Ethical clearance for this study, approval number 908, was obtained from the Institutional Research Ethics Committee of Mount Kenya University (IREC-MKU). Because this retrospective analysis of Hospice cancer records (2003–2018) did not directly involve patients or pose any risks, a patient consent waiver was requested and granted by IREC-MKU. To protect patient privacy, redacted cancer records with all personal identifying information removed, were obtained from Meru Hospice. Permission to analyze redacted patient records was granted by the Meru Hospice administration.

This retrospective study investigated cancer cases affecting residents of the larger Meru region in Eastern Kenya. The larger Meru region encompasses Meru County and the neighboring Tharaka Nithi County. Residents of this region have similar economic and cultural practices and therefore, they may likely share environmental and genetic drivers of cancer.

Meru region is located in eastern Kenya, on the eastern slopes of Mount Kenya at coordinates 0° 3′ 0″ N, 37° 38′ 0″ E. According to the 2019 Kenya population and housing census, Meru County has a land area of 7,006.3 km² and a population of 1,545,714. This population consists of 767,698 males and 777,975 females. Tharaka-Nithi County has a land area of 2,564.4 km² and 393,177 inhabitants. The population of Tharaka Nithi is made up of 193,764 males and 199,406 females ¹⁴ . The Ameru people are dominant in the region. The main economic activities in the larger Meru region are commercial farming of khat, cash crop farming, floriculture, horticulture, and subsistence farming of bananas, maize, beans and livestock keeping. The county literacy rate for Meru was 53% by 2012 ¹⁵ and the poverty level was at 15.5% as of 2016 ¹⁶ . The county literacy rate for Tharaka Nithi is 69.75% and poverty rate is about 40% ^{17, 18} .

For this study, we used all cancer cases data recorded by the Meru Hospice from 2003 to 2018. The Hospice began providing palliative care on 21 ^st January 2003 and has records for 2003 to 2018, the latest year with complete annual data. Meru Hospice provides palliative care services to residents of the wider Meru region, covering Meru and Tharaka Nithi Counties. We considered Meru Hospice cancer records to be representative of the cancer burden in this region. Being the only palliative care provider in the Meru region, the hospice sees most of the region’s cancer cases. The hospice also sporadically cares for cases from regions beyond the larger Meru region. A few patients from Embu, Nyeri, Laikipia, Marsabit, Samburu, Kirinyaga and Kitui county were also attended to at Meru Hospice during the study period. The data analyzed was collected during patient admission and included the following information: patient sex, age, diagnosis, area of residence and year of admission.

A de-identified patient register, with a total of 2424 entries was obtained from Meru Hospice. The file was first cleaned-up to exclude incomplete and non-cancer entries, a total of 39 entries. Because this analysis aimed to describe cancer patterns in the larger Meru region only, cases from Embu, Nyeri Laikipia, Marsabit, Samburu, Kirinyaga and Kitui county, 36 cases in total, that were seen at the Hospice during the study period were excluded from the analysis. This left 2349 cancer cases eligible for further analysis.

To make analysis uniform, we categorized cancer cases by type based on the diagnosis entry in the hospice cancer patient register. Cases whose raw data diagnosis column indicated colon cancer, rectal cancer or Ca rectal were together categorized as colorectal cancers. Cases affecting the head and neck region and whose raw data diagnosis column indicated tonsillar cancer, throat cancer, nasopharyngeal cancer, oral cancer, tongue cancer, laryngeal cancer, mandible cancer, squamous cell carcinoma (SCC) of the mouth, SCC of the throat, palate cancer, SCC of the hypopharynx, lip cancer, oropharyngeal carcinoma or another site in the head & neck region, were collectively grouped as the head and neck cancers. Other cancer types that were analyzed included bile duct cancer, bladder cancer, bone cancer, brain cancer, breast cancer, cancers of the uterus, cervical cancer, gall bladder cancer, Kaposi sarcoma, liver cancer, lung cancer, lymphoma, melanoma, multiple myeloma, esophageal cancer, pancreatic cancer, prostate cancer, stomach cancer and ovarian cancer. Because cancer in Kenya is often diagnosed late ^{19, 20} , we considered year of admission to be same as year of diagnosis.

Of the 2349 cancer cases eligible for further examination, entries that were incomplete for particular forms of analysis were excluded from those analysis only. For instance, entries lacking the age were excluded from age-distribution analysis but included in other forms of analysis. Cancer entries whose site was not clearly recorded, e.g. those entered as SCC without indication of the affected site, were excluded from the analysis of cancer incidence by type. Statistical analysis and graphs were done using GraphPad ^® Prism 6 statistical software. Analysis was done to describe cancer incidence in Meru by age, sex, type and age-distribution. Incidence refers to the number of cancer cases recorded at Meru Hospice between 2003 and 2018. Incidence is conveyed here as a proportion (%) of the total number of cases under consideration, i.e. the total male and female cases combined, total male cases only or total female cases only. For the age at diagnosis analysis, all cancer records were used except where the age was missing in the register.

To characterize the patterns of cancer distribution in the larger Meru region, Kenya, we analyzed cancer records at Meru Hospice for the years 2003 to 2018. In total, after excluding incomplete entries, cases from outside the larger Meru region and non-cancer cases, 2349 out of the 2424 cases seen at Meru Hospice from 2003 to 2018 were analyzed. Of the 2349 cancer cases, 1151 (49%) were male and 1198 (51%) were female ( Figure 1A), indicating an about equal distribution of cancer cases across sexes ²¹ .

We next analyzed the distribution of cancer cases by type in both sexes. From this analysis, the 10 most frequent cancers in both sexes combined, were stomach cancer (13.76%), breast cancer (11.93%), esophageal cancer (11.03%), cervical cancer (10.23%), prostate cancer (9.51%), head & neck cancers (9.21%), liver cancer (6.54%), Colorectal cancer (4.2%), pancreatic cancer (2.76%) and lung cancer (1.32%) ( Figure 1B). 30 male and 20 female cases lacking information on the cancer type or the affected sited and whose diagnosis was recorded only as SCC were excluded from this analysis.

Next, we analyzed the cancer types to establish their proportions in male vs female patients. This analysis does not apply to the sex-specific cancers, i.e. prostate cancer, cervical cancer and ovarian cancer. Stomach cancer, which is the most common cancer type in this region ( Figure 1B), was found to affect men more frequently than women, with 60% of cases occurring in men. Expectedly, our analysis revealed that breast cancer mostly affects women, with 95.7% of breast cancer cases afflicting women. Our analysis also revealed that esophageal cancer affected men more frequently than women, with 63% of cases occurring in men. Head and neck cancers are also more common in men, in whom 66% of cases occur. Likewise, liver cancer affects men more frequently than women, with 64% of cases being in men. Colorectal cancer is only slightly more frequent in men, who are affected in 54% of cases. Together, these data indicate that, as in other world regions, most cancers affect men more frequently than women ²² . The reasons for this gender bias are unclear. However, pancreatic and lung cancer appear to occur in women slightly more frequently than in men, with 55% of pancreatic and lung cancer cases affecting women ( Figure 1C). It should also be noted that our data show that the rates of these two diseases are relatively low in the Meru region. 50 cases whose diagnosis was recorded only as SCC without indication of affected site were excluded from this analysis.

We next analyzed the distribution patterns of cancer types among men and women separately. To do this, the data on cancer cases diagnosed in men were first analyzed and ranked by cancer type frequency. Frequency is expressed as the percentage of a cancer type out of the total number of male or female cancer cases seen at Meru Hospice from 2003–2018. From this analysis, prostate cancer is the most frequent in men, accounting for close to 20% of all cancers in men ( Figure 1D). It is followed by stomach cancer (17%), esophageal cancer (14%), head and neck cancers (12%), liver cancer (8%) and colorectal cancers (5%) ( Figure 1D). Other cancer types in the top 10 most frequent cancers in men are pancreatic cancer (2.52%), bone cancer (1.3%), lung cancer (1.22%) and multiple myeloma (1.22%) ( Figure 1D). 30 male cases recorded only as SCC and whose type/site was unknown were not analyzed. Among women, breast cancer was the most frequent cancer type, accounting for 22% of all cancer cases in women. It is followed by cervical cancer (20%), stomach cancer (11%), esophageal cancer (8%), head and neck cancers (6%) and liver cancer (5%) ( Figure 1E). Other cancers in the top 10 most frequent cancers in women are colorectal cancer (4%), pancreatic cancer (3%), ovarian cancer (2%) and lung cancer (1.22%) ( Figure 1E). 20 female cases recorded only as SCC and whose type/site was unknown were not analyzed. Head and neck cancer collectively refers to cancers affecting various sites in the head and neck region – including the oral cavity, nasal cavity and the pharynx ²³ . To establish the incidence of specific head and neck cancers in men, we broke down the cases by site. It was found that 58% of head and neck cancers in men affect the oral cavity ( Figure 1D’). Cancers affecting the pharynx and larynx account for 22% and 15% of all head neck cancers in men, respectively ( Figure 1D’). A similar analysis of head and neck cancer cases in women revealed that 58% of the cases affect the oral cavity, while 25% and 6% affect the pharynx and larynx, respectively ( Figure 1E’). However, the incidence of head and neck cancers is almost twice as high in men than in women ( Figure 1C).

Next, we analyzed cancer distribution by age in both sexes together as well as separately. We first analyzed all cancer cases to establish the mean age of cancer occurrence in men and women combined. The average age of cancer diagnosis in both sexes combined is 58 years. In men and women separately, the average age of cancer diagnosis is 61 and 56 years, respectively ( Figure 2A). We next analyzed the age of diagnosis for the top six cancer types (each responsible for ≥5% of the total cancer cases) in men. The average age of diagnosis for prostate cancer, stomach cancer, esophageal cancer, head and neck cancers, liver cancer and colorectal cancer was 72, 61, 65, 51, 55 and 60 years, respectively ( Figure 2B). In women, the average age of diagnosis for breast cancer, cervical cancer, stomach cancer, esophageal cancer, head and neck cancers and liver cancer was 51, 55, 61, 63, 58 and 58 years, respectively ( Figure 2C).

We then analyzed the distribution pattern of the top six cancer types (each responsible for ≥ 5% of total cancer cases) by age of diagnosis in both sexes as well as in men and women separately ( Figure 3). In men, cancer diagnosis peaked in age group 60–69, at which age 25% of all cancers in men were diagnosed ( Figure 3A). 19% of all cancers in men were diagnosed in age group 50–59 and 23% were diagnosed in age group 70–79. Together, 66% of all cancers in men were diagnosed in these three decades of life ( Figure 3A).

Next, we analyzed the age distribution of the top six cancers in men by type. The most frequently diagnosed cancer among men was prostate cancer, which accounts for 19% of cancers in men ( Figure 1D). Analysis of the age-distribution of prostate cancers indicated that its diagnosis peaked in age group 70–79, which comprised 28% of the cases ( Figure 3A–I), with 24% of prostate cancer cases diagnosed in age group 60–69 and 22% in age group 80–89. No prostate cancer cases were diagnosed in patients below 40 years ( Figure 3A–I). Only 2% and 8% of prostate cancers were diagnosed in age groups 40–49 and 50–59, respectively ( Figure 3A–I). Together, 90% of the prostate cancer cases were diagnosed after the age of 60 ( Figure 3A–I).

Stomach cancer accounts for 17% of all cancers in men ( Figure 1D). The age distribution analysis revealed that stomach cancer peaked in age group 60–69, in which 31% of the cases were diagnosed. 25% of stomach cancer cases in men were diagnosed in age group 50–59 ( Figure 3A–II), 56% were diagnosed between 50 and 69 years of age, 16% of all stomach cancer cases were diagnosed before 50 years of age ( Figure 3A–II) and 28% after the age of 70 ( Figure 3A–II).

The frequency of esophageal cancer diagnosis in men peaks in age groups 60–69 and 70–79, in which 54% of all male cases were diagnosed ( Figure 3A–III). 30% of esophageal cancers in men were diagnosed before age 60 ( Figure 3A–III) and 15% after the age of 80 ( Figure 3A–III).

Diagnosis of head and neck cancers in men peaked in age group 50–59, in which 24% of the cases were diagnosed. 37% of head and neck cancers occurred before 50 years and 16% in age group 60–69 ( Figure 3A–IV). 7% of head and neck cancers in men were diagnosed before 30 years of age and 15% past 70 years.

Liver cancer diagnosis in men peaked in age groups 50–59 (22%) and 60–69 (22%), in which 44% of liver cancers were diagnosed ( Figure 3A–V). 34% of liver cancers in men were diagnosed before the age of 50 ( Figure 3A–V) and 23% past 70 years of age.

Finally, colorectal cancer diagnosis among men peaked in age group 60–69, in which 31% of cases were diagnosed ( Figure 3A–VI). 28% of all colorectal cancers in men were diagnosed before age 50 and 31% of all colorectal cancers in men were diagnosed after 70 years of age ( Figure 3A–VI).

Age-distribution analysis revealed that cancer among women peaks a decade earlier than it does in men ( Figure 3B). 20% of cancers diagnosed in women occur in age group 40–49, 22% in age group 50–59 and 20% in age group 60–69. Together, 62% of all diagnosed cancers in women occur in these three age groups. 34% of all cancers in women were diagnosed under the age of 50 and 23% after the age of 70.

We next analyzed the age-distribution of the top six cancers in women. The most frequently diagnosed cancer in women was breast cancer, which accounts for 22% of all diagnosed cancer cases in women ( Figure 2E). Analysis of breast cancer age distribution indicated that its diagnosis peaked in age group 40–49 (30%) and age group 50–59 (20%), with 50% of all breast cancer in women diagnosed in these two age groups ( Figure 3B–I). 20% of all female breast cancer cases were diagnosed before 40 and 30% after age 60 ( Figure 3B–I).

Cervical cancer diagnosis peaked in age groups 40–49 (24%) and age group 50–59 (25%), with 49% of all cervical cancers diagnosed in these two decades ( Figure 3B–II). 12% of all cervical cancer cases were diagnosed before age 40 and 38% of cervical cancer cases were diagnosed after 60 years of age ( Figure 3B–II).

Stomach cancers in women peaked in age group 60–69 at 24%. 18% of all stomach cancers in women fell in age group 40–49, 19% in age group 50–59 and 20% in age group 70–79 ( Figure 3B–III). 7% of stomach cancer cases in women were diagnosed before 40 years of age ( Figure 3B–III). 14% of stomach cancer cases in women occurred after the age of 80 ( Figure 3B–III).

Diagnosis of esophageal cancer cases in women peaked in age groups 60–69 and 70–79, in which 53% of the cases were diagnosed. Of all esophageal cancer cases in women, 20% were diagnosed in age group 50–59 and 26% in age group 60–69 ( Figure 3B–IV). 13% of esophageal cancer cases were in women younger than 50 and 13% were diagnosed after 80 years of age ( Figure 3B–IV).

Head and neck cancer diagnoses in women peaked at age groups 50–59 and 60–69, in which 52% of the cases were diagnosed ( Figure 3B–V). 25% of head and neck cancers in women occurred before 50 years of age and 23% after the age of 80 ( Figure 3B–V).

Liver cancer in women peaked in age groups 60–69 and 70–79, in which 48% of liver cancer cases were diagnosed ( Figure 3B–VI). 28% of liver cancer cases in women occurred before 50 years of age and 18% in age group 50–59. 6% of all liver cancers in women were diagnosed after the age 80 ( Figure 3B–VI).

Five cases, one male and four female, that lacked age information were not included in the analysis of cancer distribution by age.

Outside of the top 10 most frequent cancer types, there were various other cancer types seen at the Meru hospice during the study period. These rare cancer types (responsible for ≤1.2% of cases) include bone cancers, ovarian cancer, melanoma, multiple myeloma, leukemia, lymphoma, Kaposi sarcoma, kidney cancer and uterine cancer ( Figure 5).

Cancers of the stomach are responsible for almost 800,000 annual deaths worldwide ⁴ . According to GLOBOCAN’s estimates, it is the fifth most common cancer globally, making up 5.7% of all cancers. Its incidence is reportedly highest in Eastern Asia (11% of cancers in both sexes combined) and lowest in Africa, where it constitutes only 3% of cancers in men and women combined ^{4, 32} . Our data show that in this region of Kenya, stomach cancer is the commonest cancer in both sexes combined, constituting 14% of all cancers. This rate is similar to its rates in countries in Eastern Asia with the highest stomach cancer incidence globally, including Japan, Republic of South Korea and Mongolia where it constitutes 13.1%, 13.4% and 14.4% of all cancers, respectively ^{4, 32} . From our analysis, this disease is the second most common in men, making up 17% of cancers in men and third most common in women, constituting 11% of cancers in women. These rates are markedly elevated relative to the global stomach cancer rates in men (7.2%) and women (4.1%) ³² . Together, this indicates that stomach cancer incidence is surprisingly elevated in Meru to a rate that equals stomach cancer incidence in countries with the highest rates worldwide. Interestingly, elevated stomach cancer rates were observed in this region as early as in the early 1990s ³³ . Globally, stomach cancer rates are two-fold higher in men than women ⁴ . Consistent with this observation, our analysis shows a 1.5-fold higher stomach cancer rate in men than in women.

Helicobacter pylori ( H. pylori), which infects more than 50% of the world’s population, is an established major risk factor for stomach cancer ^{34– 37} . However, only a small proportion of H. pylori positive individuals go on to develop stomach cancer and the carcinogenicity of H. pylori seems to be strain specific. While H. pylori infection rate is high in Kenya ³⁸ , the role of this pathogen in Kenyan stomach cancers has not been investigated. There is a need to establish if and how H. pylori contributes to stomach cancer in this region. Epstein-Barr virus has been associated with increased stomach cancer risk. EBV-associated gastric cancers comprise about 10% of all stomach cancer cases ³⁹ . The role of EBV in stomach cancer is poorly investigated in African populations ⁴⁰ . Given the high prevalence of EBV infection in all human populations ⁴¹ , and the high prevalence of H. pylori in Kenyan populations ³⁸ , it is curious whether co-infection with both pathogens contributes to the high stomach cancer rates seen in this population. The roles played by other risk factors, including genetic, dietary and environmental risk factors, also warrant comprehensive examination. From our analysis, the average age of stomach cancer diagnosis is 61 years in both men and women – almost a decade earlier than the world average of 69 years ⁴² . Considering that in general, cancer is diagnosed late in Kenya, stomach cancer onset is probably much earlier than revealed by our data. The reasons for the earlier age of diagnosis in Kenya are unknown and merit investigation.

In 2018, about 1.3 million new prostate cancer cases and 359,000 prostate cancer deaths were recorded worldwide ⁴ . Globally, prostate cancer is the fourth commonest cancer in men and women combined, constituting 7.1% of all cancers. It is the second commonest cancer in men, making up 13.5% of cancers affecting men ³² . Prostate cancer is the most frequently diagnosed cancer in men of most SSA countries and the leading cause of cancer mortality ⁴ . Our study shows that prostate cancer is the most common cancer in Meru men, constituting 19.4% of all cancers in men. This is a higher rate than globally estimated. The rate is also elevated when compared to GLOBOCAN’s estimates for Kenya, which reports that prostate cancer constitutes 14.9% of all cancers in Kenyan men ³² . As in other parts of the world, our analysis shows that prostate cancer manifests itself in advanced age and is on average diagnosed at 71. Despite the high incidence rates, little is understood about the etiology of prostate cancer. The best understood prostate cancer risk factors are advanced age, ethnicity and genetic predisposition ^{4, 43, 44} . Prostate cancer rates are highest in men of African descent ^{4, 45} . There is compelling evidence that obesity is an important risk factor ^{4, 46} but the importance of body weight in SSA prostate cancer is not evident. The factors leading to high prostate cancer rates in Kenya are unknown and further investigation into its etiology is required.

In 2018 there were close to 600,000 new esophageal cancer cases worldwide, resulting in over half a million deaths. This disease is the leading cause of cancer mortality in Kenya and the third leading cause of cancer deaths in East Africa. It is the sixth leading cause of cancer deaths worldwide ^{4, 32} . Our study ranks it the third most frequent cancer, making up 11% of all cancers in men and women combined. This is a markedly elevated rate relative to its global rate (3.2% incidence; seventh most common; ³² . Our data reveal a marked esophageal cancer elevation in both men (14.3% of cancers in men; third most common) and women (7.9% of cancers in women; fourth most common) when compared to the global rates of the disease in men (4.2% incidence; seventh most common) and women (2% incidence; 13 ^th most common) ³² . Esophageal cancer incidence displays a markedly variable geographic distribution ^{4, 47} . Kenya lies along Africa’s esophageal cancer corridor but as currently defined, the corridor runs from Ethiopia through Western Kenya to South Africa ²⁶ and does not include the Eastern Kenya region that includes Meru. All of Africa’s esophageal cancer hotspots, including the only characterized Kenyan esophageal cancer hotspot (Tenwek – Eldoret) lie along the Rift Valley. Our study suggests that in addition to Tenwek, there may be other high esophageal cancer incidence regions in the country, including the Meru region. Such regions may be inadequately accounted for in national esophageal cancer estimates as these estimates are based on the Nairobi and Eldoret registries. Esophageal cancer falls into two histological subtypes; esophageal adenocarcinoma (EAC) and esophageal squamous cell carcinoma (ESCC) ^{27, 48} . ESCC forms about 90% of esophageal cancers and is the predominant subtype in SSA ^{47, 48} . Although our study does not demonstrate the relative proportions of the two subtypes, it is probable that most esophageal cancers in this population are ESCCs. Future studies should investigate the relative proportions of EAC and ESCC in this and other Kenyan populations. While comprehensive research on the drivers of esophageal cancer in Kenya are lacking ^{4, 49} , several risk factors have been proposed. They include hot beverage consumption, poor oral hygiene and consumption of mursik – a fermented milk beverage popular in the rift valley region ^{50– 52} . To the best of our knowledge, all esophageal cancer studies in Kenya have been carried out in the Rift Valley region esophageal cancer hotspot. Studies of the etiological drivers of esophageal cancer in this high-incidence region are warranted.

Breast cancer is the commonest cancer in women worldwide, constituting 24.2% of cancers in women, and the leading cause of cancer deaths among women. It is the second commonest cancer in both sexes combined, making up 11.6% of all cancers. In 2018, there were over 2 million new breast cancer cases globally and more than 600,000 breast cancer deaths ^{4, 32} . According to GLOBOCAN estimates, this disease is the most common cancer in Africa and the second commonest in East Africa ³² . In our study population, breast cancer closely mirrors these trends. It is the commonest cancer among women, making up 22.4% of cancers in women. It is second commonest in men and women combined, comprising 11.9% of all cancers. These figures also conform with GLOBOCAN’s breast cancer estimates for Kenya, which rank it the most common cancer in women, with a 20.9% incidence ³² . We observe that breast cancer is frequently diagnosed at a young age (<40 years) in our study’s population. It is estimated that only about 7% of breast cancer cases are diagnosed before the age of 40 ^{53, 54} . However, our analysis shows that over 20% of breast cancer cases in this study’s population are diagnosed before 40 years of age. Considering that in Kenya cancer is often diagnosed late, the rate of breast cancer onset before age 40 could be higher. Drivers of breast cancer in Kenya are poorly investigated. Studies carried out in other regions on migrant groups indicate that non-genetic factors account for the bulk of breast cancer cases, while familial and genetic risks contribute <10% of breast cancers ⁴ . Known non-genetic breast cancer risks include early age at menarche, late age at menopause, nulliparity, late age of first child, high body weight and high alcohol consumption ^{4, 55– 58} . Well characterized genetic risks for breast cancer include mutations in BRCA1, BRCA2 and p53 ^{4, 59} . Because breast cancer risks are not well investigated in many African regions, the roles played by these factors in Kenya’s breast cancer cases are not clear. Future research should seek to better characterize the genetic and non-genetic risk factors driving this disease, given that breast cancer incidence is climbing fastest in African countries ^{4, 60} .

In 2018, there were 570,000 new cervical cancer cases and 311,000 cervical cancer deaths worldwide. It is the fourth most common cancer among women globally, at a 6.6% incidence rate ^{4, 32} . The highest cervical cancer incidence and mortality rates are seen in Africa ⁴ . Our analysis shows that cervical cancer is strikingly elevated in women from the Meru region in relation to the disease’s global estimates. Our analysis ranks this disease fourth commonest in both sexes together (10.2% of all cancers) and second most common among women (20% of cancers in women). These rates are consistent with GLOBOCAN’s estimates, which rank cervical cancer the second most common cancer in Kenyan women (18.3% of cancers in women) ³² . The vast majority of cervical cancer cases are linked to persistent infection by at least one high-risk human papillomavirus (HPV) ^{61, 62} , mainly HPV16 or HPV18 ⁶³ . It is estimated that 9.1% of Kenyan women carry cervical HPV/16 infection and more than two thirds of invasive cervical cancers are attributable to HPV16 and/or HPV18 infection ⁶⁴ . However, HPV infection alone is not sufficient to induce cervical cancer and additional risk factors influence cervical carcinogenesis ^{4, 65, 66} . Principal cofactors include cigarette smoking, multiparity, immunosuppression and the use of oral contraception ^{4, 67, 68} . While it is likely that majority of cervical cancers in Meru women are HPV-associated, further studies are needed to establish this. Comprehensive research is required to uncover the cofactors driving the high incidence of this disease and their roles.

Head and neck cancer broadly refers to malignancies that affect one or more sites of the head and neck region. This region includes the oral cavity, paranasal sinuses, pharynx, larynx, thyroid, associated lymph nodes, soft tissues and bone ⁶⁹ . Globally, head and neck cancers account for about 8% of all cancers in both sexes combined and about 5% of all cancer deaths ⁴ . Head and neck cancers disproportionately affect more men than women ⁷⁰ . From our analysis, head and neck cancers are the sixth commonest cancer type, comprising 9% of all cancers in men and women combined. This group of cancers is fourth most common in men (12.4% of cancers) and fourth most common in women (7.9% of cancers). Our study shows that the rate of these cancers is twice as high in men than women and that the oral cavity is the most frequently affected site, accounting for about 60% of all head and neck cancers in men or women. Multiple factors, including HPV infection, EBV infection, cigarette smoking, high alcohol consumption, poor oral hygiene and chewing leaves from certain plants have been implicated as head and neck cancer risk factors ^{71– 77} . Several genetic risks, including aberrations in the Notch pathway and chromatin factors ^{78, 79} , have also been implicated in the development of head and neck cancer. Further research is needed to establish the involvement of these and other factors in the etiology of head and neck cancers in this and other Kenyan regions.

In 2018, there were an estimated 841,000 new liver cancer cases worldwide and more than 780,000 liver cancer deaths. Globally, this disease ranks sixth most common in both sexes combined, making up 4.2% of all cancers, and is the fourth leading cause of cancer deaths ^{4, 32} . Accounting for 6.5% of all cancers in men and women combined, our analysis reveals a moderate elevation in liver cancer rates in the Meru region. Globally, liver cancer is the fifth commonest cancer in men (6.3% of cancers in men) and sixth commonest among women (2.8% of cancers in women) ³² . Relative to global rates, our analysis reveals a moderate elevation of liver cancer incidence among men (8.5% of cancers in men) and women (4.7% of cancers in women). The major risk factors for liver cancer include chronic hepatitis B or hepatitis C infection, aflatoxin ingestion and high alcohol consumption ^{80, 81} . Given the high prevalence of these risk factors in Kenya ^{82– 85} , they are likely to be important drivers of liver cancer in the Meru region. However, further research is necessary to establish how liver cancer incidence in Meru correlates with these and other probable risk factors.

In 2018, there were close to 1.8 million new colorectal cancer cases worldwide and nearly 900,000 colorectal cancer deaths. Globally, colorectal cancer ranks third commonest in men and women combined (10.2% of all cancers), third most common in men (10.9% of cancers in men) and second most common in women (9.5% of cancers in women) ³² . Conforming with reports that colorectal cancer rates are low in African regions, our study reveals a much lower colorectal cancer rate than estimated globally. Our analysis ranks colorectal cancer eighth most common in men and women combined (4.2% of all cancers), sixth most common in men (4.6% of cancers in men) and seventh most common in women (3.8% of cancers in women). These rates conform with GLOBOCAN’s assessment of colorectal cancer in Kenya, which estimates a 4.8% incidence in men and women combined, 5.9% incidence in men and 4.1% incidence among women ³² . However, GLOBOCAN ranks colorectal cancer the fifth most common in Kenya ³² , compared to our observation of eighth most common in the Meru region. According GLOBOCAN, colorectal cancer ranks third and fourth most common in Kenyan men and women, respectively ³² ; above its rank of sixth and seventh most common cancer in the Meru region in men and women respectively. The lower ranking of colorectal cancer in our analysis is due to elevated rates of stomach, head and neck and liver cancer. Advanced age, sedentary lifestyles, familial predisposition, cigarette smoking, heavy alcohol consumption and diet have been proposed as major colorectal cancer risk factors ^{4, 86} . The roles of these and other risks in the etiology of this disease have not been investigated in Kenyan populations. Nonetheless, we observe that most colorectal cancer cases occur after the age of 50, indicating that, as in other parts of the world, advanced age may be an important risk factor. Consistent with global reports that colorectal cancer afflicts men more than women ^{32, 87} , we observe more cases of the disease in men (54% of colorectal cases) than women (46% of colorectal cases). Comprehensive studies are warranted to establish the roles played by various risk factors in the etiology of this disease in Kenyan populations.

In 2018, there were over 450,000 new pancreatic cancer cases worldwide and more than 430,000 pancreatic cancer deaths ³² . Global pancreatic cancer incidence is estimated at 2.5% of cancers in both sexes combined (12 ^th most common), 2.6% of cancers in men (12 ^th most common) and at 2.5% of cancers in women (11 ^th most common) ³² . Consistent with global estimates, we observed that pancreatic cancer comprised 2.7% of cancers in both sexes combined (ninth most common), 2.5% of cancers in men (seventh most common) and 3% of cancers in women (eighth most common). However, these rates are about twice as high as the national pancreatic cancer estimates for Kenya. According to GLOBOCAN 2018 estimates for Kenya, pancreatic cancer makes up 1.5% of all cancers in both sexes combined (14 ^th most common), 1.8% of cancers in men (13 ^th most common) and 1.4% of cancers in women (14 ^th most common). The distribution of pancreatic cancer in Meru men and women is consistent with the global distribution pattern, which shows a moderately higher incidence in men than women ³² . It is thought that most pancreatic cancer cases are driven by environmental risks ⁸⁸ . Suspected risk factors for this disease include advanced age, cigarette smoking, high body weight, diet, diabetes and ethnicity ^{88, 89} . As with most cancer types, the causes of pancreatic cancer in Kenyan populations have not been investigated. Future research will elucidate the role played by these factors in Kenyan pancreatic cancers.

It is estimated that in 2018, there were about 2.1 million new lung cancer cases worldwide and close to 1.8 million lung cancer deaths ^{4, 32} . Globally, lung cancer is the leading cancer type in men and women combined, constituting 11.6% of all cancers. It is the leading cancer type in men, making up 14.5% of cancers in men and the third most common in women, accounting for 8.4% of cancers in women ^{4, 32} . Curiously, our analysis ranks lung cancer 10 ^th most common in men and women combined, constituting only 1.3% of all cancers. Among men it is ninth most common, making up only 1.2% of cancers in men. It is the 10 ^th most common cancer among women and makes up just 1.2% of cancers in women. The low incidence of lung cancer emerging from our study is consistent with GLOBOCAN’s low lung cancer rates estimated for Kenya in 2018 ³² . In Kenya, this disease is ranked 15 ^th most common in men and women combined, at less than 1.4% of all cancers ³² . In Kenya, lung cancer constitutes 2% of cancers in men and less than 1.3% of cancers in women. According to GLOBOCAN cancer estimates, lung cancer rates are also low in Eastern Africa (1.8% of all cancers) and Africa (3.7% of all cancers) ³² . Cigarette smoking is considered to be the most important risk factor for lung cancer and can explain up to 90% of lung cancers ⁹⁰ . Because reliable data on cigarette smoking in Kenya is not available, it is difficult to judge whether the low lung cancer incidence we report is reflective of a low prevalence of cigarette smoking. The curiously low lung cancer rates in Kenyan populations warrant further research into the etiology of this disease.

This study sought to describe the pattern of cancer in the larger Meru region. This region is thought to bear one of the heaviest cancer burdens in the country, although this has not previously been supported with empirical data. Because of limited cancer research in Kenya, the etiologies of this region’s high-incidence cancers are uncharacterized. This report offers a framework for cancer research prioritization in the region. It calls for more concerted epidemiological and basic cancer research efforts that will guide the formulation of effective cancer care and control programs.

Because the study relied on patient records, it contained only information that the hospice deems necessary during patient visits, i.e. age, sex, diagnosis, date of admission and area of residence. According to the Global Cancer Observatory, ideally, PBCRs should provide information about cancer burden as well as relevant information to facilitate the evaluation of cancer etiology in the study population and hence support studies of cancer care, screening, early detection and prevention ⁹¹ . As opposed to quality PBCRs, the patient register we analyzed does not capture important information like occupation, lifestyle (e.g. smoking or drinking) and therapeutic interventions, thereby limiting its capacity to inform about possible cancer causes and support early cancer detection, screening and prevention. Nonetheless, in the absence of quality PBCRs in most of Kenya, hospital and cancer care registers hold crucial information on the burden of cancer in the country, meeting a crucial goal of PBCRs. Evaluating the data contained in cancer patient registers can greatly inform cancer research priorities to elucidate underlying causes and formulate effective screening, early detection and prevention strategies.

Open Science Framework: The state of cancer in Meru, Kenya - a retrospective study. https://doi.org/10.17605/OSF.IO/KH649 ²¹ .

This project contains the following underlying data:

Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).