Edited by: Andrea Gallioli
Fundació Puigvert
Marco Moschini
San Raffaele Hospital
Last update: June 2025
More infoTransurethral resection of the bladder tumor (TURBT) is a standard procedure in bladder cancer (BC), which is associated with low risk of venous thrombo-embolism (VTE). The aim of this study was to find the predictors of postoperative VTE in patients undergoing TURBT for BC.
Materials and methodsIn this retrospective cohort analysis, patients aged ≥ 18 years with BC diagnosis undergoing TURBT were identified in the Merative® Marketscan® Research de-identified databases in 2007−2021. Patients with prior VTE events were excluded. Preoperative diagnostic codes and outpatient prescriptions present in at least 1% of the cohort were recorded (205 variables). Then, logistic regressions were performed including each variable separately, all variables together, as well as variables selected by stepwise and Least Absolute Shrinkage and Selection Operator (LASSO) selection methods. Adjusted odds ratios (aOR) with 95% confidence intervals (CI) were calculated.
ResultsIn total, 132,425 patients were included in this study, with 1959 (1.5%) individuals diagnosed with postoperative VTE. Various malignant neoplasms diagnosed before BC were significant risk factors of postoperative VTE, with aOR reaching up to 2.26 (95% CI: 1.96–2.61). Another strong predictor of VTE was a diagnosis of nephritis, nephrotic syndrome, and nephrosis (aOR 1.67, 95% CI: 1.48–1.87 stepwise; aOR 1.65, 95% CI: 1.46–1.85 LASSO). Also, patients with diseases of the urinary system, non-specific symptoms, diseases of the respiratory system, anemias, and other cardiovascular diseases were associated with increased VTE risk. Regarding drugs, antidiabetic agents and gastrointestinal drugs reduced the probability of VTE.
ConclusionsNumerous preoperative factors have influence on the risk of VTE after TURBT. These findings might facilitate the clinical decision about the implementation of thromboprophylaxis in the appropriate patients.
La resección transuretral del tumor vesical (RTU) es un procedimiento habitual en el cáncer de vejiga (CV), que se asocia con un bajo riesgo de tromboembolismo venoso (TEV). El objetivo de este estudio fue determinar los factores predictivos de TEV postoperatorio en los pacientes sometidos a RTU por CV.
Materiales y métodosEn este análisis de cohorte retrospectivo se identificaron pacientes de ≥18 años con diagnóstico de CV sometidos a RTU en las bases de datos desidentificados de Merative® Marketscan® Research en 2007−2021. Se excluyeron los pacientes con eventos previos de TEV. Se registraron los códigos diagnósticos preoperatorios y las prescripciones ambulatorias presentes en al menos el 1% de la cohorte (205 variables). A continuación, se realizaron regresiones logísticas incluyendo cada variable por separado, todas las variables juntas, así como variables seleccionadas por los métodos de selección escalonada (stepwise) y LASSO (Least Absolute Shrinkage and Selection Operator). Se calcularon las odds ratio ajustadas (aOR) con intervalos de confianza (IC) del 95%.
ResultadosEn total, 132.425 pacientes fueron incluidos en este estudio, con 1.959 (1,5%) individuos diagnosticados de TEV postoperatorio. Diversas neoplasias malignas diagnosticadas antes de la CV constituyeron factores de riesgo significativos de TEV postoperatorio, con una OR hasta 2,26 (IC 95%: 1,96–2,61). Otro factor predictivo importante de TEV fue el diagnóstico de nefritis, síndrome nefrótico y nefrosis (aOR 1,67; IC 95%: escalonada 1,48–1,87; aOR 1,65; IC 95%: LASSO 1,46–1,85). Asimismo, los pacientes con enfermedades del sistema urinario, síntomas inespecíficos, enfermedades del sistema respiratorio, anemia y otras enfermedades cardiovasculares se asociaron a un mayor riesgo de TEV. En cuanto a fármacos, los antidiabéticos y los gastrointestinales redujeron la probabilidad de TEV.
ConclusionesNumerosos factores preoperatorios influyen en el riesgo de TEV tras la RTU. Estos hallazgos podrían facilitar la decisión clínica sobre la implementación de profilaxis tromboembólica en los pacientes adecuados.
Transurethral resection of the bladder tumor (TURBT) is a standard diagnostic procedure in bladder cancer (BC).1 In addition, in non-muscle invasive bladder cancer (NMIBC) TURBT is a curative surgery that allows us to avoid extensive operations, such as radical cystectomy (RC). Despite its minimally invasive character, TURBT is associated with some risk of complications, including venous thrombo-embolism (VTE) that can develop in approximately 3% of the patients.2
It has been documented that the occurrence of VTE has a negative influence on perioperative outcomes of various surgeries.3,4 Therefore, it is crucial to prevent it, by administering anticoagulant medications in chosen patients. A variety of scales have been developed in order to assess the risk of VTE that help to select patients that require thromboprophylaxis, such as “EAU guidelines on thromboprophylaxis” or Caprini score.5,6 However, the 2024 EAU guidelines discontinued the “thromboprophylaxis” chapter and Caprini 2009 score has a sensitivity and specificity of 59% and 57%, respectively, in inpatients with cancer.1,7 That is why clinicians need to make arbitrary decisions in more complex cases.
In addition, fulguration is often performed in an outpatient setting, so the routine VTE risk assessment might be omitted. Furthermore, urologists might be concerned with a postoperative bleeding risk and not start thromboprophylaxis, even in patients with indications to such treatment.
The aim of this study was to find the predictors of postoperative VTE in patients undergoing TURBT for BC.
Materials and methodsThis retrospective cohort analysis is in accordance with the STROBE Statement.
Data sourceCertain data were supplied by Merative as part of one or more MarketScan Research Databases. Any analysis, interpretation, or conclusion based on these data is solely that of the authors and not Merative.
This study was based on administrative insurance claims data from the Merative™Marketscan® Research Commercial and Medicare databases (https://doi.org/10.57761/n5v8-0v21).8 They consist of individual-level, de-identified, demographic data, diagnoses, procedures, overall health-related costs, inpatient and outpatient pharmacy billing claims, and treatments allowing for longitudinal tracking of patients. International Classification of Disease Ninth and Tenth Revisions, Clinical Modification and Procedure Coding System (ICD-9-CM, ICD-10-CM, ICD-10-PCS), Current Procedural Terminology (CPT), and Healthcare Procedure Coding System (HCPCS) codes were used to identify the cohort of interest, treatments, and comorbidities. This method has been used in other studies3,4,9 and given the de-identified information, the study was deemed exempt from informed consent requirements by the Stanford University Medical Center Institutional Review Board. Data for this study was accessed using the Stanford Center for Population Health Sciences (PHS) Data Core. The PHS Data Core is supported by a National Institutes of Health National Center for Advancing Translational Science Clinical and Translational Science Award (UL1TR003142) and from Internal Stanford funding.
PatientsUsing the ICD-9/10-CM, ICD-10-PCS, and CPT diagnosis/treatment codes for BC and TURBT, data of patients at least 18 years old who underwent TURBT for BC between 2007 and 2021 was reviewed in order to create the cohort of interest. Patients selected for analysis were enrolled in the database for at least 3 months before and 3 months following initial TURBT. Time of first TURBT was designated as the index date for further assessment. Within this cohort, postoperative VTE diagnosis codes were identified and reviewed to ensure the appropriate selection of patients. VTE was defined as pulmonary embolism, lower or upper extremity deep vein thrombosis, or other peripheral or superficial phlebitis/thrombophlebitis that was present on admission. Also, patients with prior VTE events were excluded. The study cohort was further stratified into sub-groups based on the presence of VTE after TURBT (“Postop-VTE”) vs. no history of VTE after TURBT (“No Postop-VTE”). It was additionally stratified to minor TURBT vs. major TURBT. ‘Minor’ TURBT included procedures limited to cystoscopy with biopsy and/or fulguration, differently from ‘major’ which identified procedures being more involved and requiring resection. For each patient, sociodemographic data, including age at the index date, gender, US region, and insurance status, were initially recorded. Charlson Comorbidity Index (CCI) was calculated according to Charlson et al.10 and adapted according to Deyo et al.11 Also, specific comorbidity prevalence, length of stay, discharge status, additional interventions, total cost, complications, and re-hospitalizations were recorded. Treatment characteristics included the TURBT procedural data (tumor size, number of TURBTs) and the utilization of perioperative modalities relevant to TURBT, such as administration of intravesical chemotherapy. A flow chart diagram summarizing the analytical steps for the data analysis and inclusion/exclusion criteria was shown in Fig. 1 and a detailed list of implemented ICD-9/10, CPT, and HCPCS codes was presented in Supplementary Table1.
Outcome ascertainment
The main endpoint of this study was to find the main predictors of VTE occurrence after TURBT procedure in US patients diagnosed with BC. In order to achieve this goal, we examined the impact of 205 preoperative variables on the risk of postoperative VTE. The list of the preoperative variables was presented in Supplementary Table2.
Statistical analysisContinuous variables were reported using means ± standard deviations (SDs) or medians and interquartile ranges (IQR). Categorical variables were presented as counts and percentages (%). The statistical analysis involved using the Chi-square test for comparison of categorical data, the Student's t-test for age, and the Wilcoxon rank-sum test for other continuous variables all stratified by postoperative VTE occurrence.
The following steps were taken to investigate the risk factors associated with an increased probability of VTE after TURBT. First, preoperative diagnostic codes and outpatient prescriptions present in at least 1% of the cohort were recorded (205 variables). Then, univariate logistic regressions for each variable and multivariate logistic regression for all of the included variables were applied. In addition, logistic regressions were calculated for variables selected by stepwise selection method and by Least Absolute Shrinkage and Selection Operator (LASSO) selection method. The particular variables, for which regressions were inconsistent with well-established medical knowledge, were manually excluded. Finally, regressions for 15 variables from stepwise and 10 variables from LASSO selection methods were included. In both models each regression was adjusted by all of the chosen variables. All analyses were two-sided with, P < .05 considered significant, and performed using statistical software SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA).
ResultsStudy cohortIn total, a large cohort of 132,425 patients with BC diagnosis, who underwent TURBT between 2007 and 2021 in the US was included in this study. However, only 1959 (1.5%) individuals were diagnosed with postoperative VTE.
Baseline patients’ demographic, clinical, and hospital characteristics stratified into “postop-VTE” and “no postop-VTE” subgroups were presented in Table 1. Patients who presented VTE events after TURBT were significantly older (median 70 vs. 66 y.o.) and had one year shorter median follow-up (median 1.3 vs. 2.3 years). The subgroups were balanced in terms of gender, US region, data source, and insurance type. Also, postop-VTE patients had a higher median CCI (median 3 vs. 2). The analysis of specific comorbidities found that patients who developed postoperative VTE events had significantly higher rates of renal failure (27.0% vs. 13.8%) and hematologic diseases (34.5% vs. 24.4%). Interestingly, 13.5% of the patients in both subgroups were smokers. Moreover, postop-VTE patients had more large tumors (55.7% vs. 35.2%) and were treated using ‘major’ TURBT more commonly (45.7% vs. 33.8%). Finally, those with VTE after TURBT had lower rates of intravesical immunotherapy (12.8% vs. 23%), but underwent RC more often (24.4% vs. 7.4%).
Baseline patient demographic, clinical and hospital characteristics of the final cohort of the study according to the diagnosis of VTE after TURBT (postop-VTE vs. no postop-VTE).
| Postop-VTE | No Postop-VTE | P-value | |
|---|---|---|---|
| Number, (%) | 1959 (1.5) | 130,466 (98.5) | |
| Age, median (IQR) | 70 (61–80) | 66 (58–77) | <.0001 |
| 1st quartile | 388 (19.8) | 33,966 (26.0) | <.0001 |
| 2nd quartile | 445 (22.7) | 32,940 (25.2) | |
| 3rd quartile | 508 (25.9) | 31,738 (24.3) | |
| 4th quartile | 618 (31.5) | 31,822 (24.4) | |
| Follow up years, median (IQR) | 1.3 (0.6–3.1) | 2.3 (1.1–4.6) | <.0001 |
| Gender, n (%) | .5991 | ||
| Male | 1359 (69.4) | 91,223 (69.9) | |
| Female | 600 (30.6) | 39,243 (30.1) | |
| US region, n (%) | <.0001 | ||
| Northeast | 460 (23.5) | 31,100 (23.8) | |
| North Central | 593 (30.3) | 34,752 (26.6) | |
| South | 561 (28.6) | 43,485 (33.3) | |
| West | 290 (14.8) | 18,503 (14.2) | |
| Unknown | 55 (2.8) | 2626 (2.0) | |
| Data source, n (%) | <.0001 | ||
| Fee for Service | 690 (35.2) | 55,430 (42.5) | |
| Encounter | 82 (4.2) | 7097 (5.4) | |
| Medicare | 1028 (52.5) | 59,997 (46.0) | |
| Medicare Encounter | 156 (8.0) | 7799 (6.0) | |
| Insurance type, n (%) | .007 | ||
| Comprehensive | 490 (25.0) | 28,949 (22.2) | |
| HMO | 225 (11.5) | 14,082 (10.8) | |
| PPO | 966 (49.3) | 66,932 (51.3) | |
| Other | 278 (14.2) | 20,503 (15.7) | |
| CCI, median (IQR) | 3 (2–6) | 2 (1–4) | <.0001 |
| 0−1 | 364 (18.6) | 37,832 (29) | <.0001 |
| 2−4 | 916 (46.8) | 66,571 (51.0) | |
| ≥5 | 679 (34.7) | 26,063 (20.0) | |
| Prevalent comorbidities, n (%) | |||
| Obesity | 223 (11.4) | 12,136 (9.3) | .0017 |
| Diabetes | 573 (29.2) | 34,609 (26.5) | .0068 |
| Hypertension | 1333 (68.0) | 80,169 (61.4) | <.0001 |
| Renal failure | 529 (27.0) | 17,975 (13.8) | <.0001 |
| Smoking | 265 (13.5) | 17,623 (13.5) | .9799 |
| Infectious | 674 (34.4) | 37,469 (28.7) | <.0001 |
| Hematologic | 675 (34.5) | 31,778 (24.4) | <.0001 |
| Anticoagulant prophylaxis, n (%) | |||
| 90-days before TURBT | 245 (12.5) | 11,926 (9.1) | <.0001 |
| 90-days after TURBT | 922 (47.1) | 11,723 (9.0) | <.0001 |
| Tumor size, n (%) | n = 1505 | n = 90,666 | <.0001 |
| Small (≤ 2 cm) | 215 (14.3) | 22,893 (25.2) | |
| Medium (2–5 cm) | 452 (30.0) | 35,845 (39.5) | |
| Large (>5 cm) | 838 (55.7) | 31,928 (35.2) | |
| Minor TURBT, n (%) | <.0001 | ||
| No | 896 (45.7) | 44,109 (33.8) | |
| Yes | 1063 (54.3) | 86,357 (66.2) | |
| Re-TURBT < 90 days from TURBT, n (%) | 387 (19.7) | 23,178 (17.8) | .0223 |
| Time from TURBT to Re-TURBT, weeks, median (IQR) | 3.4 (1.4–6.1) | 4.1 (1.3–7.1) | .0505 |
| Intravesical chemotherapy, n (%) | 459 (23.4) | 44,081 (33.8) | <.0001 |
| PIC (≤ 14 days after TURBT) | 135 (29.4) | 11,774 (26.7) | <.0001 |
| Adjuvant CHT (> 90 days after TURBT) | 173 (37.7) | 13,207 (30.0) | |
| Intravesical BCG after TURBT, n (%) | 251 (12.8) | 30,004 (23.0) | <.0001 |
| Time from TURBT to intravesical BCG, months, median (IQR) | 2.7 (1.4–5.7) | 1.9 (1.1–5.0) | .0023 |
| RC after TURBT (i.e., Primary MIBC), n (%) | 479 (24.4) | 9679 (7.4) | <.0001 |
| Time from TURBT to RC, months, median (IQR) | 1.9 (1.2–4.6) | 4.0 (1.6–7.9) | <.0001 |
| Concomitant UTUC, n (%) | 73 (3.7) | 2543 (1.9) | <.0001 |
| RNU before TURBT, n (%) | 26 (1.3) | 1120 (0.9) | .0262 |
| Time to RNU before TURBT, months, median (IQR) | 4.7 (2.7–15.8) | 7.3 (4.3–14.2) | .0706 |
| RNU after TURBT, n (%) | 96 (4.9) | 2677 (2.0) | <.0001 |
| Time to RNU after TURBT, months, median (IQR) | 1.9 (1.0–5.3) | 3.6 (1.4–14.7) | .0004 |
VTE: venous thromboembolism; SD: standard deviation; IQR: interquartile range; n: number; US: United States; HMO: Health Maintenance Organization; PPO: Preferred Provider Organization; CCI: Charlson Comorbidity Index; TURBT: transurethral resection of bladder tumor; PIC: postoperative intravesical chemotherapy; CHT: chemotherapy; BCG: Bacillus Calmette–Guérin; RC: radical cystectomy; MIBC: muscle-invasive bladder cancer; UTUC: upper tract urothelial carcinoma; RNU: radical nephroureterectomy.
The results of stepwise and LASSO regressions for selected variables were presented in Table 2, using adjusted odds ratios (aOR) with 95% confidence intervals (CI).
Stepwise and LASSO regressions for the influence of the selected variables on the risk of postoperative venous thromboembolism.
| Variable | Stepwise regression | LASSO regression | ||
|---|---|---|---|---|
| aOR (95% CI) | P-value | aOR (95% CI) | P-value | |
| Malignant Neoplasm of Other and Unspecified Sites | 2.07 (1.78–2.40) | <.0001 | 2.26 (1.96–2.61) | <.0001 |
| Malignant Neoplasm of Genitourinary Organs | 1.23 (1.12–1.36) | <.0001 | 1.21 (1.10–1.33) | <.0001 |
| Malignant Neoplasm Of Digestive Organs And Peritoneum | 1.54 (1.26–1.87) | <.0001 | NR | NR |
| Nephritis, Nephrotic Syndrome, and Nephrosis | 1.67 (1.48–1.87) | <.0001 | 1.65 (1.46–1.85) | <.0001 |
| Other Diseases of the Urinary System | 1.47 (1.24–1.74) | <.0001 | 1.49 (1.26–1.76) | <.0001 |
| Other Disorders of the Female Genital Tract | 1.08 (0.97–1.20) | .1818 | NR | NR |
| Non-specific Symptoms | 1.47 (1.25–1.72) | <.0001 | 1.49 (1.28–1.74) | <.0001 |
| Other Diseases of the Respiratory System | 1.23 (1.1–1.37) | .0003 | 1.25 (1.12–1.40) | <.0001 |
| Chronic Obstructive Pulmonary Disease and Allied Conditions | 1.09 (0.98–1.21) | .1114 | NR | NR |
| Other and Unspecified Anemias | 1.16 (1.03–1.30) | .0151 | 1.17 (1.05–1.32) | .0066 |
| Diseases of Arteries, Arterioles, and Capillaries | 1.14 (1.03–1.27) | .0149 | 1.14 (1.02–1.26) | .0214 |
| Diseases of Veins and Lymphatics, and Other Diseases Of Circulatory System | 1.08 (0.96–1.20) | .2018 | NR | NR |
| Other Forms of Heart Disease | NR | NR | 1.14 (1.03–1.26) | .012 |
| Quinolones, NEC | 1.09 (0.99–1.20) | .093 | NR | NR |
| Antidiabetic Agents, Misc | 0.85 (0.74−0.98) | .0217 | 0.86 (0.75−0.98) | .0236 |
| Gastrointestinal Drugs Misc, NEC | 0.76 (0.68−0.86) | <.0001 | NR | NR |
aOR: adjusted Odds Ratio; CI: confidence interval; NR: not reported.
Various malignant neoplasms (of other and unspecified sites, genitourinary organs, digestive organs, and peritoneum) diagnosed before BC were significant risk factors for postoperative VTE, with aOR reaching up to 2.26 (95% CI: 1.96–2.61) in LASSO regression. Another strong predictor of VTE after TURBT was a diagnosis of nephritis, nephrotic syndrome, and nephrosis (aOR 1.67, 95% CI: 1.48–1.87 stepwise; aOR 1.65, 95% CI: 1.46–1.85 LASSO regression). Patients with diseases of the urinary system were also more likely to develop VTE (aOR 1.47, 95% CI: 1.24–1.74 stepwise; aOR 1.49, 95% CI: 1.26–1.76 LASSO regression). It is worth mentioning that non-specific symptoms were associated with up to 49% increase of VTE risk in LASSO regression. Other diagnostic codes that were significantly associated with VTE occurrence included: diseases of the respiratory system, anemias, and other cardiovascular diseases. Nonetheless disorders of the female genital tract and chronic obstructive pulmonary disease (COPD) failed to reach statistical significance as VTE risk factors.
Regarding drugs, antidiabetic agents and gastrointestinal drugs reduced the probability of VTE by 15% and 24%, respectively. On the other hand, administration of quinolones was a statistically insignificant risk factor of postoperative VTE occurrence.
DiscussionThe present study focused on preoperative factors that might influence the development of VTE after the TURBT procedure. This analysis did not provide a full overview of the predictors of postoperative VTE. Therefore, it should not be treated as a strong guideline for thromboprophylaxis implementation. Rather, it could be used as an ancillary tool in more complex clinical cases. To our knowledge, no studies regarding this particular topic are available in the literature. However, Zheng et al. found that the history of VTE, postoperative bladder hematoma, age > 65 years, and d-dimer > 1.25 mg/L were associated with a higher risk of VTE development following transurethral resection of the prostate.12 It is worth noting that the aforementioned factors were not included in this analysis.
In our research, due to a large number of identified variables, stepwise and LASSO selection methods were independently applied. As a result, complexity of the regression models was reduced and the most relevant preoperative variables were identified and included.
Even though the thrombotic complications post TURBT are rare, they might hinder the appropriate oncological treatment and deteriorate clinical outcomes. It has been proven that VTE before RC increased morbidity, hospital length of stay, rehospitalizations, and health-care costs.3 Crucially, the patients with preoperative VTE diagnoses were excluded from this study on purpose, as the risk of consequent VTE occurrence is so high that it would preclude the proper evaluation of the other variables. Nonetheless, a very strong association between preoperative and postoperative thrombotic events should be kept in mind.
Our study has brought out some interesting and novel associations post TURBT. According to the present analysis, malignant neoplasms diagnosed before the BC diagnosis were the strongest risk factor of VTE occurrence. These results are not surprising, as cancer-associated thrombosis is a common issue in clinical practice. In the literature, patients with cancer were reported to have up to 12 times higher risk of 6-month VTE development than those without neoplasms.13
Although not commonly known, nephritis, nephrotic syndrome, and nephrosis are also typical diseases that pose a risk of VTE, which is similarly in line with the results of this study.14,15 Occurrence of thrombotic events in kidney diseases is most probably caused by the imbalance of the pro/anti-thrombotic and pro/anti-coagulant factors.15
One of the most common diseases of the urinary and respiratory systems are infections. Urinary tract infections (UTI) have a reported lifetime incidence of 50%–60% in adult women, while respiratory infections are extremely common in the whole population.16,17 There is precedent in the literature for this. Along with other acute infections, UTI and respiratory tract infections were proven risk factors of VTE in a population study in Northern Denmark, which is in agreement with our results.17 The mechanism behind this phenomenon is not well elucidated, but is probably associated by the activation of platelets caused by inflammation.18
Non-specific symptoms include a variety of symptoms from all systems. In this analysis, presentation with symptoms increased the risk of VTE by 47%–49%. Our hypothesis behind this phenomenon is that these patients might have been comorbid and therefore have an initial high risk of VTE. Moreover, the symptoms might have urged the physicians to order additional tests and many of these cases were probably diagnosed incidentally.
Similarly to this research, other studies reported significantly increased the risk and incidence of VTE in patients with autoimmune hemolytic and aplastic anemias.19,20
Cardiovascular risk factors are also associated with VTE incidence.21 It is not surprising, as the relationship between cardiovascular diseases, including heart failure, and VTE is well established.22 The results of this study confirmed this association as well.
Disorders of the female genital tract and COPD were not significantly associated with VTE events. However, disorders of the female genital tract might cause elevated estrogen levels or require hormonal treatment, leading to subsequent thrombotic complications.23 Moreover, in a population-based cohort, patients with COPD were found to have 2 times higher risk of VTE than the general public.24
We have also examined the influence of drugs on VTE risk. Antidiabetic agents were protective of VTE, which was confirmed in the literature. A systematic review from 2022 found that metformin reduced the risk of VTE by 22%–58%.25 Also, gastrointestinal drugs decreased the risk of VTE by 0.76 times. Unfortunately, we were unable to provide a scientific rationale for this result. Quinolones were not a significant risk factor of VTE events. However, it has been proven that infections treated with antibiotics were associated with a 5.5-fold higher incidence of VTE in the first 2 weeks.26 Therefore, it is most probable that acute infections are a true reason for the development of thrombosis.
Our research has some limitations that need to be disclosed. It uses insurance claims data and codes which could be inaccurate, leading to coding and misclassification errors and, in turn, lack generalizability to the world over. Also, the design of this study is retrospective in nature. In addition, there is a possibility of underreporting of VTE, as asymptomatic patients might have been missed. The manual exclusion of the clinically unrelated variables may have restricted the scope of the analysis. Our paper included a large number of patients and variables, which can be both a strength and a drawback. The results of this study might be statistically significant, due to the large sample, without being clinically meaningful. On the other hand, the substantial number of included patients and resultant statistical power, allowed the assessment of post-TURBT VTE events, which are rare. Moreover, the present analysis used highly advanced stepwise and LASSO variable selection models, which increased the reliability of the results. Finally, to our knowledge, it is the first article on this important topic with significant statistical power.
ConclusionsTURBT is a standard diagnostic procedure in BC. The preoperative risk factors of VTE events after TURBT include: malignant neoplasms, nephritis, nephrotic syndrome, nephrosis, diseases of the urinary system, non-specific symptoms, diseases of the respiratory system, and cardiovascular diseases. VTE-protective variables are antidiabetic agents and gastrointestinal drugs. These findings might facilitate the clinical decision about the implementation of thromboprophylaxis in the appropriate patients.
Relevant disclosuresAuthors have nothing to disclose.
Authors have no conflict of interest to disclose.
Merative™ Marketscan® Research Databases.
The following are Supplementary data to this article:
