Introduction

Tibial plateau fractures account for 1–2% of the whole body fractures and 8% of geriatric fractures [1]. High-quality epidemiological studies can provide us essential cognitive understanding about its distribution, incidence, related risk factors and dominant affected type, and can further guide clinical treatments and improve the prognosis. China exists within a unique geopolitical context such that its huge, diverse landmass surely leads to various epidemiologic fracture features. Tibial plateau fracture was one of the severe fractures with which millions of patients were reported fully or partially disabled due to traffic injuries during recent decades [2]. Nilson et al. revealed latitude as well as UV radiation had statistically significant correlations with incidence of hip fracture in Sweden, and increased risk of hip fracture in the northern parts of Sweden was higher than the middle and southern parts [3]. Authors in our country hold the view that higher elevation and longer sunlight hours may promote spine fractures [4]. Other articles revealed bone density disparities in males and females at different ages which related to osteoporosis fracture. Some authors even pointed out that lower limbs fractures manifest intensively with regional characteristics and time difference, and we had reasons to suspect a potential relationship between altitude and tibial plateau fracture. However, no research focusing on tibial plateau fractures and altitude was done; moreover, a nationwide epidemiological study has not been demonstrated in China. Thus, our study aimed to identify the overall epidemiology characters of adult tibial plateau fractures and reveal potential epidemiological characteristics of adult tibial plateau fractures in different altitudes in China on a multicentre level.

Methods

This epidemiology study was approved by The Institutional Review Board of the third hospital of Hebei Medical University in compliance with the Helsinki Declaration. And our study was a retrospective study based upon historical X-ray films, no human participants were included. Written informed consent of each participation was not necessary. All adult fractures (≥16 years) were collected through the PACS system and case reports checking system, which were referred to 83 hospitals through multi-stage random sampling from 31 provinces in China between January 2010 and December 2012. A total of 414,935 patients (431,822 cases) were included, from which 6,227 adult tibial plateau fractures were screened and met our final eligibility criteria. For further comparison, we divided all cases into four groups based on the centre altitudes of each city, G1 = plains group (<500 m), G2 = hills group (500–1000 m), G3 = mountain areas group (1000–2000 m), G4 = plateau group (>2000 m). Age was categorized at an interval of five years and a series of 14 units were produced: 16–20 years, 21–25 years, 26–30 years, 31–35 years, 36–40 years, 41–45 years, 46–50 years,51–55 years, 56–60 years, 61–65 years, 66–70 years, 71–75 years,76–80 years and ≥80 years. Initial radiographs were reviewed and classified based on OA/OTA classification. A “41” was assigned for proximal tibial fractures and types A, B, C were classified according to the anatomic location of fractures, thereinto tibial plateau fracture represented type B and type C, and two types were subdivided into three subgroups described as 1,2,3 based on the severity of fracture. Eligibility criteria were patients over 16 years old, and definite tibial plateau fracture, i.e. those with unequivocal imaging data that can be diagnosed with tibial plateau fracture and Arbeitsgemeinschaft für Osteosynthesefragen (AO) type can be confirmed. Exclusive criteria were patients under 16 years old, imaging data absence, obsolete fracture, pathological fracture, periprosthetic fractures, and patients under unambiguous or suspicious diagnosis.

All images were identified independently by four orthopaedic specialist registrars on two occasions separated by a one-week interval based on OA/OTA classification. Assessments should maintain consistency. Once disagreement occurred, it would be transferred to a superior physician. To ensure accuracy, three arbiters including two senior orthopaedic surgeons and one senior radiologist regularly sampled radiographs. A questionable imaging would roll back to assessment directly. Quality standards were: misjudgment under 1% of each researcher (500 cases/1 researcher), misclassified ratio under 3%.

Statistical analysis

Statistical analyses were conducted by standard statistical software (SPSS Version 21.0, Chicago, IL, USA). The descriptive analyses are presented in Tables 1, 2 and 3. Continuous variable age was expressed as median and range and the test was performed using Mann-Whitney U-test. Discontinuous variables regarding age, sex and AO type were expressed as count, and chi-square test was performed. P < 0.05 was considered as statistically significant.

Table 1 Gender and AO type in the four groups
Full size table
Table 2 Age–sex distribution in the four groups
Full size table
Table 3 AO type in the four groups
Full size table

Results

A total of 6,227 adult tibial plateau fractures were collected, including 4,275 males and 1,952 females, 2704 were on the right and 3,523 were on the left. These fractures encompassed 56.90% of all proximal tibia fractures and 1.50% of the whole body fractures. The highest proportion of adult tibial plateau fractures were patients in the age of 40 to 44 years, males more than females. Peak age in males was 40–44 years and we should note a different peak age in females of 55–59 years.

Type 41-B was the most common type of adult tibial plateau fractures (3,998 cases, 64.20%), followed by 2,229 cases of type 41-C (35.80%). Incidence detailed in six subgroups were 1,435 cases in type 41-B1 (23.04%), 1375 cases in type 41-B2 (22.08%), 1206 cases in type 41-B3(19.37%), 595 cases in type 41-C1 (9.56%), 747 cases in type 41-C2 (12.00%), and 887 cases in type 41-C3 (14.24%).

Epidemical characteristics of the four groups

There were 3,178 males and 1,340 females in group 1, 632 males and 380 females in group 2, 672 males and 471 females in group 3, and 21 males and four females in group 4. The average ages were 45.37 ± 14.3, 46.33 ± 8.19, 46.50 ± 6.63, and 46.28 ± 8 years old, respectively. Difference were considered statistically significant at p < 0.05 (Table 1).

Age distribution showed no statistically significant difference in the four groups (χ2 = 44.439, P = 0.253), while same peak age showed 40–44 years in males and 55–59 years in females. Sex distribution of the four groups had statistically significant difference at 25–29, 30–34, 40–44, 65–69 intervals (P < 0.05). Note an interesting gender-inversion among patients over 60 years in the four groups, indicating an elderly female dominance and suggesting that a number of female patients gradually exceeded male patients with increasing age (Table 2, Figs. 1 and 2). Number of adult tibial plateau fractures, sex ratio as well as age distribution showed downward tendency with increased altitudes.

Fig. 1
figure 1

Frequency distribution of age range of 6,227 adult tibial plateau fractures (five-year interval)

Full size image
Fig. 2
figure 2

Sex–age distribution of 6,227 cases and four groups

Full size image

Gender-AO type distribution

Our conclusions from the 6,227 cases were that 41-B was the predominant type, and the most affected subgroup type was 41-C2 in males and 41-B2 in females (Table 3). Ratio of gender in type 41-B was higher than type 41-C, demonstrating males prone to be severely damaged.

There were statistically significant differences in AO types in the four groups, and a comparison between 41-B and 41-C was χ2 = 8.563, P = 0.036, while intergroup comparison among the six subgroups showed χ2 = 66.145, P = 0.000. Dominant AO types the of four groups were 41-B1 in group 1, 41-B2 in group 2 and 3, 41-C2 in group 4, respectively. Sex distribution of the six subgroups was statistically significant (p < 0.05), and all involved specific bimodal distribution (Figs. 3 and 4).

Fig. 3
figure 3

Gender-AO type distribution of four groups

Full size image
Fig. 4
figure 4

AO type distribution of four groups

Full size image

Discussion

The first tibial plateau fracture was reported by Sir Astley Cooper in 1825. In 1990, Donaldson et al. reported an average incidence of 26/100,000/year in a 1981 census of three years [5]. Court-Brown and Caesar revealed a higher incidence of 13.3/100,000/year in 2000 [1]. Similarly, the incidence in the North Denmark region was 10.3/100,000/year from 2005 to 2010 [6]. Court-Brown and Caesar also concluded that prevalence of fractures was changing rapidly resulting from diverse injury mechanism and aging population and revealed a 1.2% proportion of proximal tibial fractures in whole body fractures [1]. Incidence in other articles was approximately 1.3% compared to 1.5% in our study [7]. There was a slight discrepancy stated in these articles basically due to small sample size and limited single centre research.

A systematic retrospective review conducted by Tian ye et al. investigated 1,033 tibial plateau fractures in the Third Hospital of Hebei Medical University from 2003 to 2007. It demonstrated an occupation of 54.11% in all proximal tibia fractures, 10.09% of tibial and fibular fractures and 1.86% of whole body fractures [3].

Our study indicated a male predominance of tibial plateau fractures, and the same result was proven by Albuquerque et al., which referred a male predominance (70%) of tibial plateau fractures [8]. But controversy existed in Rasmus Elsoe’s cohort study, which found that a female predominance accounted for 53% [6]. Another important finding of our study was that tibial plateau fracture was primarily present in men at 40–44 years. Whereas in 1979, Schatzker et al. reported the most affected ages were 60–70; a possible explanation for this change, we believe, is the increased number of vehicles and traffic problems [9].

Number of tibial plateau fractures in the four groups in descending order were G1, G2, G3, G4. Negative correlation occurred between altitudes and the dominant fracture type, which can be explained by the restricted hospitals and few admitted patients in the plateau group; it also related to a lagging economy, less urban construction and poor communications compared to the plains group. In our four groups, age distributions were similar, while gender distribution and AO type were statistically significantly different. Frequency of severe fracture type increased along with evaluated altitude, whereby possible reasons might be altitudes have impact on bone density, daily diet, lifestyle, sunlight hours, communications, and finally, different injury mechanisms were generated, but no study had been firmly established.

Same peak ages were observed both in the overall cases and the four groups, i.e. males in the 40–44 year age group and females in the 55–59 year group. The most relevant factor we were concerned about was bone density. One research in Yunnan demonstrated that bone density of lumbar vertebrae of the general population in Kunming was higher than that in Chengdu [4]. The authors hold the view that higher elevation and longer sunlight hours may promote this occasion; when comparing it to people of north China, it was slightly lower. The researchers also did a preliminary analysis and suggested that high elevation created a depression and oxygen-deprived environments, so local population tend to have average lower body height and weight with such longer-term exposure, all of which can contribute to the lower bone density.

An very interesting inversion of sex ratio was noted among people over 60 years in our study, and we believed it had a relationship with the prevalence of osteoporosis, especially in elderly females [10, 11]. Evidence in Shanghai accessed density of tibia using quantitative ultrasound (QUS) bone measurement and concluded that SOS reached maximum at 30–39 years both in males and females. But in females, it fell rapidly at their fifth decade, while in males, SOS still sustained consistency over 50 years. This distinguished discrepancy can well explain the higher incidence of female osteoporosis fracture. Furthermore, menopause intensified this tendency [12]. Hung et al. depicted a bimodal curve of tibial cortical porosity (Ct.Po) during an adult life, abundant cortical bone was lost during people’s middle age, and cortical porosity sharply increased at the beginning of the fifth decade [13]. Indeed, elderly females became the most affected population of tibial plateau fractures .

Our study had certain limits in that the whole study was mainly based on X-ray films, and no comparisons were done with advanced imaging, i.e. CT or MRI. Elsoe reviewed the first epidemiology characters of tibial fractures according to CT [6]. Khan et al. concluded better diagnosis of plateau fractures using CT or MRI [14,15,16,17,18]. Another shortage was the lack of efficient information, i.e. injury mechanism, associated complications, intra-articular lesions of the knee, spinal cord injuries and other high-energy musculoskeletal system diseases, causing our study to only offer less useful evidence for clinics. Limitation also occurred in using the AO principle, Hohl’s [19, 20], Schatzker’s [21] and the recent “three column” concepts all have their own unique approach and may have better clinical benefits. Albuquerque et al. merged Hohl’s classification with the other two and identified a new classification presents greater interobserver accordance [8]. Yang et al. suggested a three-column classification can provide better understanding of fracture morphology and injury mechanism of tibial plateau fracture [22]. For further evaluation, a well-designed prospective cohort study is needed.