Inter-country and ethnic variation in colorectal cancer survival: Comparisons between a Philippine population, Filipino-Americans and Caucasians

Colorectal cancer (CRC) is among the most common cancers worldwide, ranking fourth in men and third in women [1]. Incidence is higher in developed countries than in less developed nations, with the ratio of colon to rectal cancer incidence being 2:1 in the high-risk countries as compared to an almost equal distribution in low-risk nations [1]. Similarly, survival estimates were also reported to be higher in developed countries, with the exception of Eastern Europe, in comparison to less developed nations [1, 2].

Within countries, incidence, mortality and survival rates were reported to vary. In the United States, incidence and mortality among Caucasians were lower than among African-Americans, but higher than among Asian and Pacific Islanders (API) and Hispanics [3]. Five-year survival was found similar in non-Hispanic whites (NHW) and Asian Americans, including Filipino-Americans [4‐6].

While international and interracial CRC survival comparisons have been reported, high resolution studies that could elucidate the role of prognostic factors, such as stage and treatment, are rare [2, 4, 5, 7]. The term "high resolution study" was first used by the EUROCARE group in comparing cancer survival estimates between populations in Europe, and aims to elucidate possible explanations for survival differences [7‐9]. It incorporates information that are usually not available in standard format in population-based registries, such as disease stage and treatment.

In previous studies, variations in access to diagnostic and treatment facilities were put forward as the reasons for survival differences between countries and populations [2, 4, 7]. However, it was also found that biological characteristics, such as age at disease onset and tumor differentiation, vary between ethnic groups [5]. Few studies tried to disentangle how much of cancer survival discrepancies between ethnic groups are explained by ethnicity or by health care conditions.

Furthermore, there is also a dearth of published population-based colorectal cancer survival estimates from developing countries, and comparisons of these with data from developed countries are limited [2, 10, 11]. Such comparisons are important to determine discrepancies in access to and utilization of cancer care services between developed and less developed nations.

From the Philippines, population-based cancer survival data were reported for the first time in the 1998 International Agency for Research on Cancer (IARC) monograph "Cancer Survival in Developing Countries" [12], and only more recently in a summary report for this study [13]. Compared to most populations from developed countries, a huge majority of Philippine residents do not have heath insurance and bear the cost of health care from out-of-pocket funds [14]. These circumstances are likely to play a significant role in access to cancer care services, and in turn, to health outcomes, and should be studied further.

To elucidate the role of various factors, including ethnicity, stage at diagnosis, and access to treatment in CRC survival, we compared Filipino-Americans with Philippine residents, who have the same ethnicity, and with Caucasians living in the US, who are exposed to similar health care systems.

The distribution of cases for each population group by specific cancer site, sex, age group, stage, histology, surgery and radiotherapy are shown in Table 1. Colon and rectum cancer were about equally common for the Philippine residents, whereas the proportion of colon cancers was much larger for both SEER populations. The sex distribution was almost equal in all groups. Caucasians on average were older than the other patient groups, with more than half aged 70 or above, while Filipinos residing in the Philippines were youngest with almost 50 percent below 60 years of age. Both US SEER populations were diagnosed at earlier stages than Filipinos in the Philippines, with larger proportions of patients presenting with localized tumors. In all three population groups, adenocarcinomas are the predominant histological type, between 80 and 90% of patients underwent surgery and less than 15% received radiotherapy. The random sample of Philippine resident patients was representative of all registered eligible CRC cases with respect to socio-demographic characteristics (data not shown).

Age adjusted overall, site- and sex-specific estimates, as well as age specific estimates of absolute and relative 5-year cancer survival are shown in Table 2. Generally, much lower survival was seen in Filipinos living in the Philippines than in the US cancer populations. No major differences were observed between Filipino-Americans and Caucasians. Within all three population groups, prognosis was very similar for colon and rectal cancer patients. Women have slightly higher survival than males in the SEER populations, but the sex difference is more pronounced among the Philippine residents.

As shown in Table 3, advanced stage at diagnosis and not receiving surgery were strongly related to the risk of dying in each of the three populations. A histologic type other than adenocarcinoma was associated with better prognosis in the US cancer populations. A sub-analysis among the Philippine residents showed that not receiving chemotherapy was also related to the risk of death (RR, 1.31; 95% CI 1.08-1.58).

In bivariate comparative survival analysis between population groups (table 4), substantial excess risk of death was seen among CRC patients from the Philippines as compared to Filipino-American patients. Excess overall mortality was further increased when controlling for age, and was substantially reduced when controlling for receipt of surgery. Nevertheless, quite substantial excess overall mortality was found (RR, 2.03; 95% CI, 1.83-2.25) after controlling for these and other factors in multivariate analysis. The small excess risk of Caucasian CRC patients compared to Filipino-American CRC patients seen in bivariate analysis was mostly explained by the age differences, but a slight excess mortality for Caucasian patients remained after controlling for age and other covariates in multivariate analysis (RR, 1.12, 95% CI 1.04-1.20).

This high-resolution study disclosed large survival deficits between Philippine resident CRC patients compared to Filipino-Americans, the latter showing a slightly more favourable prognosis in comparison to Caucasians in the US. Further analysis showed that Filipinos were less likely to be diagnosed at a localized stage, indicating delay in disease detection.

In spite of the availability of early detection methods for CRC, most patients in the Philippines are still diagnosed with an advanced stage disease. From the onset of symptoms, average delay of physician consultation was estimated to be 5 months, while average delay in diagnosis was estimated to be 7 months [36]. On the part of the patients, financial constraints were the main reasons cited for this delay [36], with patients consulting health practitioners only when symptoms become severe. Moreover, due to unawareness of the possibility of malignancy, and with the belief that signs and symptoms will eventually cease, CRC patients tend to be complacent in seeking medical consultation.

Compounding the problem of financial constraints and lack of awareness is physician delay. A slight majority of CRC patients were initially misdiagnosed by attending physicians, with Ameobiasis as the most common initial finding [36]. Due to non-specificity of signs and symptoms of early disease, CRC symptoms were sometimes mistaken as that of Ameobiasis or other gastrointestinal diseases that are common in the country.

Furthermore, Filipino patients were reported to present unique clinicopathological characteristics, including rare occurrence of polyps [37], and histological findings suggest the possibility of an alternative tumorgenesis pathway [37, 38]. This implies more difficulty in CRC detection using endoscopic or radiographic procedures, and that prevention measures through polyp removal may not be applicable. Furthermore, the unique pathology in Filipinos indicate that current screening guidelines in Western countries might not be appropriate [37]. Various recommendations for CRC screening have been put forward [39, 40], but no program currently exists in the country. Nevertheless, there are still some disagreements on what approaches and methods are most appropriate, particularly in Asian countries [39], and more research is needed to determine the best choice of screening strategy.

In our analysis, large survival differences between Philippine residents and Filipino-American patients persisted even after controlling for stage and other key prognostic factors in the analysis, suggesting an important role of other factors, including peri- and postoperative mortality. The greater proportion of Filipino CRC cases with advanced stage tumors increases the possibility of obstruction and high operative mortality. The additional risks, as well as costs, associated with such procedures contribute significantly to survival. Access to cancer care services remain a major challenge, as these are not affordable for most of the population, even though government hospitals and clinics offer subsidized services. Furthermore, the distribution of specialized centers providing cancer care is disproportionate, with most being located in major cities.

Between Filipino-Americans and Caucasians, survival differences are modest, even though very small excess risk of death for Caucasians remained after adjustment for multiple covariates including age. Previous studies have shown that Filipino-Americans have similar CRC survival as that of non-Hispanic white women [5, 6], and this could be attributed to the high level of health care access among Filipino-Americans. They are the second largest Asian-American ethnic group and among the most acculturated, with majority being either native-born or naturalized citizens [41]. Among Asian-Americans, Filipino-Americans were among those who have the highest income and educational levels, as well as health insurance and usual source of health care [41]. Although they have the lowest level of CRC screening as compared to other ethnic groups, they also have the lowest incidence and mortality rates [42], the latter indicating good access to treatment. More research though is needed to ascertain the possible sources of minor survival discrepancies between ethnic groups in the US.

Some limitations should be considered in the interpretation of our survival estimates. Our study only covered variables that were available and comparable in the databases, and not all possible factors that could affect survival were considered. Information not obtainable includes factors related to cancer services, such as organization, training and skills of health care professionals, application of diagnostic and treatment guidelines, and clinical factors and histopathologic factors such as tumor differentiation and grade.

Furthermore, in spite of exhausting measures to locate Philippine patients and to gather survival information, including personal visits to the homes of those who could not otherwise be traced, follow-up was not complete. This is primarily because of high population migration and mobility in the NCR. Patients lost to follow-up were similar to those not lost to follow-up in terms of age, sex and site-specific distributions. However, there are more patients diagnosed at later stages and with stage information unknown among those who are lost to follow-up (data not shown). It is unlikely that patients lost to follow-up have higher survival than those who were not, since it can be assumed that they have more advanced diseases. Furthermore, it can be surmised that they have not received any form of treatment, as follow-up data for these patients were not found in hospitals within the NCR, where majority of specialized cancer care facilities are located, and the availability of cancer treatment is limited in the surrounding areas. Given these circumstances, it can be deduced that the survival estimates are higher than what could be expected, should follow-up be complete for all patients.

In conclusion, the differences in CRC survival between Filipinos and Filipino-Americans underscore the importance of access to health care. Improving access to and utilization of diagnostic and treatment facilities and making them affordable in the Philippines should be emphasized. Likewise, the important role of health education and advocacy, particularly in promoting early diagnosis and prompt treatment should be given priority, for both patients and health practitioners. The comparison between Filipino-Americans and Caucasians indicates that CRC survival can be similar in different ethnic groups, given comparable access to health care.