Real-world evidence for the safety of ipragliflozin in elderly Japanese patients with type 2 diabetes mellitus (STELLA-ELDER): final results of a post-marketing surveillance study

ABSTRACT

Objective: To investigate the real-world safety of ipragliflozin in elderly Japanese patients with type 2 diabetes mellitus (T2DM).

Research design and methods: Japanese patients (≥65 years old) who were first prescribed ipragliflozin within 3 months after its launch in April 2014 were registered in this post-marketing surveillance (PMS). Final data collection was in July 2015. Survey items included demographics, treatments, adverse drug reactions (ADRs), vital signs, and laboratory variables.

Results: The PMS included 8505 patients (4181 males/4324 females). The mean age and diabetes duration were 72.3 years and 10.6 years, respectively. In 84.3% of patients, ipragliflozin was prescribed at 50 mg/day, which was continued unchanged. Overall, 16.91% of patients experienced 1880 ADRs, and 165 ADRs were classified as serious in 127 patients (1.49%). ADRs of special interest included skin complications, volume depletion, polyuria/pollakiuria, genital infection, urinary tract infection, renal disorders, hypoglycemia, cerebrovascular disease, cardiovascular disease, malignant tumor, fracture, and ketone body-related events.

Conclusions: This 1-year PMS revealed probable ADRs in elderly Japanese patients with T2DM prescribed ipragliflozin in real-world settings, with no new safety concerns. The risk factors for ADRs varied but could be rationalized. The results should help physicians to identify possible treatment-emergent ADRs in ipragliflozin-treated patients.

KEYWORDS:

• Elderly
• ipragliflozin
• post-marketing surveillance
• SGLT2 inhibitors
• type 2 diabetes mellitus

View correction statement:

Corrigendum

1. Introduction

Several sodium–glucose cotransporter 2 (SGLT2) inhibitors are available as a treatment option for type 2 diabetes mellitus (T2DM). Inhibition of SGLT2 lowers blood glucose concentrations by increasing urinary glucose excretion [1]. This glucose-lowering effect of SGLT2 inhibitors is independent of insulin signaling [1].

Ipragliflozin was the first SGLT2 inhibitor to be approved in Japan for the treatment of T2DM on the basis of several randomized placebo-controlled studies, which demonstrated that it improved glycemic control when administered as monotherapy or in combination with other oral antidiabetic drugs [2–7].

In these studies, the incidence rates of mild-to-moderate treatment-emergent adverse events (TEAEs), including genital infection and hypoglycemia, were generally low. However, some TEAEs were more common with certain treatment combinations. For example, pollakiuria and constipation were more common with metformin in combination with ipragliflozin [3], and pollakiuria and thirst were more common with sulfonylurea in combination with ipragliflozin [2] than when metformin or sulfonylurea were administered in combination with placebo.

These earlier trials enrolled relatively few elderly patients. However, more than 25% of the Japanese population is aged ≥65 years and many patients with T2DM are of an advanced age [8,9]. For these reasons, it was necessary to confirm the safety of ipragliflozin in elderly patients. Accordingly, Astellas Pharma Inc. and the Pharmaceutical and Medical Device Agency (PMDA) of Japan agreed that an all-case post-marketing surveillance (PMS) of elderly patients was necessary to confirm the safety of ipragliflozin in Japanese patients aged ≥65 years. Therefore, a 1-year PMS Specified drug use resulTs survEy of ipragLifLozin treAtment in ELDERly type 2 diabetic patients (STELLA-ELDER) was implemented in April 2014 in compliance with the ministerial ordinance on Good Post-Marketing Study Practice (GPSP). The PMS was designed to register all elderly Japanese patients with T2DM who were prescribed ipragliflozin within 3 months of its launch.

We previously reported the results of an interim analysis (STELLA-ELDER [10]) that was conducted in order to address possible safety concerns regarding SGLT2 inhibitors in real-world clinical practice in Japan, where a large number of serious adverse events (AEs) were reported in the first 3 months of using these drugs [11]. The PMS registered 8687 patients across 2634 institutions, and 6-month data were available for up to 7170 patients, while 1-year data were available for up to 1140 patients [10].

In this report, we describe the final results after completing the data collection at 1 year in a cohort of 8505 patients. We present the incidence of adverse drug reactions (ADRs), and changes in laboratory variables and vital signs. We also describe the results of multivariable logistic regression analyses, which were conducted to identify which clinical factors were potentially associated with ADRs of special interest.

2. Patients and methods

2.1. Patients and treatments

The design of STELLA-ELDER is described in more detail in the interim report [10]. Participating physicians were to register all patients with T2DM aged ≥65 years who were first prescribed ipragliflozin within 3 months after its launch in April 2014. In accordance with the approved label, patients were prescribed 50 mg ipragliflozin once daily, to be administered before or after breakfast. The label also permits a dose increase to 100 mg providing the patient’s clinical course is monitored carefully. A lower dose was also permitted according to the label, which states that a lower dose should be used with caution in patients with severe hepatic impairment. All treatment decisions were at the attending physician’s discretion. In compliance with Japanese regulations for PMS, it was not necessary to obtain informed consent from the patients. All of the medical institutions that agreed to provide data signed a contract with Astellas Pharma Inc.

2.2. Survey design

The survey was conducted between April 2014 and July 2015, and the database was locked on 16 January 2016. All data were to be recorded immediately in an electronic database using case report forms for each patient.

The survey items included the following demographic characteristics: age, sex, date of the start of administration, height, body weight, body mass index (BMI), inpatient/outpatient status, duration of diabetes, diabetic complications (neuropathy, nephropathy, retinopathy), presence of urinary tract infection, presence of genital infection, and the degree of renal and hepatic function at the start of drug administration. The survey forms also included the following medication data: the daily dose, daily administration frequency, administration period, reason of termination or discontinuation of administration of ipragliflozin, and the type and dose of concomitant medications (including antidiabetic drugs and diuretics). Vital signs and laboratory variables included blood pressure, heart rate, glycated hemoglobin (HbA1c), fasting plasma glucose, hematocrit, serum albumin, serum creatinine, and blood urea nitrogen (BUN).

AEs and ADRs that occurred during treatment with ipragliflozin were to be recorded at each visit, including the type of AE/ADR during or after drug administration, the date of onset, seriousness, outcome (resolved, improved, or worsened), causal relationship to ipragliflozin, causative relationship to factors other than ipragliflozin, and laboratory data at the time of the event. Physicians were provided with clear definitions on what constitutes an AE. To mitigate possible reporting bias, the presence/absence of AEs was verified with physicians upon receipt of every survey sheet. AEs and ADRs were categorized according to system organ class and preferred term using MedDRA version 18.1 [12].

ADRs of special interest were defined as those related to hypoglycemia, genital infection, urinary tract infection, polyuria/pollakiuria, volume depletion, renal disorders, skin complications, fracture, malignant tumor, cardiovascular disease, cerebrovascular disease, and ketone body-related events. These classes of ADRs were selected based on a risk management plan submitted to the PMDA and considering the emergence of skin-related ADRs reported shortly after the launch of ipragliflozin. For all ADRs reported, the physicians were asked whether they were possibly related to reductions in body fluid or dehydration, and these ADRs were analyzed as volume depletion-related events. Skin complications included all ADRs related to skin. Ketone body-related events included diabetic ketoacidosis, ketoacidosis, ketone body increased, ketosis, and metabolic acidosis.

2.3. Statistical analysis

All statistical analyses were performed using SAS statistical software version 9.2 or higher (BASE/SAS®, SAS/STAT®, SAS Institute Inc., Cary, NC, USA). Data are presented as the mean and standard deviation (SD) for continuous variables, and as the frequency and proportion for categorical variables. The changes in continuous variables from baseline were tested using paired t-tests. Differences in the incidences of ADRs between groups of patients were tested using Fisher’s exact test, and the trends in the incidences of ADRs across ordered categorical variables were determined using the Cochran–Armitage test. Multivariable logistic regression analysis was conducted to identify variables associated with any ADR or with ADRs of special interest. The final models were built using a stepwise procedure with p-values of <0.15 for entry and staying in the model. All p-values are derived from two-tailed tests. The sample size was not predetermined.

3. Results

3.1. Patient disposition and baseline characteristics

Overall, 8687 patients were initially registered across 2634 participating medical institutions in Japan. Survey forms were not collected for 87 patients because the physician did not submit the survey form (n = 80), the institution’s policy changed (n = 6), or the institution closed (n = 1). Survey forms were collected for 8600 patients, but 95 patients were excluded for the following reasons: the patient did not visit the institution after the initial visit (n = 77), the existence of an AE was unclear (n = 7), unable to complete the required form (n = 6), ineligible (n = 3), ipragliflozin not administered (n = 1), and duplicate patient (n = 1). Therefore, the safety population comprised 8505 patients.

Patient characteristics at baseline are shown in Table 1. There were 4181 males and 4324 females (49.2%/50.8%). The mean ± SD age, BMI, duration of diabetes, and estimated glomerular filtration rate (eGFR) were 72.3 ± 5.9 years, 27.0 ± 4.56 kg/m², 10.6 ± 7.52 years, and 69.7 ± 19.4 mL/min/1.73 m², respectively. Nearly one-third of patients (31.8%) were ≥75 years old. In the preapproval clinical trials, 69.4% of patients were male and 30.6% were female. The mean age, BMI, duration of diabetes, and eGFR of patients were 58.8 years, 25.75 kg/m², 93.8 months (7.8 years), and 83.88 mL/min/1.73 m², respectively.

Table 1. Patient characteristics at baseline.

	No. of patients (%) or mean ± SD
Safety analysis set	8505 (100.0)
Sex
Male	4181 (49.2)
Female	4324 (50.8)
Age, years	72.3 ± 5.9
Age category 1
<75 years	5800 (68.2)
≥75 years	2705 (31.8)
Age category 2
≥65 to <70 years	3287 (38.6)
≥70 to <75 years	2513 (29.5)
≥75 to <80 years	1602 (18.8)
≥80 to <85 years	802 (9.4)
≥85 years	301 (3.5)
Inpatient/outpatient status
Inpatient	147 (1.7)
Outpatient	8358 (98.3)
Duration of diabetes, years*	10.6 ± 7.52
Duration of diabetes category
<5 years	1177 (13.8)
≥5 years	4150 (48.8)
Unknown	3178 (37.4)
BMI (kg/m^2)†	27.0 ± 4.56
BMI category
<18.5 kg/m^2	73 (0.9)
≥18.5 to <22.0 kg/m^2	518 (6.1)
≥22.0 to <25.0 kg/m^2	1171 (13.8)
≥25.0 to <30.0 kg/m^2	2168 (25.5)
≥30.0 kg/m^2	1138 (13.4)
Unknown	3437 (40.4)
Complications
No	1477 (17.4)
Yes	6917 (81.3)
Unknown	111 (1.3)
Type of complication
Diabetic neuropathy	749 (8.8)
Diabetic nephropathy	990 (11.6)
Diabetic retinopathy	521 (6.1)
Urinary tract infection	38 (0.4)
Genital infection	9 (0.1)
Dyslipidemia	4294 (50.5)
Hypertension	5047 (59.3)
Hyperuricemia	532 (6.3)
Dementia	177 (2.1)
Macrovascular disorders	1701 (20.0)
Coronary artery disease	1023 (12.0)
Cerebral vascular disease	604 (7.1)
Arteriosclerosis obliterans	357 (4.2)
Other	3737 (43.9)
Hepatic function status§
Normal	7357 (86.5)
Mild impairment	764 (9.0)
Moderate impairment	55 (0.6)
Severe impairment	3 (0.0)
Unknown	326 (3.8)
Renal function status
Normal	6777 (79.7)
Mild impairment	1295 (15.2)
Moderate impairment	145 (1.7)
Severe impairment	11 (0.1)
Unknown	277 (3.3)
eGFR, mL/min/1.73 m^2‡	69.7 ± 19.4
eGFR category
≥90 mL/min/1.73 m^2	671 (7.9)
≥60 to <90 mL/min/1.73 m^2	2961 (34.8)
≥45 to <60 mL/min/1.73 m^2	1172 (13.8)
≥30 to <45 mL/min/1.73 m^2	366 (4.3)
<30 mL/min/1.73 m^2	67 (0.8)
Unknown	3268 (38.4)

*n = 5327; †n = 5068; ‡n = 5237.

§Assessed by total bilirubin, aspartate aminotransferase, and alanine aminotransferase levels.

¶Assessed by blood urea nitrogen and creatinine levels.

BMI: body mass index; eGFR: estimated glomerular filtration rate; SD: standard deviation.

3.2. Prescribed treatments

The initial dose of ipragliflozin was 50 mg in 86.3% of patients, 25 mg in 13.5% of patients, and 100 mg in 0.1% of patients. The dose was increased from 25 mg to 50 mg in 3.1% of patients and from 50 mg to 100 mg in 0.9% of patients. In 84.3% of patients, ipragliflozin was initially prescribed at 50 mg and this dose was continued unchanged. Ipragliflozin was prescribed in combination with other antidiabetic drugs in 84.6% of patients. The most commonly used antidiabetic drugs were dipeptidyl peptidase-4 (DPP-4) inhibitors (60.7%), sulfonylureas (36.3%), and biguanides (28.2%) (Table 2). Antihypertensive drugs, statins, antiplatelet drugs, antipeptic ulcer drugs, and diuretics were prescribed to 47.4%, 35.5%, 14.7%, 13.9%, and 13.4% of patients, respectively.

Table 2. Prescribed treatments.

	No. of patients (%) or mean ± SD
Safety analysis set	8505 (100.0)
Initial dose of ipragliflozin
25 mg	1152 (13.5)
50 mg	7344 (86.3)
75 mg	0 (0.0)
100 mg	9 (0.1)
Dose changes during treatment
25 mg to 25 mg	856 (10.1)
25 mg to 50 mg	263 (3.1)
50 mg to 50 mg	7168 (84.3)
50 mg to 100 mg	78 (0.9)
Other	140 (1.6)
Duration of treatment with ipragliflozin, days	277.6 ± 129.6
Concomitant antidiabetic drug
No	1257 (14.8)
Yes	7192 (84.6)
Unknown	56 (0.7)
Type of concomitant antidiabetic drug
DPP-4 inhibitor	5165 (60.7)
Sulfonylurea	3089 (36.3)
Biguanide	2400 (28.2)
α-Glucosidase inhibitor	1190 (14.0)
Insulin	914 (10.7)
Thiazolidine	898 (10.6)
Fast-acting insulin secretagogue	305 (3.6)
GLP-1 receptor agonist	143 (1.7)
Other	660 (7.8)
Concomitant diuretic drug
No	7285 (85.7)
Yes	1136 (13.4)
Unknown	84 (1.0)
Type of concomitant diuretic drug
Loop diuretic	423 (5.0)
Thiazide diuretic	346 (4.1)
Potassium-sparing diuretic	230 (2.7)
Vasopressin antagonist	5 (0.1)
Carbonate dehydratase inhibitor	2 (0.0)
Osmotic diuretic	1 (0.0)
Other	309 (3.6)
Other concomitant drugs
No	2464 (29.0)
Yes	5930 (69.7)
Unknown	111 (1.3)
Type of other concomitant drugs
Antihypertensive drugs	4031 (47.4)
ARB	2007 (23.6)
CCB	2158 (25.4)
ARB + CCB	999 (11.7)
Statin	3018 (35.5)
Antiplatelet drug	1253 (14.7)
Antipeptic ulcer drug	1181 (13.9)
Antihyperuricemia drug	447 (5.3)
Other	5704 (67.1)

ARB: angiotensin receptor blocker; CCB: calcium channel blocker; DPP-4: dipeptidyl peptidase-4; GLP-1: glucagon-like peptide-1; SD: standard deviation.

Ipragliflozin was discontinued in 3093 (36.4%) patients for the following reasons: AE (1013 patients, 11.9%), patient request (excluding AE; 682 patients, 8.0%), no change or worsening of diabetes (527 patients, 6.2%), the patient did not visit the institution during the survey period (416 patients, 4.9%), improvement in diabetes (144 patients, 1.7%), administration of ipragliflozin stopped for ≥2 consecutive months (15 patients, 0.2%), or another reason (553 patients, 6.5%). Some patients discontinued treatment because of more than two reasons.

3.3. Safety

Seventeen patients died during the study, and the causes were recorded as brain stem infarction, pancreatic carcinoma, unknown cause, and sudden death (unknown cause in one patient and suspected cardiovascular death in one patient) in two patients each, and abscess, acute cardiac failure, acute myocardial infarction, atrial fibrillation/chronic kidney disease, cardiorespiratory arrest, cerebral hemorrhage, chronic cardiac failure, sensory disturbance (suspected cerebrovascular event), and urine output decreased/chronic kidney disease in one patient each. The physician reported the relationship of these deaths to the use of ipragliflozin as unknown (n = 12) or unlikely (n = 5).

The incidence rates of ADRs (overall and by system organ class) for ipragliflozin in the current survey are summarized in Table 3. Overall, 1438 patients (16.91%) experienced 1880 ADRs, and 165 ADRs were classified as serious in 127 patients (1.49%). The most common ADRs (by system organ class) were skin and subcutaneous tissue disorders (276 patients, 3.25%), renal and urinary disorders (233 patients, 2.74%), and infections and infestations (190 patients, 2.23%).

Table 3. Adverse drug reactions.

	This survey	Preapproval clinical trials
No. of participating institutions	2588	144
No. of evaluable patients	8505	1669
Number of patients with an ADR (%)	1438 (16.91)	549 (32.89)
Number of ADRs	1880	887
Type of ADR, number of patients (%)*
Infections and infestations	190 (2.23)	64 (3.83)
Neoplasms benign, malignant, and unspecified (including cysts and polyps)	11(0.13)	4 (0.24)
Blood and lymphatic system disorders	5 (0.06)	8 (0.48)
Immune system disorders	1 (0.01)	0 (0)
Endocrine disorders	2 (0.02)	0 (0)
Metabolism and nutrition disorders	156 (1.83)	20 (1.20)
Psychiatric disorders	11 (0.13)	3 (0.18)
Nervous system disorders	144 (1.69)	42 (2.52)
Eye disorders	10 (0.12)	17 (1.02)
Ear and labyrinth disorders	2 (0.02)	5 (0.30)
Cardiac disorders	35 (0.41)	10 (0.60)
Vascular disorders	33 (0.39)	7 (0.42)
Respiratory, thoracic, and mediastinal disorders	14 (0.16)	7 (0.42)
Gastrointestinal disorders	146 (1.72)	101 (6.05)
Hepatobiliary disorders	15 (0.18)	8 (0.48)
Skin and subcutaneous tissue disorders	276 (3.25)	48 (2.88)
Musculoskeletal and connective tissue disorders	42 (0.49)	13 (0.78)
Renal and urinary disorders	233 (2.74)	176 (10.55)
Reproductive system and breast disorders	125 (1.47)	25 (1.50)
General disorders and administration site conditions	152 (1.79)	101 (6.05)
Investigations	162 (1.90)	133 (7.97)
Injury, poisoning and procedural complications	4 (0.05)	0 (0)

*By system organ class (MedDRA version 18.1).

ADR: adverse drug reaction.

The incidence rates for ADRs of special interest are shown in Table 4. The incidence of these ADRs ranged from 0.02% to 3.16%. ADRs related to hypoglycemia included hypoglycemia (57 ADRs, 0.67%) and blood glucose decreased (1 ADR, 0.01%). ADRs related to genital infection included pruritus genital (95 ADRs, 1.12%), followed by vulvovaginal candidiasis (16 ADRs, 0.19%), vulvitis (10 ADRs, 0.12%), and balanoposthitis (9 ADRs, 0.11%). ADRs associated with urinary tract infection included cystitis (52 ADRs, 0.61%) and urinary tract infection (47 ADRs, 0.55%). ADRs associated with polyuria/pollakiuria included pollakiuria (125 ADRs, 1.47%), nocturia (30 ADRs, 0.35%), polyuria (15 ADRs, 0.18%), and urine output increased (7 ADRs, 0.08%). ADRs associated with volume depletion included dehydration (49 ADRs, 0.58%), thirst (46 ADRs, 0.54%), blood urea increased (37 ADRs, 0.44%), and weight decreased (20 ADRs, 0.24%). ADRs associated with renal disorders included blood urea increased (57 ADRs, 0.67%), renal impairment (33 ADRs, 0.39%), and blood creatinine increased (23 ADRs, 0.27%). The most common ADR associated with cardiovascular disease/cerebrovascular disease was cerebral infarction (20 ADRs, 0.24%). Four serious cases of cerebral infarction were determined by the physician to be possibly related to volume depletion, but none were deemed possibly related to postural hypotension. No cases of stroke were reported to be possibly related to postural hypotension. Other ADRs included: atrial fibrillation (five ADRs, 0.06%); transient ischemic attack (four ADRs, 0.05%); angina pectoris and brain stem infarction (three ADRs, 0.04%); acute myocardial infarction, cardiac failure, cardiac failure acute, cardiac failure chronic, coronary artery stenosis, myocardial infarction, ventricular extrasystoles, cerebral hemorrhage, lacunar infarction, and thrombotic cerebral infarction (two ADRs each, 0.02%); and complete atrioventricular block, cardiorespiratory arrest, deep vein thrombosis, cerebellar infarction, intracranial aneurysm, and putamen hemorrhage (one ADR each, 0.01%). ADRs associated with skin complications included drug eruption (59 ADRs, 0.69%), rash (41 ADRs, 0.48%), eczema (40 ADRs, 0.47%), pruritus (40 ADRs, 0.47%), urticaria (18 ADRs, 0.21%), rash pruritic (17 ADRs, 0.20%), rash generalized (14 ADRs, 0.16%), erythema (11 ADRs, 0.13%), and pruritus generalized (9 ADRs, 0.11%). ADRs associated with malignant tumors included gastric cancer and pancreatic carcinoma (3 ADRs each, 0.04%), colon cancer and lung neoplasm malignant (2 ADRs each, 0.02%), and breast cancer female (1 ADR, 0.01%). Fractures occurred in two patients (0.02%), being femoral neck fracture and radius fracture in one patient each (0.01%). Three ketone body-related ADRs (diabetic ketoacidosis, ketosis, and metabolic acidosis) occurred in two patients (0.02%).

Table 4. Adverse drug reactions of special interest.

	This survey
	All	Serious	Pre-approval clinical trials*
All ADRs, no. of patients (%)	1438 (16.91)	127 (1.49)	(32.9)
Type of ADRs, no. of ADRs or patients (%)
Skin complications	269 (3.16)	11 (0.13)	(2.9)
Volume depletion	266 (3.13)	19 (0.22)	(5.8)
Polyuria/pollakiuria	170 (2.00)	1 (0.01)	(12.8)
Genital infection	166 (1.95)	1 (0.01)	(2.1)
Urinary tract infection	118 (1.39)	11 (0.13)	(2.2)
Renal disorder	118 (1.39)	7 (0.08)	(4.2)
Hypoglycemia	58 (0.68)	2 (0.02)	(1.8)
Cerebrovascular disease	36 (0.42)	31 (0.36)	(0.2)
Cardiovascular disease	24 (0.28)	16 (0.19)	(0.6)
Malignant tumor	11 (0.13)	11 (0.13)	(0.2)
Fracture	2 (0.02)	2 (0.02)	0
Ketone body-related events	2 (0.02)	1 (0.01)	0

*Definition of the ADRs at approval: skin complications, skin and subcutaneous tissue disorders; renal disorder, renal impairment + β-2 microglobulin increased + urinary β 2-microglobulin increased + β-N-acetyl-d-glucosaminidase increased + serum creatinine increased + urinary albumin-to-creatinine ratio (ACR) increased + serum alpha 1-microglobulin increased; cerebrovascular disease, cerebral infarction + transient ischemic attack; cardiovascular disease, cardiac disorder; ketone body-related events, ketoacidosis + ketone body increased + ketosis + metabolic acidosis.

ADR: adverse drug reaction.

3.4. Time to onset and outcome of ADRs of special interest

Of 1880 ADRs, 733 (38.99%) occurred within 30 days of starting treatment and 1235 (65.69%) occurred within 90 days. Meanwhile, 317 ADRs (16.86%) occurred ≥6 months after starting treatment. About one-quarter of ADRs related to polyuria/pollakiuria (24.3%) and skin complications (28.8%) occurred within 7 days of starting treatment, and over half of the ADRs related to these categories occurred within 30 days after the start of treatment (59.3% and 51.9%, respectively). For cardiovascular and cerebrovascular ADRs, there was no apparent tendency in terms of the timing of onset. All ADRs related to malignant tumor occurred 45 days after the start of treatment, except for one event with an unknown time of onset. Both ADRs related to fracture occurred at least 90 days after starting treatment. The ketone body-related events occurred within 7 days of starting treatment. Most ADRs were classified either as resolved (1259/1880, 67.0%) or in remission (436/1880, 23.2%).

3.5. Factors associated with ADRs of special interest

Table 5 shows the factors that were associated with the ADRs of special interest; each model was developed using a stepwise procedure with p-values of <0.15 for entry and staying in the model.

Table 5. Factors associated with the incidence of adverse drug reactions of special interest in the multivariable regression analyses.

ADR class	Factors included in the model*	Odds ratio (95% CI)
All	Female (vs. male)	1.549 (1.297–1.850)
	BMI ≥30.0 kg/m^2 (vs. ≥22.0 to <25.0 kg/m^2)	1.294 (1.003–1.669)
	Moderately impaired hepatic function (vs. normal)	2.298 (1.011–5.225)
	Mildly impaired renal function (vs. normal)	1.376 (1.097–1.725)
Hypoglycemia	BMI <18.5 kg/m^2 (vs. ≥22.0 to <25.0 kg/m^2)	9.356 (2.692–32.514)
	Concomitant insulin use (vs. no)	4.946 (2.540–9.629)
Genital infection	Female (vs. male)	5.927 (3.477–10.104)
	BMI ≥25.0 to <30.0 kg/m^2 (vs. ≥22.0 to <25.0 kg/m^2)	2.061 (1.139–3.728)
	BMI ≥30.0 kg/m^2 (vs. ≥22.0 to <25.0 kg/m^2)	2.363 (1.273–4.387)
Urinary tract infection	Female (vs. male)	5.063 (2.467–10.389)
	Duration of diabetes ≥5 years (vs. <5 years)	5.288 (1.637–17.082)
	Moderately impaired hepatic function (vs. normal)	6.652 (1.452–30.478)
Polyuria/pollakiuria	Concomitant antidiabetic drug use (vs. no)	3.679 (1.143–11.841)
Volume depletion	Age ≥75 years (vs. <75 years)	1.737 (1.209–2.497)
	Concomitant loop diuretic use (vs. no)	2.105 (1.141–3.881)
Renal disorder	Mildly impaired renal function (vs. normal)	3.203 (2.131–4.816)
	Moderately impaired renal function (vs. normal)	7.003 (3.579–13.703)
	Severely impaired renal function (vs. normal)	22.389 (5.412–92.617)
	Concomitant diuretic use (vs. no)	2.631 (1.754–3.949)
Skin complications	Female (vs. male)	1.428 (1.115–1.829)

*The final models were built using a stepwise procedure with p-values of <0.15 for entry and staying in the model.

ADR: adverse drug reaction; CI: confidence interval; BMI: body mass index.

Risk factors for any ADR were sex (female), BMI ≥30 kg/m², moderate hepatic impairment, and mild renal impairment. For hypoglycemia, the risk factors were BMI <18.5 kg/m² and concomitant insulin use. For genital infection, the risk factors were sex (female) and BMI (≥25.0 to <30.0 kg/m²), and BMI ≥30.0 kg/m². Sex (female), diabetes duration (≥5 years), and moderate hepatic impairment were risk factors for urinary tract infection. For polyuria/pollakiuria, the only risk factor was coadministration of an antidiabetic drug. For volume depletion-related ADRs, the risk factors were age ≥75 years and coadministration of a loop diuretic. For renal disorder-related ADRs, the risk factors were renal impairment (mild, moderate, or severe relative to normal), and coadministration of a diuretic. For skin complications, the only risk factor was sex (female).

3.6. Effects of ipragliflozin on laboratory variables and vital signs

The changes in laboratory variables and vital signs during the PMS were also evaluated as surrogate markers (Table 6). HbA1c, fasting plasma glucose, body weight, BMI, systolic blood pressure, and diastolic blood pressure decreased significantly from baseline to each visit and the final visit (all p < 0.001). Significant changes in hematocrit, serum albumin, serum creatinine, BUN, heart rate, and eGFR rate were also observed (all p < 0.001) (Table 6) at some time points.

Table 6. Surrogate markers, vital signs, and laboratory variables.

Variable	Baseline	Final visit	Change at 1 month	Change at 3 months	Change at 6 months	Change at 9 months	Change at 1 year	Change at final visit	Change at final visit in preapproval trials*
HbA1c, %	7.84 ± 1.33	7.28 ± 1.12	−0.33 ± 0.61†	−0.54 ± 0.97†	−0.64 ± 1.05†	−0.59 ± 1.01†	−0.67 ± 1.03†	−0.56 ± 1.04†	−0.51 ± 0.81†
	(6853)	(6853)	(4725)	(5719)	(5004)	(4090)	(4154)	(6853)
HbA1c (50 mg), %	7.84 ± 1.33	7.26 ± 1.11	−0.34 ± 0.61†	−0.56 ± 0.98†	−0.65 ± 1.05†	−0.59 ± 1.00†	−0.68 ± 1.01†	−0.57 ± 1.04†	−0.64 ± 0.66†
	(5751)	(5751)	(3960)	(4786)	(4245)	(3454)	(3551)	(5751)
FPG, mg/dL	164.3 ± 59.3	141.3 ± 48.2	−18.1 ± 50.7†	−22.5 ± 55.1†	−25.5 ± 54.8†	−21.9 ± 51.2†	−28.1 ± 51.1†	−23.0 ± 55.7†	−32.7 ± 31.7†
	(4154)	(4154)	(2792)	(3318)	(2895)	(2313)	(2323)	(4154)
Body weight, kg	67.64 ± 12.84	65.09 ± 12.51	−1.35 ± 1.43†	−2.20 ± 1.98†	−2.51 ± 2.51†	−2.56 ± 2.82†	−2.91 ± 3.08†	−2.56 ± 2.88†	−3.41 ± 2.30†
	(5481)	(5481)	(3845)	(4445)	(3817)	(3114)	(3149)	(5481)
BMI, kg/m^2	27.0 ± 4.54	26.0 ± 4.42	−0.53 ± 0.58†	−0.87 ± 0.79†	−1.00 ± 1.02†	−1.02 ± 1.14†	−1.14 ± 1.25†	−1.01 ± 1.16†	–
	(4734)	(4734)	(3364)	(3847)	(3307)	(2676)	(2717)	(4734)
SBP, mmHg	133.3 ± 15.3	130.0 ± 14.4	−4.1 ± 14.9†	−4.7 ± 15.1†	−2.9 ± 15.0†	−2.3 ± 15.1†	−3.7 ± 15.1†	−3.3 ± 15.6†	−6.4 ± 11.78†
	(6317)	(6317)	(4517)	(5188)	(4525)	(3725)	(3769)	(6317)
DBP, mmHg	73.4 ± 10.2	71.9 ± 9.9	−1.9 ± 9.4†	−2.1 ± 9.5†	−1.2 ± 9.9†	−1.0 ± 9.9†	−1.8 ± 9.9†	−1.5 ± 10.1†	−3.3 ± 7.82†
	(6313)	(6313)	(4512)	(5187)	(4523)	(3722)	(3768)	(6313)
Heart rate, beat/min	74.4 ± 11.1	73.9 ± 11.1	−0.1 ± 8.4	−0.5 ± 8.8	−0.5 ± 9.0	−0.4 ± 9.2	−0.9 ± 8.8†	−0.5 ± 9.2†	–
	(3760)	(3760)	(2711)	(2984)	(2616)	(2184)	(2191)	(3760)
Hematocrit, %	40.9 ± 4.5	42.6 ± 4.7	0.8 ± 2.1†	1.7 ± 2.5†	2.0 ± 2.6†	2.1 ± 3.0†	1.8 ± 3.0†	1.7 ± 3.0†	–
	(4272)	(4272)	(2531)	(3309)	(2928)	(2462)	(2479)	(4272)
Serum albumin, g/dL	4.21 ± 0.44	4.22 ± 0.42	0.04 ± 0.24†	0.01 ± 0.30	0.04 ± 0.30†	0.04 ± 0.31†	0.02 ± 0.34	0.02 ± 0.34	–
	(2343)	(2343)	(1410)	(1726)	(1505)	(1239)	(1257)	(2343)
Serum Cr, mg/dL	0.78 ± 0.27	0.80 ± 0.29	0.04 ± 0.10†	0.03 ± 0.12†	0.01 ± 0.11†	0.01 ± 0.11	0.01 ± 0.12†	0.02 ± 0.13†	–
	(4889)	(4889)	(2947)	(3745)	(3323)	(2743)	(2796)	(4889)
BUN, mg/dL	17.2 ± 5.5	18.2 ± 5.8	0.8 ± 4.0†	0.8 ± 4.4†	0.8 ± 4.1†	0.9 ± 4.2†	1.1 ± 4.3†	1.0 ± 4.6†	–
	(4054)	(4054)	(2485)	(3048)	(2716)	(2272)	(2302)	(4054)
eGFR, ml/min/1.73 m^2	69.7 ± 19.3	68.8 ± 20.8	−3.1 ± 10.3†	−2.6 ± 9.6†	−0.9 ± 9.7†	−0.2 ± 13.1	−0.6 ± 12.2	−1.0 ± 11.3†	–
	(4889)	(4889)	(2947)	(3745)	(3323)	(2743)	(2796)	(4889)

*Data from 52-week study (Kashiwagi A, et al. Japanese Pharmacology & Therapeutics 2015; 43: 85–100)

†p < 0.001 vs. before administration (paired t-test).

Values are presented as the mean ± standard deviation (number of patients).

BMI: body mass index; BUN: blood urea nitrogen; DBP: diastolic blood pressure; eGFR: estimated glomerular filtration rate; FPG: fasting plasma glucose; HbA1c: glycated hemoglobin; SBP: systolic blood pressure; Cr: creatinine.

Glucose: 1 mmol/l = 18.02 mg/dL.

Albumin: 1 μmol/l = 689.66 mg/dL.

Creatinine: 1 μmol/l = 0.01 mg/dL.

Urea nitrogen: 1 mmol/l = 2.80 mg/dL.

4. Discussion

The safety of SGLT2 inhibitors in Japanese patients with T2DM has emerged as an important issue owing to the high incidence of serious AEs, especially urogenital infections, hypoglycemia, dehydration, and skin complications, that were reported in the first 3 months of using ipragliflozin or other SGLT2 inhibitors [11]. Here, we report the final results of a PMS, which was conducted to investigate the safety of ipragliflozin administered for up to 1 year in elderly Japanese patients with T2DM.

We focused on elderly patients because many patients with T2DM in Japan are elderly, but these patients are often excluded from preapproval clinical trials, so the safety profile of ipragliflozin in this patient population is relatively unclear. To our knowledge, very few studies have enrolled patients aged >80 years. Therefore, our results should be beneficial to societies with an aging population, not just Japan. In addition, ours is one of the largest PMS of SGLT2 inhibitors in Japan.

To put the present results into context with currently available data, we compared our findings with those reported in the preapproval clinical trials. In doing so, we must consider the differences in the design of our survey with those of the clinical trials, and that there were some differences in the baseline characteristics of patients. The differences in baseline characteristics are unsurprising because the present survey focused on older patients, who likely have longer-standing diabetes and frequently other comorbidities to consider when managing their diabetes.

4.1. Safety evaluation

In the final analysis, we documented 1880 ADRs in 1438/8505 patients (16.91%) treated with ipragliflozin for up to 1 year. This compares with 898 ADRs in 721/7170 patients (10.06%) in the interim report and 887 ADRs in 549/1669 patients (32.89%) in the preapproval clinical trials.

The higher incidence of ADRs in the final analysis may reflect the greater number of patients with available data than in the interim analysis. The lower incidence in this PMS relative to the preapproval clinical trials might be due to a number of factors, including differences in reporting standards. Nevertheless, the data provide insight into the likely risk and types of ADRs in elderly patients prescribed ipragliflozin in routine clinical practice. Although there is a risk of reporting bias with respect to AEs in a study of this length, it is worth noting that the presence/absence of AEs was verified with physicians after receipt of all survey sheets. The continued high awareness of the safety of SGLT2 inhibitors among physicians, owing to the recommendations for appropriate use issued by a committee of experts on two occasions during the survey period [11] would also have served to mitigate this risk. Although the results of the final analysis are consistent with those of the interim analysis, some of the present results warrant particular mention.

In the preapproval studies, the majority of skin complications were reported as eczema, whereas drug eruption was more common in this PMS. We believe that this difference was driven by increasing awareness of the SGLT2 inhibitor-related skin complications, especially generalized rash and skin eruption, reported in the first 3 months after the launch of ipragliflozin in Japan [11]. Nevertheless, the overall incidence of skin complications in the present survey was comparable with that in the preapproval studies, which were not limited to elderly patients.

There were no cases of Stevens–Johnson syndrome or toxic epidermal necrolysis reported in the current survey. Several severe skin complications, such as Stevens–Johnson syndrome and exfoliative dermatitis, are listed in the warning section of the package insert for the DPP-4 inhibitor sitagliptin. The higher incidence of these serious events in patients treated with sitagliptin could be due to the prescription of sitagliptin and other DPP-4 inhibitors; SGLT2 inhibitors account for 5.2% of all oral antidiabetic drugs prescribed in Japan whereas DPP-4 inhibitors are much more widely prescribed (67.0%) [13].

Volume depletion-related ADRs were more frequent in the preapproval trials (5.8% of patients) than in the present PMS (3.13%). We might have expected a higher frequency of volume depletion-related ADRs in the present PMS because physicians were asked to report whether any ADR was possibly related to dehydration or a reduction in body fluid.

Another important finding is that 20 ADRs (0.24%) were reported as cerebral infarction, and 18 (0.21%) of these ADRs classified as serious. Cerebral vascular disease was reported as a complication in 604 patients (7.1%), and 8 of these patients developed cerebrovascular disease during the PMS. In the present PMS, cerebral infarction or transient ischemic attack occurred 3 days to 1 year after starting ipragliflozin.

Seventeen deaths were reported in the final analysis. Most of these patients were using multiple classes of concomitant drugs, not just antidiabetic drugs. The relationship between these deaths and the use of ipragliflozin was recorded as unknown or unlikely.

4.2. Risk factors for ADRs

In the interim analysis, we performed univariate analyses to identify factors associated with ADRs of special interest. To extend these findings, and considering the larger number of patients with available data, we repeated the analyses and also conducted multivariable logistic regression analysis, which included variables with p-values of <0.15 in univariate analyses. The objectives of these analyses were to identify possible subgroups of patients at higher risk of certain ADRs, and elucidate potential contraindications.

Hypoglycemia is a clinically important ADR, especially among elderly patients, that may limit attempts to achieve tight glycemic control or might increase the risk of other severe outcomes, especially in elderly patients [14–17]. In the present study, we noted that low BMI (<18.5 kg/m²) and the use of insulin were potential risk factors for hypoglycemia. An inverse association between BMI and hypoglycemia has been reported in Chinese patients [18], and this association might be related to altered glucose metabolism and low circulating glucose concentrations in very lean patients.

Volume depletion-related AEs are among the most common types of AEs in patients treated with SGLT2 inhibitors and the risk of these AEs might increase with age. In the present survey, age ≥75 years and concomitant diuretic use were risk factors for volume depletion-related ADRs. The relationship between diuretic use and volume depletion was expected because higher doses of loop diuretics may cause hypovolemia in patients with renal insufficiency [19], and volume depletion is a frequent adverse reaction to diuretics [20,21]. These findings are consistent with those of a pooled analysis of patients treated with 100 mg canagliflozin, 300 mg canagliflozin, or placebo/active comparators [22], where volume depletion-related AEs occurred in 4.9%, 8.7%, and 2.6% of 8949 patients aged ≥75 years, respectively, compared with 2.2%, 3.1%, and 1.4% of 490 patients aged <75 years. The risk of volume depletion-related AEs in canagliflozin-treated patients was higher in patients aged ≥65 years than in younger patients [23].

Genital infection was associated with female sex and high BMI (≥25 kg/m²) in this PMS. Several prior studies have demonstrated higher incidences of genital infections in women [24,25]. It has been proposed that women with diabetes are at greater risk of genital infection owing to higher glucose levels in the vulvovaginal tissues than occurs in non-diabetic women [26]. The likelihood of genital infection in women using an SGLT2 inhibitor may be increased further owing to the increase in urinary glucose excretion and the deposition of urine on the vulvovaginal tissues during voiding [25].

Genital infection was also associated with female sex and BMI in dapagliflozin-treated patients in a pooled analysis of 13 placebo-controlled trials [27]. Rudofsky et al. concluded that the ‘observed trend is unlikely to be clinically relevant and there is currently no need to adjust advice about the risk of genital infections in overweight or obese patients versus lean patients.’ Considering that this trend has been observed for two different SGLT2 inhibitors, further studies might be warranted to better understand the clinical relevance, if any, and putative mechanism involved in the association between genital infection and BMI in women.

Urinary tract infection, like genital infection, was more common in females, similar to prior studies of dapagliflozin [28] and canagliflozin [29], and this association is probably related to increased urinary glucose excretion in patients treated with SGLT2 inhibitors. We also found that the urinary tract infection was also associated with the duration of diabetes. Intriguingly, a prior study of postmenopausal women showed that urinary tract infection in diabetic women was associated with the duration and severity of diabetes, but not with recent glucose control (in terms of HbA1c). The authors noted that the association could not be explained by confounding due to frequency of sexual intercourse, history of urinary tract infection, postvoid residual bladder volume, ethnicity, or asymptomatic bacteriuria. Furthermore, there was no linear trend between HbA1c and the risk of urinary tract infection.

Finally, we found a significant association between skin complications and female sex. The mechanism underlying this association is elusive and we are unaware of other articles describing an association between skin complications and female sex.

The results of the present study should be considered in the light of its limitations, namely, having incomplete data on the entire cohort, possible reporting biases, and the limited generalizability of the findings.

5. Conclusions

In conclusion, this 1-year PMS has demonstrated the safety profile, in terms of the ADRs that might occur when prescribing ipragliflozin to elderly Japanese patients with T2DM. This PMS is important in that it provides evidence regarding the use of ipragliflozin in real-world clinical practice and there were no safety concerns that were not previously observed in the preapproval clinical trials. The most common types of ADRs were related to skin complications, volume depletion, polyuria/pollakiuria, genital infection, urinary tract infection, and renal disorders. Overall, 38.99% of ADRs occurred within 30 days and 65.69% within 90 days of starting treatment. The majority of ADRs resolved or were in remission at the time the case report form was completed. The risk factors for ADRs varied among the different categories, and could be rationalized based on the results of prior studies. The observed risk factors for hypoglycemia were BMI (<18.5 kg/m² vs. ≥22.0 to <25.0 kg/m²; OR, 9.356) and concomitant insulin use (vs. no; OR, 4.946), while those for volume depletion were age (≥75 years vs. <75 years; OR, 1.737) and concomitant loop diuretic use (vs. no; OR, 2.105). The results of this PMS should help guide physicians in Japan, and other countries, identify possible treatment-emergent ADRs in patients treated with ipragliflozin.

Declaration of interest

K Yokote has received grants/research support from Astellas Pharma Inc., Daiichi Sankyo Company Limited, Takeda Pharmaceutical Company Limited, Mitsubishi Tanabe Pharma Corporation; honoraria for speakers bureau and consulting fees from Astellas Pharma Inc., AstraZeneca K.K., Daiichi Sankyo Company Limited, Kowa Pharmaceutical Company Ltd., Kyowa Hakko Kirin Co., Ltd., Mitsubishi Tanabe Pharma Corporation, Mochida pharmaceuticals Ltd, Merck Sharp & Dohme K.K., Ono Pharmaceutical Co., Ltd., Pfizer Japan Inc., Shionogi & Co., Ltd, and Takeda Pharmaceutical Company Limited; funded research department from Merck Sharp & Dohme K.K. Y Terauchi has received honoraria for speakers bureau from Astellas Pharma Inc., AstraZeneca K.K., Bayer Yakuhin, Ltd., Daiichi Sankyo Company Limited, Dainippon Sumitomo Pharma Co., Ltd., Eli Lilly Japan K.K., Kissei Pharmaceutical Co., Ltd., Kowa Pharmaceutical Company Ltd., Kyowa Hakko Kirin Co., Ltd., Merck Sharp & Dohme K.K., Mitsubishi Tanabe Pharma Corporation, Nippon Boehringer Ingelheim Co., Ltd., Novartis Pharma K.K., Novo Nordisk Pharma Ltd., Ono Pharmaceutical Co., Ltd., Sanwa Kagaku Kenkyusho Co., Ltd., Sanofi K.K., Shionogi & Co., Ltd., Taisho Toyama Pharmaceutical Co., Ltd., and Takeda Pharmaceutical Company Limited; and grants from Astellas Pharma Inc., AstraZeneca K.K., Bayer Yakuhin, Ltd., Daiichi Sankyo Company Limited, Dainippon Sumitomo Pharma Co., Ltd., Eli Lilly Japan K.K., Kissei Pharmaceutical Co., Ltd., Kowa Pharmaceutical Company Ltd., Kyowa Hakko Kirin Co., Ltd., MSD K.K., Mitsubishi Tanabe Pharma Corporation, Nippon Boehringer Ingelheim Co., Ltd., Novartis Pharma K.K., Novo Nordisk Pharma Ltd., Ono Pharmaceutical Co., Ltd., Pfizer Japan Inc., Sanwa Kagaku Kenkyusho Co., Ltd., Sanofi K.K., Shionogi & Co., Ltd, Taisho Toyama Pharmaceutical Co., Ltd., and Takeda Pharmaceutical Company Limited. Ichiro Nakamura and Haruko Sugamori are employees of Astellas Pharma Inc. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. Medical writing assistance and editorial support was funded by Astellas and provided by Dr. Nicholas D. Smith of Edanz Group Ltd., and by ELMCOM™.

Acknowledgments

We would like to thank the study investigators and participating patients at the 2634 contracted institutions.

Additional information

Funding

This study was sponsored by Astellas Pharma Inc., Japan.