There is little information on oral glucocorticoid use in the general US population. Previously, there have been published estimates of glucocorticoid use in countries outside of the US. This study aimed to estimate the prevalence of glucocorticoid use, duration of use, and concomitant use of antiosteoporosis pharmaceuticals in the US population age ≥20 years.
Data from 5 cycles (1999–2008) of the National Health and Nutrition Examination Survey (NHANES) were used to provide nationally representative weighted estimates. Oral glucocorticoids and concomitant use of antiosteoporosis pharmaceuticals (bisphosphonates, calcitonin, calcium, hormone replacement therapies, teriparatide, and vitamin D) were analyzed.
There were 356 NHANES respondents ages ≥20 years who reported use of an oral glucocorticoid in the combined cycles between 1999 and 2008. The weighted prevalence of oral glucocorticoid use was 1.2% (95% confidence interval [95% CI] 1.1–1.4) from 1999–2008, corresponding to 2,513,259 persons in the US. The mean duration of oral glucocorticoid use was 1,605.7 days (95% CI 1,261.2–1,950.1), and 28.8% (95% CI 22.2–35.4) of oral glucocorticoid users reported use for ≥5 years. Concomitant use of a bisphosphonate was reported by 8.6% (95% CI 5.1–11.7) of oral glucocorticoid users, and 37.9% (95% CI 31.7–44.0) reported usage of any antiosteoporosis pharmaceutical.
Based on NHANES data from 1999–2008, it is estimated that the prevalence of glucocorticoid use in the US is 1.2%, with a long duration of use and infrequent use of antiosteoporotic medications compared to other estimates.
Oral glucocorticoids are a class of pharmaceutical that saves lives, but has significant side effects. Studies have shown that both short- and long-term use of oral glucocorticoids leads to bone loss and an increased risk of fracture (1, 2). These fractures can lead to increased pain, decreased quality of life, and in some cases, increased mortality (3). There are antiosteoporosis pharmaceuticals that have been shown to be efficacious and are recommended by clinical societies (4, 5), but studies of long-term oral glucocorticoid users have indicated their use in less than 40% of the population (6, 7).
Four peer-reviewed publications were found that estimated the prevalence of oral glucocorticoid use in a general population. Walsh and colleagues estimated oral glucocorticoid use in 0.5% of the UK population (8). Van Staa et al reported that 0.9% of the UK population age ≥20 years took an oral glucocorticoid between 1994 and 1997, with the highest age-stratified rate of 2.5% among persons ages 70–79 years (9). Gudbjornsson and colleagues reported 0.7% oral glucocorticoid use in Northeast Iceland between 1995 and 1996 (10). Most recently, Fardet et al reported 0.85% use in the UK population age ≥18 years between 1989 and 2008, with the highest sex-/age-stratified rate of 3.05% among women ages 80–89 years (11).
Previous studies related to oral glucocorticoid and antiosteoporosis pharmaceutical use in the US were limited to long-term oral glucocorticoid users only (6, 7). Their estimates provide an incomplete view of oral glucocorticoid use in the US general population and prevent a fair comparison of epidemiologic estimates between the US population and those in other countries. Therefore, this study aimed to estimate the prevalence and treatment duration for oral glucocorticoid use in the US, as well as concomitant use of antiosteoporosis pharmaceuticals by oral glucocorticoid users. To create a comparable estimate, the 1999–2008 US National Health and Nutrition Examination Survey (NHANES) was analyzed for all persons ages ≥20 years.
Significance & Innovations
It is estimated that 1.2% of the US population age ≥20 years (∼2,513,259 persons) received an oral glucocorticoid at any time point between 1999 and 2008.
The prevalence of oral glucocorticoid use in the US is greater than published estimates in other countries.
This is the first nationally representative estimate of oral glucocorticoid usage in the adult US population.
Materials and methods
The NHANES has been conducted as a continuous cross-sectional survey by the National Center for Health Statistics (NCHS) in a representative sample of the noninstitutionalized US population on 2-year cycles since 1999. Using a complex multistage stratified survey design, the NHANES aims to collect health and nutrition data that accurately represent the US population. The NHANES interview, examination, consent, and collection procedures were documented by the NCHS (12). Data from the NHANES previously have been published examining medication usage in the US (13). The publicly available data files of 5 NHANES survey cycles (1999–2000, 2001–2002, 2003–2004, 2005–2006, and 2007–2008) were combined to create the study population (14).
The study population consisted of the noninstitutionalized general US population age ≥20 years between 1999 and 2008. There were 35,209 persons who were eligible for the 1999–2008 NHANES and age ≥20 years. Interviews were conducted with 26,248 (74.5%) of these persons.
Medication usage was determined by self-report of prescription usage within the 30 days prior to the interview. Medications reported in the NHANES were converted to their generic names by the NCHS and related to the 2007 Multum Lexicon Drug Database (Cerner Multum), which provides an ingredient therapeutic category identification variable within the NHANES data releases (14). Glucocorticoids in the “glucocorticoid” Lexicon category that are administered orally were collected for this analysis. The duration of oral glucocorticoid treatment is the self-reported duration of treatment reported during the interview. The generic glucocorticoids used in this analysis are cortisone, dexamethasone, hydrocortisone, methylprednisolone, prednisolone, and prednisone.
Use of an antiresorptive agent (bisphosphonates, calcitonin, hormone replacement therapies [HRTs], teriparatide, calcium, or vitamin D) was also recorded for this analysis. Bisphosphonate and HRT use was determined using the appropriate Lexicon category. Calcitonin and teriparatide did not fall under an appropriate Lexicon category, and use was recorded by report of a medication with the corresponding generic name. Calcium and vitamin D use was determined by report of a supplement.
All data analysis was performed using PC-SAS, version 9.2 (SAS Institute). Data analysis was conducted using SAS survey procedures that account for the clusters, sample design variables, sample weighting, strata, and variances of the NHANES complex survey design using Taylor series linearization. Sample weights used in the analysis were created from the combination of interview weights from all 5 survey cycles following the NHANES analytical guidelines for combining multiple survey cycles (14). The use of variables provided by the NHANES (clusters and strata) as well as user-created variables (sample weights and sample design variables) allows the individual responses of the survey to be applicable to the general population. Prevalence estimates are presented as both the weighted prevalence from the combined study period (1999–2008) as well as the period prevalence for each of the survey cycles. Population estimates were created by multiplying the midyear Current Population Survey (15) totals for the years under study by the associated weighted prevalence rate. All data are reported as the percentage (95% confidence interval [95% CI]) unless otherwise specified. Differences in prevalence rates between survey cycles were analyzed with Wald's F test, with a P value of 0.05 or less indicating significance. This study was approved by the Cleveland Clinic Foundation Institutional Review Board.
Of the 26,248 persons ages ≥20 years and interviewed, 356 respondents reported current use of an oral glucocorticoid and, when weighted, represent 1.2% (95% CI 1.1–1.4) of the US general population, corresponding to 2,513,259 persons. Table 1 shows weighted epidemiologic estimates of oral glucocorticoid use by sex and age. When stratified by age, the highest prevalence rate (2.9%; 95% CI 2.2–3.5) was found in respondents ages ≥80 years. Women (53.3%; 95% CI 47.2–59.4) represented a larger proportion of oral glucocorticoid users than men.
Table 1. Prevalence of oral glucocorticoid usage in the US from the 1999–2008 National Health and Nutrition Examination Survey*
When stratified by sex, oral glucocorticoid use was reported by 1.3% (95% CI 1.0–1.5) of women and 1.2% (95% CI 1.0–1.4) of men. When stratified by sex and age, the prevalence rates are between 0.5% and 3.5%. Greatest use by sex and age was reported in men ages ≥80 years (3.5%; 95% CI 2.3–4.7) and women ages 70–79 years (2.7%; 95% CI 1.7–3.7).
The mean duration of oral glucocorticoid use was 1,605.7 days (95% CI 1,261.2–1,950.1), excluding 6 respondents who did not report a length of oral glucocorticoid use. Prednisone (76.6%) was the oral glucocorticoid most commonly reported, followed by methylprednisolone (9.1%) and hydrocortisone (6.7%). Of those taking oral glucocorticoids, 65.2% (95% CI 59.5–70.8) reported glucocorticoid usage ≥90 days, 42.2% reported ≥2 years, and 28.8% (95% CI 22.2–35.4) reported ≥5 years. Oral glucocorticoid users reported concomitant use of a bisphosphonate (8.6%), HRT (5.9%), calcium (22.7%), vitamin D (18.5%), or any antiosteoporosis pharmaceutical intervention (37.9%), but did not report any use of either calcitonin or teriparatide. When stratified by sex, women reported greater concomitant use of any and all antiosteoporosis pharmaceutical interventions. Prevalence rates of oral glucocorticoid and concomitant antiosteoporosis pharmaceutical use stratified by survey cycle are shown in Figure 1. Oral glucocorticoid use was greatest in 1999–2000 and concomitant use of any antiosteoporosis pharmaceutical was greatest in 2003–2004. Although usage trends for oral glucocorticoids appear to trend downward throughout the study period and for antiosteoporosis pharmaceuticals following the 2003–2004 survey, the differences between the survey cycles were not found to be statistically significant.
Our results indicate that oral glucocorticoids were likely used by the US general population age ≥20 years at a higher rate than in other countries based on previously published non-US estimates (8–11). Overall, our estimated prevalence of 1.2% is larger than previous non-US estimates, including the largest prevalence estimate of 0.9% in the UK (9). Our greatest US estimates are larger than previous non-US estimates when stratified by age (2.9% in the US versus 2.5% in the UK) (9), as well as when stratified by age and sex (3.5% in the US versus 3.05% in the UK) (11). These findings suggest that oral glucocorticoids were likely used by the US general population at a higher rate than in other countries.
For oral glucocorticoid users in the population, the mean duration of oral glucocorticoid use was estimated to be greater than 4 years, with 65.2% receiving glucocorticoid therapy for ≥90 days. Despite the long duration of oral glucocorticoid use, concomitant use of any antiosteoporosis pharmaceutical was reported by less than 40% of oral glucocorticoid users. This percentage of concomitant use is similar to previous US estimates of concomitant use of antiosteoporosis pharmaceuticals by long-term oral glucocorticoid users (6, 7). Bisphosphonates were the only medications with a Food and Drug Administration–approved label indication for glucocorticoid-induced osteoporosis prevention and treatment prior to 2008, but bisphosphonates were underutilized among the US glucocorticoid users during 1999–2008 (8.6% for both sexes and 4.8% in men). Such underutilization of bisphosphonates likely led to suboptimal interventions to reduce oral glucocorticoid–related bone loss and fracture. These findings suggest a need for additional attention to appropriate concomitant prescribing of therapies to reduce bone loss and fracture related to oral glucocorticoid use.
There was a trend toward lower utilization of oral glucocorticoids for the entire study period and following the 2003–2004 survey cycle in antiosteoporosis pharmaceuticals, although neither trend was significant. This finding is positive for oral glucocorticoid use, but troubling for antiosteoporosis pharmaceutical use, since even with new antiosteoporosis pharmaceuticals introduced during the study period, overall utilization did not increase. The possible decline in antiosteoporosis pharmaceuticals by oral glucocorticoid users would be expected to increase the risk of bone loss and fracture in the US population.
There are several limitations in this study. The NHANES does not collect information for medication doses, and has not released the self-reported reason for oral glucocorticoid use for the period 1999–2008. The NHANES does not collect information about diagnosis associated with oral glucocorticoid use. All information from the NHANES used in this study relied on self-report (e.g., medication and vitamin use in the preceding 30 days and the duration of treatment), indicating that our estimates may be affected by recall bias. The NHANES does not capture pharmaceutical use outside of the 30 days preceding the interview, which may result in an underestimation of short-term oral glucocorticoid use due to cessation before the interview. However, the NHANES was one of the best data sources available to estimate epidemiologic information in the US general population, and the impact of this underestimation and bias on our findings is likely minimal. The NHANES as a cross-sectional survey provides a nationally representative sample of the noninstitutionalized US population, allowing this study for the first time to estimate usage of oral glucocorticoid and antiosteoporosis pharmaceutical interventions in the entire US population, instead of being limited to persons with insurance or within a geographic location. Our study is also the first peer-reviewed article to estimate the period prevalence of oral glucocorticoids in the US, which can be compared to US or non-US estimates. Although we may underestimate the true prevalence for the US, our estimates are greater than previously published non-US estimates. This finding indicates that the risk of bone loss and fracture related to oral glucocorticoid use in the US is likely to be greater than previously known.
In summary, an estimated 1.2% of the US general population age ≥20 years was taking an oral glucocorticoid at any point between 1999 and 2008. This finding represents a greater prevalence compared to previous non- US estimates and reports. Many US oral glucocorticoid users reported long-term usage, yet antiosteoporosis pharmaceuticals were underutilized. The rate of oral glucocorticoid use in the US may indicate a greater risk for oral glucocorticoid–related bone loss and fracture in the US than in foreign populations, demonstrating a need for further research into oral glucocorticoid as well as antiosteoporosis pharmaceutical use in the population.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Yeh had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Overman, Yeh.
Acquisition of data. Overman.
Analysis and interpretation of data. Overman, Yeh, Deal.