Data source
This is a retrospective cohort study using a multihospital claims database provided by Medical Data Vision Co., Ltd. (Tokyo, Japan). As of April 2014, the database consisted of information on 7.39 million patients from 145 diagnosis procedure combination (DPC) hospitals across Japan and covered about 9% of all acute care hospitals located in multiple districts. DPC hospitals are acute phase hospitals administered under the Diagnosis Procedure Combination/Per-Diem Payment System. The database has two parts, i.e. administrative data on outpatients and inpatients and in-hospital data on DPC [12]. Items in the database included patient characteristics (age, sex), medical conditions (disease indicating hospital admission, comorbidities on admission, primary disease, secondary disease, and post-admission complications) which were classified according to the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10) and medical care provided (prescriptions, laboratory tests, interventional procedures such as surgery, date of admission) [13, 14]. For this database, attending physicians were in charge of recording the diagnoses through referring to medical charts. Age and gender distributions of the source population were similar to those of the national census in Japan and several epidemiologic studies using the database had been published [14, 15]. We analyzed de-identified data entered between January 1, 2003 and April 30, 2014 for this study. This study was approved by the Kyoto University Graduate School and Faculty of Medicine, Ethics Committee (E2051). This study was exempt from obtaining individual informed consent based on the Ethical Guidelines for Medical and Health Research Involving Human Subjects by the Ministry of Health, Labour, and Welfare and Ministry of Education, Culture, Sports, Science and Technology.
Study cohort and disease definition
AC is a potentially life-threatening condition and needs immediate emergency treatment. In general, it is more severe than AI, but there are no clear distinctions between them. To identify appropriate cases of AC, we used a two-step process. First, we identified candidate patients using ICD-10 in combination with a disease name that corresponded to AI (Additional file 1: Table S1). Both were recorded under administrative data consistent with AI, and those patients were enrolled in the study. We assumed that the date of recording AI was the date of diagnosis; therefore, we used that date as the index date. Second, we confirmed AC in the study cohort according to fulfillment all of the following criteria: 1) therapeutic GC administration (hydrocortisone equivalent dose ≥100 mg/day) within 3 days before or after the index date (Additional file 2: Figure S1); 2) admission within 3 days before or after the index date (Additional file 2: Figure S2); and 3) age 18 years or older at the index date.
We did not provide an exclusion criterion for treatment with high-dose GC (hydrocortisone equivalent dose >200 mg/day) because we intended to investigate practical clinical management of patients with AC. Besides, we considered the possibility that hospitals stocked a variety of dosages. Treatment with excessive GC administration may be inappropriate for AC, we assessed separately for patients receiving excessive GC administration (hydrocortisone equivalent dose >1000 mg/day) (Additional file 1: Table S2).
Predisposing conditions
Hypothalamic-pituitary-adrenal (HPA) axis dysfunction may be a risk factor for AC. We defined and classified predisposing conditions according to the pathologic mechanism of HPA axis function. The predisposing conditions were classified as either primary or central AI or “Others”.
Xem thêm : Subaru WRX STI Car Covers
Primary AI results from a disease intrinsic to the adrenal cortex. There are two major mechanisms for the development of primary AI: 1) adrenal hypofunction such as in Addison’s disease and 2) medical interventions for adrenal disease, such as surgical resection of the bilateral adrenal cortex, leading to adrenal hypofunction. Additional file 1: Table S3 and Table S4 show diseases and interventions that are risk factors for primary and central AI based on previous reports [1, 16]. For this study, we determined that patients with primary AI had to meet either of the following criteria: 1) the presence of a disease considered to be the cause of primary AI and 2) having had an adrenal disease for which an intervention involving the adrenal cortex was performed twice or more since the adrenal organ exists bilaterally.
Central AI results from impairment of the hypothalamic-pituitary axis. We identified patients with central AI by either the presence of central AI itself or having undergone an intervention involving the hypothalamic-pituitary axis at least once.
There could be a prolonged lag between the hospital visit and diagnosis of predisposing diseases [17, 18]. We collected information on predisposing conditions from administrative data recorded before the admission and during the hospitalization.
The category of “Others” included various predisposing diseases, examples of which are undiagnosed chronic AI, relative AI, and drug-induced AI. But, provision of criteria is difficult for these specific predisposing conditions. To explore these latent conditions in “Others”, we compared clinical characteristics on admission with or without the following: 1) prior admission (within 1 year before AC); 2) GC medication (including drugs interacting with GC), and 3) GC cessation (within 30 days before AC).
Disease indicating hospital admission and comorbidities on admission
Indications for hospital admission are based on a clinical diagnosis leading a physician to decide on hospitalization. Such indications would include clinical symptoms related to AI (Additional file 1: Table S5). We defined clinical symptoms related to AI based on published data [1]. From the database, we identified the following comorbidities on admission by using DPC data: cardiovascular disease, infection, sepsis, cancer, diabetes mellitus, hypothyroidism, any autoimmune disease, peptic ulcer, chronic obstructive pulmonary disease or asthma, renal failure, osteoporosis, dementia, and liver disease (Additional file 1: Table S6). In selecting these comorbidities, we referred to previous studies to confirm a possible relationship with AC [19,20,21,22].
Determination of drug and treatment regimens
To identify medications, we complied with the European Pharmaceutical Marketing Research Association (EphMRA) Anatomical Guidelines. We determined GC drugs corresponding with the H2 category, which indicates systemic corticosteroids in EphMRA. To accurately calculate the dose administered, we excluded uncountable administrations such as by a nebulizer since with this method of administration it is not possible to precisely estimate absorption. Our study did not count fludrocortisone as a GC because fludrocortisone had not been used for an anti-inflammatory effect (Additional file 1: Table S7). Additionally, we did not use fludrocortisone as a proxy of GC replacement therapy for primary AI. Japanese practice guidelines did not recommend fludrocortisone as a routine replacement therapy for primary AI because the Japanese adult population ingests a relatively large amount of salt [23].
We defined the occurrence of AC under treatment with oral GC if the index date occurred while the patient was being prescribed oral GC. Regarding inhaled GC and intra-nasal GC, we determined that AC had occurred while under treatment if the last prescription was within 90 days before the index date.
Concomitant administration of interacting drugs could modify GC effects and metabolism. As previous studies reported the interaction of various drugs with GC [1, 17, 24], we checked such medications using the combination of the EphMRA code and the actual drug name (Additional file 1: Table S8). Determination of the occurrence of AC under treatment with interacting drugs was defined by the same protocol as used for oral GC treatment.
Sensitivity analysis
To identify the study cohort more rigorously, we made an alternative (“narrowly defined”) case definition based on laboratory tests and clinical decisions. For selecting a narrowly-defined case, we required two additional restrictions: 1) having in-hospital hormone testing (cortisol, adrenocorticotrophic hormone, endocrine stimulation test, or adrenal cortex stimulation test) and 2) AI was recorded in DPC data along with any information that included primary, secondary, or a disease indicating admission, comorbidity on admission, or post-admission complication. Hormone testing is not necessary for diagnosis of AC, but Japanese practice guidelines recommended that any condition suspicious for AC should be treated immediately after taking the blood sample (cortisol, adrenocorticotrophic hormone) [23].
Statistical analyses
Characteristics of the study population were summarized using proportions for categorical variables and the median and interquartile range for continuous variables.
Nguồn: https://blogtinhoc.edu.vn
Danh mục: Info