1 Introduction

Varices are abnormally dilated submucosal veins that occur due to portal hypertension. Varices usually develop when the hepatic venous pressure gradient is above 10 mmHg, and bleeding occurs when the gradient is above 12 mmHg [1]. A gradient above 10 mmHg is clinically significant and responsible for the development of portosystemic collateral, as esophageal and gastric varices have an increased tendency to rupture and cause bleeding [2].

Variceal bleeding, a complication of portal hypertension, is the most common cause (~70%) of upper gastrointestinal bleeding (UGIB) in sub-Saharan Africa [3, 4] and is associated with higher morbidity and mortality than other causes of UGIB [1]. The risk of variceal bleeding depends on the severity of liver disease, the size of the varices, and the presence of red wale marks during endoscopic examination. Esophageal varices are the most frequent cause of variceal bleeding worldwide, with a mortality rate of about 20% [5, 6]. Gastric varices are less common compared to esophageal varices, which occur in ~20% of patients with portal hypertension [7]. However, they are responsible for 10%–30% of all variceal bleeding and tend to be severe with higher mortality [7].

Patients with bleeding varices are commonly admitted to Kilimanjaro Christian Medical Centre (KCMC), Tanzania [3]. A retrospective study conducted at KCMC on patients admitted to the ICU revealed that esophageal varices were responsible for 13.5% of UGIB cases, with a 17.2% mortality rate [8]. Another retrospective study at KCMC among patients admitted to the medical ward found patients with esophageal varices to be 2.3%, with a mortality rate of 6.0% [9]. However, in these two studies, the management outcomes were not assessed.

Management of patients with variceal bleeding includes blood transfusions, intravenous and oral medication to control the gradient, and endoscopic procedures such as endoscopic variceal ligation (EVL) and endoscopic injection sclerotherapy (EIS). Endoscopic procedures are a life-saving medical service, but the provisions in low-income countries are inadequate and lead to increased morbidity and mortality [10]. However, the management options given depend on the resources available at the hospital, as a survey in Africa indicated few centers with adequate resources [10]. A detailed survey in sub-Saharan Africa revealed that diagnostic esophagogastroduodenoscopy was performed in approximately 70% of centers, while diagnostic endoscopic retrograde cholangiopancreatography was performed in 7% [11]. Endoscopic procedures are often lacking, unaffordable for most patients, or periodically available in resource-limited settings, leading to medication therapy as the only viable option. However, treatment of variceal bleeding with medications and EVL has been shown to prevent the risk of rebleeding, the duration of hospital stay, and death. The risk of rebleeding without endoscopic intervention is about 60%, with an increased mortality rate of 33% [12].

In Tanzania, it is estimated there are 4.3 gastroenterologists per 10 000 000 people [13]. Due to the expensiveness and unavailability of specialists and endoscopy treatment procedures in Tanzania, the management of variceal bleeding in patients and their outcomes are not known. The objective of this study was to assess the management outcomes among patients with variceal bleeding at KCMC.

2 Methods

The study was a retrospective study to assess the management outcomes of variceal bleeding in patients attending the Internal Medicine department at KCMC between April 2012 and April 2022. KCMC is a teaching hospital serving a population of over 10 million people in northern Tanzania, and its catchment area extends to the central and eastern zones. Tanzania, being impoverished, has nearly 50% of its population living below the international poverty line of US$1.9 per day [14], with US$1 equivalent to 2680/− Tanzanian shillings. Endoscopy is a scarce treatment option in Tanzania, and when combined with daily out-of-pocket expenses, it easily costs over a month's wages for an average rural patient [15].

Being a resource-limited setting, the main treatment options available are conservative or endoscopic interventions. Conservative treatment is offered primarily because of the affordability of endoscopic interventions. During the study period, the therapeutic endoscopic procedures that may have been available for variceal bleeding were only EVL and EIS. Alternative treatment methods such as transjugular intrahepatic portal vein shunt, balloon-occluded retrograde transvenous obliteration, partial splenic artery embolization, and liver transplantation are not available at this tertiary hospital. The conservative treatment options offered at the hospital include blood transfusion, intravenous octreotide 50 mcg bolus followed by infusion of 25 mcg/h for 2 days; intravenous metoclopramide 10 mg thrice daily for 24 h; and intravenous ceftriaxone 1 g every 24 h for 5 days. However, not every patient can afford or continue affording the octreotide. Once the patient has no active bleeding, the patient is started on oral propranolol 20 mg twice daily and titrated up. To create a rough idea of the cost implications, an octreotide injection costs ~US$18 per vial, a metoclopramide injection costs ~US$1.2 per vial, a ceftriaxone injection costs ~US$0.4 per vial, and propranolol costs ~US$0.005 per tablet. The cost of an esophagogastroduodenoscopy is ~US$67, and performing either EVL or EIS costs over US$100 per session.

The study obtained approval from the Kilimanjaro Christian Medical College Research Ethics and Review Committee (No. PG 35/2022), as well as from the hospital and the head of the department. Confidentiality was observed, and all data were stored unlinked to patient identifiers.

The study enrolled all patients diagnosed with variceal bleeding based on diagnostic endoscopy findings and aged 18 years and older who attended KCMC, either as inpatients or outpatients, within the study period. The study excluded all patients whose medical record data was incomplete. A sample size with the power of the study at 80% was calculated [16], with a minimum sample size of 490 patients, and after accounting for a 10% loss to follow-up factor, the required sample size was 539 patients. A total of 570 patients were diagnosed with variceal bleeding; 36 patients were excluded due to missing data, and analysis was done on 534 patients.

Socio-demographic data, past medical history, clinic characteristics, treatment modalities, and outcomes of patients were collected from the medical records and the electronic hospital management system. The primary endpoint, which was the outcome of the treatment modality, included either the rebleed of the varix or mortality of the patient. As per the American Association for the Study of Liver Disease, a rebleed was defined as a recurrent bleed after at least 5 days of no bleeding following an acute variceal bleed recovery [17]. In cases where patients were treated conservatively, no evidence of hematemesis and melena, and were discharged as a result, were considered to have no bleeding.

Using an Excel data sheet, the data was collected and then transferred into SPSS version 26 and cleaned for analysis. Categorical variables were presented as frequency and percentage, and continuous variables as mean with standard deviation (SD) and median with interquartile range (IQR). Statistical tests between categorical dependent and independent variables were done using a chi-square test. Multivariable logistic regression was used to determine factors associated with rebleeding and mortality. A p-value of ≤ 0.05 was considered statistically significant.

3 Results

A total of 534 patients were enrolled in the study. As shown in Table 1, 351 (65.7%) were male. The youngest and oldest patients were 18 years and 87 years of age, respectively, with a mean age of 46 (SD 13.5) years. The most affected (37.5%) age group was 35–49 years. Most (70.1%) were peasants.

Out of 534 patients, bleeding related to esophageal varices alone was identified in 475 (88.9%) patients, gastric varices in 5 (0.9%) patients, and 54 (10.1%) patients with both. Varix size was observed as grade I in 11 (2.0%) patients, grade II in 64 (12.0%) patients, grade III in 157 (29.4%) patients, and grade IV in 296 (55.4%) patients. Four hundred and fourteen (77.5%) patients opted for conservative (medical) treatment only. Endoscopic treatment was performed in 120 (22.5%) patients; EVL in 94 (17.6%) patients; EIS in 23 (4.3%) patients; and both endoscopic procedures in 3 (0.6%) patients (Table 2).

Rebleeding occurred in 214 (40.1%) patients. The age group 35–49 years was likely to rebleed (39.3%) but showed no statistical significance (p = 0.349). The overall rebleeding rate in patients who had endoscopic therapy was 28.0%, which was 12 (5.6%) in EIS and 47 (22.0%) in EVL. However, rebleeding was 72.0% in patients who did not have any form of endoscopic therapy, though it was not statistically significant (p = 0.890) (Table 3). Factors that were associated with rebleeding were patients who were not insured (aOR: 1.85; 95% CI: 1.04–3.29; p = 0.037) and comorbidities, whereby those with no comorbidities (aOR: 0.17; 95% CI: 0.03–0.94; p = 0.042), with non-communicable diseases (aOR: 0.15; 95% CI: 0.03–0.91; p = 0.039), and chronic infection (aOR: 0.12; 95% CI: 0.02–0.86; p = 0.035) were protective for rebleeding as compared to those with malignancy (Table 3).

In-hospital mortality in the study population was 43 (8.1%). Across the age groups, mortality was highest in the 50–64 year age group at 51.2%. Mortality in males (62.8%) was higher than in females. Among the 43 deaths, 41 (95.3%) were in the group who did not undergo endoscopic treatment, and two (4.7%) did have endoscopic treatment (cOR: 5.06; 95% CI: 1.20–21.29; p = 0.027). Mortality was higher in those patients with an admission hemoglobin of < 7 g/dL (72.1%), although the difference failed to reach statistical significance (p = 0.484). Factors significantly associated with mortality were shorter length of stay (aOR 2.23; 95% CI: 1.02–4.87; p = 0.045), patients without comorbidities (aOR: 0.12; 95% CI: 0.02–0.92; p = 0.041), and patients with grade II esophageal varices (aOR: 0.03; 95% CI: 0.00–0.90; p = 0.043) (Table 4).

4 Discussion

In the current study, the patients were treated mostly with conservative treatment (77.5%), likely due to the availability and affordability of bands and sclerosants, as the endoscopic treatments were not covered by health insurance. This observation is similar to the findings in Tanzania, where 60.8% were offered conservative treatment only [4]. However, in Nigeria, conservative treatment was performed on 30.8% of patients [18]. This may be due to conservative treatment being given to patients who did not consent to or were unable to afford endoscopic treatment or who came in during a period when the services were unavailable. Evidence suggests that there is insufficient endoscopy capacity in low-income countries, as surveys have highlighted the lack of therapeutic endoscopy services, the lack of endoscopy training centers, and the need for more endoscopy training [11].

EVL was the most common modality for secondary prevention of variceal bleeding (17.6%) among patients treated endoscopically (22.5%) compared to EIS (4.3%). Our findings are comparable to a study done in Brazil, where 90% were EVL and 0.4% were EIS [19]. In China, EVL was reported in 87.4% of patients and EIS in 3.2% of patients [20]. In Yemen, EVL was applied to 72% of patients, and 28% received EIS [21]. In Nigeria, 40.4% of patients had EVL, and 28.8% had EIS [18]. In Egypt, EVL was done in 61.6%, and only 38.4% had undergone EIS [22]. In Tanzania, EVL was done in 97.3% of patients, and 2.7% of patients had EIS [4]. The trend shows that EVL is dominant over EIS, mainly as a result of the fewer sessions required to achieve variceal obliteration and the lower rate of other complications. However, other studies have also shown EIS to be dominant over EVL, as a study in Egypt showed EIS was performed in 17% of patients and EVL in 13% [23]. In Ghana, EIS was the dominant endoscopic modality (8.5%) for secondary prevention of variceal bleeding in comparison with EVL [24]. Additionally, EVL is performed in approximately 53% of centers, and EIS is performed in approximately 39% of centers [11], suggesting that EVL is more common in sub-Saharan Africa. However, treatment modality depends on the geography and availability of treatment options, as each center may provide or prefer a certain treatment modality over another.

Rebleeding in this study occurred in 40.1% of patients. Rebleeding was higher in patients who received conservative treatment (72.0%) compared to endoscopic treatment. In Nigeria, rebleeding rates were low following endoscopic therapy (5.5%) and were expectedly higher in patients who had conservative treatment (75.0%) [18]. The rate of rebleeding in our study is higher than that reported in other studies from the USA, which was 42% [25], the UK was 26% [26], Brazil was 18.5% [19], Yemen was 14.7% [21], and Morocco was 23% [27]. This difference could be due to access to healthcare and the different treatment modalities offered at each center. Notably, a large portion of the patients in this study were on conservative treatment.

In this study, the risk of rebleeding following endoscopic treatment was 22.0% in EVL and 5.6% in EIS. This finding is different from other studies, which found EIS to have a higher risk of rebleeding. In Tanzania, rebleeding was significantly higher in EIS groups than in those with EVL (23.7% vs. 7.6%) [4]. In Egypt, rebleeding was significantly less in the EVL versus the EIS groups (24.2% vs. 42.9%) [22]. A study from Sudan found that 8% of patients rebled after EIS, while 4% rebled after EVL [28]. The differences may be attributed to the number of patients that received each type of therapy, as fewer patients received EIS in this study. The comparison between these series allows us to conclude that endoscopic therapy, particularly EVL, is effective in preventing rebleeding.

This study identified patients who were not insured and comorbidities that were associated with the risk of rebleeding. Most patients (79.0%) were not insured, and since the endoscopy therapies were not covered by insurance and most of the patients got conservative treatment (77.5%), it was likely that those with no health insurance were likely to rebleed (p = 0.037). As part of health policy in Tanzania, every person should be covered with health insurance, and the therapeutic endoscopic procedures should be partly covered, if not fully. The health insurance coverage would also better protect against comorbidities that may increase the risk of rebleeding.

Patients who had no identified comorbidities (p = 0.042), patients with non-communicable diseases (p = 0.039), and patients with chronic infections (p = 0.035) were less likely to rebleed as compared to patients with malignancy. This may be linked to the severity of the illness and frailty of the patient, owing to the risk of rebleeding. Patients with comorbidities may be more aware of their underlying condition, which enables them to implore precautionary measures that prevent rebleeding. Additionally, patients with chronic infections, such as HIV and Hepatitis B, may have been taking anti-retroviral therapy, which is protective against rebleeding [29].

In-hospital mortality was 8.1% in this study and was comparatively similar to findings reported by studies from the USA, which reported the overall in-hospital mortality in patients with variceal bleeding to be between 8.7% and 9.4% [30, 31]. In the UK, the in-hospital mortality reported was 9.2% [32]. More recently, in Italy, the in-hospital mortality in patients with variceal bleeding was 11.8% [33]. In Egypt, the in-hospital mortality was between 8.5% and 11.8% [23, 34]. In Sudan, the in-hospital mortality among patients with variceal bleeding was 10% [35]. In Tanzania, the reported in-hospital mortality among patients with variceal bleeding was 10% [15]. However, the mortality after variceal bleeding in the current study was substantially lower than previously reported in Tanzania, which was 17.6% [4], and much lower than the rate reported by studies conducted in Nigeria (67%) [18], India (19.87%) [36], Yemen (15.7%) [21], and the UK (18.5%) [26]. This suggests that access to healthcare and advances in the management of variceal bleeding have improved outcomes after variceal bleeding.

This study identified patients who had no comorbidities; patients with esophageal varix grade II and a shorter length of stay were associated with mortality. Patients with no comorbidities (p = 0.041) and patients with esophageal varix grade II (p = 0.043) were less likely to suffer a fatal outcome. This means that those patients with less severe disease and generally better health were less likely to die from variceal bleeding as compared to those with comorbidities and those with higher grades of varices. Notably, patients with shorter lengths of stay had a 2.2 times higher risk of mortality. This may not be an adequate representation since the reason they had a shorter stay may be because the patients died. The difference might be due to factors such as late arrivals at the hospital and delays in intervention.

One of the challenges in Tanzania is the health services provided. As part of Sustainable Development Goal 3, a target is achieving universal health coverage. However, there is no benefit if poor quality of care leads to the reluctance of people to use services. And even if services are accessible, poor quality is undermining health outcomes. The challenges resulting in poor quality are the lack of enough, well-trained, and motivated staff with adequate financial and physical resources to provide basic health services; insufficient and/or ineffective resource allocation; assessment translation into appropriate quality improvement measures; parallel monitoring structures by donor-funded programs rather than country-owned initiatives that burden the system [37, 38]. In Tanzania, given the expansion of health services, quality of care has become a major concern, as some of the issues include poor hygiene, sanitation, and waste disposal; lack of health infrastructure; low motivation of health workers; inadequate adherence to professional and ethical conduct; as well as a knowledge gap among health workers [39].

Similarly, regarding the provision of health services for endoscopy in Tanzania, the challenges arise not only because of a lack of facilities, equipment, and funding but also because of a lack of trained personnel and limited administrative support. Additionally, less than 25% of sub-Saharan Africa is urban. Health provision and services become increasingly difficult with increasing distance from a tertiary center as these challenges become ever more existent. Providing endoscopic services in well-resourced centers is difficult, and in resource-limited settings, these challenges are greater and heightened by greater need [10].

In addition to the challenges of health services provided, a cost-effectiveness analysis has not been established in such resource-limited settings. A study in the USA found that beta-blocker treatment costs an incremental US$12400 per additional variceal bleed. Compared with the beta-blocker strategy, endoscopic treatment costs over US$175000 more per additional bleed. The study concluded that beta-blocker therapy for the primary prophylaxis of variceal bleeding was a cost-effective measure, as the use of endoscopy to guide therapy added significant costs with only a marginal increase in effectiveness [40]. Similarly, a study in Italy found that beta-blocker therapy (£1028) turned out to be more effective and less expensive than endoscopic surveillance (£1699) for primary prophylaxis of bleeding [41]. However, in resource-limited settings, treatment options are offered not based on clinician choice but rather on the financial situation of the patient.

The limitations of this study were that endoscopic therapy is an operator-dependent procedure, and treatment in our study was not done by a single operator, which can also influence the results. This was a single-center study, as these findings may not be generalizable. The data presented reflects treatment modalities and patient demographics specific to KCMC, making it difficult to draw conclusions that could be applied to a wider, more diverse population. Multi-center studies or larger sample sizes are generally needed to capture a more diverse patient demographic and to ensure broader applicability of the findings. Additionally, assessing the cost-effectiveness analysis of conservative vs. endoscopic treatments in rebleeding may help to better apply treatment options. Patients who might have had a rebleed episode or mortality outside of this hospital were not captured and may not give further details of the management outcome.

5 Conclusion

Despite the relevance of this study in a low-resource setting, treatments were guided by availability and affordability rather than guidelines. This data shows an extremely high rate of patients with variceal bleeding being treated conservatively. EVL was mostly used in secondary prevention of variceal bleeding. Mortality and rebleeding rates in our setting remain high, which appears to be due to the expensiveness and unavailability of endoscopic treatment. The predictors for rebleeding were patients who were not insured and patients with and without comorbidities. The predictors of mortality were treatment modalities, patients with no comorbidities, patients with esophageal grade II and less, and shorter length of stay. An approach to care with endoscopic services equipped with all the necessary therapeutic interventions will be vital in improving the outcomes of patients with variceal bleeding. Efforts should be made to improve the availability and accessibility of endoscopic therapy in regional hospitals in Tanzania to improve the outcomes for patients with variceal bleeding. Timely identification and aggressive management of patients at high risk for continued bleeding or rebleeding should be a major focus.

Ethics Statement

The study obtained approval from the Kilimanjaro Christian Medical College Research Ethics and Review Committee (No. PG 35/2022), as well as from the hospital and the head of the department. Confidentiality was observed, and all data were stored unlinked to patient identifiers.

Conflicts of Interest

The authors declare no conflicts of interest.