In this study, we examined 83 patients diagnosed with MTX osteopathy, compared them to age-matched osteoporosis patients with atraumatic vertebral fractures, and tracked their progress over time under specific therapy. MTX osteopathy patients exhibit significantly more stress fractures in the lower extremities, particularly near the joints, aligning with previously described pattern [3, 5]. Notably, irrespective of the chosen osteological drug, pain relief is observed in MTX osteopathy patients only with the discontinuation of MTX therapy, underlining the pivotal role of this therapeutic measure. Additionally, patients with early remobilization did more frequently receive osteoanabolic therapy.

MTX patients, which were mainly female, are characterized by a higher body weight and higher BMI in comparison to OPO patients, in line with literature [24], which may be a sign of their underlying rheumatological disease and the associated restriction of movement limiting their physical activity. No differences were seen between the two groups with respect to sex or age. In line, MTX patients with GC treatment had higher mobility restrictions than others. There is a significantly higher frequency of MTX osteopathy manifestations in the lower extremity showing a characteristic fracture appearance (Fig. 1) compared to an osteoporosis group with atraumatic vertebral fractures. This highlights the relevance of the fracture site and appearance for suspicion of MTX-osteopathy, matching the published literature [3]. Specifically, over 70% of MTX patients have stress fractures of the upper or lower ankle joint and almost 2/3 have stress fractures of the bones forming the knee. This accumulation clearly indicates the primary manifestation of MTX osteopathy in the weight bearing bones of the lower extremity. Notably, to our knowledge the presented size of our cohort from one clinical center does outperform not only several cohorts describing MTX osteopathy but also the size of assembled review cohorts[3] composed of several individual publications, increasing the evidence of the lower extremity to be the area of primary manifestation.

Blood sampling indicates a shift of bone metabolism with a decreased osteoblast action mirrored by lower BGLAP but a higher bAP level. This is of particular interest, since MTX patients did present with stress fractures, for which elevated levels of bone formation by means of alkaline phosphatase [25,26,27] and BGLAP [27, 28] in response to the fractures respectively fracture healing is known and expected. Of note, since GC treatment may introduce a low turnover, stratification for GC treatment in the MTX did not indicate any differences, suggesting the low turnover to be independent of GC and mainly caused by MTX.

Bone resorption was above reference yet did not differ from OPO patients and was only single-digit compared to usually double-digit values in fractures [27]. This aspect might point to a low turnover bone metabolism (with respect to an expected high turnover in case of a present stress fracture) and thereby inability to initiate adequate fracture healing in case of MTX treatment. This may be caused by MTX, known to decrease metabolic processes by its very own drug mechanism [29]. The sufficient vitamin D level in both groups may be explained by regular medical care and the significantly higher levels in the MTX group by a more frequent medical health care consultation for those with chronic illnesses.

Pain levels in MTX patients are significantly higher than those in OPO patients, mainly explained by the presence of non-silent fractures in the weight-bearing lower extremity in MTX osteopathy as opposed to the often-silent vertebral fractures in OPO. While bone mineral density (BMD) is negative in all MTX osteopathy patients, the T-scores in the OPO group are notably lower. This indicates that the primary factor for fracture occurrence in MTX might not be aBMD itself but rather metabolic factors as described above. Nevertheless, it's worth noting that many patients meet the WHO definition of osteoporosis [30], where osteoporosis is known to be the most frequent bone disease in rheumatoid patients [31]. Since 55% of our MTX patients do not exhibit osteoporosis, it is evident that osteoporosis is not the primary factor driving MTX osteopathy. In fact, given that all MTX patients have an osteoporosis-favoring rheumatoid disease, MTX patients without osteoporosis had relatively favorable mean T-scores (− 1.6) considering the cohort's age and underlying rheumatoid disease. However, osteoporosis might be a risk factor for developing MTX-osteopathy since it was present in 45% of the respective patients. HR-pQCT imaging did only exhibit a lower Tb. Th in the radius of OPO patients compared to MTX, thus no major differences in bone microstructure were detected, but site and thus loading specific changes may be relevant.

While the largest portion of OPO patients received an antiresorptive agent, most patients with MTX osteopathy were treated with an anabolic agent. This can be attributed to the particularly low bone turnover observed in the MTX group in response to the stress fractures. Furthermore, MTX osteopathy is linked to fracture healing issues, with teriparatide being the only treatment supported by data for such cases [32, 33]. Additionally, at the time of inclusion, national guidelines for osteoporosis treatment favored antiresorptive therapy as the first-line option for osteoporosis. Hence, T-scores did significantly increase in the course of therapy for both, MTX and OPO. Notably, there was no significant difference of bone mass gain between OPO and MTX even considering the MTX group to exhibit a larger portion of anabolic treatment. While anabolic treatment options are known to decrease fracture occurrence significantly more than antiresorptive agents [34], and increase BMD after antiresorptive treatment[35], our finding is of particular interest and potentially points to the low responsiveness of the MTX affected bone. On the other hand anabolic treatments such as teriparatide are known to affect the spine more compared to antiresorptive treatment with regard to BMD gain [36]. But this pattern was not observed in the current study comparing the treatment effects of the OPO and MTX group (Fig. 3). Furthermore, 45% of MTX patients have both MTX osteopathy and osteoporosis. This highlights the need to not only stop MTX treatment as the primary approach but also to initiate bone-specific therapy, preferably anabolic treatment, based on our data. This applies not only to those with densitometric osteoporosis (T-score ≤ − 2.5) but also to most of the patients with MTX-associated stress fractures. In addition, the large proportion of OPO patients may indicate that osteoporosis is a risk factor for the development of MTX osteopathy.

After comparing the general treatment response in both groups, we have specifically focused on the treatment effects on MTX patients. Pain reduction was only perceived in case of MTX discontinuation (Fig. 4a), as described by others [4, 16, 37], yet fracture healing may remain limited [12]. After treatment, patients mobility was significantly improved as well as a significantly higher portion of patients was able to perform muscular performance and balance assessment, indicative of treatment success beyond pain relief. In particular, patients exhibit significantly larger improvements in their mobility restriction when receiving a therapy involving an anabolic drug.

The therapy-specific difference was analyzed in more detail accordingly. There were no frequency deviations between patients with large and small improvements in pain depending on the therapeutic agent. This means that the reduction in pain seems to be independent of the choice of the therapeutic agent but relies mainly on MTX discontinuation (Fig. 4a). However, when comparing the frequency of therapy choice of patients with large improvements in mobility to small improvements, significant frequency deviations were seen involving an anabolic drug exhibiting better improvements.

From our results, it can be conducted that patient’s mobility restriction levels upon first clinical presentation do not determine the improvement in pain (Fig. 5b) yet pain levels do. On the other hand, pain and mobility restriction levels do determine the patients pain relieve (Fig. 5c) and gain in mobility (Fig. 5d) even after adjustment for primary mobility levels. Stratification for GC treatment did not indicate differences in mobility gain by treatment between patients with or without GC treatment, even given the higher restrictions in the GC subgroup at start of the therapy.

This study has limitations. Due to its retrospective design, not all data points were available for follow up measurements, yet this is the largest cohort from one specific outpatient clinic thereby limiting variations in alertness of the physicians treating the patients as well as examinations chosen and variations in such and thereby guarantee a high level of reliability compared to datasets combined of case reports. Furthermore, we are not able to directly link MTX and the clinical pattern, yet today this specific pattern is exclusively observed in patients treated with MTX and is supported by experimental evidence [38,39,40]. Thirdly, since rheumatoid diseases are challenging with respect to treatment and require tailored therapy regimes, interactions may be present, yet reporting the by far largest cohort of MTX osteopathy, such interactions may be less prominent due to the number of patients. Fourth, subgroups of treatment are small, compared to the full cohort, limiting the statistical accuracy due to interacting parameters. To fully address the above-mentioned points, longitudinal studies are needed. Furthermore, reported pain levels may be influenced by timepoint of measurement, fracture location (OPO primarily in the spine, MTX primarily at the lower extremity), the underlying rheumatoid disease and possibly taken analgesics. Patients were not randomized with respect to MTX discontinuation in this retrospective study. But it is questionable whether such a randomization would be reasonable given the already existing body of literature. Importantly, the diagnosis of MTX osteopathy cannot rely solely on the fracture site. Radiological confirmation is essential, with a typical MTX osteopathy diagnosis requiring the identification of a meander-shaped, band-like sclerosis region indicative of a stress fracture on X-ray, CT, or preferably MRI. Thus, such diagnostics should be considered in respective patients.