Cushing's syndrome (CS) describes a group of metabolic disorders including Cushing's disease (CD) that are caused by oversecretion of adrenocortical steroids (1). Corticotropic adenomas secreting adrenocorticotropic hormone (ACTH) are the most common cause of endogenous CS (1). Oversecretion of cortisol stimulated by ACTH leads to a variety of signs and symptoms of CD including acne, red cheeks, moon face, ‘buffalo hump’, thin skin, weight gain, a pendulous abdomen and swelling of the feet/legs (2). A population-based study in Denmark reported the annual incidence of CD to be 1.2–1.7 cases per million of the population (3). However, it is possible that there are many undiagnosed cases of CS/CD with poorly controlled diabetes, arterial hypertension and/or osteoporosis, particularly among younger individuals (2). CS has been associated with increased risk of fatal cardiovascular disease and infection (4). Despite the established benefits of surgical treatment in CS/CD, the frequency of obesity-related metabolic disorders (hypertension, diabetes, impaired glucose tolerance, hypercholesterolemia) and the mortality rate of CD patients remain high following surgery (5,6). The current report describes a case of CD that was difficult to treat for body weight and insulin resistance following surgery. Therapy with canagliflozin, a sodium-glucose cotransporter 2 (SGLT2) inhibitor, reduced body weight and the daily insulin dose requirement of the patient. SGLT2 inhibitors may reduce the daily insulin requirement in diabetic patients but there is little information on the clinical course of patients that exhibit successful reduction of daily insulin dose. This is the first reported case of CD-associated diabetes that presented improvement in body weight and insulin resistance following therapy with an SGLT2 inhibitor.

The patient was admitted to Mie University Hospital, Tsu, Japan, in October 2014 for treatment of obesity and diabetes. The patient was a 47-year-old obese woman with a 20-year history of type 2 diabetes mellitus associated with hypertension and dyslipidemia. In the previous hospital (Mie National Hospital, Tsu, Japan), the subject exhibited 414 mg/dl blood glucose in a random glucose test [non-diabetic level: <200 mg/dl (7)] that were difficult to control requiring high doses of insulin. The patient exhibited no diabetic nephropathy or retinopathy and was under a strict diet (1,000 kcal/day) and bicycle exercise therapy (30 min/day). The treatment that the patient was receiving was as follows: Insulin aspart 30 U/day (Novo Nordisk Pharma Ltd., Tokyo, Japan), insulin glargine 36 U/day (Eli Lilly Japan K.K., Kobe, Japan), metformin 1,500 mg/day (Sumitomo Dainippon Pharma Co., Ltd., Tokyo, Japan), sitagliptin 50 mg/day (MSD K.K., Tokyo, Japan), voglibose 0.9 mg/day (Towa Pharmaceutical Co., Ltd., Osaka, Japan), valsartan 80 mg/day (Novartis Pharma K.K.) and rosuvastatin 5 mg/day (AstraZeneca K.K., Osaka, Japan). The clinical findings on examination were as follows: Height, 155 cm; body weight, 87.5 kg, body mass index 36.4 kg/m² (definition of central obesity by International Diabetic Federation: >30 kg/m²) (8); blood pressure, 119/79 mmHg [target blood pressure level in diabetic patients: <130/80 mmHg (9)]; heart rate, 103 beats/min; body temperature, 36.4°C. Physical examination identified central obesity with waist circumference 126 cm (cut-off value of International Diabetic Federation: 80 cm) (8) and whitish striae. Laboratory data on admission including the following: Glycated hemoglobin (HbA1c), 11.2% [reference range: 4.6–6.2% (10)]; fasting blood glucose, 338 mg/dl [non-diabetic level: <126 mg/dl (7)]; plasma ACTH, 49.5 pg/ml [reference range: 7.2–63.3 pg/ml (11)]; plasma cortisol, 21.1 µg/dl [reference range 6.4–21.0 µg/dl (11)]; urinary free cortisol, 295.4 µg/day [reference range: 11.2–80.3 µg/day (11)] (Table I). There was no diurnal variation in plasma cortisol concentration and no suppression of plasma cortisol concentration during an overnight 0.5 mg dexamethasone suppression test (DST) performed for the diagnosis of CS (12). An overnight DST using a high dose of dexamethasone (8 mg) demonstrated suppression of the plasma cortisol levels to less than half of the basal level. The plasma ACTH levels did not exhibit a >50% increase at any time following 100 µg corticotropin releasing hormone (CRH) administration. In a thyrotropin-releasing hormone test, thyroid stimulating hormone exhibited a peak value of 10.34 µU/ml at 30 min [normal response: increase of >5 µU/ml (13)]. In a luteinizing hormone-releasing hormone test, luteinizing hormone and follicle stimulation hormone exhibited peak values of 92.6 mU/ml (normal response: increase of >10 mU/ml) and 94.4 mU/ml (normal response: increase of >2 mU/ml) (13), respectively. In a growth hormone-releasing protein 2 test, growth hormone exhibited a peak value of 9.421 ng/ml at 15 min (criteria for severe adult growth hormone deficiency: values of <9 ng/ml) (14) (Table I). Magnetic resonance imaging (MRI) revealed a 12-mm pituitary tumor with delayed enhancement in T1-weighted MRI (Fig. 1A). Based on the presence of central obesity, diabetes mellitus, hypertension, pituitary tumor on MRI and abnormal findings in endocrinal tests, excluding the results of the CRH test, a diagnosis of CD was made. The reported low sensitivity of the CRH test may explain why the test was negative (15). The patient was then remitted to the Neurosurgery Department for trans-sphenoidal pituitary surgery. The patient's hyperglycemia was treated with 112 U/day insulin the day before surgery and improved following the trans-sphenoidal pituitary surgery. The pathological findings of the resected specimen were consistent with the diagnosis of benign pituitary adenoma as evaluated by histological study and hematoxylin and eosin staining (Fig. 1B). The serum levels of ACTH and cortisol decreased 9 days after surgery (to 9.8 pg/ml and 2.4 µg/dl, respectively).

Following surgery, the patient was treated with 30 mg/day hydrocortisone with gradual tapering over 19 months. The dose of insulin was gradually reduced to avoid hypoglycemia via self-monitoring of blood glucose levels; the dose was reduced by 4–10 units/week at the discretion of the patient. Sitagliptin and voglibose were withdrawn following surgery. After 6 months the level of HbA1c was <7.0% and patient body weight decreased from 85.9 to 80.0 kg. The daily insulin dose was 46 U/day during the postoperative period under self-monitoring and outpatient follow-up. The patient's body weight and daily insulin dose remained stable over the following 5 months. Canagliflozin was administrated 11 months after surgery. Following the administration of 100 mg/day canagliflozin the patient exhibited loss of body weight (to 69.5 kg) and required less daily insulin dose (to 0 U/day). Therapy with insulin was withdrawn 19 months after the surgical procedure when HbA1c level was <7.0%. The dose of valsartan was decreased (from 80 to 40 mg/day) 10 months after and was withdrawn 12 months after starting therapy with canagliflozin. Remission of the CD was confirmed 19 months after the pituitary adenoma resection. In the absence of hydrocortisone, the plasma level of ACTH was 9.0 pg/ml and that of cortisol was 4.9 µg/dl. The body weight of the patient was 55.3 kg and the HbA1c level was 5.8% 31 months after the surgical procedure. This overall clinical course is depicted in Fig. 1C.

CS including CD is characterized by metabolic disorders such as obesity and insulin resistance that are caused by overproduction of steroid hormones (2). CD is associated with a high rate of mortality caused by ischemic coronary artery disease and/or infectious diseases; a recent review article reported that standardized mortality ratios of CD ranged from 1.84 to 4.25 including global retrospective analysis (16). Surgical resection of pituitary adenoma is the first-line treatment for CD but not all patients with CS exhibit improvement in metabolic function following surgical treatment (17). Obesity and diabetes are particularly difficult to control, and the amount of inter-muscular adipose tissue remains unchanged following surgery (5,18). Nevertheless, the present case exhibited loss of body weight and required lower daily insulin dose in the immediate postoperative period, and these signs remained stable 6 months after surgery.

Management of metabolic disorders in the postoperative period for patients with CD undergoing steroid therapy is a challenge. The present case required a relatively long period for withdrawal of steroid administration following surgery. Weight gain is accelerated during steroid therapy as glucocorticoids promote glucose production by the liver, peripheral glucose uptake in muscle and adipose tissues, and inhibit insulin production and secretion in pancreatic β-cells (19,20). Body weight and the daily insulin dose of the current patient remained stable with no improvement between the sixth and eleventh month of the postoperative period. Initiation of canagliflozin, a sodium glucose co-transporter inhibitor, led to decreased body weight and lower requirement of insulin dose; complete insulin withdrawal and a body weight loss of 24.7 kg were observed 8 months after initiating therapy with canagliflozin.

To the best of our knowledge, this is the first report on the therapeutic effect of a sodium glucose co-transporter inhibitor in a CD patient post-surgery. The markedly beneficial effect of canagliflozin on body weight and diabetes in the postoperative period of the present patient whilst undergoing steroid therapy may be explained by the accelerated excretory activity of carbohydrates induced by the drug. Although data supporting the beneficial effects of SGLT2 inhibitor in glucocorticoid-related metabolic disorders is limited, studies on the mechanism of action of the drug suggest its potential favorable activity (21–23). Glucocorticoids increase the expression of sodium-dependent glucose cotransporters 1 and 2 by promoting the activity of serum- and glucocorticoid-inducible kinases; canaglifozin may suppress the effects of both sodium-dependent glucose cotransporters (21). An alternative explanation for the beneficial effect of canagliflozin is its stimulatory activity on the AMP-activated protein kinase, which may increase insulin sensitivity by activating intracellular signal pathways and decrease body weight by promoting energy consumption in adipose tissues (22). Therefore, it is conceivable that canagliflozin counteracted the inhibitory effect of glucocorticoids on AMP-activated protein kinase in the adipose tissues of the current CD patient undergoing steroid therapy during the postoperative period (23).

It is noteworthy that the use of SGLT2 inhibitor in the early post-operative period may be detrimental since it may cause profound diuresis, which eventually may result in increased predisposition for deep vein thrombosis and hypotension due to decreased intravascular volume (24–26). Therefore, careful monitoring of the patient condition is required when SGLT2 inhibitor is used after surgery.

The present report demonstrates the potential of canagliflozin for the management of diabetic patients receiving steroids following surgical therapy for CD. However, it is noteworthy that other agents including glucagon-like peptide-1 agonists may also have similar beneficial effects on weight reduction in CD (27). The use of HbA1c alone and the lack of a glucose tolerance test to confirm the diabetic condition are limitations of the present case report. The blood glucose levels of the patient were measured during consultation at the outpatient department. However, the blood glucose levels measured during follow-up at the outpatient department are usually variable depending on the time from the last meal. Pasireotide has become available for the treatment of CD caused by pituitary adenoma (28). Pasireotide was not used in the present case since this medication has not been officially approved in Japan and the tumor was completely removed.

Despite the beneficial effects of canagliflozin observed in the current report, further data are required to demonstrate the effectiveness of SGLT2 inhibitor in CS patients with diabetes.