Changes in bone density and bone turnover in patients with rheumatoid arthritis treated with rituximab, results from an exploratory, prospective study
Changes in bone density and bone turnover in patients with rheumatoid arthritis treated with rituximab, results from an exploratory, prospective study
Gillian Wheater 1 2 3
Mohsen Elshahaly 0 1 3
Kamran Naraghi 3
Stephen P. Tuck 1 3 5
Harish K. Datta 1 3
Jacob M. van Laar 1 3 4
0 Rheumatology and Physical Medicine Department, Suez Canal University, Ismailia, Egypt, 4 Leeds Institute of Rheumatic and Musculoskeletal Medicine, University of Leeds , Leeds , United Kingdom
1 Institute of Cellular Medicine, Newcastle University , Newcastle-upon-Tyne , United Kingdom
2 Department of Biochemistry, The James Cook University Hospital , Middlesbrough , United Kingdom
3 Editor: Michael Nurmohamed, VU University Medical Center , NETHERLANDS
4 Department of Rheumatology and Clinical Immunology, University Medical Center Utrecht , Utrecht , The Netherlands
5 Department of Rheumatology, The James Cook University Hospital , Middlesbrough , United Kingdom
Data describing the effect of in vivo B cell depletion on general bone loss in patients with rheumatoid arthritis (RA) are limited. Given the pathogenetic role of B cells in RA, it is tempting to speculate that B cell depletion might have a beneficial effect on bone loss. We prospectively investigated the changes in bone mineral density (BMD), bone turnover, inflammation and disease activity before and after rituximab in 45 RA patients over a 12 month period, 36 patients of whom completed the study and were included in the analysis. There was no significant change in our primary endpoint; lumbar spine BMD after 12 months. However, we found a significant decrease in neck of femur (mean -0.017 g/cm2, 95% CI -0.030, -0.004 p = 0.011) and total femur BMD (mean -0.016 g/cm2, 95% CI -0.025, -0.007 p = 0.001). Additionally, there was a significant increase in procollagen type 1 aminoterminal propeptide (P1NP) and bone specific alkaline phosphatase (BAP); biomarkers of bone formation (median change from baseline to 12 months; P1NP 11.3 μg/L, 95% CI -1.1, 24.8 p = 0.025; BAP 2.5 μg/L, 95% CI 1.2, 3.6 p = 0.002), but no significant change in bone resorption or osteocyte markers. The fall in BMD occurred despite improvement in disease control. Post-menopausal women had the lowest mean lumbar spine, femoral and forearm BMD at baseline and after 12 months, additionally they had a higher level of bone turnover throughout the study. In conclusion, BMD was maintained at the lumbar spine and forearm, but fell at the femur sites. A high prevalence of vitamin D deficiency was observed and these patients had lower BMD and evidence of higher bone turnover.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information
Funding: This work was supported by [grant no.
MRA024] Roche Products Limited (Welwyn
Garden City, UK). Roche had no role in the study
design, data collection and analysis, the decision to
publish or the preparation of the manuscript.
Roche conducted a factual accuracy review on the
final article, but any decision to incorporate
comments was made solely at the discretion of the
Competing interests: JMvL has received honoraria
from Roche, Eli Lilly, MSD, Pfizer, Arthrogen and
BMS, consultancy fees from Janssen, plus funding
from Astra Zeneca, MSD and Genentech. GW, ME,
KN, SPT and HKD declare that they have no conflict
of interest. The company Roche provided funding
for the study (no salaries) and also viewed the
article prior to sending for review, but did not
change it and had no input into the conduction of
the study, analysis or writing up. This does not
alter our adherence to PLOS ONE policies on
sharing data and materials.
Rheumatoid arthritis (RA) is the most prevalent inflammatory joint disease, in which B cells
play an important role. Depletion of B cells by the anti-CD20 antibody rituximab is a highly
effective treatment of RA, which is now well established [
]. The skeletal complications of RA
consist of focal erosions of marginal and subchondral bone, juxta-articular osteoporosis and
generalised bone loss with reduced bone mass [
]. Patients with established RA have an annual
decrease of -3.9% and -2.5% of bone mineral density (BMD) at the lumbar spine (LS) and
femoral neck respectively [
]. Additionally some studies have reported that the greatest reduction
in BMD occurs at the foreram sites and that forearm BMD correlates with clinical features of
disease activity and markers of bone turnover [
]. The main cause of periarticular osteopenia
and local bone erosions is chronic inflammation of the synovial membranes, whereas several
factors such as disease activity, female gender, older age, glucocorticoid use and decreased
mobility are known to promote generalised bone loss with reduced bone mass in RA patients.
However disease activity is the major predictor and is independent of the other factors [
Furthermore, RA patients are reported to have lower levels of total 25-hydroxyvitamin D
(25OHD) and this is associated with increased disease activity and musculoskeletal pain [
Serum 25OHD is also negatively associated with DAS28 score, ESR, platelets, IL-17 and IL-23
and RA patients with osteoporosis and osteopenia have significantly lower levels of 25OHD
than those with normal BMD [
Recent investigations have provided abundant evidence for an intricate interaction between
the immune and skeletal systems [
], but the role of B cells in osteoclastogenesis is
]. Data describing the effect of in vivo B cell depletion on general bone loss in patients
with RA are still limited. We have previously presented work demonstrating a change in bone
turnover markers (BTMs) that suggests a potential favourable effect of rituximab on bone
density . We have therefore undertaken a preliminary exploratory, prospective study over 12
months to measure changes in bone density and to look at factors that may influence the
outcome such as change in disease activity and vitamin D status.
Materials and methods
This study (EudraCT no. 2010-020499-50) was approved by the NHS Health Research
Authority National Research Ethics Service (UK) Committee North East-Newcastle and North
Tyneside 1 (REC reference no. 10/H0906/57) and was sponsored by South Tees Hospitals NHS
Foundation Trust, UK. Recruitment took place in ten UK centres and patients were followed
up for 12 months. Written consent was obtained according to the Declaration of Helsinki.
Patients fulfilled the American College of Rheumatology (ACR) classification criteria 2010 for
the diagnosis of RA and the UK National Institute for Health and Care Excellence (NICE)
eligibility criteria for treatment with rituximab. Patients were excluded if they were less than
18yrs old, had previously received any B cell depleting agent or had been treated for
osteoporosis. Calcium, vitamin D, corticosteroids, non-biological DMARDs and treatment for
concomitant medical conditions were all continued throughout the study at the discretion of the
treating physician. Rituximab was administered following the recommended protocol as a 1g
intravenous infusion on days 1 and 15 in conjunction with intravenous methylprednisolone
100 mg. Patients who responded to the first rituximab course received a second course at 6
months unless they attained a state of low disease activity, as measured by DAS28< 3.2, in
accordance with clinical practice. The primary outcome measure was change in LS BMD. The
secondary outcomes were: change in mean total femur, mean neck of femur and mean forearm
2 / 15
BMD, change in BTMs, inflammatory markers and DAS28, all assessed between baseline and
Bone mineral density measurements
BMD was measured at baseline and 12 months after the first rituximab course by dual-energy
X-ray absorptiometry (DXA). Measurements were taken at LS (mean L2-L4), total and neck of
femur on both sides and ultra-distal radius (RUD) forearms. Femoral neck, total femur and
forearm BMD measurements were reported as the mean of the right and left sides. Two
different DXA machines were used across the 10 centres, GE Lunar Prodigy (Lunar, Madison,
Wisconsin, USA) and Hologic Discovery (Hologic, Waltham, Massachusetts, USA). However, the
same machine was used at baseline and follow-up measurement for each patient. The
interassay coefficient of variation (CV), measured using a local spine phantom, for the different
centres were all less than 1.8%.
Patients were assessed at baseline prior to rituximab treatment and then every 3 months over a
12 month follow up period. Routine laboratory investigations were performed locally and
clinical assessment of disease activity was undertaken using the 28-joint disease activity score and
wide range C-reactive protein. Fresh whole blood samples were analyzed using a flow
cytometer (FACS Canto II) to determine the numbers of CD19+ B cells in a subset of patients as per
the study protocol. Fasting morning blood samples were collected into EDTA and serum
separator tubes, the plasma and serum were separated within 60 mins and immediately stored at
-80ÊC until analysis. The inter-assay CV was <5% for automated and <15% for the manual
assays. All measurements were performed as per manufacturer's instructions and in a
centralized laboratory to reduce analytical variation.
Bone resorption markers. βCTx was quantified in plasma by electrochemiluminescent
immunoassay (ECLIA) on the E411 analyser (Roche Diagnostics, Lewes, UK). Tartrate
resistant acid phosphatase isoform 5b (TRAP-5b) was measured in serum by a manual
enzymelinked immunosorbent assay (ELISA) (Immuno-Diagnostic Systems Ltd, Bolden, UK).
Bone formation and osteocyte markers. P1NP was quantified in plasma by ECLIA on
the E411 analyser. Bone specific alkaline phosphatase (BAP) was quantified in serum by
chemiluminescence on the iSYS analyser (Immuno-Diagnostic Systems, Bolden, UK).
Sclerostin (SCL) and dickkopf-related protein 1 (DKK-1) levels were measured in serum using a
manual ELISA (Biomedica Group, 1210 Wien, Austria).
Inflammatory marker. Wide range CRP (CRP) was quantified in serum by a
latexenhanced immunoturbidimetric assay on the Advia 2400 analyser (Siemens Healthcare
Diagnostics, Camberley, UK).
Total 25-hydroxy vitamin D and parathyroid hormone. 25OHD was quantified in
serum by chemiluminescence on the iSYS analyser. Laboratory reference range (Deficient
<25; Insufficient 25±50; Adequate 50±75; Optimum >75 nmol/L) [
hormone (PTH) was quantified in plasma by ECLIA on the E411 analyser.
The determination of sample size was based on the primary endpoint, change in LS BMD 12
months after the first rituximab course. Assuming the true change in LS BMD is 0.01 g/cm2
and that the DXA scan is reproducible with a standard deviation (SD) of 0.02 g/cm2; then we
calculated that the study would have 80% power to detect a statistically significant difference
(at the 5% significance level) if 33 patients were included in the final analysis based on a one
3 / 15
sample t-test. To allow for a 20% dropout we estimated that we needed to recruit 42 patients.
Continuous data were presented as means and SD or medians and range depending on the
distribution. Categorical variables were displayed as absolute frequencies and percentages. The
primary and secondary endpoints were investigated using a one sample t-test or Wilcoxon
signed rank test if not normally distributed. The median percentage change from baseline
across 4 visits (3, 6, 9 and 12 months) was calculated for BTM's, inflammatory markers and
disease activity. Spearman's rank correlation coefficient (Rs) was used to correlate change in
BMD with change in BTMs and inflammatory markers. Missing BMD or biomarker
measurements were not imputed at any time point as we used a completer analysis approach. P values
0.05 were considered statistically significant. Statistical analyses were performed using
STATA v12 (STATA Corp, Texas, USA).
Demographic and clinical characteristics
A total of 45 patients met the eligibility criteria and were enrolled into the study (Fig 1). One
patient was diagnosed with chronic lymphocytic leukaemia (CLL) and excluded; therefore 44
patients received the first rituximab infusion. A total of 36 patients completed the 12 month
follow up period and were included in the analysis; 32 of these patients received a second
course of rituximab as per protocol and four patients did not (two patients had low disease
activity (DAS28<3.2); one patient had undetectable B cells and one patient refused the second
course). All patients were bisphosphonate naïve i.e. patients had not received bisphosphonates
at any time prior to the start of this study. No fractures occurred during the study period.
There was no significant difference in any baseline characteristic for those patients who
completed the study (n = 36) compared to the 9 non-completers and the following analysis
includes only those 36 patients who completed 12 month follow-up (Table 1). Briefly, 29
patients (81%) were female (23 of whom were post-menopausal) and seven (19%) patients
were male. The mean age was 58.6 ±12.1yrs and the mean disease duration was 10.4 ±7.0yrs.
Eighty-three percent (n = 30) and 76% (n = 25) were RF and ACPA positive respectively.
Eleven patients were on steroids at baseline and a further 7 started steroids during the study.
There was no significant difference between patients on steroids or not for change in BMD,
inflammatory or bone biomarkers from baseline to 12 months. Thirty-nine percent (n = 14) of
our patients had vitamin D deficiency defined as 25OHD levels below 25 nmol/L, only one of
these patients was on a calcium and vitamin D supplement at baseline, a further four patients
started on supplements during the course of the study. The numbers of CD19+ B cells were
determined in a subset of 16 patients at each visit; all had values less than 0.01 x109/L at 3
months and four patients had rapid reconstitution of their B cells at 6 months, while the others
had long-term depletion.
Continuous data are presented as means and standard deviation (SD) or medians and
range, depending on the distribution of the data set. Categorical variables are displayed as
absolute frequencies and percentages. Neck of femur, total femur and radius UD results are
reported as the mean of both sides. 25OHD: Total 25-Hydroxy vitamin D; ACPA: Anti-cyclic
Citrullinated Peptide Antibody; BAP: bone specific alkaline phosphatase; βCTx: β-isomerised
carboxy terminal telopeptide of type I collagen; BMI: Body Mass Index; CRP: C-Reactive
Protein; DAS28: Disease Activity Score using 28 tender and swollen joints; DKK-1:
dickkopfrelated protein 1; eGFR: estimated Glomerular Filtration Rate; ESR: Erythrocyte
Sedimentation Rate; FSH: Follicle Stimulating Hormone; HAQ: Health Assessment Questionnaire; LH:
Luteinising Hormone; P1NP: procollagen type 1 amino-terminal propeptide; PTH:
Parathyroid Hormone; RA: Rheumatoid Arthritis; RF: Rheumatoid Factor; SCL: sclerostin; SHBG:
4 / 15
Fig 1. Consort flow diagram. Forty-seven patients were screened; 2 patients did not meet the eligibility criteria and so 45 were enrolled
onto the study. One patient was subsequently diagnosed with chronic lymphocytic leukaemia therefore only 44 patients received the first
course of rituximab. A further 8 patients did not complete the study and a total of 36 patients were included in the final analysis; of these
only 32 received the second RTX course.
Sex Hormone Binding Globulin; SJC: Swollen Joint Count; TJC: Tender Joint Count;
TRAP5b: tartrate resistant acid phosphatase isoform 5b.
Bone mineral density
There was no significant change in our primary endpoint; lumbar spine BMD after 12 months.
However there was a significant decrease in mean neck of femur BMD (mean difference
-0.017 g/cm2, 95% CI -0.030, -0.004 a decrease of -2.0%; p = 0.011) and mean total femur BMD
(mean difference -0.016 g/cm2, 95% CI -0.025, -0.007 a decrease of -1.7%; p = 0.001) after 12
months (Table 2). Equally we found a significant decrease in both the left and right side
femoral BMD. Six patients showed a decline in LS BMD greater than the least significant change
(LSC) of 5%; 12 patients had a decrease in mean neck of femur BMD greater than 3%; 9
patients had a decline in mean total femur BMD greater than 3% and 12 patients had a mean
ultra-distal forearm BMD decline of greater than 3%. Seven centres used the GE Lunar Prodigy
5 / 15
Completed study (n = 36)
5.74 (2.24, 8.01)
32 (2, 128)
11.7 (0.4, 114.7)
30.2 (14.1, 87.7)
36.1 (12.0, 141.0)
30.1 (0.1, 70.9)
60.1 (0.6, 142.3)
12.4 (5.0, 26.0)
49.7 (18.5, 80.0)
423 (41, 1167)
39.8 (10.9, 209.9)
17.2 (8.3, 37.3)
3.0 (1.1, 4.9)
47.9 (18.5, 119.4)
53.4 (29.2, 88.2)
-0.4 (-2.6, 6.1)
0.4 (-2.7, 7.8)
-0.7 (-2.9, 1.0)
0.3 (-1.6, 1.9)
-0.5 (-2.7, 2.1)
0.3 (-1.8, 2.8)
Completed study (n = 36)
-1.1 (-3.8, 8.6)
-0.6 (-2.6, 10.4)
Advance, and 3 centres used Hologic Discovery QDR. There was no significant difference
between Lunar or Hologic machines for the change in lumbar spine, femoral or ultra-distal
forearm BMD measurement from baseline to 12 months. As menopausal status is known to
markedly affect bone turnover we also examined the data in gender- and menopausal
statusspecific manner. Mean BMD was plotted (Fig 2) at baseline i.e. before the 1st rituximab
infusion and after 12 months in all patients who completed the study (n = 36) and by gender and
menopausal status (7 males, 6 pre-menopausal and 23 post-menopausal females). We found
that post-menopausal women had the lowest mean lumbar spine, femoral and forearm BMD
at baseline and after 12 months.
Inflammation and disease activity
There was a significant reduction in inflammatory markers from baseline to 12 months (CRP
median change -3.7 mg/L, 95% CI -10.3, 0.7 p = 0.014, a median percentage decrease over 4
visits of 21%; ESR median change -8 mm/hr, 95% CI -19, 1 p = 0.006, a median decrease of 20%)
and disease activity (DAS28-CRP median change -0.84, 95% CI -1.64, -0.41 p<0.001, a median
percentage decrease of 19%) following treatment with rituximab (Table 3). Eleven percent
(n = 4) of patients achieved remission (DAS28 <2.6) and 25% (n = 9) at least low disease
activity (DAS28 <3.2) after 12 months. Furthermore, based on EULAR DAS28 response criteria
]; 39% (n = 14) of the study population experienced a good response (DAS28 improvement
>1.2) and 58% (n = 21) at least a moderate response (DAS28 improvement between 0.6 and
1.2) over 12 months. We found no significant difference between patients achieving DAS28
<3.2 or those with DAS28 >3.2 for change in BMD or BTMs over 12months. However, there
was a significant difference (p = 0.029) in DKK-1 levels between patients with good/moderate
response (mean DKK-1 change -7.6 pmol/L, 95% CI -19.9, 5.3) and those with poor EULAR
DAS28 response (mean DKK-1 change 8.7 pmol/L, 95% CI 0.01, 17.3) over 12 months.
Bone turnover markers
We found a significant increase in bone formation over 12 months; P1NP (median change
11.3 μg/L, 95% CI -1.1, 24.8 p = 0.025; a median percentage increase over 4 visits of 30%)) and
BAP (median change 2.5 μg/L; 95% CI 1.2, 3.6; p = 0.002; a median percentage increase of
Bone mineral density (BMD) was measured in lumbar spine (n = 36) mean L2-L4. Also neck of femur (n = 35), total femur (n = 35) and ultra-distal radius (n = 31), the
results were reported as mean of both sides. All measured at time 0 before the 1st rituximab infusion and after 12 months in patients who completed the study. Results
were expressed as mean and standard deviation (SD) at baseline and 12 months, the mean change from baseline was assessed by paired t-test.
7 / 15
Fig 2. The effects of gender and menopausal status on mean bone mineral density. Bone mineral density (BMD) was plotted at lumbar spine; mean L2-L4 (LS
BMD), also neck of femur (MN BMD), total femur (MT BMD) and ultra-distal radius (MRUD BMD) mean of both sides, at time 0 before the 1st rituximab
infusion and after 12 months in patients who completed the study (n = 36). Patients were also stratified by gender and menopausal status (7 males, 6
premenopausal and 23 post-menopausal females).
13%). There was no significant change in bone resorption or osteocyte markers (Table 3).
Median concentrations of BTMs were also plotted (Fig 3) for all patients (n = 34) at each study
visit and we reviewed the data by gender- and menopausal status-specific manner (6 males, 6
pre-menopausal and 22 post-menopausal females). Post-menopausal women had the highest
median concentrations of βCTx, P1NP, BAP and DKK-1 at baseline and throughout the 12
months, indicating a higher level of bone turnover.
Effect of vitamin D concentration
Thirty-nine percent (n = 14) of patients in our cohort were found to be vitamin D deficient,
this included 29% (n = 2) of males, 33% (n = 2) of pre- and 43% (n = 10) of post-menopausal
8 / 15
Markers of bone resorption (βCTx: β-isomerised carboxy terminal telopeptide of type I collagen; TRAP5b: tartrate resistant acid phosphatase isoform 5b), bone
formation (P1NP: procollagen type 1 amino-terminal propeptide; BAP: bone specific alkaline phosphatase), osteocyte markers (SCL: sclerostin; DKK-1:
dickkopfrelated protein 1), inflammatory markers (CRP: C reactive protein; ESR: erythrocyte sedimentation rate;) and disease activity (DAS28: disease activity score using 28
tender and swollen joints) were measured at baseline before the rituximab infusion and then at 3, 6, 9 and 12 months in all patients who completed the study. Results
were not normally distributed and were expressed as median and range, median change from baseline to 12 months was assessed using Wilcoxon signed rank test and
the median percentage change from baseline across all 4 visits was calculated.
females. There was no significant difference (p = 0.274) in median 25OHD between males (69
nmol/L), pre- (36 nmol/L) or post-menopausal females (30 nmol/L). However, there was a
significant difference in median PTH levels (p = 0.001) between males (19.8 ng/L), pre- (39.6
ng/L) and post-menopausal (33.0 ng/L) females. Total 25OHD was measured due to its effects
on disease activity and its critical role in the maintenance of bone health, 25OHD levels could
affect the primary and secondary endpoints and so results were stratified by 25OHD
concentration; category 1 (n = 14) included levels below 25nmol/L; category 2 (n = 22) equal to or
greater than 25nmol/L. BMD, BMT and inflammatory biomarkers were plotted at baseline
and after 12 months by 25OHD category (Fig 4) We found that vitamin D deficient patients
had the lowest mean lumbar spine, femoral and forearm BMD at baseline and after 12 months
and median concentrations of bone resorption markers (βCTx, TRAP5b) increased from
baseline to 12 months, whereas in vitamin D replete patients bone resorption decreased.
Results were categorized by 25OHD status; category 1 levels up to 24.9 nmol/L; category 2
levels greater than or equal to 25 nmol/L. Bone mineral density (BMD) was plotted for each
25OHD category at lumbar spine; mean L2-L4 (LS BMD), also neck of femur (MN BMD),
total femur (MT BMD) and ultra-distal radius (MRUD BMD) mean of both sides, at time 0
before the 1st rituximab infusion and after 12 months in patients who completed the study
(n = 36). A second rituximab cycle was given at 6 months if clinically indicated (n = 32).
Additionally for each 25OHD category, median concentrations of the following biomarkers at time
0 and 12 months; β-isomerised carboxy terminal telopeptide of type I collagen (βCTx); tartrate
resistant acid phosphatase isoform 5b (TRAP5b); procollagen type 1 amino-terminal
propeptide (P1NP); bone specific alkaline phosphatase (BAP); sclerostin (SCL); dickkopf-related
9 / 15
Fig 3. The effects of gender and menopausal status on median levels of biomarkers. Serial blood samples were taken at time 0 before the first rituximab
infusion and then at 3 monthly intervals over one year, a second rituximab cycle was given if clinically indicated at 6 months. Median concentrations were
plotted at each visit (all patients n = 34; post-menopausal females n = 22; pre-menopausal females n = 6; males n = 6) for the following biomarkers (y axis):
βisomerised carboxy terminal telopeptide of type I collagen (βCTx); tartrate resistant acid phosphatase isoform 5b (TRAP5b); procollagen type 1 amino-terminal
propeptide (P1NP); bone specific alkaline phosphatase (BAP); sclerostin (SCL); dickkopf-related protein 1 (DKK-1).
10 / 15
Fig 4. The effects of vitamin D on mean bone mineral density and median levels of biomarkers.
11 / 15
protein 1 (DKK-1); C-reactive protein (CRP); erythrocyte sedimentation rate (ESR); disease
activity score using 28 tender and swollen joints (DAS28).
We prospectively studied the effects of B cell depletion using rituximab on bone density and
bone turnover in patients with refractory RA over 12 months. We found no significant change
in bone density at the lumbar spine; our primary endpoint, or ultra-distal forearm, however
we did find a significant decrease in BMD at the femoral neck (-2%) and total femur (-1.7%).
We also found a significant increase in bone formation in both P1NP and BAP, but no
significant change in bone resorption or osteocyte markers after 12 months. Nevertheless, there was
a significant reduction in inflammatory markers and disease activity, indicating that rituximab
was effective in reducing the inflammation of RA in this patient cohort.
To our knowledge this is the first longitudinal study investigating the effects of rituximab
on BMD and bone turnover markers. The fall in BMD at the hip sites was surprising given
our previously published paper demonstrating a statistically significant 37% decrease in βCTx
mirrored by a reduction in disease activity and an increase of 13% in P1NP, in sera of 46 RA
patients 6 months after a single treatment course of rituximab, however no BMD data were
collected . There were differences in the study population between the two studies.
Compared to our current study the patients in our earlier paper had lower disease activity and bone
resorption at baseline and a greater improvement in bone turnover at 6 months. The present
study included more postmenopausal females (79% compared to 58%) and notably a higher
percentage of these women were current smokers (39% compared to 18%), above the national
average quoted as 19% in 2013 [
]. Smoking may adversely influence the severity of RA [
and the clinical effectiveness of rituximab [
], additionally tobacco reportedly increases bone
resorption and affects bone mass .
The effects of tumor necrosis factor (TNF) blockers on bone have shown contrasting
results; a study of 30 RA patients treated with TNF blockers in addition to methotrexate and
prednisolone, reported an increase in LS BMD of 0.2% and 0.1% at the hip compared to a
decrease in LS BMD of -0.8% and -0.6% at the hip in a control group of 10 patients who
received only methotrexate and prednisolone [
]. Conversely; another study found a
significant decrease in femoral neck BMD alongside an increase in BTMs in 54 RA patients treated
with various TNF blockers over 6 months [
]. A recent review of the use of TNF inhibitors
highlights this discordance [
]. However, results vary by study with regard to the magnitude
of the observed change, the time points of the DXA scanning, the number and gender of
patients and the concomitant use of prednisolone and/or anti-osteoporotic drugs [3,18±22].
Post-menopausal females in our study had the lowest mean lumbar spine, femoral and
forearm BMD at baseline and after 12 months, reflecting the fact that bone loss is more rapid in
post-menopausal women, partially due to lower circulating oestradiol levels [
]. None of our
patients were treated for osteoporosis with bisphosphonates, calcitonin, strontium ranelate,
denosumab or teriparatide prior to/ or during the study, yet BMD was maintained at lumbar
spine, although there was a decrease in BMD at the femoral sites.
DKK1 and SCL are secreted osteocyte markers acting as inhibitors to the Wnt signalling
pathway through binding to low density lipoprotein receptor-related protein 5 and 6 (LRP5/6)
and hence blocking the Wnt effects on osteoblasts decreasing bone formation . SCL
inhibition in early human trials reportedly increases BMD and bone formation and decreases bone
]. However, inhibiting SCL in a human TNF-alpha transgenic mouse model
of RA has been found to accelerate joint damage and the authors suggest that SCL therefore
has a protective role in the presence of chronic TNFα±mediated inflammation . It is
12 / 15
interesting that treatment with rituximab had no significant effect on this particular BTM,
additionally we found no significant variation in DKK-1 levels over 12 months.
Fourteen of our patients (39%) were classed as vitamin D deficient (25OHD <25 nmol/L),
they had lower mean BMD at all sites at baseline and after 12 months compared to subjects
with levels above 25 nmol/L. Additionally vitamin D deficient patients had an increase in bone
resorption (median βCTx and TRAP5b levels) after 12 months, whereas patients with higher
vitamin D levels had a reduction in bone resorption. These results are in keeping with a
Chinese study which reported that serum 25OHD levels in 130 RA patients (95 women and 35
men), were lower in those with osteopenia and osteoporosis than in those with normal BMD
]. A further study reported that higher 25OHD levels may prevent the occurrence of
osteoporosis at the femoral neck but not lumbar spine [
] and one of the most recent factors found to
be associated with autoimmunity is vitamin D deficiency [
]. However the methodology and
definition of vitamin D deficiency varies widely between studies, many quoting a cut-off value
as high as 50 nmol/L. In our cohort, although post-menopausal women had the lowest median
25OHD concentration we found no significant difference in median levels between males,
pre- or post-menopausal women, so vitamin D deficiency may not explain why the
post-menopausal women lost BMD more than pre-menopausal women or men. Vitamin D status is not
normally investigated in patients presenting with RA, but our preliminary results suggest that
vitamin insufficiency can influence outcome in terms of the response to treatment.
There are some limitations to our study; it was designed as a single treatment arm trial with
no control group. The optimal design would be a double-blind randomized comparison with
placebo. Patients with active RA would be expected to have continued bone loss and abnormal
bone turnover until the disease activity had been adequately suppressed. It is therefore possible
that the rituximab could have slowed the bone loss that was occurring, but this study would be
unable to detect this without a control arm. But rituximab is an approved treatment for
refractory RA and is already known to reduce disease activity [
], such a control arm would have had
to be matched for disease activity and it would be unethical to have an untreated arm with that
level of active disease. The duration of the study was also short at only 12 months; a longer
follow-up of patients after subsequent treatment courses may have shown improvements in BMD,
the relatively small numbers limits the power to detect smaller changes in variables and although
all our patients had high disease activity and fulfilled the treatment criteria the group was
heterogeneous, consisting of men, pre and postmenopausal women and different age groups.
In conclusion, the present study revealed that in a cohort of RA patients treated by
rituximab BMD was maintained at the lumbar spine and UD radius forearm but fell at the hip sites.
The effects of rituximab on BMD seems to be influenced by vitamin D status, gender,
menopausal status, changes in disease activity and could also be confounded by the requirement for
prednisolone. A larger study powered to take into account all these factors is required and this
will necessitate that vitamin D insufficiency or deficiency be corrected from the start.
S1 Table. Study dataset.
The authors wish to thank the patients and their families as well as the research nurses, trial
coordinators and operations staff in the 10 UK centres for their contributions, and also the
investigators whose patients were enrolled in the HORUS trial.
13 / 15
HORUS trial centres and Principal Investigators: South Tees Hospitals NHS Foundation
Trust (Stephen P Tuck); Newcastle Hospitals NHS Foundation Trust (Wan-Fai Ng); City
Hospitals Sunderland NHS Foundation Trust (David Coady); The Mid Yorkshire Hospitals NHS
Trust (Lesley Hordon); Northumbria Healthcare NHS Foundation Trust (Fraser Birrell);
County Durham and Darlington NHS Foundation Trust (Matt Bridges); Gateshead Health
NHS Foundation Trust (Saravanan Vadivelu); North Tees and Hartlepool Hospitals NHS
Foundation Trust (Dipak Ray) South Warwickshire NHS Foundation Trust (Christopher
Marguerie); Mid Staffordshire NHS Foundation Trust (Thomas Sheeran).
We acknowledge Dr M.J.H de Hair (University Medical Centre Utrecht, The Netherlands)
and Dr S Mastana (Loughborough University, UK) for their statistical advice.
Conceptualization: Stephen P. Tuck, Jacob M. van Laar.
Data curation: Gillian Wheater, Mohsen Elshahaly, Kamran Naraghi.
Formal analysis: Gillian Wheater.
Funding acquisition: Stephen P. Tuck, Jacob M. van Laar.
Investigation: Gillian Wheater, Mohsen Elshahaly, Kamran Naraghi, Jacob M. van Laar.
Methodology: Gillian Wheater, Stephen P. Tuck, Jacob M. van Laar.
Project administration: Mohsen Elshahaly, Kamran Naraghi.
Resources: Gillian Wheater, Mohsen Elshahaly.
Software: Mohsen Elshahaly, Kamran Naraghi.
Supervision: Stephen P. Tuck, Harish K. Datta, Jacob M. van Laar.
Validation: Gillian Wheater, Mohsen Elshahaly.
Visualization: Gillian Wheater, Stephen P. Tuck.
Writing ± original draft: Gillian Wheater.
Writing ± review & editing: Mohsen Elshahaly, Kamran Naraghi, Stephen P. Tuck, Harish K.
Datta, Jacob M. van Laar.
14 / 15
1. Teng YKO , Levarht EW , Hashemi M , Bajema IM , Toes REM , Huizinga TWJ , et al. Immunohistochemical analysis as a means to predict responsiveness to rituximab treatment . Arthritis Rheum . 2007 ; 56 : 3909 ± 3918 . https://doi.org/10.1002/art.22967 PMID: 18050222
2. Kelly CA , Bartholomew P , Lapworth A , Basu A , Hamilton J , Heycock C . Peripheral bone density in patients with rheumatoid arthritis and factors which influence it . Eur J Intern Med . 2002 ; 13 : 423 ± 427 . PMID: 12384130
3. Marotte H , Pallot-Prades B , Grange L , Gaudin P , Alexandre C , Miossec P. A 1-year case-control study in patients with rheumatoid arthritis indicates prevention of loss of bone mineral density in both responders and nonresponders to infliximab . Arthritis Res Ther . 2007 ; 9:R61 . https://doi.org/10.1186/ar2219 PMID: 17597527
4. Franck H , Gottwalt . Peripheral bone density in patients with rheumatoid arthritis . Clin Rheumatol . 2009 ; 28 : 1141 ± 1145 . https://doi.org/10.1007/s10067-009 -1211-2 PMID: 19529880
5. Schett G . Rheumatoid arthritis: inflammation and bone loss . Wien Med Wochenschr . 2006 ; 156 ( 1 ± 2): 34 ± 41 . https://doi.org/10.1007/s10354-005 -0244-7 PMID: 16465612
6. Kostoglou-Athanassiou I , Athanassiou P , Lyraki A , Raftakis I , Antoniadis C . Vitamin D and rheumatoid arthritis . Ther Adv Endocrinol Metab . 2012 ; 3 ( 6 ): 181 ± 187 . https://doi.org/10.1177/2042018812471070 PMID: 23323190
7. Hong Q , Xu J , Xu S , Lian L , Zhang M , Ding C . Associations between serum 25-hydroxyvitamin D and disease activity, inflammatory cytokines and bone loss in patients with rheumatoid arthritis . Rheumatol . 2014 ; 53 : 1994 ± 2001 .
8. Datta HK , Ng WF , Walker JA , Tuck SP , Varanasi SS . The cell biology of bone metabolism . J Clin Path . 2008 ; 61 : 577 ± 587 . https://doi.org/10.1136/jcp. 2007 .048868 PMID: 18441154
9. Manabe N , Kawaguchi H , Chikuda H , Miyaura C , Inada M , Nagai R , et al. Connection between B lymphocyte and osteoclast differentiation pathways . J Immunol . 2001 ; 167 : 2625 ± 2631 . PMID: 11509604 10 . Wheater G , Hogan VE , Teng YKO , Tekstra J , Lafeber FP , Huizinga TWJ , et al. Suppression of bone turnover by B-cell depletion in patients with rheumatoid arthritis . Osteoporos Int . 2011 ; 12 : 3067 ± 3072 .
11. NICE public health guidance 56 (issued November 2014) Vitamin D: increasing supplement use among at-risk groups . https://www.nice. org.uk/guidance/ph56. Cited 20 March 2015 .
12. Pearce SHS , Cheetham TD . Diagnosis and management of vitamin D deficiency . BMJ. 2010 ; 340 ( 7738 ): 142 ± 147 .
13. Fransen J , van Riel PLCM. The Disease Activity Score and the EULAR response criteria . Rheum Dis Clin North Am . 2009 ; 35 ( 4 ): 745 ± 757 . https://doi.org/10.1016/j.rdc. 2009 . 10 .001 PMID: 19962619
14. Office for National Statistics (dated November 25 2014 ) Adult smoking habits in Great Britain, 2013 . http://www.ons.gov.uk/ons/dcp171778_ 386291 .pdf. Cited 20 March 2015 .
15. Saag KG , Cerhan JR , Kolluri S , Ohashi K , Hunninghake GW , Schwartz DA . Cigarette smoking and rheumatoid arthritis severity . Ann Rheum Dis . 1997 ; 56 : 463 ± 469 . PMID: 9306868
16. Khan A , Scott DL , Batley M . Smoking, rheumatoid factor status and responses to rituximab . Ann Rheum Dis. 2012 ; 71 ( 9 ): 1587 ± 1588 . https://doi.org/10.1136/annrheumdis-2012-201758 PMID: 22532635 17 . Westhoff G , Rau R , Zink A . Rheumatoid arthritis patients who smoke have a higher need for DMARDs and feel worse, but they do not have more joint damage than non-smokers of the same serological group . Rheumatol . 2008 ; 47 : 849 ± 854 .
18. Seriolo B , Paolino S , Sulli A , Ferretti V , Cutolo M . Bone Metabolism Changes during anti-TNF therapy in patients with active rheumatoid arthritis . Ann. N.Y. Acad. Sci . 2006 ; 1069 : 420 ± 427 . https://doi.org/10. 1196/annals.1351.040 PMID: 16855169
19. Orsolini G , Adami G , Adami S , Viapiana O , Idolazzi L , Gatti D , et al. Short-term effects of TNF inhibitors on bone turnover markers and bone mineral density in rheumatoid arthritis . Calcif Tissue Int . 2016 ; 98 : 580 ± 585 . https://doi.org/10.1007/s00223-016-0114 -x PMID : 26887973
20. Manara M , Sinigaglia L . Bone and TNF in rheumatoid arthritis: clinical implications . Open . 2015 ; 1 (Suppl 1):e000065.
21. Barnabe C , Hanley DA . Effect of tumor necrosis factor alpha inhibition on bone density and turnover markers in patients with rheumatoid arthritis and spondyoarthropathy . Semin Arthritis Rheum . 2008 ; 39 : 116 ± 122 . https://doi.org/10.1016/j.semarthrit. 2008 . 04 .004 PMID: 18585759
22. Vis M , Havaardsholm EA , Haugeberg G , Uhlig T , Voskuyl AE , van de Stadt RJ , et al. Evaluation of bone mineral density, bone metabolism, osteoprotegerin and receptor activator of the NFkB ligand serum levels during treatment with infliximab in patients with rheumatoid arthritis . Ann Rheum Dis . 2006 ; 65 : 1495 ± 1499 . https://doi.org/10.1136/ard. 2005 .044198 PMID: 16606653
23. Garnero P , Sornay-Rendu E , Claustrat B , Delmas PD . Biochemical markers of bone turnover, endogenous hormones and the risk of fractures in postmenopausal women: The OFELY study . J Bone Miner Res . 2000 ; 15 ( 8 ): 1526 ± 1536 . https://doi.org/10.1359/jbmr. 2000 . 15 .8.1526 PMID: 10934651 Wheater G , Elshahaly M , Tuck SP , Datta HK , van Laar JM. The clinical utility of bone marker measurements in osteoporosis . J Transl Med . 2013 ; 11 : 201 . https://doi.org/10.1186/ 1479 -5876-11-201 PMID: 23984630
25. McClung MR . Clinical utility of anti-sclerostin antibodies . Bone . 2017 ; 96 : 3±7 . https://doi.org/10.1016/j. bone. 2016 . 12 .012 PMID: 28115281
26. Lim SY , Bolster MB . Profile of romosozumab and its potential in the management of osteoporosis . Drug Des Devel Ther . 2017 ; 11 : 1221 ± 1231 . https://doi.org/10.2147/DDDT.S127568 PMID: 28458516 24 . 27. Wehmeyer C , Frank S , Beckmann D , BoÈttcher M , Cromme C , KoÈnig U , et al. Sclerostin inhibition promotes TNF-dependent inflammatory joint destruction . Sci Transl Med . 2016 ; 8 ( 330 ): 330ra35 .
28. Yoshimura N , Muraki S , Oka H , Nakamura K , Kawaguchi H , Tanaka S , et al. Serum levels of 25-hydroxyvitamin D and the occurrence of musculoskeletal diseases: a 3-year follow-up to the road study . Osteoporos Int . 2015 ; 26 : 151 ± 161 . https://doi.org/10.1007/s00198-014 -2844-9 PMID: 25138262
29. Ritterhouse LL , Crowe SR , Niewold TB , Kamen DL , Macwana SR , Roberts VC , et al. Vitamin D deficiency is associated with an increased autoimmune response in healthy individuals and in patients with systemic lupus erythematosus . Ann Rheum Dis . 2011 ; 70 : 1569 ± 1574 . https://doi.org/10.1136/ard. 2010 .148494 PMID: 21586442