Dashed line indicates separation of ultrasonography study selection (above) from patient selection (below). DVT indicates deep vein thrombosis; PE, pulmonary embolism; and VTE, venous thromboembolism.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Utter GH, Dhillon TS, Salcedo ES, et al. Therapeutic Anticoagulation for Isolated Calf Deep Vein Thrombosis. JAMA Surg. 2016;151(9):e161770. doi:10.1001/jamasurg.2016.1770
Deep vein thrombosis (DVT) isolated to the calf veins (distal to the popliteal vein) is frequently detected with duplex ultrasonography and may result in proximal thrombosis or pulmonary embolism (PE).
To evaluate whether therapeutic anticoagulation is associated with a decreased risk for proximal DVT or PE after diagnosis of an isolated calf DVT.
Design, Setting, and Participants
All adult patients with ultrasonographic detection of an isolated calf DVT from January 1, 2010, to December 31, 2013, at the Vascular Laboratory of the University of California, Davis, Medical Center were included. Patients already receiving therapeutic anticoagulation and those with a chronic calf DVT, a contraindication to anticoagulation, prior venous thromboembolism within 180 days, or diagnosis of a PE suspected at the time of calf DVT diagnosis were excluded. Data were analyzed from August 18, 2015, to February 14, 2016.
Intention to administer therapeutic anticoagulation.
Main Outcomes and Measures
Proximal DVT or PE within 180 days of the diagnosis of the isolated calf DVT.
From 14 056 lower-extremity venous duplex studies, we identified 697 patients with an isolated calf DVT and excluded 313 of these. The remaining 384 patients were available for analysis (222 men [57.8%]; 162 women [42.2%]; mean [SD] age, 60  years). The calf DVT involved an axial vein (anterior tibial, posterior tibial, or peroneal) in 243 patients (63.2%) and a muscular branch (soleus or gastrocnemius) in 215 (56.0%). Physicians attempted to administer therapeutic anticoagulation in 243 patients (63.3%), leaving 141 control participants. Proximal DVT occurred in 7 controls (5.0%) and 4 anticoagulation recipients (1.6%); PE, in 6 controls (4.3%) and 4 anticoagulation recipients (1.6%). Therapeutic anticoagulation was associated with a decreased risk for proximal DVT or PE at 180 days (odds ratio [OR], 0.34; 95% CI, 0.14-0.83) but an increased risk for bleeding (OR, 4.35; 95% CI, 1.27-14.9), findings that persisted after adjustment for confounding factors (ORs, 0.33 [95% CI, 0.12-0.87] and 4.87 [95% CI, 1.37-17.3], respectively) and sensitivity analyses.
Conclusions and Relevance
Rates of proximal DVT or PE are low after isolated calf DVT. Therapeutic anticoagulation is associated with a reduction of these outcomes but an increase in bleeding.
When specifically assessed during lower-extremity venous duplex ultrasonographic studies, thrombosis of the deep veins distal to the popliteal vein, or calf deep vein thrombosis (DVT), accounts for approximately half of all DVTs.1 An estimated 300 000 calf DVTs are detected each year in the United States,2 with more than 100 000 hospitalizations.3 Calf DVT has long been thought to represent an initial step in the development of proximal DVT.4
Despite their common occurrence—and contrasting with the plentitude of information regarding proximal DVTs—optimal management of acute isolated calf DVT is relatively poorly understood. Approaches include no treatment or monitoring (particularly if the calf DVT is asymptomatic and provoked and the risk factors are transient), follow-up with serial imaging, and therapeutic anticoagulation of varied duration. Whereas advocates of serial imaging argue that the risk for morbidity is predominantly limited to cases involving local extension within 2 weeks,2 proponents of anticoagulation cite evidence that the rate of PE is nontrivial and might be reduced with anticoagulation.5-8
Two recent systematic reviews9,10 have identified few randomized clinical trials on this topic and scant trials in the era of duplex sonography. Both reviews concluded that, although anticoagulation appeared to reduce the risk for thrombus propagation, the methodologic quality of available studies was low. The 9th edition of the American College of Chest Physicians guidelines for antithrombotic therapy for venous thromboembolic disease11 suggests that therapeutic anticoagulation for severe symptoms or risk factors for extension, with serial imaging otherwise, are weak recommendations with low-quality evidence.
A randomized clinical trial12 evaluating the efficacy of a low-molecular-weight heparin for treatment of distal DVT is in progress. Although that study and potentially others promise eventual higher-quality information, we sought to evaluate whether therapeutic anticoagulation at our center might be associated with reduced likelihood of proximal DVT or PE.
Question Is therapeutic anticoagulation for isolated calf deep vein thrombosis (DVT) associated with reduced proximal DVT or pulmonary embolism (PE)?
Findings In this retrospective cohort study of 384 patients, intention to administer therapeutic anticoagulation was associated with a significant decrease in proximal DVT or PE within 180 days from 9% to 3%.
Meaning Therapeutic anticoagulation may be warranted for isolated calf DVT.
We conducted a single-center retrospective cohort study at the University of California, Davis, Medical Center in Sacramento. The institutional review board of the University of California, Davis, approved the study and waived the requirement for informed consent because of the minimal risk to participants.
We evaluated all lower-extremity venous duplex ultrasonography studies performed by our center’s Vascular Laboratory from January 1, 2010, through December 31, 2013, for patients 18 years or older. The Laboratory is accredited by the Intersocietal Commission for the Accreditation of Vascular Laboratories, and laboratory staff routinely evaluates calf and proximal deep veins during all venous duplex studies. We electronically screened study reports and excluded those containing standardized text indicating a negative study finding. We then individually reviewed the full text of all remaining reports to select those identifying an isolated calf DVT. We defined isolated calf DVT as a DVT involving 1 or more of the deep veins distal to the popliteal vein—whether axial (posterior tibial, peroneal, or tibioperoneal trunk) or muscular (gastrocnemius or soleal)—in the absence of a proximal DVT (common femoral, superficial femoral, deep femoral, or popliteal). Laboratory staff does not evaluate the anterior tibial veins because they are difficult to compress and partially obscured by the adjacent interosseous membrane and bones. We restricted consideration to the first isolated calf DVT per patient during the study period. We excluded patients if the report described the calf DVT as chronic; if a DVT or a PE was diagnosed within 180 days before the calf DVT diagnosis (which might interfere with ascertainment of subsequent DVT or PE, the main outcome); if a radiographically confirmed PE was diagnosed based on clinical suspicion at the time of calf DVT diagnosis (which likely would represent a coexisting condition13); if an enduring contraindication (ie, >5 days) to anticoagulation existed; or if the patient was already receiving therapeutic anticoagulation at the time of calf DVT diagnosis.
One investigator (T.S.D.) used a standardized abstraction instrument to collect information from the patients’ electronic medical records. Data included baseline characteristics (age, sex, height, weight, and care setting at time of diagnosis); risk factors for venous thromboembolism (recent operation and the anatomic region involved, recent traumatic injury, obesity, current smoking, estrogen use, pregnancy, existing cancer, ambulatory status, history of DVT or PE, personal or family history of hypercoagulability, or the presence of an inferior vena cava filter); circumstances prompting the index duplex study (symptoms, signs, and laboratory or radiographic findings); anticoagulation use before and after the diagnosis of the isolated calf DVT; subsequent testing for DVT or PE; and occurrence of DVT, PE, bleeding, and death within 180 days after diagnosis of the isolated calf DVT.
We defined the exposure of therapeutic anticoagulation as the intention to treat the isolated calf DVT—as best we could determine from the record—with therapeutic doses of unfractionated heparin, a low-molecular-weight heparin (or analogous factor Xa inhibitor), warfarin, or a direct thrombin inhibitor. We excluded from consideration any anticoagulation occurring after subsequent diagnosis of a proximal DVT or PE. We considered anticoagulants administered at standard therapeutic doses to represent therapy even if the stated purpose was prophylaxis. We did not consider acetylsalicylic acid or other antiplatelet medications to be therapy. Although we collected information on failing or ceasing to administer anticoagulants or failure to achieve therapeutic effect, we primarily focused on the intention to treat the calf DVT and considered the actual therapeutic effect in a sensitivity analysis.
We defined the primary outcome as radiographically confirmed proximal DVT or PE occurring within 180 days. We did not consider propagation within the deep calf veins to qualify as a primary outcome. Although participants in our study did not routinely undergo radiographic follow-up, we collected information on the last date of any clinical follow-up at our center and the use of venous duplex ultrasonography, chest computed tomographic angiography, ventilation-perfusion studies, and pulmonary arteriogram studies. We considered bleeding episodes (defined as clinically significant blood loss, ie, affecting evaluation or treatment of the patient, independent of transfusion), death, and a composite of proximal DVT, PE, or death as secondary outcomes. To assess the reliability of exposure and outcome classifications, we had a second blinded rater (E.S.S.) independently ascertain both for 10% of the patients selected at random.
Data were analyzed from August 18, 2015, to February 14, 2016. Assuming that physicians would intend to administer anticoagulation to 33% of patients with an isolated calf DVT and that the risk for proximal DVT or PE without anticoagulation was 15%,10 we planned to be able to detect a reduction in the absolute risk for proximal DVT or PE with therapeutic anticoagulation to no more than 5% and with 80% power at the α level of .05 if we included 111 patients receiving anticoagulation and 222 control patients for evaluation.
We compared baseline characteristics between participants intended to receive therapeutic anticoagulation vs those who were not with the χ2 or Fisher exact test for categorical variables and the 2-tailed t test for continuous variables. We used logistic regression to estimate the association between the intention to administer therapeutic anticoagulation and subsequent proximal DVT or PE within 180 days. We considered age, sex, body mass index, the aforementioned risk factors for DVT, care setting at the time of calf DVT diagnosis, presence of symptoms or signs prompting calf DVT diagnosis, type of calf DVT, and use of prophylactic anticoagulation during the 7 days before calf DVT diagnosis as factors that potentially confounded this association. In addition to age and sex, we included confounding factors in the logistic regression model if the odds ratio (OR) for the outcome changed by at least 10% with inclusion of the covariate.14
We conducted a set of sensitivity analyses to evaluate the robustness of the primary results (1) considering PEs diagnosed based on clinical suspicion at the time of calf DVT diagnosis as outcomes rather than exclusions; (2) considering patients who were administered prophylactic anticoagulation after the calf DVT diagnosis as a separate exposure category (as opposed to being grouped with patients who did not receive any anticoagulation); (3) defining exposure as actual receipt of therapeutic anticoagulation; (4) defining exposure as achievement of adequate anticoagulation (as opposed to suboptimal or no anticoagulation despite an intention to administer anticoagulation); (5) defining exposure restricted to receipt of therapeutic anticoagulation within 1 day of diagnosis of the isolated calf DVT; (6) defining exposure as intention to administer therapeutic anticoagulation solely for the isolated calf DVT (with exclusion of patients with another reason for anticoagulation); and (7) restricted to patients who underwent follow-up radiographic evaluation for DVT or PE. We performed all analyses with STATA software (version 10; StataCorp LP) using 2-tailed tests and an α level of .05.
We evaluated reports of 14 056 lower-extremity venous duplex ultrasonography studies from 2010 through 2013 (Figure), which yielded 973 patients with an isolated calf DVT. These 973 studies involved 697 unique patients (276 represented repeat studies). Of these 697 patients with a newly identified isolated calf DVT, 313 met exclusion criteria, leaving 384 patients for analysis. Interrater reliability of exposure and outcome status was high (κ = 0.94 and κ = 1.00, respectively).
The 384 patients included 243 for whom physicians intended to administer therapeutic anticoagulation (anticoagulation group) and 141 for whom they did not (control group). Patients in the 2 groups differed in some baseline characteristics (Table 1 and Table 2).
Of those intended to receive therapeutic anticoagulation, 1 patient declined treatment. Of the remaining 242 patients, 210 (86.8%) began anticoagulation within 1 day of calf DVT diagnosis and 232 (96.0%) within 7 days. Delays resulted from obtaining consultations, brief contraindications (eg, pending invasive procedure), discussions with the patient, and coordination of care for outpatients. Two patients whom physicians intended not to treat with anticoagulation subsequently received therapeutic anticoagulation in the absence of a proximal DVT or PE (1 for atrial fibrillation and 1 for an upper-extremity DVT).
Therapeutic anticoagulation mainly consisted of warfarin for 182 patients (75.2%), a low-molecular-weight heparin (enoxaparin sodium or dalteparin sodium) for 43 (17.7%), continuous heparin infusion for 15 (6.2%), and rivaroxaban and bivalirudin for 1 (0.4%) each. The only departure from standard therapeutic doses involved 1 patient who received warfarin, for whom the target international normalized ratio was 1.8 to 2.3 rather than 2.0 to 3.0.
Among 132 patients receiving therapeutic anticoagulation for whom the planned duration of treatment was described, physicians planned to administer anticoagulation to 14 (10.6%) for less than 3 months, to 55 (41.7%) for 3 months, to 32 (24.2%) for 6 months, to 3 (2.3%) for 1 year, to 26 (20.0%) indefinitely, and to 2 (1.5%) based on clinical end points. Among 156 patients not intended to receive indefinite anticoagulation and for whom the actual duration of anticoagulation was determined, the median duration was 106 (interquartile range [IQR], 44.5-184) days. For 86 patients, we were able to determine the planned and actual duration of anticoagulation; the actual duration was a median of 1.5 (IQR, −7 to 29) days longer than originally planned.
Among 221 of the 242 patients in the anticoagulation group for whom records were sufficient to characterize the adequacy of therapy, 200 (90.5%) received adequate anticoagulation for the intended duration of treatment and 21 (9.5%) did not. Ten of these patients were nonadherent to the regimen; 8 had an episode of bleeding; 2 had a risk for bleeding; and 1 had difficulty titrating the dose of warfarin.
Eighty-one controls (57.4%) received prophylactic anticoagulation after calf DVT diagnosis, including subcutaneous fixed-dose heparin injections in 50 patients; a continuous prophylactic heparin infusion in 6; a low-molecular-weight heparin (enoxaparin or dalteparin) in 22; low-dose warfarin (to maintain the international normalized ratio at 1.5-2.0) in 2; and fondaparinux sodium in 1. The median duration of prophylaxis was 9 (IQR, 4-20) days.
Eighty-three patients in the control group (59.0%) and 172 patients in the anticoagulation group (71.0%) last had interaction with health care professionals at our center at least 180 days after calf DVT diagnosis (P = .02). In the control and anticoagulation groups, 75 patients (53.2%) and 95 patients (39.3%), respectively, underwent a duplex ultrasonography study within 180 days after calf DVT diagnosis (P = .02), with 29 (20.6%) and 26 (10.7%), respectively, undergoing more than 1 study. Among those tested, the median time to the first follow-up duplex study was 8 (IQR, 6-12.5) days in the control group and 38 (IQR, 11-95) days in the anticoagulation group; the median times to the last follow-up duplex study were 16 (IQR, 8-37.5) and 73 (IQR, 23-114) days, respectively.
Among the control and anticoagulation groups, 28 patients (19.9%) and 46 patients (19.0%), respectively, underwent a study for PE within 180 days (P = .82), with 5 (3.5%) and 16 (6.6%), respectively, undergoing more than 1 study. All underwent chest computed tomographic scan with contrast except for 5 anticoagulation recipients who underwent a ventilation-perfusion scan. Among those tested, the median times to the first PE study were 22 (IQR, 5-48) days in the control group and 7.5 (IQR, 1-26) days in the anticoagulation group; the median times to the last PE study were 38.5 (IQR, 8-55) and 16 (IQR, 2-65) days, respectively.
Proximal DVT occurred in 7 control group patients (5.0%) and 4 anticoagulation group patients (1.6%; unadjusted relative risk [RR], 0.33; 95% CI, 010-1.11) (Table 3). The median time to proximal DVT diagnosis was 13 (IQR, 7-29) days. Pulmonary embolism occurred in 6 patients in the control group and 4 patients in the anticoagulation group (unadjusted RR, 0.39; 95% CI, 0.11-1.35). The median time to PE diagnosis was 15.5 (IQR, 6-40) days.
Proximal DVT or PE occurred in 13 control group patients (9.2%) and 8 anticoagulation group patients (3.3%). Intention to administer therapeutic anticoagulation was associated with a lower likelihood of proximal DVT or PE, with an RR of 0.36 (95% CI, 0.15-0.84). After adjustment for age, sex, care setting at the time of calf DVT diagnosis, existing cancer, and a history of DVT or PE, this association persisted (Table 3). With adjustment, intention to administer therapeutic anticoagulation was associated with a reduced likelihood of a composite of proximal DVT, PE, or death.
Among patients with a provoked calf DVT, the adjusted OR of proximal DVT or PE with anticoagulation was 0.35 (95% CI, 0.13-0.93); among those with single and multiple calf DVTs, the adjusted ORs were 0.27 (95% CI, 0.07-0.96) and 0.35 (95% CI, 0.06-1.95), respectively. (The association among patients with unprovoked calf DVT could not be estimated because only 1 outcome occurred.) Although therapeutic anticoagulation was not significantly associated with proximal DVT or PE among patients with an axial vein calf DVT (adjusted OR, 0.52; 95% CI, 0.14-1.90), the association was found among patients with a calf DVT of a muscular branch (adjusted OR, 0.12; 95% CI, 0.03-0.53).
Clinically significant bleeding occurred more frequently with anticoagulation (Table 3). Among 24 instances, 10 involved the gastrointestinal tract; 5, soft tissue; 4, the genitourinary tract; 4, the central nervous system; and 4, a procedure site (locations were not mutually exclusive). Among 48 patients who died within 180 days, 28 (58.3%) had cancer at the time of calf DVT diagnosis. One patient in the anticoagulation group died of a hemorrhagic stroke, and 3 patients in the anticoagulation group died possibly because PE contributed to hypoxia or circulatory insufficiency.
Inclusion of records involving a PE diagnosis based on clinical suspicion at the time of calf DVT diagnosis (rather than exclusion of such cases) only modestly affected the observed association between the intention to administer therapeutic anticoagulation and the primary outcome (Table 4). Among 18 such records, 10 involved no apparent intention to administer therapeutic anticoagulation for the calf DVT (before the diagnosis of the PE) and 8 involved an intention to administer anticoagulation for the calf DVT. In 16 instances, the PE was diagnosed within 2 days after the calf DVT diagnosis, and all 18 diagnoses were made within 6 days.
Reclassification of the exposure as actual receipt of therapeutic anticoagulation (as opposed to intention to administer anticoagulation) or as anticoagulation occurring within 1 day of diagnosis of the calf DVT did not substantially change the observed association (Table 4). Restricting the analysis to patients who underwent follow-up testing for DVT or PE also did not appreciably change the observed association. The additional sensitivity analyses similarly did not reveal large differences compared with the primary analysis.
Consensus regarding the optimal management of isolated calf DVTs has yet to coalesce, but our study may add weight to arguments for routine inclusion of the calf deep veins in ultrasonography studies and therapeutic anticoagulation for isolated calf DVTs. Similar to a previous systematic review,9 we observed a two-thirds reduction in the risk for proximal DVT or PE. The anticoagulation and control groups in our study differed in some characteristics, and specific assumptions of our analysis may have influenced the findings. Nonetheless, neither adjustment for confounders nor sensitivity analyses identified any major concerns about the robustness of the primary findings.
Although the main results were consistent with those of prior studies, our findings also differed in certain respects. Whereas some investigators believe that anticoagulation benefits mostly patients with unprovoked and/or multiple calf DVTs,15 we found anticoagulation most strongly associated with reduced thromboembolism in patients with provoked and single calf DVTs. Indeed, because we observed so few outcomes among patients with unprovoked calf DVTs, our study may have limited external validity in that population.
Prior studies have been conflicted regarding how commonly proximal DVT or PE occurs after a muscular branch calf DVT.16,17 Consistent with the findings of Lautz et al,16 we observed a relatively strong association between anticoagulation and reduced thromboembolism among patients with a calf DVT of a muscular branch. These differences are germane because the American College of Chest Physicians guidelines for treating calf DVTs did not consider provoked, multiple, or muscular branch calf DVTs as risk factors for extension.11 Other investigators18 have concluded that the risk for proximal DVT after calf DVT is reduced 10-fold with prophylactic anticoagulation, but we did not observe such a strong association in the second portion of the sensitivity analysis.
The decision to administer therapeutic anticoagulation to patients with isolated calf DVTs should weigh the risks and benefits for a given patient. Bleeding episodes, although infrequently life threatening, were clearly associated with therapeutic anticoagulation in our study, and in 1 instance, an unforeseen hemorrhagic stroke directly caused death.
The limitations of our study largely pertain to its retrospective, observational nature. Exposure status could not always be clearly ascertained. Occasionally, intention to administer anticoagulation for the calf DVT could not be unambiguously differentiated from concomitant indications (eg, atrial fibrillation). Of potentially greatest concern, testing for proximal DVT and PE was not applied to all patients, and the indications for and timing of testing were not standardized. Because our findings hinged on relatively few outcome events, increased testing for DVT and PE in the control group could have biased the results. Nonetheless, restriction of the analysis to patients who underwent follow-up testing—which presumably would diminish any association attributable to disparity in the use of follow-up testing—did not reveal different results. Residual confounding from unmeasured or unmeasureable factors may have affected our findings. Our center’s Vascular Laboratory does not assess the anterior tibial vein or record the length, diameter, or detailed location (eg, proximal vs distal calf) of calf DVTs, so we could not evaluate these characteristics.
Although radiographic end points have been widely used in studies involving DVT and PE, the extent to which such surrogate outcomes align with true patient-centered outcomes is not well understood. We did not evaluate 1 such possible patient-centered outcome, postthrombotic syndrome, because we did not think it could be ascertained well from medical records.
Outcomes detected at other hospitals may not have been captured in our center’s records. Because California maintains a statewide data repository with linked hospitalization and emergency department records, a future study may be able to capture such outcomes without having to contact patients and rely on their recall.
Considering these limitations, we conclude that therapeutic anticoagulation of patients with isolated calf DVTs may be warranted to reduce the risk for proximal venous thromboembolism. However, randomized studies are needed to draw firmer conclusions. Because the benefits of anticoagulation seem modest, we recommend attention to the risk for bleeding when determining whether anticoagulation is appropriate.
Corresponding Author: Garth H. Utter, MD, MSc, Department of Surgery, University of California, Davis, Medical Center, 2315 Stockton Blvd, Main Hospital, Room 4206, Sacramento, CA 95817 (firstname.lastname@example.org).
Accepted for Publication: April 17, 2016.
Published Online: July 20, 2016. doi:10.1001/jamasurg.2016.1770.
Author Contributions: Drs Utter and Dhillon had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Utter, Salcedo, Reynolds, Humphries.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Utter, Humphries.
Critical revision of the manuscript for important intellectual content: Dhillon, Salcedo, Shouldice, Reynolds, Humphries, White.
Statistical analysis: Utter, White.
Administrative, technical, or material support: Dhillon, Reynolds.
Study supervision: Utter, Dhillon, Salcedo, Reynolds, Humphries, White.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This paper was presented at the 87th Annual Meeting of the Pacific Coast Surgical Association, February 14, 2016; Kohala Coast, Hawaii.
Create a personal account or sign in to: