Introduction

Kashin-Beck disease (KBD), also known as “Big Bone Disease”, is a disorder affecting the bones and joints of the hands (including fingers), elbows, knees, and ankles of children and adolescents. Individuals stricken with the disease develop chronically stiff or deformed joints, and even short limbs and statures due to necrosis of the growth plates of bones and those of joint cartilage. Endemic KBD occurs mostly in China, southeastern Siberia and North Korea (Sokoloff 1989). China is the most seriously afflicted country in the world. It was reported that there are 13 provinces and 2 autonomous regions with prevalent KBD in China (Tan et al. 2002). The KBD patients in China are mainly farmers who live in remote and rural areas. According to the statistics of 1999, there are about 40 million people living in the KBD areas of China, and the total population of KBD patients is more than 1 million (Yang et al. 2006).

From the 1850s to the 1950s, Russian researchers investigated the prevalence of KBD in Russia (Hinsenkamp 2001). In China, the relevant research studies have been carried out since the 1950s. These studies focused mainly on environmental etiology of KBD, with an emphasis on the relationship between KBD prevalence and selenium levels in eco-environments or human body (Li et al. 1982, 1992; Deng et al. 1999; Guo et al. 1999; Xu et al. 2000; Zhang et al. 2001; Wang et al. 2003; Fan 2005), organic matter concentration in drinking waters (Wu et al. 1987; Peng and Xu 1988; Wang and Feng 1989; Zhang et al. 1990; Wang et al. 1991a, b) and severe contamination of grains by fungi-produced mycotoxin (Bi et al. 2001).

Due to the geographic distribution of the KBD-prevalent areas, there are few studies concerning KBD conducted by western researchers before the 1990s. Since the late 1990s, a group of Belgian researchers made a series of investigations on KBD in Tibet, China. Mathieu et al. (1997, 2001a, b) and Hinsenkamp et al. (2001) carried out clinical–radiological studies on the identification of KBD in Lhasa, Shannan, and Rikaze of Tibet. Pasteels et al. (2001) and Suetens et al. (2001) made histopathological studies of the KBD sufferers in Lhasa of Tibet, and considered that KBD is induced by the peroxidation of chondrocytes as the antioxidant system of human body is impaired. Moreno-Reyes et al. (1998, 2001, 2003) investigated the relationship among the serum selenium concentration, thyroid function, and KBD status in children with ages between 5 and 15 living in the villages around Lhasa, and evaluated the effects of selenium and iodine supplements on KBD control. Chasseur et al. (1997, 2001) and Haubruge et al. (2001, 2003) analyzed the impact of food crop contamination by fungi-produced mycotoxin on the KBD prevalence in Lhasa and Rikaze, and accordingly recommended some prevention and control measures for KBD.

From the above overview of previous work, one notes that much effort has been made to explain the prevalence of KBD in terms of environmental etiology and pathology. However, there are so far only a few systematic studies that have focused on the relationship between KBD occurrence and drinking water chemical composition. The aims of this study are: (1) to investigate the hydrochemistry of major and trace elements in drinking waters in four counties of Tibet as study areas where KBD is prevalent; and (2) to delineate any hydrochemical anomalies of the water samples collected from the study areas. This study will not only be a reference for future etiological studies of KBD, but help to guide the prevention and control of KBD in Tibet and other KBD-afflicted areas.

KBD in Tibet

Tibet is where KBD disease is the most widely distributed, serious and active in China (Institute of Kashin-Beck Disease of Harbin Medical University 2000), and there have been reports of KBD cases in all seven of its administrative divisions (Lhasa, Changdu, Linzhi, Shannan, Rikaze, Naqu and Ali). A survey made in the 1980s indicates that there were 13 KBD counties in Tibet (Wang et al. 1985) (Fig. 1). However, 20 new KBD counties were discovered during an investigation made in the late 1990s (Yang et al. 2003) (Fig. 1). Altogether, KBD disease was present in 379 KBD villages distributed in the above 33 counties (Gong et al. 2004a, b), and the number of individuals diagnosed with KBD was 18,000, equivalent to 0.69% of the total population of Tibet (Zaxi et al. 2003). The number of villages with KBD incidence rates greater than 40% reached 113 and accounted for 30% of the total of KBD villages (Zaxi et al. 2007). Moreover, children and adolescents who were less than 25-year old accounted for 50% of the total KBD patients in Tibet (Gong et al. 2004a). The most serious KBD symptoms, such as shortened limbs and statures, were observed in children 5–10 years of age (Institute of Kashin-Beck Disease of Harbin Medical University 2000).

Fig. 1
figure 1

Geographical distribution of the KBD counties in Tibet

Full size image

Sampling and analysis

The study areas selected for this investigation were the Sangri, Nimu, Xietongmen, and Gongbujiangda counties of Tibet. A total of 30 water samples were collected from springs, rivers, and streams that serve as drinking water sources for local residents. In each county, several KBD-affected and KBD-unaffected villages were chosen for drinking water sample collection and analysis. The sampling locations of all water samples were listed in Table 1.

Table 1 Sampling locations and hydrochemical types of water samples collected from the endemic and non-endemic villages of Sangri, Nimu, Xietongmen and Gongbujiangda counties
Full size table

All water samples were filtered through 0.45 μm membranes on site. Samples were collected in 300 ml polyethylene bottles, which had been rinsed with deionized water twice before sampling. For cation analysis, reagent-grade HNO3 with molar concentration up to 14 M was added to one sample collected at each site to bring the pH below 1. Unstable hydrochemical parameters including water temperature, pH, and electrical conductance (EC) were measured on site using a portable pH/EC/temperature meter. Alkalinity was measured using the Gran titration method (Appelo and Postma 1996). The concentrations of SO4 2−, Cl−, and NO3 − were determined on an unacidified sample by Ion Chromatography. Se was analyzed by Atomic Fluorescence Spectrometry, and other chemical constituents by ICP-MS within 2 weeks of sampling. The HCO3 −, CO3 2− and CO2 concentrations of all samples were calculated from the measured alkalinity and pH using PHREEQC. The hydrochemical properties and major ion concentrations of all 30 water samples are listed in Table 2, and the trace element concentrations in Table 3.

Table 2 Average, maximum, and minimum values of hydrochemical properties and major constituent concentrations of water samples (in mg/L except for water temperature, T, in °C, electrical conductivity, EC, in μS/cm and pH)
Full size table
Table 3 Average, maximum, and minimum values of trace element concentrations of water samples (in μg/L)
Full size table

Results and discussion

Major constituents

The hydrochemical ‘type’ of all collected water samples was determined from the concentrations of seven major constituents (HCO3 −, SO4 2−, Cl−, Ca2+, Mg2+, Na+, and K+), and are listed in Table 1. HCO3–Ca, HCO3–Ca·Na, or HCO3·SO4–Ca are the hydrochemical types of almost all water samples, except for samples GBJD04-A, XTM01 and XTM02 from the KBD-affected villages, which are SO4–Ca and HCO3–Ca·Mg type, respectively. In other words, there is basically no difference in hydrochemical type between water samples from KBD-affected areas and those from KBD-unaffected areas. In fact, the concentrations of most major constituents and the TDS values of the endemic samples do not show anomalies as compared to the non-endemic samples (see Table 2). For instance, the average TDS values of water samples from the KBD-affected villages of Sangri and Nimu County are 54.6 and 41.4 mg/L, respectively, lower than those from the KBD-unaffected villages (91.5 and 75.8 mg/L, respectively). However, in the KBD-affected and KBD-unaffected villages of Xietongmen County, the average TDS values of the drinking water samples were 129.2 and 71.3 mg/L, respectively, the former being evidently higher than the latter. In Gongbujiangda County, the average TDS values in drinking water are almost the same between the KBD-affected and KBD-unaffected villages (40.0 and 43.0 mg/L, respectively). It is worth noting that a Russian geochemist Vinogradov (1949) proposed that KBD is caused by deficiency of calcium and excess of strontium in drinking water and other ingested media in the environment. However, this proposed idea is not supported by this study. In Sangri, Nimu and Gongbujiangda counties, the average concentrations of both Ca and Sr in the endemic samples were lower than those in the non-endemic samples, whereas in Xietongmen County, the situation was the reverse.

Although there are no significant differences in the concentrations of most major constituents between the endemic samples and the non-endemic samples, Na, Al, and Fe are exceptions. It can be seen from Fig. 2a that the average Na concentrations in the water samples from the KBD-affected villages of Sangri, Nimu, Xietongmen, and Gongbujiangda counties are all less than those from the KBD-unaffected villages of the corresponding county. Similarly, in every county, the average Al and Fe concentrations in the endemic samples are higher compared to the non-endemic samples (Fig. 2b, c). The average Al and Fe concentrations of the endemic samples collected from Xietongmen County are especially high and 33.3 and 32.3 times those of the non-endemic samples, respectively. It is possible that these elements may have relations to the KBD occurrence in the study areas. Interestingly, M. Aiiso and K. Hiyeda have also indicated that chronic intake of Fe from drinking water or food may result in KBD, but the mechanism by which excessive Fe intake induces KBD has not been determined (Guo 2008).

Fig. 2
figure 2

Box and whisker plots of the Na (a), Al (b), and Fe (c) concentrations for water samples of different type. The meanings of type S-E, S-N, N-E, N-N, X-E, X-N, G-E and G-N are as follows. S-E samples collected from the endemic villages of Sangri County, S-N samples collected from the non-endemic villages of Sangri County, N-E samples collected from the endemic villages of Nimu County, N-N samples collected from the non-endemic villages of Nimu County, X-E samples collected from the endemic villages of Xietongmen County, X-N samples collected from the non-endemic villages of Xietongmen County, G-E samples collected from the endemic villages of Gongbujiangda County, G-N samples collected from the non-endemic villages of Gongbujiangda County. The boxes show the mean value minus standard error, the mean value, and the mean value plus standard error. The smallest and largest values are indicated by the small horizontal bars at the end of the whiskers. Bullet represents outlier values

Full size image

Selenium

Whether Se deficiency in the environment can cause KBD is still under debate. In 1972, Professor D. Mo of Xi’an Medical University of China reported for the first time that the selenium concentrations in drinking waters and food crops in some KBD areas were lower than those in KBD-unaffected areas (Mo et al. 1975). Since then, the relationship between KBD prevalence and selenium deficiency in environmental media or human organs has been intensively investigated in China. The most significant result of these studies is that KBD-affected areas of China are generally located along a northeast–southwest belt where the selenium concentrations of surface soils are usually lower than 0.125 mg/kg (Tan et al. 2002). However, there are many exceptions. For example, selenium levels are very low in Yulin and Luonan of Shaanxi Province, China, but there is no KBD occurrence in these two areas. Conversely, in many endemic KBD areas, such as Yidu of Shandong Province and Banma of Qinghai Province, soil selenium levels are relatively high.

To evaluate the relationship between selenium concentration of drinking water and KBD occurrence in Tibet, we compared the average selenium concentrations of the endemic water samples and those of the non-endemic water samples. The results indicate that although the average selenium concentrations in the water samples from the endemic villages of Sangri and Gongbujiangda counties (0.70 and 0.27 μg/L, respectively) are slightly lower than those from the non-endemic villages (0.84 and 0.31 μg/L, respectively), the endemic samples from Xietongmen County (1.03 μg/L) have higher average selenium concentrations as compared to the non-endemic samples (0.74 μg/L). In Nimu County, there is little difference in average selenium concentration between the endemic samples (0.22 μg/L) and the non-endemic samples (0.23 μg/L). In fact, no significant statistical difference in selenium concentration was found between the endemic water samples and the non-endemic water samples in each of the four counties (the P levels are much higher than 0.05). So the selenium concentration of drinking water does not correlate with the prevalence of KBD in the study areas. Further study should focus on inspecting whether or not the prevalence of KBD in Tibet results from selenium deficiencies in other environmental media (such as soil and food crop).

Other trace elements

Similar to Se, some other trace elements in drinking water, including Pb, Cr, Ni, V, Zn, As, and Li, do not correlate with KBD occurrence in the study areas. However, Mn in the drinking water samples is different. The average Mn concentration of the endemic samples is higher than that of the non-endemic samples (Fig. 3a). Specifically, the drinking water samples collected from the endemic areas of Sangri, Nimu, Xietongmen, and Gongbujiangda counties have average Mn concentrations of 2.3, 3.6, 37.3 and 4.2 times higher than those from the corresponding KBD-unaffected areas, which is in accordance with the research results of Voshchenko and Ivanov (1990). They concluded that manganese may play an important role in the etiology of KBD occurring in Russia, and proposed that excessive intake of manganese from drinking water can induce osteoclast activation, which in turn inhibits bone growth and causes KBD.

Fig. 3
figure 3

Box and whisker plots of the Mn (a), Cd (b), Co (c), Cu (d), Ba (e) and Mo (f) concentrations for water samples of different type. The meanings of the types and the legend of the figure are the same as in Fig. 2

Full size image

Besides manganese, the average concentrations of Cd, Co, Cu and Ba in the drinking water samples from the KBD-affected and KBD-unaffected villages were also compared, and the results are the same as for Mn. As Fig. 3b, c, d, and e shows, the water samples from the KBD-affected villages have higher average Cd, Co, Cu and Ba concentrations than those from the KBD-unaffected villages. Furthermore, contrary to Mn, Cd, Co, Cu and Ba, the average Mo concentration in the water samples from the KBD-affected areas of each county was lower than those from the corresponding KBD-unaffected areas (Fig. 3f). It is interesting that Cu and Ba were more enriched in drinking waters and soils in the KBD areas of Russia as compared to the background values as well (Muchkin 1967). However, whether or not the Cd, Co, Cu, Ba and Mo concentration anomalies in drinking water are related to the occurrence of KBD has not been ascertained through pathological studies. Thus, more investigations in other KBD areas are needed to further reveal the relationship between the prevalence of KBD and the Cd, Co, Cu, Ba and Mo concentrations in drinking water.

Conclusions

In Sangri, Nimu, Xietongmen and Gongbujiangda counties, there are no distinct differences in hydrochemical type between the drinking water samples from the KBD-affected villages and those from the KBD-unaffected villages. Almost all the water samples are of HCO3–Ca, HCO3–Ca·Na, and HCO3·SO4–Ca types. Also most of the major constituents in drinking water, including HCO3 −, SO4 2−, Cl−, NO3 −, CO3 2−, CO2, Ca, Mg, K, Sr, SiO2, and HBO2, are not correlated with KBD occurrence in the study areas. However, the Na, Al and Fe concentrations of drinking water in the KBD-affected villages of Sangri, Nimu, Xietongmen and Gongbujiangda were significantly different from that in the KBD-unaffected villages of the corresponding county, indicating that the anomalous concentrations of these constituents may be related to the prevalence of KBD.

The Se concentrations in drinking water are not correlated to the occurrence of KBD. In Xietongmen County, the average Se concentration of the endemic water samples is even higher than that of the KBD-unaffected samples. Among other trace elements, Mn, Cd, Co, Cu and Ba are of higher average concentration in the endemic samples than in the non-endemic samples, whereas the results for Mo are lower. In other words, the enrichment of Mn, Cd, Co, Cu and Ba and the deficiency of Mo in drinking water may be related to the prevalence of KBD in the study areas. However, the anomalous differences in these elements must be further substantiated in other areas, and pathological studies are required to substantiate any link between KBD disease and these elemental anomalies.