Journals Library

Incidence of osteoporosis and fracture in common inflammatory diseases in primary care

Project title

Incidence of osteoporosis and fracture in common inflammatory diseases in primary care

Project reference


Final report date

11 December 2017

Project start date

01 October 2015

Project end date

30 September 2016

Project duration

12 months

Project keywords

Fracture, polymyalgia rheumatic, gout, giant cell arteritis

Lead investigator(s)

Dr Zoe Paskins, Senior Lecturer and Honorary Consultant Rheumatologist  Keele University and Haywood Rheumatology Centre

  • Rebecca Whittle, Research Associate

  • Alyshah Abdul Sultan, Research Fellow

  • Sara Muller, Senior Research Fellow

  • Milica Bucknall, Lecturer in Statistics

  • Toby Helliwell, NIHR Academic Clinical Lecturer in General Practice

  • Samantha Hider, Reader and Honorary Consultant Rheumatologist

  • Edward Roddy, Reader and Honorary Consultant Rheumatologist

  • Christian Mallen, NIHR Research Professor in General Practice

  • Jon Packham, Consultant Rheumatologist


Project objectives

The primary aim of this study is to estimate the effect of inflammatory conditions (Rheumatoid Arthritis, Ankylosing Spondylitis, Systemic Lupus Erthematous, Inflammatory Bowel Disease, gout, Polymyalgia Rheumatica & psoriasis) on risk of osteoporosis and fragility fracture.

  1. Secondary aims:

    1. To investigate the effect of treatment as an effect modifier e.g. whether and to what extent use of glucocorticoids alters this risk in each condition; methotrexate alters risk in RA; allopurinol alters risk in gout
    2. To investigate whether and to what extend other common drugs such as proton pump inhibitors, alters this risk in each condition.

Changes to project objectives

We included a further inflammatory condition: giant cell arteritis, which is closely related to polymyalgia rheumatica.

We amended the title to Risk of fragility fracture across inflammatory conditions: a UK population study, removing the term osteoporosis. This is because the diagnosis of osteoporosis is subject to both poor recording and surveillance bias.

Due to recent publications relating to some of our conditions of interest, we focused our analyses on the four conditions where a risk of fragility fracture was not previously reported or clear cut (polymyalgia rheumatica, giant cell arteritis, gout and psoriasis).

Brief summary

Materials and Method

We used data from the Clinical Practice Research Datalink (CPRD), a large UK primary care medical record database of anonymised patients that covers more than 6.9% of the UK population and is representative in terms of age and sex distribution. Practices contributing data to CPRD receive training on recording information and the database is subjected to quality checks, with data from a practice only being used when it has reached a certain standard of quality, defined as being up-to-standard (UTS).

An exposed population was defined as patients aged over 40 years with an incident (new onset) diagnosis (Read code) of one of the eight conditions of interest (Rheumatoid arthritis, Systemic Lupus Erythematous, ankylosing spondylitis, gout, psoriasis, inflammatory bowel disease, polymyalgia rheumatic and giant cell arteritis) between 1990 and 2004. Each patient was assigned an index date corresponding to the date of their disease diagnosis.

Within each disease cohort, for each exposed patient, up to four controls were randomly selected matched on age (three year age bands), gender and general practice, with the controls’ index date anchored to that of their matched exposed patient. Controls were defined as those without psoriasis and without other common inflammatory musculoskeletal conditions prior to their index date. 

The study end date was defined as the earliest date of: the patient’s death; date the patient transferred out of the practice; date of last data collection from that practice; 31st August 2015 and date of first fracture.

The event of interest was time from the index date until the first diagnosis of fracture. Read codes for fractures at sites of major osteoporotic fracture were selected (vertebrae, humerus, wrist and hip) in addition to Read codes for fragility fractures of unspecified site. Patients with the following were excluded: Read code for fracture (as previously defined) prior to their index date; Read code for fracture in the first six months of their registration with the practice (assumed to be prevalent cases); less than 12 months of UTS follow-up prior to index date and less than 3 years of UTS follow-up after index date.

We extracted information on patient demographics (age and gender) at their index date, on lifestyle-related characteristics (body mass index (BMI), smoking status and alcohol consumption) using the measurement nearest to their index date (ever prior to index and up to 1 year after) and on comorbidities (using the Charlson comorbidity index(14)) and the prescription of medications (corticosteroids, methotrexate, allopurinol, bisphosphonates and proton-pump inhibitors (PPI)) prior to the outcome for both exposed and non-exposed. Those with missing information on BMI, smoking and alcohol were included in separate categories.

Incidence rates were expressed as the number of incident fractures per 10,000 person-years. Cox proportional hazards models were used to obtain estimates of hazard ratios with 95% confidence intervals to assess the association between psoriasis exposure and time to first fracture, based on robust standard errors to account for matching. Unadjusted estimates were obtained, followed by adjustment for confounding factors which affected estimates by >10% (to avoid over-adjustment) from age, gender, BMI, alcohol, smoking, Charlson comorbidity index, bisphosphonate, glucocorticoid and PPI use. The proportional hazards assumption was tested using Schoenfeld residuals. Subgroup analyses were conducted by fracture site, disease severity, age and gender. The effect of methotrexate use on fracture risk within psoriasis and PMR/GCA patients was evaluated by estimating hazard ratios within the exposed group, comparing patients with any methotrexate use to those with none. Among patients with gout, we assessed the impact of ULT on fracture and utilised landmark analysis and propensity score matching to account for immortal time bias and confounding by indication.

Further sensitivity analyses were conducted for the gout, PMR/GCA and psoriasis cohort. All analyses were performed using Stata/MP 14.2 (Stata Corporation, TX. USA). This study was approved by the Independent Scientific Advisory Committee of CPRD (protocol 15_165RA).


Our initial analysis showed the absolute risk of fracture varied between 0.056 (for gout) and 0.139 (for PMR/ GCA). Five inflammatory conditions were associated with increased risk of fragility fracture: RA, GCA, PMR, IBD and psoriasis.


12,136 and 2,673 cases of PMR and GCA respectively were identified. The incidence rate of fracture was 148.05(95% CI 141.16 to 155.28) in PMR and 147.15(132.91 to 162.91) in GCA per 10,000 person-years. Risk of fracture was increased by 63% in PMR (adjusted hazard ratio (aHR) 1.63(95%CI 1.54 to 1.73)) and 67% in GCA (1.67; 1.49 to 1.88) compared to the control populations. Fewer than 13% of glucocorticoid treated cases were prescribed bisphosphonates. There were no significant differences in risk of fracture between those prescribed, or not prescribed methotrexate in either group (PMR: aHR=0.96 (95%CI 0.75,1.21), GCA: 1.36 (0.84,2.19))


We identified 31,781 patients with incident gout matched to 122,961 controls. The absolute risk of fracture was similar in both cases and controls (AR=53 and 55 per 10,000 person-years respectively) corresponding to the HR of 0.97 (95%CI;0.92-1.02). Our finding remained unchanged when we stratified our analysis by age and gender. We did not observe statistically significant differences in the risk of fracture among those prescribed ULT within 1 and 3 years after gout diagnosis.

Late onset Psoriasis

24,219 patients with late onset psoriasis and 94,820 controls were identified. The absolute rate of fracture in psoriasis patients was 58 per 10,000 person-years (95% confidence interval (CI): 55, 61) and 53 per 10,000 person-years in the matched controls ((CI): 52, 54). Psoriasis patients had a 10% increased risk of fracture compared to their matched controls (unadjusted hazard ratio (HR) = 1.10; 95% CI: 1.04, 1.16).


This study reports for the first time, a similar increase in fracture risk for patients with PMR and GCA. More needs to be done to improve adherence to guidelines to co-prescribe bisphosphonates. Further research needs to identify whether lower glucocorticoid starting doses and/or aggressive dose reduction reduces fracture risk.

We found no excess risk of fragility fractures among patients with gout. Our findings remained consistent when we stratified our analysis by age, gender and fracture site. Our propensity score matched landmark analyses showed that those prescribed at least 6 months of ULT within one and three years of their initial gout diagnosis had neither beneficial nor adverse effects on long term risk of fracture. These findings should be reassuring to patients, healthcare commissioners and clinicians.

Identifying additional clinical factors that are independently associated with an increased fracture risk is an important step in improving fracture risk stratification. This study suggests that further work is needed to determine the relationship between age of onset of psoriasis and fracture risk, explore causative explanations and identify if existing fracture risk stratification tools underestimate fracture risk in patients with psoriasis.

Plain English summary

A number of inflammatory conditions such as rheumatoid arthritis (RA) and ankylosing spondylitis (AS) are associated with an increased risk of breaking a bone and/or osteoporosis (fragile bones). This is thought to be because inflammation reduces bone strength meaning that bones break more easily than normal. Patients at risk should have a full fracture risk assessment and/or medication which prevents fractures. As such it is important to identify conditions that are associated with the highest risk of fracture.

For some inflammatory conditions, such as RA, AS, inflammatory bowel disease (IBD) and systemic lupus erythematous (SLE), medical treatment has improved substantially over the last few years meaning that fracture risk may have reduced. For other, more common inflammatory conditions, such as psoriasis, polymyalgia rheumatica (PMR), giant cell arteritis (GCA) and gout, when we started this study, it was unclear if these conditions were associated with a higher risk of fracture or not.

The aim of this study was therefore to compare the risk of fracture across a number of inflammatory conditions. This study used data from a large database of anonymised GP records to see how frequently fragility fractures occur in patients with these 8 different inflammatory conditions.

We found that fractures were more common in all the conditions we studying, with the exception of AS, SLE and gout. The level of risk in conditions such as RA was similar to that found in other previous studies. We focused our detailed analysis on the conditions where no previous fracture risk had been shown. Over a ten-year period, 14% of people with PMR and GCA had a fracture. We found that people with these conditions were between 63 and 67% more likely to have a fracture than similar people without these conditions. Less than 20% of people with this condition were on medicines to reduce the risk of fracture, despite national guidelines stating all patients with GCA should receive these. In patients with late-onset psoriasis (occurring over the age of 40) we found 6.5% had a fracture over 10 years. These patients were 10% more likely to have a fracture than similar people without psoriasis. We will use these results to raise awareness of the need for medication to reduce fracture risk in PMR and GCA, and to highlight the need for screening (of fracture risk) in patients with psoriasis.  This in turn will hopefully prevent fractures in these patients.


Accepted articles

Sultan AA, Whittle R, Muller S,Roddy E,Mallen CD,Bucknall M,Helliwell T, Hider S,Paskins Z , Risk of fragility fracture among patients with gout and the effect of urate-lowering therapy

(Undergoing 2nd round of revisions (pending final publication decision) Canadian Medical Association Journal, Risk of fragility fracture among patients with gout and the impact of urate lowering therapy Alyshah Abdul Sultan1 PhD, Rebecca Whittle1 MSc, Sara Muller1 PhD, Edward Roddy1,2 PhD, Christian D Mallen1 PhD, Milica Bucknall1 PhD, Toby Helliwell1 PhD, Samantha Hider1,2 PhD,  Zoe Paskins1,2 PhD).

Planned articles

To be submitted to Canadian Medical Association Journal shortly

Risk of fragility fracture among patients with late-onset psoriasis: a United Kingdom population-based study

Zoe Paskins PhD1,2, Rebecca Whittle MSc1, Alyshah Abdul Sultan PhD1, Sara Muller PhD1, Milica Blagojevic-Bucknall PhD1, Toby Helliwell PhD1, Jon Packham MD1,2, Samantha Hider PhD1,2, Edward Roddy DM1,2, Christian Mallen PhD1

Public involvement

PPI was engaged at the start and end of this study.

The PPI group confirmed the need for this study and gave advice about how best to describe risk in lay terms and gave advice on dissemination strategies. In particular, they felt patients needed to know that bisphosphonates were recommended for all with GCA, and this should be added to patient leaflets. We are currently exploring how to take this forward with our implementation unit. 


An ongoing research study led by co-investigator SH is investigating primary care management and screening of co-morbidities, including osteoporosis (and fracture risk) in inflammatory conditions (INCLUDE).

Implementation of guidelines for bone sparing agents needs to be addressed in PMR and GCA.

Future guidelines for Psoriasis may include attention to fracture risk.

This project was funded by the National Institute for Health Research School for Primary Care Research (project number 256)

Department of Health Disclaimer

The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR School for Primary Care Research, NIHR, NHS or the Department of Health.