The Incidence and Timing of Unprovoked Venous Thromboembolism Prior to the Diagnosis of Common Cancers.

Background. It is not clear how frequently unprovoked venous thromboembolism (VTE) reflects the presence of an underlying occult malignancy.

Objective: To compare the expected and observed incidence of unprovoked VTE during a one year period immediately preceding the diagnosis of common types of cancer

Design: Retrospective population-based cohort study using the California Cancer Registry linked with the California Patient Discharge Data Set.

Setting: All non-federal hospitals in California.

Patients: All cases over the age of 17 that were diagnosed with one of 19 common cancers between Jan. 1993 and Dec. 1995.

Measurement. VTE was identified and classified as provoked (after surgery, pregnancy, trauma, medical hospitalization etc.) or unprovoked using an array of validated specific ICD-9-CM codes. VTE cases first diagnosed before the cancer diagnosis of cancer were labeled as having preceding VTE, and VTE cases diagnosed at the time of the cancer diagnosis were labeled as having concurrent VTE cases. Cancer cases with common malignant histologies were identified using the registry. The standardized incidence ratio (SIR) of unprovoked VTE was determined by calculating the expected incidence in the cancer cohorts using the 1995 and 1996 age, race and sex specific incidence rates of unprovoked VTE among California residents.

Results. There were 250585 cancer cases; mean age was 66 years, 52% were women, 35% had metastatic or unknown stage cancer at the time of diagnosis; A total of 385 cancer cases had VTE in the preceding year, and 221 of these cases (0.09%) had unprovoked VTE, compared to 216 expected cases, SIR= 1.02 (95% CI= 0.9–1.2, p=NS). 50% of the cases with preceding unprovoked VTE had metastatic or unknown stage cancer (p<0.001). Only cases with lymphoma, renal cell cancer and acute myelogenous leukemia had an SIR significantly greater than 1.0 (SIR range = 2.1–5.3). Among the cases with preceding unprovoked VTE that were later diagnosed with metastatic or unknown stage cancer, there were 30 more VTE cases than expected, SIR = 1.4 (CI = 1.2–1.6), and a significantly greater proportion of these cases were diagnosed with VTE within 91 days of being diagnosed with cancer (p =0.001). 832 (0.33%) cases were diagnosed with VTE during the hospital stay when cancer was diagnosed, and 168 (20%) of the cases were admitted with a principal diagnosis of VTE; 68% of these 832 cases with concurrent VTE had metastatic or unknown stage cancer at the time of diagnosis (p<0.001 vrs entire cohort)

Limitations. This was a retrospective study that relied on administrative data to identify cases and outcomes. The extent of any evaluation for cancer that might have occurred among the cases with preceding unprovoked VTE could not be determined.

Conclusions: In this large population based cohort of cancer cases, the incidence of unprovoked VTE in the year prior to the diagnosis of cancer was not significantly higher than expected in the general population. Approximately 10% to 15% of the cases diagnosed with preceding unprovoked VTE had features suggesting a causal link between thromboembolism and the undiagnosed cancer. Many patients were diagnosed concurrently with cancer and VTE during the same hospitalization.