Abstract
Background
Venous thromboembolism (VTE) is a sex-specific disease that has different presentations between men and women. Women with uterine leiomyoma can present with VTE without exhibiting the traditional risk factors. We investigated the relationship between a history of uterine leiomyoma and the risk of VTE using the National Health Insurance Research Database (NHIRD).
Methods
We conducted a retrospective, nationwide, population-based case-control study using the NHIRD. We identified 2,282 patients with diagnosed VTE and 392,635 subjects without VTE from 2000 to 2013. After development of an age and index diagnosis year frequency-matched model and propensity score-matched model, 2 models with a case-to-control ratio of 1 to 4 were established. Using the diagnosis of uterine leiomyoma as the exposure factor, conditional logistic regression was performed to examine the association between uterine leiomyoma and VTE. Multiple logistic regression analysis was used to investigate the joint effect of uterine leiomyoma and comorbid diseases on the risk of VTE.
Results
A strong association was observed between uterine leiomyoma and VTE in the overall patient model, frequency-matched model and propensity score-matched model [p < 0.0001, odds ratio (OR): 1.547; p = 0.0005, OR: 1.486; p = 0.0405, OR: 1.26, respectively]. In the subgroup analyses, women with uterine leiomyoma who were ≥ 45 years old were less likely to experience VTE, but women with uterine leiomyoma and anemia, cancer, coronary artery disease or heart failure were more likely to experience VTE.
Conclusions
Women with uterine leiomyomas have an increased risk of developing VTE, especially during reproductive periods or in the presence of specific diseases.
Keywords: Uterine leiomyoma, Venous thromboembolism
INTRODUCTION
Venous thromboembolism (VTE) presents clinically as deep venous thrombosis (DVT), pulmonary embolism (PE) or both.1 VTE is associated with significant morbidity and mortality and a large socioeconomic burden. VTE has been shown to exhibit sex-specific differences,2 and women exposed to reproductive-related risk factors such as oral contraception, postmenopausal hormone therapy, and pregnancy have been reported to be at a higher risk of VTE than men.1,3 Therefore, women are at a higher risk of developing VTE during their fertile years,4,5 and men are at a higher risk of VTE than women at an older age.
Uterine leiomyomas, monoclonal tumors that arise from uterine smooth muscle tissue, are the most common benign tumors of the genital organs in women of childbearing age. They can cause significant morbidity, including menstrual abnormalities, iron deficiency anemia, and bulk symptoms with increased pelvic pressure.6 To the best of our knowledge, few case reports have described an association between uterine leiomyomas and VTE.7-14 In addition, some of these cases of uterine leiomyomas causing VTE have lacked the traditional risk factors for VTE.13,14
As no large-scale studies have investigated the association between uterine leiomyoma and VTE, we conducted this study to investigate the relationship between a history of uterine leiomyoma and the risk of VTE in Taiwanese women using data from the National Health Insurance Research Database (NHIRD).
METHODS
Data source and study patients
Data were retrieved from the NHIRD, which includes all claims data from the National Health Insurance program from 1996 to 2012. One million patients were randomly sampled from those enrolled in the Longitudinal Health Insurance Dataset (LHID) 2005, which is a subset of the NHIRD. We identified patients diagnosed with VTE in the LHID2005 dataset from 1996 to 2013. All diseases were identified according to International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes.
A flowchart of the subject selection process is shown in Figure 1. Patients diagnosed with VTE, including PE (ICD-9-CM code 415.1) and DVT (ICD-9-CM code 453.8), with at least three records in the claims data from January 1, 2000, to December 31, 2013, were identified and included in this study. The index date was defined as the date when VTE was first diagnosed. Patients < 18 years and > 100 years of age or with incomplete demographic information were excluded. A total of 2,282 female patients diagnosed with VTE and 392,635 subjects without VTE were included. To reduce bias due to confounding variables, two matched study cohorts were established using random frequency-matched (study cohort 1) and propensity score-matched methods (study cohort 2). For each identified VTE patient, four controls were frequency-matched by age and index year. In total, 2,282 patients with VTE and 9,128 age- and index year-matched subjects without VTE were included in study cohort 1. In addition, to balance the measured covariate distribution in the case-control patients, a propensity score for each patient was calculated. The propensity scores were calculated using multivariate logistic regression to predict the probability of the occurrence of VTE. Based on the propensity score, 1:4 propensity score matching was used for study cohort 2. A total of 2,227 patients with VTE and 8,690 propensity score-matched subjects without VTE were included in study cohort 2.
Figure 1.
Flowchart of patients with VTE included from the NHIRD and selection of the control population using age and frequency matching (analysis method 1) or propensity score matching (analysis method 2). DVT, deep venous thrombosis; PE, pulmonary embolism.
Supplementary Table 1. ICD-9-CM codes used to identify uterine leiomyoma, venous thromboembolism, and comorbid conditions.
| Diseases | Corresponding ICD-9-CM codes |
| Uterine leiomyoma | 218.0, 218.1, 218.2 and 218.9 |
| Venous thromboembolism | |
| Pulmonary artery embolism | 415.1 |
| Deep vein thrombosis | 453.8 |
| Co-morbid diseases | |
| Hypertension | 401.x–405.x |
| Diabetes mellitus | 250.x |
| Hyperlipidemia | 272.x |
| Chronic kidney disease | 580.x-589.x |
| Coronary artery disease | 410.x–414.x |
| Congestive heart failure | 428.x |
| Cardiac dysarrhythmia | 427.x |
| Peripheral artery occlusive disease | 443.x–444.x |
| COPD | 416.x, 490.x -505.x, 506.4, 508.x |
| Anemia | 280.0, 280.9, 285.x |
| All cancer | 140.x -208.x |
| Stroke | 430.x–438.x |
| Lower leg fracture or surgery | 820.x, 821.x |
| OP icd-9: 81.51, 81.52, 81.53, 81.54 |
COPD, chronic obstructive pulmonary disease; ICD-9-CM, International Classification of Disease, 9th Revision, Clinical Modification.
Outcome measures and relevant variables
Uterine leiomyoma exposure
A diagnosis of uterine leiomyoma was identified using ICD-9-CM codes (ICD-9 codes 218.0, 218.1, 218.2 and 218.9) and included submucosal leiomyoma of the uterus (ICD-9 code 218.0), intramural leiomyoma of the uterus (ICD-9 code 218.1), subserosal leiomyoma of the uterus (ICD-9 code 218.2), and unspecified leiomyoma of the uterus (ICD-9 code 218.9). In addition, the diagnosis of uterine leiomyoma was confirmed by the presence of at least three records containing a uterine leiomyoma diagnostic code as indicated by a gynecologist. To increase the reliability of the uterine leiomyoma diagnosis, we identified the patients who also received a gynecologic ultrasound examination (NHI procedure code: 199003 C) during the study period. We also performed sensitivity analysis to test the robustness of the diagnoses of uterine leiomyoma.
Other relevant variables
Major comorbid diseases diagnosed in at least three records in the claims data 1 year before the index date were defined as baseline comorbidities, and included hypertension, diabetes mellitus, hyperlipidemia, chronic kidney disease (CKD), chronic obstructive pulmonary disease (COPD), stroke, cardiac dysrhythmia, peripheral artery occlusive disease (PAOD), anemia, all cancers, coronary artery disease, heart failure and lower leg fractures or surgery. The ICD-9 codes are shown in the supplemental file. In addition, the long-term use of medications that were thought to be associated with VTE including anti-diabetic drugs, statins, hormone replacement agents, ferric agents, ferrous agents, and anti-hypertensive drugs, were also assessed.
Statistical methods
The demographic and clinical characteristics of the patients with VTE and the controls without VTE were summarized using numbers and percentages (%) for ca-tegorical variables, and means ± standard deviation (SD) for continuous variables. The chi-square test and t test were used to compare the distributions of discrete and continuous variables, respectively. Conditional logistic regression analysis was used to examine the association between uterine leiomyoma and VTE. Confounders including age, monthly income, major comorbidities, and long-term use of medications were adjusted for in the multiple conditional logistic regression analysis to estimate the adjusted odds ratios (aORs) and 95% confidence intervals (CIs). To improve the reliability of the results, we used three study cohorts and two different adjusted models including all of the clinical variables and covariate adjustments using the propensity scores. The joint effects of uterine leiomyoma and comorbid diseases on the risk of VTE between different groups are presented as aORs and 95% CIs, and multiple logistic regression analysis with backward elimination was performed. A two-tailed p value of < 0.05 was considered to be statistically significant. All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, NY).
RESULTS
Characteristics of the study population
The characteristics of the study population are listed and compared in Table 1. When considering the overall patients and the frequency-matched model, consistent differences existed in the baseline characteristics of cases and controls. The patients with VTE were more likely to be older and to have a lower monthly income than those without VTE. In addition, the patients with VTE were also more likely to have hypertension, hyperlipidemia, diabetes mellitus, obesity, CKD, COPD, coronary artery disease (CAD), heart failure, PAOD, cardiac arrhythmia, cancer, anemia, and lower leg fractures or surgery than those without VTE. Medications including anti-diabetic drugs, statins, hormonal replacement therapeutic agents, ferric tablets, ferrous tablets, and antihypertensive drugs were more frequently prescribed to the patients with VTE than those without VTE. In the propensity score-matched model, age and monthly income were similar between the groups. The prevalence rates of comorbidities, including hypertension, hyperlipidemia, diabetes mellitus, obesity, and stroke were similar in the patients with and without VTE. Compared to the patients without VTE, the patients with VTE were more likely to have CKD (16.66% vs. 12.51%, p < 0.001), COPD (20.25% vs. 18.09%, p = 0.019), CAD (26% vs. 23.62%, p = 0.019), heart failure (19.08% vs. 14.45%, p < 0.001), PAOD (5.52% vs. 3.25%, p < 0.001), cardiac dysrhythmia (14.55% vs. 12.23%, p = 0.003), cancer (17.83% vs. 14.26%, p < 0.001), anemia (16.26% vs. 12.62%, p < 0.001), lower leg fractures and surgery (15.49% vs. 12.61%, p < 0.001) and a history of long-term ferric or ferrous agent use. The use of anti-diabetic drugs, statins, hormone replacement therapeutic agents, and antihypertensive drugs was similar in both groups. Of the 2,282 patients diagnosed with VTE, 1,612 (70.64%) received anticoagulation therapy. Overall, 17,071 women were diagnosed with uterine leiomyomas, including 16,955 cases in the group without VTE and 116 cases in the group with VTE.
Table 1. Characteristics of cases and controls.
| Before matching | Frequency matching by age and index year | Propensity score matching | |||||||
| No PE/DVT (n = 392614) | PE/DVT (n = 2282) | p value | No PE/DVT (n = 9128) | PE/DVT (n = 2282) | p value | No PE/DVT (n = 8690) | PE/DVT (n = 2227) | p value | |
| Age | 43.15 ± 16.83 | 64.05 ± 16.11 | < 0.001 | 63.38 ± 16.16 | 64.05 ± 16.11 | 0.076 | 63.6 ± 16.08 | 64.01 ± 16.05 | 0.281 |
| Income | 15921 ± 13267 | 12339 ± 11740 | < 0.001 | 130233 ± 12668 | 12339 ± 11740 | 0.014 | 12473.84 ± 12295.71 | 12415 ± 11794.15 | 0.839 |
| Comorbidity disease at baseline | |||||||||
| Hypertension | 54776 (13.95%) | 1318 (57.76%) | < 0.001 | 3512 (38.48%) | 1318 (57.76%) | < 0.001 | 4914 (56.55%) | 1266 (56.85%) | 0.799 |
| Hyperlipidemia | 27001 (6.88%) | 581 (25.46%) | < 0.001 | 1481 (16.22%) | 581 (25.46%) | < 0.001 | 2100 (24.17%) | 550 (24.7%) | 0.602 |
| Diabetes mellitus | 24248 (6.18%) | 638 (27.96%) | < 0.001 | 1557 (17.06%) | 638 (27.96%) | < 0.001 | 2210 (25.43%) | 602 (27.03%) | 0.123 |
| Obesity | 754 (0.19%) | 26 (1.14%) | < 0.001 | 26 (0.28%) | 26 (1.14%) | < 0.001 | 74 (0.85%) | 22 (0.99%) | 0.539 |
| CKD | 7455 (1.9%) | 413 (18.1%) | < 0.001 | 475 (5.2%) | 413 (18.1%) | < 0.001 | 1087 (12.51%) | 371 (16.66%) | < 0.001 |
| Stroke | 11095 (2.83%) | 415 (18.19%) | < 0.001 | 876 (9.6%) | 415 (18.19%) | < 0.001 | 1390 (16%) | 388 (17.42%) | 0.104 |
| COPD | 19973 (5.09%) | 474 (20.77%) | < 0.001 | 1029 (11.27%) | 474 (20.77%) | < 0.001 | 1572 (18.09%) | 451 (20.25%) | 0.019 |
| CAD | 17199 (4.38%) | 612 (26.82%) | < 0.001 | 1247 (13.66%) | 612 (26.82%) | < 0.001 | 2053 (23.62%) | 579 (26%) | 0.019 |
| CHF | 5937 (1.51%) | 464 (20.33%) | < 0.001 | 483 (5.29%) | 464 (20.33%) | < 0.001 | 1256 (14.45%) | 425 (19.08%) | < 0.001 |
| PAOD | 1423 (0.36%) | 150 (6.57%) | < 0.001 | 97 (1.06%) | 150 (6.57%) | < 0.001 | 282 (3.25%) | 123 (5.52%) | < 0.001 |
| Cardiac dysarrhythmia | 8970 (2.28%) | 342 (14.99%) | < 0.001 | 578 (6.33%) | 342 (14.99%) | < 0.001 | 1063 (12.23%) | 324 (14.55%) | 0.003 |
| Cancer | 8217 (2.09%) | 428 (18.76%) | < 0.001 | 440 (4.82%) | 428 (18.76%) | < 0.001 | 1239 (14.26%) | 397 (17.83%) | < 0.001 |
| Anemia | 14468 (3.69%) | 397 (17.4%) | < 0.001 | 581 (6.37%) | 397 (17.4%) | < 0.001 | 1097 (12.62%) | 362 (16.26%) | < 0.001 |
| Lower leg fracture or surgery | 8123 (2.07%) | 363 (15.91%) | < 0.001 | 567 (6.21%) | 363 (15.91%) | < 0.001 | 1096 (12.61%) | 345 (15.49%) | < 0.001 |
| Medication use | |||||||||
| Diabetic drugs | 17475 (4.45%) | 466 (20.42%) | < 0.002 | 1143 (12.52%) | 466 (20.42%) | < 0.001 | 1624 (18.69%) | 437 (19.62%) | 0.315 |
| Statin | 15102 (3.85%) | 449 (19.68%) | < 0.001 | 943 (10.33%) | 449 (19.68%) | < 0.001 | 1543 (17.76%) | 419 (18.81%) | 0.246 |
| Hormonal replacement therapy | 22738 (5.79%) | 318 (13.94%) | < 0.001 | 979 (10.73%) | 318 (13.94%) | < 0.001 | 1201 (13.82%) | 310 (13.92%) | 0.903 |
| Ferric or Ferrous tablet | 2223 (0.57%) | 81 (3.55%) | < .0001 | 92 (1.01%) | 81 (3.55%) | < .0001 | 187 (2.15%) | 72 (3.23%) | 0.0028 |
| Antihypertensive drugs | 45928 (11.7%) | 1207 (52.89%) | < 0.001 | 2943 (32.24%) | 1207 (52.89%) | < 0.001 | 4389 (50.51%) | 1156 (51.91%) | 0.238 |
| Exposure | |||||||||
| Leiomyoma | 16955 (4.32%) | 116 (5.08%) | 0.073 | 330 (3.62%) | 116 (5.08%) | 0.001 | 356 (4.10%) | 111 (4.98%) | 0.065 |
| Propensity score | 0.01 ± 0.02 | 0.05 ± 0.08 | < 0.001 | 0.01 ± 0.04 | 0.05 ± 0.08 | < 0.001 | 0.03 ± 0.05 | 0.04 ± 0.06 | < 0.001 |
CAD, coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DVT, deep venous thrombosis; PAOD, peripheral arterial occlusive disease; PE, pulmonary embolism.
The association between leiomyoma and VTE
Table 2 shows the results of multivariate logistic regression analysis of predictive ability of uterine leiomyoma for VTE. Before matching the data, the multivariate logistic regression analysis showed that the patients with uterine leiomyomas were more likely to suffer from VTE (p < 0.0001, OR: 1.547, 95% CI: 1.27-1.88). When the frequency-matched and propensity score-matched data were included, the multivariate logistic regression analysis also demonstrated that the women with uterine leiomyomas were more susceptible to the development of VTE (p = 0.0005, OR: 1.486, 95% CI: 1.19-1.86; and p = 0.0405, OR: 1.26, 95% CI: 1.01-1.57, respectively, Table 2). Of the 17,071 subjects diagnosed with uterine leiomyoma, 15,586 (91.3%) received gynecologic ultrasound examinations. To test the robustness of the uterine leiomyoma diagnoses, we also performed sensitivity analysis of uterine leiomyoma and VTE, as identified using stringent criteria (subjects who received gynecologic ultrasound examinations and had at least three records of a uterine leiomyoma diagnostic code indicated by a gynecologist). This analysis showed that the association between uterine leiomyoma and VTE was in agreement with the results in Table 2 (Supplementary Table 2).
Table 2. Multivariate logistic regression model to assess the associations between uterine leiomyoma and VTE.
| VTE case | Control | Adjusted odds ratioModel 1 (95 % CI) | p value | Adjusted odds ratioModel 2 (95 % CI) | p value | |
| Before matching data | ||||||
| Non-leiomyoma | 2166 (94.92%) | 375659 (95.68%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma | 116 (5.08%) | 16955 (4.32%) | 1.464 (1.2-1.78) | 0.0001 | 1.547 (1.27-1.88) | < 0.0001 |
| Frequency matching data | ||||||
| Non-leiomyoma | 2166 (94.92%) | 8798 (96.38%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma | 116 (5.08%) | 330 (3.62%) | 1.485 (1.17-1.88) | 0.0011 | 1.486 (1.19-1.86) | 0.0005 |
| Propensity score matching data | ||||||
| Non-leiomyoma | 2116 (95.02%) | 8334 (95.90%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma | 111 (4.98%) | 356 (4.10%) | 1.287 (1.03-1.61) | 0.025 | 1.26 (1.01-1.57) | 0.0405 |
Model 1 adjusted for uterine leiomyoma and the significantly different variables in Table 1. Model 2 adjusted for uterine leiomyoma, age and propensity score.
VTE, venous thromboembolism.
Supplementary Table 2. Multivariate logistic regression model to assess the associations between uterine leiomyoma and VTE. We carried out a sensitivity analysis in diagnosis of uterine leiomyoma and VTE identified using a stringent criterion. The analysis showed that the association between uterine leiomyoma and VTE was in agreement with the results in Table 2. Sensitivity table:
| VTE case# | Control | Adjusted odds ratioModel 1 (95 % CI) | p-value | Adjusted odds ratioModel 2 (95 % CI) | p-value | |
| Before matching data | ||||||
| Non-leiomyoma | 1550 (96.15%) | 377760 (96.05%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma* | 62 (3.85%) | 15524 (3.95%) | 1.417 (1.087-1.846) | 0.01 | 1.444 (1.107-1.883) | 0.0067 |
| Frequency matching data | ||||||
| Non-leiomyoma | 1550 (96.15%) | 6243 (96.82%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma* | 62 (3.85%) | 205 (3.18%) | 1.645 (1.190-2.274) | 0.0026 | 1.371 (1.015-1.851) | 0.0397 |
| Propensity score matching data | ||||||
| Non-leiomyoma | 1536 (96.3%) | 6158 (97.13%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma* | 59 (3.7%) | 182 (2.87%) | 1.456 (1.071-1.979) | 0.0164 | 1.464 (1.081-1.982) | 0.0137 |
Model 1 adjusted for uterine leiomyoma and the significantly different variables in Table 1. Model 2 adjusted for uterine leiomyoma, age and propensity score.
* The diagnosis of leiomyoma was confirmed by gynecologic ultrasound. # The diagnosis of VTE was confirmed using patients received anticoagulation therapy.
VTE, venous thromboembolism.
Leiomyoma and comorbid disease status and the risk of VTE
The subgroup analyses are shown in Figure 2. We first performed subgroup analysis using the frequency-matched model. In the patients with uterine leiomyoma, no significant differences were noted in any of the data between those with an index age ≥ 45 years and those with an index age < 45 years. In the random frequency-matched model, the patients with an index age ≥ 45 years had a lower risk of VTE than those aged < 45 years (p = 0.009, OR: 0.498, 95% CI: 0.29-0.84). In addition, the patients with anemia and those receiving ferric or ferrous tablets were more likely to have VTE than those without anemia in both the patients overall (p = 0.003, OR: 1.996, 95% CI: 1.27-3.15) and in the random frequency-matched model (p = 0.017, OR: 2.12, 95% CI: 1.14-3.93). The patients with cancer were more likely to suffer from VTE than those without cancer in both the patients overall (p < 0.0001, OR: 6.030, 95% CI: 3.79-9.57) and in the random frequency-matched model (p < 0.0001, OR: 5.071, 95% CI: 2.44-10.54). The patients with CAD had a higher risk of VTE than those without CAD in both the overall patient model (p = 0.0005, OR: 2.853, 95% CI: 1.58-5.14) and the random frequency-matched model (p = 0.0096, OR: 2.791, 95% CI: 1.28-6.07). The patients with congestive heart failure (CHF) had a higher risk of VTE than those without CHF in both the patients overall (p < 0.0001, OR:3.852, 95% CI: 1.79-8.27) and the random frequency-matched model (p = 0.0042, OR: 6.172, 95% CI: 1.78-21.43). The patients with PAOD had a higher risk of VTE than those without PAOD (p = 0.0002, OR: 6.634, 95% CI: 2.45-17.89). However, no significant difference was observed between the patients with PAOD and those without PAOD in the random frequency-matched model (p = 0.2003, OR: 2.862, 95% CI: 0.57-14.3). In the subgroup analysis, no significant difference was observed between the patients with uterine leiomyoma receiving hormone therapy and those not receiving hormone therapy in the overall data analysis and in the random frequency-matched model.
Figure 2.
The subgroup analysis is shown in the forest plot. Abbreviations are in Table 1.
Leiomyoma in different age groups of patients not receiving hormone therapy and the risk of VTE
To clarify the association between uterine leiomyoma and VTE in different age groups of patients not receiving hormone therapy, we performed subgroup analyses of those aged ≤ 30, 31-40, 41-50, 51-60, and ≥ 61 years, as shown in Table 3. The patients with uterine leiomyoma aged ≤ 30 and 41-50 years had a higher risk of VTE than those without uterine leiomyoma before the data were matched (p = 0.0247, OR: 4.602, CI: 1.21-17.44; p = 0.0129, OR: 1.535, CI: 1.10-2.15). Both the results using the frequency-matched model and the propensity score-matched model were consistent with the results obtained in the overall patient sample.
Table 3. The risk of VTE in patients with uterine leiomyomas in different age group after excluding those receiving hormone therapy.
| Subgroup | Before matching data | Frequency matched data | Propensity score matched data | |||
| Adjusted odds ratio (95% CI) | p value | Adjusted odds ratio (95% CI) | p value | Adjusted odds ratio (95% CI) | p value | |
| ≤ 30 | 4.602 (1.21-17.44) | 0.0247 | 6.407 (1.57-26.2) | 0.0097 | 5.927 (1.45-24.24) | 0.0133 |
| 31-40 | 1.212 (0.55-2.69) | 0.636 | 1.361 (0.61-3.05) | 0.4534 | 0.951 (0.42-2.16) | 0.9049 |
| 41-50 | 1.535 (1.10-2.15) | 0.0129 | 1.607 (1.04-2.47) | 0.031 | 1.468 (1.05-2.05) | 0.0231 |
| 51-60 | 1.203 (0.83-1.74) | 0.3268 | 1.213 (0.84-1.75) | 0.3034 | 1.183 (0.82-1.70) | 0.3626 |
| ≥ 61 | 1.076 (0.60-1.93) | 0.8053 | 1.056 (0.47- 2.37) | 0.8943 | 0.956 (0.53-1.74) | 0.956 |
Leiomyoma in the patients without cancer who did not receive hormone therapy and the risk of developing VTE
Due to concerns about the joint effect of hormone therapy and cancer on VTE, we performed sensitivity analysis by excluding those receiving hormone therapy and cases with a diagnosis of cancer (Table 4), and the results remained consistent with the results in Table 2.
Table 4. Multivariate logistic regression model to assess the associations between uterine leiomyoma and VTE after excluding the patients with hormone therapy and cancer diagnosis in uterine leiomyoma comparison.
| VTE case# | Control | Adjusted odds ratioModel 1 (95 % CI) | p value | Adjusted odds ratioModel 2 (95 % CI) | p value | |
| Before matching data | ||||||
| Non-leiomyoma | 1539 (95.47%) | 348283 (95.97%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma* | 73 (4.53%) | 14623 (4.03%) | 1.632 (1.283-2.076) | < 0.0001 | 1.579 (1.245-2.003) | 0.0002 |
| Frequency matching data | ||||||
| Non-leiomyoma | 1539 (95.47%) | 7523 (96.61%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma* | 73 (4.53%) | 264 (3.39%) | 1.615 (1.220-2.137) | 0.0008 | 1.650 (1.256-2.168) | 0.0003 |
| Propensity score matching data | ||||||
| Non-leiomyoma | 1518 (95.47%) | 6249 (96.35%) | 1 (reference) | 1 (reference) | ||
| Leiomyoma* | 72 (4.53%) | 237 (3.65%) | 1.387 (1.056-1.823) | 0.0188 | 1.341 (1.023-1.758) | 0.0334 |
Model 1 adjusted for uterine leiomyoma and the significantly different variables in Table 1. Model 2 adjusted for uterine leiomyoma, age and propensity score.
VTE, venous thromboembolism.
DISCUSSION
In this study, we found that Taiwanese women with uterine leiomyoma were susceptible to venous thromboembolism. Strong correlations were consistently observed in the overall population before matching, and also after age- and index year-matching and propensity score-matching. Although the association between uterine leiomyoma and VTE has not been frequently reported, our findings support several previous case reports and case series studies. Previous studies have also reported that uterine leiomyomas, especially those with a large size, may cause DVT because of venous stasis due to mechanical compression of the iliac veins or inferior vena cava. The most frequent site of venous compression is the common iliac vein, regardless of whether on the left, right or both.14 Once DVT occurs in the large veins of the lower extremities, the thrombus is likely to detach and embolize, resulting in a PE sequence.1 Furthermore, in a case series study, Barsam et al. found that the site of the uterine leiomyoma was not always correlated to the site of DVT.13 It is possible that direct compression of the uterine leiomyoma is not the only cause of DVT. As pregnancy causes physical enlargement of the uterus but does not usually cause DVT, and as many women apparently have uterine leiomyomas without DVT, it seems that the size of the uterus may not be the only factor predisposing women to DVT. However, the presence of uterine leiomyomas may cause non-uniform enlargement of the uterus accompanied by shape changes, and this could easily predispose the uterus by impinging on the veins.15 The reason for the higher risk of VTE in the patients with uterine leiomyomas could also be explained by a hormone-related trend toward VTE in patients with uterine leiomyomas.13 Wolanska et al. found that an increased expression of type 1 basic fibroblast growth factor receptors, which act as a heparin-binding growth factor, in the endometrium of uterine leiomyoma may interfere with normal coagulation.16 An elevated expression of thrombospondin 1, a glycoprotein involved in the platelet adhesion response, has also been observed in uterine leiomyoma compared to normal endometria.17 Gokdeniz et al. identified a marked expression of endothelial nitrogen oxide synthase in uterine leiomyoma tissue compared with normal myometrium.18 They also found that nitrogen oxide could contribute to VTE by either mediating the growth-promoting effects of estrogen on uterine leiomyoma or by enhancing alterations in pelvic blood flow.18
The findings of the current study revealed that the patients with uterine leiomyomas who were older than 45 years of age had a lower risk of VTE than those with uterine leiomyomas who were younger than 45 years of age. Based on observational studies, the incidence of uterine leiomyomas initially increases during puberty, is most frequent during perimenopause, and then decreases rapidly after menopause.6 The incidence of uterine leiomyoma peaks at 40-44 years of age because of the natural regression in uterine leiomyomas after menopause.19,20 The decreasing incidence and impact of uterine leiomyomas in women older than 45 years of age may then contribute to the lower susceptibility to VTE. In our study, we found that in the women aged ≤ 30 and 40-50 years who did not receive hormone therapy, uterine leiomyomas were significantly associated with the risk of VTE. The possible explanations are as follows. The mean age of menopause in Taiwan is 49 years, but in some women, menopause can start from 40 years of age.21 Gurka et al. showed that women had a rapid increase in the risk of cardiovascular atherosclerosis during the menopausal transition, and this increased cardiovascular risk during this transitional period led to a greater susceptibility to arterial or venous thrombosis.22 In contrast, women aged ≤ 30 years with uterine leiomyomas in the current study had the highest risk of developing VTE (aOR: 5.927-6.407). The underlying mechanism of this finding is unknown, however it raised concerns about anticoagulation therapy in young subjects with uterine leiomyomas.
In subgroup analysis, the patients with uterine leiomyoma and comorbid diseases including anemia, cancer, CAD and CHF had an increased risk of VTE. We also found that the patients with uterine leiomyoma and anemia were more susceptible to VTE, regardless of whether anemia was managed with ferric or ferrous tablets. In patients with anemia, low blood viscosity may cause decreased secretion of antithrombotic mediators resulting in increased blood clotting.23 In addition, the severity of anemia has been shown to be correlated to the diameter of uterine leiomyomas and the menorrhagic period.24 Anemia may be related to occult cancer, and patients with cancer also exhibit a higher risk of developing VTE. We also demonstrated that the patients with both uterine leiomyoma and cancer had a greater risk of developing VTE. Cancer is well known to increase the risk of VTE, and approximately 15-18% of all cases of VTE are associated with cancer.25,26 VTE is also a preclinical marker of cancer, especially in the first year after the diagnosis of VTE.27 Anemia has been shown to increase the risk of VTE after a cancer diagnosis, further supporting this connection.28
The association between venous thromboembolism and atherosclerosis has been clarified in recent studies, and these two diseases are no longer considered to be separate clinical entities.29-31 Atherosclerosis is believed to be associated with the activation of platelets and blood coagulation and an increase in fibrin turnover to induce a pro-thrombotic state in the slow-flowing venous system.32 This study supported the strong correlation between CAD and VTE.
CHF is well known to be associated with the risk of VTE, and also to be an independent risk factor for VTE.33,34 Women with severe heart failure, as assessed from either the lower left ventricular ejection fraction or a higher serum N-terminal pro-brain natriuretic peptide level, appear to have a higher risk of VTE than those with less severe or no heart failure.35,36 Among patients with heart failure, both decreased cardiac output and patient immobility produce the negative effects in Virchow’s classic triad, namely, blood flow stasis, endothelial dysfunction, and abnormalities in blood constituents.33 Patients with heart failure are predisposed to thrombus formation because of their chronic inflammatory state.37 Consistent with previous studies, we demonstrated that the women with uterine leio-myoma and HF had a higher risk of VTE.
The management of uterine leiomyomas is well described in clinical practice guidelines and should be individualized depending on the patient’s symptomatology, size and location of the uterine leiomyoma, patient’s age, and concern for the preservation of fertility.38 The present study specifically addressed the risk of VTE in women with uterine leiomyomas and may provide therapeutic references for those women. In clinical application, if women are diagnosed with VTE during reproductive age and lack the traditional risk factors, a gynecological work-up should be performed to determine the presence of uterine leiomyoma.
Study limitations
There are several limitations to this study. The first limitation pertains to the use of an administrative database and the case control study. This limitation has been well described in other population-based studies using the NHIRD.39 Further prospective studies are needed to confirm our observations. Second, the accuracy and validity of diagnoses based on ICD-9 codes may be an issue that needs to be addressed. To achieve quality and accuracy in the diagnosis of a disease, the diagnoses of uterine leiomyoma and VTE in this study were determined according to the presence of at least three records of the appropriate diagnostic code as coded by specialists. Finally, the prevalence of uterine leiomyoma may be underestimated. Without menstrual abnormalities or bulk symptoms, women with uterine leiomyomas will not seek medical aid and will therefore not have a diagnosis of uterine leiomyoma recorded in NHIRD. In addition, women of reproductive age are more likely to seek medical aid because of menstrual abnormalities and pain symptoms related to uterine leiomyomas than older women.
CONCLUSIONS
Taiwanese women with uterine leiomyomas are susceptible to VTE. Women with specific comorbid diseases including anemia, cancer, coronary artery disease or heart failure and a younger age have a greater risk of developing VTE.
CONFLICTS OF INTEREST
None.
REFERENCES
- 1.Wang KL, Chu PH, Lee CH, et al. Management of venous thromboembolisms:part I. The consensus for deep vein thrombosis. Acta Cardiol Sin. 2016;32:1–22. doi: 10.6515/ACS20151228A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Marshall AL, Bartley AC, Ashrani AA, et al. Sex-based disparities in venous thromboembolism outcomes: a National Inpatient Sample (NIS)-based analysis. Vasc Med. 2017;22:121–127. doi: 10.1177/1358863X17693103. [DOI] [PubMed] [Google Scholar]
- 3.Goldhaber SZ. Risk factors for venous thromboembolism. J Am Coll Cardiol. 2010;56:1–7. doi: 10.1016/j.jacc.2010.01.057. [DOI] [PubMed] [Google Scholar]
- 4.Bleker SM, Coppens M, Middeldorp S. Sex, thrombosis and inherited thrombophilia. Blood Rev. 2014;28:123–133. doi: 10.1016/j.blre.2014.03.005. [DOI] [PubMed] [Google Scholar]
- 5.Roach RE, Lijfering WM, Rosendaal FR, et al. Sex difference in risk of second but not of first venous thrombosis: paradox explained. Circulation. 2014;129:51–56. doi: 10.1161/CIRCULATIONAHA.113.004768. [DOI] [PubMed] [Google Scholar]
- 6.Sparic R, Mirkovic L, Malvasi A, Tinelli A. Epidemiology of uterine myomas: a review. Int J Fertil Steril. 2016;9:424–435. doi: 10.22074/ijfs.2015.4599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Yun JK, Kim JB. Pulmonary thromboembolism caused by huge uterine myoma. Asian Cardiovasc Thorac Ann. 2015;23:1003. doi: 10.1177/0218492314532109. [DOI] [PubMed] [Google Scholar]
- 8.Khademvatani K, Rezaei Y, Kerachian A, et al. Acute pulmonary embolism caused by enlarged uterine leiomyoma: a rare presentation. Am J Case Rep. 2014;15:300–303. doi: 10.12659/AJCR.890607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nishikawa H, Ideishi M, Nishimura T, et al. Deep venous thrombosis and pulmonary thromboembolism associated with a huge uterine myoma--a case report. Angiology. 2000;51:161–166. doi: 10.1177/000331970005100210. [DOI] [PubMed] [Google Scholar]
- 10.Bonito M, Gulemi L, Basili R, et al. Thrombosis associated with a large uterine myoma: case report. Clin Exp Obstet Gynecol. 2007;34:188–189. [PubMed] [Google Scholar]
- 11.Yamaga J, Takahashi H, Ishihara A. Autopsy case of sudden death due to acute diffuse pulmonary thromboembolism that was caused by giant uterine myoma. Nihon Naika Gakkai Zasshi. 1999;88:1521–1523. [PubMed] [Google Scholar]
- 12.Ogawa N, Hayashi Y, Maehara T, et al. A surgically treated case of acute pulmonary embolism owing to deep vein thrombosis of the leg mainly caused by uterine myoma. Kyobu Geka. 1992;45:631–634. [PubMed] [Google Scholar]
- 13.Barsam S, Bagot C, Patel R, et al. Extrinsic venous compression: a sufficient explanation for venous thromboembolism due to massive fibroids? Thromb Haemost. 2006;96:694–696. [PubMed] [Google Scholar]
- 14.Fernandes FL, Dinardo CL, Terra-Filho M. Uterine myoma as a cause of iliac vein thrombosis and pulmonary embolism: common disease, rare complication. Respirol Case Rep. 2014;2:132–134. doi: 10.1002/rcr2.79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fletcher H, Wharfe G, Williams NP, et al. Venous thromboembolism as a complication of uterine fibroids: a retrospective descriptive study. J Obstet Gynaecol. 2009;29:732–736. doi: 10.3109/01443610903165545. [DOI] [PubMed] [Google Scholar]
- 16.Wolanska M, Bankowski E. Fibroblast growth factors (FGF) in human myometrium and uterine leiomyomas in various stages of tumour growth. Biochimie. 2006;88:141–146. doi: 10.1016/j.biochi.2005.07.014. [DOI] [PubMed] [Google Scholar]
- 17.Bodner-Adler B, Nather A, Bodner K, et al. Expression of thrombospondin 1 (TSP 1) in patients with uterine smooth muscle tumors: an immunohistochemical study. Gynecol Oncol. 2006;103:186–189. doi: 10.1016/j.ygyno.2006.02.023. [DOI] [PubMed] [Google Scholar]
- 18.Gokdeniz R, Mizrak B, Ozen S, Bazoglu N. Endothelial nitric oxide synthase expression in leiomyoma and parental myometrium. Gynecol Obstet Invest. 2000;49:132–136. doi: 10.1159/000010231. [DOI] [PubMed] [Google Scholar]
- 19.Wise LA, Palmer JR, Stewart EA, Rosenberg L. Age-specific incidence rates for self-reported uterine leiomyomata in the Black Women’s Health Study. Obstet Gynecol. 2005;105:563–568. doi: 10.1097/01.AOG.0000154161.03418.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kim DH, Kim ML, Song T, et al. Is myomectomy in women aged 45 years and older an effective option? Eur J Obstet Gynecol Reprod Biol. 2014;177:57–60. doi: 10.1016/j.ejogrb.2014.04.006. [DOI] [PubMed] [Google Scholar]
- 21.Chow SN, Huang CC, Lee YT. Demographic characteristics and medical aspects of menopausal women in Taiwan. J Formos Med Assoc. 1997;96:806–811. [PubMed] [Google Scholar]
- 22.Gurka MJ, Vishnu A, Santen RJ, DeBoer MD. Progression of metabolic syndrome severity during the menopausal transition. J Am Heart Assoc. 2016;5 doi: 10.1161/JAHA.116.003609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cagdas Can HT, Reyan Ucku. Investigation of relationship between blood hemoglobin level and acute pulmonary embolism in emergency setting. International Medical Journal. 2013;20:1–3. [Google Scholar]
- 24.Yang JH, Chen MJ, Chen CD, et al. Impact of submucous myoma on the severity of anemia. Fertil Steril. 2011;95:1769–1772. doi: 10.1016/j.fertnstert.2011.01.142. [DOI] [PubMed] [Google Scholar]
- 25.Heit JA, O'Fallon WM, Petterson TM, et al. Relative impact of risk factors for deep vein thrombosis and pulmonary embolism: a population-based study. Arch Intern Med. 2002;162:1245–1248. doi: 10.1001/archinte.162.11.1245. [DOI] [PubMed] [Google Scholar]
- 26.Caine GJ, Stonelake PS, Lip GY, Kehoe ST. The hypercoagulable state of malignancy: pathogenesis and current debate. Neoplasia. 2002;4:465–473. doi: 10.1038/sj.neo.7900263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sorensen HT, Svaerke C, Farkas DK, et al. Superficial and deep venous thrombosis, pulmonary embolism and subsequent risk of cancer. Eur J Cancer. 2012;48:586–593. doi: 10.1016/j.ejca.2011.10.032. [DOI] [PubMed] [Google Scholar]
- 28.Sanden P, Svensson PJ, Sjalander A. Venous thromboembolism and cancer risk. J Thromb Thrombolysis. 2017;43:68–73. doi: 10.1007/s11239-016-1411-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Milan M, Vedovetto V, Bilora F, et al. Further evidence in support of the association between venous thrombosis and atherosclerosis: a case-control study. Thromb Res. 2014;134:1028–1031. doi: 10.1016/j.thromres.2014.09.007. [DOI] [PubMed] [Google Scholar]
- 30.Andrei MC, Andercou A. Is there a link between atherothrombosis and deep venous thrombosis? Maedica. 2014;9:94–97. [PMC free article] [PubMed] [Google Scholar]
- 31.Jezovnik MK, Poredos P, Lusa L. Idiopathic venous thrombosis is associated with preclinical atherosclerosis. J Atheroscler Thromb. 2010;17:304–311. doi: 10.5551/jat.3079. [DOI] [PubMed] [Google Scholar]
- 32.Prandoni P, Bilora F, Marchiori A, et al. An association between atherosclerosis and venous thrombosis. N Engl J Med. 2003;348:1435–1441. doi: 10.1056/NEJMoa022157. [DOI] [PubMed] [Google Scholar]
- 33.Dean SM, Abraham W. Venous thromboembolic disease in congestive heart failure. Congest Heart Fail. 2010;16:164–169. doi: 10.1111/j.1751-7133.2010.00148.x. [DOI] [PubMed] [Google Scholar]
- 34.Tang L, Wu YY, Lip GY, et al. Heart failure and risk of venous thromboembolism: a systematic review and meta-analysis. Lancet Haematol. 2016;3:e30–e44. doi: 10.1016/S2352-3026(15)00228-8. [DOI] [PubMed] [Google Scholar]
- 35.Dries DL, Rosenberg YD, Waclawiw MA, Domanski MJ. Ejection fraction and risk of thromboembolic events in patients with systolic dysfunction and sinus rhythm: evidence for gender differences in the studies of left ventricular dysfunction trials. J Am Coll Cardiol. 1997;29:1074–1080. doi: 10.1016/s0735-1097(97)00019-3. [DOI] [PubMed] [Google Scholar]
- 36.Mebazaa A, Spiro TE, Buller HR, et al. Predicting the risk of venous thromboembolism in patients hospitalized with heart failure. Circulation. 2014;130:410–418. doi: 10.1161/CIRCULATIONAHA.113.003126. [DOI] [PubMed] [Google Scholar]
- 37.Chong AY, Lip GY. Viewpoint: the prothrombotic state in heart failure: a maladaptive inflammatory response? Eur J Heart Fail. 2007;9:124–128. doi: 10.1016/j.ejheart.2006.05.009. [DOI] [PubMed] [Google Scholar]
- 38.Vilos GA, Allaire C, Laberge PY, Leyland N. The management of uterine leiomyomas. J Obstet Gynaecol Can. 2015;37:157–178. doi: 10.1016/S1701-2163(15)30338-8. [DOI] [PubMed] [Google Scholar]
- 39.Sung SH, Chen TC, Cheng HM, et al. Comparison of clinical outcomes in patients undergoing coronary intervention with drug-eluting stents or bare-metal stents: a nationwide population study. Acta Cardiol Sin. 2017;33:10–19. doi: 10.6515/ACS20160608A. [DOI] [PMC free article] [PubMed] [Google Scholar]