Spontaneous superficial venous thrombophlebitis: Does it increase risk for thromboembolism?

Methods

In the Dutch health care system, all citizens are enlisted with a family physician of their choice. This family physician functions as a gate-keeper to secondary care, evaluating all health related complaints and answering questions before referral is considered.

We carried out a historic 6-month follow-up study. The exposed cohort consisted of patients who had presented to their family physician with an uncomplicated and spontaneous SVTP. Nonexposed patients had no history of SVTP, but were matched for practice, age (within 1 year), and sex. For exposed and unexposed patients alike, the occurrence of all venous and arterial thromboembolic events were tallied over a follow-up period of 6 months using the electronic patient records of 5 health care centers. All centers participate in the registration network of the Academic Medical Centre–University of Amsterdam in Amsterdam, the Netherlands.

The database

Family physicians electronically register all their enlisted patients and their medical problems. Medical problems are coded using the International Classification of Primary Care (ICPC). Furthermore, at each health care center all physician-patient contacts (including reason for encounter, diagnosis, and treatment/prescription) are recorded in an electronic medical database and available for further analysis.

The anonymous data are stored centrally at the Department of General Practice. The system has been in use since 1995 and gradually extended to include more health care centers in the hospital’s catchment area. For this study, the data of all 40,013 enlisted patients of 34 family physicians in the period 1995 to 2002 were used.

Identifying exposure

In the central electronic medical database, possible consultations concerning SVTP were identified using truncated keywords. After testing in another time period in the same database, successful truncated keywords were t(h)romb*, phleb* and fleb*. Afterwards, full consultations were assessed to identify patients with SVTP from the initial potential cases. SVTP was diagnosed clinically according to the rules of the ICPC: “signs of inflammation along a superficial vein.”¹¹

As SVTP could be misdiagnosed, we scrutinized all consultations occurring within a month after the event for misclassified cases and excluded them. For each patient with thrombophlebitis, 2 nonexposed patients matched for center, age, and sex were randomly selected from the entire (nonexposed) enlisted population. The date of diagnosis of the thrombophlebitis was taken as index date.

The medical records of patients and controls were prospectively searched both electronically and manually by the same investigator for new thromboembolic events (DVT, PE, stroke, and myocardial infarction [MI]) during the 6 months following the index date. As all patients with any of these events were hospitalized, none of these diagnoses was made without adequate diagnostic procedures.

DVT was diagnosed when noncompressibility of the vein in the groin or fossa poplitea was observed with serial compression ultrasound (on day 1 or day 7). Problem lists contained information about the previous medical history.

Statistics

We calculated the 6-month cumulative incidences of thromboembolic events and their 95% confidence intervals (CIs). Logistic regression was used to quantify the associations between SVTP and the 6-month cumulative incidences of DVT, PE, stroke, and MI, respectively. Odds ratios are adjusted for age and sex (matched design) and for a history of the particular outcome event at issue. Clustering by health center was accounted for by using robust variance estimation commands in Stata (version 8.2).

Results

Initially, 206 patients with thrombophlebitis of the leg were identified. Twenty patients were excluded as the thrombophlebitis was induced by an intravenous catheter. One 97-year-old male patient was excluded as 2 matched controls could not be found.

Thus, for this analysis, data were available for 185 patients with a spontaneous episode of thrombophlebitis of the lower limb and 370 age and sex-matched nonexposed patients. The mean age of the patients was 63.2 years (range, 20–96; standard deviation, 15.7), and 77% were female.

TABLES 1 AND 2 show the distribution of the types of venous and arterial sequelae during the 6 months follow-up period in the 2 cohorts as well as the odds ratios (OR). DVT occurred in 2.7% of the SVTP patients as compared to 0.2% in the controls (OR=10.2; 95% CI, 2.0–51.6). Correction for DVT in the past decreased the OR to 7.1. ORs for PE (1.0), stroke (0.7), and MI (2.0) were smaller and had wider confidence intervals.

As expected, few patients received specific treatment (TABLE 3). If treatment was given, NSAIDs were most commonly prescribed. Occasionally compression stockings were advised.

TABLE 1
6-mo incidence of thromboembolic events in patients with and without SVTP