Краткосрочные факторы риска потери костей при множественном склерозе: перспективное исследование и анализ литературы.

Abbreviations

bone mineral density

body‐mass‐index

disease‐modifying treatment

dual‐energy X‐ray absorptiometry

Expanded Disability Status Scale

femoral neck

follow‐up

glucocorticoid

healthy controls

highly effective treatment

interferon

lumbar spine

male

moderately effective treatment

month

methylprednisolone

multiple sclerosis

no

not available

postmenopausal

premenopausal

people with multiple sclerosis

per years¸ Vit, vitamin

versus

yes

years

Introduction

Reduced bone mass and an increased risk of osteoporosis compared to healthy controls (HC) are a known issue in chronic diseases and have also been described in people with multiple sclerosis (pwMS) [ 1 ]. Although bone loss can already occur in the early stage of multiple sclerosis (MS) without significant disability, several risk factors for reduced bone mineral density (BMD) or osteoporosis have been previously identified [ 2 , 3 , 4 , 5 , 6 ]. In particular, the reduction in weight‐bearing activity, which is reflected in the degree of disability, plays a role in the development of osteoporosis in pwMS [ 3 , 4 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 ]. The effect of the short‐term, high‐dose glucocorticoid (GC) therapy, used for relapse treatment in MS, is less clear in the development of osteoporosis; data regarding both the cumulative lifetime dose and short‐term effects are contradictory [ 1 , 4 , 10 , 16 , 18 , 19 , 20 ].

In a previous study, we have examined osteoporosis risk factors in pwMS under 65. We identified age, sex, menopausal status, body‐mass index (BMI), smoking, and degree of disability as key risk factors [ 16 ].

Understanding the dynamics of bone health in pwMS is essential for developing targeted interventions to prevent osteoporosis and associated increased morbidity and mortality and to improve overall quality of life and health. As prospective longitudinal studies on progressive bone loss have included only a limited number of patients or have had only very short observation periods, we performed the present study with a larger cohort and prospective follow‐up after 2 years [ 8 , 21 ], examining how disease characteristics and GC treatment as well as medications with possible positive effects on bone health influence bone density changes. In addition, we present a literature review of longitudinal data on BMD in MS, highlighting risk factors and implications for clinical practice.

Methods

Study Design and Participants

Initially, premenopausal and postmenopausal women, and men with MS diagnosed according to the 2017 McDonald criteria [ 22 ] aged between 18 and 65 years were included at the Department of Neurology of the Medical University Innsbruck in a prospective observational study initiated in 2020 [ 16 ] Exclusion criteria were an ongoing pregnancy as a contraindication for densitometry (at inclusion or during the scheduled follow‐up visit), conditions that confound dual‐energy X‐ray absorptiometry (DXA) such as a lumbar spinal fusion surgery, lack of documented history of cumulative GC dose, and a history of osteoporosis at screening.

Demographic data, weight, height, BMI, abdominal circumference, smoking status, alcohol consumption, date of menopause, family status, occupational status, education, parental hip fracture, and familial osteoporosis were documented. The disease course, comorbidities, concomitant medication, disease duration, and past and/or current disease‐modifying treatment (DMT) were assessed at baseline and during follow‐up. DMT was categorized into moderately effective treatment (mDMT) including glatiramer acetate, interferon‐beta, teriflunomide, dimethyl fumarate; and highly effective treatment (hDMT) including alemtuzumab, cladribine, sphingosine‐1‐phosphate receptor modulator, natalizumab, anti‐CD20 antibodies; or no treatment. The lifetime methylprednisolone dose was recorded as previously described [ 16 ]. GC therapy and the number of relapses during follow‐up were prospectively documented by the treating physician.

Disability was assessed by the Expanded Disability Status Scale (EDSS) at baseline and at least every 6 months during the 2‐year follow‐up [ 23 ]. Disability worsening was defined as an increase in EDSS score for ≥ 1.0 point (EDSS ≤ 5.5) or ≥ 0.5 point (EDSS ≥ 6.0) and confirmed after 24 weeks [ 24 ].

Postmenopausal status was defined as a 12‐month secondary amenorrhea confirmed by hormonal status. Physical activity was recorded using the Baecke physical activity questionnaire [ 25 ].

Densitometry

BMD (g/cm 2 ) was assessed of lumbar spine (total LWS) and right hip (total hip) by DXA (Hologic Discovery A) at the Department of Nuclear Medicine, Medical University Innsbruck according to standard protocol at baseline and at year 2. Analysis of results was performed by a blinded rater of the Department of Nuclear Medicine. For the assessment of osteoporosis, the T score was calculated, and osteoporosis was defined according to the WHO criteria [ 26 , 27 ] in study participants with T score ≤ −2.5 in at least one skeletal site; osteopenia was defined as a T‐score of −1.0 to −2.4. Bone loss was defined as a decrease > 0.34% [ 28 ].

Laboratory Measures

Levels of 25‐hydroxy‐vitamin D were analyzed by Central Institute of Medical and Chemical Laboratory Diagnostics (ZIMCL), University Hospital of Innsbruck, using high performance liquid chromatography UV (Chromsystems).

Statistical Analysis

Distribution of data was assessed graphically and by the Kolmogorov–Smirnov test, and data were displayed as median and interquartile range or range, as appropriate. For independent and paired group comparisons, the Wilcoxon test was applied, or, as appropriate, McNemar's test.

To identify predictors of bone loss over 2 years, logistic regression was employed with BMD loss as the dependent binary variable (BMD loss > 0.34%). Current smoking status (binary: smoking vs. non‐smoking), BMI (continuous), baseline 25‐hydroxy vitamin D levels (continuous), patient group (categorical: premenopausal women, postmenopausal women, men ≤ 45 and men > 45 years), EDSS at baseline (continuous), disability worsening (binary: disability worsening versus stable disease), GC exposure during follow‐up (binary: exposure versus non‐exposure) and use of medication with an effect on bone health (binary: medication vs. no medication) were employed as independent variables [ 16 ].

Two‐sided p values < 0.05 were considered statistically significant. Statistical analyses were performed using R [ 29 ].

Ethics and Consent

The study was approved by the ethics committee of the Medical University of Innsbruck (approval number: 1261/2018). Written informed consent was obtained from all study participants.

Literature Review

We conducted a literature search using MEDLINE and Google Scholar (November 2024). Search terms were: “risk factors reduced bone mineral density multiple sclerosis”, “risk factors bone loss multiple sclerosis”, “multiple sclerosis AND osteoporosis”, “multiple sclerosis AND bone mineral density”, “risk factors for bone loss in multiple sclerosis”, “risk factors for bone loss in multiple sclerosis longitudinal”, “risk factors for bone loss in multiple sclerosis follow‐up”. Abstracts that did not primarily address longitudinal data on osteoporosis, BMD and MS were excluded. The reference lists of selected studies were also screened and, if relevant, data were included in the review.

Results

Of the 159 pwMS initially included, 139 (87%) underwent a 2‐year follow‐up with BMD measurement. In one patient, DXA of only the lumbar spine could be carried out due to a bilateral hip replacement. For the detailed reasons for the lack of follow‐up see Figure  1 . Table  1 shows the demographic and clinical data at baseline and after 2 years. The demographic data between the total baseline cohort and pwMS with existing follow‐up is comparable (Table  S1 ). At the time of the 2‐year visit, pwMS had a median age of 50 years (IQR 39–57) and showed a female predominance of 69%. Of the 48 premenopausal women, 6 (13%) converted to a postmenopausal status.

Inclusion flow chart. DXA, dual‐energy X‐ray absorptiometry.

TABLE 1 Characteristics of 139 pwMS with 2‐year follow‐up. Baseline After the 2 year follow‐up
Demographics
Age, years 47 (37–55) 50 (39–57) Sex (females), n (%) 96 (69) 96 (69)
Clinical characteristics
BMI 24 (21–28) 25 (22–28) Abdominal circumference (cm) 90 (80–101) 91 (82–104) Alcohol consumption (glasses per week) 0 (0–2) 0 (0–2) Current smoking, n (%) 39 (28) 35 (25) Ever smoking, n (%) 94 (68) 94 (68)
Socioeconomic status
Occupational status, n (%) Working 94 (68) 89 (64) Retired 33 (24) 43 (31) Unemployed 12 (9) 7 (5) Family status, n (%) Single 36 (26) 35 (25) Divorced 15 (11) 18 (13) Widowed 3 (2) 3 (2) Relationship 16 (11) 17 (12) Married 69 (50) 66 (49) Educational grade, n (%) ≤ 9 years of schooling 18 (13) 12 (9) Secondary schooling 39 (28) 35 (25) Apprenticeship 41 (30) 45 (32) Tertiary degree 36 (26) 37 (27) Others 5 (4) 10 (7) Baecke questionnaire total 2.7 (2.2–3.2) 2.6 (2.0–3.1) Work activity 2.1 (1.3–2.7) 2.1 (1.4–2.7) Sport activity 3.0 (2.0–3.8) 2.5 (2.0–3.0) Nonsport leisure activity 3.0 (2.3–3.7) 3.0 (2.3–3.7)
Comorbidities and concomitant medication , n ( % ) Depression 23 (17) 20 (14) Diabetes type I 1 (1) 1 (1) Diabetes type II 1 (1) 3 (2) Hyperlipidemia 15 (11) 22 (16) Arterial hypertension 22 (16) 28 (20) Thyroid disease 32 (23) 32 (23) Rheumatoid arthritis 0 (0) 0 (0) Cardiovascular disease 6 (4) 6 (4) Anti‐seizure treatment 15 (11) 24 (17) Anti‐spasmodic treatment 18 (13) 26 (19) Vitamin D intake 20 (14) 59 (42) Medication with effect on bone health 20 (14) 60 (43) Familial osteoporosis 49 (35) 49 (35)
Fractures , n ( % ) Fractures in the last 12 month before screening 0 (0) NA Fractures during follow‐up NA 8 (5.8) Of those, atraumatic fractures NA 2 (1.4) Ever fractures 70 (51) 70 (51) Parental hip fracture 14 (10) 14 (10)
MS characteristics
Disease duration, years 10 (3–16, range 0–35) 12 (5–19, range 2–37) EDSS score 2.0 (1–4, range 0–8) 2.5 (1–5, range 0–7.5) Confirmed disability worsening, n (%) NA 29 (20.9) Relapses since diagnosis 3 (1–6, range 0–18) 3 (2–6, range 0–18) PwMS with relapse during follow‐up NA 25 (18.0) Relapse during follow‐up NA 0 (0–0, range 0–2) Cumulative lifetime GC dose in g 11.5 (4–20, range 0–48) 12 (6–21, range 0–48) GC dose during follow‐up in g NA 3 (3–5, range 3–8) GC during follow‐up, n (%) NA 24 (17.3)
DMT
Number of different DMT 1 (1–2, range 0–5) 2 (1–3, range 0–5) No DMT, n (%) 45 (32) 49 (35) Moderately effective DMT, n (%) 35 (25) 30 (22) Highly effective DMT, n (%) 59 (42) 60 (43)
Note: Data are given as median, interquartile range and range, unless specified otherwise directly in the table. Abbreviations: BMI, body‐mass index; DMT, disease‐modifying treatment; EDSS, Expanded Disability Status Scale; g, gram; GC, glucocorticoid; MS, multiple sclerosis; n , number; NA, not applicable; PwMS, people with multiple sclerosis.

Characteristics of 139 pwMS with 2‐year follow‐up.

Note: Data are given as median, interquartile range and range, unless specified otherwise directly in the table.

Abbreviations: BMI, body‐mass index; DMT, disease‐modifying treatment; EDSS, Expanded Disability Status Scale; g, gram; GC, glucocorticoid; MS, multiple sclerosis; n , number; NA, not applicable; PwMS, people with multiple sclerosis.

At baseline, 102 (73%) were classified as relapsing disease course, while 37 (27%) had a progressive disease course. During the follow‐up, 10 converted into a secondary progressive disease course, ending up with a total of 47 (33%) patients with progressive and 92 (67%) with relapsing disease course. The median disease duration was 12 years at follow‐up, with a range of 2–37 years. The median EDSS at baseline was 2.0 and slightly increased to 2.5 at follow‐up.

After 2 years, pwMS had received a median of two different DMTs, in contrast to the baseline visit with 1 median DMT. At the time of the 2‐year assessment, 49 (35%) were untreated, 30 (22%) and 60 (43%) received mDMt or hDMT, respectively.

During the 2‐year observation period, 24 (17%) pwMS had a relapse and received high‐dose GC therapy. The median dose was 3 g methylprednisolone (IQR 3–5, range 3–8). Long‐term low‐dose GC therapy was not used in any subjects. In total, 29 (21%) of the patients suffered disability worsening during the 2‐year follow‐up.

During the follow‐up, medication with an influence on bone density was taken by 60/139 pwMS. Only five patients took osteoporosis‐specific medication; the remaining patients took either vitamin D or vitamin D combined with calcium. At baseline, only 24 (15%) used vitamin D supplementation. Median levels of 25‐hydroxy vitamin D were 58 nmol/L (IQR 38–75).

Eight (6%) patients suffered a fracture between the baseline and follow‐up visit; those were significantly older compared to patients without a fracture (median 59, 46–62 compared to median 47, 37–54, p  = 0.030). The median EDSS was significantly higher in pwMS with a fracture than without a fracture (median 4.4, 2.8–5.8 compared to median 2.0, 1.0–4.0, p  = 0.026). Lumbar spine BMD (pwMS without fracture: 0.994, 0.878–1.106, pwMS with fracture: median 0.887, 0.764–1.056, p  = 0.140) and hip BMD (pwMS without fracture: median 0.909, 0.828–1.015 compared to pwMS with fracture: median 0.760, 0.546–0.890, p  = 0.024) were lower in patients with fractures than without fractures.

Bone Loss Over 2 Years in pwMS

Whereas no BMD loss was observed in the lumbar spine, there was a significant loss of BMD in the total hip (Table  2 ). The median total hip BMD was 0.898 g/cm 2 (0.808–1.014) and decreased to 0.882 g/cm 2 (0.784–1.01, p  < 0.001). Overall, total hip BMD loss was observed in 101 (73%) of the patients. In these 101 subjects, the median BMD loss was 3.5%. In addition, there was a significant decrease in total hip T ‐score during the follow‐up. In the patient group with DXA follow‐up after 2 years, 41/139 had osteoporosis at baseline, and after 2 years, 43/139 were diagnosed with osteoporosis.

TABLE 2 Bone mineral density changes during a follow‐up of 2 years in 139 people with multiple sclerosis. Baseline Year 2
p
Lumbar spine BMD (g/cm 2 ) 0.991 (0.872–1.105) 0.993 (0.868–1.106) 0.614 Total hip BMD (g/cm 2 )
a


0.898 (0.808–1.014)

0.882 (0.784–1.01)

< 0.001
Lumbar spine T‐score −0.6 (−1.8–0.4) −0.6 (−1.8–0.4) 0.785 Total hip T‐score
a


−0.5 (−1.2–0.4)

−0.6 (−1.5–0.1)

< 0.001
Osteopenia, n (%) 58 (41.7) 57 (41.0) 0.924 Osteoporosis, n (%) 41 (29.5) 43 (30.9) 0.754
Note: Data are shown as median and IQR unless specified otherwise. p ‐values < 0.05 were considered statistically significant and are marked bold.
a
Data are available of 138 patients.

Bone mineral density changes during a follow‐up of 2 years in 139 people with multiple sclerosis.

Note: Data are shown as median and IQR unless specified otherwise. p ‐values < 0.05 were considered statistically significant and are marked bold.

Data are available of 138 patients.

Multivariable logistic regression revealed that disability worsening was a significant predictor of hip bone loss, while the remaining variables had no statistically significant impact. Individuals with disability worsening had a 7 times higher probability to experience bone loss in the hip compared to those without worsening (Table  3 ).

TABLE 3 Logistic regression analysis identifying predictors of short‐term bone loss in people with multiple sclerosis. Estimate Standard Error P value Odds ratio 95% CI Constant −0.542 1.266 0.668 0.581 0.046 6.793 Current smoking
a

−0.092 0.448 0.838 0.912 0.275 1.809 BMI 0.051 0.046 0.269 1.052 0.952 1.146 Postmenopausal women
b

−0.023 0.495 0.963 0.977 0.368 2.600 Men ≤ 45 years
b

0.641 0.612 0.296 1.898 0.597 6.838 Men > 45 years
b

−0.468 0.704 0.506 0.626 0.157 2.591 EDSS score 0.019 0.111 0.865 0.999 0.866 1.386 Disability worsening
a


1.951

0.784

0.013

7.034

1.857
46.587 High‐dose GC therapy during follow‐up
a

−0.530 0.538 0.325 0.589 0.206 1.739 Medication with effect on bone health
a

0.058 0.431 0.893 1.060 0.457 2.495
Note: McFadden's R
2  = 0.084. p ‐values < 0.05 were considered statistically significant and are marked bold. Abbreviations: BMI, body‐mass index; EDSS, Expanded Disability Status Scale; GC, glucocorticoid.
a
Reference category = no.
b
Reference category = premenopausal women.

Logistic regression analysis identifying predictors of short‐term bone loss in people with multiple sclerosis.

Note: McFadden's R
2  = 0.084. p ‐values < 0.05 were considered statistically significant and are marked bold.

Abbreviations: BMI, body‐mass index; EDSS, Expanded Disability Status Scale; GC, glucocorticoid.

Reference category = no.

Reference category = premenopausal women.

PwMS with disability worsening experienced a significantly more pronounced median bone mass change of −3.8% (−8.1 to −1.5), compared to −1.9% median change (−4.9 to −1.2) in those without disability worsening (Figure  2 ).

Percentage of total hip bone loss in people with multiple sclerosis without or with disability worsening.

Furthermore, when baseline 25‐hydroxy vitamin D levels were added as an additional independent variable, the results remained unchanged; the data supporting this analysis are presented in the supplemental material (Table  S2 ).

Literature Review

We identified one retrospective and 10 prospective longitudinal studies on bone loss in pwMS. Data are shown in Table  4 and Table  S3 .

TABLE 4 Results of studies examining the long‐term changes in bone health in individuals with multiple sclerosis. Author Number of pwMS Study design Disability (EDSS) Follow‐up in years Bone loss y/n Main results of longitudinal data on bone health in MS Cosman [ 8 ] 54 Prospective, longitudinal study


c
6.5 [prem]
6.3 [postm]
7.0 [m]
2 y
(1) Significant more frequent atraumatic fractures in pwMS ( n  = 12, 22%) than in HC (1, 2%).
(2) PwMS lost more BMD in the FN than HCs (3% and 6% p.y. in prem PwMS and postm. PwMS vs. 0.5% and 0.8% in HCs; 7.3% p.y. in mPwMS versus 1.6% p.y. in HC)
(3) Bone loss rates: FN 3‐6x higher in pwMS than in HC, LS: Higher rates only in women with MS
(4) Bone loss rates were associated with EDSS in both hip and spine (spine: r  = −0.40, p  = 0.008; hip: r  = −0.38, p  = 0.01)
(5) Trend of a correlation of GC treatment beyond 5 months with BMD loss
(6) Serum vitamin D levels: FN BMD loss: Non significant slight difference low Vit. D group (< 20 ng/mL) compared to normal Vit. D (5.6% p.y. vs. 4.3% p.y.); spine BMD loss: low Vit. D (1.9% per year, p  < 0.04), normal Vit. D levels no decrease
Dovio [ 30 ] 10 Prospective, single arm

e
1.9 ± 0.3 0.5 n (1) No change in BMD was observed 6 months after GC treatment McKenna [ 6 ] 29 Longitudinal cohort study

d
2.9 ± 1.9

d
4.0 ± 0.4 y
(1) No BMD change at LS
(2) Decrease hip BMD average 2.9% ( p  = 0.002) (3) No correlation of disease duration, EDSS, change in EDSS, and GC treatment and change in LS or hip BMD
Ravnborg [ 31 ] 338 Multicentre, double‐blind, randomized, placebo‐controlled, parallel‐group trial to assess effect of GC in combination with INF


a
2 (0–6) [MP]
2 (0–7) [Placebo]
3.25 y (1) Slight decrease in lumbar BMD, no difference between the treatment and placebo group (0.00% and −0.99%) Schwid [ 32 ] 17 Prospective, single arm

d
5.2 ± 1.2 0.5 n
(1) LS BMD: 1.7% increase compared to baseline ( p  = 0.02)
(2) FN BMD: overall no change, loss of 1.6% BMD in pwMs and EDSS > 5.5 ( n  = 10, p  = 0.04)
(3) No evidence of effect of iv MP on BMD change
Simeakis [ 33 ] 10 Prospective observational pilot study, single arm Inclusion criteria EDSS ≤ 1 2 n (1) No BMD loss 24 months after baseline at any site Sorensen [ 34 ] 130 Randomized, placebo‐controlled trial to assess the effect of oral Methylprednisolone as add‐on therapy to INF


d
2.5 ± 1.3 [MP]
2.9 ± 1.4 [Placebo]
1.85 y
(1) Osteoporosis was recorded as an adverse event in three patients in each group (4.5% and 4.7%).
(2) LS mean BMD reduction of 2.8% ± 6.3 [MP], reduction of 0.6% ± 5.9 [Placebo]
Steffensen [ 35 ] 68 Randomized, placebo‐controlled phase II trial


a
2.5 (0–4.5) [Vit D3]
2.0 (0–4.5) [Placebo]
1.85 y
(1) BMD decrease: total hip by 1.4% (SD 2.7) in the placebo group (significant) and by 0.7% (2.7) in the treatment group, LS non‐significant (placebo: −0.1%, 3.6; treatment group: −0.3%; 2.7)
(2) No effect of intervention on BMD loss
(3) No significant associations BMD percentage change and sex, age, BMI, baseline BMD, baseline/change serum vitamin D, serum testosterone in men at week 96, and EDSS.
(4) Higher BMD change in perimenopausal women
Stepan [ 36 ] 70 Longitudinal cohort study

d
4.4 ± 1.9

d
1.8 ± 0.8 years y (1) Rate of bone loss at FN and TF: significant correlation with disability Then Bergh [ 37 ] 9 Prospective, open, single‐cross‐over study

e
1.5 ± 0.3 1 n (1) No BMD loss at LS Tyblova [ 7 ] 438 Retrospective analysis of prospectively collected data

b
3.5 (2–5.5)

d
5 ± 2.5 y
(1) Osteopenia and osteoporosis prevalence increased to 47%–56% and 9%–16%, respectively
(2) BMD significantly decreased in the FN and TF ( p  ≤ 10–11)
(3) Univariable models: significant associations for BMD changes and disability change or increment of cumulative steroid dose ( p  ≤ 0.005), with exception of EDSS and BMD changes at the LS ( p  = 0.8)
(4) Multivariable models: independent negative associations between EDSS category change and BMD changes at the FN and TF ( p  ≤ 0.003)
(5) Fractures: 79 low‐trauma fractures in 51 pwMS, cumulative hazard association with time from menopause ( b  = 0.2, p  = 10 14), age ( b  = 0.03, p  = 0.008) and EDSS 6 ( b  = 13, p  = 0.02), but not with cumulative steroid dose ( p  = 0.9)
Abbreviations: BMD, bone mineral density; EDSS, expanded disability status scale; FN, femoral neck; FU, follow‐up; GC, glucocorticoid; INF, interferon; LS, lumbar spine; m, male; mo, month; MP, methylprednisolone; n, no; na., not available; p.y., per year; postm, postmenopausal; prem, premenopausal; pwMS, people with multiple sclerosis; TF, total femur; Vit, vitamin; vs., versus; y, yes.
a
Median + range.
b
Median + interquartile range.
c
Mean.
d
Mean, standard deviation.
e
Mean, standard error of the mean.

Results of studies examining the long‐term changes in bone health in individuals with multiple sclerosis.

c
6.5 [prem]

6.3 [postm]

7.0 [m]

(1) Significant more frequent atraumatic fractures in pwMS ( n  = 12, 22%) than in HC (1, 2%).

(2) PwMS lost more BMD in the FN than HCs (3% and 6% p.y. in prem PwMS and postm. PwMS vs. 0.5% and 0.8% in HCs; 7.3% p.y. in mPwMS versus 1.6% p.y. in HC)

(3) Bone loss rates: FN 3‐6x higher in pwMS than in HC, LS: Higher rates only in women with MS

(4) Bone loss rates were associated with EDSS in both hip and spine (spine: r  = −0.40, p  = 0.008; hip: r  = −0.38, p  = 0.01)

(5) Trend of a correlation of GC treatment beyond 5 months with BMD loss

(6) Serum vitamin D levels: FN BMD loss: Non significant slight difference low Vit. D group (< 20 ng/mL) compared to normal Vit. D (5.6% p.y. vs. 4.3% p.y.); spine BMD loss: low Vit. D (1.9% per year, p  < 0.04), normal Vit. D levels no decrease

(1) No BMD change at LS

(2) Decrease hip BMD average 2.9% ( p  = 0.002) (3) No correlation of disease duration, EDSS, change in EDSS, and GC treatment and change in LS or hip BMD

a
2 (0–6) [MP]

2 (0–7) [Placebo]

(1) LS BMD: 1.7% increase compared to baseline ( p  = 0.02)

(2) FN BMD: overall no change, loss of 1.6% BMD in pwMs and EDSS > 5.5 ( n  = 10, p  = 0.04)

(3) No evidence of effect of iv MP on BMD change

d
2.5 ± 1.3 [MP]

2.9 ± 1.4 [Placebo]

(1) Osteoporosis was recorded as an adverse event in three patients in each group (4.5% and 4.7%).

(2) LS mean BMD reduction of 2.8% ± 6.3 [MP], reduction of 0.6% ± 5.9 [Placebo]

a
2.5 (0–4.5) [Vit D3]

2.0 (0–4.5) [Placebo]

(1) BMD decrease: total hip by 1.4% (SD 2.7) in the placebo group (significant) and by 0.7% (2.7) in the treatment group, LS non‐significant (placebo: −0.1%, 3.6; treatment group: −0.3%; 2.7)

(2) No effect of intervention on BMD loss

(3) No significant associations BMD percentage change and sex, age, BMI, baseline BMD, baseline/change serum vitamin D, serum testosterone in men at week 96, and EDSS.

(4) Higher BMD change in perimenopausal women

(1) Osteopenia and osteoporosis prevalence increased to 47%–56% and 9%–16%, respectively

(2) BMD significantly decreased in the FN and TF ( p  ≤ 10–11)

(3) Univariable models: significant associations for BMD changes and disability change or increment of cumulative steroid dose ( p  ≤ 0.005), with exception of EDSS and BMD changes at the LS ( p  = 0.8)

(4) Multivariable models: independent negative associations between EDSS category change and BMD changes at the FN and TF ( p  ≤ 0.003)

(5) Fractures: 79 low‐trauma fractures in 51 pwMS, cumulative hazard association with time from menopause ( b  = 0.2, p  = 10 14), age ( b  = 0.03, p  = 0.008) and EDSS 6 ( b  = 13, p  = 0.02), but not with cumulative steroid dose ( p  = 0.9)

Abbreviations: BMD, bone mineral density; EDSS, expanded disability status scale; FN, femoral neck; FU, follow‐up; GC, glucocorticoid; INF, interferon; LS, lumbar spine; m, male; mo, month; MP, methylprednisolone; n, no; na., not available; p.y., per year; postm, postmenopausal; prem, premenopausal; pwMS, people with multiple sclerosis; TF, total femur; Vit, vitamin; vs., versus; y, yes.

Median + range.

Median + interquartile range.

Mean.

Mean, standard deviation.

Mean, standard error of the mean.

Discussion

In this 2‐year prospective study, BMD significantly decreased in the total hip, with 101 (73%) of pwMS experiencing a significant loss (median 3.5%). Disability worsening was the key predictor of hip BMD loss, increasing the likelihood by 7 times. PwMS who had fractures during follow‐up were older, more disabled, and had lower baseline BMD, indicating a link between lower BMD and higher fracture risk.

In another prior 2‐year prospective study, hip BMD loss was significantly higher in 54 pre‐ (3% per year) and postmenopausal women (6% per year) and men (7% per year) with MS, compared to controls. A smaller difference was observed in the spine, with significant changes noted only in women. Risk factors for BMD loss included GC treatment for over 5 months, disability, and reduced ambulatory ability. In postmenopausal women and men, BMD loss was almost twice as high as in premenopausal women [ 8 ].

Whereas in our cohort, postmenopausal women and men had a significantly higher risk of osteoporosis at baseline [ 16 ] loss of BMD over 2 years was predominantly driven by disability worsening. A plausible explanation for our findings is that disability worsening, as a dynamic measure, captures change over time in contrast to baseline EDSS, which is a static measure reflecting disability cross‐sectionally at a single timepoint. Disability worsening may have a higher impact on BMD, as this variable provides further information on disease progression. Disability worsening may reflect ongoing inflammatory activity impacting bone metabolism, an effect not captured by baseline EDSS alone.

The discrepancies with variable bone loss rates in our study compared to previous reports likely arise from differences in study populations, particularly with regard to disease severity with a significantly higher degree of disability with an average EDSS (6.3–7.0) [ 8 ]. A 96‐week phase II trial involving pwMS randomized to receive vitamin D3 or placebo, combined with 500 mg calcium daily, demonstrated a significant decrease in hip BMD loss (1.4% in the placebo group and 0.7% in the treatment group), but no change in spine BMD [ 35 ]. While the median EDSS in our study was similar, our cohort included a wider range of disability (EDSS 0–8), potentially allowing for a greater likelihood of disability worsening, particularly in more advanced disease. In contrast to our study, the trial focused on fully ambulatory (EDSS < 4.5), younger 18–50, and premenopausal participants, which may contribute to the differences in BMD loss observed between our studies.

A greater short‐term loss of bone mass in the femur, in which bone density is particularly determined by weight bearing activity compared to the spine, is probably due to the stronger effect of impaired ambulation on the femur. A larger retrospective study with 438 pwMS over 5 years also found that reduced ambulation (defined as EDSS ≥ 6) impacted BMD more significantly in weight‐bearing bones (femur) compared to the lumbar spine. Similarly to our results, an association of BMD change with change in disability adjusted for follow‐up duration and age was shown in a multivariable regression analysis including cumulative steroid dose change and interaction of change in EDSS category and cumulative steroid dose [ 7 ].

In the only other study examining the association of BMD loss and disability change over time, a cohort of 29 pwMS with low‐effective DMT was assessed over 4 years [ 6 ]. A modest BMD loss of 2.9% at the total hip was observed, with no significant associations found between BMD changes and factors such as sex, age, MS duration, DMT duration, GC therapy, EDSS or EDSS change, or vitamin D levels. A notable limitation of the study was the relatively small sample size, which was likely underpowered to detect statistically significant associations between changes in BMD and factors such as disability worsening or demographic variables. This is particularly important given the subtle nature of the BMD changes observed, with only a modest loss in total hip BMD over 4 years. Additionally, the absence of a standardized definition of disability worsening in both analyses limits the comparability with our findings, where we adhered to a well‐established definition of disability worsening, which allowed for a more precise correlation between disability status and BMD changes over time. We were unable to confirm findings by Cosman et al. that low baseline vitamin D levels significantly influence BMD loss, as their study showed a 5.6% annual loss in pwMS with low vitamin D compared to 4.3% with high levels [ 8 ].

In addition, no effect of an intervention with vitamin D 3 once a week and 500 mg calcium daily on BMD at any site was observed after 96 weeks in a phase II study [ 35 ]. Even though our study was not randomized, there was also no significant effect of vitamin D supplementation (with or without calcium) on BMD observed. This suggests that the short‐term impact of vitamin D on BMD changes in pwMS may be minimal, and confounding factors such as sunlight exposure and physical activity might contribute to the results of vitamin D levels and BMD loss found in other studies [ 8 , 38 ].

Our study found no association between GC treatment during the 2‐year follow‐up and bone loss. We also observed no negative impact of cumulative lifetime GC dose on baseline osteoporosis risk; no patients received long‐term GC therapy [ 16 ]. Two clinical trials evaluating the effect of high‐dose monthly GC treatment as an add‐on to interferon beta showed minimal changes in lumbar spine BMD, suggesting no significant effect of short‐term GC use [ 31 , 34 ]. Similarly, a study on high‐dose GC treatment in newly diagnosed pwMS found no impact on BMD after 24 months, though the small sample size ( n  = 10) limits the generalizability of these results [ 33 ]. The results of no relevant effect of intermittent short‐term high‐dose GC are confirmed by several cross‐sectional studies [ 4 , 6 , 9 , 10 , 14 , 16 , 30 , 32 , 39 ], but not all [ 7 , 8 , 18 ]. For a detailed overview of the literature review of all longitudinal studies on BMD changes in pwMS and the influencing factors considered, see Table  4 and Table  S3 .

There are several limitations of our study. Firstly, a relatively large proportion of pwMS used medications affecting bone health during the study period. While the prospective design and high baseline osteoporosis rates necessitated treatment, no significant short‐term effect was reported in the aforementioned phase II study, and these medications were accounted for in regression analyses [ 35 ].

Secondly, long‐term low dose GC treatment could not be included in the analysis, although this is not very relevant as long‐term low dose GC treatment is not used in MS. Additionally, disability range and treatment patterns may differ from other MS populations. Finally, the relatively short follow‐up duration limits our ability to fully assess osteoporosis risk factors, given its gradual progression. This restricts our ability to capture long‐term changes in bone health or identify risk factors with smaller effect sizes. However, a longer follow‐up is currently ongoing, which will provide these aspects in thefuture.

In conclusion, this study provides first evidence for the association of disability worsening: and bone density reduction in pwMS over a short‐term period. This relationship highlights the importance of osteoporosis screening in patients at high risk to reduce fracture risk and improve quality of life.

Author Contributions

Anne Zinganell: conceptualization, writing – original draft, writing – review and editing, investigation, project administration. Harald Hegen: writing – original draft, writing – review and editing, data curation, formal analysis, conceptualization, investigation, supervision, visualization, methodology. Janette Walde: formal analysis, writing – review and editing. Robert Barket: investigation, writing – review and editing. Klaus Berek: investigation, writing – review and editing. Michael Auer: investigation, writing – review and editing. Martin Schmidauer: investigation, writing – review and editing, project administration. Gabriel Bsteh: writing – review and editing. Alexander Stephan Kroiss: investigation, writing – review and editing. Andrea Griesmacher: writing – review and editing, investigation. Birgit Waldner: project administration, writing – review and editing, investigation. Alexander Tschoner: writing – review and editing. Thomas Berger: writing – review and editing. Florian Deisenhammer: writing – review and editing. Franziska Di Pauli: writing – review and editing, conceptualization, investigation, writing – original draft, methodology, validation, formal analysis, project administration, supervision, data curation.

Conflicts of Interest

A.Z. has participated in meetings sponsored by, received speaking honoraria or travel funding from Biogen, Merck, Novartis, Sanofi‐Genzyme, Janssen, Bristol Myers Squibb, and Teva. H.H. has participated in meetings sponsored by, received speaker honoraria or travel funding from Bayer, Biogen, Bristol Myers Squibb, Horizon, Janssen, Merck, Novartis, Sanofi‐Genzyme, Siemens, Teva, and received honoraria for acting as a consultant for A‐med, Biogen, Bristol Myers Squibb, Novartis, Roche, Sanofi‐Genzyme, and Teva. J.W. has nothing to disclose. R.B. has participated in meetings sponsored by or received travel grants from Novartis, Janssen‐Cilag, Merck, and Sanofi‐Genzyme. Received honoraria from Janssen‐Cilag and Biogen. K.B. has participated in meetings sponsored by and received travel funding or speaker honoraria from Roche, Teva, Merck, Biogen, Sanofi, and Novartis. He is an associate editor of Frontiers in Immunology/Neurology, Section Multiple Sclerosis and Neuroimmunology. M.A. received speaker honoraria, honoraria for consulting and/or travel grants from Biogen, Novartis, Merck, Sanofi, Horizon Therapeutics/Amgen, and Zentiva. M.S. has participated in meetings sponsored by or received travel grants from Novartis, Sanofi‐Genzyme, and Amgen. G.B. has participated in meetings sponsored by, received speaker honoraria or travel funding from Biogen, Celgene/BMS, Janssen, Lilly, Medwhizz, Merck, Novartis, Roche, Sanofi‐Genzyme, and Teva, and received honoraria for consulting Adivo Associates, Biogen, Celgene/BMS, Janssen, Merck, Novartis, Roche, Sanofi‐Genzyme, and Teva. He has received unrestricted research grants from Celgene/BMS and Novartis. He serves on the Executive Committee of the European Committee for Treatment and Research in Multiple Sclerosis (ECTRIMS). A.S.K. has nothing to disclose. A.G. has nothing to disclose. B.W. has nothing to disclose. A.T. has nothing to disclose. T.B. has participated in meetings sponsored by and received honoraria (lectures, advisory boards, consultations) from pharmaceutical companies marketing treatments for MS: Amgen, Allergan, Bayer, Biogen, Bionorica, BMS, Genesis, GSK, GW/Jazz Pharma, Horizon, Janssen, MedDay, Merck, Neuraxpharma, Novartis, Octapharma, Roche, Sandoz, Sanofi, Teva, TG Therapeutics, and UCB. His institution has received financial support in the past 12 months by unrestricted research grants (Biogen, Bayer, BMS, Merck, Novartis, Roche, Sanofi, Teva) and for participation in clinical trials in multiple sclerosis sponsored by Alexion, Bayer, Biogen, Merck, Novartis, Octapharma, Roche, Sanofi, and Teva. F.D. has participated in meetings sponsored by or received honoraria for acting as an advisor/speaker for Alexion, Almirall, A‐Med, Amgen/Horizon, Biogen, BMS, Sanofi, Janssen, Laurea Group, Medwhizz, Merck, Novartis Pharma, Neuraxpharm, Roche, Sandoz, and Teva. He is section editor of the MSARD Journal (Multiple Sclerosis and Related Disorders) and review editor of Frontiers in Neurology. F.D.P. has participated in meetings sponsored by, received honoraria (lectures, advisory boards, consultations), or travel funding from Amgen, Bayer, Biogen, Celgene, BMS, Horizon, Merck, Novartis, Sanofi‐Genzyme, Teva, and Roche.

Supporting information

Table S1: Comparison of baseline characteristics of pwMS with and without 2‐year follow‐up.

Table S2: Logistic regression analysis identifying predictors of short‐time bone loss in people with multiple sclerosis.

Table S3: Detailed information on the studies included in the literature review of bone health dynamics in multiple sclerosis.

Acknowledgement

Open access funding provided by Medizinische Universitat Innsbruck/KEMÖ.

Harald Hegen and Franziska Di Pauli contributed equally to this work.

Contributor Information

Harald Hegen, Email: harald.hegen@i-med.ac.at.

Franziska Di Pauli, Email: franziska.dipauli@i-med.ac.at.

Data Availability Statement

Data supporting the findings of this study are available from the corresponding author upon reasonable request by a qualified researcher and upon approval by the ethics committee of the Medical University of Innsbruck.

References

Associated Data

Supplementary Materials

Table S1: Comparison of baseline characteristics of pwMS with and without 2‐year follow‐up.

Table S2: Logistic regression analysis identifying predictors of short‐time bone loss in people with multiple sclerosis.

Table S3: Detailed information on the studies included in the literature review of bone health dynamics in multiple sclerosis.

Data Availability Statement

Data supporting the findings of this study are available from the corresponding author upon reasonable request by a qualified researcher and upon approval by the ethics committee of the Medical University of Innsbruck.