Trends in the incidence of acquired immunodeficiency syndrome–related malignancies in Thailand

Authors

  • Hutcha Sriplung M.D.,

    1. Epidemiology Unit, Faculty of Medicine, Prince of Songkla University, Songkhla, Thailand
    Search for more papers by this author
  • D. Maxwell Parkin M.D., Ph.D.

    Corresponding author
    1. Unit of Descriptive Epidemiology, International Agency for Research on Cancer, Lyon, France
    • Unit of Descriptive Epidemiology, International Agency for Research on Cancer, 150 Cours Albert Thomas, 69372 Lyon Cedex 08, France
    Search for more papers by this author
    • Fax: (011) 33 472 73 86 96


Abstract

BACKGROUND

Thailand was one of the first Asian countries to be affected by the epidemic of human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS). Furthermore, Thailand possesses a national cancer surveillance system based upon regional cancer registries.

METHODS

Data from five population-based cancer registries, covering one-fifth of the national population, were used to study trends in the incidence of malignancies related to HIV/AIDS in Thailand during the period 1989–2001.

RESULTS

Although the incidence of Kaposi sarcoma (KS) increased slightly from 1989–1991 to 1995–1997, KS remains a very rare malignancy in Thailand compared with other countries in which the prevalence of HIV/AIDS is much lower. The authors reported a marked increase in the incidence of non-Hodgkin lymphoma (NHL), and particularly high-grade/diffuse NHL. However, the largest increases in incidence were noted among individuals age ≥ 55 years and in regions with a relatively low prevalence of HIV/AIDS.

CONCLUSIONS

The rarity of KS presumably reflected the low prevalence of the causative agent (i.e., KS-associated herpesvirus) in the Thai population. The increasing incidence of NHL may be related to the AIDS epidemic, although a similar increase is observed in many countries worldwide and is not specifically linked to the HIV/AIDS epidemic. Cancer 2004. © 2004 American Cancer Society.

Thailand has been affected quite severely by the acquired immunodeficiency syndrome (AIDS) epidemic. The estimated prevalence of infection with human immunodeficiency virus (HIV) at the end of 1999 was 2.2% among adults. In Asia, this is second only to Cambodia, but it is approximately 4 times the prevalence in the United States and 10 times that in Europe, albeit 4 times less than the prevalence in sub-Saharan Africa.1

The risk of developing several malignancies is increased in patients with HIV/AIDS. Kaposi sarcoma (KS), non-Hodgkin lymphoma (NHL), and cancer of the cervix uteri are AIDS-defining illnesses—i.e., their occurrence in an HIV-positive individual implies a diagnosis of AIDS. In prospective (cohort) studies, the risk of developing KS is increased 200–1000-fold in patients with HIV/AIDS, while the risk of NHL is increased by a factor of 70–300.2–4 In contrast, an increase in the risk of developing invasive carcinoma of the cervix due to HIV infection has not been demonstrated convincingly,5 and the risk of developing Hodgkin lymphoma is increased by a factor of approximately 10.3 Marked changes in the incidence of KS and NHL in the general population have been reported in the United States and Africa as a consequence of the AIDS epidemic.6, 7

In Thailand, there have been several case reports of malignancies in patients with HIV/AIDS,8–12 although to date, there has been no information on the impact of the AIDS epidemic on the incidence of malignant disease in the general population. A network of population-based cancer registries located in each of the major regions of the country has been established for the purposes of cancer surveillance.13 Data became available beginning in 1989 (i.e., before the impact of the AIDS epidemic was felt in Thailand), and have allowed us to examine time trends in the incidence of AIDS-related malignancies in the 12 years that followed.

MATERIALS AND METHODS

Cancer cases registered between 1989 and 2001 were obtained from the five population-based cancer registries responsible for cancer surveillance in Thailand.14 These registries cover the provinces of Chiang Mai and Lampang in the north, Khon Kaen in the northeast, and Songkhla in the south of Thailand, as well as the metropolitan area of Bangkok. The total population served (in mid-2000) was 11,592,000 (18.9% of the national population). Data from Bangkok (which had a population of 6.3 million in 2000) were considered to be complete only for the period 1991–1999,14 and cases were only included for this 9-year period. The data collected on each case of malignant disease included anatomic site and histology, along with patient characteristics such as sex, age, place of residence, and vital status. Cancer site and histology are coded according to the Second Edition of the International Classification of Diseases for Oncology (ICD-O).15 Cases of Hodgkin lymphoma (ICD-O codes 9650–9667) were analyzed separately, and NHL cases were grouped into six categories, as recommended by Groves et al.16: small lymphocytic lymphoma (ICD-O codes 9670–9671), follicular lymphoma (ICD-O codes 9690–9698), diffuse lymphoma (ICD-O codes 9593, 9595, 9672–9677, 9680–9683, 9688, 9710–9711, and 9715), high-grade lymphoma (ICD-O codes 9594 and 9684–9687), T-cell lymphoma (ICD-O codes 9670–9671), and NHL not otherwise specified (NOS). The quality and completeness of diagnostic information varied from registry to registry, so that the percentage of cases classified as NOS (ICD-O codes 9590–9592) varied from 24% (in Songkhla) to 63% (in Bangkok). This variation limits the accuracy of reported incidence rates for individual histologic subtypes and of comparisons between regional registries. Diffuse and high-grade lymphomas were considered as a single group, because it is these subtypes that are associated with AIDS; the likelihood of developing small lymphocytic or follicular lymphoma is not affected by HIV infection.17 Within the category of diffuse/high-grade lymphoma, Burkitt lymphoma (ICD-O code 9687) was examined separately, as were lymphomas located in the central nervous system (CNS).

Using 1990 and 2000 census data, age- and sex-specific annual incidence rates were calculated on the basis of person-years at risk.18, 19 Age-specific incidence rates were calculated for the age groups 0–19, 20–44, 45–54, and ≥ 55 years. Age standardization was performed by the direct method, using the world standard population.20

Data on Human Immunodeficiency Virus/Acquired Immunodeficiency Syndrome

A comprehensive HIV/AIDS surveillance system was introduced in Thailand in 1989. National surveillance data are regularly reported by the Bureau of AIDS, Tuberculosis and Sexually Transmitted Diseases to the Joint United Nations Programme on HIV/AIDS (UNAIDS).21

RESULTS

HIV/AIDS

National surveillance data indicate that HIV prevalence peaked among antenatal clinic attendees in the mid-1990s and has since been slowly decreasing. Data obtained via surveillance of women attending clinics in different parts of the country are presented in Figure 1. At the time of diagnosis, 87% of men and 83% of women with AIDS were aged 20–44 years.22

Figure 1.

Human immunodeficiency virus prevalence in pregnant women in five provinces of Thailand, 1989–2001 (Thailand Bureau of Acquired Immunodeficiency Syndrome).

Cancer

Only 27 cases of KS were recorded in the 5 registries over the 13-year period examined (Table 1). Twenty-two cases occurred in men, and five in women. The average age-standardized incidence rate for the period was 0.02 per 100,000. Only 1 case of childhood KS was documented (in a patient age 4 years), whereas 21 of the 26 cases for which patient age was recorded occurred in individuals ages 15–49 years. The mean and median ages of affected patients were 32.5 and 35.8 years, respectively.

Table 1. Kaposi Sarcoma Cases Reported to Cancer Registries, 1989–2001
YearChiang MaiLampangKhon KaenBangkokSongkhlaAll registries
No. of casesASR (95% CI)
  • ASR: age-standardized rate (per 100,000).

  • a

    Excluding Bangkok.

1989000110.028a
1990000000.000a
19910001010.007
19920000000.000
19930001010.013
19942001030.022
19951001020.015
19960006060.053
19971001020.012
19981003040.025
19990003030.025
2000000000.000a
2001400040.073a
1989–19910001120.010 (0.008–0.012)
1992–19942002040.012 (0.009–0.015)
1995–199720080100.026 (0.021–0.032)
1998–200150060110.028 (0.026–0.030)
1989–200190016127 

Over the entire period studied, the change in annual incidence was not significant (P for trend = 0.14). However, incidence rates over 3-year periods increased significantly from 1989–1991 to 1995–1997 (Table 1) but remained stable thereafter.

When all types of NHL (including NHL NOS) were considered, incidence over the 13-year study period increased by 5.9% per year (95% confidence interval [CI], 3.8–8.1%) (men, 6.3% per year; women, 5.5% per year), from 2.7 per 100,000 for the period 1989–1991 to 4.8 per 100,000 for the period 1999–2001 (Fig. 2). Examining trends according to histologic subtype revealed that the incidence rates of follicular lymphoma and small cell lymphoma were stable during this period, whereas the incidence of diffuse/high-grade lymphoma increased at a rate of approximately 10.6% per year (95%CI, 8.6–12.7) (Fig. 2).

Figure 2.

Age-standardized incidence rates (ASR), per 100,000 for all types of non-Hodgkin lymphoma (NHL), and for diffuse intermediate/high-grade, follicular, and small lymphocytic subtypes, compared with chronic lymphocytic leukemia (CLL).

Twenty-nine cases of Burkitt lymphoma were registered during the 13-year study period (1.2% of all cases of NHL). There was no apparent change in incidence over this period (Table 2). Sixty cases of NHL (2.5%) were localized in the CNS. There was no evidence of an increase in the incidence of these tumors (Table 2).

Table 2. Incidence of Burkitt Lymphoma and CNS Lymphoma for Both Sexes
YearsBurkitt lymphomaCNS lymphoma
No. of casesASRb (95% CI)No. of casesASRb (95% CI)
  • ASR: age-standardized rate; CNS: central nervous system; CI: confidence interval.

  • a

    Excludes Bangkok in 1989–1990 and 2000–2001.

  • b

    Per 100,000.

1989–1991a60.12 (0.02–0.23)50.06 (0.01–0.11)
1992–199460.03 (0.01–0.05)190.07 (0.04–0.10)
1995–1997100.03 (0.01–0.05)240.09 (0.05–0.12)
1998–2001a70.02 (0.01–0.04)120.04 (0.02–0.06)

With regard to age group, the rate of increase in incidence of NHL among patients ages 20–44 years was 3.4% per year (95% CI, 0.0–6.8%) in the areas of higher HIV prevalence (Chiang Mai and Lampang) and 4.5% per year (95% CI, 1.3–7.6%) in the areas with lower HIV prevalence (Khon Kaen, Bangkok, and Songkhla). In all registries, the rate of increase was highest among patients age ≥ 55 years (7.0% per year [95% CI, 5.5–8.5%] in higher-prevalence areas and 8.4% per year [95% CI, 5.0–11.9%] in lower-prevalence areas).

Figure 2 also illustrates the incidence of chronic lymphocytic leukemia (CLL; ICD-O code 9823) over the same period. It appears unlikely that the increasing rates of NHL incidence can be ascribed to changing terminology or diagnostic criteria (the recording of cases formerly diagnosed as CLL as small B-cell lymphocytic lymphoma [ICD-O code 9670]), since such changes would have been expected to result in a decrease in CLL incidence.

There was no increase in the incidence of Hodgkin lymphoma during the study period (P for trend = 0.55 and 0.85 for males and females, respectively) (Fig. 3). In Lampang, incidence rates decreased by 12.1% annually (95% CI, 1.4–22.8%), and there also was a nonsignificant decrease in incidence in Khon Kaen (8.0% per year; 95% CI, 0.3–16.2% per year).

Figure 3.

Age-standardized incidence rates (per 100,000) of Hodgkin lymphoma by sex in Thailand, 1989–2001.

DISCUSSION

The analysis of trends in incidence was based on data from the five population-based cancer registries in Thailand. The data were of sufficient quality, with respect to completeness and validity, for inclusion in the series “Cancer Incidence in Five Continents.”23 Although detailed specification of the histology of NHL was incomplete, the proportion of cases with specified histology was constant over the period of observation within each individual registry, and there is no reason why the recording of high-grade and low-grade disease would have changed in a reciprocal fashion. It appears reasonable to conclude, therefore, that the increasing incidence of high-grade NHL is genuine.

KS is one of the most striking manifestations of AIDS. In fact, it was the appearance of an aggressive form of KS in 1981 in the United States among homosexual men that heralded the onset of the HIV/AIDS epidemic.22 It has been suggested that KS was caused by a combination of a sexually transmitted agent and immune suppression.24 Increases in the incidence of KS in relation to the emerging AIDS epidemic have been observed in the United States, Europe, Australia, and Africa.6, 25, 4, 7 In Thailand, there have been various case reports of KS in patients who were HIV positive—e.g., four cases were reported at the teaching hospital in Chiang Mai, one of the provinces most heavily affected by the HIV epidemic.9 Nevertheless, the incidence in Thailand remains relatively low (0.026 per 100,000), as in other Asian countries, and is much lower than in many other countries in which HIV is considerably less prevalent (Fig. 4).

Figure 4.

Age-standardized incidence rates of Kaposi sarcoma (KS) (per 100,000) around 199523 versus prevalence of human immunodeficiency virus (HIV) infection1 in adults in 17 populations. 1: Thailand (5 registries); 2: Korea (4 registries); 3: Hong Kong; 4: Japan (6 registries); 5: Singapore; 6: New Zealand; 7: Italy (15 registries); 8: Sweden; 9: Spain (11 registries); 10: The Gambia; 11: Denmark; 12: France (9 registries); 13: Australia (8 registries); 14: United States (National Cancer Institute Surveillance, Epidemiology, and End Results Program); 15: Harare, Zimbabwe; 16: Kampala, Uganda; 17: Bamako, Mali.

A recently discovered human herpesvirus, Human Herpesvirus-8 (HHV-8), or KS-associated Herpesvirus (KSHV), has been consistently linked with KS and is now considered to be the principal cause of the disease.26 A number of assays have been developed to measure the presence of antibodies against HHV-8 in human sera for diagnostic and epidemiologic purposes. There is variation in both the sensitivity and the specificity of these assays,27 but in general, HHV-8 seroprevalence in ‘normal’ populations (i.e., blood donors, patients without KS, and population samples) broadly reflects the risk of KS in both normal and HIV-infected populations. Thus, HHV-8 seroprevalence appears to be highest in Africa (10–100%, depending on the assay used), intermediate in the Middle East and Mediterranean littoral, and lowest in northern Europe.28 In 2 separate studies, HHV-8 seroprevalence in the normal Thai population in the late 1990s was reported to be 0.6%29 and 4.0%,30 respectively. Although variations in the study populations and the assay techniques used probably underlie the difference in results, the low prevalence of HHV-8 infection in the normal Thai population is confirmed by these findings. This is in keeping with the observation of a very low incidence of KS in the Thai population, even during the HIV/AIDS epidemic. In Western countries, the incidence of KS has declined since the introduction of highly active antiretroviral therapy (HAART).6, 31 This may reflect a decrease in the incidence of new HIV and/or HHV-8 infections among homosexual men (the group with the highest risk of developing KS), or it may be the result of widespread therapy-induced slowing of the progression of HIV infection to clinical AIDS.

The worldwide incidence of NHL was increasing by 3–4% per year before the start of the HIV/AIDS epidemic, and this increase was more marked among older individuals (age ≥ 55 years).32 After the onset of the AIDS epidemic, the incidence of NHL in the United States increased more rapidly during the early 1990s but began to decline after 1995, when HAART was introduced.6 The increase and subsequent decrease in incidence were most marked for diffuse high-grade subtypes and more moderate for diffuse intermediate-grade subtypes of NHL, and were observed only among men age < 55 years.6

In Thailand, cases of NHL in adults8, 11 and children10 infected with HIV have been reported in case series from various parts of the country. Burkitt lymphoma and diffuse large cell lymphoma appear to be the most common types, whereas T-cell lymphoma occurred more rarely. Sukpanichnant et al.33 reported an increasing frequency of NHL cases diagnosed in the pathology department of Siriraj Hospital (Bangkok, Thailand) during the period 1993–1995 relative to the period 1957–1991. In that series of 255 cases, 71% were B-cell neoplasms (of which 75% were high- or intermediate-grade lesions), and there were 8 cases of primary CNS lymphoma; only 5 cases of AIDS-associated lymphoma were encountered.

The incidence of NHL in Thailand, like elsewhere in the region, is relatively low (age-standardized incidence between 1993 and 1997, 4.2 per 100,000).23 The increase in the incidence of all subtypes of NHL during the 13-year period from 1989 to 2001 was 5.9% per year, so that the absolute incidence rate almost doubled. The increase appeared to be attributable primarily to diffuse and high-grade subtypes, although throughout the study period, a large but constant percentage of registered cases (51.3%) were not subtyped. An increase of this magnitude might be anticipated as a result of the AIDS epidemic. In the United States and Europe, approximately 5–10% of all HIV-infected individuals will develop a lymphoma, and NHL is the AIDS-defining illness in approximately 3% of the HIV-infected population.34 Twenty-six thousand new cases of AIDS were identified in Thailand in 1999,21 but this figure probably is an underestimate, as UNAIDS estimated that there were 66,000 AIDS-related deaths in Thailand that year.1 With one-fifth of the national population covered by the sentinel cancer registries, the expected number of registered AIDS cases would be 5000–13,000, with 150–390 of these cases being associated with NHL (assuming that 3% of patients with AIDS develop NHL). Based on the observed incidence rates, the increase in the annual number of new cases encountered in the areas covered by the sentinel registries from 1989–1991 to 1999–2001 (i.e., 284 cases) would be of this order of magnitude. Alternatively, the HIV-attributable proportion of cases can be estimated as 50%, based on a population prevalence of 2% and a relative risk of 50.35

Nonetheless, we cannot unequivocally ascribe the increase in NHL incidence to the HIV/AIDS epidemic. The rate of increase was not higher in the provinces most affected by HIV/AIDS (Chiang Mai and Lampang). Furthermore, the most pronounced increase in incidence occurred among patients age ≥ 55 years, rather than among patients ages 20–44 years, despite the fact that 86% of all HIV cases in Thailand occur among individuals ages 20–44 years, and only 2.5% among individuals aged ≥ 55 years.21 There also was no evidence of an increase in the incidence of Burkitt lymphoma or CNS lymphoma, both of which are strongly associated with HIV/AIDS.2, 4, 7 It is difficult, therefore, to be certain as to whether the observed increase in incidence is attributable to the HIV epidemic or is part of the generalized increase observed in many other countries.35 Artifactual results caused by general improvements in cancer registration seem less likely, since the average annual increase in the incidence of all malignancies combined was only 0.65% between 1989–1991 and 1999–2001.

In conclusion, the epidemic of HIV/AIDS in Thailand has not resulted in an increase in the incidence of KS. Marked increases in the incidence of NHL (particularly high-grade/diffuse subtypes) are not focalized to the age groups affected by AIDS, nor are they more pronounced in the regions of highest HIV prevalence. Although HIV seroprevalence in the Thai population is relatively high, the low prevalence of other known (e.g., HHV-8) and unknown cofactors may underlie these findings.

Acknowledgements

The authors thank the directors and staff at cancer registries across Thailand for making their data available for the current study.

Ancillary