30 per cent of the patients with acute appendicitis but in an insignificant percentage of patients with other diagnoses. Although acute appendicitis, acute cholecystitis, and intestinal obstruction were associated with high white blood cell counts (Table V), the white
Associated Symptoms in Patients with Abdominal Pain Associated
Diagnosis Acute (43
Pain
appendicitis cases)
Abdominal pain of unknown cause (406 cases)
Symptoms
Pain Followed by Vomiting
Anorexia
Vomiting
33
(72%)
21
(49%)
21/21
135
(100%)
(33%)
105
(26%)
21/105
Cholecystitis (26 cases)
18 (69%)
20
(77%)
20/20
(100%)
Gastroenteritis (68 cases)
48
(70%)
68
(100%)
16/68
(24%)
Intestinal obstruction (27 cases)
22 (81%)
20
(74%)
20/20
(100%)
Volume 131, Fetmiaq 1976
(20%)
Diarrhea
Obstipation
5 (15%)
5 (15%)
30 (0.7%)
22 (0.5%)
2 (8%) 68
0
Similar Symptoms Previously
3 (9%)
138
(33%)
11 (42%)
(100%)
1 (1.4%)
6 (12%)
4 (15%)
5 (18%)
13 (48%)
221
Brewer et al
TABLE
IV
Abdominal
Physical
Findings
in Patients with Abdominal
Pain
Abdominal Physical Finding
Appendicitis (43 cases) Abdominal pain of unknown cause (406 cases) Cholecystitis (26 cases) Gastroenteritis (68 cases) Intestinal obstruction (27 cases)
TABLE
V
White
Rebound Tenderness
Guarding
Rigidity
Diagnosis
Abnormal Bowel Sounds
Rectal Tenderness
13 (30%)
3 (7%)
27 (63%)
36 (84%)
32 (77%)
4 (1%)
65 (16%)
50 (12%)
221 (50%)
17 (4.3%)
1 (3.9%)
15 (58%)
9 (35%)
10 (39%)
1 (3.9%)
0
10 (15%)
5 (7.5%)
45 (67%)
1 (1.5%)
6 (22%)
25 (93%)
1 (3.7%)
9 (33%)
2 (7.4%)
Blood Cell Count
in Patients with Abdominal
Pain White Blood Cell Count (oer mm?
Diagnosis Acute
20.000
37 (86%)
2 (4.6%)
228 (56%)
121 (26%)
2 (5.0%)
6 (23%)
17 (66%)
3 (12%)
appendicitis 3 (7%)
0
(43 cases) Abdominal pain of unknown cause
27 (6%)
(406 cases) Acute cholecystitis (26 cases) Gastroenteritis
0
(68 cases) Intestinal obstruction
2 (2.9%)
31 (46%)
29 (43%)
0
1 (4%)
12 (45%)
14 (56%)
0
(27 cases)
TABLE
VI
Temperature in’patients with Abdominal Pain Temperature (“C)
Diagnosis Acute
36-37.9
38-38.9
39-39.9
Comments
appendicitis
(43 cases) Abdominal pain of unknown cause (406 cases) Acute cholecystitis (26 cases) Gastroenteritis (68 cases) Intestinal obstruction (27 cases) -__----
24 (56%)
12 (28%)
7 (16%)
337 (73%)
50 (12%)
3 (0.7%)
18 (69%)
4 (15%)
4 (15%)
55 (81%)
8 (11%)
6 (9%)
24 (89%)
3 (11%) 0 ---__-
blood cell count was elevated in 26 per cent of the patients in whom a diagnosis was never established and in almost one-half the patients with gastroenteritis. Table VI shows the recorded temperatures in these patients. A temperature elevation is expect-
222
ed with acute surgical illness and may occur in patients with nonsurgical causes of abdominal pain with considerable frequency.
In patients in whom a specific diagnosis could be assigned, it was of interest that pelvic inflammatory disease, urinary tract infection, and ureteral calculi were all more common than conditions requiring immediate operation, such as acute appendicitis. The high incidence of ureteral calculi in this series probably reflects the high incidence of this disease in Virginia compared with other parts of the country [I]. It is even more noteworthy, however, that a cause for abdominal pain was not elucidated in more than 40 per cent of the patients in this series. Even though some of these patients may have sought treatment subsequently at another hospital, it is more likely that the pain subsided and did not return. Brooke and his associates [2,3] cite a similarly low diagnostic and ther-
The American Journal of Surgery
Abdominal
apeutic yield in emergency room patients with subacute gastrointestinal symptoms. Although similar in some respects to our own survey, these studies analyzed the overall health care and costeffectiveness of an emergency facility. To our knowledge, the present study represents the first attempt to categorize a large number of emergency room patients with respect to diagnosis, signs, and symptoms. A false-positive diagnostic rate of 2.0 per cent occurred in this series, in that twenty patients who did not require immediate surgical treatment underwent operation. Most frequently the preoperative diagnosis was acute appendicitis or tubo-ovarian abscess, and in this situation an occasional negative exploration is justified. In this series of patients, no mortality or significant morbidity occurred. Diagnostic errors resulting in inappropriately delayed surgery were most often due to early acute appendicitis and small bowel obstruction. The elusive nature of early acute appendicitis is well known, and frequent reexamination in the hospital or outpatient department is the mainstay of sound surgical practice. Acute small bowel obstruction was the third most common disease for which immediate operation was required in this series. Patients with abdominal pain who have had intra-abdominal surgery are at risk in that they may have acute small bowel obstruction despite normal roentgenograms early in their illness. Patients with pain of less than 48 hours’ duration have a significantly higher incidence of surgical disease than do those with pain of longer duration. Although pain of short duration identifies patients at high risk, pain of several days’ duration is not reassuring from the standpoint of excluding an acute surgical abdominal condition. The elderly patient with abdominal pain is more likely to have a surgical disease than is the younger counterpart. This group requires particularly careful evaluation and frequent reevaluation. Although fever and leukocytosis frequently accompany acute surgical disease, both are often present in patients with nonsurgical conditions. Both parameters added little more than supportive evidence for the diagnosis of an acute surgical abdomen in this series of patients. Normal values are not reassuring, are common in the elderly, and cannot be the determinants of conservative treatment. No consistent relation between the symptoms of anorexia, vomiting, and obstipation appeared in this study to clearly identify those patients with acute surgical disease. Moreover, many patients
Volume 131, February la76
Pain
with surgical illness reported previous episodes of similar symptoms, and a prior episode is of little help in excluding the diagnosis of an acute surgical disease. The only significant feature in the comparison of symptoms was the initial onset of abdominal pain followed by vomiting, which appeared to be uniform in patients with acute surgical illness who experienced vomiting. Pain followed by vomiting suggests obstruction of a hollow viscus and may be helpful in diagnosis, whether it is due to a ureteral calculus, a stone in the cystic or common bile duct, an adhesion obstructing the small intestine, or a fecolith in the appendix. Guarding and rebound tenderness were frequent in patients requiring operation and relatively infrequent in patients with abdominal pain from nonsurgical causes. Summary and Conclusions In the majority of patients in this series of 1,000, acute abdominal pain was due to conditions that required neither surgical intervention nor hospitalization. Eleven of the 1,000 patients had an early missed diagnosis in the emergency clinic for which a subsequent operation was needed, and twenty underwent an operation which subsequent diagnosis showed was not required. All false-negative evaluations occurred in patients with early appendicitis or small bowel obstruction. Most falsepositive results were due to acute infections of the female genitourinary tract in patients operated on to exclude appendicitis or a tubo-ovarian abscess. The following factors help identify the high risk patient with an acute surgical abdomen: (1) pain for less than 48 hours; (2) pain followed by vomiting; (3) guarding and rebound tenderness on physical examination; (4) advanced age; (5) a prior surgical procedure. The presence of these features demands careful evaluation and a liberal policy of admission and observation. White blood cell counts, body temperature, and abnormal abdominal roentgenograms may add confirmatory evidence but are not particularly helpful as screening devices. References 1. Boyce WH. Garvey FK. Strawcutter HE: Incidence of urinary calculi among patients in general hospitals. 1948-1952. JAMA 161: 1437, 1956. 2. Brooke RH, Berg MH, Schechter PA: Effectiveness of nonemergency care via an emergeqcy room. Ann Intern &ted 78: 333, 1973. 3. Brooke RH, Stevenson RL: Effectiveness of patient care in an emergency room. N Engl J Med 283: 904, 1970.
223
Cases in a University Hospital Emergency
Virginia Virginia
David C. Hitch, MD,* Charlottesville,
Virginia
Leslie E. Rudolf, MD, Charlottesville,
Virginia
Stephen L. Wangensteen, MD, Charlottesville,
Virginia
Abdominal pain unrelated to trauma has been the presenting complaint in one of every twenty patients seen in the emergency clinic at the University of Virginia Medical Center. A retrospective study of 1,000 consecutive cases of abdominal pain has been performed to (1) delineate the causes of abdominal pain in a large number of patients, (2) determine the efficiency of an active emergency facility in the evaluation of abdominal pain, and (3) outline those features frequently associated with the acute surgical abdomen and assess their potential value as screening devices in the selection of patients who may require immediate operation. Design of Study The University of Virginia Medical Center, located in Charlottesville, Virginia, serves a local city-county population of ,approximately 90,000 and is the primary hospital for an additional 70,000 persons in seven surrounding counties. In this emergency service, all patients who complain of abdominal pain are examined by a junior surgical resident, who may consult with the senior resident at his discretion. The charts of 1,000 consecutive patients who presented with abdominal pain between July 1, 1971 and January 15, 1972 were reviewed; excluded were patients under fifteen years of age and thosewith a recent history of abdominal trauma. During this six month period, 19,876 patients were treated for all causes in this emergency clinic. To insure uniformity of data abstraction, a questionnaire containing eightynine items of information was constructed and each patient’s chart was reviewed in detail. An initial diagnosis was made during the first visit in the emergency room; the final diagnosis was established by confirmatory eviFrom the Department of Surgery, University of Virginia Medical Center. Charlottesville, Virginia. Reprint requests should be addressed to Gerald T. Golden, MD, Box 71, Department of Surgery, University of Virginia Hospital, Charlottesville, Virginia 22901. = Present address: Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada.
Vokmo 131, February 1076
dence obtained
at surgery, by subsequent
cultures, by
roentgenograms, or by laboratory data. There were three exceptions. First, in patients with cramping abdominal pain, diarrhea, and vomiting, whose symptoms subsided within 48 hours, a final diagnosis of gastroenteritis was assigned without further supporting evidence. Second, in female patients with a history, physical findings, and clinical course typical of pelvic inflammatory disease, this diagnosis was assigned without obtaining positive cervical cultures. Third, in patients in whom no diagnosis was established, a diagnosis of abdominal pain of unknown cause was assigned. All information was entered into a computer program to facilitate handling of data.
Results The age and race distributions of this group of patients are presented in Figures 1 and 2. More than 40 per cent of the patients were between fifteen and twenty-four years of age. This constitutes a disproportionate number of young patients compared with the general population distribution in this area, even though the University students are included in the distribution. The twenty most common diagnoses in this group of 1,000 patients with abdominal pain are as follows: abdominal pain of unknown cause, 41.3 per cent gastroenteritis, 6.9 per cent pelvic inflammatory disease, 6.7 per cent urinary tract infection, 5.2 per cent ureteral stone, 4.3 per cent appendicitis, 4.3 per cent acute cholecystitis, 2.5 per cent intestinal obstruction, 2.5 per cent constipation, 2.3 per cent duodenal ulcer, 2.0 per cent dysmenorrhea, 1.8 per cent simple pregnancy, 1.8 per cent pyelonephritis, 1.7 per cent gastritis, 1.4 per cent
219
Brewer
et al
AGE RANGE
Figure 1. Population distribution by age.
TABLE
I Surgery for a Nonsurgical Condition (20 patients)
Number of Patients 1
4 2 1
3 1 1
Acute Acute
appendicitis appendicitis
Acute Acute Acute
appendicitis abdomen abdomen
2
Acute abdomen Ectopic pregnancy Ectopic pregnancy
1
Ovarian
1
Diverticulitis perforation Tubo-ovarian abscess
3
220
Preoperative Diagnosis
cyst with
Postoperative Diagnosis Mesenteric lymphadenitis Abdominal pain of unknown cause Acute salpingitis Pyelonephritis Abdominal pain of unknown pause Acute salpingitis Acute salpingitis Abdominal pain of unknown cause Abdominal pain of unknown cause Acute salpingitis Acute
salpingitis
chronic cholecystitis, 1.2 per cent ovarian cyst, 1.0 per cent incomplete abortion, 1.0 per cent pancreatitis, 0.9 per cent abdominal aneurysm, 0.7 per cent epididymitis, 0.7 per cent In more than 40 per cent of the patients, the cause of abdominal pain was not determined. Although 12 per cent of this group were reevaluated in the hospital or outpatient department, the majority did not return after the initial evaluation in the emergency room. Gastroenteritis, pelvic inflammatory disease, urinary tract infections, and ureteral calculi accounted for the difficulty in the majority of patients in whom a diagnosis could be assigned whereas acute appendicitis was the most common disease for which immediate operation was required. Of the 1,000 patients in this series, 274 were ultimately admitted to the hospital. Of this group, 124 (45 per cent) were treated by observation, appropriate diagnostic tests, and nonoperative management. The additional 150 patients underwent operation during their hospitalization. In 20 patients, a condition that did not require immediate surgical treatment was found at operation. (Table I.) The remaining 130 patients all had conditions that required immediate surgical intervention. In eleven cases, an acute surgical condition was not recognized initially and the patients were discharged from the emergency room. Early acute appendicitis was present in eight of these patients. (Table II.) In only one patient did significant morbidity occur, presumably as a result of such a diagnostic error. This patient had perforation of the appendix and a prolonged hospitalization but survived without sequelae. Three patients had early acute intestinal obstruction. All had undergone intra-abdominal operations, and all had normal abdominal roentgenograms when first evaluated. Three hundred eighty-six patients had had abdominal pain longer than 48 hours when first exTABLE
No. of Patients 1 4 3 1 2
II
Diagnostic Errors in 11 Patients Initially Discharged from the Emergency Room Initial Diagnosis Endometriosis Gastroenteritis Abdominal pain of unknown cause Urinary tract infection Abdominal pain of unknown cause
Correct Diagnosis Acute Acute Acute
appendicitis appendicitis appendicitis
Acute intestinal obstruction Acute intestinal obstruction
The American Journal GI Surgery
Abdominal
amined in the emergency clinic. Of this group, 82 (21 per cent) were admitted to the hospital and 43 (11 per cent) were operated on. In 614 patients with symptoms less than 48 hours, 192 (31 per cent) were hospitalized and 107 (18 per cent) were operated on. Although patients who presented with abdominal pain of less than 48 hours’ duration were hospitalized more often and were operated on with greater frequency (chi square analysis, p < O.OOl), the differences were not striking. Of the patients sixty-five years of age or older, a total of 51 per cent were admitted to the hospital and 33 per cent required immediate operation. By contrast, only 16 per cent of patients less than sixty-five years of age underwent a surgical procedure while in. the hospital (p < 0.001). Moreover, during hospitalization the mortality for patients older than sixty-five years was 8.4 per cent compared with 0.9 per cent for patients less than sixtyfive years (p < 0.001). A hematocrit value and white blood cell count were obtained in 95 per cent of the patients. Forty per cent of the blood leukocyte counts were greater than 10,006 per mm3. Of patients who had a disease requiring immediate surgical intervention, only 39 per cent of those older than sixty-five years had a white blood cell count greater than 10,000 per mm3 compared with 71 per cent of those patients under sixty-five. Abdominal roentgenograms, obtained in 427 patients, were considered abnormal in 174 (38 per cent). In only 58 patients did roentgenographic findings provide positive information concerning a specific diagnosis such as appendolith, mechanical small bowel obstruction, or a visible ureteral calculus. In no case was the clinical diagnosis changed by examination of the roentgenograms. In the remaining cases, the findings were not diagnostic,
TABLE I II
consisting usually of a nonspecific ileus pattern. Chest roentgenograms were abnormal in 15 per cent of the 320 patients in whom they were obtained, but in only 3 per cent did this examination yield specific diagnostic information, such as free air under the diaphragm, that was helpful in the diagnosis of the abdominal problem. The symptoms, signs, and laboratory data in patients with abdominal pain of unknown cause and gastroenteritis were compared with those in patients with acute appendicitis, acute cholecystitis, and acute mechanical small bowel obstruction. Table III lists the frequency of associated complaints in patients with these conditions. Anorexia and vomiting were common in all diagnostic categories. In patients with abdominal pain of unknown cause and gastroenteritis, the temporal relation between pain and vomiting was erratic. In patients with surgical illnesses, however, the onset of vomiting invariably followed the onset of pain. The abdominal physical findings in patients with these conditions are listed in Table IV. Rigidity, defined as intense, involuntary muscle spasm, was an infrequent finding. Guarding, defined as a moderate resistance to palpation, and rebound tenderness were associated with surgical disease in a larger proportion of patients. Abnormal bowel sounds (hypoactive, hyperactive, or obstructive) were associated with a high percentage of illnesses in every category but, as would be expected, were most frequent in patients with acute intestinal obstruction (93 per cent). Tenderness on rectal examination was found in 30 per cent of the patients with acute appendicitis but in an insignificant percentage of patients with other diagnoses. Although acute appendicitis, acute cholecystitis, and intestinal obstruction were associated with high white blood cell counts (Table V), the white
Associated Symptoms in Patients with Abdominal Pain Associated
Diagnosis Acute (43
Pain
appendicitis cases)
Abdominal pain of unknown cause (406 cases)
Symptoms
Pain Followed by Vomiting
Anorexia
Vomiting
33
(72%)
21
(49%)
21/21
135
(100%)
(33%)
105
(26%)
21/105
Cholecystitis (26 cases)
18 (69%)
20
(77%)
20/20
(100%)
Gastroenteritis (68 cases)
48
(70%)
68
(100%)
16/68
(24%)
Intestinal obstruction (27 cases)
22 (81%)
20
(74%)
20/20
(100%)
Volume 131, Fetmiaq 1976
(20%)
Diarrhea
Obstipation
5 (15%)
5 (15%)
30 (0.7%)
22 (0.5%)
2 (8%) 68
0
Similar Symptoms Previously
3 (9%)
138
(33%)
11 (42%)
(100%)
1 (1.4%)
6 (12%)
4 (15%)
5 (18%)
13 (48%)
221
Brewer et al
TABLE
IV
Abdominal
Physical
Findings
in Patients with Abdominal
Pain
Abdominal Physical Finding
Appendicitis (43 cases) Abdominal pain of unknown cause (406 cases) Cholecystitis (26 cases) Gastroenteritis (68 cases) Intestinal obstruction (27 cases)
TABLE
V
White
Rebound Tenderness
Guarding
Rigidity
Diagnosis
Abnormal Bowel Sounds
Rectal Tenderness
13 (30%)
3 (7%)
27 (63%)
36 (84%)
32 (77%)
4 (1%)
65 (16%)
50 (12%)
221 (50%)
17 (4.3%)
1 (3.9%)
15 (58%)
9 (35%)
10 (39%)
1 (3.9%)
0
10 (15%)
5 (7.5%)
45 (67%)
1 (1.5%)
6 (22%)
25 (93%)
1 (3.7%)
9 (33%)
2 (7.4%)
Blood Cell Count
in Patients with Abdominal
Pain White Blood Cell Count (oer mm?
Diagnosis Acute
20.000
37 (86%)
2 (4.6%)
228 (56%)
121 (26%)
2 (5.0%)
6 (23%)
17 (66%)
3 (12%)
appendicitis 3 (7%)
0
(43 cases) Abdominal pain of unknown cause
27 (6%)
(406 cases) Acute cholecystitis (26 cases) Gastroenteritis
0
(68 cases) Intestinal obstruction
2 (2.9%)
31 (46%)
29 (43%)
0
1 (4%)
12 (45%)
14 (56%)
0
(27 cases)
TABLE
VI
Temperature in’patients with Abdominal Pain Temperature (“C)
Diagnosis Acute
36-37.9
38-38.9
39-39.9
Comments
appendicitis
(43 cases) Abdominal pain of unknown cause (406 cases) Acute cholecystitis (26 cases) Gastroenteritis (68 cases) Intestinal obstruction (27 cases) -__----
24 (56%)
12 (28%)
7 (16%)
337 (73%)
50 (12%)
3 (0.7%)
18 (69%)
4 (15%)
4 (15%)
55 (81%)
8 (11%)
6 (9%)
24 (89%)
3 (11%) 0 ---__-
blood cell count was elevated in 26 per cent of the patients in whom a diagnosis was never established and in almost one-half the patients with gastroenteritis. Table VI shows the recorded temperatures in these patients. A temperature elevation is expect-
222
ed with acute surgical illness and may occur in patients with nonsurgical causes of abdominal pain with considerable frequency.
In patients in whom a specific diagnosis could be assigned, it was of interest that pelvic inflammatory disease, urinary tract infection, and ureteral calculi were all more common than conditions requiring immediate operation, such as acute appendicitis. The high incidence of ureteral calculi in this series probably reflects the high incidence of this disease in Virginia compared with other parts of the country [I]. It is even more noteworthy, however, that a cause for abdominal pain was not elucidated in more than 40 per cent of the patients in this series. Even though some of these patients may have sought treatment subsequently at another hospital, it is more likely that the pain subsided and did not return. Brooke and his associates [2,3] cite a similarly low diagnostic and ther-
The American Journal of Surgery
Abdominal
apeutic yield in emergency room patients with subacute gastrointestinal symptoms. Although similar in some respects to our own survey, these studies analyzed the overall health care and costeffectiveness of an emergency facility. To our knowledge, the present study represents the first attempt to categorize a large number of emergency room patients with respect to diagnosis, signs, and symptoms. A false-positive diagnostic rate of 2.0 per cent occurred in this series, in that twenty patients who did not require immediate surgical treatment underwent operation. Most frequently the preoperative diagnosis was acute appendicitis or tubo-ovarian abscess, and in this situation an occasional negative exploration is justified. In this series of patients, no mortality or significant morbidity occurred. Diagnostic errors resulting in inappropriately delayed surgery were most often due to early acute appendicitis and small bowel obstruction. The elusive nature of early acute appendicitis is well known, and frequent reexamination in the hospital or outpatient department is the mainstay of sound surgical practice. Acute small bowel obstruction was the third most common disease for which immediate operation was required in this series. Patients with abdominal pain who have had intra-abdominal surgery are at risk in that they may have acute small bowel obstruction despite normal roentgenograms early in their illness. Patients with pain of less than 48 hours’ duration have a significantly higher incidence of surgical disease than do those with pain of longer duration. Although pain of short duration identifies patients at high risk, pain of several days’ duration is not reassuring from the standpoint of excluding an acute surgical abdominal condition. The elderly patient with abdominal pain is more likely to have a surgical disease than is the younger counterpart. This group requires particularly careful evaluation and frequent reevaluation. Although fever and leukocytosis frequently accompany acute surgical disease, both are often present in patients with nonsurgical conditions. Both parameters added little more than supportive evidence for the diagnosis of an acute surgical abdomen in this series of patients. Normal values are not reassuring, are common in the elderly, and cannot be the determinants of conservative treatment. No consistent relation between the symptoms of anorexia, vomiting, and obstipation appeared in this study to clearly identify those patients with acute surgical disease. Moreover, many patients
Volume 131, February la76
Pain
with surgical illness reported previous episodes of similar symptoms, and a prior episode is of little help in excluding the diagnosis of an acute surgical disease. The only significant feature in the comparison of symptoms was the initial onset of abdominal pain followed by vomiting, which appeared to be uniform in patients with acute surgical illness who experienced vomiting. Pain followed by vomiting suggests obstruction of a hollow viscus and may be helpful in diagnosis, whether it is due to a ureteral calculus, a stone in the cystic or common bile duct, an adhesion obstructing the small intestine, or a fecolith in the appendix. Guarding and rebound tenderness were frequent in patients requiring operation and relatively infrequent in patients with abdominal pain from nonsurgical causes. Summary and Conclusions In the majority of patients in this series of 1,000, acute abdominal pain was due to conditions that required neither surgical intervention nor hospitalization. Eleven of the 1,000 patients had an early missed diagnosis in the emergency clinic for which a subsequent operation was needed, and twenty underwent an operation which subsequent diagnosis showed was not required. All false-negative evaluations occurred in patients with early appendicitis or small bowel obstruction. Most falsepositive results were due to acute infections of the female genitourinary tract in patients operated on to exclude appendicitis or a tubo-ovarian abscess. The following factors help identify the high risk patient with an acute surgical abdomen: (1) pain for less than 48 hours; (2) pain followed by vomiting; (3) guarding and rebound tenderness on physical examination; (4) advanced age; (5) a prior surgical procedure. The presence of these features demands careful evaluation and a liberal policy of admission and observation. White blood cell counts, body temperature, and abnormal abdominal roentgenograms may add confirmatory evidence but are not particularly helpful as screening devices. References 1. Boyce WH. Garvey FK. Strawcutter HE: Incidence of urinary calculi among patients in general hospitals. 1948-1952. JAMA 161: 1437, 1956. 2. Brooke RH, Berg MH, Schechter PA: Effectiveness of nonemergency care via an emergeqcy room. Ann Intern &ted 78: 333, 1973. 3. Brooke RH, Stevenson RL: Effectiveness of patient care in an emergency room. N Engl J Med 283: 904, 1970.
223
