ORIGINAL ARTICLE

Effectiveness of plain radiography in diagnosing hollowviscus perforation: study of 1,723 patientsof perforation peritonitis

Jyoti Bansal & Raj Kamal Jenaw & Jagdeep Rao &

Jeevan Kankaria & Nilesh N. Agrawal

Received: 19 September 2011 /Accepted: 23 November 2011 /Published online: 6 December 2011# Am Soc Emergency Radiol 2011

Abstract Gastrointestinal perforations remain the mostcommon cause of surgical pneumoperitoneum since timeimmemorial. The aim of this study was to find out theeffectiveness of plain radiography in diagnosing hollowviscous perforation. A prospective analysis of a total of1,723 patients of perforation peritonitis between January2009 and June 2011, confirmed by exploratory laparotomy,was worked out in the study. All these patients had under-gone either an upright chest or erect abdominal or bothradiographs before undergoing operative procedure. Pneu-moperitoneum was evaluated, and the findings were com-pared with that of exploratory laparotomy. Out of the 1,723patients of documented perforation on intraoperative find-ing, 1,537 patients showed pneumoperitoneum on preoper-ative plain radiography. The overall positivity rate of plainradiography in detecting pneumoperitoneum was 89.20%.The positivity rate was highest for stomach and duode-nal perforation (94.19%) and the least for appendicularperforation (7.69%) with highly significant difference(p value, <0.001). In developing world, where there islimited availability of resources and overburden of patients,imposing a limitation in adapting advanced radiologicaltechnique as a first line of investigation, plain radiographymay be considered as a valuable screening tool in detectingpneumoperitoneum with high positivity rate.

Keywords Efficacy . Sensitivity . Perforation peritonitis .

Pneumoperitoneum . Gas under diaphragm . GUD

Background

Gastrointestinal perforations remain the most commoncause of surgical pneumoperitoneum since time immemori-al. Perforation is said to occur when pathology has breachedthrough all the layers of the hollow viscus with resultantescape of intraluminal content into the peritoneal cavity andperitoneal contamination. This necessitates the patient toattend the emergency department usually with the featuresof peritonitis either localized or generalized.

It has been stated that the presence of pneumoperitoneumreflects visceral perforation in 85% to 90% of all occurrence[1–3], and only 50% to 70% cases of hollow viscous perfo-rations show presence of pneumoperitoneum at first instanceon plain radiography [4–7]. Sensitivity varied from 50% to98%, depending upon the type of radiograph that has beencaptured (upright chest, erect abdomen, left lateral decubitus,supine abdomen film) and an additional postural manoeuvrewas taken into account to increase the sensitivity in detectingpneumoperitoneum [5–11]. With the advent of newer diag-nostic modalities, the value of plain radiograph is overshad-owed to detect the pneumoperitoneum in current literature.A number of researches have shown better sensitivity andspecificity of ultrasonography (USG) [12–15] and computedtomography scan [16–18] in detecting pneumoperitoneumalong with their site of perforation in some cases.

The aim of present study was to evaluate hollow viscusperforation patients, those operated as provisional diagnosis ofperforation peritonitis and documented perforation intraoper-atively, showing pneumoperitoneum or not in preoperativeupright chest and/or plain erect abdominal roentgenogram.

J. Bansal (*) : R. K. Jenaw : J. Rao : J. Kankaria :N. N. AgrawalGeneral Surgery Department, Sawai Man Singh Medical College,Jaipur, Rajasthan, Indiae-mail: jyotibbansal@gmail.com

J. Bansal105, Trimurthy Apartment, D-185, Bhrigu Marg,Kanti Chandra Road, Bani Park,Jaipur 302016, India

Emerg Radiol (2012) 19:115–119DOI 10.1007/s10140-011-1007-y

So that plain radiography may be continued as a standard partof preoperative assessment of those patients suspicious ofperforation as a first line of investigation in developing coun-try with limited resource availability.

Material and methodology

Between January 2009 and June 2011, patients admitted andreferred to the general surgery department of Sawai ManSingh Hospital, Jaipur were prospectively analyzed. Patientscoming in casualties or in the outpatient department withcomplaint of abdominal pain, with or without abdominaldistension or vomiting, were ordered an upright chest and/orerect abdominal radiograph. Few patients that had beenreferred to the general surgery department, during theirhospital stay in another department for some other illness,underwent roentgenogram later on. This was the standardpart of investigation for all patients without fail. After inter-relating complaints, positive clinical findings, and results ofradiographs, a decision was made either to admit the patientor not. If radiographs were suggestive of or came out withgas under the right dome of the diaphragm, a patient wasadmitted or transferred to the general surgery departmentimmediately for further line of management. When no ab-normality was appreciable on radiographs but the patient’sgeneral condition and clinical finding necessitated furtherdiagnostic workup, they were also taken into considerationfor close observation and further workup.

All patients were assessed and evaluated by a seniorconsultant surgeon of the general surgery department and adecision for operative procedure was taken by them. Onlythose patients in the study who had confirmed hollow vis-cous perforation at the time of exploratory laparotomy wereincluded. Patients of whom no perforation was found onexploratory laparotomy irrespective of either pneumoperito-neum was present or not on radiography were also excludedfrom study.

Inclusion criteria:

1. Patients age >12 years2. Patient evident of hollow viscus perforation resulting

from underlying disease process3. Patients of blunt trauma abdomen (BTA)4. Patients of therapeutic and/or diagnostic gastrointestinal

endoscopic procedure related perforation

Exclusion criteria (irrespective of whether pneumoperi-toneum was present or not):

1. Patient age ≤12 years2. Patient of gunshot injury or penetrating injury to the

abdomen3. Patients who had undergone laparotomy or laparoscopic

procedure in the previous 30-day period

4. Patient who had intraoperative iatrogenic perforation5. Patient of nonsurgical pneumoperitoneum6. Female who encountered genital injury either during

diagnostic or therapeutic procedure7. Those who had no preoperative radiographs

For analysis, perforations were categorized in two ways:first as per site of perforation and second as per aetiology ofperforation. Sites were subclassified as stomach and duode-num, jejunum, ileum, appendix, and colon. On an aetiolog-ical basis, they were subclassified into BTA, iatrogenicinjury (diagnostic or therapeutic endoscopic procedure),inflammatory (peptic, enteric, appendicular, tubercular, gan-grenous, ulcerative colitis etc.), malignant and other, orunknown category.

Desired information about the finally selected cases werecollected and entered in Microsoft Word Office Excel 2007.Data thus collected was classified and analyzed with thehelp of Microsoft Word Office Excel 2007 and PrimerStatistical Software. Proportions were analyzed using chi-square test. Level of significance was decided with the helpof Primer Statistical Software as per p value, i.e., if the pvalue was ≥0.05 then the difference was not significant, ifthe p value was <0.05 then the difference was significant,and if the p value was <0.001 then the difference was highlysignificant.

Observations

A total of 1,723 patients were incorporated for final analy-sis. Out of this, 286 (16.60%) were females and 1,437(83.40%) were males. The mean age of the patients was39.19±17.08 years. Out of the 1,723 patients of documentedperforation on intraoperative finding, 1,537 patients showedpneumoperitoneum on preoperative radiography either up-right chest or erect abdomen or both. So, the overall posi-tivity rate of the plain radiography in detecting gas under thediaphragm (GUD) was observed to be 89.20%. The rest,10.80%, i.e., 186 patients failed to show pneumoperitoneumon radiography and were found with hollow viscus perfora-tion on exploratory laparotomy).

The most common site of perforation was the stomachand duodenum (49.91%). The positivity rate of plain radi-ography in detecting pneumoperitoneum for this site wasfound to be 94.19% (Table 1). The most common underly-ing aetiology remained at this site was peptic disease, i.e.,stomach and the first part of the duodenum (98.37% of totalperforation of the stomach and duodenum) with positivityrate of radiography was 95.15% (Table 2).

The second most common site of perforation was theileum with a positivity rate of the plain radiography foundto be 91.35% (Table 1). The most common cause of

116 Emerg Radiol (2012) 19:115–119

perforation at this site remained an enteric (typhoid) perfo-ration (93.16% of the total ileal perforations) with a positiv-ity rate of 92.52% (Table 2). The positivity rate for plainradiography in detecting perforation at other sites (appen-dicular, jejunal, and colonic) was quite low, but the perfo-ration at these sites constituted only 5% of the total studypopulation (Table 1).

This difference in the positivity rate for plain radiographyamong different sites of perforation was highly significantly(p value, <0.001). The positivity rate was more than 90% indetecting the ileal, stomach, and duodenal perforation, asthis site constituted 94.89% of the total number of patients.In detecting pneumoperitoneum for jejunal perforation, itwas not as effective as the previous one (positivity rate,76.92%), but in the case of appendicular and colonic perfo-ration, its role is doubtful (Fig. 1).

Based on aetiology, inflammatory pathology remainedthe predominant mode of perforation that was observed in95.82% patients. The positivity rate for plain radiographywas found to be 89.95% (Table 3). A difference in positivityrate for plain radiography among the different aetiologies ofperforation was also observed to be highly significant

(p value, <0.001). Positivity rate was almost 90% for in-flammatory hollow viscus perforation so it can be used as aneffective diagnostic tool in that. In detecting BTA and ma-lignant perforation, it was quite effective (positivity rate,74.62% and 75%, respectively), but in the case of iatrogenicperforation, its role was doubtful. Among inflammatorypathology, peptic, enteric, and appendicular perforation con-stituted almost 95% of the total number of patients, andpositivity rate of radiography to detect GUD remained alto-gether 90.5% for them (p value, <0.0001; Table 2).

A difference in positivity rate for plain radiography indetecting perforation was found just significant as per the age(p value00.013; Table 4), but it was not found significant at

Table 1 Association of gas under diaphragm with site of perforation

Serial no. Site ofperforation

Gas underdiaphragm

Total patients(%)

Positivityrate (%)

Absent Present

1. Stomach andduodedum

50 810 860 (49.91) 94.19

2. Jejunal 3 10 13 (0.75) 76.92

3. Ileal 67 708 775 (44.98) 91.35

4. Appendix 60 5 65 (3.77) 7.69

5. Colonic 6 4 10 (0.58) 40.00

Grand total 186 1,537 1,723 (100) 89.20

Chi-square test value 501.54 at DF, 4; p value<0.0001; LS 0 HS

DF degree of freedom, LS level of significance, HS highly significant

Table 2 Various aetiology of inflammatory hollow viscus perforation

Serial no. Etiology Gas underdiaphragm

Totalpatients (%)

Positivityrate (%)

Absent Present

1. Peptic 41 805 846 (51.24) 95.15

2. Enteric 54 668 722 (43.73) 92.52

3 Appendicular 60 5 65 (3.93) 7.69

4. Tubercular 7 5 12 (0.7) 41.66

5. Miscellaneous 4 2 6 (0.36) 33.33

Grand total 166 1,485 1,651 89.95

Chi-square test value 518.48 at DF, 4; p value<0.0001; LS 0 HS

DF degree of freedom, LS level of significance, HS highly significant

Table 3 Association of gas under diaphragm with aetiology ofperforation

Serialno.

Mode ofperforation

Gas underdiaphragm

Total patients(%)

Positivityrate (%)

Absent Present

1. Inflammatory 166 1,485 1,651 (95.82) 89.95

2. Blunt traumaabdomen

16 47 63 (3.66) 74.6

3. Iatrogenic 3 1 4 (0.23) 25

4. Malignancy 1 3 4 (0.23) 75

5. Other/unknown 0 1 1 (0.06) 100

Grand total 186 1,537 1,723 89.20

Chi-square test value 35.88 at DF, 4; p value<0.0001; LS 0 HS

DF degree of freedom, LS level of significance, HS highly significant

Fig. 1 Positivity rate of gas under diaphragm among different sites ofperforation

Emerg Radiol (2012) 19:115–119 117

all as per the sex (p value00.62; Table 5). The positivity rateor radiography to detect pneumoperitoneum varies from86.74% to 95.94% and increases as the age advances.

Discussion

The overall positivity rate of plain radiography in detectingpneumoperitoneum for hollow viscus perforation was89.20%. Whatever site or aetiology remained behind theperforation, the positivity of plain radiography in detectingpneumoperitoneum was well above 90%. Taj MohammadKhan et al. [19] demonstrated that GUD was present only in60% the cases (84 out of 140) of typhoid perforation. USGhas shown a sensitivity of 92% [14] and 85% [20] indetecting pneumoperitoneum in patients of acute abdomenand victim of BTA, but the major drawback of both thesestudies was that results were totally operator dependent. Inthe first study, USG was done by a blinded ultrasonographerwho was either a staff surgeon or a staff emergency physician,while in the second study, examinations were performed by ageneral surgeon experienced in gastrointestinal surgery, othergeneral surgery, or trauma surgery. Here, it can be observedthat plain radiography is well comparable (89% vs 85% or92%) with no significant difference (p value, >0.05). Anotherchoice is to go with CT scan to increase the sensitivity todetect pneumoperitoneum that will neither be cost effective

nor safe (unnecessary 35 times extra radiation exposure).This will also lead to more dependency on diagnostic mo-dalities and loss of clinical acumens.

Our aim of this study was not to compare advancedimaging modality with conventional plain radiography butto evaluate and to provide evidence of how well plainradiography performs as a screening test, particularly in adeveloping country where resource availability is limited (interms of both trained personnel and advanced imaging at theprimary health care level). As we all know, today’s practicein medicine is an “Evidence Based Practice,” and the plainradiography was found significantly satisfactory in detectinghollow viscus perforation as a first line investigation forpatients suspicious of perforation peritonitis. Plain radiog-raphy is less operator dependent, cost effective, and lessharmful than CT scan and can be made available at theprimary health care level very easily. Of the cases, 10.80%that were not showing pneumoperitoneum indicates that thisproportion of patients can be subjected to advanced imagingmodality for definitive preoperative diagnosis rather thanoffering higher investigation for all patients.

Absence of GUD might be because colon and duodenum(second to fourth part) is retroperitoneal in most of their partand evacuation of air from there into the peritoneal cavity isquite impossible until there is breach in the retroperitoneum.In the case of appendicular perforation, the omentum andsurrounding bowel loops adhere to it. Owing to a smalllumen, the volume of air which can remain in it is very smallso that, even if air escapes from the appendicular perforation,it is very small in amount and easily gets trapped within theadhered omentum and bowel loop.

The absence of pneumoperitoneum, in those cases ofhollow viscous perforation covered by the peritoneum, canbe a result of an insufficient amount of air leak to be detectedon roentgenogram, delayed presentation to the hospital dur-ing which period escaped gas might get absorbed, adherenceof perforation from surrounding viscus, obliteration of sub-diaphragmatic space by a pathological process [21], adhe-sion, absence of intraluminal gas at the site of perforation,filling of lumen of perforated viscous by fluid, the pluggingof the perforation by food or redundant mucosa, sealing ofthe perforation by the omentum or peritoneum before theescape of intraluminal gas in an amount which could bedemonstrated by the radiological method [8].

Conclusion

In a developing world where limited availability of resour-ces and an overburden of patients impose a limitation inadapting advanced radiological technique as a first line ofinvestigation, plain radiography either upright chest X-rayor erect abdomen should be considered as a valuable

Table 4 Association of gas under diaphragm with age of patient

Serial no. Age group(years)

Gas underdiaphragm

Totalpatients (%)

Positivityrate (%)

Absent Present

1. >12–30 90 599 689 (39.99) 86.94

2. 31–50 66 537 603 (35.00) 89.05

3. 51–70 28 356 384 (22.29) 92.71

4. >70 2 45 47 (2.73) 95.74

Grand total 186 1,537 1,723 89.20

Chi-square test value 10.67 at DF, 3; p value, 0.017; LS0S

DF degree of freedom, LS level of significance, S significant

Table 5 Association of gas under diaphragm with sex of patients

Serial no. Sex Gas underdiaphragm

Total patients(%)

PositivityRate (%)

Absent Present

1. Male 158 1,279 1,437 (83.40) 89.00

2. Female 28 258 286 (16.60) 90.21

Grand total 186 1,537 1,723 89.20

Chi-Square test value 0.245 at DF, 1; p value, 0.620; LS 0 NS

DF degree of freedom, LS level of significance, NS not significant

118 Emerg Radiol (2012) 19:115–119

screening tool in detecting pneumoperitoneum with a posi-tivity rate of 89.20% in detecting hollow viscous perforationwhich will be more cost effective, better interpretative, lessoperator dependent, less harmful than CT scan and can bemade easily available at primary health care level.

Acknowledgments I gratefully acknowledge IASG (Indian Associ-ation of Surgical Gastroenterology) for accepting this work as theposter presentation in IASGCON 2011 and it has been awarded secondprize. I am deeply grateful to Dr. Kusum Gaur, M.D., Department ofPreventive and Social Medicine, SMS Medical College, Jaipur for theirvaluable support in the statistical analysis.

References

1. McGlone FB, Vivion CG, Meir L (1966) Spontaneous pneumo-peritoneum. Gastroenterology 51(3):393–398

2. Roh JJ, Thompson JS, Harned RK et al (1983) Value of pneumo-peritoneum in the diagnosis of visceral perforation. Am J Surg146:830–833

3. Winek TG, Mosely HS, Grout G, et al (19880 Pneumoperitoneumand its association with ruptured abdominal viscus. Arch Surg123:709–712

4. Cho KC, Baker SR (1994) Extraluminal air: diagnosis and signif-icance. Radiol Clin North Am 32:829–844

5. Levine MS, Scheiner JD, Rubesin SE, Laufer I, Herlinger H (1991)Diagnosis of pneumoperitoneum on supine abdominal radio-graphs. AJR Am J Roentgenol 156:731–735

6. Phatak MG, Frank SJ, Ellis JJ (1984) Computed tomography ofbowel perforation. Gastrointest Radiol 9:133–135

7. Rice RP, Thompson WM, Gedgaudas RK (1982) The diagnosisand significance of extraluminal gas in the abdomen. Radiol ClinNorth Am 20:819–837

8. Miller RE, Nelson SW (1971) The roentgenologic demonstrationof tiny amounts of free intraperitoneal gas: experimental andclinical studies. AJR 112:574–585

9. Paster SB, Brogdon BG (1976) Aoentgenographic diagnosis ofpneumoperitoneum. JAMA 235:1264–1267

10. Shafer HA Jr (1992) Perforation and obstruction of the gastroin-testinal tract: assessment by conventional radiology. Radiol ClinNorth Am 30:405–426

11. Woodring JH, Heiser MJ (1995) Detection of pneumoperitoneumon chestradiographs: comparison of upright lateral and posteroan-terior projections. Am J Roentgenol 165:45–47

12. Chen CH, Yang CC, Yeh YH (2001) Role of upright chestradiography and ultrasonography in demonstrating free air ofperforated peptic ulcers. Hepatogastroenterology 48(40):1082–1084

13. Chen SC, Wang HP, Chen WJ et al (2002) Selective use ofultrasonography for the detection of pneumoperitoneum. AcadEmerg Med 9(6):643–645

14. Chen S-C, Yen Z-S, Wang H-P, Lin F-Y, Hsu C-Y, Chen W-J(2002) Ultrasonography is superior to plain radiography in thediagnosis of pneumoperitoneum. Br J Surg 89(3):351–354

15. Karahan OI, Kurt A, Yikilmaz A, Kahriman G (2004) New methodfor the detection of intraperitoneal free air by sonography: scissorsmaneuver. J Clin Ultrasound 32(8):381–385

16. Stapakis JC, Thickman D (1992) Diagnosis of pneumoperitoneum:abdominal CT vs upright chest film. J Comput Assist Tomogr16:713–716

17. Earls JP, Dachman AH, Colon E, Garrett MG, Molloy M (1993)Prevalence and duration of postoperative pneumoperitoneum: sen-sitivity of CT vs left lateral decubitus radiography. AJR 161:781–785

18. Hainaux B, Agneessens E, Bertinotti R, De Maertelaer V, RubesovaE, Capelluto E et al (2006) Accuracy of MDCT in predicting site ofgastrointestinal tract perforation. AJR Am J Roentgenol 187:1179–1183

19. Mohammad Khan T et al (1997) Experience with typhoid ilealperforation. JPMI 11(2):175–181

20. Moriwaki Y, Sugiyama M et al (2009) Ultrasonography for thediagnosis of intraperitoneal free air in chest–abdominal–pelvicblunt trauma and critical acute abdominal pain. Arch Surg 144(2):137–141

21. Williams N, Everson NW (1997) Radiological confirmation ofintraperitoneal free gas. Ann R Coll Surg Engl 79:8–12

Emerg Radiol (2012) 19:115–119 119