Schmorl’s nodes and vertebral fractures: A diagnostic dilemma

Abstract

A previously healthy 42-year-old man presented with sudden onset back pain whilst playing football. A T8 vertebral fragility fracture was diagnosed following X-ray and MRI thoracic spine, which demonstrated grade 1 compression. In addition, osteopenia was reported on dual-energy X-ray absorptiometry scan (DXA), which revealed a T-score of −1.8 in the lumbar spine. A follow-up MRI carried out after 18 months, reviewed by the metabolic bone clinic (MBC)-radiology multidisciplinary team, suggested features compatible with a ruptured Schmorl’s node. The DXA was re-reviewed by an MBC specialist and revealed bone density within expected range for age based on a Z-score above −2.0 using the ISCD 2023 recommended diagnostic criteria for men aged under 50. A conservative treatment approach was taken, sparing the patient from bisphosphonate therapy. This case highlights the importance of considering a ruptured Schmorl’s node in the differential of vertebral compression, especially in younger patients with no risk factors, and assessing bone density using Z-score definition rather than T- scores.

Introduction

Vertebral fragility fractures (VFF) are the commonest form of osteoporotic fractures, with prevalence studies suggesting 12% of women aged 50-79 years have vertebral fractures [1]. As they are strong predictors for future fractures, it is important to start bone protection treatment, once diagnosis is confirmed [2].

However, non-fracture vertebral deformities such as juvenile kyphosis (Scheuermann’s disease) and Schmorl’s nodes can mimic VFF radiologically, especially in younger patients with no risk factors [3]. Schmorl’s nodes are herniations of nucleus pulposus through the bony and cartilaginous end plate into the adjacent vertebral body. Schmorl’s nodes are common and usually picked up incidentally on imaging in asymptomatic patients. Rarely, Schmorl’s nodes can occur acutely, causing pain and inflammation [4,5]. They are also common in Scheuermann’s disease and form part of the diagnostic criteria [6].

Case report

We describe the case of a 42-year-old man who had a delayed presentation to medical services with an 8-month history of back pain which started immediately after playing football.

He denied any traumatic injury during the match. He was lactose intolerant and was on vitamin D supplements, for vitamin D deficiency. He was a non-smoker and denied any excess alcohol or illicit drug use and had no other risk factors for osteoporosis such as corticosteroid use or family history. He was an office worker and exercised regularly prior to the injury.

Examination revealed mild thoracic postural kyphosis which was non-tender. His body mass index was 23 kg/m².

He initially saw a physiotherapist, however due to a lack of improvement after 6 months, an MRI was requested, which was reported as showing a subacute T8 vertebral body wedge fracture, evident by mild depression in the superior endplate of the vertebrae (Fig. 1) but no posterior wall retropulsion.

Fig. 1.

Sagittal T1 (A), T2 (B), STIR (C) and axial T2 (D) images from December 2021 demonstrate mild subacute depression of the superior endplate of T8 vertebrae with mild bone marrow oedema and fatty changes at the fracture site, which were deemed to represent subacute wedge compression fracture of T8 body on the preliminary MRI report.

After an urgent neurosurgical review, where the patient was deemed to not meet the threshold for surgical intervention, he was referred for a DXA scan which was reported as osteopenia based on a T-score of −1.8 in both the lumbar spine (LSP) and femoral neck (FN).

An osteoporotic fracture was suspected, so an endocrinology assessment was recommended to determine the cause of osteopenia. Clinical assessment and blood tests to exclude secondary causes of osteoporosis including full blood count, creatinine, liver function, thyroid function, bone profile, vitamin D, ferritin, coeliac screen, testosterone and calculated free androgen index, were normal. Bone turnover markers, namely procollagen type 1 N-propeptide (P1NP) and C-terminal telopeptide of type 1 collagen (CTX) were also within the normal range.

On reviewing his imaging at the metabolic bone clinic-radiology multidisciplinary team meeting (MDT), the possibility of an acquired Schmorl’s node was discussed. No additional Schmorl’s nodes were identified at other levels. Given the entire superior endplate of the T8 did not show any bony end plate irregularity (only chronic signal changes) on all axial and STIR sequence imaging, with only central marrow involvement, this would be atypical of a compression fracture. Repeat MRI was recommended to further assess this possibility.

The repeat scan 10 months later again showed some chronic fatty changes involving the superior endplate. The previously seen central oedema showed near complete resolution (Fig. 2) but very minimal central irregularity on STIR sequence, which can also happen in Schmorl's nodes.

Fig. 2.

Sagittal T1 (A), T2 (B), STIR (C) and axial T2 (D) images from follow-up MRI in October, 2022 demonstrate chronic fatty endplate changes involving the superior endplate but there has been near complete resolution of the bone marrow oedema, when compared with the previous scan approximately 10 months ago.

A third MRI a year later showed healing in the bone marrow, with some patchy chronic fatty changes involving the endplate (Fig. 3). Serial imaging confirmed that the superior endplate surface was grossly intact. Given the lack of trauma or localized tenderness and minimal high signal change returns on STIR sequences throughout, the MDT favored the diagnosis of an interval progressive acquired Schmorl’s node, over a primary superior endplate depression fracture.

Fig. 3.

Sagittal T1 (A), T2 (B), STIR (C) and axial T2 (D) images from follow-up MRI in October, 2023 demonstrate reduction in signal intensity demonstrating with patchy changes chronic fatty endplate changes involving the superior endplate, when compared with the previous scan approximately 12 months ago. After this scan and multidisciplinary team discussion, it was raised the possibility of an interval progressive Schmorl’s node rather than primary superior endplate depression fracture as there has been no clear history of trauma or localized tenderness throughout but minimal high signal change returns on STIR sequence.

In addition, his DXA scan results were re-evaluated by a metabolic bone specialist, using the International Society for Clinical Densitometry (ISCD) 2023 diagnostic criteria for men under 50 years old. His Z-scores, which should have been the preferred unit of measurement rather than the T-scores, were above −2.0 and within the expected range for his age, which was also confirmed on repeat DXA after 1 year (−1.6 Lumbar Spine, −1.3 Femoral Neck, −0.9 Total Hip).

The patient continued on vitamin D supplements. Although he continued to report some moderate back pain, this was deemed positional and from paraspinal muscle fatigue, which was improving with physiotherapy exercises. Bone sparing treatment was not indicated, and he was discharged from follow up.

Discussion

This case highlights the challenges in distinguishing non-fracture vertebral deformities like Schmorl’s nodes from vertebral fractures, and common pitfalls when interpreting DXA imaging.

Vertebral wedging, as seen in this case, is most often attributed to osteoporotic vertebral fractures, particularly when there is no history of axial loading or trauma. The T8 height loss was first reported as a grade 1 vertebral fracture when assessing using the Genant semiquantitative (SQ) method [7]. However, in younger patients without underlying osteoporosis (as later clarified using Z-scores in our case) or major trauma, the presence of mild-moderate wedging should prompt careful reassessment before a diagnosis of compression fracture is made.

An alternative to SQ assessment is by using an algorithm-based qualitative (ABQ) approach. Wáng et al. describes several morphologic features that are characteristic of osteoporotic vertebral fractures which may help diagnose/rule out a fracture deformity when using an ABQ approach, such as central endplate depression, cortical step defects, posterior wall retropulsion, etc [8]. The Royal Osteoporosis Society’s guidance on identifying vertebral fractures also highlights the risk of over-diagnosing vertebral fractures in the presence of other pathologies through SQ methods, and the need for adjudication from experienced readers, for example specialized musculoskeletal radiologists in a multidisciplinary team [9]. Noticing the importance of analyzing vertebral morphology, Diacinti et al. proposed to first diagnose/rule out a vertebral fracture through detecting morphologic signs by expert readers, before applying SQ methods to grade the severity in order to improve the diagnostic accuracy [10].

Although Schmorl’s nodes are common, and often found incidentally on MRI, they can be associated with pain, with higher incidence noted in patients with back pain, especially in the context of degenerative disc disease, trauma, and osteoporosis [3]. Acute presentations in younger patients with no underlying causes are also described [4,5].

Symptomatic Schmorl’s nodes are associated with bone marrow oedema and localized inflammation, most notably seen in STIR/T2-weighted sequences [11]. Subtle vertebral body changes may also be seen, for example vertebral wedging, which can mimic traumatic osteoporotic vertebral fractures. Rarely, cases of painful Schmorl’s nodes associated with vertebral fractures have also been observed in literature, particularly in relation to pre-existing osteoporosis [12,13].

In our case, 2 key features suggest a non-fracture aetiology and process. Firstly, using a more qualitative approach, deformities may be excluded from the fracture classification if morphologic signs like collapse or cortical disruption are missing, particularly when vertebral height loss is considered mild (<20-25%) [10]. Non-fracture related height loss is also well-recognized in literature. Ferrar et al observed up to 37% of premenopausal women having short vertebral height, which was not associated with low bone mineral density [14]. Non-traumatic wedging of vertebral bodies has been reported in patients with endplate deformities including Schmorl’s nodes in literature, which can mimic vertebral fractures and lead to misdiagnoses [15,16].

Another reason vertebral fracture was suspected initially was the result of misinterpretation of the first DXA scan. Although DXA is widely used to diagnose osteoporosis, the results should be interpreted with caution. Technical issues related to poor positioning, incorrect identification of areas of interest, operator error and scan calibration differences can all affect precision and accuracy of scan [17]. Patient factors such as short stature and degenerative changes can also artificially affect measurements especially in patients with underlying secondary causes such as thalassemia, where elevated marrow iron content and altered vertebral shape may alter the results [18]. Using T-scores, which references BMD of a young, healthy female adult, can also result in over-estimation of osteoporosis risk in men [19]. Z-scores, however, compare to BMD of an age-, ethnicity- and sex-matched population, and are considered to be a more accurate assessment of bone mass in younger patients and was more appropriate for our patient. The ISCD (2023) recommends using Z-scores for premenopausal women and men under 50 years old [20]. Using this approach, our patient’s BMD fell within expected range for his age.

This had significant implications for treatment, in that bone sparing treatment such as alendronate was avoided. Bone-sparing treatment can post potential significant side effects, including hypocalcemia, osteonecrosis of the jaw, oesophagitis and atypical femoral fractures.

With regards to determining the origin of our patient’s singular Schmorl’s node, uncertainty remains. However, on review of literature and balance of probability, an acquired aetiology seems more likely. Wáng et al. observed acquired Schmorl’s nodes were more commonly found in the mid-thoracic spine, and in the superior rather than inferior endplate, and had less solidly defined borders on imaging [21]. This is in contrast with those secondary to developmental/congenital causes, which may be found in atypical sites and in greater numbers and could form part of other congenital disorders for example Scheuermann’s disease.

Many theories have been proposed to explain the pathogenesis of Schmorl’s nodes and they may be multifactorial in origin. Williams et al showed a hereditary factor of over 70% in twin studies, suggesting intrinsic vertebral end plate weakness may be the cause [22]. Dar et al proposed an axial load model in which micro-traumas can accumulate with time, leading to the formation of Schmorl’s nodes [23]. We have therefore highlighted some possible contributing factors in our case, primarily a history of vitamin D deficiency and repetitive microtrauma and axial loading from sports, that is football.

We acknowledge there are some limitations in our case which would have provided valuable information for comparison. A CT scan could have provided better spatial resolution including superior anatomical detail, particularly when evaluating endplate integrity, presence of sclerosis or other bony changes. However, in consulting our orthopedic specialists, treatment would likely have remained conservative given the extent of the pathology, therefore a CT scan was not performed in order to minimize unnecessary radiation exposure (patient aged <50 years) as per Ionizing Radiation (Medical Exposure) Regulations 2017 [24]. This is also supported by radiological evidence that the Schmorl’s node, most appreciable in Fig. 2, appears to have almost completely resolved in Fig. 3, which likely reflects healing of the ruptured node.

It would also be useful to have previous imaging to compare, even a simple lateral view of a chest radiograph. Unfortunately, lateral view chest radiographs are rarely performed in the United Kingdom, with CT/MR providing much greater sensitivity and anatomical details and being relatively accessible.

Overall, our case highlights the importance of interpreting imaging findings within clinical context and making individualized treatment decisions through multidisciplinary discussion. This is more important than ever given artificial intelligence is likely to play an increasing role in radiology. We also hope to shed light on the challenges of diagnosing osteoporosis and fragility fracture in young individuals, particularly when the data on treatment efficacy is limited and can have life-long implications. This case report will be extremely valuable for radiologists when interpreting plain films, MRI, CT scans, and DXA, as incorrectly labelling a fracture may lead to unnecessary treatment and potential adverse side effects.

Patient consent

Written informed consent was obtained from the patient for the purpose of publication in scientific journal.