Use of oral corticosteroids in the community and the prevention of secondary osteoporosis: a cross sectional study

Introduction

Oral corticosteroids have a major and often essential role in the treatment of several diseases. The cost of long term treatment in terms of side effects such as skin changes, truncal obesity, diabetes, hypertension, and osteoporosis can, however, be large. Little can be done to prevent or treat some of these adverse effects, whereas others—such as osteoporosis—may be amenable to prevention.

Osteoporosis constitutes a major public health problem through its association with fractures, the annual cost of which has been estimated at £614m for England and Wales.1 Although osteoporosis is inextricably associated with aging, any measure that alters the risk of osteoporosis has large public health implications. As established osteoporosis is difficult to treat, the aim should be prevention, including the identification and prevention of secondary causes. Steroid treatment is the commonest cause of osteoporosis and accounted for more than half the osteoporosis identified in young people in one study.2 Over 5.5 million prescriptions for systemic corticosteroids were issued by general practitioners in the United Kingdom in 1993,3 but no information exists on how these were prescribed. We carried out a community survey in Nottinghamshire to determine the prevalence of continuous use of oral steroids, the conditions for which they were prescribed, and the concurrent use of treatment to prevent osteoporosis.

Methods

The four year survey covered a population of 65 786 in eight general practices in central and southern Nottinghamshire. We initially approached practices that fulfilled two requirements—more than 5000 patients in the practice and computerised medical records—having obtained this information from the Nottingham Family Health Services Authority. We sent a letter to the 46 practices that fulfilled these requirements, inviting them to participate in the study and giving details of our computer requirements. Eight practices with computerised patient records for at least four years agreed to participate; two practices were rural, one was in the inner city, and five were in suburban Nottingham.

Patients currently registered with the practice who had received an oral corticosteroid within the previous year were identified from a computer search of practice records. Each patient's computer record was then reviewed by a medically qualified person to document the patient's details; the dose, duration, and type of oral steroid treatment; the condition for which the steroids were required; and the use of hormone replacement therapy or other bone modifying treatment during the study period. In six of the practices the computer record included the patient's history, details of all drugs prescribed over the preceding four years, and major diagnoses before the computer record; in the other two practices the computer contained a full drug record for at least three years with more limited patient information. In these two practices and in two of the six with complete records the written record was also reviewed so that we had complete data for four years from all practices. It also allowed us to determine the duration of oral steroid treatment in these four practices. Paper records were also reviewed for new patients who had computerised records for less than four years and whenever there was uncertainty with the computer record.

A history of current oral corticosteroid treatment of at least three months' duration was defined as “continuous treatment.” The median dose for all subjects was determined from individual mean doses during the four years. The diagnoses of asthma and chronic obstructive pulmonary disease were amalgamated as it was sometimes difficult to separate the two. The study was given ethical approval by Nottingham City Hospital's ethics committee and was approved by Nottingham Family Health Services Authority. Descriptive data were obtained with SPSS/PC+ 4.0 statistical package (SPSS, Chicago, Illinois).

Results

From a total population of 65 786 (52% women) covered by the eight practices, 303 patients (0.5%) were currently taking continuous oral corticosteroid treatment, of whom 65% (197) were women. The age range of those taking continuous treatment was 12-94 (median 69) years (fig 1). Of the population aged 55 years or more, 1.4% (245/17 114; 1.73% (166/9601) of women) were currently taking continuous oral steroids. The most commonly prescribed steroid was prednisolone (in 97% (294/303) of cases), with a mean dose of 8.0 mg/day (median 6.8 mg/day). The most common conditions requiring continuous oral steroid treatment were rheumatoid arthritis (23%), polymyalgia rheumatica (22%), and asthma or chronic obstructive pulmonary disease (19%) (table 1). In the four practices in which the written record was reviewed (149 patients) the median duration of continuous oral steroid treatment was three (range 0.3-37) years; 22 patients had received oral steroids for over 10 years and eight patients for over 20 years (fig 2).

Fig 1

Age-sex distribution of patients taking continuous oral corticosteroids

• Download figure
• Open in new tab
• Download powerpoint

Fig 2

Duration of oral corticosteroid treatment. Data are expressed as cumulative percentage of patients taking oral steroid treatment against time

• Download figure
• Open in new tab
• Download powerpoint

Only 41 patients (37 women) had received treatment for the prevention of osteoporosis over the past four years. Treatment comprised hormone replacement therapy (24 patients); calcium alone (7) or with vitamin D (6) or bisphosphonates (3); and vitamin D alone (1). Only 10% (18/181) of women aged over 45 years taking continuous oral steroids were currently taking hormone replacement therapy. A further 3% (6/18) had taken hormone replacement therapy during the four years but were not taking it at the time of the survey.

Discussion

This is the first reported survey of the prevalence of use of oral corticosteroids in a community population. The practices covered urban and rural areas and seemed to provide a representative sample of the local population as they were sited in areas with Townsend scores ranging from −4.5 at the affluent end of the scale to 11.6 at the deprived end, with a median value of 1.3.4 (The Townsend scores in Nottinghamshire range from −7.3 to 11.6.) Many of the patients in the survey were attending hospital outpatient clinics, and the findings therefore reflect prescribing practice in primary and secondary care.

Every effort was made to ensure that all patients were identified in the initial computer search, and in the two fully computerised practices in which computerised and written records were compared there were only minor discrepancies in patient information, such as an occasional missing record of a home visit. The computerised record usually contained more information on drug treatment as repeat prescriptions are often not reported in written records. Most prescribing for acute conditions was documented, and, although a course of steroids may occasionally have been omitted, this would be unlikely to affect the detection of continuous use of oral steroids.

The figure of at least three months was chosen as the defining duration for “continuous treatment” in order to exclude discrete courses of steroids, which rarely last more than four to six weeks. Our figures will underestimate the total use of oral steroids as some patients, particularly those with asthma and chronic obstructive pulmonary disease, had intermittent courses of oral prednisolone rather than continuous treatment. Two thirds of continuous use of oral steroids was accounted for by rheumatoid arthritis, polymyalgia rheumatica, and asthma or chronic obstructive pulmonary disease. Our data do not allow us to determine whether oral steroids were being used appropriately, although we know that 85% of the patients with asthma or chronic obstructive pulmonary disease were also taking an inhaled corticosteroid.