A total of 2,402 arseniasis patients in a few villages in Southwest Guizhou Buyi and Miao Ethnic Minority Autonomous Prefecture, China represent a unique case of endemic arseniasis related with indoor combustion of high arsenic coal (Jin et al. 2003; Zheng et al. 2005; Liu et al. 2002; Zhou et al. 1994, 1997). Unlike the situation in most other arseniasis endemic areas in this country and around the world as well, farmers in this area are exposed to indoor combustion of local high arsenic coal. The exposure was given via different routes, including inhalation of As polluted indoor air, ingestion of As-contaminated food and, also possible, dermal uptake.

Since the 1960s, as local forest had gradually faded out, farmers turned to burn local high As containing coal. The highest As concentrations in local coal were once detected as 3.2–3.5 wt% (Ding et al. 2001). That coal was used in poorly or unventilated traditional stoves (without chimney) for cooking, heating and drying crop and food over the stoves. Soon afterwards, in the early 1970s, hundreds of arseniasis cases appeared. All the cases concentrated in three small isolated areas in the prefecture. The township Jiao-le was the first one to be reported (877 cases, 1976). Most of the cases diagnosed and confirmed so far (1,386 out of 2,241 cases) are clustered in this township (Jin et al. 2003). The only overall epidemic investigation was held in 1991, when 8,958 persons in this township and its abutting villages were surveyed and 1,548 arseniasis cases were diagnosed and registered, representing a prevalence of 17.3% (Zhou et al. 1994, 1997).

This endemic community is a multiethnic one. Various ethnic minority groups have intermingled with local ethnic majority Han people (ethnic majority in China, represents 93% of the population in this country) in the same townships or even in the same villages for generations. The house architecture styles and many aspects of daily life in various ethnic groups have been largely harmonized. The intra-marriage (only marriage in the same ethnicity) is still strictly followed in some ethnic groups (e.g., Hmong).

Differences in individual susceptibility to chronic arsenic poisoning have been suggested (Vahter 2001). In our field survey conducted in 2002 and in 2004, remarkable clan-related (or ethnicity-related) differences in the prevalence of skin lesions (including palm or sole hyperkeratosis and, at a much lesser extent, spotted hyper or hypopigmentation of the body trunk) have been observed among the clans with different ethnic origin in one of the hyperendemic villages (Lin et al. 2003). For instance, ethnic Hmong clan P members displayed a much lower prevalence of skin lesions, compared with neighbouring Han clan G1 members (5.9 vs. 32.7%, OR 0.13, 95% CI 0.06–0.27, P = 3 × 10−10) although they have been exposed to indoor combustion of local high As coal for a quite similar time period (23.2 ± 13.5 years for ethnic Hmong clan P vs. 25.3 ± 12.9 years for Han clan G1, F = 2.181, P > 0.05; Lin et al. 2006). While the exact causes of this variability are still largely unknown, the possibility that genetic variations might play a role in the etiology was postulated. In our recent work, we observed that the status at the locus of GSTP1 A 1578 G (Ile 105 Val) in the ethnic Han clan members, but not in the Hmong clan members, modulated the susceptibility to arsenic-related skin lesions (Lin et al. 2006).

Glutathione S-transferases M1 and T1 play a key role in cellular detoxification, protecting biomacromolecules from attacks by reactive electrophiles. Due to its importance in cellular detoxification of xenobiotics, numerous studies on the association of GST M1 and T1 genotypes with risks of urinary bladder, breast, lung, colon, brain cancers and various non-malignant diseases have been reported (Bolt and Thier 2006; Parl 2005). In a study conducted in a total of 422 unrelated arsenic-exposed subjects (244 subjects with arsenic-related skin lesions and 178 subjects without arsenic-related skin lesions) in West Bengal, India, Ghosh et al. (2006) found that the population frequency of the GSTM1 0/0 genotype was significantly higher in the asymptomatic group. Individuals with GSTM1-positive genotypes (GSTM1 0/1 or 1/1) had significantly higher risks of arsenic-induced skin lesions. A possible protective role of GSTM1 0/0 genotype in arsenic toxicity was proposed by the authors. Hayakawa et al. (2005) reported that As–GSH complexes, such as arsenic triglutathione (ATG) or monomethylarsonic diglutathione (MADG), are substrates of the human Cyt19/ arsenic (+3 oxidative state) methyltransferase (AS3MT) that catalyzes the transfer of a methyl group from S-adenosyl-l-methionine to arsenic and produces monomethyl and dimethyl arsenicals. Thus, a study on the association of GSTT1 and M1 polymorphisms with the individual susceptibility to arsenic-related skin lesions as well as hair and urine As concentrations was conducted in two, relatively, genetically isolated ethnic clans in a remote mountain village as a part of a molecular-based epidemiologic investigation in this hyperendemic township.

Subjects and methods


The members of two clans of different ethnic origin (Han and Hmong) living for generations in an arseniasis-endemic village in Southwest Guizhou Ethnic Minority Autonomous Prefecture, Guizhou Province, China were included.

  1. 1.

    Ethnic Han clan G1. Intact clan genealogy records have been issued since mid-nineteenth century. The ancestor couple moved from adjoining Yunnan Province and settled in the village in 1822. At the time of investigation, 153 descendants (88 males, 65 females), divided into two lineages and four sub-lineages, still lived in the village. Our questionnaire investigation data showed that the marriage distance for most couples in the clan was within 10 km.

  2. 2.

    Ethnic Hmong (also named as “Miao” in China) clan P. The clan does not have a written family genealogy. The clan belongs to a special Hmong offset, called “Wan-Shu-Miao” (Bent-Comb Hmong). The name originates from a unique hairstyle. In former time, women of this offset inserted a bent wooden comb on their top hair bun. Their language is almost unintelligible for local Hei-Miao (Black Hmong) people, another major Hmong offset in the same county. There were 170 members (male 81, female 89) living in the village. The tradition of intra-marriage inside the Wan-Shu-Miao offset is still strictly kept. Daily life is largely influenced by their Han neighbors (clothing, housing, kitchen structure, arrangement in the bedroom, etc.). The clan P members are all bilingual and speak fluent local Han dialect, which is the common language in the area. Hmong language is used only inside the families.

In both clans, only the consanguineous members and their spouses were included in the present study. The other clan members who were adsorbed by adopting or any other non-consanguineous reason were excluded.

Collection of the samples

The environmental samples were collected simultaneously during the process of the questionnaire-based query in April 2004. Hair, urine and blood samples were donated by the investigated clan members on a voluntary basis. Informed consent was provided by the donors.

The present study, including its ethics aspects, was approved by local public health authority and met all the legal requirement of Chinese laws and regulations.

Analyses of environmental and biological samples

A spectrophotographic procedure using silver diethyldithiocarbamate (AgDDTC) as the complexing agent was applied for the determination of total arsenic concentration in all the environmental and human samples collected from the village. All the measurements were performed according to good laboratory practice as defined by the Chinese National Standards (GB) or the Technical Standard issued by Chinese Ministry of Health (WS/T) such as: “Water quality-determination of total arsenic–silver diethyldithiocarbamate spectrophotometric method” (GB 7485-1987) and “Determination of total arsenic in urine-silver diethyldithiocarbamate spectrophotometric method” (WS/T28-1996). Total As content in hair samples was determined according to the procedure specified in “Standard for identification of area of chronic arsenic poisoning caused by environmental arsenic pollution” (WS/T183-1999).

GSTT1 and GSTM1 genotyping

Genotyping of GSTT1 and GSTM1 using a one-tube procedure was described in detail elsewhere (Ma et al. 2002). In short, GSTM1 gene fragment of 215 bp was amplified (with a set of primers: 5′-GAA CTC CCT GAA AAG CTA AAG C -3′ and 5′-GTT GGG CTC AAA TAT ACG GTG G -3′) together with 480 bp fragment of 3′ part of GSTT1 gene (with a pair of primers: F1143: 5′-TTC CTT ACT GGT CCT CAC ATC TC-3′ and F1144: 5′-TCA CCG GAT CAT GGC CAG CA-3′A) and a 268 bp fragment of human β-globin gene serves as internal control (primers: PC04: 5′-CAA CTT CAT CCA CGT TCA CC-3′ and GH20: 5′-GAA GAG CCA AGG ACA GGT AC-3′). Reaction was conducted as initial melting 94°C for 5 min followed by 35 cycles consisting of melting 94°C, 30 s, annealing 60°C, 1 min, and extension 72°C, 1 min. A final extension taking 10 min at 72°C ends the procedure.

Statistic analyses

All the statistic analyses were conducted using the software package SPSS 10.0 (SPSS Inc., Chicago, Il). The data of polymorphisms of GST T1 and M1 as well as chemical analyses of As contents in various environmental and human samples were pooled. One-way ANOVA analysis was performed to identify significant differences of the variables, except those where the heterogeneity of variances was proved by Levene’s Test (P < 0.05). In the latter cases Mann–Whitney U test was used for non-parametric testing of two independent samples. The level of statistical significance was accepted as P < 0.05. Chi-square test was used to determine differences in population frequencies of assumed “risk” genotypes (homozygous “null” genotypes, GSTT1 0/0, M1 0/0 and the combination GSTT1 0/0 and GSTM1 0/0) between the diagnosed arseniasis patients and asymptomatic clan members. Stratification by ethnicity (clan) and gender was also made. P values less than 0.05 were regarded as significant. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were also calculated to estimate the risk due to different genotypes.


Total arsenic contents in environmental samples of the homes and in human biospecimens in the village

Drinking water samples were collected from all water sources for both clans. As concentrations detected were within the range of 0.014–0.025 mg/L (0.0181 ± 0.0049 mg/L) which meets the Chinese National Criteria for drinking water (threshold limit <0.05 mg/L, “The sanitary standard of drinking water”, GB5749-1985). However, the mean value detected slightly exceeded the WHO provisional guideline value for drinking water (threshold limit <0.01 mg/L) issued in 2004 (WHO 2004).

Average total As contents in exposed farmers’ (both clans combined) urine and hair samples were detected as 0.15 ± 0.22 mg/L and 11.58 ± 12.03 μg/g, respectively, which was a tremendous decrease, compared with the values recorded 13 years ago in the same village (Table 1). Unfortunately, the subjects investigated in the survey of 1991 cannot be tracked back any more. The exposure conditions in the village have been improved in the last decades or so, since a series of countermeasures have been implemented in the endemic villages by local authorities and some other agencies. Farmers in the endemic area have learnt a lot how to protect their families from the hazardous impacts of burning of high As coal inside the homes. The total urine As and hair As contents displayed no significant differences between the genders: average As content in urine samples: 0.099 (0.159) mg/L [median (interquartile)] for women versus 0.077 (0.129) mg/L for men, > 0.05; in hair samples: 12.79 ± 17.79 versus 11.79 ± 12.24 μg/g, P > 0.05. In the case of urine As, the heterogeneity of variances was proved by Levene’s Test (P < 0.05); Mann–Whitney U test was used for nonparametric testing of two independent samples (men and women).

Table 1 Arsenic content in human samples of different sampling years

Polymorphisms of GSTT1 and GSTM1

No statistic significant deviations in the population frequencies of the 0/0 genotype of GSTT1 and of GSTM1 could be found between the diagnosed arseniasis patients and asymptomatic individuals. This was also the case when combined homozygous deletion genotype GSTT1 0/0–GSTM1 0/0 which stands for a least conjugation activity was compared in both groups of villagers. The healthy Hmong individuals in clan P displayed a significantly higher presentation of GSTM1 0/0 carriers than their Han neighbors (71 versus 47%, OR 2.7, 95% CI 1.27–5.84, χ2 = 6.809, P = 0.009) (Table 2).

Table 2 Polymorphisms of GSTT1 and GSTM1 and prevalence of skin lesions among the members of ethnic Hmong clan P and Han clan G1

Comparison of total arsenic contents of biospecimens of members of ethnic Hmong clan P and Han clan G1 and its association with GSTT1 and GSTM1 genotypes

Significantly higher As concentrations in urine samples as well as in hair samples were found more frequently in GSTM1 0/0 carriers than in non 0/0 carriers (i.e., carriers of GSTM1 1/1 and 0/1 genotypes combined) in the village. In the case of GSTT1 genotype, no statistically significant deviations could be observed. The subjects with the lowest conjugation ability (i.e., with GSTM1 0/0 and GSTT1 0/0 status combined) showed elevated As contents only in hair samples, not in urine samples. When the stratification was made by clans and gender, only in ethnic Han clan G1 and in all male farmers (including the males of both clans), statistically significant higher hair As contents were confirmed in GSTM1 0/0 carrier subgroups. Significance was reached neither in Hmong clan P members nor in all female villagers (Table 3).

Table 3 Association of GSTM1 and GSTT1 polymorphisms and As body burden in two ethnic clans exposed to indoor combustion of high As coal


The homozygous deletion frequencies of GSTT1 and GSTM1 in asymptomatic members of ethnic Han clan G1 (as the control group of ethnic Han individuals) were detected as 41.8 and 47.3%, respectively which were within the normal range in general Han population (40–60%). These findings were comparable with the proportions of homozygous deletion of GSTT1 and GSTM1 in the normal Han population in different parts in this country, e.g. in normal Han people in Guangdong province (southern China, which is situated in the low-reaching area of the Pearl River, while the target arseniasis endemic area of this paper is in the up-reaching area of the same river basin), where GSTT1 0/0 was 48.5% (χ2 = 0.763, P = 0.382), and GSTM1 0/0 was 56.1% (χ2 = 1.355, P = 0.244) (Zhong et al. 2006), in the normal Han population in Shanghai (Yangtze Delta, in middle coastal area) where GSTT1 0/0 and GSTM1 0/0 were 48.4% (χ2 = 0.724, = 0.395) and 54.4% (χ2 = 0.860, P = 0.354), respectively (Shen et al. 1998), or with those in Beijing (northern China) where the proportions of GSTT1 0/0 and GSTM1 0/0 were reported as 51% (χ2 = 1.199, P = 0.274) and 44.9% (χ2 = 0.080, P = 0.777) (Wang et al. 2006). The genotype distribution in ethnic Han population in different parts of China seems to be highly homogeneous.

The GSTM1 0/0 frequency in asymptomatic ethnic Hmong clan P members was also found higher than that in normal populations of ethnic Han origin in southern China (71.0 versus 56.1%, OR 1.9, 95% CI 1.03–3.54, χ2 = 4.314, P = 0.038). It would be noteworthy to mention that the unusual high frequency of GSTM1 0/0 in healthy individuals of ethnic Hmong (in clan P) is comparable with that of 81.5% observed in a group of babies of Hmong origin in Minnesota, USA (Kiffmeyer et al. 2004) whose families have immigrated from Southeast Asian countries during the Vietnam War (71.0 versus 81.5%, χ2 = 3.165, P = 0.075). Their forefathers had left Chinese southwest provinces (including Guizhou province) and had moved south during the last couple of centuries. Nevertheless, no reference data for the very low proportion of GSTT1 0/0 carriers in normal Wan-Shu-Miao people are available so far.

In the present study, statistically significant elevated concentrations of total As either in urine or in hair were found to be associated with the GSTM1 0/0 genotype, but not with the GSTT1 0/0 genotype among the clan members of both ethnicities, Han and Hmong.

Chiou et al. (1997) have reported that in 115 farmers living in the southeast coastal area of Taiwan island and exposed to arsenic from artisan well water, GSTM1 0/0 carriers provided an increased percentage of inorganic arsenic in urine, while no significant associations were observed regarding arsenic contents in hair and toenails and GST M1 and T1 polymorphisms. Kile et al. (2005) reported a quite different result. They found in a population (n = 223 in 48 families) residing in an arseniasis-endemic region of Bangladesh individuals possessing GSTT1 0/0 genotypes had significantly higher As contents in their toenails, compared to GSTT1 wild-type individuals and suggested that GSTT1 modifies the relationship between inorganic As exposure and toenail As content. All three investigations provided disparate pictures. The different exposure situation in different populations might have caused the conflicting results. The real biological function of GSTT1 and GSTM1 proteins in the biotransformation of As species remains, largely, presumptive so far.

The data of the present investigation showed a marked decrease of body burden of total As (expressed as hair As content). It must be clearly stated that hair As contents detected in the present investigation do not represent the real internal As doses of the subjects at the time the skin lesions appeared or at the time of diagnosis.

The exposure route for indoor combustion of high As coal is much more complicated than that of As contaminated well water alone. The complicated exposure route consists of inhalation of polluted indoor air, ingestion of As contaminated grain (corn or rice) and other food (includes dried vegetables, home made hams and sausages, etc.) and, also possible, dermal uptake. The exposure parameters varied from house to house, from person to person and from time to time. No attempt for a reasonable assessment of real exposure dosage for each individual or for each family in the endemic village was ever made.