Professional Documents
Culture Documents
Inter 3
Inter 3
131–136
Copyright 䉷 2001 by The American Society of Tropical Medicine and Hygiene
Abstract. Blood samples were collected from healthy subjects, aged 9 months–29 years in urban and rural com-
munities from 4 distinct regions in Thailand, to determine the seroprevalence rate of varicella-zoster virus (VZV)
antibody and its relationship with demographic, climatic, and socioeconomic factors. The overall seroprevalence rate
was 52.8% and increased from 15.5% in the 9-month to 4-year-old group to 75.9% in the 20–29 year-olds. The age-
adjusted seroprevalence was significantly higher in the cooler than in the warmer regions. In the warmer regions
only, the age-specific seroprevalence was significantly higher in the urban population than in the rural population. In
Thailand, climate is the main determinant of VZV seroprevalence. The delayed onset of natural immunity is more
marked in warmer climate areas. Population density is a secondary determinant; in the warmer areas, the pattern of
adolescent and adult susceptibility was greater in rural than in urban areas.
131
132 LOLEKHA AND OTHERS
1,015
243
279
251
242
Urban
20,000, 20,001–50,000, ⬎ 50,000 baht (now is @43 baht ⫽
US$1), as well as any history of chickenpox. As most of the
Address
information regarding socioeconomic background in the
Northeastern area was not completed, this area was excluded
from the analysis for this parameter. A positive varicella his-
1,078
244
297
287
250
Rural
tory was defined as either personal recollection (memory) or
hearsay (being told that he or she had varicella in childhood).
Blood samples were taken from each subject, and sepa-
rated serum was stored at ⫺20⬚C until testing. The testing
was done at the Laboratory of Pediatric Infectious Unit, at
the Ramathibodi Hospital of Mahidol University in Bang-
Male
229
215
229
198
871
kok, Thailand. Specific VZV immunoglobulin (Ig)G anti-
bodies were determined using a commercial enzyme-linked
Gender
immunosorbent assay (ELISA), Enzygnost anti-VZV/IgG
(Dade Behring, Marburg Germany) according to the manu-
facturer’s instructions.
Female
1,222
258
361
309
294
Delta absorbance readings of ⱖ 0.2 were taken as positive
and indicated immunity to VZV infection. Values ⬍ 0.1 were
considered negative and indicated susceptibility to VZV in-
fection. Values between 0.1 and 0.2 obtained twice on the
20–29 yr
96
121
99
95
411
All data were entered into an Oracle database, and statis-
tical analysis was performed using Statistical Analysis Sys-
tem (SAS), version 6.12 (SAS Institute Inc., Cary, NC). Uni-
variate logistic regression models were built with each ex-
planatory variable. Significance (P ⫽ 0.25) of explanatory
15–19 yr
98
108
102
97
405
variables was tested using the loglikelihood statistic. A mul-
TABLE 1
10–14 yr
96
120
117
106
439
RESULTS
100
114
118
94
426
subjects
2,093
487
576
538
492
of
Center
South
North
ters and 4 study regions in the year 1997 are shown in Table
4 and Table 5.
TABLE 2
Varicella-zoster virus (VZV) prevalence by age group and centers
VZV sero- 95% CI VZV sero- 95% CI VZV sero- 95% CI VZV sero- 95% CI VZV sero- 95% CI
Age prevalence prevalence prevalence prevalence prevalence
(yr) (%) LL UL (%) LL UL (%) LL UL (%) LL UL (%) LL UL
9 months–4 11.3 5.8 19.4 6.2 2.5 12.3 33.3 24.3 43.3 12 6.4 20.0 15.5 12.2 19.4
5–9 36.0 26.6 46.2 26.3 18.5 35.4 72.9 63.9 80.7 54.3 43.7 64.6 47.7 42.8 52.5
10–14 51.0 40.6 61.4 45.8 36.7 55.2 75.2 66.4 82.7 71.7 62.1 80.0 61.1 56.3 65.6
15–19 50.0 39.7 60.3 58.3 48.5 67.8 80.4 71.4 87.6 65.0 54.6 74.4 63.5 58.6 68.2
TABLE 3
Age-specific varicella-zoster virus (VZV): prevalence rate by center, in an urban or rural setting
VZV sero- VZV sero- VZV sero- VZV sero- VZV sero-
prev- 95% CI prev- 95% CI prev- 95% CI prev- 95% CI prev- 95% CI
Age alence alence alence alence alence
Address (yr) (%) LL UL (%) LL UL (%) LL UL (%) LL UL (%) LL UL
Rural 9 mo.–4 8.51 2.36 20.38 7.35 2.43 16.33 26.42 35.26 40.33 12.73 5.27 24.48 13.45 9.26 18.64
5–9 24.49 13.34 38.87 13.79 6.15 25.38 81.84 69.97 90.08 61.36 45.50 75.64 46.30 39.51 53.19
10–14 41.67 27.61 56.79 35.09 22.91 48.87 79.03 66.82 88.34 76.92 63.16 87.47 58.90 52.08 65.49
15–19 26.53 14.95 41.08 45.83 31.37 60.83 80.77 67.47 90.37 66 51.23 78.79 55.28 48.08 62.31
20–29 60.78 46.11 74.16 71.21 58.75 81.70 83.64 71.20 92.23 87.76 75.23 95.37 75.57 69.35 81.08
Urban 9 mo.–4 14 5.82 26.74 4.44 0.54 15.15 40.82 27 55.79 11.11 3.71 24.05 17.99 12.79 24.22
5–9 47.06 32.93 61.54 39.29 26.50 53.25 62.26 47.89 75.21 48 33.66 62.58 49.05 42.10 56.02
10–14 60.42 45.27 74.23 55.56 42.49 68.08 70.91 57.10 82.37 66.67 52.53 78.91 63.18 56.44 69.57
15–19 73.47 58.92 85.05 68.33 55.04 79.74 80 66.28 89.97 63.83 48.52 77.33 71.36 64.67 77.43
20–29 71.11 55.69 83.63 69.09 55.19 80.86 79.55 64.70 90.20 86.96 73.74 95.06 76.32 69.62 82.17
CI ⫽ confidence interval; LL ⫽ 95% confidence interval lower limit; UL ⫽ 95% confidence interval upper limit.
133
134 LOLEKHA AND OTHERS
FIGURE 1. Varicella-zoster virus prevalence by age and region in four study areas.
Association between history and seropositivity. Avail- increasing to 75.9% in the 20–29-year-old group. These val-
able data on 705 subjects showed an association between the ues are similar to those obtained in the Philippines9 and in
past history of VZV and seropositivity. There was a positive a previous study in Bangkok, Thailand, where the seroprev-
correlation between a history of varicella and seropositivity; alence was 69.2% at 20–29 years of age and increased to
85.2% (601 of 705) of the participants who could recall a 96.1% at 30–39 years of age.13 Consequently, our study was
clinical history of chickenpox tested positive for VZV anti- restricted to subjects up to 29 years of age. Varicella vaccine
bodies. However, for 33.9% (425 of 1,255) of subjects who has been available in Thailand only since 1997 and is limited
were seropositive, there was no history of clinical varicella. to the private market with no current plans for inclusion in
Analysis of data in a logistic regression model for the the national expanded program of immunization (EPI) pro-
Central and Southern areas showed age (P ⬍ 0.005), urban gram.
or rural address (P ⬍ 0.005), and previous occurrence of It has been suggested that the transmission potential of
varicella (P ⬍ 0.005) as significant variables but not income/ the VZV virus might be adversely affected by a combina-
socioeconomic background (P ⬎0.05). tion of high ambient temperatures and humidity in tropical
regions.12 This is supported by the seasonal and regional
DISCUSSION variations in acute disease found within some Southeast
Asian countries. A review of data from Southeast Asia
The overall age-specific VZV seroprevalence rate in our showed that the peak incidence of varicella infection oc-
study is similar to previous reports from other tropical coun- curred during cooler months and in cooler, more temperate
tries in which seroprevalence reaches ⬎ 90% only in those regions.11 In Kerala, India, the highest levels of varicella-
over 30 years of age.9–12 Among the 15–19-year-old group, associated mortality occur in January, the coolest part of
less than 65% of the study population were seropositive, the year.12 In Thailand the 1995, 1996, and 1997 Annual
TABLE 4
The geographic data of the four study centers in the year 1997
TABLE 5
The climatological data of the four study centers and four regions in the year 1997
†Mean †Mean
*Mean *Mean extreme extreme
maximum minimum maximum minimum ‡Annual ‡Mean
temperature temperature temperature temperature rainfall humidity
Study site/region (⬚C) (⬚C) (⬚C) (⬚C) (mm) (%)
Epidemiological Surveillance Reports showed a seasonal 22–23⬚C) are comparable with the tropical climate of Central
pattern, with a higher incidence in the cooler months of the and Southern areas. However, in contrast, the population
year.16–18 These surveillance reports also included reported density in Calcutta, West Bengal, (5,400/square kilometer)21
cases by age group in each of the study areas (Table 6). is higher than in Thailand and may thus have contributed to
Between 1995–1997, some 35%–40% of reported cases the high seroprevalence rate observed by Mandel and oth-
were of adolescents aged 15 years or more in the Central ers.15 It is widely accepted that there is a high correlation
and Southern areas compared to 15%–20% in the Northern between a history of varicella infection and seropositivi-
and Northeastern areas.16–18 The results of our study are ty,9,10,13,22 and this was also the case in our study. In line with
comparable with these reports. We found that the Northern another study from Thailand,13 we found family income was
area had the highest seroprevalence rate of varicella infec- not significantly associated with the varicella seroprevalence
tion in all age groups. The susceptibility in adolescents and rate.
adults was still higher in the Northern area than in temper- In conclusion, similar to other reports from tropical
ate countries. Our data confirm the significantly higher se- countries, the prevalence pattern showed seroconversion in
roprevalence rates in the Northern as compared to the late childhood or young adulthood, which is often associ-
Southern and Central area of Thailand. With its higher al- ated with greater morbidity and mortality. Furthermore, this
titude, the cooler Northern areas are more comparable to study clearly showed a regional variation most likely due
temperate regions. Assessing the climatological character- to climatic differences in 4 regions of Thailand. In the
istics of the 4 regions, the mean of the extreme minimum warmer areas a marked delay in natural infection was found
temperature (mainly due to relative altitude) is the most with high extreme minimum temperature being the most
important factor contributing to an increase of VZV sero- critical climatic factor. In addition, especially in these re-
prevalence. In fact the Southern area with the highest mean gions, the pattern of adolescent and adult susceptibility to
extreme minimum temperature of 22⬚C had the lowest se- VZV is more distinct in rural areas. Thus, our study results
roprevalence rate of 41.7% decreasing to 42.7% in the Cen- suggest that increased adolescent and adult susceptibility to
tral area (19.2⬚C), 57.9% in the Northeastern area (12.3⬚C) VZV in the tropics is related to both climate and population
and 69.0 % in Northern areas (10.1⬚C). density.
The second most important factor of delayed onset of var- Consequently, a vaccination program in Thailand and in
icella was rural versus urban living. The most likely expla- other tropical countries should include administration to
nation for this is population density. In urban settings, higher susceptible adolescents and adults who are at high risk of
population densities may partially overcome the transmis- developing severe varicella with complications. Routine
sion-interrupting effect of a tropical climate. In India, prior varicella vaccination programs should also be directed at
to large scale immunization, a similar phenomenon was de- children, in whom seroprevalnce is low, accessibility is
scribed for measles,19 which is even more infectious than generally high, and one dose of the vaccine is sufficient.
varicella. The empirical observation of varicella affecting
adults in rural settings of tropical countries dates back to
Acknowledgments: We are grateful to Miss Dusadee Chareonpipop
1976;20 however, epidemiologic comparisons with urban for performing the anti-VZV serology assay.
populations were published only 20 years later.15 In a study
from West Bengal, India, nearly all urban adults were im- Financial support: This study was supported by a grant from
SmithKline Beecham Biologicals, Rixensart, Belgium.
mune, and only adults in rural areas were susceptible to var-
icella infection.15 The authors concluded that the delayed on- Authors’ addresses: Somsak Lolekha, Chief of Infectious Disease,
set of varicella in tropical areas is a purely rural phenome- Department of Pediatrics, Faculty of Medicine, Ramathibodi Hos-
pital, Rama VI Road, Mahidol University, Bangkok 10400, Thai-
non. The current study of a larger sample size confirmed the land. Watcharee Tanthiphabha, Department of Pediatrics, Faculty of
increased risk for delayed-onset varicella in rural populations Medicine, Chiang Mai University, Chiang Mai 50002, Thailand.
of tropical areas. The high seroprevalence rate in urban West Penpark Sornchai, Pediatrics Unit, Nakornping Hospital, Maerim
Bengal of 96% at the age of 25 years can obviously not be District, Chiang Mai 50002, Thailand. Pensri Kosuwan and Sumit
Sutra, Department of Pediatrics, Faculty of Medicine, Khon Kaen
explained by impact of climate alone as both maximum and University, Khon Kaen 40002, Thailand. Boonyarat Warachit and
minimum temperatures of West Bengal (average maximum Suda Chup-Upprakarn, Pediatrics Unit, Had Yai Regional Hospital,
temperature is 37–38⬚C, average minimum temperature is Had Yai, Song Kla 90110, Thailand. Yanee Hutagalung, John Weil
136 LOLEKHA AND OTHERS
14,332
3.30
21.40
39.10
19.20
11.20
5.70
1997
(%)
–
Reprint requests: Somsak Lolekha, Chief of Infectious Disease, De-
partment of Pediatrics, Faculty of Medicine, Ramathibodi Hospital,
Mahidol University, Rama VI Road, Bangkok 10400, Thailand Tel/
fax 662-201-1674, E:mail address: rasll@mahidol.ac.th.
Northeast
14,105
3.40
21.30
41.90
19.30
9.20
4.80
1996
(%)
–
REFERENCES
15,842
3.50
20.00
38.40
22.30
10.10
5.70
0.03
135: 896–899.
1995
(%)
3.40
24.70
33.40
19.90
12.10
6.50
9,698
851.
8. Taylor-Wiedeman J, Yamashita K, Miyamura K, Yamazaki S,
1989. Varicella zoster virus prevalence in Japan: no signifi-
cance change in a decade. Jpn J Med Sci Biol 42: 1–11.
3.10
16.80
26.00
17.90
20.90
15.20
0.01
6,825
9. Barzaga NG, Roxas JR, Florese RH, 1994. Varicella zoster prev-
1997
TABLE 6
(%)
5,220
South
1996
(%)
5,352
1995
(%)
9,698
1997
⬍25