Our results show a dramatic increase in the number of overall hospitalizations of sarcoidosis patients in the U.S. between 1998 and 2008. We found the highest increase in the rate of hospitalizations among African Americans, females, older patients, and those with insurance other than Medicaid. However, for those admitted with sarcoidosis listed as the primary diagnosis, the incidence was stable during the study period with no significant differences in trends between races, gender, or insurance status. Thus, it appears that this increasing trend is driven by admissions for people with sarcoidosis, rather than because of sarcoidosis.
Although sarcoidosis was first described over 100 years ago, the cause is still unknown, treatment is still controversial, and epidemiologic data are difficult to collect due to the rarity of the disease and diverse clinical presentations. Our findings are important as there are few published data on the burden of hospitalizations in this disease in the past ten years. Our results clearly demonstrate that there is an increasing burden of inpatient medical care required by sarcoidosis patients, and that this trend is greater in older patients, women, and African Americans.
With respect to race, gender, and age, our results are similar to those published by Foreman et al who reported finding hospitalizations more frequently among females, older patients, and African Americans . Sarcoidosis is generally more common in African Americans and women [16–18]. Also, African Americans and women tend to have more severe chronic disease and extrapulmonary involvement [3, 4]. Differences in ascertainment and access to preventative and primary care may also be playing a role in this case of disparity of outcomes among African Americans . Further, other investigators have reported that socioeconomic status plays a significant role in medical care and outcomes for patients with sarcoidosis . In previous studies of sarcoidosis patients, less access to medical care has been associated with depression, and lower family income and lack of private or Medicare insurance is associated with increased severity of disease [20, 21].
In contrast to previous work, however, we found that patients with insurance other than Medicaid actually had a higher increase in the rate of hospitalization compared to Medicaid patients. We speculate that this difference may be due to differential practice patterns or access to care/travel issues between patients with Medicaid and those with alternative insurance. For instance, those with insurance may be receiving more aggressive immunosuppression with corticosteroids, potentially more use of costly immunosuppressive alternatives to prednisone (with increased risky side effects), more diagnostic procedures, or may be able to more easily access care (whether geographically or financially), thereby influencing treatment and monitoring decisions by practitioners. Advancing radiographic imaging technology has also made it easier to find potential sarcoidosis cases and identify more subtle lung findings, which may have in turn led to more aggressive treatment approaches than in times of monitoring only with chest x-rays. Thus, the use of more sensitive imaging in insured populations may contribute to more aggressive immunosuppression.
Our hypothesis that the increase in hospitalizations is associated with more aggressive immunosuppressive treatments is further supported by the observation that hospitalizations with a primary sarcoidosis diagnosis are not increasing; increasing admission rates due to primary diagnoses other than sarcoidosis suggest that factors other than sarcoidosis itself are driving this trend. In addition, a recent retrospective study of healthcare utilization showed that sarcoidosis patients with greater corticosteroid use had a greater number of unscheduled emergency department visits . Further support to our hypothesis is provided by the increasing body of literature showing that the widespread introduction of steroid-sparing immunosuppressive medications increases hospitalizations from other diseases, especially inflammatory bowel disease [23, 24].
Because other studies have suggested a seasonal peak of diagnosis of sarcoidosis cases in the early spring, we investigated the seasonality of hospital admissions among patients with a sarcoidosis diagnosis [25, 26]. Generally, seasonality of a disease or outcome implies that the driving force is an environmental influence, such as a seasonal viral infection or exposure. However, when we looked at patients specifically admitted for a primary diagnosis of sarcoidosis, we did not find any evidence of a seasonal pattern. We did find a seasonal pattern (peaking in January, February, and March) for admissions due to all causes. This could be associated with the timing of the influenza season. However, influenza cannot account for the dramatic eleven-year overall increase in the rate of hospitalizations of sarcoidosis patients. We performed further analyses to investigate whether the upward trend in sarcoidosis over the study period is still statistically significant after adjusting for influenza. This analysis was performed on the overall series and on the subseries comprised of patients 55 and older (data not shown). Specifically, we added a series representing national influenza activity as an explanatory variable in our seasonal ARIMA models for these series. In both models, the linear trend remains highly statistically significant (both p-values < 0.0001). Thus, adjustment for influenza does not impact the trend in sarcoidosis admissions.
There are several limitations to our study. First, a major limitation is that we use that we use administrative data, and do not have access to laboratory, radiographic or medication data. Second, we cannot confirm our hypothesis that immunosuppressive drugs are driving the change in epidemiology, nor can the coded diagnosis of sarcoidosis be confirmed. We also cannot report the burden of repeated hospitalizations. Due to the sampling design in the AHRQ database, there are no individual identifiers that can link patients across multiple admissions or over multiple years. Third, it is possible that medical advances are sustaining patients for longer, leading to increased number of hospitalizations that are not necessarily related to the severity or treatment of sarcoidosis. Fourth, it is also possible that subgroups with limited access to outpatient care (whether for financial or geographic reasons) may be admitted more frequently. Fifth, racial data in large databases may be missing or inaccurate. Last, it is possible that the increase in hospitalizations is due to an increase in the prevalence of sarcoidosis during the study period. However, hospitalizations specifically for a primary diagnosis have remained stable, suggesting that increased incidence of sarcoidosis is not the driving factor for hospitalization. The incidence in populations in the United Kingdom and Australia (both with similar incidence data to the United States) show that the diagnosis of sarcoidosis is not increasing over time [27, 28].