A review of the management of perforated duodenal ulcers at a tertiary hospital in south western Nigeria

Background: Gastro-duodenal perforations are common and may complicate peptic ulcer disease. Management is often by surgical closure. Objective: To determine the patterns of presentation and mode of management of duodenal ulcer perforations. Methods: Retrospective review of patients with duodenal ulcer perforations seen at the Obafemi Awolowo University Teaching Hospital between June 2001 and July 2011. Patients’ records were reviewed for demography, duration of disease, probable risk factors, type of surgery and complications. Data obtained was analyzed using SPSS 15.0. Result: Fortyfive patients were reviewed. There were 37 males (82.2%). Mean age was 39.7years (range 15-78years). There were 10 (22.6%) students and 8(17.8%) farmers. NSAIDs abuse (11), previous peptic ulcer disease (2), and no prior dyspeptic symptoms (20) constituted 24.4%, 4.4% and 44.4% respectively of cases. Seven (16%) patients presented less than 24 hours of onset of illness. Forty one perforations (91.1%) involved the first part of duodenum. Twenty two (49%) patients had Graham’s omental patch. We had one (2.2%) failed repair and six (13.3%) mortalities. Conclusion: Late presentation of duodenal ulcer perforation is common with high mortality. Pragmatic surgical intervention with Graham’s omentopexy with broad spectrum antibiotics is still commonly practiced.


Introduction
Gastro-duodenal perforations are common in surgical practice and do occur as a complication of peptic ulcer disease (PUD), abuse of non-steroidal anti-inflammatory drugs (NSAIDs) and gastric cancer [1][2][3][4][5][6] .Alcoholics and smokers are at higher risk 7,8 .Management is quite challenging as patients present late; with septicaemia, fluid and electrolyte derangements, shock and/or systemic inflammatory response syndrome.
Three decades of advances in drug treatment of peptic ulcer disease (PUD) has led to less need of elective surgery 9 .Emergency surgical treatment is often reserved for complications of the disease and not necessarily to cure the ulcer.These complication includes upper gastrointestinal tract (GIT) bleeding; perforation of the stomach, duodenum or sites of ectopic acid production; and gastric outlet obstruction.Helicobacter pylori (H.pylori) infection has come to play a pivotal role in the aetiopathogenesis of the disease 10 and its eradication is associated with better prognosis 11 .
This study aimed at determining the patterns of presentation and methods of management of duodenal ulcer perforation and the outcome, at the Obafemi Awolowo Teaching Hospital Complex, south western Nigeria over, a 10year period.

Methods
The names and hospital identification numbers of all patients who had surgery for duodenal ulcer perforation between June 2001 and July 2011 were extracted from the theatre register.With this list, the case files of these patients were retrieved from the medical records department of the hospital.
The files were reviewed for patients' bio data (sex, age, occupation and tribe), history of regular use of non-steroidal anti-inflammatory medications, duration of dyspeptic symptom, endoscopic findings (if previously done prior to occurrence of the perforation), the symptoms each patient presented with, the interval between the time of diagnosis and time of surgery; the pre-operative medications administered to each patient prior to the surgery; the operative findings and the mode of closure of the perforation.In addition, number of days spent in the post-operative period before discharge and occurrence of any post-operative complications were also reviewed.Note is made of any re-operation.Their take-home medications, duration of their follow-up, any long term complications and post-operative endoscopic findings after discharge were also noted.
The data retrieved were entered on a spread sheet and analyzed for frequencies using the statistical packages for the social science (SPSS 15.0).Ethical clearance was obtained from the Research and Ethics Committee of the teaching hospital.

Results
A total number of 60 patients were treated within the period only 45 had complete data for the review.There were 37 males (82.2%) and 8 females (17.8%).Their ages ranged from 15years to 78years with a mean age of 39.7years.Majority of the patient were in age group 21-44years (table1).Sixteen patients (35.6%) presented in shock states with six of them having features of systemic inflammatory response syndrome.Of the latter, one was less than 45years old three were 45-64years while the remaining two were above 64years.Thirty six (80%) patients were classified, using the American Society of Anaesthesiologist (ASA), as ASA 4E while nine (20%) were classified as ASA 5E.Most patients (37, 82.2%) had surgery within 24hours of diagnosis and admission while six (13.3%) and two (4.4%) respectively had surgery on the second and third day post diagnosis.
Intra-operative findings included perforation in anterior aspect of first part of the duodenum in 41(91.9%) of cases; with size of the perforations ranging from 0.5cm -2cm in widest diameter.Twenty two (49%) patients had Graham's omental patch repair, 13(29%) had simple closure with omental reinforcement while 10(22%) had just simple closure.One (2.2%) patient had pyloroplasty.None of the patients had anti-ulcer surgery.Thirteen (29%) patients had any form of medical treatment for PUD before surgery while 14(31.1%)had the treatment after surgery.Parenteral ciprofloxacin and metronidazole were the drugs of choice in the perioperative period.
Post-operative hospital stay (table 3) ranged from 6 days to 33 days, with about 16(35.6%)patients spending less than 10 days.Four patients were re-operated (for intra-abdominal collections and failed primary repair).Follow-up after discharge from the hospital ranged from 1 week to 2 years with an average of 24 weeks.Only 2 patients had post-operative upper G.I. endoscopy as a follow-up investigation, which was normal.

Discussion
Peptic ulcers are circumscribed loss of epithelial tissue of the gastro-intestinal tract due to corrosive effect of the gastric juice.They occur when there is a breech in the normal mucosal barrier, when excessive acid is produced, or normal gastric juice comes in contact with a mucosa not adapted to its effects.Breech in mucosa barrier may be from Helicobacter pylori (H.pylori) infection, diet (alcohol), social habits (smoking) or drugs (NSAIDs).Cigarette smoking and the indiscriminate use of ulcerogenic drugs (like non-steroidal anti-inflammatory drugs and corticosteroids) have be implicated as probable causes in as much as 75% and 30% respectively 1 .Additional risk factors include advanced age, male gender, alcohol abuse, and medically debilitating comorbidities 12,13 .Duodenal perforation is a common complication of PUD.This study shows a prevalence rate of about 4-5cases per year in the study period.This is similar to observation earlier in Ile-Ife 14 and other centres 15,16 .This may be attributable to indiscriminate use of antibiotics and proton pump inhibitor (PPI) which are effective in ulcer care; and are common over the counter (OTC) drugs in our environment.There has been a steady increase in number of private health facilities, some of them offering specialist surgery and medical services.However, in general, the incidence rate of perforated duodenal ulcer appears not to have changed 14,17 .It is more common in males than females [17][18][19] .In our study, it was five times more common in males than females, and this is similar to findings in previous reviews 4,2 , but at variance with findings by Watt et al 20 .It appears to be a disease of the young and middle especially males, as our study suggests (93.3%).We observed that 27(60%) and 15(33.3%)were in the young and middle age groups respectively; with mean age between 39.7years: collaborating earlier works 14,[16][17][18][19] .However Ohene-Yeboah 21 reported a mean age of 64.8years.It was also seen in women because of increasing rate of smoking amongst them (particularly elderly women): this however is not common in our environment.
Also implicated in aetiopathogenesis is the use of NSAIDs in all the age groups 1,2,20,22,23 .
Students formed the modal occupation group; it is still not clear if the was a chance finding, or what the risk factors in them were.It may however, be attributed to use of ulcerogenic substances like xanthine and hypoxanthine present in coffee 24 commonly used as stimulant for reading.This contrasts with earlier findings, in this community, of higher prevalence amongst working class individuals 14 .Consistent with findings elsewhere, fewer patients had dyspepsia prior to presentation compared to those who did not 14,15,25 .Use of ulcerogenic drugs like NSAIDS was low compared to the reports from Ghana 2, 21 .This may be due to the younger age groups in our study: long term NSAID use is common in the elderly for care of osteoarthritis.Abdominal pain was the most common presenting complaint.
We found a sizeable proportion having centrally located pain; this may be due to gastromegaly which has been noted in patients with chronic duodenal ulcer 26 .Seven (15.6%) patients presented within 24hours from time of onset of symptom(s); this is low when compared with findings of 46.2% and 47.3% elsewhere 2,15,19 .Thirty seven (82.2%) patients had surgery on or before the 24 th hour on admission which is fair compared to a study in southwest Nigeria 14 but a far cry from better result by Bin-Taleeb et al 19 .Surgical care of PUD used to be fashionable in times past, but has been relegated to the background in the face of great success with medical therapy (mainly PPIs and H. pylori eradication protocols).Surgery in now required for complications like perforation (closure or omental plug), gastric outlet obstruction (drainage) and massive haemorrhage (under-running of bleeder) not amenable to endoscopic intervention.C.J. Cellan-Jones in 1929 reported 51 cases of perforated duodenal ulcer treated with a patch of omentum sutured over the perforation 27 .Later, in 1937, Roscoe Graham of Toronto reported 51 cases treated with omental patch closure in essentially the same fashion, though his technique was a transverse rather than a longitudinal positioning of the omental patch (by Cellan-Jones) along the long axis of the duodenal perforation, a modification thought to reduce the incidence of duodenal stenosis 28 .Dragstedt, in 1949, came up with the finding that vagotomy could reduce gastric acid, which paralleled a significant reduction in peptic ulcer diseases 29 .However the addition of a definitive acid-reducing procedure at same surgery with the repair of duodenal perforation seems to add to the risk for mortality and morbidity without appreciably improving the long-term outcome. 9Gupta et al identified three distinct types of perforated duodenal ulcers 4 .The first are small perforations that are easy to manage with low morbidity and mortality.The second are "large" perforations that are not uncommon, and which still have the best of results even with omental patch repair.The third are "giant", which exceed 3cm in diameter, and are extremely uncommon.In an emergency setting, the patients are often critically ill and definitive anti-acid procedure along with closure of perforation is not advisable.Omental patch repair is simple, and still remains dependable for closure of even large perforations up to 3cm in size 4 .Patch closure alone is enough for duodenal perforation in light of the low recurrence rate of PUD 16 especially when H. pylori eradication is achieved.It has been noted that eradicating H. pylori infection reduced significantly ulcer recurrence at a year later following repair of perforation 11 .
No record was seen of attempts at H. pylori eradication therapy in our study.Haphazard attempts were made at eradication, though not documented as such: as antibiotics were given alongside cimetidine, ranitidine or omeprazole.By Gupta's classification our patients had either type 1 or type 2, but none had type 3 perforation.The mortality rate of 13.3% (6/45) is higher than reported by Dakubo, Plummer, Keremu and Otu 2,18,25,30 but lower than findings by Lawal and Nuhu. 14,15 ve out of the six patients that died had Graham's repair; this result is poor compared to 95.8% success rate by Abid 31 .These patients presented in shock states, with SIRS and preoperatively were classified as ASA 5E.These can be attributed to, amongst other probable causes 32 , late presentation (outside the golden 8 hours when bacteria peritonitis succeeds chemical peritonitis).Post-operative wound infection rate of 18.9% was not un-expected as the wound was classified 'dirty' ab initio.This is again was higher than what Abid et al reported 31 .Association between the duration of the perforation before surgery on the one hand, and period of post-operative stay on the other hand was not statistically significant (P > 0.05) and had a weak correlation (r = 0.283), suggesting that length of post-operative stay for these patients was probably not dependent on the delay before surgery.
Association of duration of the perforation before surgery with post-operative complications such as wound infections was also not statistically significant (P>0.05) with a weak correlation (r = 0.217), suggesting that wound infection occurrence was independent of delay before surgery.However, there was a strong association between the interval from diagnosis to surgery and peri-operative death (P < 0.05; r = 1.000) as was the findings by Barut et al 33 and James Y Lau et al 34 .We had follow-up period ranging from 1week to 2 years (with average of 24 weeks).This is rather short follow up period may probably be due to financial difficulty on the part of our patients to sustain regular hospital follow-up visit.
Our patients were mainly from the low socio-economic class based on their occupation.This will no doubt impacted negatively on available resources to access quality care.As noted earlier, delayed presentation, regular use of NSAIDs and other OTC drugs were common.Though the study did not seek for prevalence of H.pylori, it has been noted to be high amongst the low socio-economic group.Also the erroneous belief amongst patients that there is really no need for regular follow-up once the operative wound has healed with little or no immediate or early post-operative sequels may also account for the low follow-up period.Moreover, despite the availability of endoscopy services, patronage was low.This again could be due to financial constraint on the part of the patients, or clinicians not insisting enough, or probable poor communication between patients and their surgeons.

Limitations
This is a retrospective review and vulnerable to poor record keeping as noted in the incomplete data in 25% of the total patients managed.The rather short mean follow-up period would not allow categorical statement on outcome of care: a randomized prospective study (on-going) would address this.

Conclusion
DU perforation is predominantly a disease of middle aged men.Our patients presented late.Graham's omentopexy or simple closure with omentum reinforcement, were the common repair methods deployed.Utilization of endoscopic services was low within the study period.We recommend a prospective study, preferable a multicentre one, to ascertain the epidemiology, risk and prognostic factors of the disease.Health education aimed at improving health care seeking habit of the population would also be beneficial.

Figure 3 :
Figure 3: Effect of duration of perforation at presentation on duration of hospital stay