Daily Adjustment of Glucocorticoids by Patients With Adrenal Insufficiency

Christof Schöfl; Bernhard Mayr; Nicole Maison; Felix Beuschlein; Gesine Meyer; Klaus Badenhoop; Tina Kienitz; Marcus Quinkler


Clin Endocrinol. 2019;91(2):256-262. 

In This Article



Patients with AI were recruited in four specialized tertiary endocrine centres across Germany, which are using patients' diaries in their everyday clinical practice for patients with AI. In 2014, a total of 100 patients were asked to participate by returning their patient's diaries. Finally, 80 patients returned their diaries and were included in this study. The patients were seen every 4–6 months by experienced endocrinologists (CS, BM, FB, GM, KB, TK and MQ). The endocrinologists evaluated GC replacement therapy according to best clinical practice guidelines:[2,4,13] the monitoring of GC replacement predominantly relied on the clinical assessment including the search for symptoms and signs of over- or under-replacement. Patients with secondary AI and additional pituitary hormonal insufficiencies received appropriate hormonal replacement therapy according to current guidelines and were on stable replacement doses during the study. Patients' histories were reviewed and data extracted from medical records regarding sex, age at begin of study, age at diagnosis, duration of AI, type of AI (primary, secondary), weight, height, BMI and medication. The study was approved by the ethics committee of the Charité—University Medicine Berlin (permit No. ES1/037/06). Patients' characteristics are shown in Table 1.


Patients were instructed to record information by ticking a box for each single day when they experienced either general discomfort (question: "Did you experience any health-related problems today?"—answer: "yes" or "no") or adjusted their usual GC dosage (question: "Did you change your GC replacement dose today?"—answer: "yes" or "no"). For each of these days answered with "yes," the patients noted presence or absence of gastrointestinal (GI) infection (ticking a box), other infection (ticking a box and then free text), fever (ticking a box), vomiting (ticking a box) and other symptoms (ticking a box and then free text). In addition, patients recorded their type of GC, their normal daily dose and their mineralocorticoid dose in the case of PAI. On days with GC dose adjustment, the patients filled in the additional dose ("additional daily dose: …..mg") and whether the dose adjustment was advised or administered by a physician (ticking a box).

The start date and end date of each diary were noted. Total observation period in days, number of days with general discomfort and both general discomfort and dose adjustment were determined from the diaries. Possible overdosage (days with dose adjustment but no indication of discomfort) was not evaluated because signs and symptoms of elevated GC need or GC deficiency may be under-reported and some patients may have increased GC therapy temporarily to treat symptoms from unrelated diseases such as osteoarthritis or strenuous physical activity. The daily records of the diaries were evaluated by a trained endocrinologist (BM). The following data were determined from yes/no answers and free-text entries in the context of all available medical information including the medical history of the patient.

Analysis of Signs and Symptoms

Signs and symptoms that may cause intermittent elevated GC need or indicate intermittent GC deficiency were infections, fever, diarrhoea, anorexia or nausea, emesis, fatigue, dizziness or trembling as a possible sign for hypotension, hypoglycaemia, hyponatraemia; or documented hypotension, hypoglycaemia, hyponatraemia as well as muscle, joint or abdominal pain (but not headache). If signs and symptoms including the free text indicated a specific diagnosis such as urinary tract infection or broken ankle, a defined intercurrent illness was recorded. In case of documented emesis, it was assumed that nausea was also present. When patients documented "GI infection," it was classified as infection, intercurrent illness and diarrhoea. When intercurrent illnesses were not infectious episodes (eg epileptic attacks, sprained ankle or radius fracture), the physical stress level (little, medium, severe) was noted. In all but one event, the physical stress level was low. In a single case of hip fracture, immediate hospitalization was necessary and GC therapy was determined by physicians; therefore, this event is not included in the single day evaluation. The presence of symptoms was not regarded as sign of elevated GC need or GC deficiency, if a confirmed chronic condition or unrelated cause was present, for example bone pain after minor injury, back pain in osteoporosis, joint pain in osteoarthritis or chronic diarrhoea in a patient with inflammatory bowel disease. However, acute exacerbations of chronic diseases, for example acute episodes of bronchitis in patients with chronic obstructive lung disease or epileptic attacks, were included.

We created an unvalidated sign and symptom score (ssScore) to sum up and weighed the number of signs and symptoms present during an episode of general discomfort and/or intercurrent illness (Table 2). For all days with general discomfort, the number of days with different ssScore and the percentage of days with dose adjustments were identified and calculated for each patient. In addition, the number of days with fever and GI infection and the percentage of days with dose adjustments were calculated for each patient. Adrenal crisis was defined according to Allolio:[14] profound impairment of general health and at least two of the following conditions: hypotension (systolic blood pressure <100 mm Hg), nausea or vomiting, severe fatigue, hyponatraemia, hypoglycaemia and hyperkalaemia, triggering subsequent parenteral GC administration.

Statistical Methods

Statistical analysis was performed using the Statistical Package for the Social Sciences 19 (SPSS 19). Data generally were not normally distributed. To draw comparisons between variables, Mann-Whitney U tests were used, if data were interval- or ordinal-scaled. To evaluate relationships between two variables, Spearman's rank order correlation was assessed. All results were tested for two-tailed significance. Test results were considered statistically significant if P < 0.05.