A pharmacoepidemiological nested case-control study of risk factors for venous thromboembolism with the focus on diabetes,cancer,socioeconomic group,medications,and comorbidities

Abstract

Objectives

A pharmacoepidemiological study to assess VTE risk factors in a diabetes-rich population.

Methods

The study comprised 299,590 individuals. We observed 3450 VTEs and matched them with 15,875 controls using a nested case-control approach and collected data on comorbidities and prescriptions. By multivariable conditional logistic regression, we calculated ORs with 95%CIs for comorbidities and medications to evaluate their associations with VTE.

Results

Diabetes (aOR 2.16; 95%CI 1.99–2.34), inflammatory bowel disease (1.84; 1.27–2.66), and severe psychiatric disorders (1.72; 1.43–2.05) had the strongest associations among the non-cancer comorbidities. Pancreatic (12.32; 7.11–21.36), stomach (8.57; 4.07–18.03), lung and bronchus (6.26; 4.16–9.43), and ovarian (6.72; 2.95–15.10) cancers were ranked as high-risk for VTE. Corticosteroids, gabapentinoids, psychotropic drugs, risedronic acid, and pramipexole were most strongly associated (aOR exceeding 1.5) with VTE. Insulin (3.86; 3.33–4.47) and sulphonylureas (2.62; 2.18–3.16) had stronger associations than metformin (1.65; 1.49–1.83). Statins and lercanidipine (0.78; 0.62–0.98) were associated with a lowered risk of VTE.

Conclusions

In this cohort, with 50% diabetes prevalence, pancreatic, stomach, lung and bronchus, and ovarian cancers were strongly associated with VTE. Corticosteroids, gabapentinoids, and psychotropic medications had the strongest associations with VTE among medications. This may be valuable for generating hypotheses for the further research. Lercanidipine may be a novel protective medication against VTE.

Keywords

Venous thromboembolism pulmonary embolism deep vein thrombosis diabetes cancer-associated thrombosis prothrombotic drugs

Introduction

Venous thromboembolism (VTE), with its typical manifestations of deep vein thrombosis (DVT) and pulmonary embolism (PE), is a major global health issue, with an incidence of 1–2 per 1000 person-years.^1–5 Over time, the understanding of VTE has developed from a complication of major surgery or terminal illness to a multifactorial disease with several associated inherited, acquired, environmental, and social risk factors.^1–5 VTE is a highly heterogeneous disease, ranging from asymptomatic radiological co-incidence to immediately life-threatening acute massive PE with up to a 10% acute-phase direct thrombosis-related case mortality.⁶ Accompanied by common post-thrombotic consequences that compromise either the acute-phase survivors´ quality of life^7,8 or long-term prognosis,⁹ the burden of VTE at the population level is substantial. Despite all landmark epidemiological research,^1–5 predicting who will develop VTE remains challenging, as some people with a notable risk factor burden will never develop VTE, while many cases (25–50% of first VTE incidents) are idiopathic and have no identifiable risk factors.⁴

Regarding non-cancer comorbidities, many conditions are associated with increased VTE risk, including inflammatory bowel disease, chronic heart failure, neurologic paralytic conditions, autoimmune disorders like systemic lupus erythematosus or systemic vasculitis, hyperthyroidism, or chronic kidney disease.^1–5 Diabetes mellitus with its complications is an emerging and burdensome public health problem.¹⁰ Atherosclerosis and diabetes have a well-established connection, and prothrombotic alterations in coagulation cascade and fibrinolytic function have been reported in diabetes.^11,12 VTE is not infrequent among diabetes patients, but it is not proven if diabetes by itself is an independent risk factor for VTE^13,14 after adjustment for diabetes-related confounders and co-morbidities. Traditional general VTE risk factors like major surgery or hospitalization also predispose diabetes patients to thrombosis,¹⁵ and additionally, type 1 diabetes¹⁶ may carry stronger VTE risk compared to type 2 diabetes,¹⁶ and female patients with diabetes may be more prone to VTE than male diabetes patients.¹⁷ Impact of the glycemic control with the risk of VTE is not clear¹⁸ whereas occurrence of hypoglycemia seems to predispose to long-lasting prothrombotic state.¹⁹ Of the common acquired risk factors of VTE, cancer makes a significant contribution, and cancer-associated thrombosis has attracted extensive scientific interest in recent years.²⁰ The risk of VTE varies considerably depending on the tumor type, and is the highest for pancreatic, brain, lung, and hematologic cancers, whereas breast and prostate cancers generally have the lowest thrombotic potential.²⁰ Among cancer-free diabetes patients with first-time episode of VTE, incidence of gallbladder and biliary tract cancer, pancreas cancer, and ovarian cancer were highest in 1-year study period.²¹

Certain medications are associated with an increased risk of VTE incidence, notoriously estrogen-containing oral agents.^22,23 Glucocorticoids,^22,24 and many oncological medications^22,25,26 are also prothrombotic. Additionally, exposure to antipsychotics or antidepressants^22,27 are associated with VTE, although the causation and specific mechanisms remain uncertain. Beyond actual antithrombotic drugs, statins have been studied as VTE-protecting agents with promising,^28,29 but also insignificant³⁰ results. With respect to diabetes medications, metformin may have a protective effect against VTE³¹ while SGLT-2 and DDP4 inhibitors seems to be neutral.^32,33 It is noteworthy that these drugs do not predispose to hypoglycemia.

We conducted this nested case-control study to assess potential VTE risk factors and prothrombotic or protective medications in specific diabetes-rich population with an average age of 60. As our data source we used previously described FIN-CARING2 retrospective cohort³⁴ consisting of almost 400,000 individuals, of which 50% had diabetes and circa 3400 developed VTE in 16 years of follow-up.

Methods

Study design and population

The original study population of FIN-CARING2 was based on the multinational CARING project. FIN-CARING2 focused on cancer risk and insulin analogues in Finland and consisted of 398,708 individuals (50% with diabetes), as earlier described by Niskanen et al.³⁴ (Figure 1). The beginning of the follow-up was the date of the first diabetes medication prescription and the study period of the original cohort was 1996–2017. In our study, we excluded people with the start of the follow-up before 2001 (N = 97,700), leading to 16 years of follow-up. Exclusion was carried out in order to guarantee at least 5-year history of records before the start of the follow-up. People who had received hospital or tertiary care during the 5 years before the start of follow-up with ICD-codes I26 Pulmonary embolism, I63.6 Cerebral infarction due to cerebral vein thrombosis, non-pyogenic, and I80 Diseases of veins, lymphatic vessels, and lymph nodes, not elsewhere classified (N = 1418) were also excluded to rule out recurrent VTE events. People with recurrent VTE are a special high-risk group, and their inclusion could have confounded the associations of our study.^1–5 Thus, the final study population consisted of N = 299,590 individuals. The definition of VTE in our study included any of the diagnosis I26, I63.6 and I80. Our national real-world practice is to use the ICD-codes I80 for DVT, and I26 for PE, rather than I82, which captures acute embolism and thrombosis of other specified or unspecified veins. In addition, our previous publication used the same diagnosis codes.²³ Our study was accepted by the Helsinki University Ethics Committee (REF 02/2012).

Figure 1.

Flow chart of the study. Original cohort, exclusions, final study population, cases, and controls are presented.

Nested case-control design

We observed 3450 (1.15% of study population) VTE cases and sampled them using the nested case-control approach aiming five controls for each case, matched by age (accuracy of 1 year), sex, hospital district, and the year of the start of the follow-up. Forty-six cases could not be matched because data did not include any potential individuals with correct matching criteria, resulting in one control for 124 cases, two for 61 cases, three for 144 cases, four for 112 cases, and five controls for 2930 cases. Thus, the number of included VTE cases was 3404, and the total number of controls was 15,875. We acquired the baseline data on age, sex, socioeconomic group, cancer diagnosis and type (10-year period before VTE incidence). The data of prevalent comorbidities were based on special reimbursement rights, and the following comorbidities were included: hypothyroidism (reimbursement code 104 including ICD-10 C73, E03, E89.0), severe psychiatric disorders (reimbursement code 112 including ICD-10 F01, F03, F06.0-F06.3, F20–F25, F28, F29, F30.1, F30.2, F31, F32.3, F33.3, F84, G10, G20, G30.0, G30.1, G30.8, G30.9, G31.0, G35, G40.9), chronic heart failure (reimbursement code 201 including ICD-10 I11.0, I13, I50, I97.1, P29.0), coronary artery disease (reimbursement code 206 including ICD-10 I20–I22, I24.0, I25), hypertension (reimbursement code 205 including ICD-10 I10–I13, I15, I27.0, I27.2), chronic arrhythmias (reimbursement code 207 including ICD-10 I47–I49), and inflammatory bowel disease (reimbursement code 208 including ICD-10 K50, K51) (Table 3). The data were obtained from registries of The Social Insurance Institution of Finland (SII) (permission Kela 16/522/2012, providing medications and reimbursement rights), Statistics Finland (permission TK-53-214-12) and the Finnish Cancer Registry (THL/264/5.05.00/2012). There was no missing data in used variables.

Medications and predefined exposures

We also collected data on the ambulatory use of drugs by anatomical therapeutic chemical (ATC) classification A08A, A10, C, G03, H01-03, L01-04, M05, N03-07, using the SII registry (permission KELA 16/522/2012) as the data source (Supplemental Table 1). A specific drug was included in the analyses if there were at least 50 cases and controls in both the exposed and unexposed groups. To be classed as exposed, an individual had to have redeemed at least three prescriptions in the year before the index date. We also applied another approach to variable selection. Among the available prescription data, we rated certain predefined agents as high-risk medication for VTE and certain agents as protective, based on the previous literature,^21–29 and analyzed the associations with VTE incidence of these predefined drugs (Supplemental Table 2).

Statistical modeling

We analyzed the data using multivariable conditional logistic regression, which takes matching into account. Each drug exposure was analyzed separately with and without adjustment for the background variables: socioeconomic group, cancer diagnosis and type; prevalence of comorbidities, including diabetes, hypothyroidism, severe psychiatric disorders, chronic heart failure, coronary artery disease, hypertension, chronic arrhythmias, and inflammatory bowel disease. These variables we added into models, as potential confounders that may have an effect both on exposure and outcome. The results were reported as odds ratios (OR) with 95% confidence intervals (CI). The data were analyzed by R (R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing, 2021 https://www.R-project.org/). BioRender.com® was used for Figure 1.

Results

Baseline characteristics

The mean age of our study population was 66, and 50% were female (Table 1). Regarding comorbidities, diabetes, severe psychiatric disorders, chronic heart failure, hypertension, and coronary artery disease were more prevalent in the case group. Cancer was diagnosed in 18.4% of cases (10-year period before VTE incidence), whereas the cancer prevalence in the control group was 8.5% (Table 2). Because of the nature of the original cohort, the total prevalence of diabetes was high, 50%.

Table 1.

Basic characteristics of study population.

		Controls	Cases	Overall
n		15,875	3404	19,279
Mean age (SD)		65.9 (12.2)	66.3 (12.7)	66.0 (12.3)
Sex (female, %)		7954 (50.1)	1715 (50.4)	9669 (50.2)
Diabetes (yes, %)		7338 (46.2)	2322 (68.2)	9660 (50.1)
Socioeconomic group (%)	Upper-level employees	920 (5.8)	109 (3.2)	1029 (5.3)
	Self-employed	728 (4.6)	119 (3.5)	847 (4.4)
	Lower-level employees	1363 (8.6)	218 (6.4)	1581 (8.2)
	Manual workers	1584 (10.0)	299 (8.8)	1883 (9.8)
	Students	83 (0.5)	26 (0.8)	109 (0.6)
	Pensioners	10177 (64.1)	2384 (70.0)	12561 (65.2)
	Others	1020 (6.4)	249 (7.3)	1269 (6.6)
Thyroid insufficiency (yes, %)		638 (4.0)	171 (5.0)	809 (4.2)
Severe psychotic and other severe mental disorders (yes, %)		508 (3.2)	215 (6.3)	723 (3.8)
Chronic heart failure (yes, %)		619 (3.9)	221 (6.5)	840 (4.4)
Chronic hypertension (yes, %)		5194 (32.7)	1330 (39.1)	6524 (33.8)
Coronary artery disease (yes, %)		2405 (15.1)	729 (21.4)	3134 (16.3)
Chronic arrhythmias (yes, %)		635 (4.0)	143 (4.2)	778 (4.0)
Ulcerative colitis and crohn’s disease (yes, %)		113 (0.7)	45 (1.3)	158 (0.8)

Table 2.

Cancer prevalence.

		Controls (%)	Cases (%)	Overall (%)
n		15,875	3404	19,279
Cancer^a		1350 (8.5)	628 (18.4)	1978 (10.3)
No cancer		14525 (91.5)	2776 (81.6)	17301 (89.7)
Cancer type	Prostate cancer	401 (2.5)	95 (2.8)	496 (2.6)
	Breast cancer	209 (1.3)	53 (1.6)	262 (1.4)
	Other cancers	129 (0.8)	91 (2.7)	220 (1.1)
	Skin cancer	148 (0.9)	44 (1.3)	192 (1.0)
	Colon cancer	96 (0.6)	43 (1.3)	139 (0.7)
	Bronchus and lung cancer	49 (0.3)	55 (1.6)	104 (0.5)
	Cancer of hematopoietic and reticuloendothelial system	55 (0.3)	44 (1.3)	99 (0.5)
	Cancer of corpus uteri	52 (0.3)	29 (0.9)	81 (0.4)
	Bladder cancer	55 (0.3)	23 (0.7)	78 (0.4)
	Pancreatic cancer	19 (0.1)	50 (1.5)	69 (0.4)
	Rectum cancer	41 (0.3)	22 (0.6)	63 (0.3)
	Kidney cancer	42 (0.3)	20 (0.6)	62 (0.3)
	Lymphoma	29 (0.2)	25 (0.7)	54 (0.3)
	Stomach cancer	14 (0.1)	19 (0.6)	33 (0.2)
	Ovarian cancer	11 (0.1)	15 (0.4)	26 (0.1)

^a10-year period before incident VTE was diagnosed.

Socioeconomic group, medical comorbidities, cancer, and associations with VTE

By conditional logistic regression analysis, we calculated adjusted ORs (aORs) with 95% CI for incidence of VTE among socioeconomic groups and medical comorbidities, including cancers (Table 3, Figure 2). We adjusted for all background variables in these analyses. Students (aOR 2.38; CI 1.36–4.19) and pensioners (2.31; 1.81–2.94) had the strongest associations with VTE among the socioeconomic groups. The strongest associations among non-cancer comorbidities were with diabetes (2.16; 1.99–2.34), inflammatory bowel disease (1.84; 1.27–2.66), and severe psychiatric disorder (1.72; 1.43–2.05). With respect to the cancer types, pancreatic (12.32; 7.11–21.36), stomach (8.57; 4.07–18.03), ovarian (6.72; 2.95–15.10), and bronchus and lung (6.26; 4.16–9.43) cancers had the highest aORs for VTE incidence (Figure 1). In contrast, breast, skin, and prostate cancers were associated with lower risk compared to other cancer types.

Table 3.

Associations of socioeconomic groups and comorbidities with VTE.

		aOR^a (95% CI^b)
Socioeconomic group	Self-employed	1.39 (1.04–1.86)
	Lower-level employees	1.35 (1.05–1.74)
	Manual workers	1.49 (1.16–1.90)
	Students	2.38 (1.36–4.19)
	Pensioners	2.31 (1.81–2.94)
	Others	1.92 (1.49–2.47)
Diabetes (yes vs no)		2.16 (1.99–2.34)
Hypothyroidism (yes vs no)		1.11 (0.92–1.33)
Severe psychotic and other severe mental disorders (yes vs no)		1.72 (1.43–2.05)
Chronic heart failure (yes vs no)		1.38 (1.15–1.65)
Chronic hypertension (yes vs no)		1.10 (1.01–1.19)
Coronary artery disease (yes vs no)		1.41 (1.27–1.56)
Chronic arrhythmias (yes vs no)		0.83 (0.68–1.02)
Ulcerative colitis and crohn’s disease (yes vs no)		1.84 (1.27–2.66)

^aOdds ratio adjusted for all background variables (cancer diagnosis and type; prevalence of comorbidities, including diabetes, hypothyroidism, severe psychiatric disorders, chronic heart failure, coronary artery disease, hypertension, chronic arrhythmias, and inflammatory bowel disease).

^bConfidence interval.

Figure 2.

Cancer types and their associations with VTE incidence. Adjusted odds ratios with 95% confidence intervals. Pancreatic, stomach, ovarian, bronchus, and lung cancers have the strongest associations.

Diabetes medications and VTE

We calculated crude ORs and ORs adjusted for the socioeconomic group for different diabetes medication combinations (Table 4). Insulin-containing regimens had the strongest associations with VTE (insulin aOR 3.86; 3.33–4.47, insulin + sulphonylurea aOR 3.55; 1.91–6.58, insulin + sulphonylurea + metformin 3.59; 1.85–6.97) (Table 4). Diabetes without medication had an aOR of 2.60 (2.34–2.90) while metformin monotherapy had the lowest aOR (1.65; 1.49–1.83) among diabetes medication regimens.

Table 4.

Diabetes medications and associations with VTE.

	Controls (%)	Cases (%)	Overall	OR^a (95%CI^b)	aOR^c (95%CI)
n	15,875	3404	19,279
No diabetes	8537 (53.8)	1082 (31.8)	9619 (49.9)	1.00 (1.00–1.00)	1.00 (1.00–1.00)
Diabetes, no medication	2127 (13.4)	762 (22.4)	2889 (14.6)	2.67 (2.40–2.96)	2.61 (2.34–2.90)
Insulin	669 (4.2)	368 (10.1)	1037 (5.4)	3.98 (3.44–4.60)	3.86 (3.33–4.47)
Sulphonylurea	509 (3.2)	195 (5.7)	704 (3.7)	2.71 (2.25–3.26)	2.62 (2.18–3.16)
Insulin+sulphonylurea	31 (0.2)	18 (0.5)	49 (0.3)	3.80 (2.15–7.38)	3.55 (1.91–6.58)
Metformin	3004 (18.9)	678 (19.9)	3682 (19.0)	1.68 (1.51–1.87)	1.65 (1.49–1.83)
Insulin+metformin	532 (3.4)	179 (5.3)	711 (3.7)	2.55 (2.12–3.07)	2.48 (2.06–2.99)
Sulphonylurea+metformin	436 (2.7)	109 (3.2)	545 (2.8)	1.91 (1.53–2.39)	1.86 (1.48–2.33)
Insulin+ sulphonylurea +metformin	30 (0.2)	13 (0.4)	43 (0.2)	3.49 (1.80–6.74)	3.59 (1.85–6.97)

^aOdds ratio.

^b95% confidence interval.

^cOdds ratio adjusted for socioeconomic group.

Medications and associations with VTE

After analyzing diabetes medications, we calculated crude ORs and ORs adjusted for all background variables for other ambulatory medications (Table 5). The detailed data on drug exposures and the associations with VTE incidence reveal a strong association between oral corticosteroids (methylprednisolone (aOR 5.37; 95% CI 3.30–8.73), prednisolone (4.06; 3.35–4.91) and prednisone (3.96; 2.44–6.44)) and VTE incidence (Figures 3 and 4). Gabapentin, an antiepileptic and widely used neuropathic pain management agent, had an aOR of 3.16 (2.13–4.70) and another gabapentinoid, pregabalin, had an aOR of 2.21 (1.73–2.83). The loop-diuretic furosemide had a clear association with VTE incidence with an aOR of 2.05 (1.83–2.30).

Table 5.

Ambulatory medications of the study population and associations with VTE.

	Controls (%)	Cases (%)	Overall (%)	OR^a (95%CI^b)	aOR (95%CI)^c
n	15,875	3404	19,279
C01AA05 (digoxin)	451 (2.8)	101 (3.0)	552 (2.9)	0.79 (1.24–0.81)	0.72 (0.57–0.92)
C01DA02 (glyceryl trinitrate)	185 (1.2)	78 (2.3)	263 (1.4)	1.48 (2.56–1.75)	1.55 (1.16–2.01)
C01DA08 (isosorbide dinitrate)	429 (2.7)	168 (4.9)	597 (3.1)	1.43 (2.09–1.56)	1.35 (1.10–1.66)
C01DA14 (isosorbide mononitrate)	1037 (6.5)	368 (10.8)	1405 (7.3)	1.52 (1.98–1.53)	1.35 (1.17–1.56)
C02AC05 (moxonidine)	151 (1.0)	55 (1.6)	206 (1.1)	1.24 (2.31–1.42)	1.36 (0.98–1.90)
C03AA03 (hydrochlorothiazide)	181 (1.1)	39 (1.1)	220 (1.1)	0.70 (1.41–0.87)	0.88 (0.61–1.27)
C03CA01 (furosemide)	1518 (9.6)	718 (21.1)	2236 (11.6)	2.36 (2.91–2.16)	2.05 (1.83–2.30)
C03DA01 (spironolactone)	256 (1.6)	113 (3.3)	369 (1.9)	1.68 (2.65–1.72)	1.50 (1.18–1.91)
C03EA01 (hydrochlorothiazide and potassium-sparing agents)	500 (3.1)	94 (2.8)	594 (3.1)	0.70 (1.10–0.81)	0.85 (0.68–1.08)
C03EB01 (furosemide and potassium-sparing agents)	148 (0.9)	49 (1.4)	197 (1.0)	1.04 (2.04–1.22)	1.23 (0.87–1.74)
C07AA05 (propranolol)	118 (0.7)	35 (1.0)	153 (0.8)	0.96 (2.07–1.24)	1.29 (0.86–1.92)
C07AB02 (metoprolol)	1561 (9.8)	422 (12.4)	1983 (10.3)	1.13 (1.44–1.15)	1.06 (0.94–1.20)
C07AB03 (atenolol)	215 (1.4)	41 (1.2)	256 (1.3)	0.65 (1.27–0.82)	0.81 (0.56–1.16)
C07AB07 (bisoprolol)	2956 (18.6)	827 (24.3)	3783 (19.6)	1.29 (1.54–1.24)	1.14 (1.04–1.26)
C07AG02 (carvedilol)	221 (1.4)	58 (1.7)	279 (1.4)	0.90 (1.62–1.02)	1.03 (0.75–1.40)
C08CA01 (amlodipine)	1646 (10.4)	358 (10.5)	2004 (10.4)	0.90 (1.15–0.91)	0.88 (0.78–1.01)
C08CA02 (felodipine)	396 (2.5)	88 (2.6)	484 (2.5)	0.83 (1.34–0.92)	0.93 (0.83–1.19)
C08CA05 (nifedipine)	180 (1.1)	52 (1.5)	232 (1.2)	0.98 (1.84–1.31)	1.24 (0.89–1.72)
C08CA13 (lercanidipine)	549 (3.5)	100 (2.9)	649 (3.4)	0.68 (1.05–0.77)	0.78 (0.62–0.98)
C08DB01 (diltiazem)	144 (0.9)	48 (1.4)	192 (1.0)	1.15 (2.24–1.59)	1.35 (0.95–1.92)
C09AA02 (enalapril)	867 (5.5)	196 (5.8)	1063 (5.5)	0.88 (1.22–0.93)	0.91 (0.77–1.07)
C09AA04 (perindopril)	176 (1.1)	43 (1.3)	219 (1.1)	0.81 (1.59–0.99)	0.90 (0.63–1.30)
C09AA05 (ramipril)	1330 (8.4)	368 (10.8)	1698 (8.8)	1.20 (1.54–1.17)	1.23 (0.99–1.29)
C09BA02 (enalapril and diuretics)	445 (2.8)	87 (2.6)	532 (2.8)	0.73 (1.16–0.81)	0.82 (0.64–1.04)
C09CA01 (losartan)	735 (4.6)	182 (5.3)	917 (4.8)	1.00 (1.40–1.04)	1.00 (0.84–1.19)
C09CA03 (valsartan)	289 (1.8)	65 (1.9)	354 (1.8)	0.79 (1.37–0.93)	0.91 (0.69–1.22)
C09CA06 (candesartan)	597 (3.8)	146 (4.3)	743 (3.9)	0.95 (1.38–1.06)	1.03 (0.85–1.25)
C09CA07 (telmisartan)	176 (1.1)	35 (1.0)	211 (1.1)	0.65 (1.36–0.86)	0.97 (0.66–1.42)
C09DA01 (losartan and diuretics)	455 (2.9)	92 (2.7)	547 (2.8)	0.76 (1.20–0.81)	0.83 (0.65–1.05)
C09DA03 (valsartan and diuretics)	202 (1.3)	46 (1.4)	248 (1.3)	0.80 (1.52–1.01)	1.04 (0.74–1.46)
C09DA06 (candesartan and diuretics)	313 (2.0)	69 (2.0)	382 (2.0)	0.83 (1.40–0.98)	1.06 (0.81–1.39)
C10AA01 (simvastatin)	3116 (19.6)	636 (18.7)	3752 (19.5)	0.87 (1.05–0.82)	0.79 (0.72–0.88)
C10AA03 (pravastatin)	146 (0.9)	46 (1.4)	192 (1.0)	1.02 (2.01–1.36)	1.34 (0.94–1.90)
C10AA04 (fluvastatin)	228 (1.4)	41 (1.2)	269 (1.4)	0.59 (1.15–0.73)	0.70 (0.49–0.99)
C10AA05 (atorvastatin)	1186 (7.5)	256 (7.5)	1442 (7.5)	0.90 (1.20–0.91)	0.84 (0.72–0.97)
C10AA07 (rosuvastatin)	374 (2.4)	76 (2.2)	450 (2.3)	0.75 (1.25–0.85)	0.81 (0.63–1.06)
G03CA03 (estradiol)	500 (3.1)	77 (2.3)	577 (3.0)	0.56 (0.93–0.78)	0.78 (0.60–1.01)
H02AB04 (methylprednisolone)	31 (0.2)	47 (1.4)	78 (0.4)	4.70 (11.81–5.59)	5.37 (3.30–8.72)
H02AB06 (prednisolone)	276 (1.7)	251 (7.4)	527 (2.7)	3.80 (5.44–4.29)	4.05 (3.35–4.91)
H02AB07 (prednisone)	44 (0.3)	36 (1.1)	80 (0.4)	2.67 (6.62–4.26)	3.96 (2.4–6.44)
H03AA01 (levothyroxine sodium)	1238 (7.8)	364 (10.7)	1602 (8.3)	1.26 (1.63–1.34)	1.36 (1.17–1.59)
L04AX03 (methotrexate)	114 (0.7)	35 (1.0)	149 (0.8)	1.02 (2.19–1.41)	1.54 (1.03–2.29)
M04AA01 (allopurinol)	454 (2.9)	164 (4.8)	618 (3.2)	1.39 (2.01–1.34)	1.20 (0.99–1.47)
M05BA04 (alendronic acid)	201 (1.3)	54 (1.6)	255 (1.3)	0.94 (1.73–1.32)	1.35 (0.98–1.86)
M05BA07 (risedronic acid)	75 (0.5)	29 (0.9)	104 (0.5)	1.14 (2.73–1.84)	1.88 (1.19–2.96)
N03AE01 (clonazepam)	51 (0.3)	25 (0.7)	76 (0.4)	1.38 (3.65–2.02)	1.90 (1.13–3.22)
N03AG01 (valproic acid)	91 (0.6)	36 (1.1)	127 (0.7)	1.25 (2.74–1.45)	1.13 (0.73–1.72 (
N03AX12 (gabapentin)	66 (0.4)	49 (1.4)	115 (0.6)	2.44 (5.16–3.22)	3.16 (2.13–4.70)
N03AX16 (pregabalin)	220 (1.4)	118 (3.5)	338 (1.8)	2.04 (3.23–2.28)	2.21 (1.73–2.83)
N04BC05 (pramipexole)	116 (0.7)	42 (1.2)	158 (0.8)	1.20 (2.44–1.73)	1.86 (1.28–2.70)
N05AH03 (olanzapine)	84 (0.5)	52 (1.5)	136 (0.7)	2.13 (4.34–2.49)	1.94 (1.30–2.90)
N05AH04 (quetiapine)	200 (1.3)	98 (2.9)	298 (1.5)	1.77 (2.91–1.94)	1.72 (1.31–2.25)
N05AX08 (risperidone)	181 (1.1)	78 (2.3)	259 (1.3)	1.49 (2.57–1.81)	1.57 (1.19–2.14)
N05BA01 (diazepam)	98 (0.6)	36 (1.1)	134 (0.7)	1.22 (2.63–1.41)	1.30 (0.87–1.95)
N05BA04 (oxazepam)	276 (1.7)	136 (4.0)	412 (2.1)	1.94 (2.96–2.18)	1.80 (1.60–2.50)
N05BA06 (lorazepam)	86 (0.5)	36 (1.1)	122 (0.6)	1.34 (2.95–1.92)	1.80 (1.19–2.73)
N05CD07 (temazepam)	262 (1.7)	105 (3.1)	367 (1.9)	1.51 (2.43–1.71)	1.61 (1.25–2.06)
N05CF01 (zopiclone)	826 (5.2)	239 (7.0)	1065 (5.5)	1.18 (1.60–1.29)	1.22 (1.04–1.43)
N05CF02 (zolpidem)	194 (1.2)	43 (1.3)	237 (1.2)	0.74 (1.43–0.97)	0.97 (0.68–1.37)
N06AB04 (citalopram)	339 (2.1)	119 (3.5)	458 (2.4)	1.30 (2.01–1.52)	1.39 (1.11 (1.75)
N06AB10 (escitalopram)	196 (1.2)	61 (1.8)	257 (1.3)	1.10 (1.98–1.39)	1.38 (1.02–1.87)
N06AX11 (mirtazapine)	358 (2.3)	162 (4.8)	520 (2.7)	1.81 (2.66–2.11)	1.94 (1.58–2.38)
N06AX16 (venlafaxine)	99 (0.6)	49 (1.4)	148 (0.8)	1.66 (3.32–1.96)	1.94 (1.35–2.78)
N06DA02 (donepezil)	269 (1.7)	86 (2.5)	355 (1.8)	1.16 (1.92–1.58)	1.63 (1.26–2.12)
N06DA03 (rivastigmine)	155 (1.0)	46 (1.4)	201 (1.0)	0.94 (1.85–1.30)	1.33 (0.93–1.89)
N06DX01 (memantine)	272 (1.7)	106 (3.1)	378 (2.0)	1.42 (2.29–1.80)	1.84 (1.44–2.36)

^aOdds ratio.

^bConfidence interval.

^cOR adjusted for all background variables (socioeconomic group, cancer diagnosis and type; prevalence of comorbidities, including diabetes, hypothyroidism, severe psychiatric disorders, chronic heart failure, coronary artery disease, hypertension, chronic arrhythmias, and inflammatory bowel disease).

Figure 3.

Medications and their associations with VTE incidence. Adjusted odds ratios with 95% confidence intervals are shown. Corticosteroids and gabapentinoids show the strongest associations.

Figure 4.

Medications and their associations with VTE incidence. A volcano plot highlights the strongest and statistically most significant associations. Adjusted odds ratio is given on X-axis, −log10 of p-value on Y-axis.

Otherwise, many psychotropic agents from different drug classes had the strongest positive associations: Various benzodiazepines (oxazepam, clonazepam, lorazepam, temazepam), antipsychotics (olanzapine, quietapine, risperidone), dementia medications (memantine, donepezil), and an antidepressant agent venlafaxine all had aORs exceeding 1.5. Risedronic acid (osteoporosis agent) (1.88; 1.19–2.96) and a dopamine agonist pramipexole (1.86; 1.28–2.70) also exhibited a positive association with VTE incidence.

In contrast, the usage of statins was associated with a lower VTE incidence in our study (simvastatin 0.79; 0.72–0.88, fluvastatin 0.70; 0.49–0.99, atorvastatin 0.84; 0.77–0.97). Also, a dihydropyridine class calcium-channel blocker (DHB-CCB) and antihypertensive agent lercandipine had a significant protective association, with an aOR of 0.78 (0.62–0.97). Other DHB-CCBs, namely amlodipine and felodipine, also had aORs below 1, but the CIs were too wide to reach statistical significance. Digoxin, an agent that controls heart rate in atrial fibrillation, showed a protective association with an aOR of 0.72 (0.57–0.92).

Predefined exposures and associations with VTE

As predefined, we conducted a second analysis of the study drugs with selected variables (Supplemental Table 2). Due to the infrequent exposure, we were unable to draw precise conclusions. However, a long-acting granulocyte colony-stimulating factor pegfilgrastim had a high aOR for VTE incidence (17.17; 4.80–6.12E1) with 21 exposures among the cases and three in the control group.

Discussion

In our study, diabetes was an independent risk factor for VTE after adjustment for background variables. Additionally, inflammatory bowel disease and severe psychiatric disorders had significant associations with VTE in our diabetes-rich population, both of which are known risk factors in the general population.^1–5 Among socioeconomic groups, students and pensioners had the strongest associations. The reasons for that remain speculative. Pensioners may also be less mobile than the occupational population, and their age exceeds 60 years, which has been shown the enhance the propensity for VTE.^1–5 Diabetes medications that predispose to hypoglycemia (insulins and sulphonylureas) had stronger associations than metformin, which may reflect the reported prothrombotic effect of hypoglycemia.¹⁹ Our large, nested case-control study supports that the previously reported cancer type-related variation of thrombotic risk is also relevant in the Finnish diabetes-rich cohort. Many factors play a part in cancer-associated thrombosis, including the tumor cell biology, anatomic expansivity to the main vessels, therapy-related risks, and baseline characteristics. The most thrombotic cancer types in our study (pancreatic, stomach, ovarian, bronchus, and lung) are notorious for their extensive tissue factor expression, and in many cases, are diagnosed at the advanced stage, which also augments thrombogenicity.²⁰

Our pharmaco-epidemiological data suggested some previously unrecognized risk medications (such as gabapentin, pregabalin, pramipexole, memantine, donepezil) for VTE. Indeed, both widely used gabapentinoids, gabapentin and pregabalin, had a significant positive association with VTE incidence. As these agents are mainly used for chronic pain syndromes (e.g., diabetic neuropathy), this association may be a surrogate for pronounced immobilization and other thrombotic lifestyle habits, as well as serious baseline comorbidities that were not adjusted for. The pharmacodynamic effects of gabapentinoids include binding to the von Willebrand factor (VWF)-containing calcium-channel alfa-2 delta subunits (found in the central nervous system, and skeletal and cardiac muscle cells), the potential hemostatic effects of which have not been studied.^35,36 No data on the link between pramipexole and VTE is available, but the positive association in our study may reflect a thrombotic risk of Parkinson’s disease itself.³⁷ However, pramipexole is also widely used among the non-Parkinson’s population.³⁸ Memantine and donepezil, agents used for major cognitive impairment, have not previously been reported to carry thrombotic risk, but VTE has been reported as a major cause of death among people with advanced dementia, and immobilization and hypovolemia are overrepresented in this patient group.³⁹

Our data highlight the already known prothrombotic effects of exogenous glucocorticoids,^22,24 based on the increased levels of Factor VIII (FVIII) and VWF, leading to a hypercoagulable state.^22,24 Corticosteroids are also used for several systemic diseases with thromboinflammatory properties. Granulocyte colony-stimulating factors promote neutrophilia and monocytosis, which both associate with thrombogenesis.^20,26 These are usually used as supportive adjuncts in cancer care, but after adjustment for background variables, including cancers, pegfilgrastim still presented with a high aOR for VTE incidence, reflecting independent prothrombotic properties. Another positive association with risedronic acid may reflect the already reported thrombotic risk among osteoporosis patients that did not depend on whether they received pharmacological treatment.⁴⁰ Furosemide continued to have a significant association with VTE incidence after adjustment for chronic heart failure. It is also widely used for non-cardiogenic edema, and the positive association of furosemide may reflect baseline non-adjusted thrombogenic conditions such dehydration, nephrotic syndrome, or chronic venous insufficiency.^1–5

Severe psychiatric disorder as a baseline condition was associated with an increased risk of VTE. In accordance with this, psychotropic agents from different drug classes had significant positive associations with VTE. Exposure to antipsychotics and its relationship to VTE is an early observation, but the mechanisms and causation are unclear and complex. Sedation, obesity, and immobilization may play a central part, and drug-induced hyperprolactinemia is reported to enhance coagulation and possibly also platelet aggregation.^22,27,41,42 In our study, olanzapine, quietapine, and risperidone were all associated with VTE incidence. In addition to antipsychotics, several benzodiazepines had a positive, previously reported association.⁴³ Immobilization and sedation may be potential contributors to the thrombosis risk. All in all, thrombotic risk among patients with severe psychiatric disorders and psychotropic medications is important in clinical practice, and the threshold of adequate diagnostics should be low in cases of potential thrombotic symptoms. Considering the mean age of our cohort, psychotropic medications and benzodiazepines may also have been used to control dementia-related symptoms, which partly may explain the positive associations with VTE.³⁸

With respect to likely protective medications, our data align with the earlier positive results of statin being associated with lower VTE incidence.^28,29 Fluvastatin and simvastatin had the most significant protective associations. In diabetes population, lipophilic statin usage has been previously associated with marginally lower risk of VTE without statistical significance.⁴⁴ In addition, DHB-CCB lercanidipine was also associated with a significantly lower risk. Other DHB-CCBs, namely amlodipine and felodipine, also exerted non-significant protection, which suggests a potential group effect. Calcium ions are the major regulators of coagulation cascade, and previous evidence suggests that CCBs may have a protective association with stent thrombosis in coronary artery disease,⁴⁵ and vice versa, may potentiate the risk of GI bleeding in hypertensive patients.⁴⁶ However, the relationship between DHP CCBs and VTE has not been previously reported.

Our study has certain limitations. Prevalence of diabetes was high, and the mean age of the cohort was 65, which considerably limits the generalizability of the results. In our study, the definition of diabetes was based on the reimbursement and redeeming of at least one insulin product or oral antidiabetic agent. It is possible that some patients with early diabetes were treated only with life-style interventions, not with medication. However, as behavioral modification is challenging and may just delay the start of drug treatment, the Finnish Current Care guideline (original version published in 2007 and thereafter updated several times) recommends that metformin should be initiated concomitantly with life-style interventions unless contraindicated. These guidelines have been extensively implemented in primary care,⁴⁷ and therefore it is likely that those clinically diagnosed with diabetes were included in the study. We included the diagnoses I26, I63.6 and I80 due to their common real-world national use in clinical practice, but it is possible that some DVT diagnoses are not included in the I80 category. However, the large mortality study of PE, Finnish data included, reveals that among fatal DVT ICD codes, I80 predominates while I82 diagnoses are rare.⁴⁸

In some of the subgroups analyzed (especially regarding some cancer types and ambulatory medications), the sample sizes were small, compromising the statistical certainty of the results. For example, 95% CIs for the most prothrombotic cancers (pancreatic, stomach, ovarian, and bronchus and lung) were quite wide. All thrombogenicity-related baseline conditions (e.g., body mass index, autoimmune conditions, dyslipidemia, chronic kidney disease, hereditary and acquired thrombophilia)^1–5 could not be adjusted because of unavailable data. Additionally, information about major transient risk factors for VTE, like major surgery, infections or immobilization to name a few, could not be acquired in the frames of this study. We were unable to obtain data on antithrombotic medications, but we did adjust for chronic arrythmias and coronary artery disease, as many people with anticoagulation or antiplatelet medication have these diagnoses. We also excluded people with previous VTE events, and thus possible ongoing anticoagulation. As said, we excluded people with a history of VTE requiring hospital or tertiary care to make the associations less confounded by the high-risk patients with recurrent VTE. However, some less severe VTE events are diagnosed and treated by primary health care providers and could not be excluded. Positive associations between drugs and VTE incidence may sometimes instead reflect the thrombogenicity of the possible unadjusted baseline condition rather than the drug itself. Further research is required to establish the causation of the specific drugs and thrombosis.

Conclusion

VTE risk varied considerably according to the tumor type, and pancreatic, stomach, ovarian, bronchus, and lung cancers are associated with the highest VTE risk among a patient cohort of which 50% had diabetes. The thrombogenicity among people with major psychiatric disorders and/or psychotropic medication were highlighted in our cohort.

As novel observations, we found some previously unrecognized potential risk medications for VTE, but also a possible protective association of lercandipine and VTE, these data will require confirmation in other patient cohorts.

Supplemental Material

Supplemental Material - A pharmacoepidemiological nested case-control study of risk factors for venous thromboembolism with the focus on diabetes, cancer, socioeconomic group, medications, and comorbidities

Supplemental Material for A pharmacoepidemiological nested case-control study of risk factors for venous thromboembolism with the focus on diabetes, cancer, socioeconomic group, medications, and comorbidities by Lasse Myllylahti, Leo Niskanen, Riitta Lassila and Jari Haukka in Journal of Diabetes & Vascular Disease Research.

Footnotes

Author contributions

LM: conceptualization, writing original draft, review, and editing. LN: review and editing. RL: conceptualization, review, and editing. JH: data curation, conceptualization, statistical analysis, visualization, review, and editing.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Open Access was funded by Helsinki University Library.

ORCID iD

Lasse Myllylahti

Data availability statement

We are open to data sharing upon reasonable request.

Supplemental Material

Supplemental material for this article is available online.

References

Anderson

Jr Spencer

. Risk factors for venous thromboembolism. Circulation 2003; 107(23 Suppl 1): I9–16.

Kearon

. Natural history of venous thromboembolism. Circulation 2003; 107(23 Suppl 1): I22–I30.

Heit

Silverstein

Mohr

, et al. Risk factors for deep vein thrombosis and pulmonary embolism: a population-based case-control study. Arch Intern Med 2000; 160(6): 809–815.

White

. The epidemiology of venous thromboembolism. Circulation 2003; 107(23 Suppl 1): I4–I8.

Heit

Spencer

White

. The epidemiology of venous thromboembolism. J Thromb Thrombolysis 2016; 41(1): 3–14.

Laporte

Mismetti

Décousus

RIETE Investigators , et al. Clinical predictors for fatal pulmonary embolism in 15,520 patients with venous thromboembolism: findings from the registro informatizado de la enfermedad tromboembolica venosa (RIETE) registry. Circulation 2008; 117(13): 1711–1716.

Galanaud

Monreal

Kahn

. Epidemiology of the post-thrombotic syndrome. Thromb Res 2018; 164: 100–109.

Klok

van der Hulle

den Exter

, et al. The post-PE syndrome: a new concept for chronic complications of pulmonary embolism. Blood Rev 2014; 28(6): 221–226.

Kim

Delcroix

Jais

, et al. Chronic thromboembolic pulmonary hypertension. Eur Respir J 2019; 53(1): 1801915.

10.

Zimmet

Alberti

Magliano

, et al. Diabetes mellitus statistics on prevalence and mortality: facts and fallacies. Nat Rev Endocrinol 2016; 12(10): 616–622.

11.

Poznyak

Grechko

Poggio

, et al. The diabetes mellitus-atherosclerosis connection: the role of lipid and glucose metabolism and chronic inflammation. Int J Mol Sci 2020; 21(5): 1835.

12.

Alzahrani

Ajjan

. Coagulation and fibrinolysis in diabetes. Diabetes Vasc Dis Res 2010; 7(4): 260–273.

13.

Bell

Folsom

Lutsey

, et al. Diabetes mellitus and venous thromboembolism: a systematic review and meta-analysis. Diabetes Res Clin Pract 2016; 111: 10–18.

14.

Gariani

Mavrakanas

Combescure

, et al. Is diabetes mellitus a risk factor for venous thromboembolism? A systematic review and meta-analysis of case-control and cohort studies. Eur J Intern Med 2016; 28: 52–58.

15.

Heit

Leibson

Ashrani

, et al. Is diabetes mellitus an independent risk factor for venous thromboembolism? A population-based case-control study. Arterioscler Thromb Vasc Biol 2009; 29(9): 1399–1405.

16.

Hinton

Nemeth

de Lusignan

, et al. Effect of type 1 diabetes and type 2 diabetes on the risk of venous thromboembolism. Diabet Med 2021; 38(5): e14452.

17.

Deischinger

Dervic

Nopp

, et al. Diabetes mellitus is associated with a higher relative risk for venous thromboembolism in females than in males. Diabetes Res Clin Pract 2022; 194: 110190.

18.

Charlier

Meier

Jick

, et al. Association between glycemic control and risk of venous thromboembolism in diabetic patients: a nested case-control study. Cardiovasc Diabetol 2022; 21(1): 2.

19.

Chow

Iqbal

Walkinshaw

, et al. Prolonged prothrombotic effects of antecedent hypoglycemia in individuals with type 2 diabetes. Diabetes Care 2018; 41(12): 2625–2633.

20.

Timp

Braekkan

Versteeg

, et al. Epidemiology of cancer-associated venous thrombosis. Blood 2013; 122(10): 1712–1723.

21.

Christensen

Horváth-Puhó

Thomsen

, et al. Venous thromboembolism and risk of cancer in patients with diabetes mellitus. J Diabet Complicat 2016; 30(4): 603–607.

22.

Girolami

Cosi

Tasinato

, et al. Drug-induced thrombophilic or prothrombotic states: an underestimated clinical problem that involves both legal and illegal compounds. Clin Appl Thromb Hemost 2017; 23(7): 775–785.

23.

Stegeman

de Bastos

Rosendaal

, et al. Different combined oral contraceptives and the risk of venous thrombosis: systematic review and network meta-analysis. OM BMJ 2013; 347: f5298.

24.

Johannesdottir

Horváth-Puhó

Dekkers

, et al. Use of glucocorticoids and risk of venous thromboembolism: a nationwide population-based case-control study. JAMA Intern Med 2013; 173(9): 743–752.

25.

Shah

Capanu

Soff

, et al. Risk factors for developing a new venous thromboembolism in ambulatory patients with non-hematologic malignancies and impact on survival for gastroesophageal malignancies. J Thromb Haemostasis 2010; 8(8): 1702–1709.

26.

Zhang

. Risks of venous thromboembolism, stroke, heart disease, and myelodysplastic syndrome associated with hematopoietic growth factors in a large population-based cohort of patients with colorectal cancer. Clin Colorectal Cancer 2015; 14(4): e21–e31.

27.

Adelborg

Sundbøll

Videbech

, et al. The risk of thromboembolism in users of antidepressants and antipsychotics. Adv Exp Med Biol 2017; 906: 351–361.

28.

Lassila

Jula

Pitkäniemi

, et al. The association of statin use with reduced incidence of venous thromboembolism: a population-based cohort study. BMJ Open 2014; 4(11): e005862.

29.

Squizzato

Galli

Romualdi

, et al. Statins, fibrates, and venous thromboembolism: a meta-analysis. Eur Heart J 2010; 31: 1248–1256.

30.

Rahimi

Bhala

Kamphuisen

, et al. Effect of statins on venous thromboembolic events: a meta-analysis of published and unpublished evidence from randomised controlled trials. PLoS Med 2012; 9(9): e1001310.

31.

Sha

Zhang

, et al. Association of metformin use with risk of venous thromboembolism in adults with type 2 diabetes: a general-population-based cohort study. Am J Epidemiol 2022; 191(5): 856–866.

32.

Wang

Yang

Wang

, et al. Effects of sodium-glucose cotransporter 2 inhibitors on risk of venous thromboembolism in patients with type 2 diabetes: a systematic review and meta-analysis. Diabetes Metab Res Rev 2020; 36(1): e3174.

33.

Xin

Sun

Wang

, et al. Dipeptidyl peptidase 4 inhibitors and venous thromboembolism risk in patients with type 2 diabetes: a meta-analysis of cardiovascular outcomes trials. Thromb Haemostasis 2021; 121(1): 106–108.

34.

Niskanen

Partonen

Auvinen

, et al. Trends, and predictors in all-cause and cause-specific mortality in diabetic and reference populations during 21 years of follow-up. J Epidemiol Community Health 2020; 74(11): 950–956.

35.

Offord

Isom

. Drugging the undruggable: gabapentin, pregabalin and the calcium channel α2δ subunit. Crit Rev Biochem Mol Biol 2015; 51(4): 246–256.

36.

Cantí

Nieto-Rostro

Foucault

, et al. The metal-ion-dependent adhesion site in the Von Willebrand factor-A domain of alpha2delta subunits is key to trafficking voltage-gated Ca²⁺ channels. Proc Natl Acad Sci U S A 2005; 102(32): 11230–11235.

37.

Yamane

Kimura

Unoda

, et al. Postural abnormality as a risk marker for leg deep venous thrombosis in Parkinson’s disease. PLoS One 2013; 8(7): e66984.

38.

Gossard

Trotti

Videnovic

, et al. Restless legs syndrome: contemporary diagnosis and treatment. Neurotherapeutics 2021; 18(1): 140–155.

39.

Nuñez

Villalba

Cebrián

RIETE Investigators , et al. Venous thromboembolism in immobilized patients with dementia. Findings from the RIETE registry. Thromb Res 2012; 130(2): 173–177.

40.

Breart

Cooper

Meyer

, et al. Osteoporosis and venous thromboembolism: a retrospective cohort study in the UK general practice research database. Osteoporos Int 2010; 21(7): 1181–1187.

41.

Ishioka

Yasui-Furukori

Sugawara

, et al. Hyperprolactinemia during antipsychotics treatment increases the level of coagulation markers. Neuropsychiatric Dis Treat 2015; 11: 477–484.

42.

Chen

Shen

, et al. Antipsychotic use and risk of venous thromboembolism: a meta-analysis. Psychiatr Res 2021; 296: 113691.

43.

Chen

Winkelman

Mao

, et al. Real-world evidence on the use of benzodiazepine receptor agonists and the risk of venous thromboembolism. J Thromb Haemostasis 2020; 18(11): 2878–2888.

44.

Lin

. Association of venous thromboembolism between hydrophilic and lipophilic statin users among diabetic subjects. Medicine 2022; 101(36): e30542.

45.

Nagano

Hokimoto

Nakao

Kumamoto Intervention Conference Study KICS Investigators , et al. Relation between stent thrombosis and calcium channel blocker after drug-eluting stent implantation: kumamoto intervention conference study (KICS) registry. J Cardiol 2015; 66(4): 333–340.

46.

Kaplan

Heckbert

Koepsell

, et al. Use of calcium channel blockers and risk of hospitalized gastrointestinal tract bleeding. Arch Intern Med 2000; 160(12): 1849–1855.

47.

Niskanen

Hahl

Haukka

, et al. Type 2 diabetes and treatment intensification in primary care in Finland. Acta Diabetol 2018; 55(11): 1171–1179.

48.

Barco

Mahmoudpour

Valerio

, et al. Trends in mortality related to pulmonary embolism in the European region, 2000–15: analysis of vital registration data from the WHO mortality database. Lancet Respir Med 2020; 8(3): 277–287.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.43 MB