Effects of delayed treatment on perforated peptic ulcers at Kenyatta National Hospital (KNH)

Nasio NA1 MBChb, MMed (Surg), Saidi H1,2 BSc (Anat), MBChB, MMed (Surg), FCS (ECSA), FACS. 1. Egerton University, 2. Department of Human Anatomy, University of Nairobi Correspondence to: Nasio A. Nasio, Email: nasioabner@yahoo.com


Background: Perforations complicate up to 5-10% of peptic ulcer diseases. Mortality following peptic ulcer perforation can peak 29%. Of the factors that influence the outcome of peptic ulcer perforation, treatment delay is most important and modifiable. This study reviewed delay and how it affected outcome in patients treated for perforated peptic ulcers at the Kenyatta National Hospital.


Methods: Patient’s files for the period January 2002 to December 2007 were reviewed and direct interviews carried out for patients seen from January to December 2008. Data sought included patient demographics, clinical presentation, time from symptom onset to presentation at casualty, time from presentation at casualty to surgical treatment and the treatment outcomes. Data was entered using a structured data sheet /questionnaire. The effect of delay as a determi-nant of outcome was evaluated using univariate analysis.

Results: One hundred and ninety three patients were evaluated. Twenty four patients (12.4%) died. Sixty one patients (31.6%) developed complications post-operatively. Thirty patients were re-operated for the complications. No patient treated within 24 hours died. Complications rate was 0 %, 1.5% and 29.5% for patients treated within 24 hours, 24-48 hours and after 48 hours respectively. Delay >48 hours was significantly associated with increased mortality (p value <0.001), morbidity (p value <0.001), and surgical site infections (p value <0.001). The mean length of hospital stay for patients with delay <48 hours and over 48 hours was 7.22 (+ 1.9) and 19.7 days (+ 19.1) respectively (p<0.001).


Conclusion: Delay of more than 48 hours is associated with high morbidity and mortality. Efforts should be made to reduce the amount of pre-treatment delay to less than forty eight hours.


Peptic ulcer perforation is the second most frequent abdominal emergency that requires surgery. Perfora-tion occurs in up to 10% of patients with peptic ulcer disease (1). Perforation is predominantly a surgical dis-ease and surgery should proceed as soon as a patient is resuscitated. However, recent data indicates a worrying trend towards longer treatment delays when compared to earlier studies (1). In Norway, Svanes et al have shown a steady increase in treatment delay between the years 1935-1990, especially so for in-hospital delay (2).


Delays of more than 24 hours increase surgical mortality seven-eight fold, complication rate three fold and length of hospital stay two fold in the Western world (3). The patterns and adverse effects of treatment delay have not been evaluated in Kenya. An understanding of this aspect is potentially relevant in guiding the treatment protocols at our institutions.


Patients and Methods

This descriptive study involved consecutive selection of patients treated surgically for perforated peptic ulcer disease at KNH between Jan 2002-Dec 2008. The study was approved by the KNH ethics board. The information sought included patient characteristics, treatment delays, treatment complications and length of hospital stay.


Treatment delay encompassed both pre-hospital and intra-hospital delay. Pre-hospital delay was the time in hours from onset of pain to presentation at the hospital’s casualty department. Intra-hospital delay was defined as the time in hours from arrival at casualty to operative treatment. Simple repair and Graham patch formed the mainstay of treatment for patients with peptic ulcer per-forations (190 of 193). Only three patients had a de-finitive repair done for their perforation. No patient had a laparoscopic surgical repair done for their perforated peptic ulcer.


The information sought was collected by the first author by means of a structured data sheet and the analysis per-formed using SPSS program version 11.5. The total delay time was used to stratify the patients into three delay groups: <24 hours, 24-48 hours and over 48 hours. For descriptive purposes, data were presented as means with standard deviation for continuous variables or as absolute and relative frequencies for qualitative vari-ables. The Student t test was used for the comparison of continuous variables. Categorical and binary variables were tested by Fisher’s exact test/chi square test. A P val-ue of <0.05 was accepted as significant.




One hundred and ninety three (193) patients were re-cruited during the study period. One hundred and fifty one (151) patients had their records reviewed (2002-2007) while forty two patients were directly interviewed by the researcher (January 2008 to December 2008). An annual incidence of 27.5 patients per year was recorded in this study period. Most (90.7%) patients were male (Table 1).


The patient’s ages ranged from 16 to 84 years. The mean age for women was 35.78 (+ 15.9) years while that for men was 34.35 (+ 10.8) years. Younger patients were predominantly affected (Fig. 1); patients younger than 40 years (especially in the third and fourth decades) ac-counted for 74.6% of all cases while those above forty years of age formed 25.4%.


The anterior duodenum was the most common (78.2%) site of perforation Table 1). while the descriptive nature of the perforation was acute, chronic or unspecified in 31.6%, 19.7% and 48.7% of cases.


Click to view table 1

to the medical ward, instances where patients were too ill and needed prolonged “stabilization” before under-going surgery and for other unstated reasons.

Patterns and impact of treatment delays

Patients presented to hospital within a range of 1-168 hours from approximated time of perforation. Most (63.2%) presented to hospital within 24 hours from the start of their symptoms. The mean time to presenta-tion to hospital was 36.05 hours (Std deviation 31.14 hours, median 23 hours). Males presented to hospital earlier than their female counterparts by a difference of up to eighteen hours (34.3 hours vs. 52.5 hours) (p value 0.012). The intra hospital delays were similar for the genders.


The total time from onset of symptoms to treatment ranged from a minimum of two hours to a maximum of 240 hours with a mean time to treatment of 58.03 hours (Std deviation 43.52 hrs). Seventy four percent (74.5%) of patients were operated on within twenty four hours of hospitalization. The rest were delayed due to several reasons including delayed diagnosis, initial admission


Sixty one patients (31.6%) developed complications af-ter treatment. The most common complication was sur-gical site infections (intestinal leak, wound dehiscence, wound sepsis, deep seated infections like peri-hepatic and intra abdominal abscesses, peritonitis) in 49 pa-tients (25.8%). Other post-operative complications in-cluded intestinal obstruction in two patients, acute renal failure in four patients who died, poor reversal from an-esthesia leading to admission into the intensive care unit (two patients), paralytic ileus (one patient) and pneu-monia (one patient). The mean pre-treatment delay was 105.4 hours for the complications group (Standard deviation 45.52 hrs, and 35.4 hours in the group with no complications (std dev 15.32 hrs, 1.36 hrs) (p <0.001).In 22 patients with pre treatment delays less than 24 hours, none recorded any complication while sixty of 164 patients with delays >


Click to view figure 1

Click to view figure 1


24 hours developed complication (p < 0.001). Delays < 48 hours were associated with morbidity in three (2.6%) patients while 57 patients (78.0%) with delays > 48 hours developed complications (p value <0.001). There was also a statistically significant (p value< 0.001) effect of delay over 48 hours on surgical site infections. Those who delayed more than 48 hours were 2.6 times more likely to develop surgical site infections compared to those who presented earlier. Three patients (3/112) who presented within 48 hours suffered surgical site in-fections compared to their counterparts (46/71 patients).


Length of hospital stay

The length of hospital stay ranged from 2-136 days while the average stay was 12.08 days. Patients with delay of <48 hours had a mean hospital stay of 7.22 days compared to 19.7 days for those receiv-ing treatment after 48 hours ( p < 0.001).



Twenty four patients died in hospital before discharge representing 12.4% of all patients treated. Patients who presented to hospital and were operated within 24 hours of the start of their symptoms recorded no death as com-pared to those operated > 24 hours where 23 of 164 pa tients died (p= 0.045). There was no mortality for those treated less than 48 hours with all mortality occurring to those treated after 48 hours i.e. 23 of 73 patients (30.1%) p value <0.001 (table 2). Those patients who were operated less than 48 hours since their symptoms started had fewer surgical site in-fections compared to those operated after 48 hours from the time their symptoms started. P value<0.001



This study has documented an increasing disease bur-den, predominant involvement of young males and pro-longed pre-treatment durations. The annual incidence of 27.5 is higher than the rate in a 1980 study (4). The actual rate may be closer to the 40 recorded in the con-trolled and prospective arm of the current study. Several studies in Africa have documented increasing incidence of perforated PUD in younger persons (4-6), possibly due to increasing risk factor levels (7). Although the over involvement of males may be attributable to protective influence of estrogens (8), the Kenyan male in this co-hort is likely to be smoking, taking alcohol or unem-ployed (9), a risk factor profile similar to elderly females who seem to be the affected cohort in the West (1,10).


The respective overall morbidity and mortality rates of 31.6% and 12.4% are consistent with previous accounts that put morbidity and mortality associated perforations at 30-50% and 5-31% respectively (10-14). Several fac-tors have been associated with increased complications and death including presence of co-morbidity, shock, ASA class and delay in treatment. The latter was the focus of this study. Following duodenal perforation, the secre-tions are sterile in the first 12 hours. As the time of delay increases infection sets in causing more surgical site in-fections, septicaemia, shock and possibility of death (3). The characteristic patterns of delay in our study include pre-hospital delays longer than 24 hours and gender dis-parity in pre-hospital delay.



The average time to presentation was 36.3 hours, three times longer than the 12 hour cut-off used in Western literature to define delay (3). Only one patient received treatment within the 12 hour definition with 11.8% re-ceiving treatment within 24 hours of symptom onset. Such treatment delays are not surprising within the trop-ical environment, occasioned by poor infrastructure (7, 15, and 19) and poverty.


Post-surgical complications were significantly associated with the pre-treatment delays. Morbidity and mortality rates were higher and the lengths of hospital stay longer for female patients who also happened to present late t