American Indians in the U.S. have a high incidence of gallbladder carcinoma (GBC). Furthermore, American Indians in New Mexico (NM) have the highest incidence rate of GBC in the U.S. The epidemiology of GBC in NM has not been studied in the past 3 decades.
By using the NM Tumor Registry (NMTR) and the Surveillance, Epidemiology, and End Results (SEER) database, age-specific incidence rates, average annual age-adjusted incidence rates, and incidence rate time trends of GBC were compared among the three major ethnic groups in NM: American Indians, Hispanics, and non-Hispanic whites, for the period 1973–2001.
A sharp increase in GBC incidence occurred with advanced age and started earlier in American Indians (at age 55) than Hispanics (at age 60) than non-Hispanic whites (at age 65). GBC occurred more frequently in females than in males in all ethnic groups. In females, the incidence of GBC was the highest in American Indians (14.5 per 100,000) followed by Hispanics (6.8 per 100,000) and non-Hispanic whites (1.4 per 100,000). Similarly in males, American Indians had the highest incidence rate of GBC (7.8 per 100,000), followed by Hispanics (2.0 per 100,000), and non-Hispanic whites (1.0 per 100,000). The time trend analysis revealed that there has been a disproportionate decline in the incidence of GBC in the three ethnic groups, with the greatest drop in American Indians and Hispanics followed by non-Hispanic whites. Despite the decline, American Indians continued to have the highest incidence rate of GBC in NM.
Gallbladder carcinoma (GBC) in the U.S. is a rare but fatal disease characterized by poor prognosis and absence of effective therapy. In the U.S., The median survival is < 6 months, the overall 5-year survival rate is < 5%, and most patients die within 1 year of the diagnosis.1 Females have been found to have a two to six times higher incidence of GBC than males.2 GBC incidence increases steadily with age; the mean age is 65 years.1–3
GBC has a striking geographic and ethnic variation, with pockets of high prevalence scattered throughout the world.1–4 Bolivia and Chile have the highest incidence rates in the world (15 and 13 per 100,000 population, respectively) compared with other countries.4–6 Intermediate incidence rates (3.7–9.1 per 100,000) were reported from Eastern Europe (such as in Poland and the Czech Republic), and Israel.4–9 Low incidence rates (< 3 per 100,000) were reported in New Zealand, the UK, and Spain.4, 8 In the U.S., the overall incidence of GBC is low, at 1.2 per 100,000.2, 10–12 However, the incidence rate of GBC in American Indian females in New Mexico (NM) was noted to be among the highest in the world (11.3 per 100,000).4, 12, 13 Hispanics in NM also have a high incidence rate, although less than American Indians.12–14 In addition, among all the cancers having a higher incidence rate in American Indians than in non-Hispanic whites, GBC ranks as the most common cancer.15 Despite advances in diagnostic technology, in the majority of cases GBC is diagnosed at advanced stages and the prognosis remains dismal.16 Therefore, it is important to develop a preventive strategy to screen high-risk populations and to detect precancerous lesions or early stage GBC where medical intervention can have an impact on survival.
The incidence pattern of GBC in NM has not been updated since the 1970s. NM has a unique tricultural population comprised of 44.7% non-Hispanic whites, 42.1% Hispanics, 9.5% American Indians, and 3.7% other minorities.17 American Indians in NM represent 10% of the total American Indians in the U.S. Thus, this unique population distribution presents an unusual opportunity to study the changes in the epidemiology of GBC in three distinct racial groups having a widely varying propensity to develop GBC within the same region over time. In this study, the New Mexico Tumor Registry (NMTR) and the Surveillance, Epidemiology, and End Results (SEER) program databases were used to analyze the incidence time trends and the pattern of incidence changes of GBC in NM among the main three major ethnic groups during the most recent 29-year period (1973–2001) for which data are available.
MATERIALS AND METHODS
Data was collected from the NMTR, which is one of the 14 registries in the SEER program of the National Cancer Institute (NCI). The population census data was obtained from the NMTR database. The study focused on the major three ethnic groups in NM, non-Hispanic whites, Hispanics, and American Indians, and covered the time period 1973–2001. The age-adjusted annual incidence rates by gender and race/ethnicity were calculated using the 2000 standard population. Because the number of GBC cases is relatively small, we used the average 9-year moving midpoint (taking the average of 9-year values and attaching it to the midpoint year, then moving the average and midpoint in an overlapping window) to calculate the age-adjusted incidence time trend per gender and race/ethnicity. The age-specific incidence rate per race/ethnicity was also calculated. The changes in incidence of GBC in NM over time were analyzed per ethnicity, age, and gender.
The accuracy of the ethnic classification has been a major challenge to interpreting cancer surveillance statistics for American Indians in the U.S. However, the NMTR has worked closely with the Indian Health Service (IHS) over the last 30 years and their records are routinely linked to identify and minimize misclassification of American Indians. In a recent report,18 the misclassification rate of American Indians in NMTR was 6%, one of the lowest rates of misclassification in the SEER program. Therefore, it is unlikely that the trends in GBC presented in this article can be attributed to misclassification.
It is also unlikely that systematic improvements in population estimates, which serve as denominators in cancer incidence rates, could fully account for the trends in GBC that are described in the present report. For example, the absolute number of GBC cases dropped dramatically during the period of this study. Furthermore, during the time period of the present study the incidence rates for some cancers (i.e., female breast, colorectal, liver) actually increased among the American Indian population, while rates of GBC declined.
To provide statistical P-values and 95% confidence intervals (CIs), and to consider the simultaneous effect of age, gender, ethnicity, and calendar time, we used multivariate Poisson regression methods (SAS program, GENMOD).19 A P-value < 0.05 was considered statistically significant. Because the link function for Poisson regression is a natural logarithm, we computed the relative risk (RR) ratio and 95% CI for a binary explanatory variable by exponentiating the estimated coefficient and its 95% CI.
A total of 773 cases of GBC were reported in NM between 1973 and 2001. In all, 193 (25%) cases were in males and 580 (75%) cases were in females. Of the male cases, 52 (27.9%) were in American Indians, 75 (38.8%) were in Hispanics, and 66 (34.2%) were in non-Hispanic whites. Of the female cases, 116 (20%) were in American Indians, 303 (52.2%) were in Hispanics, and 161 (27.7%) were in non-Hispanic whites.
Average Annual Age-Adjusted Incidence Rate
The average annual age-adjusted incidence rate of GBC per gender and ethnicity in NM between 1973 and 2001 is displayed in Figure 1. During the combined 29-year period, American Indian females had the highest incidence rate (14.5 per 100,000), which was 2 times the incidence rate of Hispanic females (6.8 per 100,000) and 10 times the incidence rate of non-Hispanic white females (1.4 per 100,000). American Indian males had an incidence rate 3.8 times the rate of Hispanic males and 7.8 times the rate of non-Hispanic white males (7.8, 2.0, and 1.0 per 100,000, respectively). Because the female-to-male incidence ratio was quite variable depending on ethnicity, Poisson regressions were done separately in each ethnicity. In all three ethnic groups the incidence of GBC was significantly higher in females than in males, with an average female-to-male ratio of 2.4 (95% CI, 2.0–2.8, P < 0.001). In American Indians, the female-to-male incidence ratio for GBC was 1.8 (95% CI, 1.3–2.5, P < 0.001), 3.3 in Hispanics (95% CI, 2.6–4.3, P < 0.001), and 1.7 in non-Hispanic whites (95% CI, 1.3–2.3, P < 0.001). When compared with other SEER registries, NM has a higher incidence rate of GBC in females (3.4 vs. 1.8 per 100,000) and males (1.5 vs. 1.0 per 100,000). These findings were expected because of the higher proportion of the American Indian population in NM.
Age-Specific Incidence Rate
The average age-specific incidence rate of GBC per ethnicity between 1973 and 2001 is shown in Figure 2. These incidence rates were combined and averaged over calendar time (1973–2001). The effect of age on incidence rates was not confounded by calendar time because there was no age/calendar time interaction (P < 0.50). No GBC cases were diagnosed before the age of 30 in any of the ethnic groups during the study period. In all three ethnic groups the incidence of GBC was relatively rare before age 55 (< 3.5/100,000 population). The incidence rates remained low in non-Hispanic white and Hispanics up to the age of 60; however, there was a sharp rise in the GBC incidence rate in American Indians beginning in the 55–59 age group, which continued to increase sharply with age (Fig. 2). A sharp increase in GBC incidence rate in Hispanics started at the 60–64 age group, while a more rapid rise in non-Hispanic whites did not occur until the age group of 65–69. The incidence rates progressively increased with age for all ethnic groups (RR: 1.092 per 5 years, 95% CI, 1.087–1.098, P < 0.001). With aging, the GBC incidence rates were higher in American Indians (RR: 9.8, 95% CI, 8.0–11.9, P < 0.001) and Hispanics (RR: 3.9, 95% CI, 3.3–4.7, P < 0.001) compared with non-Hispanic whites. By the age of 80, the annual incidence rate in American Indians was about 75 per 100,000 compared with 26 and 9 per 100,000 for Hispanics and non-Hispanic whites, respectively. The cumulative incidence of GBC by age 85 as an average over calendar time (1973–2001) was very small in non-Hispanic whites (0.0014 or 1 in 720), small but higher in Hispanics (0.0054 or 1 in 184), and still small but 10 times higher in American Indians (0.013 or 1 in 74) compared with non-Hispanic whites.
Incidence Time Trends
The 9-year average moving midpoints of the age-adjusted incidence rates were computed per gender and ethnicity in NM between 1973 and 2001. The GBC incidence time trend in NM females based on ethnicity is shown in Figure 3. The time trend curves showed that the incidence of GBC in females has declined between 1973 and 2001 in all ethnic groups (Poisson regression, P < 0.001). American Indian females had the greatest decline (70.4%) followed by Hispanic females (68.2%) and non-Hispanic white females (41.2%) (Fig. 3). The GBC incidence time trend in NM males based on ethnicity is shown in Figure 4. In comparison with females, the decline in males was more modest in American Indians and Hispanics (36.6% and 41.6%, respectively) and similar in non-Hispanic whites (50%). Because of the greater overall rate of decline in GBC incidence per 100,000 in American Indians and Hispanics between 1973 and 2001, the ethnic differences in the incidence of GBC in NM have sharply narrowed over time, particularly in females (Figs. 3, 4). Despite this decline, American Indians continued to have the highest incidence of GBC in NM. In this study, we found female-to-male ratios of 4.5, 2.4, and 1.4, for Hispanics, American Indians, and non-Hispanics in 1973, respectively. However, as the overall incidence of GBC dropped at a much greater rate in American Indian females compared with males from 1973 to 2001, by the year 2001 the female-to-male ratio had narrowed from 2.4 to 1.1 in American Indians. Similarly, the female-to-male ratio also progressively narrowed with time in Hispanics, from 4.5 in 1973 to 2.4 in 2001. Interestingly, such narrowing of the female-to-male ratio did not occur in non-Hispanic whites. In fact, in this subpopulation the rate of decline for males was greater than that for females, and the female-to-male ratio increased slightly from 1.4 in 1973 to 1.7 in 2001. Thus, between 1973 and 2001 the female-to-male ratio of GBC decreased by about half in American Indians and in Hispanics, while this ratio increased slightly in non-Hispanic whites (Figs. 3, 4).
GBC is rare in the U.S.; however, the increased incidence in American Indians is well documented. In NM, American Indians have a remarkably increased incidence of GBC and American Indians females in NM have one of the highest incidence rates of GBC in the world.4, 12, 13 This study describes the changes that affected the incidence pattern of GBC in this interesting region from 1973 to 2001.
The analysis of the GBC incidence data from NMTR and the SEER program over a 29-year period (1973–2001) revealed significant changes in the incidence rates of GBC in the three major ethnic groups in NM: American Indians, Hispanics, and non-Hispanic whites. There were three studies in the 1970s that looked at the epidemiologic characteristics of GBC in NM. However, since the 1970s there has been no update on the changing pattern of GBC in NM. Morris et al.,12 using the NMTR database between 1969 and 1975, reported that American Indian females in NM had a significantly higher incidence rate of GBC than Hispanic and non-Hispanic white females (21.1, 10.5, 1.4 per 100,000 in American Indian, Hispanic, and non-Hispanic white females, respectively), and that American Indian males in NM also had a higher incidence rate than Hispanic and non-Hispanic white males (5.1, 2.0, and 0.8 per 100,000 in American Indian, Hispanic, and non-Hispanic white males, respectively). Black et al.13 also examined the GBC incidence in NM between 1969 and 1972 using the NMTR database and found that GBC was the third most common cancer site in American Indian females, accounting for 8.5% of total cancers, exceeded only by cancers of the uterine/cervix (16.3%) and breast (15.7%). Devor and Buechley14 also studied the GBC incidence in Hispanics in NM between 1957 and 1977 and found that Hispanics in NM had an intermediate incidence rate of GBC that was higher than non-Hispanic whites but lower than American Indians.
In this study, we updated the epidemiological changes in GBC in NM between 1973–2001 using the NMTR and SEER databases. During the combined 29-year study period, American Indians continued to have the highest incidence rate of GBC in NM, followed by Hispanic, then non-Hispanic whites. The incidence rate of GBC was always higher in females than males across the three ethnic groups; the female-to-male ratio was 3.3 in Hispanics, 1.8 in American Indians, and 1.7 in non-Hispanic whites. By comparison, a previous study reporting data between 1957 and 1977 found a female-to-male ratio of 3.4 in Hispanics, 3.0 in American Indians, and 1.6 in non-Hispanic whites.14 Thus, these findings indicate a large variation in relative risk to develop GBC based on gender and ethnicity.
The time trend analysis of GBC incidence in NM indicated that the GBC incidence rates decreased in all three ethnic groups over time between 1973 and 2001 (Fig. 3). However, the decline in incidence per 100,000 population was greatest in American Indians and Hispanics, followed by non-Hispanic whites. This is in contrast to colorectal carcinoma, in which the incidence rates have increased in Hispanics and American Indians and decreased in non-Hispanic whites during the same period.10 The ethnic and gender differences in the incidence of GBC have sharply narrowed over time from 1973 to 2001 because of the greater decline in GBC incidence in American Indians and Hispanics than non-Hispanic whites and in females more than males. Despite this decline, American Indians continued to have the highest incidence of GBC in NM.
The reason for the disproportionately greater decline in GBC incidence in American Indians and Hispanics during the 29-year study period is unclear. Lazcano-Ponce et al.4 previously reported that the incidence and mortality rates of GBC were declining in most of the world, including the U.S., with the notable exception of Chile. An inverse relationship between an increase in cholecystectomy rates and a decrease in incidence and mortality of GBC has been well demonstrated.20–23 In the recent Strong Heart Study by Everhart et al.,24 cholecystectomy rates were significantly higher in American Indians than in other ethnic groups, suggesting that the greater decline in GBC in American Indians may be in part because of the increase in the cholecystectomy rates.
The reason American Indians in NM have such a higher incidence rate remains unclear. There are several factors that have been identified to be associated with an increased incidence of GBC. GBC has been shown to be associated with the presence of gallstone disease in 75–85% of cases.25, 26 The high prevalence of cholelithiasis in American Indians has been well documented. In 1970, Sampliner et al.27 reported a prevalence rate of 48.6% in his study on Pima Indians using cholecystography. Recently, Everhart et al.24 also evaluated the prevalence rate of gallstone disease in 3296 participants from 13 American Indian tribes in the Strong Heart Study. In their study, they found that 64.1% of females and 29.5% of males had gallbladder disease. Previous studies from NM have also reported a high prevalence rate of gallstone disease in American Indians and Hispanics.12
Epidemiological studies indicate that American Indians carry dominant lithogenic genes and have a higher tendency to develop cholesterol gallstones.28 Carey and Paigen29 speculated that this predisposition is polygenic, involving ‘thrifty’ genes that conferred survival advantages when Paleo-Indians migrated from present-day Siberia to the Americas during the last Great Ice Age, approximately 50,000 to 10,000 years ago. Maringhini et al.30 followed 2583 residents of Rochester, Minnesota, who had gallstones initially diagnosed during the years 1950–1970, and found an association between gallstones and GBC, but only in men. In addition, association between the duration of chronic cholelithiasis (> 20 years) and the stone size (> 3 cm) and increased risk for GBC development has been reported.31–33
Despite the strong association between GBC and the presence of gallstones, the causal relationship between gallstone disease and GBC has not been established. Only 1–3% of patients with gallstones go on to develop GBC, and 15–25% of GBC patients do not have a prior history of gallstones.4 Thus, many patients without gallstones develop GBC. It has been postulated that the chronic inflammation and trauma of stones-containing gallbladder may lead to dysplasia, and subsequent progression to GBC.31
It has also been suggested that bacteria may play a role in the development of GBC by producing carcinogens.34 There is a recent growing interest in the association of Helicobacter species and hepatobiliary malignancies.35, 36 Studies from Japan and Chile, where the risk of GBC is high, showed that patients with GBC have a significantly higher frequency of harboring Helicobacter species (H. bilis and H. pylori) in their bile samples or gallbladder tissues than the general population.37, 38 The causative effect of Helicobacter species in the etiology of GBC is still not proven. To date, no such studies have been done to evaluate this association in American Indians in the U.S.
Finally, it is speculated that American Indians may have a genetic predisposition that in the presence of chronic gallstone disease may promote other carcinogenic and environmental factors and lead to development of GBC.2, 21
In summary, our updated study of the incidence patterns of GBC in NM over the last three decades using the NMTR and SEER databases showed that there has been a significant change in the incidence patterns of GBC in NM. Between 1973 and 2001, the incidence of GBC has disproportionately decreased among the three major ethnic groups. The greatest overall decrease occurred in American Indians and Hispanics, followed by non-Hispanic whites. Thus, the differences in GBC incidence per 100,000 population has narrowed between the three groups during this period. Moreover, the incidence of GBC in American Indian and Hispanic females has declined at a much more rapid rate than in males during this period, such that by 2001 the incidence rates of GBC in American Indian females were almost equal to that of males. In contrast, the female-to-male ratio has increased slightly in non-Hispanic whites. Despite the decline, American Indians in NM continued to have the highest risk of GBC. The reason for the disproportionate decline in GBC incidence based on race and gender remains unclear. It was postulated that the increased rate of cholecystectomy at a higher rate in American Indians may explain in part the disproportionate decline in GBC rates. Further studies are needed to delineate the precise factors involved. Those studies will be very essential to design an effective preventive plan that targets the high risk population. We are currently in the process of evaluating the role and the cost effectiveness of early detection of gallstones, regular surveillance of the at-risk population with ultrasound, and prophylactic cholecystectomy in preventing GBC in the high-risk groups in NM.