Article Text

Download PDFPDF

Defining conditions where long-term glucocorticoid treatment has an acceptably low level of harm to facilitate implementation of existing recommendations: viewpoints from an EULAR task force
  1. Cindy Strehl1,
  2. Johannes W J Bijlsma2,3,
  3. Maarten de Wit4,
  4. Maarten Boers3,5,
  5. Nele Caeyers6,
  6. Maurizio Cutolo7,
  7. Bhaskar Dasgupta8,
  8. William G Dixon9,
  9. Rinie Geenen10,
  10. Tom W J Huizinga11,
  11. Alison Kent12,
  12. Annette Ladefoged de Thurah13,
  13. Joachim Listing14,
  14. Xavier Mariette15,16,
  15. David W Ray17,
  16. Hans U Scherer11,
  17. Raphaèle Seror15,16,
  18. Cornelia M Spies1,
  19. Simon Tarp18,
  20. Dieter Wiek19,
  21. Kevin L Winthrop20,
  22. Frank Buttgereit1
  1. 1Department of Rheumatology and Clinical Immunology, Charité University Medicine, Berlin, Germany
  2. 2Department of Rheumatology and Clinical Immunology, University Medical Centre Utrecht, Utrecht, the Netherlands
  3. 3Amsterdam Rheumatology& Immunology Center, VU University Medical Center, Amsterdam, the Netherlands
  4. 4Medical Humanities, VU Medical Centre, Amsterdam, the Netherlands
  5. 5Department of Epidemiology & Biostatistics, VU University Medical Center, Amsterdam, the Netherlands
  6. 6Patient Research Partner, Mol, Belgium
  7. 7Department of Internal Medicine, Research Laboratory & Academic Division of Clinical Rheumatology, University of Genoa, Genova, Italy
  8. 8Southend University Hospital NHS Foundation Trust, Westcliff-on-Sea, Essex, UK
  9. 9Arthritis Research UK Centre for Epidemiology, Centre for Musculoskeletal Research, Institute of Inflammation and Repair, University of Manchester, Manchester, UK
  10. 10Department of Clinical & Health Psychology, Utrecht University, Utrecht, the Netherlands
  11. 11Department of Rheumatology, Leiden University Medical Center, Leiden, the Netherlands
  12. 12Salisbury Foundation Trust NHS Hospital, Wiltshire, UK
  13. 13Department of Rheumatology, Aarhus University Hospital, Aarhus, Denmark
  14. 14Epidemiology Unit, German Rheumatism Research Centre, Berlin, Germany
  15. 15Department of Rheumatology, Assistance Publique–Hopitaux de Paris, Hôpitaux Universitaires Paris-Sud, Le Kremlin Bicêtre, France
  16. 16Université Paris-Sud; INSERM U1184, Le Kremlin Bicêtre, France
  17. 17Faculty of Medical and Health Sciences, University of Manchester, Manchester, UK
  18. 18Department of Rheumatology, The Parker Institute, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Denmark
  19. 19Deutsche Reuma-Liga, Bonn, Germany
  20. 20Divisions of Infectious Diseases, Oregon Health & Science University, Portland, USA
  1. Correspondence to Dr Cindy Strehl, Department of Rheumatology and Clinical Immunology, Charité University Hospital, Charitéplatz 1, Berlin 10117, Germany; cindy.strehl{at}


There is convincing evidence for the known and unambiguously accepted beneficial effects of glucocorticoids at low dosages. However, the implementation of existing recommendations and guidelines on the management of glucocorticoid therapy in rheumatic diseases is lagging behind. As a first step to improve implementation, we aimed at defining conditions under which long-term glucocorticoid therapy may have an acceptably low level of harm. A multidisciplinary European League Against Rheumatism task force group of experts including patients with rheumatic diseases was assembled. After a systematic literature search, breakout groups critically reviewed the evidence on the four most worrisome adverse effects of glucocorticoid therapy (osteoporosis, hyperglycaemia/diabetes mellitus, cardiovascular diseases and infections) and presented their results to the other group members following a structured questionnaire for final discussion and consensus finding. Robust evidence on the risk of harm of long-term glucocorticoid therapy was often lacking since relevant study results were often either missing, contradictory or carried a high risk of bias. The group agreed that the risk of harm is low for the majority of patients at long-term dosages of ≤5 mg prednisone equivalent per day, whereas at dosages of >10 mg/day the risk of harm is elevated. At dosages between >5 and ≤10 mg/day, patient-specific characteristics (protective and risk factors) determine the risk of harm. The level of harm of glucocorticoids depends on both dose and patient-specific parameters. General and glucocorticoid-associated risk factors and protective factors such as a healthy lifestyle should be taken into account when evaluating the actual and future risk.

  • Corticosteroids
  • Rheumatoid Arthritis
  • Osteoporosis
  • Infections
  • Cardiovascular Disease

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Several guidelines and recommendations on the management of glucocorticoid therapy in rheumatology exist,1–6 but their implementation is hampered by persisting uncertainty about the actual benefit–risk balance. Despite convincing evidence that low-dose glucocorticoids can be beneficial in rheumatoid arthritis7–10 and other rheumatic diseases, fear of harm categorises patients and physicians as glucocorticoid ‘supporters’, ‘opponents’ or ‘neutrals’ (not favouring a pro or con position).3 ,11–16 On the background of the known beneficial effects of glucocorticoids,17–20 the aim of this Task Force was to achieve consensus on specific conditions where long-term glucocorticoid therapy has an acceptably low risk of harm in the treatment of chronic inflammatory rheumatic diseases. Achieving consensus is considered an essential first step towards a better implementation of existing recommendations among professionals and patients with rheumatic conditions.


The Task Force comprised rheumatologists, epidemiologists, an endocrinologist, an infection disease specialist, a pharmacist, health professionals, four patients with rheumatic diseases and a fellow in charge of literature research. They were assigned to four breakout groups, each including at least one supporter, opponent, ‘neutral’, health professional and patient, respectively, focusing on the four most worrisome glucocorticoid adverse effects: osteoporosis, hyperglycaemia/diabetes mellitus (DM), cardiovascular diseases (CVDs) and infections (according to ref. 21; see online supplementary table S1).

Systematic literature search

Prior to the group meeting, a systematic literature review was performed (see online supplementary text).

Expert discussion

The meeting started with an overview of the results of the search (see online supplementary figure S1 and tables S2–S6). After individual group discussion, the relevant evidence was summarised and presented to the whole group according to a structured questionnaire for final discussion and consensus (see online supplementary box S1).


Risks of long-term glucocorticoid therapy are defined by both drug-specific (dose, duration) and patient-specific characteristics

On the background of a general agreement with current guidelines stating there is convincing evidence for the beneficial effects of low-dose glucocorticoids, the Task Force reviewed and discussed the harms of glucocorticoids. Studies on harm are often of low quality: observational designs with high risk of bias (especially confounding by indication), poor documentation of glucocorticoid exposure and differing models of risk attribution. Moreover, clinical trials of glucocorticoids have often been small, short and with limited assessment of adverse events. Consequently, results can be contradictory and their interpretation is sometimes biased (supporters may trivialise and opponents may overvalue harm). There was also consensus for the term ‘long-term glucocorticoid usage’ to refer to (3-) 6 months or more, based on information on glucocorticoid-induced osteoporosis.22–24

After critical appraisal of available evidence, the group agreed on the following statements on long-term glucocorticoid therapy in chronic inflammatory rheumatic diseases (with the exception of, eg, patients with a poor or absent response to glucocorticoids in terms of both therapeutic and adverse effects due to (partial) signalling-dependent glucocorticoid resistance25) (figure 1A).

  • At ≤5 mg/day, there is an acceptably low level of harm for the specified outcomes (with the exception of patients at high risk for CVD who may require preventive measures).

  • At >10 mg/day, the risk of harm is elevated.

  • At dosages between >5 and ≤10 mg/day, uncertainty still exists and, consequently, patient-specific characteristics need particular consideration to interpret and estimate the individual risk of harm.

Figure 1

The level of harm of long-term glucocorticoid therapy in rheumatic diseases. Bearing in mind the beneficial effects of glucocorticoids, the Task Force members agreed that (A) at dosages of ≤5 mg/day prednisone equivalent, there is an acceptably low level of harm that is elevated at dosages of >10 mg/day. At dosages between >5 and ≤10 mg/day, there still exists uncertainty and therefore patient-specific characteristics (ie, disease activity, the presence of additional risk factors) need consideration when estimating the risk of harm. These patient-specific factors can shift the level of harm towards the (B) better or (C) worse. ACPA, anti-citrullinated peptide/protein antibodies; CV, cardiovascular; RF, rheumatoid factor.

The actual risk of harm is patient-specific, that is, it depends on individual risk factors and/or preventive measures

Certain factors may increase the susceptibility to specific diseases in general, even without glucocorticoid exposure. For example, elderly people have an increased incidence of adverse events, including each of the harms discussed here. Also, smoking, high alcohol consumption or bad nutrition negatively influences individual health. The person can minimise these risks of harm by adopting a healthier lifestyle (including healthy nutrition and appropriate exercise).

Long-term glucocorticoid therapy potentially represents an additional risk factor, and, therefore, the glucocorticoid-associated risk of harm is patient-specific. As a general rule, early diagnosis, low disease activity, low cumulative glucocorticoid dosage, and monitoring and treatment of additional risk factors and comorbidities reduce the glucocorticoid-associated risk of harm26–30 (figure 1B). Comprehensive information on glucocorticoid treatment is essential, and risks should be discussed with patients (and their family and/or carers, including healthcare professionals) before glucocorticoid therapy is started.1 ,2

Risk factors for osteoporosis

Glucocorticoid treatment can lead to rapid loss of bone mineral density and bone quality, and to an increased fracture risk.26 Elderly patients receiving long-term glucocorticoid therapy have an increased risk for osteoporosis.23 ,31 Additional risk factors for osteoporosis include female sex, low body weight, low bone mineral density, prevalent fractures and a family history of osteoporosis (figure 1C).23 ,26 ,32–37 Patients can reduce their risk, for example, by appropriate exercise; glucocorticoid therapy may also support this by reducing disease activity, allowing patients to become more physically active. Guidelines also recommend sufficient vitamin D/calcium intake and treatment with bisphosphonates, osteoanabolic drugs or selective oestrogen receptor modulators (on indication)36 ,38–42 (figure 1B).

Risk factors for infections

In the general population, age (>60 years) and male gender increase the risk of infection, and these may add to the risk conferred by glucocorticoids,43–45 with the current glucocorticoid dose being more risk-relevant than former usage or cumulative dose.45–47 Comorbidities such as chronic lung, heart, renal and certain neurological diseases, peripheral vascular diseases, DM, hepatitis C and leucopenia probably confer additive risks.28 ,47–51 Also, patients with a history of prior serious infections are at increased risk. Screening for and/or vaccination against, for example, influenza, pneumococci or herpes zoster, is proposed.50 ,52 Risk scores are useful to estimate potential risks and support clinical decision-making before the start of therapy53 (figure 1B).

Risk factors for hyperglycaemia/DM

A genetic disposition, age and obesity as well as chronic inflammation are risk factors for the development of hyperglycaemia/DM, and glucocorticoids represent an additional factor, sometimes interacting with chronic inflammation54–57 (figure 1C). Weight loss for obese patients, healthy diet, appropriate exercise and hydroxychloroquine as a therapeutic measure reduce the DM risk58 ,59 (figure 1B).

Risk factors for glucocorticoid-induced CVD

Age, male sex, obesity, hypertension, diabetes and dyslipidaemia increase the risk for CVD. Glucocorticoid therapy may have an additional negative impact on the cardiovascular system.29 ,60–71 Patients with active disease, extra-articular manifestations and positive rheumatoid factors and/or antibodies to citrullinated proteins have an increased risk regardless of therapy.29 ,69 ,72–74 Patients with high disease activity or functional impairment both have an increase in cardiovascular mortality and are more likely to be treated with glucocorticoids.29 ,61 ,68–70 ,73 ,74 Thus, glucocorticoid association with CVD might partly be explained by confounding by indication.

Preventive measures include healthy diet (low in saturated fat and calories), appropriate physical activity, sodium restriction and cessation of smoking.60 ,70 ,75–78 Preventive use of statins or ACE inhibitors or other antihypertensive drugs has been recommended on indication2 ,60 ,77 (figure 1B).


This Task Force concluded that the level of harm conveyed by long-term glucocorticoid therapy is dose-dependent with dosages of ≤5 mg/day producing an acceptably low level of harm (with the exception of patients at high CV risk), whereas at >10 mg/day the risk is elevated. There was consensus that patient-specific parameters clearly modify the actual risk of harm and that these need consideration when evaluating actual and future benefit–risk balance of long-term glucocorticoid therapy, especially at dosages between >5 and ≤10 mg/day.

High-quality data on glucocorticoid harm are limited, and study results are often contradictory, have a high risk of bias or are evaluated inconsistently. Heterogeneous study designs (differences in glucocorticoid dosages, treatment periods, patient characteristics and total durations of therapy including former glucocorticoid usages) often lead to incomparable results and therefore sometimes contradictory conclusions. For example, a recent publication revealed that observational studies but not randomised controlled trials suggest an increased infection risk associated with glucocorticoid therapy.46

A specific conundrum refers to the ‘Janus-head-like behaviour’ of glucocorticoids39 ,56 ,79: the pro-inflammatory state of rheumatoid arthritis itself induces insulin resistance (glucocorticoids reduce the disease activity, thus counteracting this negative impact of the disease on glucose metabolism); however, glucocorticoids intrinsically also induce insulin resistance.56 Similar scenarios exist for osteoporosis and effects on the cardiovascular system, making the impact of long-term glucocorticoid therapy a complicated study subject since exact quantification of benefits and harms is not possible with currently existing data. Consequently, patients who are uncertain about the glucocorticoid-associated risks should be thoroughly informed that an untreated disease also negatively influences other body functions. Another need is to implement existing recommendations on study design, performance, data analysis and publication, to enable future high-quality data synthesis.3 Clearly, future study designs need improvement with regard to documentation, reproducibility and comparability (box 1).

Box 1

Points that should be addressed in future studies

In order to improve future studies, it is necessary to unify study design and outcome parameters.


Gapless documentation of exact glucocorticoid dosages (including start and end date of treatment, date specific changes in dosages), day time when glucocorticoid dose is administered (eg, glucose metabolism is influenced by the circadian clock)


Glucocorticoid dosage ranges need to be reconciled


Screening of published data (eg, for adaptation of dosage ranges), analysis of unpublished data with the help of the corresponding authors

Risk scores:

Improvement, adaptation and usage of risk scores44 ,48 ,53


Standardisation of questionnaires and glucocorticoid-specific core outcome parameter for all clinical studies3

In conclusion, the actual level of harm of long-term glucocorticoid treatment represents both a drug-specific (dose, duration) and patient-specific condition. Given the known beneficial effects of glucocorticoids, the consensus on the risks of long-term treatment achieved here represents a first step towards a better implementation of existing recommendations on glucocorticoid therapy in rheumatology. The next, ongoing step of this European League Against Rheumatism Task Force is to develop modern tools, for example, education material for patients and professionals, to achieve this goal.


View Abstract


  • Handling editor Tore K Kvien

  • Twitter Follow William Dixon at @WGDixon

  • Contributors We declare that all authors included on this paper fulfil the criteria of authorship.

  • Funding The activities of this task force are financially supported by EULAR (project number CLI073).

  • Competing interests JWJB reports personal fees from Mundipharma International Ltd, Horizon Pharma and Sun. MdW received consultancy fees, honoraria, travel expenses or grant support from AbbVie, Bristol-Myers Squibb GmbH & Co. KGaA, Eli-Lilly and Company, Novartis and Roche Pharma AG. MB received consultancy fees from Mundipharma International Ltd. MC received honoraria from Horizon Pharma and Mundipharma International Ltd, and grant research support from Horizon Pharma. BD reports personal fees from GSK, Servier, Mundipharma International Ltd and Sobi and grants from Napp. JL received consultancy fees from Horizon Pharma and Pfizer. HUS reports personal fees from Pfizer and Actelion and FB received consultancy fees, honoraria and travel expenses from Horizon Pharma, Pfizer and Mundipharma International Ltd, and grant support from Horizon Pharma.

  • Provenance and peer review Not commissioned; externally peer reviewed.