Osteoporosis and fractures contribute to damage in one of the most frequently involved organ systems in patients with systemic lupus erythematosus (SLE): the musculoskeletal system. In recent studies, a high frequency of low bone mineral density (BMD) and both peripheral and vertebral fractures has been demonstrated in patients with SLE. Moreover, a relative high frequency of undertreatment of SLE patients with manifest osteoporosis or at high risk of the development of osteoporosis has been reported. These findings highlight the need to develop better strategies for the prevention and treatment of osteoporosis and fractures as important disease complications in SLE.
SLE is a chronic autoimmune disease that usually affects women. Because survival of SLE patients has improved over the last decades, attention is now focused on complications that contribute to increased morbidity and mortality. The etiology of bone loss and the occurrence of fractures in SLE is supposed to be multifactorial, involving both non–disease-related and disease-related factors. Importantly, several risk factors associated with reduced BMD and fractures in SLE are modifiable by lifestyle measures or medication. This review discusses the etiology of osteoporosis and fractures in patients with SLE and proposes measures for the prevention and treatment of these complications.
Significance & Innovations
The etiology of bone loss in systemic lupus erythematosus (SLE) is supposed to be multifactorial, including traditional osteoporosis risk factors, inflammation, metabolic factors, and medication-induced adverse effects.
The occurrence of symptomatic fractures is increased 5-fold in women with SLE.
Prevalent vertebral fractures are present in 20–26.1% of the SLE patients, and 1 in 3 of these patients has normal bone mineral density at the lumbar spine and hip.
Several risk factors associated with osteoporosis and fractures in SLE are modifiable by lifestyle measures or medication.
Osteoporosis in SLE
Osteopenia is reported in 25–74% (1, 2) and osteoporosis is reported in 1.4–68% (3, 4) of SLE patients. These frequencies differ widely as a consequence of differences in size, age, sex, ethnicity, disease severity, and medication use between the patient groups investigated.
Etiology of osteoporosis in SLE.
The etiology of bone loss in SLE is supposed to be multifactorial, including traditional osteoporosis risk factors, inflammation, metabolic factors, hormonal factors, serologic factors, and medication-induced adverse effects (Figure 1).
Traditional osteoporosis risk factors.
Similar to findings in the general population, age (5–7), postmenopausal status (5, 8), and low body weight (7) or low body mass index (BMI) (5, 7, 8) are independent risk factors for osteoporosis in SLE patients. In addition, reduced lean body mass is associated with low BMD in SLE (9).
The influence of sex on BMD in SLE is unclear, since the majority of studies investigated (almost) exclusively female patients. Of the few studies performed in male patients, one demonstrated reduced BMD in men with SLE compared to healthy controls (10).
In the general population, white ethnicity is associated with an increased risk of osteoporosis, whereas African American ethnicity is associated with a higher peak bone mass (11). In 2 studies in SLE patients, white (12) or non–African Caribbean ethnicity (13) was a risk factor for osteoporosis, while another study showed reduced BMD in African American women as compared to whites (14). The reasons for these conflicting results are unclear and warrant further research on this subject. In general, the reported prevalence of osteoporosis in premenopausal Asian patients with SLE is lower (1.4–6.1%) than that reported in white patients (4, 15), but BMD was significantly lower in premenopausal Asian patients who had had glucocorticoid (GC) treatment than in those who had not (4). In postmenopausal Asian women, the frequency of osteoporosis (48%) is in the same range as whites (2, 4).
Smoking and excessive alcohol use have adverse effects on bone. Smoking is not reported as a risk factor for bone loss in SLE (5, 8, 10, 13, 16), but has been identified as a risk factor for osteoporotic fractures in a large study from the Hopkins Lupus Cohort (17), while the percentage of current smokers among SLE patients is not different from healthy controls (10, 18). Alcohol use was associated with bone loss in male lupus patients in Hong Kong (10).
Role of inflammation.
Systemic inflammation in SLE may contribute to bone loss by increasing osteoclastic bone resorption and by reducing osteoblastic bone formation. In patients with active disease, increased serum levels of tumor necrosis factor (TNF) (19) and elevated levels of oxidized low-density lipoprotein (LDL) (20) are demonstrated. Oxidized lipids can activate T cells, which may induce increased production of RANKL and TNF. Both RANKL and TNF increase the maturation and activity of osteoclasts (19). Moreover, oxidized LDL might also negatively influence bone formation by reducing osteoblast maturation (21). Decreased serum levels of osteocalcin, a marker for bone formation, and increased crosslinks excretion in the urine, a marker for bone resorption, were demonstrated in premenopausal women with untreated recently diagnosed SLE (22).
In the Hopkins Lupus Cohort, low serum complement C4 levels (a measure of active disease) were a predictor of low lumbar spine BMD (7). In patients with prolonged periods of disease activity, organ damage will accumulate. Several studies report a relationship between higher organ damage index and reduced BMD (5, 12, 23, 24), which suggests that disease activity adversely affects bone mass in SLE. However, clinical studies fail to demonstrate an association between disease activity score and BMD (4, 8, 12, 24), which might be explained by the cross-sectional design of these studies.
Inflammation-induced lupus nephritis occurs in 50–60% of SLE patients ever during the disease course and can result in renal failure (25). In severe renal failure, bone loss is increased by the development of secondary hyperparathyroidism and increased osteoclastic bone resorption, and by the development of low 1,25-dihydroxyvitamin D (1,25[OH]2D) serum levels, which might impair intestinal calcium absorption. However, an association between impaired renal function and low BMD was reported in only 1 study, in older female lupus patients (5). Importantly, the majority of studies performed employed decreased renal function as an exclusion criterion. Therefore, the impact of renal disease on bone mass in SLE is still partly unclear.
Hyperhomocysteinemia and vitamin D deficiency are metabolic conditions associated with SLE that can induce bone loss.
Hyperhomocysteinemia adversely affects bone quality by stimulating bone resorption and by disturbing collagen crosslinking (26). In the general population, high homocysteine levels have been associated with bone loss (27). In SLE patients, an increased prevalence of hyperhomocysteinemia was demonstrated, but homocysteine levels were not associated with BMD change (18).
Studies in different geographic regions show an increased prevalence of vitamin D deficiency among SLE patients (28–30). Moreover, low 25-hydroxyvitamin D (25[OH]D) levels were associated with low BMD in the lumbar spine in Dutch SLE patients (8).
Several factors might negatively influence vitamin D status in SLE: photosensitivity, dark skin pigment, use of sunscreen, renal failure, GC use, and probably hydroxychloroquine (HCQ) use, disease activity, and anti–vitamin D antibodies.
Ultraviolet light intolerance is highly frequent in SLE patients and leads to avoidance of sun exposure and subsequent reduced de novo vitamin D synthesis in the skin.
Patients with darker skin pigment might be at an increased risk for the development of vitamin D deficiency, since melanin blocks vitamin D synthesis in the skin. Kamen and colleagues demonstrated that African American patients indeed had significantly lower vitamin D levels than whites (31).
Patients with SLE are strongly advised to use sunscreen continuously, which can lead to decreased and sometimes completely stopped de novo vitamin D synthesis in the skin. Studies have shown that continuous use of sunscreen can reduce serum vitamin D levels (32).
Renal failure induces low 1,25(OH)2D levels by loss of the renal enzyme hydroxylase α1. Recently, an association between high serum creatinine and low 1,25(OH)2D levels in SLE patients was reported (30).
GC treatment might reduce vitamin D levels by suppressing the intestinal calcium absorption. Toloza and colleagues found an association between cumulative GC exposure and low serum levels of both 25(OH)D and 1,25(OH)2D (30).
The antimalarial HCQ, which is frequently used for the treatment of SLE, is supposed to inhibit the conversion of 25(OH)D to 1,25(OH)2D by inhibiting hydroxylation α1. A cross-sectional study in SLE patients demonstrated lower 1,25(OH)2D levels in HCQ users as compared to nonusers, while 25(OH)D levels were not different between both groups (29). However, another study showed higher 25(OH)D levels in HCQ users (33).
In contrast to a cross-sectional study demonstrating a negative association between vitamin D levels and disease activity (28), a longitudinal study did not confirm this relationship (34).
The role of anti–vitamin D antibodies in SLE is unclear. These antibodies were detected in 4% of SLE patients, but were not related to vitamin D serum levels (35).
SLE is characterized by a relatively high estrogenic and low androgenic state. In patients with SLE, increased hydroxylation α16 of estradiol, increased testosterone oxidation, and a decrease in dehydroepiandrosterone (DHEA) have been described (36). In addition, an association between low dehydroepiandrosterone sulfate (DHEAS) levels and low BMD in patients with SLE has been reported (37).
Mok and colleagues studied the relationship between serologic profile and BMD in postmenopausal SLE patients in Hong Kong and revealed that the presence of anti-Sm and absence of anti-Ro were associated with a higher BMD at the lumbar spine (2). The relationship between the absence of anti-Ro and higher bone mass might be explained by the fact that patients with positive anti-Ro are generally advised against sun exposure (2).
Role of medication.
GCs are extensively used for the treatment of SLE disease flares and complications. GCs predominantly affect trabecular bone and the cortical rim of the vertebral body. However, in prolonged use, GCs can also affect the cortex of long bones. On the other hand, GCs might also have beneficial effects by reducing the adverse effects of systemic inflammation on bone. Cross-sectional studies on the relationship between GC use and BMD in SLE show conflicting results. In contrast to studies reporting an association between GC use and low BMD in the lumbar spine or at the hip (12, 16), other studies did not demonstrate any relationship between GC use and BMD (5, 8). Only a few longitudinal studies on BMD change and its determinants, in small groups of SLE patients and with a relative short followup duration, were performed (3, 38–41) (Table 1). The studies by Kipen et al and Jardinet et al demonstrate that lumbar spine bone loss occurs exclusively in patients treated with a mean prednisone dosage of >7.5 mg/day (39, 40). Boyanov et al showed increased bone loss in both the lumbar spine and hip in patients treated with GC as compared to patients who had never taken GC (3).
Table 1. Longitudinal studies on BMD change in patients with systemic lupus erythematosus*
Author, year (ref.)
No. of patients (no. of GC users)
Mean followup duration, months
Daily prednisone dosage in GC users, mg/day
Change in lumbar spine BMD
Change in hip BMD
Influence of GCs on BMD change
Values are the mean ± SD unless otherwise indicated. BMD = bone mineral density; GC = glucocorticoid; NS = not significant.
Prednisone users → increased ↓ BMD lumbar spine and hip
HCQ use was associated with higher BMD in the lumbar spine (2, 12) and at the hip (12) in 2 cross-sectional studies in female SLE patients. In contrast, male SLE patients who had ever used HCQ had significantly lower BMD both in the lumbar spine and at the hip as compared to patients who had never taken HCQ (10). Clearly, further studies of the relationship between HCQ use and BMD in large groups of SLE patients as well as in patients with other diseases treated with HCQ are needed.
Fractures in SLE
Epidemiology and etiology of symptomatic fractures.
In a population-based cohort of 702 women with SLE, fracture occurrence was nearly 5-fold compared with healthy women of a similar age (42). The combined risk for hip and vertebral fractures was increased with an odds ratio (OR) of 2.9 in a Swedish study (43). Symptomatic fractures since lupus diagnosis occur in 6–12.5% of patients (7, 13, 42, 44), and the hip/femur, vertebra, rib, foot, ankle, and arm are the most frequent sites of fractures (7, 42, 44). Age (7, 13, 42), postmenopausal status (7), smoking (17), disease duration (44), renal insufficiency (17), Raynaud's syndrome (17), lupus anticoagulant (17), and reduced BMD (13) are reported as risk factors for symptomatic fractures in SLE patients. In addition, GC treatment (both cumulative and highest dose [7, 45]) is a predictor of osteoporotic fractures, and longer use of GCs is associated with time from lupus diagnosis to fracture (42).
Epidemiology and etiology of (prevalent) vertebral fractures.
In a case–control study, the risk for (symptomatic) vertebral fractures in SLE patients was increased with an OR of 2.2 (43). However, studies focusing on symptomatic fractures have a disadvantage in that only one-third of all vertebral fractures come to clinical attention (46). Identifying prevalent vertebral fractures is important, since prevalent vertebral deformities are associated with reduced quality of life (47), an increased mortality rate, and an increased risk of future fractures in the general population (48).
Only a few studies on prevalent vertebral fractures (using a standardized method of scoring vertebral deformities) were performed, which demonstrated the presence of at least 1 vertebral fracture in 20–26.1% of the SLE patients (8, 24, 49, 50) (Table 2). Low BMD (24, 49, 50), age (24, 50), previous use of intravenous methylprednisolone (8), male sex (8), and, surprisingly, higher BMI (50) were associated with vertebral fractures.
Table 2. Cross-sectional studies on prevalent vertebral fractures in patients with systemic lupus erythematosus*
Author, year (ref.)
No. of patients
% with ≥1 vertebral fracture
Values are the mean ± SD unless otherwise indicated. MP = methylprednisolone; IV = intravenously; BMD = bone mineral density.
Importantly, among the patients with vertebral fractures, 29–35.8% had normal BMD (24, 50). This finding is in line with results from studies in the general population reporting that the proportion of fractures attributable to osteoporosis is only 10–44% (51).These data point to the limited value of BMD measurement in the assessment of future fracture risk.
Fall risk might be increased in SLE patients due to fatigue, muscle weakness (due to GC-induced myopathy and/or vitamin D deficiency), arthritis, neuropathy, epilepsy, or visual impairment, but studies on fall risk in lupus patients have not been published yet.
Prevention and treatment of osteoporosis and fractures
First, lifestyle measures are important, which include avoiding smoking, limiting alcohol intake, maintaining a normal body weight, avoiding falling, and performing regular weight-bearing exercise.
Second, attention must be paid to an adequate calcium intake and sufficient serum 25(OH)D levels, which are important for bone mineralization and neuromuscular function. The total calcium intake (dietary intake plus supplement) should be at least 1,000 mg daily and in GC users, 1,200–1,500 mg/day is recommended (52). The cutoff value of adequate serum 25(OH)D levels is still under debate. A level of at least 50 nmoles/liter has been proposed, but probably a level of at least 75 nmoles/liter is necessary in the context of optimal muscle function (53). Vitamin D supplementation (800–1,000 IU/day) is recommended in patients with insufficient 25(OH)D levels and patients receiving GCs (52).
Third, the prescription of immunosuppressive medication, to reduce inflammation-induced bone loss, and minimizing dose and duration of GC treatment are essential measures.
The etiology of osteoporotic fractures is multifactorial, including bone- and fall-related factors. The identification of patients with increased fracture risk solely using BMD measurements has several disadvantages, such as its inaccuracy in measuring bone quality and its age dependency (54). Several algorithms for calculating the absolute fracture risk for the individual patient have been developed in the last decade. A disadvantage of using the absolute fracture risk is the lack of consensus regarding at which cutoff point treatment should be started. A frequently used model is the Fracture Risk Assessment (FRAX) tool as proposed by the World Health Organization (55). The FRAX tool takes into account BMD and family history, but does not include the evaluation of the risk factors of falls and the presence or absence of prevalent vertebral deformities. The latter is a great disadvantage in lupus patients, since prevalent vertebral fractures are frequent in patients with SLE and are often present in patients with a normal or only slightly reduced BMD. In addition, the FRAX tool takes GC usage into account, but not GC dose. Furthermore, using this algorithm is not appropriate in the large subgroup of premenopausal women with SLE, since the FRAX tool has been developed for postmenopausal women.
Drug treatment of osteoporosis.
For SLE patients with osteoporosis, with a previous fragility fracture, and/or receiving GCs, it is important to consider the prescription of an antiresorptive agent. The bisphosphonates alendronate, risedronate, and zoledronic acid are recommended for the prevention and treatment of osteoporosis in GC-treated individuals without renal impairment (52). Bisphosphonates should not be prescribed to premenopausal women planning a pregnancy, since these agents are associated with fetal abnormalities in animal studies (56). In addition, no data are presently available on the effectiveness of bisphosphonate treatment in premenopausal women. Therefore, bisphosphonates should only be prescribed in premenopausal women with severe osteoporosis or at a high risk for severe bone loss (e.g., those who are likely to be treated with high GC doses for a prolonged period of time). These women should have completed their families or they should have decided not to become pregnant for several years. Furthermore, physicians should remind patients that osteonecrosis of the jaw (57, 58) and atypical femoral shaft fractures (59) are potential complications of bisphosphonate therapy, especially in patients with concomitant GC treatment.
The anabolic agent teriparatide can be considered for the treatment of postmenopausal women and men ages ≥50 years with osteoporosis and high fracture risk (52).
The use of estrogen-containing agents to prevent osteoporosis in SLE is not recommended, since these drugs are associated with an increased thrombotic risk in the general population and an increased flare rate in SLE patients (60). The estrogen receptor modulator raloxifene may be a useful alternative in postmenopausal women with inactive lupus and without antiphospholipid antibodies or a history of thromboembolism (61).
The low DHEAS serum levels, which are correlated with low BMD in patients with SLE, offer the opportunity for the therapeutic use of DHEA to prevent bone loss. Studies have demonstrated that DHEA treatment offers mild protection against bone loss in postmenopausal and GC-treated patients with active disease (6, 62). This therapeutic option is relevant for patients in whom treatment with bisphosphonates or raloxifene is contraindicated or is not tolerated.
Recent studies on the molecular pathways underlying bone metabolism have identified potential novel therapeutic targets for the management of osteoporosis.
It may be expected that new drugs (denosumab, cathepsin K inhibitors, and monoclonal antibodies against sclerostin) interfering with the Wnt signaling pathway or the RANKL/osteoprotegerin pathways may prove more effective than currently used drugs (e.g., bisphosphonates) in the treatment of osteoporosis, especially in GC-induced osteoporosis (54). Denosumab, a monoclonal antibody against RANKL, has recently become available for the treatment of osteoporosis. This drug is an attractive new therapeutic agent for the subgroup of SLE patients with impaired renal function, since denosumab may be prescribed in patients with renal failure.
Recently, quality indicators for osteoporosis prevention and treatment in SLE were published, recommending screening for osteoporosis by BMD testing in GC-treated patients (63). However, in line with the American College of Rheumatology recommendations for prevention and treatment of GC-induced osteoporosis, additional vertebral fracture assessment must be considered since vertebral fractures are frequent in SLE and might change treatment recommendations for GC-treated patients (52). Moreover, osteoporosis screening should also be considered in subgroups of SLE patients without GC treatment, i.e., postmenopausal women and patients with other osteoporosis risk factors.
Dr. Bultink drafted the article, revised it critically for important intellectual content, and approved the final version to be published.