Diagnosis, Management and Outcomes of Intestinal Obstruction at an Urban Tertiary Hospital in Sub Saharan Africa: a Cross-Sectional Study.


 BACKGROUND: Intestinal Obstruction (IO) is among the commonest causes of acute abdomen worldwide and globally it remains a challenge because it is a major cause of morbidity and surgical financial expenditure. Clinically it presents with nausea, vomiting, colicky abdominal pain and cessation of bowel movements or passage flatus and stool. Diagnosis can be clinical but is confirmed usually by radiologic imaging. We studied the current diagnosis, management and outcomes of IO in Mulago HospitalMATERIALS AND METHODS: This was a Prospective Descriptive Study in all the surgical units of Mulago from January to May 2014. Ethical approval was got in line with Helsinki declaration and then a pretested and validated questionnaire was used to collect data. Informed consent was got with eligible and consenting/assenting patients recruited among those patients of all ages and sex presenting with suspected Intestinal Obstruction. Uni-variate and bi-variate of the variables plus measurements of associations were done.RESULTS: We enrolled 135 patients, excluded 25 and recruited 110 patient with more males than females i.e. 71.8% males and 28.2% females. Colicky abdominal pain, abdominal distension, and vomiting were the 3 commonest symptoms with abdominal distension, increased bowel sounds and abdominal tenderness as commonest signs. Majority of the patients were diagnosed radiologically (51%) and the rest (48.2%) clinically diagnosed accounting. “Dilated bowel loops” was the most frequent radiological sign. Return of bowel sounds occurred within 5 days of the POD, while opening of bowels on average, occurred on the 3rd POD. The mean day of discharge was the 5th POD and 73% of the patients were discharged by the 7th Post-Operative Day. The commonest unfavourable management outcome noted was prolonged hospital stay followed by wound sepsis (Surgical Site Infection) and MortalityCONCLUSION: Majority of the patients were diagnosed radiologically (51%). surgical management was done for 72.7% of the cases and 27.3% conservatively managed. Prolonged hospital stay was the commonest unfavorable outcome of management

sign. Return of bowel sounds occurred within 5 days of the POD, while opening of bowels on average, occurred on the 3 rd POD. The mean day of discharge was the 5 th POD and 73% of the patients were discharged by the 7 th Post-Operative Day. The commonest unfavourable management outcome noted was prolonged hospital stay followed by wound sepsis (Surgical Site Infection) and Mortality CONCLUSION: Majority of the patients were diagnosed radiologically (51%). surgical management was done for 72.7% of the cases and 27.3% conservatively managed. Prolonged hospital stay was the commonest unfavorable outcome of management Background Intestinal Obstruction (IO) refers to the interruption of forward ow of intestinal contents and can occur at any point from mouth to anal canal with clinical symptoms varying based on the level of obstruction [1].
It is among the commonest causes of acute abdomen worldwide and is characterized by abdominal pain which in 2006 was the most common reason for a visit to the emergency department in the USA accounting for 8 million (7%) of the 119 million visits [2]. IO is a common surgical emergency that is frequently encountered in abdominal surgery [3,4] and it remains a challenge because it is a major cause of morbidity plus nancial expenditure worldwide [5]. Clinical presentation of IO generally includes nausea, vomiting, colicky abdominal pain and cessation of bowel movements or passage atus and stool, although the severity of these clinical symptoms varies based on the acuity and anatomic level of obstruction [1,6]. Diagnosis is con rmed usually by radiologic imaging although, with negative radiography and high index of suspicion, non-contrast Computed Tomography (CT) scan is recommended [1]. Management of uncomplicated obstructions can be conservative by uid resuscitation to correct the metabolic derangements, intestinal decompression, and bowel rest. If there's evidence of vascular compromise or perforation then management is by surgical intervention [6].
Recent studies on IO in Uganda have only been documented in 2 non-governmental hospitals in 2 regions of Uganda [7,8], yet the tertiary health care institution Mulago National Referral and Teaching Hospital (MNRTH) which offers free services to patients doesn't have recent studies with the last documented IO study done by Macadam in 1961 [9]. Also raw data from the Accident and Emergency (A&E) revels that an average of 50 patients are admitted per month with a suspected diagnosis of IO [10].
We studied the current diagnosis, management and outcomes of IO in MNRTH due to the paucity of data to guide current policy on decision making process, resuscitation measures, timing of surgery and choice of surgical procedure in our setting.

Materials And Methods
Study Setting MNRTH, located in Kampala, the Capital City of Uganda, is a 1500 bed tertiary hospital caring for approximately 140,000 patients annually, with an annual average of 48,000 patients transitioning through the A&E [11]. It has three surgical wards, with an A&E that triages patients into elective or emergency cases done in casualty theatre that records about 31 operations monthly for the relief of IO [10]. Also A&E records show that on average about 7 to 10 patients' with IO are seen per week by the Consent was got from all patients and for subjects below 18 years, a parent, legal guardian or Next Of Kin gave consent on their behalf. Eligible and consenting patients were recruited with details of their demographics, symptoms and signs recorded.
Aspects of their resuscitation, investigations and treatment were recorded and monitored till discharge.
The causes of IO, and the early management outcomes for the rst 7 post-operative/post admission days, including deaths or up to discharge, whichever came rst, were recorded.

Study Periods
The study was conducted from January 2014 and May 2014 where we recruited consented patients that met the inclusion criteria.

Study variables
The main variables of interest were clinical examination, radiologic ndings, and management options focusing on favorable and unfavorable outcomes in the Post-Operative Day (POD) window.

Data management and analysis
A structured pretested and validated questionnaire was used to collect data. The data was cleaned, backed up and later analyzed in STATA version10.1. Study statistics were reported using proportions, means, medians and inter-quartile ranges. Uni-variate and bi-variate analysis of the variables plus measurements of associations were done using Pearson Chi-X² values, P values and logistic regression. Presentation of data was by bar graphs, pie charts, Box-plots and Tables.

Results
We recruited 110 patients that met criteria or had outcomes of interest i.e. investigations suggestive of obstruction, or con rmation of diagnosis at laparotomy with favorable and unfavorable outcomes. Demographically there were more males than females i.e. 71.8% males and 28.2% females in a ratio of 2.6:1. The 3 commonest symptoms reported by the participants were; colicky abdominal pain, abdominal distension, and vomiting. The 3 commonest examination signs were; abdominal distension, increased bowel sounds and abdominal tenderness. Majority of the patients were diagnosed radiologically (51%) with clinical diagnosis accounting for 48.2%. However for the 80 cases eventually operated a higher majority (58.8%) were diagnosed clinically. See Table1 0.001* "Dilated bowel loops" was the most frequent sign found on both imaging modalities. Multiple air uid levels were observed in 12.7% of the radiographs and faecal impaction was noted in 5.5% of the x-rays. On Sonography, reversed peristalsis was noted in 10% of the cases. The other signs were seen 5.5% of the cases. Ultra sound Sonography exerted more in uence on the choice of management when compared to plain radiography. (p = 0.001 vs. p = 0.013). Shown in Table 2  Outcomes were categorized as favourable or unfavorable with the favourable outcomes being: return of bowel sounds, opening of bowel, NG-tube removal, drain removal and discharge by the 7th post-operative or post admission day (for non-operative management). The unfavorable outcomes included: wound sepsis, systemic sepsis, anastomotic leak, anemia, chest infections, prolonged hospital stay and death.
For the favourable outcomes; most post-operative events occurred within the rst 10 days. Return of bowel sounds occurred within 5 days of the POD, while opening of bowels on average, occurred on the 3rd POD. The mean day of discharge was the 5th POD and 73% of the patients were discharged by the 7th Post-Operative Day (Fig. 1).
The commonest unfavourable management outcome noted was prolonged hospital stay followed by wound sepsis (Surgical Site Infection) and Mortality (Fig. 2).
Factors that were persistently associated with favorable outcomes, in the Uni-variate model, were fever (RR = 0.19, p = 0.09), Increased bowel sounds (RR = 4.3, p = 0.001). Absent bowel sounds and abdominal tenderness (p = 0.001&p = 0.014) showed association with favorable outcomes at lower odds. Small Bowel Obstruction (SBO) is 3.17 likely to be associated with favorable outcomes (p = 0.013). See Table 3. The choice of management, i.e. non-operative versus operative, greatly in uenced outcome with more of the patients undergoing non-operative management experiencing favourable outcomes.
The results in Table 4 below show the recomputed odds ratio in a multivariate regression and unlike in the binary logistic regression, none of the factors independently predict the outcome, implying that they most likely act in association or in combination with one another, as their Relative Risk changes and the P value becomes statistically insigni cant.

Discussion
We studied the diagnosis, management and early outcomes IO in our local setting at MNRTH. Baseline demographics showed more males than females i.e. 71.8% males and 28.2% females. Radiological imaging was done for 51.8% while 48.2% were diagnosed clinically, this signi cant role of clinical diagnosis is comparable to that in other African studies [12,13,14]. The modality of diagnosis in uenced choice of management and the diagnostic value of radiographs was depicted in the strong statistical correlation between imaging and choice of management (Table 2), with X-rays having a p = 0.013, while Sonography had a p = 0.001. Studies have shown that there is also increasing reliance on radiological investigations when the immediate choice of management is conservative rather than operative because it's shown that a proportion of cases will resolve spontaneously [15].
We noted that Eighty (72.7%) of the participants who presented with IO were managed operatively with resection and anastomosis as the commonest procedure done on 28 (35.0%) followed by hernia repair at 23.8% and then Simple colostomy at 15% as the 3rd commonest procedure and laparotomy with colostomy was the least common procedure in the study. Seventy three (66.4%) cases had favourable outcomes while 37(33.6%) had associated morbidity and mortality. Prolonged hospital stay (30.4%) which was the commonest adverse outcome was comparable to regional studies that had similar ndings [7,8,16,17,18].
The procedure most associated with prolonged hospital stay was resection and anastomosis, followed by ileostomy and we posit that this was so because cases diagnosed as "tumour" had to undergo radiotherapy treatment or staging investigations prior to de nitive management. We also found out that mortality was the 3rd commonest postoperative complication at a frequency of 8.8% which was comparatively low compared to similar studies that showed rates of 12.9%, 19.7% and 20% [8,17,19,20].
Our rates of non-operative management in this study (27.3%) compare differently to those in developed countries like USA (73%) and we postulate that could probably be attributed to the lack of alternative methods of non-operative management such as Gastrogra n use in adhesions, hydrostatic or pneumatic reduction of intussusception, as well as the lack of investigative capacity to con rm or rule out bowel ischemia in our setting [16,21,22].
This study was subject to limitations such as recall bias as some patient data was from records and POD follow up interviews of the operating surgeon or SHO, this may have affected some study variables collected.

Conclusion
In DUPLICATE PUBLICATION: This is an original manuscript whose work is not under consideration elsewhere, copyright has not been breached in seeking publication and the publication has been approved by all co-authors and responsible authorities at MNRTH AUTHOR CONTRIBUTION: Dr. AMN: developed this idea from conception, proposal development, data collection and manuscript writing as this was her thesis for her masters in Surgery.
Dr. SCK and PO(RIP) gave surgical supervisory oversight that helped oversee the proposal development and nal proposal preparation Dr. TSL worked with Dr. AMN worked hand in hand to develop this into a nal proposal through to data collection and manuscript writing for this submission. ACKNOWLEDGEMENT: To all the colleagues in the Department of Surgery at Makerere University whose guidance and instruction was invaluable especially my supervisors for their unwavering commitment.
To all the SHO's graduate colleagues with unique personalities, individual strengths and " ne hands who walked the data collection journey and made it lighter".
The tireless and dedicated staff of Mulago National Referral and Teaching Hospital who despite the odds stacked against them keep on keeping on with working of saving lives