Title: Concurrent use of oral anticoagulants and sulfonylureas in individuals with type 2diabetes and risk of hypoglycaemia: A UK population-based cohort study.

Type of Manuscript: Original Article

Running Title: Oral anticoagulants and sulfonylureas and risk of

Abstract

OBJECTIVE: To investigate the association of concurrent use of oral anticoagulants (OACs) and sulfonylureas andthe risk of hypoglycaemia in individuals with type 2 diabetes mellitus (T2DM). RESEARCH DESIGN ANDMETHODS: A retrospective cohort study was conducted between (2001-2017) using The Health ImprovementNetwork (THIN) database. Individuals with T2DM who received OAC prescription and sulfonylureas wereincluded. We compared the risk of hypoglycaemia with sulfonylureas and OACs using propensity score matchingand Cox regression. RESULTS: 109,040 individuals using warfarin and sulfonylureas and 77,296 using DOACs andsulfonylureas were identified and included. There were 285 hypoglycaemia events in the warfarin withsulfonylureas group (incidence rate = 17.8 per 1000 person-years), while in the sulfonylureas only, 304hypoglycaemia events were observed (incidence rate = 14.4 per 1000 person-years). There were 14hypoglycaemic events in the DOACs with sulfonylureas group (incidence rates = 14.8 per 1000 person-years),while in the sulfonylureas alone group, 60 hypoglycaemia events were observed (incidence rate =23.7 per 1000person-years). Concurrent use of warfarin and sulfonylureas was associated with increased risk of hypoglycaemiacompared with sulfonylureas alone, (HR 1.38; 95% CI 1.10–1.75). However, we found no evidence of anassociation between concurrent use of DOACs and sulfonylureas and risk of hypoglycaemia (HR 0.54; 95% CI,0.27–1.10) when compared with sulfonylureas only. CONCLUSIONS: We provide real-world evidence of possibledrug-drug interactions between warfarin and sulfonylureas. The decision to prescribe warfarin with coexistentsulfonylureas to individuals with T2DM should be carefully evaluated in the context of other risk factors ofhypoglycaemia,

Editorial membersDiabetes and Metabolism Journal Editorial Office101-2104, Lotte Castle President, 109 Mapo-daero, Mapo-gu, Seoul 04146, KoreaTEL: +82-2-714-9064FAX: +82-2-714-9084E-mail : diabetes@kams.or.krWebsite: http://submit.e-dmj.org/

101-2104, Lotte Castle President, 109 Mapo-daero, Mapo-gu, Seoul 04146, Korea

Copyright© Korean Diabetes Association.

Diabetes

and Met

aboli

sm Jo

urna

l

1

ABSTRACT (257 words) 1

OBJECTIVE: To investigate the association of concurrent use of oral anticoagulants (OACs) 2

and sulfonylureas and the risk of hypoglycaemia in individuals with type 2 diabetes mellitus 3

(T2DM). 4

RESEARCH DESIGN AND METHODS: A retrospective cohort study was conducted 5

between (2001-2017) using The Health Improvement Network (THIN) database. Individuals 6

with T2DM who received OAC prescription and sulfonylureas were included. We compared 7

the risk of hypoglycaemia with sulfonylureas and OACs using propensity score matching and 8

Cox regression. 9

RESULTS: 109,040 individuals using warfarin and sulfonylureas and 77,296 using DOACs 10

and sulfonylureas were identified and included. There were 285 hypoglycaemia events in the 11

warfarin with sulfonylureas group (incidence rate = 17.8 per 1000 person-years), while in the 12

sulfonylureas only, 304 hypoglycaemia events were observed (incidence rate = 14.4 per 1000 13

person-years). There were 14 hypoglycaemic events in the DOACs with sulfonylureas group 14

(incidence rates = 14.8 per 1000 person-years), while in the sulfonylureas alone group, 60 15

hypoglycaemia events were observed (incidence rate =23.7 per 1000 person-years). Concurrent 16

use of warfarin and sulfonylureas was associated with increased risk of hypoglycaemia 17

compared with sulfonylureas alone, (HR 1.38; 95% CI 1.10–1.75). However, we found no 18

evidence of an association between concurrent use of DOACs and sulfonylureas and risk of 19

hypoglycaemia (HR 0.54; 95% CI, 0.27–1.10) when compared with sulfonylureas only. 20

CONCLUSIONS: We provide real-world evidence of possible drug-drug interactions between 21

warfarin and sulfonylureas. The decision to prescribe warfarin with coexistent sulfonylureas to 22

individuals with T2DM should be carefully evaluated in the context of other risk factors of 23

hypoglycaemia, and availability of alternative medications. 24

Diabetes

and Met

aboli

sm Jo

urna

l

2

Keywords: Diabetes mellitus, Oral anticoagulants, Drug-drug interactions, Sulfonylureas 25

BACKGROUND 26

Individuals with type 2 diabetes mellitus (T2DM) are often suffering from cardiovascular 27

complications (1). Cardiovascular diseases (CVDs) such as atrial fibrillation (AF), venous 28

thromboembolism, and stroke are amongst the leading causes of mortality worldwide (2). 29

Individuals with T2DM are initially treated with metformin therapy to control the blood 30

glucose levels. However, for many individuals with T2DM, this type of treatment fails to 31

control blood glucose levels and they are often prescribed additional or alternative therapies 32

(3). The most common second-line therapy for T2DM is sulfonylureas (4). 33

T2DM individuals are at an increased risk of developing adverse drug reactions due to their 34

multiple comorbidities, renal impairment and multiple uses of drugs (5). In a systematic review, 35

reporting adverse drug reaction (ADR)s hospitalization worldwide, Hamid and colleagues 36

reported that 33.9% and 9% of all ADRs were due to cardiovascular and antidiabetic 37

medications (6). Similarly, the United States health and human service department reported 38

that “ Warfarin accounts for about 33% of all hospitalization due to adverse drug events (ADEs), 39

while it is 10.7% for oral hypoglycaemic agents” (7). 40

Oral anticoagulant medications (OACs) including warfarin and direct oral anticoagulants 41

(DOACs) are widely prescribed for the prevention and treatment of CVDs (8). However, their 42

use may be considered a public health concern, especially due to their high prescription 43

frequency, high probability of drug-drug interactions, and their association with serious 44

adverse events, such as bleeding and hypoglycaemia (9). In the last few years, DOACs have 45

been introduced for the treatment of different CVDs, and many guidelines recommend them as 46

a first-line therapy for the management of CVDs (8, 10). DOACs have several advantages 47

over warfarin including: less drug-drug interactions (due to their predominantly renal 48

Diabetes

and Met

aboli

sm Jo

urna

l

3

metabolism), and predictable pharmacokinetics and pharmacodynamics. However, they are 49

expensive and there is no standardised test to monitor blood levels (11). 50

Hypoglycaemia is an expected dose-related complication of sulfonylureas (12), which can be 51

potentiated by several risk factors including drug-drug interaction via inhibition of hepatic 52

cytochrome P450 (CYP) enzymes which are responsible for the metabolism of sulfonylureas 53

(13). Drug-drug interactions resulting in the risk of hypoglycaemia are widely documented in 54

the literature (14). 55

Previous pre-clinical studies have reported a possible drug-drug interaction between warfarin 56

and sulfonylureas through the displaced plasma protein binding and hepatic metabolism 57

CY450 (15). In addition, two case reports, showed a possible drug-drug interaction between 58

warfarin and glibenclamide (16, 17). Two previous studies reported a significant association 59

between warfarin and sulfonylurea (18, 19), however, both studies used data only from the 60

United States. 61

Previous studies have demonstrated that the prevalence of AF in individuals with T2DM 62

ranges from 8% to 14.9% and that individuals with T2DM have 40% higher risk of 63

developing AF compared with individuals without T2DM (20, 21). Given the fact that 64

diabetes and AF are highly prevalent, and that antidiabetic and anticoagulant medications are 65

largely prescribed concomitantly and the possibility of developing drug-drug interactions (9, 66

22, 23), and the lack of published evidence, we aimed to investigate the association between 67

concurrent OAC and sulfonylurea use and risk of hypoglycaemia. 68

RESEARCH DESIGN AND METHODS 69

Study design 70

This was a population-based cohort study in individuals with T2DM from 2001 to 2017. 71

Diabetes

and Met

aboli

sm Jo

urna

l

4

Data source 72

The Health Improvement Network (THIN) database, now known as IQVA Medical Research 73

Data (IMRD) UK Database was used for this study. THIN is a UK primary care database 74

containing anonymised administrative, clinical, and prescribing data from over 587 practices 75

with more than 13 million individuals (24, 25). THIN is one of the largest sources for primary 76

care data in the UK and has been validated for epidemiological research purposes (22, 23, 25). 77

It holds data on personal information, health related behaviours, and diagnoses information 78

which is recorded and identified using Read codes. THIN contains records of prescriptions 79

issued only by GPs and recorded in the individual’s records (24, 25). Data from practices that 80

met the acceptable mortality reporting (AMR) measures of quality assurance for THIN data 81

were used in this study (26). 82

This study was reviewed, and scientific approval was obtained by THIN SRC in 2018 83

(18THIN046). The research was reported in accordance with strengthening the reporting of 84

observational studies in epidemiology (STROBE) Statement. 85

Study cohort 86

The study population included people aged at least 18 years old, with a recorded diagnosis of 87

T2DM. To ensure accurate measurement of medical history, we included individuals only if 88

they had an observation period of at least 12 months before the first T2DM diagnosis and were 89

registered with the general practice during the study period. Individuals were followed up until 90

the end of September 2017 and were censored if they experienced the outcome of interest, died, 91

left their general practice during the study period, or stopped medications during the 92

overlapping period of both medications, whichever came first. 93

The study covered a period of 17 calendar years (2001-2017), and two separate but similar 94

analyses were carried out. The first and second analyses comprised individuals with T2DM 95

Diabetes

and Met

aboli

sm Jo

urna

l

5

who received at least one prescription for one of the OACs of interest (including warfarin 96

(Analysis 1) or DOACs (Analysis 2)) and sulfonylureas and were identified using drug codes 97

recorded in THIN. The index date (start date) was defined as the date on which users were first 98

co-exposed (concurrently) to both medications (OACs and sulfonylureas), regardless of which 99

medication was first. Individuals with records of prescriptions of both medications during the 100

follow up were included in the study. The index date was defined as: a) if sulfonylurea 101

prescription was between the first and last warfarin prescriptions then the date of sulfonylurea 102

was accounted as the index date) if warfarin prescription was between the first and last 103

sulfonylurea prescriptions then the date of warfarin was accounted as the index date. 104

Individuals who did not meet these criteria were categorised as ‘no overlap’ and excluded. We 105

also excluded individuals if they had less than 12 months observation period, were <18 years 106

old, were diagnosed with malignancy or metastatic tumours (as they are at higher risk of 107

glucose level fluctuations due to anticancer therapy) and individuals with incomplete data for 108

transfer out or death data. Details of the exclusion criteria are provided in figures 1-2. 109

Exposure definition 110

The exposure of interest was person-time concomitantly exposed to OACs and sulfonylureas 111

after the diagnosis of T2DM. OACs included warfarin and DOACs (dabigatran, apixaban, 112

rivaroxaban and edoxaban). Individuals with records of acenocoumarol or phenindione only 113

were not included, because of the very low number of individuals and because it is very 114

unlikely to be prescribed in the UK. Sulfonylureas were of a second generation and included: 115

gliclazide, glipizide, glyburide and glimepiride. For analysis 1, we compared individuals using 116

warfarin and sulfonylureas concurrently, versus individuals using sulfonylureas only, and the 117

index date for this group was the first prescription of sulfonylureas). For analysis 2, we 118

compared individuals using DOACs and sulfonylureas concurrently, versus individuals using 119

sulfonylureas only, and the index date for this group was the first prescription of sulfonylureas. 120

Diabetes

and Met

aboli

sm Jo

urna

l

6

only individuals who were new users from 2011 onwards were included in analysis 2, as 121

DOACs were first approved on the market from 2011. 122

Study outcomes 123

The study outcome was incident hypoglycaemia during the follow up time defined as the first 124

record of hypoglycaemia after the index date. 125

Study covariates 126

Demographics, comorbidities, and medications associated with developing hypoglycaemia 127

were included as covariates, based on previous studies (18, 19, 27) and clinician 128

recommendation. They included; age, gender, smoking status (never-smoked, ex-smoker and 129

current smoker), body mass index (BMI), alcohol consumption (never-drink, ex-drinker and 130

current drinker), Townsend deprivation score, CVDs disease, history of hyperglycaemia, 131

hyperlipidaemia, hypertension (HTN), AF, stroke/ transient ischaemic attack (TIA), deep vein 132

thrombosis (DVT), peripheral vascular disease (PVD), chronic kidney disease (CKD), liver 133

disease, anxiety, depression, chronic obstructive pulmonary disease (COPD), use of multiple 134

antidiabetics (intensification), beta-blockers (BBs), angiotensin converting enzyme inhibitor 135

(ACEIs), angiotensin II receptor blocker inhibitor (ARBs), calcium channels blockers (CCBs), 136

statins, corticosteroids, , antiplatelets. Information for comorbidities was evaluated at any time 137

on/or before the index date and was identified using the Read codes. Medications were 138

identified using the drug codes within 180-days on/or before the start date. We did not account 139

for international normalised ratio (INR) results because the data was not complete. Details of 140

the Read codes are provided in supplement. 141

Propensity score matching 142

Diabetes

and Met

aboli

sm Jo

urna

l

7

To minimise potential bias due to non-randomised, a propensity score matching cohort (within 143

±0.05, comprising 1:1 controls were created (28). To address confounding by indication, when 144

individuals with T2DM with more severe illness are likely to be treated with insulin, 145

sulfonylurea or multiple antidiabetic medications or individuals with T2DM using OACs might 146

be prescribed OACs for different indications, we included the following variables (AF, DVT 147

and stroke/TIA) in the PS model (29). We used a logistic regression model to estimate the 148

propensity score for each individual and variables listed in the previous section were included. 149

Absolute Standardised Differences (ASD) for all baseline variables were calculated to assess 150

the differences in individuals’ characteristics and covariates balance between treatment groups 151

before and after the propensity score matching. We used a cut-off point of 0.2 ASD (30-32). 152

Therefore, characteristics with an ASD greater than 0.2 after propensity score matching were 153

included as covariates in the subsequent regression model. 154

Statistical analysis 155

Individuals characteristics were presented as number (percentage) for categorical variables and 156

as mean (±SD) for continuous variables. Crude incidence rates were calculated by dividing 157

the number of hypoglycaemia events by person-time at risk and were expressed as rates per 158

1000 person-years. Person-year was calculated as the time from the index date until the end of 159

follow up. The Cox proportional hazard model before and after propensity score was used to 160

estimate the time to an event, and to investigate the association between the use of OACs with 161

co-existent sulfonylurea and the risk of hypoglycaemia; results were presented as HR with 95% 162

confidence interval (CI). We performed Kaplan-Meier survival curves plots on the matched 163

datasets to compare outcomes between the cohorts over time. The hazard assumption was 164

examined by both visual graphs and by applying tests using ph statement (proc PHREG) in 165

SAS statistical software. All analysis was performed using SAS version 9.4 (SAS Institute). 166

Diabetes

and Met

aboli

sm Jo

urna

l

8

Sensitivity analysis 167

To confirm the robustness of our results we conducted five sensitivity analyses. First, we re-168

analysed the data taking into account subsequent changes in the exposure status. Individuals 169

were censored if they stopped the treatment. The gap between expected prescription end date 170

and the start date of any subsequent prescription were no more than 90 days. 171

Second, we repeated the analysis using the IPTW method to address the risk difference between 172

groups at baseline (33). The stabilised IPTW was calculated by multiplying the IPTW by the 173

marginal probability of receiving the actual treatment (33). Stabilisation was used to reduce the 174

variability of the IPTW weights (34). The predictor variables inserted in the IPTW models 175

included the same covariates as in the PS matching. PS trimming for the IPTW was also 176

conducted. We re-analysed the IPTW based on (1st percentile of the PS distribution in exposure 177

and the 99th percentile of the PS in non-exposed treated) (35). Third, the missing data on 178

smoking status, BMI, alcohol consumption and Townsend scores were imputed by the multiple 179

imputation (MI) method (36, 37). All PS covariates, the outcome and survival time were 180

included in the MI model, and 25 imputed datasets were generated, analysed separately and 181

then combined using Rubin’s rule. 182

Patient and public involvement 183

It was not appropriate or possible to involve individuals or the public in the design, or conduct, 184

or reporting, or dissemination plans of this research. 185

Data and Resource Availability 186

Data access is through permission from THIN only. 187

RESULTS 188

Individuals characteristics 189

Diabetes

and Met

aboli

sm Jo

urna

l

9

A total of 418,613 individuals with T2DM were identified, of whom only 178,084 received a 190

prescription for sulfonylureas and OACs at some point during the study period between 2001 191

and 2017. We excluded 49,317 (22%) individuals in the warfarin group (Analysis 1) and 192

16,413 (17.5%) individuals in the DOACs group (Analysis 2) for missing data in the main 193

analysis. Finally, after we applied exclusion criteria, 109,040 using warfarin and sulfonylureas 194

(Analysis 1) and 77,296 using DOACs and sulfonylureas (Analysis 2) were included for the 195

analyses. Details of the identification of the study cohort, including the study cohort of 196

individuals included in Analyses 1 and 2 are presented in Figures 1-2. 197

At baseline, before propensity score matching, compared to individuals who received 198

sulfonylureas only, individuals who received warfarin with sulfonylureas were older (mean 199

age: 73.3 vs 61.0), had a higher cardiovascular profile: CVDs (14.7% vs 4.5%), HTN (68% vs 200

50%), stroke/TIA (19% vs 5.7%), AF (61% vs 1.8%), and DVT (21% vs 2%). At least 50% of 201

the study population had a BMI ≥30, with nearly similar BMI ratios in individuals who received 202

warfarin with sulfonylureas compared to individuals who received sulfonylureas only. 203

However, gender and the mental health profile including depression and anxiety showed little 204

difference between the two groups (61% females vs 57%females, 20% vs 23% and 13% vs 205

15%, respectively). After matching, all baseline individuals’ characteristics had standardised 206

differences less than 0.2 (Table 2). Details of the study characteristics, including individuals’ 207

characteristics of Analysis 2 (DOACs with sulfonylureas vs sulfonylureas), are presented in 208

Table S1. 209

Hypoglycaemia 210

Warfarin with sulfonylureas vs sulfonylureas 211

There were 578 hypoglycaemia events in the warfarin and sulfonylureas group (crude 212

incidence rates = 16.7 per 1000 person-years), while 7307 hypoglycaemia events were recorded 213

Diabetes

and Met

aboli

sm Jo

urna

l

10

in the sulfonylureas only group with a total follow up of 526,422.49 person-years, (crude 214

incidence rates = 13.8 per 1000 person-years). The risk of developing hypoglycaemia was 215

higher for individuals receiving warfarin with sulfonylureas compared to individuals receiving 216

sulfonylureas alone (HR 1.20; 95% CI, 1.10 – 1.30), p<0,0001), Table 3. 217

DOACs with sulfonylureas vs sulfonylureas 218

Individuals using DOACs with sulfonylureas concomitantly contributed to a total of 957 219

person-years, during which 14 hypoglycaemic events were recorded (crude incidence rates = 220

15.0 per 1000 person-years), while in the (sulfonylureas only group, a total of 246,345.65 221

person-years, with4,514 hypoglycaemia events were recorded (crude incidence rates = 18.0 per 222

1000 person-years). The risk of developing hypoglycaemia was lower for individuals receiving 223

DOACs with sulfonylureas compared to individuals receiving sulfonylureas alone. However, 224

this was not statistically significant (HR 0.66; 95% CI, 0.39 – 1.11, p=0.118), Table 3. 225

Propensity score matched analysis 226

Warfarin with sulfonylureas vs sulfonylureas 227

After matching, 5,379 individuals were included in each group. There were a total of 15,959.04 228

of concomitant exposure and, 285 hypoglycaemia events in the warfarin with sulfonylureas 229

group (incidence rates = 17.8 per 1000 person-years), while users of sulfonylureas only 230

contributed to 21,028.52 person-years, during which 304 hypoglycaemia events were observed 231

(incidence rates = 14.4 per 1000 person-years). The risk of hypoglycaemia was 38% higher in 232

individuals with concomitant use of warfarin with sulfonylureas, compared to sulfonylureas 233

alone users (HR 1.38; 95% CI, 1.10 – 1.76, P= 0.010), Table 4. Kaplan-Meier curves for the 234

incidence of hypoglycaemia of analysis 1 are shown in Figure 3. 235

DOACs with sulfonylureas vs sulfonylureas 236

Diabetes

and Met

aboli

sm Jo

urna

l

11

A total of 1,027 in each group were included in the analysis, with a total of 942, and 2,532 237

person-years were recorded, 14 and 60 hypoglycaemic events happened in the DOACs with 238

sulfonylureas and sulfonylureas alone groups, respectively (incidence rates =14.8 per 1000 239

person-years and 23.7 per 1000 person-years, respectively), Table 4. The risk of developing 240

hypoglycaemia was again lower for individuals receiving DOACs with sulfonylureas 241

compared to individuals receiving sulfonylureas alone (HR 0.54; 95% CI, 0.27 – 1.10, 242

P=0.091). However, this was not statistically significant (Table 4). Kaplan-Meier curves for 243

the incidence of hypoglycaemia of analysis 2 are shown in Figure 4. 244

Sensitivity analysis 245

The results of the sensitivity analyses are presented in Tables S2-4. When re-analysing the data 246

taking into account only individuals who received subsequent prescriptions within no more 247

than 90-days, the risk of hypoglycaemia was higher in individuals receiving warfarin with 248

sulfonylureas compared to individuals receiving sulfonylurea alone (HR 1.14; 95% CI, 1.01 – 249

1.30, P= 0.032), Table S2. However, the results of the matching analysis based on 90-days 250

grace period showed non-statistically significant results between both groups (HR 1.10; 95% 251

CI, 0.71 – 1.22, P= 0.634), Table S2. 252

Results from the inverse probability of treatment weighting (IPTW) were similar to main 253

analyses. Individuals had a higher risk of hypoglycaemia by 24% when receiving warfarin with 254

sulfonylureas compared to individuals receiving sulfonylureas alone (HR 1.24; 95% CI, 1.20- 255

1.25, P <0.0001) (Table S3). The results of IPTW were different from main analysis when 256

comparing individuals receiving DOACs and sulfonylureas concomitantly against individuals 257

receiving sulfonylureas alone. The risk of hypoglycaemia was again lower for individuals 258

receiving DOACs and sulfonylureas, but significant in the IPTW analysis (HR 0.56; 95% CI, 259

Diabetes

and Met

aboli

sm Jo

urna

l

12

0.36 – 0.90, P= 0.011). Results from the inverse probability of treatment weighting (IPTW), 260

including PS trimming, are presented in Table S3. 261

Results from multiple imputations were also similar to the main analyses, demonstrating 262

warfarin with sulfonylurea treatment was associated with a higher risk of hypoglycaemia 263

compared to sulfonylureas alone (HR=1.08; 95% CI, 1.03 – 1.13; p=<0.001), Table S4. 264

DISCUSSION 265

In this large population-based study in UK primary care, we investigated the association 266

between the concurrent use of OACs and sulfonylureas and the risk of hypoglycaemia in 267

individuals with T2DM. This study found that warfarin was associated with 38% increased risk 268

of hypoglycaemia when used with sulfonylureas concurrently in individuals with T2DM, (HR 269

1.38; 95% CI, 1.10 – 1.75). We found no evidence of an association between the use of DOACs 270

and sulfonylureas concurrently and the risk of hypoglycaemia (HR 0.54; 95% CI, 0.27 – 1.10). 271

The study results support the hypothesis that warfarin is associated with an increased risk of 272

hypoglycaemia when given concomitantly with sulfonylureas in individuals with T2DM. 273

Comparison with other studies 274

There are limited studies that explored the safety of the concurrent use of warfarin with 275

sulfonylureas. Previous RCTs focused on younger population with fewer number of 276

comorbidities. While the probability of adverse events was more common among elderly. This 277

could be related to the difference in excretion rates between different age groups which is vital 278

to maintain correct therapeutic levels (38). 279

The findings of increased risk of hypoglycaemia among the use of warfarin with sulfonylureas 280

are consistent with two previous large database studies using the Medicare claims and 281

Medicaid programmes in the United States (18, 19). Romley et al. reported that the concurrent 282

use of warfarin and sulfonylureas was associated with a 22% higher risk for hypoglycaemia 283

Diabetes

and Met

aboli

sm Jo

urna

l

13

(18) in a cohort study of 465,918 individuals. Similarly, in a self‐ controlled case series design, 284

Nam et al. reported an elevated rate of serious hypoglycaemia when warfarin was given 285

concomitantly with sulfonylureas (glipizide, glyburide, glyburide), (RR, 1.72; 95% CI, 1.29 – 286

2.29), (RR, 1.57; 95% CI, 1.15 – 2.15), and (RR, 1.56; 95% CI, 0.97 – 2.50), respectively (19). 287

However, in this study, we included DOAC use and had a longer follow-up than the other two 288

studies. 289

Our findings indicated that there is no evidence of an association found between the use of 290

DOACs and sulfonylureas concurrently and the risk of hypoglycaemia (HR 0.54; 95% CI, 0.27 291

– 1.10). We suggest that our estimate did not reach the significant level due to the small sample 292

size in the compared group and ultimately number if events. DOAC therapy has multiple 293

advantages over warfarin therapy, which include lack of the need for regular monitoring of the 294

degree of anticoagulation and wider therapeutic range (39, 40). 295

Potential mechanisms 296

The underlying mechanism of action for the concurrent use of warfarin with sulfonylureas and 297

increased risk of hypoglycaemia is unclear. Previous pre-clinical hypotheses suggest some 298

potential mechanisms for this association through a drug-drug interaction between the warfarin 299

and sulfonylureas. First, there may be a displaced protein binding mechanism, where 300

interaction between warfarin and sulfonylureas may occur on the site of the protein binding, 301

and thus warfarin may enhance the plasma concentration of sulfonylureas in the blood, and 302

hence increase its activity and risk of hypoglycaemia (15). However, previous studies and 303

reviews have described this mechanism of drug interaction to be overestimated and not to have 304

meaningful clinical effects, as it only applies to data from in vitro studies, or to drugs that are 305

given through loading intravenous doses (41, 42). Second, a drug-drug interaction through 306

inhibition of the cytochrome CYP2C9 hepatic metabolic pathway has also been suggested. 307

Diabetes

and Met

aboli

sm Jo

urna

l

14

Warfarin, glimepiride, and glipizide are all largely metabolised by hepatic cytochrome 308

CYP2C9 (15), and therefore, warfarin may limit the rate at which the sulfonylurea can be 309

metabolised in the liver (43). This mechanism may explain the findings of this study, especially 310

the fact that widely used drug references warn that the concurrent use of warfarin with 311

sulfonylureas may increase the risk of bleeding (44). However, no previous human studies exist 312

to validate this hypothesis, and future studies to investigate this association are needed. 313

Furthermore, pre-clinical studies have suggested that osteocalcin which is one of the important 314

bone proteins produced by the bone (45), is involved in the metabolism of glucose and insulin 315

sensitivity through the process of bone mineralisation and formation, which requires high 316

energy (46). It has been postulated that the uncarboxylated form of osteocalcin enhances the 317

glucose tolerance by the beta cells in the pancreatic islets and increases the insulin sensitivity 318

in peripheral tissues (46). However, this function is mainly dependent on Vitamin K, and 319

therefore, administration of warfarin may block the activation of the carboxylation forum of 320

osteocalcin and thus increase the uncarboxylated forum in the plasma (46). In vitro studies 321

have suggested stimulation in the production of undercarboxylated osteocalcin by insulin 322

through a positive feedback mechanism between the pancreas, adipose tissue and bone, which 323

in turn enhances insulin production and sensitivity (47). 324

Meaning of the study 325

Warfarin therapy is indicated as a prophylaxis and treatment for a wide range of life-threatening 326

health conditions including venous thrombosis, pulmonary embolism, and thromboembolic 327

complications associated with AF and cardiac valve replacement. Additionally, warfarin 328

therapy proven to reduce the risk of death, recurrent MI and thromboembolic events (48). 329

Nonetheless, warfarin therapy needs precise dosing regimen and follow-up by the patients and 330

their healthcare professionals due to its associated potential drug-drug interaction and 331

Diabetes

and Met

aboli

sm Jo

urna

l

15

contraindications. In specific clinical settings, warfarin therapy has higher risk of 332

complications and adverse events, these include hepatic impairment, moderate to severe 333

hypertension, and diabetes mellitus. Given the fact that diabetes and AF are highly prevalent, 334

and that antidiabetic and anticoagulant medications are largely prescribed concomitantly (22, 335

23), this urge us to weight the benefit-risk ratio in the case of the concurrent use of antidiabetic 336

and anticoagulant medications therapy. Several important drug references have warned that 337

warfarin may have some drug interactions when given with sulfonylureas. However, this 338

possible drug interaction has not been studied extensively or appreciated in the literature. This 339

study provides the first evidence for this drug interaction of two widely used medications in 340

the UK, and it is also consistent with the findings from the previous studies in the US. In 341

addition, this study found an insignificant reduced risk of hypoglycaemia when DOACs are 342

used concurrently with sulfonylureas. 343

Implication of the study 344

Individuals with T2DM receiving OACs are older and more susceptible to polypharmacy and 345

drug-drug interactions (Mallet et al., 2007)(22, 23). Doctors and clinical pharmacists must be 346

vigilant when prescribing warfarin with sulfonylureas and must be alert to both immediate and 347

delayed-onset hypoglycaemia when prescribing this drug combination. Clinical surveillance, 348

frequent blood glucose measurements, INR monitoring, diet changes and patient education 349

may be necessary to reduce the risk of hypoglycaemia if individuals are prescribed these 350

medications together (49, 50). 351

Additionally, given that DOACs are widely available nowadays, these medications may be an 352

alternative therapy when OACs and sulfonylureas are indicated in individuals with T2DM. 353

DOACs have a more predictable pharmacokinetic profile and have less drug-drug interactions. 354

Diabetes

and Met

aboli

sm Jo

urna

l

16

This research has also highlighted a possible protective effect of DOACs against 355

hypoglycaemia when prescribed with sulfonylureas; however, the sample size was small, and 356

we did not have a long follow-up time. Therefore, considering these results in the context of 357

the currently available literature, we underline the need for future research with a longer follow-358

up time, and large sample sizes to examine the association of DOACs and sulfonylureas and 359

the risk of hypoglycaemia or bleeding in individuals with T2DM. 360

Strengths and limitations of the study 361

Our study has several strengths. First; this study had a longer follow up period compared to the 362

previous studies (18, 19). Second, this study used the population representative data from the 363

UK primary care database – THIN. It is therefore, reasonable to assume that our findings are 364

generalised and may broadly reflects real world practice in the UK and the world. Third, we 365

used advanced statistical method (i.e. PS matching and IPTW) to address the measured 366

confounder issue at baseline. In addition, to confirm the robustness of our results, we conducted 367

several sensitivity analyses that suggest our results are robust, including multiple prescriptions 368

grace periods, IPWT and multiple imputations. 369

This study has some limitations. First, this study is an observational cohort study, and unlike 370

RCTs, the residual confounding cannot be excluded. Second, due to the emergency nature of 371

the outcomes of interest (hypoglycaemia), we did not have access to hospital data, also, mild 372

hypoglycaemia may not be reported to the doctors, and this could lead to underestimations of 373

the cases. Despite that wide range of confounders were adjusted for in our analyses, which is 374

potentially justified, this could increase the risk of the model fit being affected when there is a 375

small sample size or frequency of outcomes. THIN is an administrative database and therefore, 376

data on medication adherence, the actual ingestion of medications or diet is lacking. However, 377

we tackled this by conducting a sensitivity analysis to account only for prescription refills of 378

fewer than 90 days. 379

Diabetes

and Met

aboli

sm Jo

urna

l

17

CONCLUSION 380

In summary, the study found that warfarin was associated with an increased risk of 381

hypoglycaemia when given concomitantly with sulfonylureas in individuals with T2DM. 382

Concurrent use of DOACs with sulfonylureas however was associated with an insignificant 383

reduced risk of hypoglycaemia. The decision to prescribe warfarin with coexistent 384

sulfonylureas to individuals with T2DM should be carefully evaluated in the context of other 385

risk factors of hypoglycaemia, and the availability of alternative medications. Future studies 386

are needed to validate the finding of the protective effect of DOACs and sulfonylureas on 387

hypoglycaemia on larger sample size and longer follow-up periods. 388

Funding. 389

Alwafi’s PhD project was supported by a scholarship from the Saudi Arabian Ministry of 390

Higher Education. 391

Conflict of interest 392

The authors declare no conflict of interest. 393

Author Contributions. 394

HA, IW and LW had full access to all the data in the study and take responsibility for the 395

integrity of the data and the accuracy of the data analysis. Concept and design: HA, IW and 396

LW. Acquisition, analysis, or interpretation of data: HA, IW, AB, PM, CW and LW. Drafting 397

of the manuscript: HA, IW, AB, AYN, CW and LW. Critical revision of the manuscript for 398

important intellectual content: All authors. Statistical analysis: HA, PM and LW. 399

Administrative, technical, or material support: HA. Study supervision: IW, CW and LW. 400

Ethics approval. 401

Diabetes

and Met

aboli

sm Jo

urna

l

18

The present study is based on anonymised and unidentifiable THIN data; thus, the need for 402

informed consent was waived by the THIN scientific review committee (SRC). This study 403

was reviewed and scientific approval was obtained by THIN SRC in 2018 (18THIN046). 404

References 405

1. Celis-Morales CA, Petermann F, Hui L, Lyall DM, Iliodromiti S, McLaren J, et al. 406

Associations Between Diabetes and Both Cardiovascular Disease and All-Cause Mortality 407

Are Modified by Grip Strength: Evidence From UK Biobank, a Prospective Population-408

Based Cohort Study. Diabetes care. 2017;40(12):1710-8. 409

2. Nichols M, Townsend N, Scarborough P, Rayner M. Cardiovascular disease in 410

Europe 2014: epidemiological update. European heart journal. 2014;35(42):2950-9. 411

3. National Institute for Health and Care Excellence. Type 2 diabetes in adults: 412

management.NICE guidline.[NG28]. 2015. 413

4. Sharma M, Nazareth I, Petersen I. Trends in incidence, prevalence and prescribing in 414

type 2 diabetes mellitus between 2000 and 2013 in primary care: a retrospective cohort study. 415

BMJ open. 2016;6(1):e010210. 416

5. Mangoni AA, Jackson SH. Age-related changes in pharmacokinetics and 417

pharmacodynamics: basic principles and practical applications. British journal of clinical 418

pharmacology. 2004;57(1):6-14. 419

6. Al Hamid A, Ghaleb M, Aljadhey H, Aslanpour Z. A systematic review of 420

hospitalization resulting from medicine-related problems in adult patients. British journal of 421

clinical pharmacology. 2014;78(2):202-17. 422

7. U.S Department of Health and Human Services. National Action Plan for Adverse 423

Drug Event Prevention. Washington, DC. 2014. 424

8. National Institute for Health and Care Excellence. Atrial fibrillation management. 425

NICE guideline (CG180) 2014 [Available from: https://www.nice.org.uk/guidance/cg180. 426

Diabetes

and Met

aboli

sm Jo

urna

l

19

9. Ament P BJ, Liszewski J. . Clinically Significant Drug Interactions. American Family 427

Physician. 2000;15(61):1745-54. 428

10. January TC, Wann LS, Calkins H, Chen Lin Y, Cigarroa Joaquin E, Cleveland Joseph 429

C, et al. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS Guideline for 430

the Management of Patients With Atrial Fibrillation: A Report of the American College of 431

Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the 432

Heart Rhythm Society in Collaboration With the Society of Thoracic Surgeons. Circulation. 433

2019;140(2):e125-e51. 434

11. Mekaj YH, Mekaj AY, Duci SB, Miftari EI. New oral anticoagulants: their 435

advantages and disadvantages compared with vitamin K antagonists in the prevention and 436

treatment of patients with thromboembolic events. Therapeutics and Clinical Risk 437

Management. 2015;11:967-77. 438

12. Jennings AM, Wilson RM, Ward JD. Symptomatic hypoglycemia in NIDDM patients 439

treated with oral hypoglycemic agents. Diabetes care. 1989;12(3):203-8. 440

13. Kirchheiner J, Roots I, Goldammer M, Rosenkranz B, Brockmöller J. Effect of 441

genetic polymorphisms in cytochrome p450 (CYP) 2C9 and CYP2C8 on the 442

pharmacokinetics of oral antidiabetic drugs: clinical relevance. Clinical pharmacokinetics. 443

2005;44(12):1209-25. 444

14. May M, Schindler C. Clinically and pharmacologically relevant interactions of 445

antidiabetic drugs. Therapeutic Advances in Endocrinology and Metabolism. 2016;7(2):69-446

83. 447

15. Triplitt C. Drug Interactions of Medications Commonly Used in Diabetes. Diabetes 448

Spectrum. 2006;19(4):202-11. 449

16. Namazi S aRG. Case Presentation of a 45 Years Old Woman with Hypoglycemia and 450

Bleeding. Iranian Journal of Pharmaceutical Sciences. 2005;9(3):183-8. 451

Diabetes

and Met

aboli

sm Jo

urna

l

20

17. Naganuma M, Hashimoto Y, Matsuura Y, Terasaki T, Uchino M. A case of sustained 452

hypoglycemia induced by taking glibenclamide and warfarin 453

subtitle_in_Japanese. Nosotchu. 2003;25(3):334-7. 454

18. Romley JA, Gong C, Jena AB, Goldman DP, Williams B, Peters A. Association 455

between use of warfarin with common sulfonylureas and serious hypoglycemic events: 456

retrospective cohort analysis. BMJ. 2015;351:h6223. 457

19. Nam YH, Brensinger CM, Bilker WB, Leonard CE, Han X, Hennessy S. Serious 458

Hypoglycemia and Use of Warfarin in Combination With Sulfonylureas or Metformin. 459

Clinical Pharmacology & Therapeutics. 2018;0(0). 460

20. Movahed MR, Hashemzadeh M, Jamal MM. Diabetes mellitus is a strong, 461

independent risk for atrial fibrillation and flutter in addition to other cardiovascular disease. 462

International journal of cardiology. 2005;105(3):315-8. 463

21. Murphy NF, Simpson CR, Jhund PS, Stewart S, Kirkpatrick M, Chalmers J, et al. A 464

national survey of the prevalence, incidence, primary care burden and treatment of atrial 465

fibrillation in Scotland. Heart (British Cardiac Society). 2007;93(5):606-12. 466

22. Alwafi H, Wei L, Naser AY, Mongkhon P, Tse G, Man KKC, et al. Trends in oral 467

anticoagulant prescribing in individuals with type 2 diabetes mellitus: a population-based 468

study in the UK. BMJ open. 2020;10(5):e034573. 469

23. Alwafi H, Wong ICK, Banerjee A, Mongkhon P, Whittlesea C, Naser AY, et al. 470

Epidemiology and treatment of atrial fibrillation in patients with type 2 diabetes in the UK, 471

2001–2016. Scientific Reports. 2020;10(1):12468. 472

24. Blak BT, Thompson M, Dattani H, Bourke A. Generalisability of The Health 473

Improvement Network (THIN) database: demographics, chronic disease prevalence and 474

mortality rates. Informatics in primary care. 2011;19(4):251-5. 475

Diabetes

and Met

aboli

sm Jo

urna

l

21

25. Lewis JD, Schinnar R, Bilker WB, Wang X, Strom BL. Validation studies of the 476

health improvement network (THIN) database for pharmacoepidemiology research. 477

Pharmacoepidemiology and drug safety. 2007;16(4):393-401. 478

26. Maguire A, Blak BT, Thompson M. The importance of defining periods of complete 479

mortality reporting for research using automated data from primary care. 480

Pharmacoepidemiology and drug safety. 2009;18(1):76-83. 481

27. Leonard CE, Bilker WB, Brensinger CM, Han X, Flory JH, Flockhart DA, et al. 482

Severe hypoglycemia in users of sulfonylurea antidiabetic agents and antihyperlipidemics. 483

Clinical pharmacology and therapeutics. 2016;99(5):538-47. 484

28. Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of 485

Confounding in Observational Studies. Multivariate Behav Res. 2011;46(3):399-424. 486

29. Okoli GN, Sanders RD, Myles P. Demystifying propensity scores. Br J Anaesth. 487

2014;112(1):13-5. 488

30. Lau WC, Chan EW, Cheung CL, Sing CW, Man KK, Lip GY, et al. Association 489

Between Dabigatran vs Warfarin and Risk of Osteoporotic Fractures Among Patients With 490

Nonvalvular Atrial Fibrillation. Jama. 2017;317(11):1151-8. 491

31. Normand S-LT, Landrum MB, Guadagnoli E, Ayanian JZ, Ryan TJ, Cleary PD, et al. 492

Validating recommendations for coronary angiography following acute myocardial infarction 493

in the elderly: A matched analysis using propensity scores. Journal of clinical epidemiology. 494

2001;54(4):387-98. 495

32. Austin PC. Balance diagnostics for comparing the distribution of baseline covariates 496

between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083-497

107. 498

Diabetes

and Met

aboli

sm Jo

urna

l

22

33. Austin PC, Stuart EA. Moving towards best practice when using inverse probability 499

of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects 500

in observational studies. Stat Med. 2015;34(28):3661-79. 501

34. Xu S, Ross C, Raebel MA, Shetterly S, Blanchette C, Smith D. Use of stabilized 502

inverse propensity scores as weights to directly estimate relative risk and its confidence 503

intervals. Value in health : the journal of the International Society for Pharmacoeconomics 504

and Outcomes Research. 2010;13(2):273-7. 505

35. Lee BK, Lessler J, Stuart EA. Weight trimming and propensity score weighting. PloS 506

one. 2011b;6(3):e18174-e. 507

36. Pedersen AB, Mikkelsen EM, Cronin-Fenton D, Kristensen NR, Pham TM, Pedersen 508

L, et al. Missing data and multiple imputation in clinical epidemiological research. Clin 509

Epidemiol. 2017;9:157-66. 510

37. Marshall A, Altman DG, Holder RL, Royston P. Combining estimates of interest in 511

prognostic modelling studies after multiple imputation: current practice and guidelines. BMC 512

Med Res Methodol. 2009;9:57-. 513

38. Abraham NS, Singh S, Alexander GC, Heien H, Haas LR, Crown W, et al. 514

Comparative risk of gastrointestinal bleeding with dabigatran, rivaroxaban, and warfarin: 515

population based cohort study. BMJ : British Medical Journal. 2015;350:h1857. 516

39. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al. Rivaroxaban 517

versus warfarin in nonvalvular atrial fibrillation. The New England journal of medicine. 518

2011;365(10):883-91. 519

40. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM, Hanna M, et al. 520

Apixaban versus warfarin in patients with atrial fibrillation. The New England journal of 521

medicine. 2011;365(11):981-92. 522

Diabetes

and Met

aboli

sm Jo

urna

l

23

41. McElnay JC, D’Arcy PF. Protein Binding Displacement Interactions and their 523

Clinical Importance. Drugs. 1983;25(5):495-513. 524

42. Rolan PE. Plasma protein binding displacement interactions--why are they still 525

regarded as clinically important? British journal of clinical pharmacology. 1994;37(2):125-8. 526

43. Bibi Z. Role of cytochrome P450 in drug interactions. Nutrition & Metabolism. 527

2008;5(1):27. 528

44. Glipizide oral (drug facts and comparisons [Internet]. 2018 [cited March 2020]. 529

Available from: https://fco.factsandcomparisons.com/lco/action/home. 530

45. Wei J, Karsenty G. An overview of the metabolic functions of osteocalcin. Rev 531

Endocr Metab Disord. 2015;16(2):93-8. 532

46. Booth SL, Centi A, Smith SR, Gundberg C. The role of osteocalcin in human glucose 533

metabolism: marker or mediator? Nature reviews Endocrinology. 2013;9(1):43-55. 534

47. Zanatta LCB, Boguszewski CL, Borba VZC, Kulak CAM. Osteocalcin, energy and 535

glucose metabolism %J Arquivos Brasileiros de Endocrinologia & Metabologia. 536

2014;58:444-51. 537

48. Witt DM, Clark NP, Kaatz S, Schnurr T, Ansell JE. Guidance for the practical 538

management of warfarin therapy in the treatment of venous thromboembolism. Journal of 539

Thrombosis and Thrombolysis. 2016;41(1):187-205. 540

49. Wolpert HA. Use of continuous glucose monitoring in the detection and prevention of 541

hypoglycemia. Journal of diabetes science and technology. 2007;1(1):146-50. 542

50. Snipelisky D, Kusumoto F. Current strategies to minimize the bleeding risk of 543

warfarin. J Blood Med. 2013;4:89-99. 544

545

546

TABLES LEGENDS 547

Diabetes

and Met

aboli

sm Jo

urna

l

24

TABLES 548

549

Diabetes

and Met

aboli

sm Jo

urna

l

25

550

551

Table 1. Individuals’ characteristics among the cohort of first analysis (sulfonylureas and warfarin versus sulfonylureas only) 552

Variable Before propensity score matching

No. (%) of participant

After propensity score matching

No. (%) of participant

Sulfonylurea

s + warfarin

(n= 10,076)

Sulfonylurea

s

(n= 98,964)

Crude ASD Sulfonylurea

s + warfarin

(n= 5,379)

Sulfonylurea

s

(n= 5,379)

Matched

ASD

Demographics

Age mean (SD) 73.3 (10.0) 61.0 (12.8) 0.918 69.9 (10.8) 70.0 (11.8) 0.054

Male, n (%) 6,918 (61.1) 63,965

(57.0)

0.083 3,114

(57.89)

3041 (56.53) -0.027

BMI 0.037 0.038 Diabetes

and Met

aboli

sm Jo

urna

l

26

BMI < 25 1,715 (15.5) 18,491

(16.5)

- 877 (16.3) 923 (17.1) -

BMI 25-30 3,942 (34.8) 38,351

(34.2)

- 1837 (34.1) 1894 (35.2) -

BMI 30 5,665 (50.0) 55,341

(49.3)

- 2665 (49.5) 2562 (47.6) -

Smoking 0.228 -0.011

Non-smokers 10,068

(88.9)

90,649

(80.8)

- 4673 (86.8) 4652 (86.4) -

Smokers 1,254 (11.0) 21,534

(19.2)

- 706 (13.3) 727 (13.5) -

Alcohol 0.035 0.006

Non-drinker 3,604 (31.8) 33,899

(30.2)

- 1740 (32.3) 1757 (32.6) -

Drinker 7,718 (68.2) 78,284

(69.8)

- 3639 (67.7) 3622 (67.3) - Diabetes

and Met

aboli

sm Jo

urna

l

27

Townsend 0.053 0.036

1 (Least deprived) 2026 (20.1) 19174 (19.4) - 1075 (19.9) 1021 (19.0) -

2 2140 (21.2) 19555 (19.7) - 1091 (20.3) 1057 (19.6) -

3 2168 (21.6) 21773 (22.0) - 1149 (21.4) 1203 (22.4) -

4 2167 (21.5) 21523 (21.7) - 1199 (22.3) 1222 (22.7) -

5 (Most deprived) 1575 (15.6) 16939 (17.2) - 865 (16.0) 876 (16.3) -

Comorbid conditions, n (%)

CVDs 1670 (14.7) 5039 (4.5) 0.353 684 (12.7) 656 (12.2) -0.016

Hypertension 7693 (68.0) 56550 (50.4) 0.363 3399 (63.2) 3370 (62.6) 0.011

Stroke/TIA 2154 (19.0) 6429 (5.7) 0.412 838 (15.6) 866 (16.1) -0.014

Bleeding 2112 (18.6) 15302 (13.6) 0.137 1009 (18.7) 970 (18.0) -0.019

Hyperlipidaemia 2652 (23.4) 19747 (17.6) 0.144 1187 (22.0) 1194 (22.2) -0.003

AF 6952 (61.4) 1973 (1.8) 1.673 1862 (34.6) 1750 (32.5) 0.044

DVT 2403 (21.2) 2300 (2.0) 0.627 1453 (27.0) 1571 (29.2) -0.049

Chronic kidney disease 3016 (26.6) 10991 (9.8) 0.447 1099 (20.4) 1105 (20.5) -0.003

COPD 1208 (10.7) 5081 (4.5) 0.233 481 (9.0) 518 (9.6) -0.024

Diabetes

and Met

aboli

sm Jo

urna

l

28

Hyperglycaemia 422 (3.7) 3451 (3.0) 0.036 212 (3.9) 216 (4.0) -0.004

Liver diseases 59 (0.5) 678 (0.6) 0.011 36 (0.7) 49 (0.9) -0.027

Depression 2323 (20.5) 25934 (23.1) 0.063 1202 (22.3) 1254 (23.3) -0.023

Anxiety 1542 (13.6) 17554 (15.6) 0.057 811 (15.0) 843 (15.7) -0.017

Baseline medication use, n (%)

Aspirin use 4264 (37.6) 34978 (31.8) 0.137 2223 (41.3) 2480 (46.1) -0.096

Antiplatelet drugs use 716 (6.3) 3636 (3.2) 0.145 368 (6.8) 394 (7.3) -0.019

Beta blockers use 5264 (46.5) 23663 (21.0) 0.558 1943 (36.1) 36.12 (38.8) -0.055

ACEs /ARBs use 8061 (71.2) 50245 (44.8) 0.555 3337 (62.0) 3328 (61.9) 0.003

Corticosteroids use 1276 (11.2) 6854 (6.1) 0.184 578 (10.7) 616 (11.4) -0.022

Multiple antidiabetic

medications use

(intensification)

7705 (68.0) 76835 (68.5) 0.009 3559 (66.1) 3425 (63.7) 0.052

553

554

555

Diabetes

and Met

aboli

sm Jo

urna

l

29

Table 2. Number of events, incidence rates and crude HR, for risk of hypoglycaemia 556

Exposure group No. of event Person-years

at risk, year

IR, per 1000

person-years

(95% CI)

Crude HR

(95% CI)

Sulfonylurea+warfarin

(n=10,076)

578 34422.40 16.7 1.20

(1.10 – 1.30)

Sulfonylurea only

(n=98,964)

7307 526422.49 13.8

Sulfonylurea+DOAC

*(n=1,045)

14 956.9 15.0 0.66

(0.39- 1.11)

Sulfonylurea only *

(n=76251)

4514 246345.6 18.0

* This analysis included individuals only from 2011 onward. 557

558

559

560

561

562

563

564

565

566

Diabetes

and Met

aboli

sm Jo

urna

l

30

567

568

Table 3. Number of events, incidence rates and HR, for risk of hypoglycaemia for the 569

matched cohort 570

Exposure group No. of event Person-years

at risk, year

IR, per 1000

person-years

(95% CI)

Matched HR

(95% CI), p-

value

Sulfonylurea+warfarin

(n=5379)

285 15959.04 17.8

1.38

( 1.10-1.75)

Sulfonylurea only

(n=5379)

304 21028.52 14.4 1.00

Sulfonylurea+DOAC

*(n=1027)

14 942.2 14.8

0.54

(0.27- 1.10)

Sulfonylurea only *

(n=1027)

60 2532.0 23.7 1.00

* This analysis included individuals only from 2011 onward. 571

572

573

574

Diabetes

and Met

aboli

sm Jo

urna

l

31

575

576

577

578

579

580

Diabetes

and Met

aboli

sm Jo

urna

l

32

581

Figure 1. Flow chart of the study cohort (analysis 1) 582

583

584

585

Diabetes

and Met

aboli

sm Jo

urna

l

33

586

587

588

589

Figure 2. Flow chart of the study cohort (analysis 2) 590

591

Diabetes

and Met

aboli

sm Jo

urna

l

34

592

Figure 3. Kaplan-Meier curves for the incidence of hypoglycaemia during the follow-up 593

period (sulfonylureas and warfarin versus sulfonylureas only). 594

595

596

Diabetes

and Met

aboli

sm Jo

urna

l

35

597

598

Figure 4. Kaplan-Meier curves for the incidence of hypoglycaemia during the follow-up 599

period (sulfonylureas and DOACs versus sulfonylureas only). 600

601

602

603 Diabetes

and Met

aboli

sm Jo

urna

l