Colorectal cancer is increasing in rural Kenya: challenges and perspectives

Graphical Abstract

Perspectives

Colorectal cancer (CRC) is the third most common cancer worldwide and the second leading cause of cancer death^1,2. Although CRC incidence and mortality are decreasing in some high-income countries (HIC)³, CRC rates in low- and middle-income countries (LMIC) are on the rise⁴. Historically, CRC was thought to be an uncommon diagnosis in much of sub-Saharan Africa, attributed to the presence of a high-fiber diet⁵. Although reliable data from sub-Saharan Africa are limited^4,6, there does appear to be a recent increase in CRC rates⁶ (Table 1). With the data available, CRC is the fifth most common cancer in sub-Saharan Africa⁷.

Table 1:

Age Standardized Rates of Colorectal Cancer per 100,000 in multiple countries in sub-Saharan Africa

	1998–2002	2003–2007	2008–2012	2013–2017
Kenya, Bomet
Male	2.0	3.3	4.9	10.0
Female	1.9	3.7	2.2	9.3
Total	2.0	3.5	3.5	9.6
Kenya, Nairobi*
Male	-	12.1**	13.0	-
Female	-	12.4**	12.7	-
Total	-	12.3**	12.9	-
Uganda, Kyadondo*
Male	7.6	8.2	7.8	-
Female	8.2	9.9	7.6	-
Total	7.9	9.1	7.7	-
Zimbabwe, Harare*
Male	7.4	13.3	12.9	-
Female	6.3	11.3	11.7	-
Total	6.9	12.3	12.3	-
South Africa*
Male	-	2.4	2.5	-
Female	-	1.2	1.8	-
Total	-	1.8	2.2	-

^*Registry data reviewed from IARC⁷

^**Data from 2004–2008⁸

At our institution, Tenwek Hospital, we have observed an increased burden of CRC. Tenwek Hospital is a 300-bed faith-based hospital in Bomet County in rural, southwestern Kenya. It serves as a surgery and endoscopy referral center for the South Rift Valley including the surrounding counties and beyond. Anecdotally, it is the personal observation of our senior surgeon (author R.W., who first worked at Tenwek Hospital in 1992) that CRCs were relatively uncommon in the early 1990s, but have become much more common in recent years. As a rough calculation of the incidence of CRC at Tenwek, we calculated the age-standardized rate (ASR) of CRC per 100,000 population using the world standard population and the 2009 Kenya census data for the district (now county) of Bomet. During the 20-year era of 1993 to 2012, the ASR was 2.9 cases per 100,000 people per year. During the recent 5 years of 2013 to 2017, the ASR was 9.6 cases per 100,000 people per year. This trend is consistent with Saidi et al,⁹ who reported a 2.7-fold increase between 1993 and 2005 in Nairobi, the capital of Kenya. Registry data from Nairobi shows that CRC accounts for 6% of all incident malignancies^8,10. Mounting CRC rates should caution the medical community to no longer assume CRC to be uncommon in places like rural Kenya. This realization has implications for other resource-constrained settings on how to prioritize early diagnosis and management of CRC.

Our observation of more CRC cases over time may suggest an actual increase in CRC, but with the limitations of the calculated incidence rates, it may also reflect multiple other factors, including increasing or shifting populations, greater acceptance of health care and greater CRC awareness in our patient population, enhanced detection and specialist referral by clinicians, better recognition of warning symptoms by patients and doctors, or possibly a stronger reputation for CRC care at our hospital. During this period, life expectancy has risen and there has been a 38% increase¹¹ in the national population of Kenya, but CRC case numbers have increased more than this. In addition, although there have been more referrals from far distances in recent years, there have also been more cases from the traditional catchment area of the hospital. We have also observed more hospital admissions throughout the last 25 years, but CRC has risen at higher rates. Diagnostic endoscopy and surgical treatment have been available at Tenwek Hospital throughout the past 25 years, so availability of these services should not have affected the CRC rates. Finally, another reason that has been suggested for increasing rates of CRC in sub-Saharan Africa is the introduction of a “Western diet.”^6,9,12,13 Indeed, it has been proposed that in a progressively more global world, with rising obesity, physical inactivity, and tobacco use, CRC risk may escalate in LMICs to levels as high as those seen in HICs¹⁴. As diets change with higher incomes¹³, this does not imply that the new diet is causative, just as the previous high-fiber East-African diet was postulated to reduce CRC has not been confirmed. Rather, growing incomes could result in both a more Western-style diet and greater exposure to other more important risk factors. All of the possible causes for increasing rates should be further evaluated in Africa, and, regardless of the causes of the trends, healthcare professionals and institutions should be prepared to address the increasing burden of CRC in sub-Saharan Africa.

A limitation in retrospectively reviewing CRC experience in resource-constrained settings is the difficulty in identifying cases and ascertaining complete data upon review of medical charts¹⁵. This is a well-documented limitation of conducting research in resource-limited settings¹⁶. Incomplete medical records, especially in retrospective studies, can lead to many missing data points. Additionally, our figures from Tenwek Hospital include only pathologically-diagnosed cases, which underestimates true incidence. Thus, these estimates must be taken with caution as we work toward better understanding of CRC in our rural region. Despite the limitations in documentation, the 25-year experience at our institution is likely similar to many other comparable settings. These expected limitations should not silence resource-constrained institutions from being included in the conversation of how to improve care of CRC at the global level.

Future data collection at our hospital should be improved. In 2013, we installed electronic medical records, which should increase the completeness of our medical records, allow for ICD-10 coding of diseases, and allow us to include all clinically diagnosed cases, not just those that were pathologically confirmed. This new record system should also allow us to include documentation such as the type of bowel prep used and all polyps that are endoscopically removed. Our institution also implemented a hospital-based cancer registry in 2015. The spread of such data collection capabilities benefit future studies of CRC within the region.

The scarce research on CRC available from countries in sub-Saharan Africa does demonstrate some common themes. Countries throughout sub-Saharan Africa have reported younger ages of CRC diagnosis than are typically seen in HICs^6,17,18. The proportion of cases less than 40 years of age from single centers around sub-Saharan Africa has been reported to be between 19% and 38%⁶. In our experience, we see patients as young as 14 years old and over 20% of our cases are under 40 years of age. For comparison and despite a recent trend toward a younger age distribution of CRC cases in HICs¹⁹, only 3% to 7% are reported to be under 40 years of age^20,21. In our region, other GI malignancies, especially esophageal cancer, are often diagnosed at young ages^22,23. The overall younger distribution of age within the population is one possible explanation. But, the rising incidence of CRC in young patients may also be due to different factors, such as a distinct molecular profile^24,25. CRC in young people is typically associated with a poorer prognosis, faster progression of the disease, a higher stage at presentation, and increased frequency of metastasis^20,26,27. These challenges make the treatment of CRC in LMICs possibly even more difficult.

The tumor site of CRC is another major difference between cases in HIC and LMIC. High-income countries have reported a right-shift of CRC incidence, in which lesions have been found more frequently in the proximal colon over time²⁸ . This is suspected to be due to a decrease in incidence of distal lesions accompanied by an aging population, in whom proximal lesions are more common²⁹. Publications from sub-Saharan Africa typically report a higher proportion of distal colon and rectal cases than do articles from high-income countries³⁰. These left-sided colon and rectal cancers are more often symptomatic, presenting with pain, obstruction, and/or hematochezia than right colon cancers, which tend to grow to a much larger size before symptoms develop.

Essentially all of our patients present with symptoms as asymptomatic patients are currently very rarely screened for CRC in Kenya. The majority of patients present with multiple symptoms. Our patients also experience symptoms suggestive of illness for a long duration before seeking or being referred for definitive care. There are strong correlations between late-stage colorectal cancer and time since the onset of symptoms³¹, so many of our patients present at late stages. There are numerous possible reasons for such a delay in receiving therapy in low-resource settings such as ours, particularly regarding access to care, patient education, expense and difficulty of travel, and the frequent need for referral to distant surgical centers. In order to improve CRC diagnosis and management in resource-constrained settings, these factors warrant further investigation and attention. Further, the prolonged duration of symptoms and advanced stage of disease at presentation¹⁷ suggest opportunities for developing meaningful strategies for earlier diagnosis and referral to centers capable of care.

At our hospital, there have been no cases of CRC detected on routine screening. As the gastroenterology referral center for a large population, we do minimal screening with lower endoscopy and we would anticipate a similar situation for a great deal of rural Africa. Symptoms consistent with a diagnosis of CRC should certainly warrant colonoscopy, and educating referring providers and patients is paramount to reduce the long duration between onset of symptoms and presentation. However, further investigation as to the appropriateness and feasibility of screening programs in Kenya must also be pursued. Utilizing mathematical modeling, Ginsberg et al. found that screening through colonoscopy would be a highly cost-effective strategy to combat CRC in sub-Saharan Africa³², and endoscopic screening is known to impact survival in other settings³³. But in our region, endoscopic capabilities are limited, and need to be further developed through proper training and infrastructure³⁴ Our institution has been training endoscopists to perform routine endoscopy and helping to cultivate providers’ abilities to do therapeutic maneuvers such as polypectomy. But, these efforts need to be scaled up to meet the demand. One such scenario to consider is that the region could potentially benefit from training non-physician endoscopists to perform screening colonoscopies³⁵, but this would need to be further explored in our context. Additionally, with the high percentage of left-sided colon and rectal cancers that we see, our population may be more amenable to screening. However, the long, tortuous sigmoid colons, which are prone to volvulus in our area³⁶, may make screening colonoscopies more challenging.

Over the last 25 years, our institution has relied upon pathologists at other institutions for histological diagnosis. This process often takes 2–6 weeks for results, which is also a limiting factor for our ability to screen. Other low-cost solutions, such as high-resolution microendoscopy, could possibly improve our ability to detect gastrointestinal neoplasia in a more timely manner³⁷.

In Kenya, the Ministry of Health and other stakeholders (including author M.M.) are in the process of developing new guidelines for the screening of CRC in our country. These guidelines in addition to the expansion of insurance coverage, will hopefully improve access to endoscopic screening for gastrointestinal malignancies. Learning from high-resource investigators is important. But local solutions and adaptations can also help to quickly advance Kenya and other countries in sub-Saharan Africa. Similar to how most of sub-Saharan Africa has skipped the technology of telephone landlines to become leaders in mobile phone technology³⁸, we hope that the same ingenuity can expand CRC detection and care. Interestingly, the current discrepancy between CRC incidence in HICs and LMICs greatly diminishes when looking at CRC mortality¹. This is all the more reason to expand infrastructure and capacity to care for CRC in sub-Saharan Africa.

There are numerous challenges to determining the true burden of CRC in settings with resource constraints. However, these difficulties should not limit the inclusion of institutions from LMICs in the discussion of how to address CRC in their population. Regardless of the reason, colorectal cancer is increasingly diagnosed at our hospital in rural Kenya, and this is consistent with the limited reports from around sub-Saharan Africa. These findings have implications not only for our area, but also for Kenya, East Africa, and the wider global community. With the observed shifts, improvements in the ability to screen for, diagnose, and treat CRC by increasing endoscopic and surgical capacity should be public health priorities.

CRC Acronyms

ASR

Age Standardized Rates

CRC

Colorectal cancer

HIC

High-income countries

LMIC

Low- and middle-income countries