Appendicitis Mortality in a Resource-Limited Setting: Issues of Access and Failure to Rescue

Abstract

Introduction:

Appendicitis is one of the most common emergency surgery conditions worldwide, and the incidence is increasing in low- and middle-income countries (LMIC). Disparities in access to care can lead to disproportionate morbidity and mortality in resource-limited settings; however, outcomes following an appendectomy in LMIC remain poorly described. Therefore, we aimed to describe the characteristics and outcomes of patients with appendicitis presenting to a tertiary care center in Malawi.

Methods:

We conducted a retrospective analysis of the Kamuzu Central Hospital (KCH) Acute Care Surgery database from 2013–2020. We included all patients ≥ 13 years with a postoperative diagnosis of acute appendicitis. We performed bivariate analysis by mortality, followed by a modified Poisson regression analysis to determine predictors of mortality.

Results:

We treated 214 adults at KCH for acute appendicitis. The majority experienced pre-hospital delays to care, presenting at least one week from symptom onset (n = 99, 46.3%). Twenty (9.4%) patients had appendiceal perforation. Mortality was 5.6%. The presence of a postoperative complication the only statistically significant predictor of mortality (RR 5.1 [CI 1.13–23.03], p = 0.04) when adjusting for age, shock, transferring, and time to presentation.

Conclusion:

Delay to intervention due to inadequate access to care predisposes our population for worse postoperative outcomes. The increased risk of mortality associated with resultant surgical complications suggests that failure to rescue is a significant contributor to appendicitis-related deaths at KCH. Improvement in barriers to diagnosis and management of complications is necessary to reduce further preventable deaths from this disease.

Introduction

Acute appendicitis is a common condition requiring emergency surgery worldwide, yet outcomes following an appendectomy in low- and middle-income countries (LMIC) remain understudied.¹ While the incidence of appendicitis in Western countries has remained stable at approximately 100 per 100,000 person-years, estimates of the incidence in sub-Saharan Africa, ranging from 15 to 28 per 100,000 person-years, and are increasing.²

There is a low incidence rate of acute appendicitis in sub-Saharan Africa and other LMIC.^3,4 One hypothesis holds that the high dietary fiber content of people in these parts of the world makes for low fecal transit time and would account for a low incidence of a fecalith in the lumen of the appendix as a causal agent in people living in these parts of the world.⁵ However, the burden of appendicitis is on the rise in developing countries where access to surgical care is limited, and delays in presentation and surgical intervention is not uncommon.^2,6

A disproportionate number of both years of life lost and disability adjusted life years are attributed to appendicitis in LMIC when compared to HIC.⁷ A recent systematic review estimates that post-surgical mortality in sub-Saharan Africa is 54.0 per 1,000 appendectomies compared to 3.03 per 1,000 appendectomies performed in a high-income country (HIC)⁸ The predictors of mortality and specific barriers to surgical care in this setting are not well delineated. Therefore, we sought to describe the characteristics and outcomes of patients with appendicitis presenting to a tertiary care center in Malawi.

Methods

We performed a retrospective analysis of prospectively collected data in the Kamuzu Central Hospital (KCH) acute care surgery (ACS) registry in Lilongwe, Malawi. KCH is a 900-bed tertiary hospital that serves eight districts in the central region of Malawi, and a catchment population of approximately 6 million people. The ACS registry captures the demographic, clinical, surgical, and operative outcome data of all patients presenting to KCH for emergency surgical care. All patients ≥ 13 years of age who underwent surgery at KCH from September 2013 to March 2020 and had a postoperative diagnosis of acute appendicitis were eligible for inclusion. No patients were excluded. The diagnosis of appendicitis was obtained based on the intraoperative assessment by the surgeon and not histologically confirmation. At KCH, the appendix specimen is not routinely sent for histopathology unless there is a high suspicion for other diagnoses. All patients were managed with laparotomy as laparoscopy is not currently available at our center.

We utilized summary statistics to describe patients’ demographic and clinical characteristics presenting with acute appendicitis to KCH. Variables included age, sex, known comorbidities, vital signs, transfer status, time from symptom onset to presentation to KCH, time from the presentation at KCH to surgical intervention, type of surgery performed, postoperative complications, length of stay, and in-hospital mortality. A dichotomous variable for time from symptom onset to presentation at KCH was created using greater or less than 2 days, as an interval greater than 2 days has been shown to significantly increase risk of appendiceal perforation and postoperative complications.^9,10

To determine the factors associated with postoperative mortality, we performed a bivariate analysis by the mortality. Variables were compared using student’s t-test for parametric continuous variables, Kruskal-Wallis rank test for non-parametric continuous variables, and Fisher’s exact test for categorical variables. We then conducted a modified multivariable Poisson regression to determine predictors of mortality in our population.¹¹ Age, shock, time to presentation, and presence of a postoperative complication were included in the regression a priori. Perforation of the appendix was included in the full model; however, was not found to significantly contribute to the model after reduction.

We performed all analyses using STATA v16.0 (StataCorp, College Station, TX). We report confidence intervals at 95%, and we set statistical significance at α = 0.05. The Malawi National Health Science Research Committee and the University of North Carolina Institutional Review Boards approved this study and the need for informed consent was waived.

Results

Over the study period, 214 adults presented to KCH with acute appendicitis, representing 5.5% of the ACS population. The median age was 30 years (IQR 20–41) and the majority were male (n = 159 [74.3%]) (Table 1). Most did not have any known comorbidities (n = 133 [62.2%]), although a small proportion did have a history of malaria (n = 15 [7.0%]), human immunodeficiency virus (n = 6 [2.8%]), diabetes (n = 2 [1.0%]), tuberculosis (n = 1 [0.5%]), and cancer (n = 1 [0.5%]).

Table 1.

Characteristics of adult patients presenting with acute appendicitis

	Acute Appendicitis (n = 214)
Age (years), median (IQR)	30 (20-41)
Male, n (%)	159 (74.3)
Comorbidities, n (%)
History of TB	1 (0.5)
Recent Malaria Treatment	15 (7.0)
Diabetes	2 (1.0)
Cancer	1 (0.5)
HIV	6 (2.8)
None	133 (62.2)
Missing	45 (21.0)
Initial HR, n (%)
≤ 100	91 (42.5)
> 100	59 (27.6)
Missing	64 (29.9)
Initial SBP, n %
≤ 90	4 (1.9)
> 90	144 (67.3)
Missing	66 (30.8)
Shock index, mean (SD)	0.8 (0.3)
Shock index, n (%)
Shock (shock index ≥ 1)	24 (11.2)
Not in shock (shock index < 1)	124 (57.9)
Missing	66 (30.8)
Transferred, n (%)	163 (76.2)
Admission disposition, n (%)
Ward	212 (99.1)
HDU	1 (0.5)
ICU	1 (0.5)
Pre-op imaging, n (%)
Ultrasound	9 (4.2)
X-ray	107 (50.0)
CT	0 (0)
None	98 (45.8)
Days to KCH from symptom onset, median (IQR)	3 (2-6)
Days to KCH from symptom onset, n (%)
0-1 days	35 (16.4)
2-3 days	46 (21.5)
4-6 days	34 (15.9)
≥ 7 days	99 (46.3)
Days to OR from presentation at KCH, median (IQR)	0 (0-1)
Days to OR from presentation at KCH, n (%)
0 days	122 (57.0)
1-2 days	76 (35.5)
≥ 3 days	16 (7.5)
Perforated appendicitis, n (%)	20 (9.4)
Surgical intervention, n (%)	202 (100)
Appendectomy	200 (93.5)
Abdominal washout	14 (6.5)
Intraoperative Complications, n (%)
Small bowel resection	4 (1.9)
Small bowel repair	6 (2.8)
Ileostomy creation	1 (0.5)
Postoperative Complications, n (%)
Re-laparotomy	3 (1.4)
Wound infection	5 (2.3)
Dehiscence	1 (0.5)
Remained intubated	2 (1.0)
Length of stay, median (IQR)	6 (4-9)
Mortality, n (%)	12 (5.6)

The majority of patients were hemodynamically stable on admission, with only 59 (27.6%) presenting with tachycardia (heart rate > 100 bpm) and 4 (1.9%) hypotensive (systolic blood pressure < 90 mmHg) (Table 1). Based on the shock index, which is a useful predictor of poor hemodynamic status when ≥ 1 in septic patients, only 24 (11.2%) patients with acute appendicitis presented in shock.¹² Unfortunately, approximately 30% of vital sign data was missing, which is not uncommon in the LMIC setting. Sensitivity analysis showed no significant difference in the proportion of shock index missing by mortality (p= 0.4).

Most patients were transferred from outside facilities (n = 163 [76.2%]) with the majority taking at least one week from symptom onset to arrive at KCH (n = 99 [46.3%]) (Table 1). None underwent surgical intervention prior to transfer. Upon arrival, the vast majority were admitted to the ward (n = 212 [99.1%]) with only 2 patients requiring high dependency unit (HDU) or intensive care unit (ICU) care (n = 2 [1.0%]). Over half of the cohort underwent preoperative imaging consisting of either plain x-ray (n = 107 [50.0%]) or ultrasound (n = 9 [4.2%]). Most were able to undergo surgery on the day of arrival at KCH (n = 122 [57.0%]) or within 1–2 days (n = 76 [35.5%]).

An open appendectomy was performed in 200 (93.5%) patients (Table 1). The remaining 14 (6.5%) patients underwent an abdominal washout, consisting of exploratory laparotomy, irrigation of contamination, and drain placement. Washouts were typically performed in the appendiceal perforation setting when only an abscess is identified, presumably due to liquefactive necrosis of the appendix. Other procedures performed were likely secondary to iatrogenic injury or serosal tears, including 4 (1.9%) small bowel resections, 6 (2.8%) small bowel repairs, and 1 (0.5%) ileostomy creation. Twenty (9.4%) patients had perforated appendicitis. Eight (4.0%) had at least one postoperative complication, including re-laparotomy (n = 3 [1.4%]), wound infection (n = 5 [2.3%]), wound dehiscence (n = 1 [0.5%]), and respiratory failure (n = 2 [1.0%]). The 3 relaparotomies were performed for ileostomy creation (n =1), abdominal washout (n = 1), and one coded as “other” without further information. The median length of stay was 6 days (IQR 4–9). Overall mortality in patients presenting with acute appendicitis was 5.6%.

Bivariate analysis by mortality showed that non-survivors were older (median age 35.5 years [IQR 23.5–45.5] vs. 29.5 years [IQR 20.0–41.0], p = 0.3), male (n = 10 [83.3%] vs 149 [73.8%], p = 0.7), and more often in shock (mean shock index 1.0 (SD 0.4) vs. 0.8 (SD 0.2), p = 0.06) (Table 2). Most patients who died were transferred to KCH from an outside facility (n = 11 [91.7%]). The median time to KCH from the onset of symptoms was longer in non-survivors (6 days [IQR 3–10] vs. 3 days [IQR 2–6]), p = 0.3); however, the median time to surgery once arrived was 0 days (IQR 0–1) in both groups. Non-survivors had a higher proportion of appendiceal perforation (n = 2 [16.7%] vs. 18 [8.9%], p = 0.3) and post-operative complications (n = 2 [16.7%0 vs. 7 [3.5%], p = 0.08), though not statistically significant. In multivariable Poisson regression analysis, being in shock, transferring, and delay to surgery greater than 2 days was not statistically significantly associated with greater risk of mortality (RR 2.1 [CI 0.46–9.27], RR 1.4 [CI 0.2–10.7], and RR 1.7 [CI 0.2–13.2], respectively); however, the presence of a post-operative complication resulted in a statistically significant increase in risk of mortality (RR 5.1 [CI 1.1–23.0], p = 0.04) (Table 3).

Table 2.

Bivariate analysis of the appendicitis population by mortality

	Survived n = 202 (94.4%)	Died n = 12 (5.6%)	p-value
Age (years), median [IQR]	29.5 (20.0-40.0)	35.5 (23.5-45.5)	0.3
Male sex, n (%)	149 (73.8)	10 (83.3)	0.7
Shock index, mean (SD)	0.8 (0.2)	1.0 (0.4)	0.06
Transferred to KCH, n (%)	152 (76.0)	11 (91.7)	0.3
Time to KCH from symptom onset (days), median (IQR)	3 (2-6)	6(3-10)	0.3
Time to OR from presentation at KCH, median (IQR)	0 (0-1)	0 (0-1)	0.9
Perforated appendicitis, n (%)	18 (8.9)	2 (16.7)	0.3
Had a post-op complication, n (%)	7 (3.5)	2 (16.7)	0.08

Table 3.

Multivariable Poisson regression of predictors of mortality in patients presenting with acute appendicitis

	RR	95% CI	p-value
Age	1.0	0.96-1.07	0.6
Transferred	1.4	0.17-10.67	0.8
Presence of shock
Not in shock	ref
Shock	2.1	0.46-9.27	0.3
Days from symptom onset to KCH
0-2 days	ref
> 2 days	1.7	0.22-13.18	0.6
Had a postoperative complication	5.1	1.13-23.03	0.04

Discussion

As the incidence of acute appendicitis rises in LMIC, surgical capacity building to address the surgical burden of disease is needed. However, the necessary postoperative outcomes data from sub-Saharan Africa remains poorly described. In our study of the general surgery population at a tertiary care center in Malawi, mortality following the postoperative diagnosis of acute appendicitis is 5.6%. Most notably, over 40% of the population took at least 7 days from symptom onset to arrive at KCH. The median time to presentation in non-survivors was double that of those that survived. Additionally, postoperative complications were associated with a 5-fold higher risk of mortality.

Mortality in our cohort was higher than that of HICs and as well as other LMIC. In the United States, mortality following appendectomy is 0.1%. Worldwide it is estimated to be 0.3%.^13,14 Limited data exist on surgical outcomes in sub-Saharan Africa; however, mortality rates from similar hospital-based studies in Nigeria and South Africa ranged from 0–2% and a multicenter study of countries with a low human development index found 24-hour and 30 day mortality rates following appendectomy to be as low as 0% and 0.6%, respectively.^15–17 These differences in mortality are likely a result of delays to care and failure to rescue.

Delay to intervention has been shown to produce worse surgical outcomes in the appendicitis population.¹³ An interval greater than 48 hours from the onset of symptoms to surgery portends a significant increase in the risk of appendiceal perforation and a two-fold increase in the risk of postoperative complications.^9,10 The potential consequences of this delay include loculated intra-abdominal abscesses, generalized peritonitis, and sepsis.¹⁸ In HICs, patients with appendicitis presenting > 7 days from symptom onset would likely undergo a computed tomography scan and managed with a percutaneous drain rather than operative intervention to avoid potential intra-operative complications like an iatrogenic injury. However, in the absence of adjunctive imaging and interventional radiology, patients in Malawi must be managed based solely on clinical signs and symptoms. Despite delayed presentation and known associated risks, exploratory laparotomy is mandatory for patients presenting with peritoneal signs in Malawi and other resource-limited settings. Although there may be a consideration for damage control laparotomy in the setting of delayed presentation and intra-abdominal sepsis, damage control surgery has a minimal role in a resource-poor environment with high mortality. The increased mortality may be attributable to lack of a robust blood bank, insufficient critical care bed access, minimal ventilators, and limited ICU expertise.

While delay to surgery has been found to increase risk of mortality in some studies, others have found that delay is solely a risk factor for perioperative complications.¹³ Similarly in our analysis, we found that delay to surgery did not directly increase risk of mortality, but given the length of symptoms before presentation, both real and perceived delays to surgery combined with the natural history of the disease are likely driving postoperative complications. These complications, however, should not be fatal. As the only significant predictor of mortality in our population was the presence of a postoperative complication, these results suggest that increased mortality in the appendicitis population at KCH is related to failure to rescue.¹⁹

Failure to rescue, or death secondary to an adverse event or complication that was unrecognized or not promptly treated, is an evolving quality metric in the emergency general surgery population and often reflects disparities in surgical care.²⁰ Factors that are known to contribute to failure to rescue include, staff availability, nursing ratios, limited training and education, lack of technology for patient monitoring, lack of diagnostic adjuncts, insufficient critical care facilities and access to the operating room, and hospital and surgeon volume.^19,21 Specific to our institution is the limited surgical and health care workforce, specifically the limited nurses in Malawi. The most recent review of the country’s healthcare workforce revealed a ratio of only 0.24 nursing professionals per 1,000 population.²² General and specialist medical practitioners’ ratio was even more staggering at 0.02 and 0.007 per 1,000 population, respectively.²² As rates of emergency surgery conditions rise, continued investment in increasing the healthcare workforce is imperative to decrease nursing and other health care provider workloads, improve patient safety, and decrease the incidence of failure to rescue.

There are several limitations owing to the use of a limited hospital-based registry in an LMIC. In our study, pertinent information, such as reasons for the delay to presentation, symptoms, antibiotic use, wound management, and more detailed intraoperative findings, were not collected. Preoperative and intraoperative characteristics such as generalized peritonitis and level of contamination were not captured and thus, were not able to be included in our model as potential predictors of mortality. It is possible that these factors also contributed to mortality rather than solely failure to rescue; however, these factors may have also just driven complication rates rather than independently predicting mortality. Additional limitations include the likely underreporting of perforation rates and limited collection of vital sign data. While vital signs were utilized in our cohort’s clinical management, the vital sign data was not captured in the registry for one-third of the population, a phenomenon that is not uncommon in resource-poor settings. However, missing vital sign data did not vary by mortality as shown by our sensitivity analysis and complete case analysis was used in our Poisson regression model. Finally, only in-hospital complications and mortality were collected, preventing analysis of long-term complications and 30-day mortality. However, we were able to provide a relatively comprehensive description of the acute appendicitis population in Malawi and highlight some of the barriers to care.

Conclusion

Acute appendicitis represents a significant proportion of the emergency surgery burden of disease in Malawi. Pre-hospital delays to care and postoperative failure to rescue contribute to the high mortality in our population. Improvement in barriers to prompt identification and management of postoperative complications is necessary to reduce further preventable deaths from this disease.

Highlight.

• Mortality following appendectomy in our resource limited setting Malawi is 5.6%.

• Most patients with appendicitis present after 7 days of symptom onset.

• Postoperative complications are associated with a 5-fold higher risk of mortality.

• Failure to rescue is a contributor to appendicitis-related deaths in this setting.