Prevalence of hand osteoarthritis and knee osteoarthritis in Kashin-Beck disease endemic areas and non Kashin-Beck disease endemic areas: A status survey

Osteoarthritis (OA) is a considerable health problem worldwide, and the prevalence of OA varies in different regions. In this study, the prevalence of OA in Kashin-Beck disease (KBD) and non-KBD endemic areas was examined, respectively. According to monitoring data, 4 types of regions (including none, mild, moderate and high KBD endemic areas) in Heilongjiang and Jilin provinces were selected. All local residents were eligible for inclusion criteria have undergone X-ray images of hands and anteroposterior image of knees. A total of 1673 cases were collected, 1446 cases were analyzed after removing the KBD patients (227). The overall hand OA and knee OA detection rates were 33.3% (481/1446) and 56.6% (818/1446), respectively. After being standardized by age, the detection rate of hand OA in the KBD endemic areas was significantly higher than that in the non-endemic endemic areas. Differently, there was no significant difference in the detection rates of knee OA between the KBD endemic areas and the non-endemic area. The correlation coefficient between the severity of OA and the severity of knee OA was 0.358 and 0.197 in the KBD and non-KBD endemic areas, respectively. Where the KBD historical prevalence level was higher, the severity of the residents’ hand OA was more serious. The detection rates of hand OA and knee OA increased with age. The detection rate of knee OA increased with the increase in body mass index. The prevalence of hand OA was closely related to the pathogenic factors of Kashin-Beck disease, and the prevalence of knee OA had no significant correlation with KBD pathogenic factors.

Introduction Detection equipment and methods Detection equipment. High frequency portable digital medical diagnostic X-ray image DR system (Beijing Longan Imaging Technology limited company).
(1) X-ray generator Exposure time: 0.01s-4.0s Focus size: 2.3mm mAs adjustment range: 0.5-160 mAs KV adjustment range 40-90 KV (2) Flat panel detectors Pixels: 2208×2688 Pixels Image size: 356×427mm (14×17 in) (3) Portable video workstation Computer model: Lenovo computer Y450 Detection method. Detection site: metacarpophalangeal joint, knee joint (1) Shoot the hands of the image: palm down, fingers close together, straight, flat on the DR system image receiving screen, wrist and hand to be kept in a straight line.
(2) Shoot the knees of the image: the legs straight, knee flat on the DR system image receiving screen.
Instrument parameter setting. The parameters of the all-digital multi-function radiography system (DR) were: (1) Shoot the hands of the image Exposure conditions: 50KV, 2.86 mAs Projection distance: 100 cm (2) Shoot the knees of the image Exposure conditions: 60KV, 4.20 mAs Projection distance: 100 cm

Quality control
In order to ensure the quality of the images, all abnormal X-ray images were examined by the experts; if the results were inconsistent with the image again, the findings were considered diagnostic. The results were combined and underwent statistical analysis.

Judgement standard
Investigate village selection criteria. KBD investigation of the village selection principle: ① history of being a KBD endemic area; ② KBD monitoring data complete; ③ X-ray positive detection rate of children <3%; ④ the population of residents over age 40 can not be less than 100 people.
Non-KBD investigation of the village selection criteria: ① no history of KBD; ② the population of residents over age 40 can not be less than 100 people; ③ similar eating habits and economic level to KBD villages.
KBD endemic area identification and division criteria. Mild endemic area: prevalence of the local residents of clinical I grade and above or X-ray detection rate of 10% in 7-12-year-old children.
Moderate endemic area: prevalence of the local residents of clinical I degrees and above or X-ray detection rate > 10% and 20% in 7-12-year-old children.
Severe endemic area: prevalence of the local residents of clinical I degrees and above or Xray detection rate> 20% in 7-12-year-old children.

KBD diagnostic criteria
Kashin-Beck disease was diagnosed in accordance with the "Kashin-Beck disease diagnosis" standard (WS / T 207-2010) for the diagnosis of Kashin-Beck disease. Osteoarthritis was diagnosed in accordance with the American College of Rheumatology 1995 diagnostic criteria.
KBD is an osteoarthropathy with joint injury of limbs. The X-ray image of hand can reflect the degree of pathological lesions and the damage of the joints in the whole body, and there is no difference between the change of large joints, such as the elbow in KBD and that in OA. Therefore, we can regard the X-ray image of hand as the basis of diagnosis and differential diagnosis. KBD patients with hand X-ray images often manifested as multiple, symmetrical finger joint thickening or short finger deformity, severe appearance of short stature. KBD patients with the characteristics of the hand X-ray images for basal broadening or spindleshaped of the proximal end of phalanx, horn mouth shape change of proximal end of phalanx, collapse repair phase of phalanx, which is not available in the hand X-ray of patients with osteoarthritis.

Image classification standards
Palm finger joint imaging classification criteria. Mild: ① joint stenosis; ② distal phalanx, proximal margin of the formation of the angle or small spur or small low density proliferation; ③subchondral bone sclerosis; ④subchondral cystic change; ⑤ cortical collapse; ⑥ carpal crowded, disorder. (Fig 1a) Moderate: ① distal phalanx, proximal margin of the formation of bone crest, bone spine; ② metacarpal head sickle shadow; ③ carpal edge hardening, not the whole or hyperplasia. (Fig 1b) Severe: ① distal brachial flexion and subluxation, metacarpophalangeal subluxation; ② metacarpal hypertrophy with sickle-like; ③ carpal deformation. (Fig 1c) Classification of knee imaging Mild: intercondylar uplift slightly ossification, no clinical symptoms or mild clinical symptoms. (Fig 2a) Moderate: intercondylar interria obvious ossification, joint space narrowing or uneven width, distal femur or proximal tibial lip-like hyperplasia, clinical symptoms were obvious. (Fig 2b) Severe: joint swelling and deformation, intercondylar intervertebral ossification prominent, joint space is cloudy, with narrow stenosis or uneven width, distal femur or proximal tibial liplike hyperplasia, difficult to walk. (Fig 2c) Statistical methods SPSS 18.0 statistical software was used for statistical analysis. The detection rate of arthritis was compared using a chi-square test. Furthermore, the difference between the two groups was analyzed using the card separation method; the mean number was chosen to be t test, and the test level was chosen to be α = 0.05.
The study was performed in accordance with the Declaration of Helsinki and approved by the Human Ethics Committee of Endemic Disease Center of Harbin Medical University, PR of China. Written informed consent was also obtained from the subjects S1 File. Table 1 provides the overall inspection results for each region. In this study, 1673 effective samples were collected from 12 villages in 2 provinces. A total of 227 KBD patients were detected, and the detection rate was 13.6%. Including KBD patients, hand OA was detected in 708 people, with a detection rate of 42.3%. Knee OA was detected in 1045 people, with a detection rate of 62.5%. Excluding KBD patients, the remaining sample included 1446 people; 481 of these patients were found to have hand OA, a detection rate of 33.3%. The detection rate in  the KBD endemic areas was significantly higher than in the non-KBD endemic area (χ 2 = 79.910, P < 0.001). There were 818 knee OA patients with a detection rate of 56.6%; the detection rate in the KBD endemic areas was significantly higher than in the non-KBD endemic area (χ 2 = 215.789, P < 0.001). Table 2 provides the basic characteristics of the subjects after removal of KBD.

Gender differences in hand OA and knee OA
In the KBD endemic area, the male detection rate of hand OA was 50.7% (143/282), and the female detection rate was 33.1% (174/526). In the non-endemic area, the male detection rate was 33.3% (75/225), and the female detection rate was 21.5% (89/413) (Fig 3). There were no statistically significant differences in gender composition between the two groups (χ 2 = 0.021, P = 0.885). Regardless of the status as endemic area or non-endemic area, the detection rate of hand OA in male patients was higher than in females (χ 2 = 23.932, P < 0.001; χ 2 = 10.590, P = 0.001).
In the KBD endemic area, the male detection rate of knee OA was 64.2% (181/282), and the female detection rate was 51.7% (272/526). In the non-endemic area, the male detection rate of knee OA was 64.4% (145/225), and the female detection rate was 53.3% (220/413) (Fig 4). Regardless of the status as an endemic area or non-endemic area, the detection rate of knee OA in male patients was higher than in females (χ 2 = 11.595, P = 0.001; χ 2 = 7.431, P = 0.006).

Age difference in hand OA and knee OA
There was a significant difference in the detection rate of hand OA between the KBD endemic area and the non-endemic area (χ 2 = 89.790, P < 0.001; χ 2 = 48.888, P < 0.001), and the detection rate of OA increased with age ( Fig 5). To compare the detection rate of hand OA at the same age between the KBD endemic area and the non-endemic area, the rate of hand OA detection in the 40-49-year-old age group was not statistically different (χ 2 = 2.602, P = 0.107); however, the hand OA detection rates of three age groups were higher in the endemic area than in the non-endemic area (χ 2 = 17.867, P < 0.001; χ 2 = 18.869, P < 0.001; χ 2 = 3.926, P = 0.048).
There was a significant difference in the detection rate of knee OA between the KBD endemic area and the non-endemic area (χ 2 = 35.248, P < 0.001; χ 2 = 38.908, P < 0.001), and the detection rate of OA increased with age (Fig 6). When comparing the detection rate of knee OA at the same age between the KBD endemic area and the non-endemic area, the detection rate of the endemic area was slightly higher than in the non-endemic area, but there was no significant difference in the detection rate of knee OA (χ 2 = 0.264, P = 0.608; χ 2 = 0.416, P = 0.519; χ 2 = 0.452, P = 0.501; χ 2 = 0.014, P = 0.904).

Relationship between knee OA and body mass index
By statistical test, the body mass index of knee OA patients was higher than in the non-knee OA group (t = -4.900, P < 0.001) ( Table 3). With an increase in BMI, the detection rate of knee OA gradually increased (Table 4).  Kashin -Beck disease endemic areas and hand OA and knee OA

Age of the standardized population hand OA and knee OA detection rates
The total number of people surveyed was the standard population. From this number, the number of expected detections and the standardized hand OA detection rate were calculated ( Table 5). The detection rate of hand OA in the KBD endemic area was higher than in the non-endemic area, and the difference was statistically significant (χ 2 = 79.910, P < 0.001).
After age was standardized (Table 6), the detection rate of hand OA in the four groups was statistically different (χ 2 = 656.092, P < 0.001). Hand OA was detected most frequently in the seriously KBD endemic area and was the lowest in the non-endemic area. Furthermore, the chi-square segmentation method was used to compare the two; the detection rate of hand OA in the seriously endemic area was higher than in the moderate endemic area (χ 2 = 185.856, P < 0.001). The detection rate of the moderate endemic area was higher than in the mild endemic area (χ 2 = 194.327, P < 0.001). The mild endemic area was higher than the nonendemic area (χ 2 = 191.186, P < 0.001).
After age was standardized (Table 7), the detection rate of knee OA in the KBD endemic area and the non-endemic area was no significant difference (χ 2 = 0.114, P = 0.736). The four regions were standardized according to age; the highest detection rate was in the moderate area, and the lowest detection rate was in the mild endemic area (Table 8). A further chi-square segmentation method was used to compare the two, and the detection rate of knee OA in the moderate area was higher than in the non-endemic area (χ 2 = 215.789, P < 0.001); the non-endemic area was higher than the severe endemic area (χ 2 = 183.721, P < 0.001); and the severe endemic area was higher than the mild endemic area (χ 2 = 276.237, P < 0.001).

Age of the standardized endemic areas and non-endemic area hand OA and knee OA severity of the detection of the situation
The detection rate of mild, moderate and severe hand OA in the KBD endemic area was 28.6%, 5.9% and 5.9%, respectively. The detection rate of mild, moderate and severe hand OA in the non-KBD endemic area was 19.2%, 3.8% and 1.8%, respectively. The detection rate of mild, moderate and severe hand OA in the two regions was statistically different (χ 2 = 34.655, P < 0.001; χ 2 = 7.165, P = 0.007; χ 2 = 32.612, P < 0.001), and the detection rate of mild, moderate and severe hand OA in the KBD endemic area was significantly higher than in the nonendemic area (Fig 7).
The two categories of regions were standardized by age, and the detection rate of mild, moderate and severe knee OA in the KBD endemic area was 37.0%, 18.9% and 0.9%, respectively. The detection rate of mild, moderate and severe knee OA in the non-endemic area was 46.8%, 8.6% and 0.6%, respectively. There were statistically significant differences in the detection rate of knee OA between the two groups (χ 2 = 28.639, P < 0.001; χ 2 = 64.820, P = 0.007), and there was no significant difference in the detection rate of severe knee OA (χ 2 = 0.733, P = 0.392). The detection rate of mild knee OA was significantly lower than in the nonendemic area, and the detection rate of moderate knee OA was significantly higher than in the non-endemic area (Fig 8).

Correlation between the severity of hand OA and the severity of knee OA
A list of 808 subjects in the wrist and knee X-ray reading results were analyzed, and the statistical analysis showed that the severity of hand OA in the KBD endemic area and the severity of knee OA were statistically different (χ 2 = 121.131, P < 0.001) ( Table 9). There was a linear trend relationship between the severity of hand OA and the severity of knee OA, and the correlation coefficient was r = 0.358.
A list of 638 subjects in the wrist and knee X-ray reading results were analyzed. The statistical analysis showed that the severity of hand OA in the non-endemic area and the severity of knee OA were statistically different (χ 2 = 30.424, P < 0.001) ( Table 10). There was a linear trend relationship between the severity of hand OA and the severity of knee OA, and the correlation coefficient was r = 0.197.

Discussion
The aim of this study was to investigate the effect of KBD pathogenic factors on the development of OA. The results of this survey showed that the detection rate of OA in the KBD endemic areas was higher than in the non-endemic areas, and the heavier the KBD history of the endemic areas, the higher was the detection rate of OA. The results were consistent with those of Yang Jianbo [11]. The detection rate of mild, moderate and severe hand OA in the endemic areas was significantly higher than in the non-endemic areas (Fig 7), indicating that KBD pathogenic factors have a significant impact on the occurrence and development of hand OA. KBD pathogenic factors can lead directly to KBD in childhood in some people, while KBD changes do not occur in early childhood in others. The elderly with pathogenic factors, in addition to advancing age and articular cartilage function decline are more likely to suffer from OA. This study also found that the detection rate of knee OA in the KBD areas was slightly higher than that in the non-endemic areas, but there was no significant difference, indicating that the effect of KBD pathogenic factors on the occurrence of knee OA was not obvious. The reason for this finding may be that KBD pathogenic factors are not significantly related to the injury of large joints, and factors such as lifestyle and labor intensity have a greater effect on knee OA. However, this study found that the incidence of mild knee OA was significantly lower than in the non-endemic areas, and the moderate knee OA detection rate was significantly higher than in the non-endemic areas (Fig 8), illustrating that pathogenic factors of KBD can impact the development of knee OA.
The results of this survey showed that the severity of hand OA was linearly correlated with the severity of knee OA in the same area, but the severity of hand OA was not representative of the severity of knee OA. Correlation between the severity of hand OA and the severity of knee OA in the KBD areas were higher than in the non-endemic areas, which may be because the prevalence of knee OA in the non-endemic area was higher than hand OA. In the KBD areas, KBD pathogenic factors on the injury of the hand and bone joints increased the severity of hand OA, although the effect on knee OA was not obvious. Therefore, the severities of hand OA and knee OA in the KBD areas were relatively highly correlated. Kashin -Beck disease endemic areas and hand OA and knee OA Estrogen protects articular cartilage, which can directly inhibit the damage of bone and cartilage from cytokines. Studies have shown that in women after menopause, the level of estrogen and progesterone in the body changes significantly, the balance between the two hormones is altered, and the prevalence of OA in postmenopausal women increases [13]. Rong et al showed that the average prevalence of OA was 11.1% and that the average prevalence in women was 26.8% [14]. The survey by Yu Wei et al also showed that the prevalence of knee OA was higher than men [15]. The results of an epidemiological survey of the elderly in six of China's cities by Li Ninghua et al showed that the prevalence of OA in women in all regions was higher than in men [16]. However, the results that found that the detection rate of hand and knee OA were higher in women, may have surveyed populations with relatively small numbers of healthy men, considering the male labor load in rural areas.
Domestic and foreign surveys show that age is a major risk factor for OA [17,18]. The greater the age, the greater is the risk of developing OA. There is a trend toward the development of OA characteristics in younger people, as in the study by Zang Changhai et al on the Taiyuan area, which found that after 40 years of age, the prevalence of OA significantly increased [19]. The OA survey of the rural elderly in Shanghai Fengxian by Wang Yongbin et al showed that with increasing age, OA prevalence increased. At 40-50 years old, the rate was 37.6%; at 50-60 years old, the rate had reached 62.7%; and at 60-70 years old and 70 years of age, the rates were 72.7% and 74.6%, respectively [20]. A survey of the Minhang District in Shanghai by Gu Ming Shi et al found 613 cases of OA in residents 61-81 years of age, translating to a knee OA prevalence of up to 60.5% [21]. The results also showed that the detection rate of hand and knee OA significantly increased with age. And the prevalence was heaviest in people over 60.
Obesity has long been identified as a risk factor for knee OA [22]. In a meta-analysis, those individuals who were obese or overweight had a 2.96 times higher risk of incident knee OA compared with those who were normal weight (95% CI 2.56-3.43) [23]. Assuming the prevalence of obesity in a hypothetical population to be 25%, the population attributable risk percent due to obesity would therefore be 29% (95% CI 24-34%); this rate would be higher where the prevalence of obesity is higher [24]. Furthermore, those individuals who were only overweight (not obese) had over 2 times the chance of developing knee OA compared with their normal weight counterparts [23]. Risk of incident knee OA increases with increasing BMI, regardless of knee alignment [25]. Decreasing BMI by 2 units or more over 10 years (~5 kg) was associated with a 50% lower risk of developing symptomatic knee OA among women [26], findings that were supported by a recent meta-analysis [27]. The duration of exposure to high BMI during adulthood confers risk of incident knee OA, suggesting the importance of weight control throughout life as a means of primary prevention of knee OA [28]. The results of this study also showed that the BMI of the knee OA population was higher than that of the non-knee OA population (t = -4.900, P = 0.000), and with the increase in BMI, the detection rate of knee OA gradually increased (Tables 3 and 4). Our study has some merits. The knee joint is one of the load-bearing joints that has a significant amount of daily activity, is a common joint affected by OA, and is therefore more commonly researched than hand OA both at home and abroad. To date, research on the prevalence of OA in the KBD areas has been minimal. In this study, we investigated the prevalence of OA in the KBD areas and the non-endemic areas by comparing the prevalence of OA in the KBD areas and the non-endemic areas and the differences in the severity of OA in each region, providing the foundation for research on the etiology of KBD and OA. At the same time, we also explored the relationship between the prevalence of knee OA and the prevalence of hand OA.  Kashin -Beck disease endemic areas and hand OA and knee OA Some limitations are present in this study. Because in rural areas, men 40-60 years of age go out to work more, the survey population therefore included a relatively small number of healthy men. In addition, according to historical survey data, there were more male KBD patients than female in the KBD areas.

Conclusion
The present study found that the residents' hand OA detection rate in the KBD area is higher than in the non-endemic area. Where the KBD historical prevalence level was higher, the residents' hand OA was more serious. There was no significant difference in the knee OA detection rates between the KBD endemic area and the non-endemic area. The detection rates of hand OA and knee OA increased with age. The detection rate of knee OA increased with an increase in body mass index. The incidence of hand OA was closely related to the pathogenic factors of KBD and the incidence of knee OA had no significant correlation with KBD pathogenic factors.
Supporting information S1