|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2021 | Volume
: 27
| Issue : 2 | Page : 116-121 |
|
Diagnostic validity of x-rays in detection of hollow viscus perforation
Ajay M Jose, MP Saanida, EV Gopi
Department of General Surgery, Medical College Hospital, Kozhikode, Kerala, India
| Date of Submission | 09-Apr-2021 |
| Date of Decision | 10-Apr-2021 |
| Date of Acceptance | 11-Apr-2021 |
| Date of Web Publication | 15-Nov-2021 |
Correspondence Address:
Dr. Ajay M Jose
Medical College Hospital, Kozhikode, Kerala
India
 Source of Support: None, Conflict of Interest: None
DOI: 10.4103/ksj.ksj_23_21
Introduction: Identification of a small amount of free intra-abdominal gas remains one of the most significant signs in a perforated viscus. This paper tried to evaluate the sensitivity of chest radiographs in the detection of hollow viscus perforation and to compare the sensitivity of left lateral decubitus view and chest radiograph posteroanterior (PA) view in the detection of pneumoperitoneum. Methodology: Diagnostic validation study was conducted in sixty patients with clinical signs of peritonitis and strongly suspected to have perforation were evaluated with chest radiograph PA view and left lateral decubitus view in search of pneumoperitoneum and confirmed during laparotomy. Results: Out of sixty patients, 54 had gastrointestinal (GI) perforation and 42 out of those 54 were diagnosed in plain chest radiograph. Rest 12 required contrast enhancement computerised tomogam for diagnosis. There was only one case reported with chest radiograph showing pneumoperitoneum and the patient without any breach in the GI tract. Thirty patients with perforation took a left lateral decubitus view which showed free air only in 23 cases. Conclusion: Sensitivity of left lateral decubitus was around 76.6%. Overall, the sensitivity of plain chest radiograph in diagnosing perforation was 77% with specificity 83%, positive predicative value 97.6%, negative predicative value 29.4% and an accuracy of 79.6%.
Keywords: Contrast enhancement computerised tomogam, crescent sign, gastrointestinal perforation, left lateral decubitus view, plain radiograph, pneumoperitoneum
How to cite this article:
Jose AM, Saanida M P, Gopi E V. Diagnostic validity of x-rays in detection of hollow viscus perforation. Kerala Surg J 2021;27:116-21 |
| Introduction | |  |
Perforations in gastrointestinal (GI) tract may be due to peptic ulcer, inflammatory bowel disease, iatrogenic factors, foreign body or malignancy. Gastroduodenal perforation is about 32% of the total. It is important to identify the presence, location and the cause of perforation correctly for planning and appropriate management. Treatment of almost all perforation cases is emergency surgical intervention and it cannot be delayed. In the elderly, mortality associated with perforation may be as high as 50%.[1] Hence, early diagnosis of perforation is at most importance as the management should be initiated before the bacterial peritonitis and sepsis settles in.
Usually, the standard radiograph to detect the perforation remains the chest radiograph posteroanterior (PA) view. Another view which used in moribund patients is the left lateral decubitus view. Careful examination of the medial aspects of the cupola, particularly on the right side may be necessary to detect a thin film of air. If there is genuine doubt and there are no pressing indications for exploration, then it is possible again to gain further information with the use of a left lateral decubitus view or a carefully performed water-soluble contrast study, in which the contrast medium may be either swallowed or injected down the nasogastric tube.
The gold standard investigation in the detection of hollow viscus perforation is multidetector computed tomography (CT) scan. It can pick the minimal amount of pneumoperitoneum and can throw light into the source of give away in the gastrointestinal tract. Ultrasound scan is mostly operative dependent and can supplement the radiographs in the detection and not replace it.
This paper evaluated the sensitivity of radiographs in the detection of hollow viscus perforation, to compare the sensitivity of left lateral decubitus view and chest radiograph PA view in the detection of penumoperitoneum, to identify cases of pneumoretroperitoneum and to assess the predictive value of free gas under the diaphragm diagnosing as gastrointestinal perforation.
| Methodology | | |
This was a diagnostic validation study conducted in a Medical College. Sample size was calculated as sixty using formula. All patients with a history of trauma and presenting with abdominal pain and abdominal signs suspicious GI perforation attending the emergency department from August 2019 to August 2020 were included. Sampling procedure was consecutive cases till the sample size was reached. Patients presenting with abdominal pain and with no abdominal signs and stable vitals and those with appendicular perforations.
Data were collected after proper history taking, clinical examination, blood tests, radiographs, contrast enhancement computerised tomogam (CECT) abdomen and intraoperative finding. Outcome measurements included comparison of sensitivity, specificity, negative predicative value (NPV), positive predicative value (PPV) of radiographs and various views in the detection of hollow viscus perforation. After recording vitals and resuscitating the unstable ones, the patients were subjected to diagnostic investigations, namely, erect chest radiograph showing both the domes of diaphragm clearly. Routinely, an erect radiograph abdomen was taken along with a chest radiograph to include the entire bowel within the diagnostic range. We also included the left lateral decubitus view. If any of these radiographs revealed pneumoperitoneum in the form of free gas under diaphragm, the patient was diagnosed to have a breach in the continuity of GI tract and is taken for emergency laparotomy and proceed accordingly. If none of radiographs demonstrated free gas, then in order to rule out acute pancreatitis and acute cholecystitis serum amylase, lipase levels and (Ultrasonogram) USG abdomen were ordered. If still the diagnosis was uncertain, a CECT abdomen was ordered. Since the sensitivity of CECT in picking perforation is highest, it was the last resort in diagnosis. All the patients with signs of peritonitis and diagnosed to have perforation will have to undergo an emergency surgery in the form of exploratory laparotomy. Data collected were analysed using appropriate statistical tests.
| Results | | |
Sixty patients who presented to emergency department with symptoms and signs of peritonitis and strongly suspected to have GI perforation and were evaluated. The diagnostic validity of radiographs in the detection of perforation was studied. Age ranged from 16 to 85 years. The highest frequency was in 50–59 years – 16 (26.7%). Mean age of presentation was 45 years and median 49 years with standard deviation 18.057 [Table 1].
Amongst the patients, 52 (87%) were males and 8 (13%) were females. Out of these, two females and four males had features of peritonitis and were diagnosed to have no perforation. One showed air under diaphragm and had no perforation intraoperative.
All sixty patients presented with features of peritonitis had abdominal pain as a constant symptom either throughout or at some points of time during the course. Thirty-four (56%) had nausea or vomiting accompanying pain and only 11 (18.3%) each had constipation or fever. Amongst the 54 patients which had perforation, seven (14.8%) had a history of trauma.
Out of sixty patients who presented with peritonitis, 51 (85%) had rigidity as a clinical sign. Fifty-four had bowel perforations, and surprisingly, nine out of 54, that is 16% of patients had only localised signs. Forty-three (71.7%) patients had tachycardia and 16 (26.7%) presented in unstable condition with tachycardia and hypotension and had to be resuscitated. Fourteen (23.3%) patients out of sixty had diabetes and 21 (35%) had hypertension. Ten (16.7%) had both comorbidities.
Associated risk factors in the patients included primarily smoking and alcohol intake. Thirty-six (60%) of patients had smoking as a risk factor. Eighteen (30%) patients consumed alcohol 17 (28.4%) had smoking along with it. Sixteen (26.7%) had a history of taking non-steroidal anti-inflammatory drugs. Thirteen (21.6%) patients with perforation had a history of acid peptic disease.
Diagnosis of perforation in 42 (77.8%) out of 54 patients with perforation of GI tract were picked by radiograph, but CECT was required to detect the rest 12 (22.2%).
Sites of perforation in the 54 perforations were 26 (43.3%) gastric perforation, 9 (15%) in duodenum, 10 (16.7%) in small intestine mostly ileum, 7 (11.7%) had colon perforation and 2 (3.3%) rectal perforation. Grossly gastroduodenal perforations cumulative frequency was 58.3%.
The diagnostic validation and contingency matrix are given in [Table 2] and [Table 3]
Most commonly seen radiological sign in radiographs was crescent sign (subdiaphragmatic air) in 34 (56.7%). Double bowel sign was seen in 12 (20%), but Rigler's sign alone was not used to take decision on diagnosis; a well-exposed abdominal radiograph revealed continuous diaphragm sign in 7 (11.67%) radiographs which showed crescent sign. Foot ball sign (massive pneumoperitioneum) was seen in one (1.67%). Most patients come as referred cases with a radiograph. If further radiograph is taken, sometimes, it is found that the free air might have disappeared. It can due to the sealing of perforation with omentum or surface of the liver and the already escaped free gas getting absorbed. This is depicted in [Figure 1] and [Figure 2], where radiographs of the same person taken at two different points of time show free air at 4 h of onset of abdominal pain but vanished in the chest radiograph taken 12 h after the incident.
Adequate exposure while taking the radiograph is one of the important factors in detecting free gas. Chest radiograph PA view is the first and most important view which throws light into the diagnosis of pneumoperitoneum. Abdominal radiograph requires more exposure and is valuable in detecting the named signs but is never superior to a chest radiograph which shows domes of diaphragm clearly.
The left lateral decubitus view is an additional view that is used for diagnosis. It is most preferred for bedridden patients and those who cannot stand erect. In most of the studies conducted prior, it was found that the left lateral decubitus view was superior in diagnosing free air than an erect chest radiograph. In the present study, most of the patients were referred with a chest radiograph and diagnosis, so the rest of all patients presented to the emergency took a left lateral view along with it [Figure 3]. The study failed to highlight the superiority of left lateral decubitus view over chest radiograph, as none of the patients had a positive lateral decubitus radiograph with a negative chest radiograph. However, it was found that left lateral view made the free gas more prominent. Here, in the above-displayed image, free gas was not very evident in the chest radiograph but became very clear in the left lateral decubitus view.
Pneumoretroperitoneum is always abnormal and has a relatively few differential diagnoses. Pneumoretroperitoneum is best appreciated by CT, however, can also be detected by plain abdominal radiograph and even by transabdominal ultrasound. Generally, the air is most commonly seen surrounding the kidneys in the right and left upper quadrants of the abdomen. There may also be a loss of the normal psoas muscle shadow and aorta may be lined with air [Figure 4] and [Figure 5]. However, the detection of pneumoretroperitoneum in radiographs is not conclusive of perforation and should be confirmed with CECT before planning any operative procedure. In the present study, three cases of pneumoretroperitoneum were confirmed with CECT abdomen before exploration.
Various named signs were found in the routinely taken radiographs. Amongst them, the most common and diagnostic sign was the crescent sign which was seen in 34 patients out of 42 diagnosed to have a perforation in radiograph. The continuous diaphragm sign [Figure 6] and Riglers sign [Figure 7] were seen in an adequately exposed radiograph. Very rarely we picked the football sign in massive pneumoperitoneum cases [Figure 8].
In the 54 patients, who had GI perforation and 42 were diagnosed in a plain radiograph. Rest 14 required CECT for diagnosis. There was one single case reported with radiograph showing pneumoperitoneum and the patient without any breach in the GI tract. Thirty patients with perforation had a left lateral decubitus view which showed free air only in 23 cases. Hence, the sensitivity of left lateral decubitus was around 76.6%. Overall, the sensitivity of plain radiograph film in diagnosing perforation was 77% with specificity 83%, PPV 97.6%, NPV 29.4% and an accuracy of 76.66%.
| Discussion | | |
'In a patient who develops sudden onset of severe generalized abdominal pain due to the irritant effect of gastric acid on the peritoneum, bacterial peritonitis supervenes over few hours usually accompanied by deterioration in patients' clinical condition. Initially, patient may be shocked with tachycardia, but pyrexia settles in late. The abdomen exhibits board-like rigidity and does not move with respiration'.[2]
The present study aimed at determining the diagnostic validity of radiographs in the detection of perforation by finding the sensitivity, specificity and positive predictive value and analysing the role of left lateral decubitus view in diagnosis.
There are some factors that have widely influenced the role of radiographs in detection of perforation, time of reporting and adequate exposure. The present study was conducted amongst sixty patients with peritonitis suspected to have perforation, in which 52 were males. Hence, the incidence of perforation is five times more in males than females. Out of these, 54 patients in 60 who was diagnosed to have perforation traumatic pathology was found only in 14.8% of people. Almost all patients had abdominal pain at some point during their presentation. Another common symptom was vomiting. Eighty-five percentage of patients had rigidity as a clinical sign and 71.7% presented with tachycardia. Nearly 26.7% of patients were unstable at the time of presentation. The most common risk factor associated with perforation was smoking and about 61.1% of people were smokers.
The results obtained from the study were compared with the results of three studies – Shekhar and Guntpalli-index study,[3] Roh et al.[4] and Woodring and Haiser.[5]
The results obtained from contingency matrix tables of three studies showed that the first study by Shekhar et al.,[3] showed sensitivity of 83.92% and specificity of 75% with PPV value of 97.91 and negative predictive value of 25%. The most common site of perforation was stomach (48.33%), followed by duodenum 20%, small intestine 20% and colon 5%. The study also compared the sensitivity of radiographs with USG and analyzed. It was found that, even though USG is a valuable tool in the detection of pneumoperitoneum, radiograph has more sensitivity.
In the second study, Roh et al.[4] viewed records of 89 patients suspected to have a perforation. In those, 82 patients had perforation. However, radiographs taken during the 1970s, which may be the conventional ones only revealed 42 out of 82 perforations. That is in about 51%. In the study, seven patients with pneumoperitoneum had so signs of visceral perforation even intra-operative. In this study, 46% of patients had gastroduodenal perforation, 20% small intestine and 34% was colon.
In the third study, Woodring and Haiser[5] evaluated 100 patients with known pneumoperitoneum. The upright lateral radiograph showed pneumoperitoneum in 98% of cases and PA chest radiograph showed free air in 80% of cases. The left lateral radiograph picked all seven patients with pneumoperitoneum due to perforated viscus, but the chest radiograph PA view missed two cases. Sixty-one cases both view showed pneumoperitoneum equally, in 12, the lateral chest radiograph showed free air better than the PA chest radiograph did; in five cases, the PA chest radiograph showed free air better than the lateral chest radiograph did. The greater sensitivity of the upright lateral chest radiograph in showing pneumoperitoneum compared with the upright PA chest radiograph was found to be statistically significant (P < 0.01).
| Conclusion | | |
In this study conducted among 60 patients with peritonitis, the sensitivity of plain radiograph chest PA view to detect pneumoperitoneum and thereby perforation was found to be about 77% with a specificity of 83%, the positive predictive value of about 97.6%, and accuracy of 79.66%. The left lateral decubitus view showed the pneumoperitoneum clearer than chest radiograph PA view with sensitivity similar to chest radiographs of about 76.6%. It is concluded that early diagnosis of perforation from the radiograph is of foremost importance as the management should be initiated as soon as possible before the setting of bacterial peritonitis and sepsis.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Lagoo J, Pappas TN, Perez A. A relic or still relevant: The narrowing role for vagotomy in the treatment of peptic ulcer disease. Am J Surg 2014;207:120-6.  |
| 2. | Rosenblatt E. Peritonitis. In. Infectious Diseases Emergencies. Eds Chanmugam AS, Rothman R, Desai S, Putman S. Oxford University Press. 2016; DOI: 10.1093/med/9780199976805.001.0001. |
| 3. | Shekhar R, Guntupalli S, Inamdar PS. Comparative study of ultrasonography with plain radiography in the diagnosis of pneumoperitoneum. IJSS J Surg 2017;3:76-80. |
| 4. | Roh JJ, Thompson JS, Harned RK, Hodgson PE. Value of pneumoperitoneum in the diagnosis of visceral perforation. Am J Surg 1983;146:830-3. |
| 5. | Woodring JH, Heiser MJ. Detection of pneumoperitoneum on chest radiographs: Comparison of upright lateral and posteroanterior projections. AJR Am J Roentgenol 1995;165:45-7. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8]
[Table 1], [Table 2], [Table 3]
|
Search Pubmed for