Osteoporosis is Associated with Varus Deformity in Postmenopausal Women with Knee Osteoarthritis: A Case-Control Study

Background: Varus deformity of the knee is a common pathological characteristic in knee osteoarthritis (KOA), and not enough attention has been given to the relationship between knee varus deformity and the state of systemic bone mass. The purpose of this study is to evaluate the potential relationship between bone mineral density (BMD) and varus deformity in postmenopausal women with KOA. Methods: This case-control study included 202 postmenopausal women with KOA with or without varus deformity. The hip-knee-ankle angle of the lower extremity (HKA), medial distal femoral angle (MDFA), medial proximal tibial angle (MPTA), and the angle of the joint line (JLCA) were measured in all patients. The BMD of the lumbar (L1-L4), left femoral neck, and left hip were measured by dual-energy X-ray absorptiometry in all patients. The difference in BMD between the knee varus deformity group and the normal limb alignment group was compared, and the relationship between the different angles of limb alignment and the BMD values at different sites was evaluated. Results: The BMD at different sites of the knee varus deformity group was lower than of the normal limb alignment group, and the incidence of osteoporosis was higher. Pearson correlation analysis showed that BMD of the lumbar spine (L1-L4), left femoral neck and left hip joint were positively correlated with the HKA, but negatively correlated with JLCA. MPTA was positively correlated with the left femoral neck and left hip joint BMD, but not correlated with lumbar bone density. Furthermore, in the normal limb alignment group, the HKA was only negatively correlated with JLCA, but not signicantly correlated with MDFA and MPTA. In the varus deformity group, the HKA was not only negatively correlated with JLCA but also positively correlated with MDFA and MPTA. Conclusions:


Introduction
Knee osteoarthritis (KOA) and osteoporosis are two common diseases in orthopedics. Because the two diseases share common pathogenic factors, such as age, gender, heredity and in ammation, they often coexist in the same patient, especially middle-aged and elderly postmenopausal women [1,2]. It is worth mentioning that the two diseases are also the main diseases affecting the quality of life of middle-aged and elderly people. Experiencing long-term pain and limited activity lead to a decline in the labor force, which imposes an additional economic burden on society [3].
KOA is no longer considered simple cartilage degeneration but also includes subchondral bone changes, osteophyte formation, synovial in ammation, and changes in ligaments and muscles [4]. Among these effects, increasing attention has been paid to the role of subchondral bone changes in the pathological process of KOA. In previous studies, researchers only focused on the sclerosis of the subchondral plate and the formation of surrounding osteophytes in patients with advanced KOA, suggesting that knee osteoarthritis is associated with high bone density. However, in recent years, more studies have found that in the early stage of KOA, subchondral bone resorption increases and bone strength decreases, which in turn cause degeneration of the cartilage above [5,6]. Nonetheless, when osteoporosis and KOA coexist in the same patient, there is debate about whether systemic low bone mass exacerbates the progression of KOA.
Varus deformity of the knee joint is not only a key node in the pathological process of knee osteoarthritis but also an important factor for the rapid degeneration of the knee joint. Therefore, the prevention of varus deformity can greatly delay the speed of joint degeneration, which is the key to prevent and treat severe KOA. It is well known that varus deformity may result from congenital anatomical abnormalities, or it may be secondary to medial articular space narrowing after cartilage wear and subluxation of the tibiofemoral joint. However, this cannot fully account for all clinical varus deformity [7]. Although much attention has been given to the effect of periarticular bone mineral density (BMD) on limb alignment [8,9], the effect of systemic BMD, which represents the level of general bone mass, on limb alignment has not been adequately studied.
Full-length, load-bearing anteroposterior computed radiography images of the lower extremities were taken to measure the different angles of limb alignment, such as the hip-knee-ankle angle ( (HKA)), distal femoral medial angle (MDFA), proximal tibial medial angle (MPTA) and the angle of the joint line (JLCA).
The purpose of this study was to investigate the relationship between the different angles of the limb alignment and the general bone mass in postmenopausal women with KOA. Our hypothesis was that systemic low bone mass may cause or exacerbate varus deformity in KOA, thereby accelerating the progression of joint degeneration.

Materials And Methods
We included postmenopausal female patients hospitalized for knee pain in orthopedics at the First A liated Hospital of Guangzhou University of Chinese Medicine from January 2018 to June 2020. All of these patients met the American College of Rheumatology diagnostic criteria for KOA [10]. After admission, full-length weighted X-rays of bilateral lower extremities were taken for all patients, and BMD of the lumbar spine (LI-L4), left femoral neck and left hip were measured by dual-energy X-ray absorptiometry. The Kellgren-Lawrence (KL) grade was evaluated by a bilateral lower extremity X-ray.
Clinical data, such as age, duration of pain, menstrual history, trauma history, and recent medication use, were collected using an inpatient electronic medical record system. Height and weight were measured prior to the bone density test, and body mass index (BMI) was calculated using height and weight. To reduce the interference of confounding factors in this study, the following exclusion criteria were used: rst, patients with knee valgus deformity, as knee valgus is relatively rare clinically and the number of cases is relatively small; second, patients who had taken drugs affecting bone metabolism, such as bisphosphonates and hormones, in the last half year; third, a full-length X-ray of both lower limbs in nonstandard anteroposterior standing position to exclude the measurement error of limb alignment angles caused by rotation of lower extremities; fourth, patients with extra-articular malformations that may affect the distribution of the force lines of the knee, such as severe scoliosis, pelvic tilt, hip dysplasia, and angular malformation of the femoral shaft or tibial shaft, etc.; and fth; patients with pain so severe they cannot walk to reduce the effects of disuse osteoporosis. Finally, 202 postmenopausal women with knee osteoarthritis were enrolled, with an average age of 65.5 ± 7.4 years (49-83 years), an average BMI of 26.2 ± 3.8 kg/m2, a median pain duration of 60 months (0.5-360 months), 44 cases of grade KL2, 79 cases of grade KL3, and 79 cases of grade KL4.

Measurement of different angles of limb alignment
After admission, full-length weight-bearing X-ray lms of both lower limbs were taken, and four different angles of HKA, MDFA, MPTA and JLCA that could re ect the limb alignment were measured. If patients had been diagnosed with bilateral KOA, the more severe side was measured. The mechanical axis of the femur was de ned as the connection between the center of the femoral head and the center of the femoral condyle, while the mechanical axis of the tibia was de ned as the connection between the center of the tibial plateau and the center of the ankle joint. The degree of knee varus was de ned by the HKA, i.e., the medial intersection angle between the mechanical axis of the femur and the mechanical axis of the tibia. The smaller the HKA was, the more serious the varus deformity was. According to the standard range of the physiological HKA of Chinese women [11], we de ned 175.3°≤ HKA ≤ 180.3° as normal limb alignment and an HKA < 175.3° as varus deformity. The other three angles were de ned as follows: MDFA as the medial angle between the mechanical shaft of the femur and the tangential line of the femoral condyle, MPTA as the medial angle between the mechanical shaft of the tibia and the tangential line of the tibial plateau, and JLCA as the lateral angle between the tibial plateau tangent and the femoral condyle tangent (Fig. 1).
To reduce measurement errors, the same researcher (the rst author) took all measurements using the same imaging system successively at two-week intervals. The investigator was unaware of the patient's bone mineral density and other clinical data prior to measurement.

Measurement of bone density
After admission, BMD was measured at the left femoral neck, left hip and lumbar spine (L1-L4) by dual energy X-ray absorptiometry (GE Lunar Prodigy, USA). The BMD of these three regions can su ciently re ect the bone mass of the whole body and is the gold standard for the diagnosis of osteoporosis. According to WHO diagnostic criteria, T≤-2.5 is diagnosed as osteoporosis, -2.5 < T<-1 is diagnosed as osteopenia, and ≥-1 is normal bone density. Measurement parameters and machine quality are regularly checked according to the instructions provided by the manufacturer.

Statistical analysis
After the normal distribution test of the numerical variables, the Student 's t-test was used for the intergroup comparison of age, BMI and BMD values at different sites. Conversely, the Mann-Whitney U test was used for the intergroup comparison of the duration of pain. The chi-square test was used to compare the prevalence of osteoporosis. Binary logistic regression analysis was used to determine the risk factors affecting varus deformity. The Pearson correlation coe cient was used to evaluate the correlation between BMD at different positions and different angles of limb alignment. Similar to age, BMI and duration of pain may affect the HKA. Multiple linear regression was used to evaluate the in uence of BMD values at different sites on the HKA after adjusting for confounding factors, such as age, BMI and duration of pain. Finally, the Pearson correlation coe cient was used to evaluate the correlation between the HKA and other angles in the varus group and the normal limb alignment group.
All data analyses were performed with IBM SPSS 24.0 statistical software, and a P-value less than 0.05 was considered statistically signi cant in all tests.

Results
In the varus deformity group (HKA < 175.3°), there were 144 cases, including 69 cases of osteoporosis, 60 cases of osteopenia and 15 cases of normal bone mass. There were 58 cases in the normal limb alignment group (175.3°≤HKA ≤ 180.3°), including 15 cases of osteoporosis, 34 cases of osteopenia, and 9 cases of normal bone mass. There were no statistically signi cant differences in age, BMI and pain duration between the two groups. The BMD of the lumbar spine (L1-L4), left femoral neck and left hip of the varus deformity group was signi cantly lower than that of the normal limb alignment group, and the incidence of osteoporosis (47.9%) was signi cantly higher than that of the normal limb alignment group (25.9%, P < 0.05) ( Table 1).    (Table 3). That is, with increasing BMD, the degree of varus deformity would be reduced. Finally, in the normal limb alignment group, Pearson correlation analysis showed that the HKA was only negatively correlated with JLCA but had no signi cant correlation with MDFA and MPTA, whereas in the varus deformity group, HKA was not only negatively correlated with JLCA but also positively correlated with MDFA and MPTA (Fig. 3).

Discussion
Studies on the relationship between osteoporosis and osteoarthritis are ongoing. In early studies, high bone mineral density was generally believed to be related to the incidence and progression of osteoarthritis. Although recent studies began to realize that osteoporosis plays an important role in the pathological process of KOA, the relationship between osteoporosis and KOA is still controversial. Part of the reason may be that patients in different studies were at different stages of disease and had different indicators for measuring the progression of KOA [12]. KL grade pays more attention to the formation of osteophyte and the narrowing of joint space [13], while limb alignment can more directly re ect the incidence of varus deformity, which is key to the rapid progression of KOA.
Previous studies have primarily focused on the relationship between the subchondral BMD and limb alignment. Yoshinori et al. found that in patients with medial compartment KOA, the medial condyles had signi cantly higher BMD than the lateral condyles in both the femur and tibia, and the medial-to-lateral BMD ratio was negatively correlated with the HKA and the inclination angle of the tibial plateau [14]. Similarly, Grace H. Lo reported that in patients with KOA, the severity of knee varus deformity was positively correlated with the subchondral BMD of the medial tibia platform [15]. However, these studies can only re ect the results of the change in periarticular BMD caused by the uneven change in mechanical stress, and it is di cult to explain the in uence of BMD on the change in limb alignment. Moreover, local BMD could not represent the general bone mass, and it was di cult to reveal the relationship between the general bone mass state and limb alignment.
In this study, we measured the BMD at the lumbar spine, hip joint and femoral neck to represent the state of general bone mass and measured the different angles of limb alignment to evaluate the relationship between knee varus deformity and systemic BMD. We found that the BMD values of the above three sites in varus deformity patients were signi cantly lower than those in the normal limb alignment group, and the incidence of osteoporosis was higher than that in the normal limb alignment group. Furthermore, after adjusting for confounding factors, such as age, BMI, and pain duration, regression analysis showed that the BMD at the lumbar spine ( In recent years, substantial evidence has shown that subchondral bone and cartilage are a closely related integrated unit and mutually adaptable [17][18][19]. In the early stage of KOA, the absorption of subchondral bone is enhanced, and the mechanical strength of the bone is reduced. Improper stress stimulation also induces subchondral microfracture, which causes bone marrow edema and then degeneration of the cartilage covered on the surface [20,21]. Y. Chen's study on tibial plateau specimens of patients with knee osteoarthritis found that subchondral bone cysts degeneration was caused by increased local bone remolding, and the structural change in the cysts region accelerated the destruction of the overlying cartilage [22]. Zhu et al. even reported that the source of pain of KOA might be related to the abnormal activity of subchondral osteoclasts, which secreted Netrin-1 and induced neuronal axonal growth [23]. Therefore, we speculated that if patients with KOA also have osteoporosis, their subchondral mechanical properties will be further reduced, which will accelerate the progress of knee osteoarthritis and even varus deformity.
MDFA re ects the bone morphology of the medial femoral condyle, while MPTA re ects the inclination of the medial tibial plateau and JLCA is related to the narrowing degree of the medial articular space. Interestingly, in our study, BMD at the lumbar spine, femoral neck and hip joint is not only signi cantly correlated with the HKA, but signi cantly positively correlated with MTPA and negatively correlated with JLCA, indicating that with decreasing BMD, the more severe the inclination of the medial tibial plateau and the narrower the medial articular space would be. In addition, in the normal limb alignment group, the HKA was only negatively correlated with JLCA but not signi cantly correlated with MDFA and MPTA. In contrast, HKA was not only negatively correlated with JLCA but also positively correlated with MDFA and MPTA angles in the varus deformity group. The above results indicate that varus deformity of the knee is not only related to the narrowing of the medial articular space but may also be related to the change in the bone structure around the joint.
In fact, subchondral bone provides structural support for articular cartilage and plays a role in shock absorption, and its mechanical structure changes are closely related to cartilage degeneration [24].
Considering the increased fragility of the subchondral bone in osteoporosis patients, subchondral bone remolding will be further enhanced in the early stage of KOA [25]. At the same time, the stress of the medial platform is higher than that of the lateral platform under physiological conditions [26]. Therefore, we hypothesized that when KOA patients have osteoporosis, due to the weakening of the subchondral bone strength and improper stress stimulation, the medial platform may subside, thus aggravating the existing varus deformity caused by cartilage wear. In this study, MTPA was positively correlated with BMD values of the three different sites, which can con rm our inference to a certain extent. With the progression of knee varus, the stress in the medial compartment will further increase, which will aggravate the cartilage damage, thus forming a vicious circle. One study also reported that osteoporosis causes attrition of bone structure in advanced KOA, aggravating the deformity of joints [27]. Therefore, these results suggest that maintaining the systemic bone mass of patients may effectively delay the progression of KOA, especially in patients with osteoporosis.
There are several limitations in this study. First, this is a cross-sectional, retrospective study design in which we were unable to longitudinally observe the dynamic changes in the in uence of osteoporosis on varus deformity. Second, as a de ciency of estrogen is a common cause of osteoporosis and KOA, an osteoporosis clinical merger with KOA is often seen in postmenopausal women, so we only included postmenopausal Chinese women. To extend the conclusion to men and other races, research with a large sample of other populations should be conducted to con rm our ndings. Third, we failed to observe the changes of BMD in the subchondral bone, as the BMD around the knee joint is not the regular detection area of BMD in our hospital. Therefore, it is necessary to further observe the correlation between periarticular BMD and systemic BMD in patients with KOA in future studies. Despite these limitations, this study still revealed the relationship between the systemic bone mass status and different angles of limb alignment and proposed that osteoporosis may be an independent risk factor for varus deformity. Our study results will also provide a reference to further reveal the internal relationship between osteoporosis and knee osteoarthritis.

Conclusion
In summary, the incidence of knee varus deformity is not only related to the narrowing of the joint space but also to the change in the bone structure around the joint, which may be an important factor.
Maintaining systemic bone mass may help prevent varus deformity from progressing, but more longitudinal studies are needed to con rm this conclusion.