Oesophageal Cancer Magnitude and Presentation in Ethiopia

Background Although oesophageal cancer is a public health problem in Ethiopia, there are limited data on the magnitude and distribution of the disease in the country. The aim of this study is to assess the magnitude, socio-demographic and clinical characteristics of oesophageal cancer patients in selected referral hospitals of Ethiopia. Methods A retrospective document review was employed in ten referral hospitals in different regions of Ethiopia. Data were extracted from clinical care records of all clinically and pathologically conrmed oesophageal cancer patients who were diagnosed and treated in those hospitals from 2012 - 2017. A structured data extraction tool was used to collect important variables. Descriptive statistics such as median, interquartile range (IQR) and percentages were computed to describe the regional and national distribution of the disease. ESCC: Oesophageal Squamous Cell Carcinoma, HMIS: Health Management Information System, IQR: Interquartile Range, IRB: Institutional Review Board and SNNPR: Southern Nations’ Nationality of People Region.


Background
Oesophageal cancer is the leading cause of cancer related mortality due to its subtle disease course and poor prognosis [1]. According to GLOBOCAN 2018 report approximately 572,034 new cases and 508,585 deaths from oesophageal cancer were estimated worldwide [2]. Esophageal cancer incidence varies globally, with its highest across the "esophageal cancer belt" (Countries of the East and South Africa and Asia), [3,4]. The two main types of oesophageal cancers are squamous cell carcinoma and adenocarcinoma [5]. Squamous cell carcinoma is more common in Central, Eastern, and Southern parts of Africa; with the African oesophageal squamous cell carcinoma (ESCC) corridor stretching from the Southern part of Sudan to the Eastern Cape province of South Africa [6,7]. Ethiopia is one of oesophageal cancer belt countries. A study in the largest referral hospital in Ethiopia indicated that, oesophageal cancer was more common among patients from Arsi and Bale areas as compared to other regions [6].
Stage of cancer at diagnosis determines the disease prognosis. The overall oesophageal cancer survival rate is low because ESCC is often diagnosed late because the disease is asymptomatic at its early stage.
In nearly 50% of the cases the lesions are metastasized at the time of diagnosis [24]. The delay at diagnosis is even more prominent in developing countries with a limited access to cancer care. At one of the referral hospitals in Ethiopia, oesophageal cancer operation rate was 56%, where only 24% were suitable for esophagectomy and the post-operative mortality was also considerably high at 28% [25]. Coupling early case identi cation with prompt treatment helps improve patient survival and quality of life.
Despite its high prevalence little is known about oesophageal cancer in Ethiopia. The existing few facilitybased studies are inconclusive and don't give a clear account of the disease at a national level. This study aims to generate a national estimate of the magnitude of oesophageal cancer, describe the clinical presentation and identify the treatment outcomes in tertiary hospitals.

Study design and setting
The study team went to twelve selected hospitals all over Ethiopia and reviewed the registry books of these hospitals in order to identify oesophageal cancer cases over the past six years. Out of these hospitals, data were collected from ten referral hospitals which are located in seven regional states and one city administration; Amhara, Oromia, Southern Nations' Nationality People Region (SNNPR), Tigray, Afar, Harari, Somali and Dire Dawa. The largest referral hospital was selected for those regions which have more than one referral hospitals. From Oromia regional state, three hospitals by name, Aira general hospital, Goba referral hospital and Arsi University referral hospital were purposively selected since previous studies indicated that oesophageal cancer is common in those areas. Addis Ababa city administration was not included in this study because a similar study was conducted at Tikur Anbessa specialized hospital one year ago. Initially, it was also planned to include two general hospitals from Benishangul and Gambella regional states. However, it was not possible to nd a single oesophageal cancer case either suspected or referred despite all the efforts exerted in reviewing the logbooks in collaboration with the matron nurse, Health Management Information System (HMIS) focal person and surgeons in the two hospitals. Therefore, no data was collected from these two sites and they were not included in the analysis of the present study since the reason for non-documentation was not clear (lack of cases or un-documented cases).

Data sources and study population
Secondary data were collected from the aforementioned hospitals registry books and oesophageal cancer patients' cards from 2012 to 2017. The inclusion criteria were all clinically and pathologically con rmed oesophageal cancer cases that were diagnosed and treated in those selected hospitals during the study period. Suspected cancer cases were included in the study because majority of the hospitals do not have con rmatory diagnostic techniques for oesophageal cancer. Data was not extracted from oesophageal cancer patients´ card with incomplete basic personal information such as age, sex, date of rst diagnosis and type of treatment received during data collection.

Data collection tools and procedures
A structured data extraction tool was used to collect important variables like socio demographic factors (age, sex, marital status, level of education, religion, residence, occupation) clinical characteristics of patients such us date of diagnosis, chief compliant at rst presentation, stage and histologic type of cancer, type of treatment provided, current status and last date of follow up. Health professionals who have rst degree and work experiences in data collection were recruited for data collection and supervision. A day long training was given for these health professionals on the purpose of the study, content of the questionnaire and eld procedures. In addition to this, supervisors were trained in data quality control procedures and eldwork coordination. At the end of the training, each data collector conducted pre-tests in order to make him/herself familiar with the data extraction tool.

Data Management and Analysis
Various precaution measures were taken to ensure the quality of the data. Field supervisors and coordinators were checking the completeness and consistency of the collected data at the end of the day during the data collection phase. The data were entered and cleaned using Epi data version 3.

Results
Socio-demographic characteristics The median age of total 777 oesophageal cancer patients whose cards were reviewed was 55 years with IQR 19 years. Afar regions 34 (4.4%) ( Figure 1).    were found in the middle thoracic in 98(31%), followed by lower thoracic with 75(23.7%) patients. From the total 123 patients whose cancer stage was recorded, 98(80%) of them were at the stage III and IV during their rst presentation (Table 2).

Treatment pattern and outcome
The patient cards' review revealed that 210 (27.0%) patients had surgical treatment and the common type of surgery performed was feeding tube 156 (74.3%) followed by trans-hiatal oesophagostomy 23(10.9%) and thoracotomy/laparotomy 19 (9.0%). However, near to one third 246 (31.0%) of the patients' le did not describe whether they had surgery or not. Diaphragm 17 (2.2%), pericardium 9 (1.2%), pleura 8 (1%) were the most frequently reported adjacent structures based on the intra operative nding or imaging results of the patients' pro le. However, it was found that most 718 (92.4%) of the patients' pro le had no information related to adjacent site involvement.  (Table 3). In most cases oesophageal cancer is three to four times more common in male than females [25,26]. A systematic review conducted in sub-Saharan African countries also showed that male predominance in oesophageal cancer cases with ratio of 2:1 [26]. A systematic review and a meta-analysis conducted in 36 African countries revealed that being male was a risk factor for oesophageal cancer cases (1.7; 95% CI: 1.4, 2.0) [27]. In opposite to these ndings, a study conducted in Sudan showed that 1:1.8 male to female ratio of oesophageal cancer cases [28]. In the present study male to female ratio is 1: 1.3 which was not as high as studies conducted in many developed countries and African countries. The difference could be in most countries the risk factor for oesophageal cancer is tobacco smoking and alcohol. While the prevalence of cigarette smoking and alcohol consumption were low among oesophageal cancer patients in Ethiopia with 5% and 2% respectively [29].
While most oesophageal cancer patients in Ethiopia were not smokers and alcohol users. Based on this study nding, the reason why still a high proportion of men were diagnosed for oesophageal cancer could be related to better health seeking behaviour of men compared to females.
This study indicated that the mean age of the study participants was 52.9 years. A systematic review in Sub Sahara Africa region indicated that the median age of the oesophageal cancer patients was 59 years [26]. Another study conducted in Ghana indicated the mean age of the oesophageal cancer patients was 57.8 years [30]. This nding can be explained by the fact that the chance of getting oesophageal cancer cases increase with age. However, generally the mean age of oesophageal cancer patients can vary from country to country as often the mean age highly depends on the underlying population structure.
In the present study, of the 118 results analysed for histopathological subtype, squamous cell carcinoma consists of 67 (56.7%). This nding is consistent with studies conducted in Kenya, North Sudan, Uganda which all showed a higher prevalence of ESCC as compared to Adenocarcinoma [31,32]. A systematic review in Africa region showed that squamous cell sub type is the pre-dominant histology in Africa [27].
Generally, from the two common sub-types of oesophageal cancer, squamous cell carcinoma is the commonest worldwide and especially it is common among black men [33].
In this study almost 50% of oesophageal cancer cases were found from Oromia region in which for this study data were collected from Bale, Arsi and Wellega. Ethiopia is one of oesophageal cancer belt countries; Arsi and Bale are part of hot spot areas [6] which is similar to the study done at Tikur Anbessa hospital where most oesophageal cancer cases were higher among patients from Arsi and Bale areas as compared to other regions [34]. In Ethiopia most oesophageal cancer patients come from rural area mainly from South and Eastern part of Ethiopia. Most of the risk factors indicated for oesophageal cancer like smoking and alcohol were not present in most oesophageal cancer cases in Ethiopia [29]. This could indicate potentially preventable nutritional factors such as porridge and kocho as a cause. This could be the main reason why most cases are from the similar region and it needs a further study for the risk factors of the diseases in these places.
Of the patients who underwent endoscopy, the majority of patients 209 (66.1%) site of mass was located.
Of the site of mass located, oesophageal cancers were found in the middle thoracic in 98 (31%) and this nding is similar with ve years card review study at Tikur Anbessa Hospital where the middle lower part was the site for the mass in 49% of oesophageal cancer cases [31]. The approximate anatomical distribution of tumours within the oesophagus is <20% in the upper third, 30-70% in the middle third and 20-50% in the lower third [35][36]. The middle third of the oesophagus is the commonest site for squamous cell carcinoma and the lower third is the commonest site for adenocarcinoma [37][38][39][40]. In contrary to this a study in Ghana showed that 84.9% of cases anatomical site for oesophageal cancer, the mass was found at the distal third part [31].
Of the total patients seen during the study period, 692 (90.6%) presented to the hospital with complaint of di culty of swallowing, followed by weight loss 374 (49%). Studies in different part of Africa, including Ethiopia revealed that most patients come to health facility with di culty of swallowing and weight loss [37][38]. The disease is asymptomatic at early stage and most patients seek health care after experiencing increase of clinical manifestation. This is also true for other type of cancers, especially in country where awareness and access to health facilities is low.
From the patient whose cancer stage was recorded, 98(80%) of the patients were at the stage 3 and 4 during their rst-time presentation. This nding is similar with the study in Kenya where 70-80% of patients diagnosed at late stage [27]. This could be because most oesophageal cancer were asymptomatic at early stage and diagnosed at late stage when the outcome of the treatment is poor [33]. A review on data from the six continents indicated that more than 50% of oesophageal cancer patients come to the health facilities when they have metastatic disease [40]. In Ethiopia patients diagnosed at black lion hospital also revealed the same problem. A ve-year review of the oesophageal cancer showed that 56% were operated and only 24% were suitable for esophagectomy; and the mortality after operation was 28% [34].

Strength And Limitation Of The Study
In Ethiopia there are few studies conducted on oesophageal cancer; most of the studies are based on one hospital data. This study was conducted at 10 hospitals which are found in 6 regions and one city administration of Ethiopia. The results from this study can give a better picture about the magnitude of oesophageal cancer in major hospitals in the country. One of the limitations of this study was the inclusion of suspected cancer cases because of unavailability of con rmatory diagnostic techniques in the majority of the hospitals included in this study.

Conclusion
Since most of oesophageal cancer cases were diagnosed at late stage of the disease, palliative treatment options such as oesophageal stent implantation is urgently needed. This applies especially for the regions where higher numbers of patients are received in the hospitals such as Oromia region. More public awareness could possibly lead to early detection with higher cure rates even though the disease has high fatality rates even in higher resource settings.

Recommendation
Further research is required to assess the risk factors for oesophageal cancer in Oromia region where the disease is more prevalent. A data recording system should be improved in all hospitals in Ethiopia.