Although the first cases of AIDS in Thailand were reported among male homosexuals beginning in 1984, extensive transmission of HIV began among injecting drug users (IDU) in Bangkok in 1988 (1). By 1991, about a third of the estimated 36,000 IDUs in this capital city of 6 to 8 million people were infected with HIV (2). A much larger epidemic of heterosexual transmission began in 1989 (1). Transmission was facilitated, in large part, by male patronage of female prostitutes (1,3). By 1993 it was estimated that more than 80% of the 500,000 to 600,000 HIV infections in Thailand had resulted from heterosexual transmission (4). Through 1993, the cumulative numbers of persons with reported cases of AIDS and AIDS-related complex (ARC) were 6,829 and 2,730, respectively (5). Of these, 2,078 (30.4%) with AIDS had died (5). Little is known about the natural history and survival of AIDS patients in Thailand.
Most data on survival of AIDS patients are from developed countries (6-17). Limited data are available from South America (18) and Africa (19,20). There are no such data from Asia, where the HIV/AIDS epidemic is advancing in the 1990s (21). However, available data suggest that AIDS patients in developing countries have shorter survival than do those in developed countries (18-20). We conducted this study to describe the survival time from diagnosis of AIDS to death and to explore factors associated with survival in AIDS patients in the early phase of the epidemic in Thailand.
Data were abstracted from the medical records of all adult (≥13 years old) Thai nationals with AIDS who attended Bamrasnaradura Hospital, a public tertiary-care infectious disease hospital of the Ministry of Public Health, from February 1987 to February 1993. This hospital, one of the first to treat AIDS patients in Thailand, is located in a Bangkok suburb. We used the 1989 Thai AIDS case definition (22), which modified the 1987 Centers for Disease Control (CDC) definition (23) to include Penicillium marneffei infection as an AIDS-defining condition.
Data included age at AIDS diagnosis, sex, martial status, residence, reported HIV risk categories, presenting diagnoses, dates of first AIDS diagnosis and admission, history of zidivudine treatment, date and status of last follow-up, and cause of death. Laboratory data included complete blood counts and, for some patients, CD4 and CD8 counts by flow cytometry at the date of first AIDS diagnosis, defined as the date of a laboratory report confirming the presence of, or presumptive diagnosis of, an AIDS-defining condition. In the survival analysis, each patient was considered as a censored observation (24) at the time of last contact with the hospital if lost to follow-up or known to be alive on February 28, 1993.
Data were double-entered into a computer data base (Epi Infoversion 5.01b) (25) and corrected for errors and inconsistencies. To compare these patients with those reported nationally (26), goodness of fit chi-square (χ2) was used to compare distribution of sex and risk categories, and a 1-sample t-test was used to compare the mean ages (27). Kaplan-Meier survival function (24) was used to estimate survival time from AIDS diagnosis to death. Cox's proportional hazards model (28) was used for multivariable analyses to examine factors associated with survival. Cumulative hazard function was plotted against time to determine if the assumption of proportional hazard was met (29). zidovudine therapy was treated as a time-dependent variable in Cox's model to adjust for the variation in time of initiation of therapy related to AIDS diagnosis. Survival analyses were performed with EGRET (30) software.
We analyzed data from 329 Thai nationals representing 21.0% of Thailand's reported AIDS cases through February 1993 (26). Data from nine AIDS patients from other countries were not included. Of the Thai patients, 152 (46.2%) were known to have died. Compared with adult (aged ≥15 years) patients reported nationally (26), the AIDS patients at this hospital were slightly older (mean ± SD = 34.2 ± 10.4 years versus 32.3 ± 9.7 years, p < 0.01) and more likely to be IDUs (22.6% versus 10.0%, p < 0.01), but they were similarly likely to be male (93.0% versus 91.9%, p = 0.44). Eighty-nine patients (27.1%) were lost to follow-up; they were similar to those who completed follow-up with respect to age, gender, marital status, reported residence, reported HIV risk categories, number and types of presenting symptoms, and baseline laboratory data (p > 0.07 for any variable). For all patients, the median age was 31.5 years (range, 18-74); 306 (93.0%) were male, 174 (53.5%) were single, and 167 (50.9%) reported Bangkok as their place of residence (Table 1). About half (52.9%) presented with a single diagnosis. Only 127 (38.6%) were treated with zidovudine. Of all these patients, 212 (64.3%) had AIDS diagnosed on their date of admission, 74 (22.5%) within 1 week before admission, and 17 (5.2%) after admission. The most common diagnoses at the time of death among these patients were cryptococcal meningitis (26.3%), extrapulmonary tuberculosis (EPTB, 19.7%), Pneumocystis carinii pneumonia (PCP, 13.2%), wasting syndrome (5.3%), and other (35.5%).
The median lymphocyte count for 305 patients was 904 cells/μl (range, 0-6,565). Median CD4 cell counts for the 55 patients (16.7%) who were tested at AIDS diagnosis was 30 cells/μl (range, 0-915). The CD4 cell count was less than 100 cells/μl for 67.3%, 100 to 399 cells/μl for 20%, and more than 400 cells/μl for 12.7%. Patients whose CD4 cell counts were examined on their date of AIDS diagnosis were not different (p > 0.1 for any variable) from those without available counts with respect to the following variables: age, death rate, sex, area of residence, HIV risk categories, number of symptoms, zidovudine treatment, or baseline laboratory parameters.
Median survival time estimated by the Kaplan-Meier method was 7.0 months for all patients (range, 0 days to 42.8 months) (Table 1, Fig. 1); 1-year survival probability was 39.2% (95% confidence interval [CI] = 31.5-46.9%). The 2-year cumulative probability of survival was 19.9% (95% CI = 11.2-30.5%). It should be noted, however, that only about 10% of the original sample was under observation beyond 12 months. On bivariable analysis, improved survival (p < 0.05) was related to age and presenting diagnosis (Table 1). Patients aged 26 to 35 years had the longest survival times (median, 10.6 months). The presenting single diagnosis associated with the longest survival time (median, 19.9 months) was EPTB. The shortest survival (1.1 months) was associated with cryptococcal meningitis. Survival time was not significantly related to sex, marital status, reported residence, reported risk categories, number of presenting symptoms, zidovudine treatment, year of diagnosis, or CD4 and lymphocyte counts on bivariable analysis (p > 0.05).
On multivariable analysis using Cox's proportional hazards model, three factors were associated with survival: age, reported risk category, and presenting diagnosis (Table 2). Patients aged 26 to 35 years had longer survival (relative hazard [RH] = 0.61, 95% CI = 0.44, 0.85, referent: others). Patients with sexual risk factors survived longer (RH = 0.59, 95% CI = 0.40-0.78, referent: IDUs and others), as did patients with a single presenting diagnosis of EPTB (RH = 0.55, 95% CI = 0.35-0.86, referent: other initial diagnoses). Patients with a single diagnosis of cryptococcal meningitis had slightly higher mortality that approached statistical significance (RH = 1.50, 95% CI = 0.92, 2.44). Other variables, including zidovudine treatment, CD4 count, and lymphocyte count, were not associated with length of survival in this analysis (p > 0.05).
These data suggest that the survival time of AIDS patients attending this tertiary-care hospital in the early phase of the HIV epidemic in Thailand was shorter than that of patients in industrialized countries. The median survival time of 7.0 months was shorter than that of 11 to 20 months reported from the United States (7,10-13), United Kingdom (14,15), Italy (16) and 10 months from New Zealand (17). However, survival time in these Thai patients was similar to that of AIDS patients in Brazil (5.1 months) (18) and the Gambia (6 months) (20). A likely partial explanation for this observation in Thailand, suggested by the low total lymphocyte counts found at AIDS diagnosis, is late presentation to the health care system and late recognition of AIDS-related conditions.
Improved survival of patients in this study was independently associated with age (26-35 years), presenting a diagnosis of EPTB only, and sexually acquired infection. Better survival in patients 26 to 35 years old has also been observed in the United States (10,11), where higher mortality among very young and very old patients results in a “Ushaped” curve (8). In Thailand, this finding may reflect different age-specific treatment-seeking patterns, access to care, or some other factors not identified by our retrospective data.
A diagnosis of EPTB was associated with a two-fold longer survival than other conditions; a Spanish study found a similar outcome (9). Tuberculosis (pulmonary and/or extrapulmonary) has been identified as the most common diagnosis (58.3%) among Thai AIDS patients at this hospital (31). In many other countries, longer survival was associated with Kaposi's sarcoma or PCP (6-8,10-12,18). In the current study, the median CD4 cell count was 110 cells/μl for AIDS patients with EPTB but less than 30 cells/μl for other AIDS patients (p < 0.05). This may be a partial explanation for the longer survival of AIDS patients with EPTB.
Patients who reported sexual contact as their HIV risk also survived longer than IDUs and others. This finding is similar in some respects to the results from studies among homosexual men in the United States (6,11), and New Zealand (17) but contradicts data from Brazil (18), where IDUs survived longer. Although the HIV epidemic among Bangkok IDUs is due largely to HIV-1 subtype B (formerly Thai genotype B) and the Thai heterosexual epidemic is predominantly due to strains of subtype E (formerly Thai genotype A) (32,33), the available data are insufficient to attribute a difference in survival times to HIV-1 subtype.
This study was limited by the data available from medical records. Ascertainment of deaths may have been incomplete because of patients moving away from Bangkok; however, this should not bias our findings, as those who were lost to follow-up were demographically and clinically similar to those who remained in the study. Data on zidovudine treatment must be interpreted with caution because such treatment was not systematic; use of zidovudine was related to patients' ability to pay for the drug and their availability for clinic follow-up. Also, documentation of zidovudine treatment outside of this hospital may have been incomplete, but such treatment apart from this referral center was unlikely to have been common in the early phase of the Thai epidemic. Additionally, CD4 cell counts were available for only 55 patients (16.7%) because of the hospital's limited financial resources. Finally, these data from an urban tertiary-care hospital may not be representative of all of Thailand, as noted in age and risk category distributions; a large component of the epidemic is in rural areas, especially in the north.
In conclusion, we found the survival time of AIDS patients in the early phase of the epidemic in Thailand to be shorter than that of patients in industrialized countries and comparable to that of patients from other developing countries. This finding may be explained by late recognition of AIDS-related conditions in the early phase of an epidemic and delayed medical care for patients. To improve survival, greater attention to the early diagnosis and treatment of opportunistic infections is required. In particular, early diagnosis and treatment of tuberculosis, which is the most common diagnosis among these patients, are necessary to prevent a resurgence of HIV-related tuberculosis.
Acknowledgment: We thank Dr. Chana Tanchanpong, Director, Bamrasnaradura Hospital, for supporting this study and his staff for assisting with data collection; Dr. Alvaro Muñoz, School of Hygiene and Public Health, The Johns Hopkins University; Dr. Phillip I. Nieburg and Dr. Bruce G. Weniger, Division of HIV/AIDS Prevention, Centers for Disease Control and Prevention; Dr. Junya Pattara-arechachai, Department of Biostatistics, Faculty of Public Health, Mahidol University; and Dr. Kumnuan Ungchusak, Division of Epidemiology, Ministry of Public Health. We also thank Rameth Sinjermsiri for assisting with data management and Janjao Witta and Nuanphen Suk-aram for their help with graphics and manuscript preparation.
1. Weniger BG, Limpakarnjanarat K, Ungchusak K, et al. The epidemiology of HIV infection and AIDS in Thailand. AIDS
2. Mastro TD, Kitayaporn D, Weniger BG, et al. Estimating the number of HIV-infected injection drug users in Bangkok: a capture-recapture method. Am J Public Health
3. Mastro TD, Satten G, Nopkesorn T, Sangkharomya S, Longini IM Jr. Probability of female-to-male transmission of HIV-1 in Thailand. Lancet
4. Thai Ministry of Public Health. HIV/AIDS situation in Thailand—Update: December 1994
. Bangkok: Department of Communicable Disease Control, Ministry of Public Health, 1995.
5. Division of Epidemiology. Weekly Epidemiological Surveillance (Thailand)
6. Harris JE. Improved short-term survival of AIDS patients initially diagnosed with Pneumocystis carinii
pneumonia, 1984 through 1987. JAMA
7. Lemp GF, Payne SF, Neal D, Temelso T, Rutherford GW. Survival trends for patients with AIDS. JAMA
8. Rothenberg R, Woelfel M, Stoneburner R, Milberg J, Parker R, Truman B. Survival with the acquired immunodeficiency syndrome, experience with 5,833 cases in New York City. N Engl J Med
9. Martinez ML, Laguna F, Adrados M, Gazapo E, Garcia Aguado C, Gonzalez Lahoz JM. Description and survival analysis of 401 cases of AIDS in Madrid. Rev Clin Esp
10. Oregon Health Division. Survival after AIDS diagnosis in Oregon. CD Summary (State Health Department Newsletter)
11. Seage GR III, Oddleifson S, Carr E, et al. Survival with AIDS in Massachusetts, 1979 to 1989. Am J Public Health
12. Chang HG, Morse DL, Noonan C, et al. Survival and mortality patterns of an acquired immunodeficiency syndrome (AIDS) cohort in New York state. Am J Epidemiol
13. Chaisson RE, Stanton DL, Gallant JE, Rucker S, Bartlett JG, Moore RD. Impact of the 1993 revision of the AIDS case definition on the prevalence of AIDS in a clinical setting. AIDS
14. Johnson AM, Shergold C, Hawkins A, Miller R, Adler-MW. Patterns of hospital care for patients with HIV infection and AIDS. J Epidemiol Community Health
15. Whitmore-Overton SE, Tillett HE, Evans BG, Allardice GM. Improved survival from diagnosis of AIDS in adult cases in the United Kingdom and bias due to reporting delays. AIDS
16. Monteforte AD, Mainini F, Bini T, et al. Survival differences for 547 AIDS cases in Milan. J Acquir Immune Defic Syndr
17. Dickson NP, Sharples K, Carlson RV, McDermott J, Ellis-Pegler RB, Thomas MG. Outcome of adults with acquired immunodeficiency syndrome in Auckland 1983-9. N Z Med J
18. Chequer P, Hearst N, Hudes ES, et al. Determinants of survival in adult Brazilian AIDS patients. AIDS
19. Mbaga JM, Pallangyo KJ, Bakari M, Aris EA. Survival time of patients with acquired immune deficiency syndrome: experience with 274 patients in Dar-es-Salaam. East Afr Med J
20. Whittle H, Egboga A, Todd J, et al. Clinical and laboratory predictors of survival in Gambian patients with symptomatic HIV-1 or HIV-2 infection. AIDS
21. Editorial. AIDS: the third wave. Lancet
22. Division of Epidemiology. Weekly Epidemiological Surveillance (Thailand)
23. Centers for Disease Control and Prevention. Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome, MMWR
24. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc
25. Dean AG, Dean AJ, Burton AH, Dicker RC. Epi Info, version 5.01b:A Word Processing, Database, and Statistics Program for Epidemiology on Microcomputers
. Atlanta: Centers for Disease Control and Prevention, 1990.
26. Division of Epidemiology. Weekly Epidemiological Surveillance (Thailand)
27. Daniel WW. Biostatistics: A foundation for analysis in the health sciences, 5th ed.
New York: John Wiley & Sons, 1991;195-206.
28. Cox DR, Oakes D. Analysis of survival data
. London: Chapman and Hall, 1984;91-111.
29. Lee ET. Statistical methods for survival data analysis,
2nd ed New York: John Wiley & Sons, 1992;261.
30. EGRET version 0.26.6
. Seattle: Statistics and Epidemiology Research Corporation, 1993.
31. Tansuphaswadikul S, Limpakarnjanarat K, Lohsomboon P, Weniger BG. The clinical presentation of AIDS in Thailand. In: Proceedings of the VIII International Conference on AIDS. Amsterdam, July 1992 (abstract PoC 4078).
32. Ou C-Y, Takebe Y, Luo C-C, et al. Wide distribution of two subtypes of HIV-1 in Thailand. AIDS Res Hum Retroviruses
33. Ou C-Y, Takebe Y, Weniger BG, et al. Independent introduction of two major HIV-1 genotypes into distinct high-risk populations in Thailand. Lancet