Identification of vitamin D and other bone metabolism parameters as risk factors for primary bone marrow oedema syndrome

Bone marrow oedema (also known as bone marrow lesion) is characterised by the excessive accumulation of fluid within the bone that causes nonspecific pain and high-signal-intensity alterations on magnetic resonance fluid-sensitive sequences [1]. A number of these lesions are secondary to various pathologies, including trauma, osteoarthritis, inflammatory arthritis, benign and malignant neoplasms, infections and metabolic disorders [2, 3]. However, primary bone marrow oedema lesions that lack an obvious cause also exist. These lesions are most commonly found in the bones around the hip, knee, ankle and foot and have been variously termed regional migratory osteoporosis, transient osteoporosis, and, most recently, bone marrow oedema syndrome (BMES) [4, 5]. The multiple names, and the fact that it is a diagnosis by exclusion, reflect the prevailing uncertainty about the aetiology of BMES [4]. However, several hypotheses about BMES pathogenesis have been proposed. One suggests that it is caused by nerve compression, another that it is the result of increased intramedullary pressure due to altered venous outflow [6]. However, more recently, clinical and histological studies on bone metabolism in BMES suggested that deterioration of bone mineralisation due to vitamin D deficiency may contribute to BMES [7‐10]. However, these studies involved small patient cohorts and/or only assessed a few diagnostic aspects such as bone mineral density (BMD) or vitamin D deficiency. Thus, the aetiology and pathogenesis of BMES remain poorly understood.

To improve our understanding of why and how BMES develops, we conducted a retrospective study with a large cohort of patients with BMES in which we assessed the relationship between a wide variety of bone metabolism, bone quality, and potential risk factors and BMES.

The current study showed that, in patients with newly diagnosed BMES, age at onset exhibited two peaks, one in adolescents (11–20 years of age) and one in older/elderly patients (51–70 years of age). The majority (61.5%) were female, a fifth had thyroid disorders (21.5%), another fifth (21.4%) had secondary hyperparathyroidism and another fifth (19.6%) had elevated DPD levels, which indicate increased bone resorption. In addition, the vast majority (91.7%) of the patients who had not recently undergone vitamin D supplementation exhibited vitamin D insufficiency or deficiency. A peak of disease onset was seen during the winter time were vitamin D levels are generally lower than in summer months. Furthermore, the patients exhibited alterations in trabecular and cortical bone structure.

Numerous studies have assessed the disease characteristics of BMES. They showed that male patients predominate and the vast majority of lesions are found on the lower extremity (98%), with the proximal femur being the most common site, followed by the knee and the ankle/foot [3, 4, 6, 15‐19]. By contrast, our patient cohort showed a female predominance. However, we did find, like the other studies, that the BMES lesions in our cohort showed a strong predilection for the lower weight-bearing extremities (99%). This suggests that external load is a predisposing factor for BMES [4]. However, unlike previous observations, the lesions in our patients were markedly concentrated in the foot and ankle region, followed by the knee and then the proximal femur. This may reflect the fact that the other studies focused on one selected joint region, and included subchondral fractures (43% in their study group) at the hip [15] or including subchondral fractures (11.2%) and articular collapse (24.5%) at the knee [16]. This limits our ability to compare our data with those of these studies. Our findings are being supported by the study of Cahir et al.: they assessed multiple regions in sequential scans for regional migratory osteoporosis and found that the ankle and foot and the knee were most affected when the lesions recurred [17].

Previous studies showed that the mean age of disease onset in BMES is 30–60 years [19]. While we also detected a large peak of BMES cases in 51–70-year-olds, we also observed a small peak of BMES cases in adolescents. This interesting observation most likely reflects the fact that pubertal growth and aging both associate with relatively insufficient bone mineralisation that can increase bone fragility [20, 21].

The fact that our BMES cases are prevalent in a state of insufficient bone metabolism and high bone turnover is consistent with observations of this study. First, we found that a relatively large proportion of our patients (12%) were on long-term systemic corticosteroid therapy. This increases bone fragility by promoting bone resorption [22]. Second, a surprisingly large number of our patients (21.5%) had been diagnosed previously with thyroid disorders (almost exclusively hypothyroidism) and altogether 16% were being treated with thyroid hormone at the time BMES was diagnosed. Several other studies reported that the frequencies of thyroid disorders in their BMES patient cohorts ranged from 4 to 19% and that 10% were taking thyroid hormone at BMES diagnosis [23‐25]. While hyperthyroidism is well known to induce catabolic bone metabolism, there is also some suspicion that hypothyroidism can promote bone fragility by reducing bone turnover and prolonging the bone remodelling cycle [26]. Moreover, thyroid hormone therapy associates with an increased risk of fracture [26], although the underlying mechanism remains to be elucidated. Third, there was a high frequency of secondary hyperparathyroidism in our cohort (21.4%). In normal populations, 2.7–6.6% have secondary hyperparathyroidism [27, 28]. The high frequency of secondary hyperparathyroidism in our cohort is consistent with the low average calcium serum levels, relatively high PTH levels and average DPD levels that exceeded the upper reference value. Fourth, we found that 5% of our patients had manifest PPI-induced hypochlorhydria, which associates with bone fragility [29]. Fifth, we found reduced trabecular thickness and cortical volumetric BMD. Similar alterations in bone structure were found in patients with primary hyperparathyroidism [30] and may display osteomalacia. All of these findings suggest that patients with BMES have high bone turnover. If this is causal for the development of BMES or a consequence of remains unclear. Berger et al. found that, although patients with BMES had normal serum bone turnover variables (BAP, osteocalcin and C-terminal crosslinking telopeptide), aspirates of the cancellous bone of the BMES lesions had high levels of these markers [9]. Overall, these previous observations are consistent with the fact that anti-resorptive therapy (i.e., bisphosphonates, denosumab) was found to be effective in the treatment of BMES [31‐33].

The possibility that BMES associates with impaired calcium metabolism is further supported by our finding that vitamin D insufficiency and deficiency were very common in our cohort (frequencies of 92 and 58% in the patients without and with recent vitamin D supplementation, respectively). Our findings are supported by Horas et al. (n = 31) and Sprinchorn et al. (n = 10), who also found that patients with BMES frequently had vitamin D insufficiency (defined as < 30 μg/l, 84 and 90%, respectively) and a low average vitamin D level (19 μg/l) [7, 8]. While recent epidemiological data show that these high frequencies of vitamin D insufficiency are not dissimilar to those in the general German population (88% have 20–30 μg/l, 62% have < 20 μg/l and the mean level is 18 μg/l) [34], such extensive vitamin D insufficiency is likely to contribute to the derangement of bone metabolism/bone mineralisation that is seen in BMES.

Notably, given the high prevalence of vitamin D deficiency in our cohort, it is likely that this is largely responsible for the high prevalence of secondary hyperparathyroidism. Other factors such as PPI-induced hypochlorhydria may also contribute. It should be noted that, although 36% of our patients were taking vitamin D supplements at the time BMES was diagnosed (and PTH was measured), this supplementation only started very recently in most of these patients. Consequently, the PTH levels of the cohort patients had not yet had the time to respond to the supplementation in most cases.

BMES lesions were originally called transient osteoporosis. However, this name was dropped because DXA studies showed that patients with bone marrow lesions do not have always decreased BMD [8, 15, 16]. This is supported by the present study. Although we found that 47% of our patients had osteopenia and 18% had osteoporosis, several studies reported that up to 15% of the general population have osteoporosis [35, 36]. Thus, our cohort did exhibit marginal reduced BMD compared to the general population. Although metabolism derangement does appear to associate strongly with BMES, the pathogenesis of this disease does not seem to require substantial alterations in bone mineral density as reflected by DXA.

To sum up, this study showed that BMES lesions, which almost exclusively affect weight-bearing and thus load-affected joints, may be caused by impaired bone metabolism during adolescence and older age that is the result of high systemic bone turnover, as possible consequence of reduced vitamin D levels. We thus hypothesise that BMES pathogenesis is a multifactorial process in which local or systemic phases of accelerated bone turnover cause bone microdamage, even under physiological load, and that this bone microdamage may predispose the individual to the development of BMES. This hypothesis is entirely consistent with the increasingly accepted pathophysiology-based concept, namely, that diseases such as BMES may be “regional acceleratory phenomena” where a noxious stimulus such as bone tissue microdamage in vitamin D deficient bone causes normal biological processes, including blood flow, cell metabolism and tissue remodelling processes, to accelerate. This in turn leads to a transient shift to unmineralised bone [37]. This possibility is supported by histological and biochemical findings of BME lesion specimens [5, 9, 10]. Thus, BMES may be the result of decoupling between the self-repairing capacity of the bone tissue and the accumulation of bone microdamage due to impaired bone metabolism [4].

This study has a number of limitations. First, its cross-sectional design makes it difficult to identify a causal association between BMES and impaired calcium metabolism and bone mineralisation. Second, the data are retrospective, which introduces the possibility of information bias. Third, there may have been some selection bias, which may limit the generalisability of our findings to other BMES populations. However, these limitations are unlikely to affect the main conclusions of the study.

We conclude that patients diagnosed with primary bone marrow lesions should be screened carefully for bone metabolism disorders, especially vitamin D levels, and their potential risk factors. Moreover, our findings suggest that a comprehensible treatment for BMES is a therapy that compensates for elevated bone resorption.