Thrombosis Research 134 (2014) 846–850
Contents lists available at ScienceDirect
Thrombosis Research journal homepage: www.elsevier.com/locate/thromres
Regular Article
Overuse of Compression Ultrasound for Patients with Lower Extremity Cellulitis☆ Craig G. Gunderson ⁎,1, John J. Chang 1 Yale School of Medicine, Department of Internal Medicine, Veteran’s Administration Healthcare System, 950 Campbell Ave, West Haven CT 06516
a r t i c l e
i n f o
Article history: Received 14 May 2014 Received in revised form 28 July 2014 Accepted 4 August 2014 Available online 15 August 2014 Keywords: Cellulitis Venous thrombosis Ultrasonography
a b s t r a c t Background: Compression ultrasound (CUS) is often ordered in hospitalized patients with cellulitis to assess for deep vein thrombosis (DVT). Despite this common practice, the rate of use and utility of CUS has not been well described. Methods: We conducted a retrospective cohort study of adult patients with lower extremity cellulitis hospitalized between October 1, 2008 and September 30, 2013 at an academic medical center. Cases meeting inclusion criteria were reviewed for the use of CUS, the indication for CUS, the occurrence of DVT, and the 3 month follow-up occurrence of DVT after discharge. Results: A total of 239 patients were identified using ICD-9 coding data with a discharge diagnosis of cellulitis or abscess of leg. Of these, 183 met criteria for inclusion in the study, 133 of whom had CUS to assess for DVT (73%). Of the 133 who received CUS, 11 studies found DVTs (8%). Of the 11 DVTs, 8 had been previously diagnosed, and 3 were new. Of the new DVTs, only one was ipsilateral to the leg with cellulitis. Conclusion: Most patients admitted with lower extremity cellulitis received CUS to assess for DVT. Despite this common practice, the rate of acute ipsilateral DVT was low and matched the rate of acute contralateral DVT. Previously diagnosed DVTs were commonly re-imaged. Overall the use of CUS had minimal impact on patient management and the routine use of CUS to assess for DVT in hospitalized patients with cellulitis appears to be unnecessary. Published by Elsevier Ltd.
Introduction Cellulitis is a common type of skin and soft tissue infection resulting in more than 600,000 hospitalizations per year in the United States [1]. The occurrence of deep vein thrombosis (DVT) is often considered in patients with cellulitis because of the shared presentation of unilateral limb swelling, erythema and pain [2,3], and compression ultrasound (CUS) is often ordered in hospitalized patients with cellulitis. We recently performed a systematic review of the risk of DVT in patients with cellulitis and found an overall pooled rate of proximal DVT of 2.1% (95% CI, 0.5%-9.1%) [4]. In our review we only found two studies that reported rates of CUS use in patients with cellulitis; one from Denver reported that 42% of patients admitted with cellulitis received ultrasounds, but did not individually review the cases to determine
the indication for the ultrasound or report the rate of DVT [5]. The other study from New Zealand reported that 16% of patients with cellulitis admitted to the hospital received CUS, but also did not individually review the cases to determine the indication for the study [6]. This study reported a low rate of proximal DVT (1.25%). In summary, based on our recent systematic review of the literature, no study carefully reports the frequency and indication for CUS in patients with cellulitis. The purpose of the current study is to review a cohort of patients hospitalized with cellulitis to describe the frequency of CUS as well as to record the rate of positive studies and impact on management. Our hypothesis is that the majority of patients being admitted with cellulitis has CUS to assess for DVT and that the rate of DVT and impact on management is low. Methods
Abbreviations: CUS, compression ultrasound; DVT, deep vein thrombosis; WHVA, West Haven Connecticut Veteran’s Administration Hospital; ICD-9, International Classification of Disease, 9th Revision; RVO, residual vein obstruction; ACCP, American College of Chest Physicians; IDSA, The Infectious Disease Society of America. ☆ Funding: None. ⁎ Corresponding author at: Department of Internal Medicine, West Haven Veteran's Hospital, West Haven, CT 06516. E-mail address:
[email protected] (C.G. Gunderson). 1 Tel.: +1 203 932 5711x2616; fax: +1 203 937 3425.
http://dx.doi.org/10.1016/j.thromres.2014.08.002 0049-3848/Published by Elsevier Ltd.
Study Design The West Haven Connecticut Veteran’s Administration Hospital (WHVA), located outside New Haven, Connecticut is a 228 bed tertiary care center affiliated with Yale Medical School. All patients discharged from the WHVA with an International Classification of Disease, 9th Revision (ICD-9) primary diagnosis of cellulitis or abscess of the leg
C.G. Gunderson, J.J. Chang / Thrombosis Research 134 (2014) 846–850
(682.6) between October 1, 2008 and September 30, 2013 were reviewed for possible inclusion in the study. For patients with multiple hospitalizations for cellulitis at the WHVA during the study period only the first episode was included. Patients’ records were then reviewed using a standardized data extraction form for demographics, length of stay, type and location of infection, admitting service, putative risk factors for cellulitis and deep vein thrombosis, whether previous admissions for cellulitis had occurred, prior compression ultrasounds, use of anticoagulation and indication, D-dimer, whether CUS was ordered and indication and result. Post-discharge records were reviewed for the occurrence of deep vein thrombosis at three months. All patients with documented cellulitis of the lower extremity were included. Patients were excluded if they had infection involving a different location or if they had a different type of soft tissue infection that did not also involve a spreading infection of the leg, such as diabetic foot ulcers, abscesses without cellulitis, peripheral artery disease with gangrene and necrotizing soft tissue infections. The study was approved by the WHVA Institutional Review Board. Study Definitions Leg cellulitis was defined as any spreading infection involving the leg [7]. In the manner of Jenkins et al. [5], cases of cellulitis were divided into cases with complicating factors such as abscess, peripheral artery disease or chronic ulcers and cases without such factors. Risk factors for DVT were defined by the Wells Criteria [8]; active cancer was defined as receiving treatment for cancer within the previous six months or currently receiving palliative care, excluding basal-cell and squamous-cell carcinoma of the skin; bedridden was defined as recently bedridden for 3 or more days; major surgery was defined as major surgery within the past 12 weeks requiring general or regional anesthesia. DVTs were defined by evidence of thrombosis of the popliteal or more proximal veins on compression ultrasonography. New DVTs were assumed when there was no known prior history of leg DVT. Prior DVTs were defined as cases with previously diagnosed leg DVTs as documented in notes and confirmed by prior CUS at our institution. These were further classified as recurrent DVTs when there was proximal extension of prior thrombosis into a previously fully compressible proximal segment or an increase in size of thrombus as commented by the radiology report compared with old studies. If there was a prior DVT and the thrombus had not extended proximally or increased in size this was classified as residual vein obstruction (RVO) in accordance with the American College of Chest Physicians (ACCP) guidelines on antithrombotic therapy and prevention of thrombosis [9]. In cases with prior DVT in which the original and new CUS were not directly comparable because one had been done outside our institution, the DVT was recorded as uncertain recurrent or RVO. Data Analysis The primary outcomes of the study were the frequency of patients with leg cellulitis that had CUS done to evaluate for DVT and the frequency of studies that were positive for DVT. Patients were grouped by whether they had CUS done and compared in terms of demographics and relevant clinical characteristics. For comparison of dichotomous variables, chi-square or Fisher’s exact tests were used. For continuous variables, ttests or Mann–Whitney U tests were performed. Variables with p values less than 0.2 on bivariate analysis were included in a multivariate logistic regression analysis. P values b0.05 were considered significant. All analyses were performed using SPSS version 21.0 (Chicago, Illinois). Results 239 patients were identified during the study period with a discharge diagnosis of cellulitis or abscess of the leg. In total, 56 patients were excluded; 17 cases had more than one admission for leg cellulitis during
847
the study period and the later admission excluded; 17 cases had abscesses without cellulitis; 7 had stasis dermatitis and antibiotics discontinued; 4 had prepatellar bursitis; 2 had diabetic foot infections without leg cellulitis; 2 had chronic ulcers without cellulitis; 3 had miscellaneous noncellulitic infections, including a case each of pneumonia, chronic osteomyelitis, and disseminated mycobacterium marinum. Fig. 1 diagrams the overall results of the study. Of the 183 included patients, 133 (72.7%) had CUS done to exclude DVT. Of the 133 studies, 11 (8.3%) were positive for thrombosis including 3 (2.3%) new diagnoses and 8 (6.0%) patients with prior ipsilateral leg DVTs that were re-imaged. Table 1 shows the clinical characteristics of the 11 patients with DVT. Of the new diagnoses, one patient had an ipsilateral DVT, one had a contralateral DVTs compared with the leg with cellulitis and one case had bilateral cellulitis with a unilateral DVT. Of note, the one patient found to have a new ipsilateral DVT refused anticoagulation and may in fact have had a chronic DVT as he told the physician that he had been recommended to take anticoagulation for his legs in the past and refused. We chose to include him as a new DVT however given the lack of confirmatory documentation. Of the eight patients who had prior known leg DVTs, six of the eight imaged DVTs were ipsilateral to the side with cellulitis, one was contralateral and one patient had bilateral DVTs with unilateral cellulitis. Six of the eight old DVTs were classified as residual thrombus since clot had not extended proximally or grown, one of the eight was classified as recurrent and one was uncertain due to lack of prior imaging at our institution. Five of the eight were already on anticoagulation for DVT while three were restarted although two of these patients were classified as residual vein obstruction rather than recurrent DVT. Interestingly, none of the treating teams distinguished between residual vein obstruction and recurrent DVT and all DVTs, whether old or new, prompted resumption or continuation of anticoagulation. Table 2 shows the demographic and clinical characteristics of the included cases categorized by whether CUS was done to assess for DVT. Both groups were similar in terms of age, gender, admitting service, year of admission and whether they were on anticoagulation and whether they had had previous CUS of the same limb to assess for DVT. Patients who had CUS were significantly more likely to have cellulitis without a complicating factor, chronic edema, and risk factors for DVT, including prior DVT and being bedridden. There were also nonsignificant trends for patients who had CUS to have longer lengths of stay and paresis. Patients with traumatic wounds were less likely to have CUS ordered. After multivariate adjustment, only cellulitis without complicating factor was significantly associated with having CUS performed (odds ratio 2.6, 95% confidence interval: 1.2-5.9). Table 3 shows the demographic and clinical characteristics of the patients that had CUS categorized by the result of the CUS. Compared with the group that did not have DVT, the patients with DVT were significantly more likely to have a prior history of DVT. There were also nonsignificant trends for the group with DVT to have chronic edema and leg ulcers and to already be on anticoagulation. Lastly, there was a trend for the frequency of DVT to vary by year. Multivariate analysis was not performed due to the limited number of patients with DVT. Of the 183 patients included in the study, 181 had documented follow-up 3 months or later in clinic and were without newly diagnosed DVT. Two patients were lost to follow-up post discharge. Discussion The primary finding of our study is that the majority of admissions with leg cellulitis at our facility are being screened for DVT by CUS. Why this is being done is uncertain. The Infectious Disease Society of America’s (IDSA) Guideline on Skin and Soft Tissue Infections does not even mention CUS as a test to consider in patients with cellulitis [10]. Our recent systematic review of the rate of DVT in patients with cellulitis concluded that although the quality of the literature was limited the available evidence indicated a relatively low rate of proximal DVT in patients with cellulitis, comparable to the rate of patients in the “low risk”
848
C.G. Gunderson, J.J. Chang / Thrombosis Research 134 (2014) 846–850
Fig. 1. Flow diagram of study results.
group according to the Wells criteria and comparable to rates of unselected hospitalized patients screened for DVT (Table 4). In our present study, patients were more likely to have CUS ordered if they had chronic edema, a risk factor for DVT, or cellulitis without a complicating factor. On multivariate analysis however only cellulitis without a complicating factor significantly predicted having a CUS performed. Presumably patients with cellulitis with a complicating factor such as a chronic ulcer or traumatic wound are less likely to have CUS because of the assumption that there is already a reason for the cellulitis to have originated. Other factors that might discourage a physician from ordering an ultrasound such as a negative prior CUS, already being on anticoagulation, or the rapid resolution of symptoms as defined by one day admission were not significant predictors. Our study found a relatively low rate of newly diagnosed DVT, with only 1 newly diagnosed ipsilateral DVT, which matched the rate in the
contralateral leg. Previous studies that have reported rates of DVT in cohorts with cellulitis have found variable rates, ranging from 0.5% to 12.5%, but these studies have varied by inclusion of distal DVTs, clinical setting and patient inclusion factors [4]. Only one prospective study of consecutive patients hospitalized with leg cellulitis and reporting proximal DVT has been done previously, and this found only one ipsilateral DVT among 200 patients for an overall rate of 0.5%, a similar rate to that in the contralateral leg and a similar rate of new ipsilateral DVT to that in our present study [11]. Overall given the results of this prospective study, our current study and our previous systematic review, we do not believe there is evidence for a relationship between cellulitis and acute DVT. We were surprised by the relatively high rate of patients with cellulitis that had previously diagnosed DVTs. Overall eight patients (6%) had previous leg DVTs, six of which were ipsilateral to the leg with cellulitis.
Table 1 Characteristics of 11 Patients with DVT. Laterality No Prior DVT: 1. Contralateral 2. Ipsilateral 3. Uncertain† Prior Leg DVT: 4. Ipsilateral 5. Ipsilateral 6. Ipsilateral 7. Ipsilateral 8. Bilateral 9. Contralateral 10. Ipsilateral 11. Ipsilateral
Prior CUS
Current CUS
Thrombus Description
A/C
RVO?
3/20/05 none 11/14/05
8/22/13 10/13/11 11/23/09
new occlusive, femoral and popliteal veins non occlusive, SFV and popliteal new non-occlusive, popliteal through femoral
started refused⁎ started
n/a n/a n/a
6/8/11 10/24/02 8/9/10 10/12/11 2/11/04 9/4/09 7/27/07 12/1/09
4/15/13 10/11/12 1/10/13 6/25/12 11/1/11 5/25/12 10/28/11 8/22/11
n/a non-occlusive with webbing, popliteal (unchanged) non-occlusive thrombus, same segments (diminished) non-occlusive thrombus, same segment (diminished) new right SFV and new left popliteal thrombus non-occlusive popliteal (unchanged) non-occlusive thrombus, same segments (diminished) non-occlusive thrombus, same segments (unchanged)
continued continued restarted continued restarted restarted continued continued
uncertain‡ RVO RVO RVO Recurrent§ RVO RVO RVO
Abbreviations: A/C, anti-coagulation; RVO, residual vein obstruction, N/A, not applicable; SFV, superficial femoral vein; Nl, normal. ⁎ Patient was found to have DVT ipsilateral to leg with cellulitis but refused anticoagulation. Patient told team he had been previously diagnosed with DVT at outside facility and had refused anticoagulation. Case included as new DVT because of lack of documentation of prior diagnosis. † Patient admitted with bilateral cellulitis, found to have unilateral DVT. ‡ Patient had extensive DVT diagnosed in 2011. Had repeat CUS prior to current admission but at outside facility and chart documentation insufficient to distinguish RVO from recurrent. § Patient had remote bilateral DVTs in 1999, then follow-up CUS in 2004 that showed residual thrombus on the right but not the left, then repeat bilateral CUS at admission for cellulitis with bilateral thrombosis that had progressed from 2004, therefore bilateral recurrent DVT.
C.G. Gunderson, J.J. Chang / Thrombosis Research 134 (2014) 846–850 Table 2 Demographic and Clinical Characteristics of 183 Patients Admitted with Leg Cellulitis.
Age Male Fiscal year 2009 2010 2011 2012 2013 Admitting service Medicine Surgery Length of Stay Length of Stay one day Type of SSTI Cellulitis Abscess w/ cellulitis Cellulitis w/ complication Risks for Cellulitis DM Chronic Edema Chronic Ulcer Traumatic wound Prior Cellulitis No Risk for Cellulitis Risks for DVT Prior DVT Active Cancer Paresis Bedridden Major Surgery Elevated D-Dimer On anticoagulation Sub-Therapeutic INR Previous CUS
No CUS, n = 50 (27.3%)
CUS, n = 133 (72.7%)
P Value
67.3 + −12.0 50 (100%)
68.3 + −11.8 130 (97.8)
0.62 0.56 0.18
14 (43.8) 9 (20.9) 8 (22.2) 9 (23.1) 10 (30.3)
18 (56.3) 34 (79.1) 28 (77.8) 30 (76.9) 23 (69.7) 0.22
46 (92.0) 4 (8.0) 5.1 + −5.6 6 (12.0)
129 (97.0) 4 (3.0) 6.9 + −6.6 14 (10.5)
31 (62.0) 2 (4.0) 17 (34.0)
109 (82.0) 6 (4.5) 18 (13.5)
16 (32.0) 31 (62.0) 9 (18.0) 12 (24.0) 12 (24.0) 1 (2.0) 9 (18.0) 3 (6.0) 5 (10.0) 0 (0) 0 (0) 1 (2.0) 2 (2/2) 9 (18.0) 1/9 21 (42.0)
35 (26.3) 103 (77.4) 16 (12.0) 12 (9.0) 33 (24.8) 8 (6.0) 48 (36.1) 27 (20.3) 11 (8.3) 9 (6.8) 10 (7.5) 2 (1.5) 10 (10/11) 33 (24.8) 9/33 55 (41.3)
0.05 0.78 0.007
0.45 0.04 0.30 0.007 0.91 0.45 0.02 0.01 0.45 0.05 0.04 0.62 0.85 0.33 0.30 0.94
Abbreviations: CUS, compression ultrasound; SSTI, skin and soft-tissue infection; INR, international normalized ratio.
Six of these met criteria for residual vein thrombosis. According to the ACCP, approximately 80% of patients with proximal DVT have residual thrombus at 3mos and 50% still have thrombosis at one year. The recommended approach to evaluating recurrent thrombosis is with a combination of D-dimer and CUS or serial CUS done one week apart. In patients with stable thrombus size which has not extended proximally, the ACCP does not recommend further anticoagulation. Of the eight patients with previous DVT in our study, only one appeared to have recurrent DVT which would appropriately have merited anticoagulation. The rest had RVO and most were already on anticoagulation. Interestingly, an association between old DVTs and cellulitis has not been previously reported to our knowledge but makes intuitive sense given the reported association between chronic edema and cellulitis [12,13]. The only significant predictor of a positive CUS for DVT was a prior history of DVT. Additionally there were non-significant trends for the group with DVT to have chronic edema and leg ulcers and to be already on anticoagulation. There was also a trend for the frequency of DVT to vary by year, with no DVTs in FY2009 and relatively more DVTs later in the study period. We can think of no a priori reason that the DVT rate would vary by year other than the fact that in FY2009 there were fewer CUS done to assess for DVT. The result may also just reflect a chance association. What does our study mean for ordering CUS in patients hospitalized with cellulitis? Overall our results when combined with previously published studies do not show a relationship between acute ipsilateral DVT and cellulitis and we do not recommend screening for acute DVT in patients with cellulitis. The one possible exception is for patients with previously diagnosed ipsilateral DVTs. In this group, one of our patients had recurrent DVT and it may therefore be reasonable to assess for DVT in this subgroup if the result would change management, such as by
849
Table 3 Demographic and Clinical Characteristics of 133 Patients with Leg Cellulitis that had CUS.
Age Male sex Fiscal year 2009 2010 2011 2012 2013 Admitting service General Medicine Surgical Length of Stay Length of Stay one day Type of SSTI Cellulitis Abscess w/ cellulitis Complicated Cellulitis Risks for Cellulitis Diabetes Chronic Edema Chronic Ulcer Traumatic wound Prior Cellulitis No Risk for Cellulitis Risks for DVT Prior DVT Active Cancer Paresis Bedridden Major Surgery Elevated D-Dimer On anticoagulation Sub-Therapeutic INR Previous CUS
No DVT, n = 122
DVT, n = 11
68.1 + −11.9 119 (97.5)
70.7 + −11.0 11 (100)
P Value
18 33 27 25 19
0 1 1 5 4
118 (96.7) 4 (3.3) 7.0 + −6.8 12 (9.8)
11 (100) 0 (0) 5.2 + −3.8 2 (18.1)
98 (80.3) 6 (4.9) 18 (14.8)
11 (100) 0 0
33 (27.0) 92 (75.4) 13(10.7) 11 (9.0) 31 (25.4) 8 (6.6)
2 (18.1) 11 (100) 3 (27.3) 1 (9.1) 2 (18.2) 0 (0)
0.73 0.07 0.13 1.0 0.73 1.0
19 (15.6) 11 (9.0) 7 (5.7) 10 (8.2) 2 (1.6) 9/10 28 (23.0) 8/26 51 (41.8)
8 (72.7) 0 (0) 2 (18.2) 0 (0) 0 (0) 1/1 5 (45.4) 1/5 4 (36.3)
b0.001 0.60 0.16 0.41 0.81 0.91 0.14 0.54 1.0
0.48 1.0 0.06
1.0
0.47 0.33 0.27
Abbreviations: CUS, compression ultrasound; SSTI, skin and soft-tissue infection; INR, international normalized ratio.
prompting the resumption anticoagulation. The overall effect of CUS on patient management in our study was minimal. As can be seen in Table 1, of the eleven DVTs found by screening patients with leg cellulitis, only one new ipsilateral DVT was diagnosed (and this patient refused anticoagulation, as he had in the past), eight old DVTs were reimaged which resulted in one patient being appropriately anticoagulated for a recurrent DVT, two patients being inappropriately anticoagulated for residual thrombosis, and five patients with residual thrombosis who were already on anticoagulation which was simply continued. Over the past decade numerous authors and professional groups have deplored the use of medical services that offer minimal benefits. Such practices have been variably labeled overuse [14], unnecessary care, and waste [15]. Medical groups such as the American College of Physicians [16], the American Board of Internal Medicine [17] and the National Institute of Health and Clinical Excellence in the United Kingdom [18] have developed and promoted lists of medical practices that appear to offer minimal benefit or harm exceeding benefit. A recent
Table 4 Benchmark Rates of Proximal DVT in select populations. Overall average for patients referred for CUS [20] ICU patients screened for DVT [21] Hospitalized with acute stroke, screened [22] Low risk patients per Wells’ criteria [20] Patients with cellulitis or erysipelas [4] Hospitalized patients without suspected DVT, screened [23] Low risk per Wells’ criteria and normal D-dimer [24]
24% 16% 11.4% 6.5% 2.1% 1.8% 0.7%
(3.2%-11.4%) (0.5%-9.1%)
(0.3%-1.3%)
Abbreviations: DVT, deep vein thrombosis; CUS, compression ultrasound; ICU, intensive care unit.
850
C.G. Gunderson, J.J. Chang / Thrombosis Research 134 (2014) 846–850
systematic review of overuse of health care services in the United States concluded that there was a relative paucity of literature on the subject and called for increased research in the area [19]. Based on the current study, we believe that the routine use of CUS to assess for DVT in patients hospitalized with cellulitis is an example of overuse. Although there is seemingly little harm in compression ultrasound (other than patient discomfort with compression of a cellulitic leg perhaps), it is notable that medical teams consistently did not distinguish between residual vein obstruction and recurrent DVT. All cases were simply labeled as DVT and restarted or continued on anticoagulation. A typical example is patient number 6 in Table 1, who had a DVT three years previously and completed a year of anticoagulation. At the time of his original CUS he had occlusive thrombus involving the popliteal and superficial femoral veins. When admitted for cellulitis he had a repeat CUS which now showed non-occlusive thrombus in the same venous segments without proximal extension, consistent with RVO. The team nevertheless restarted anticoagulation despite ACCP recommendations against treating RVO with anticoagulation. Although the proper duration of anticoagulation in deep vein thrombosis is uncertain, the incorrect diagnosis of recurrent DVT unnecessarily exposes the patient to prolonged and possibly lifelong anticoagulation with its attendant risks and inconvenience. Our study has several important limitations. First, given its retrospective design it is possible that we missed cases of DVT in patients with cellulitis if they were coded as DVT instead of cellulitis. The fact that we found such a high rate of DVT, albeit chronic DVT, indicates that this bias may not have been present. Our study was also done at only one institution, and there is likely wide variation in the use of CUS in different areas of the country and different health care systems. In summary, our data does not support a relationship between hospitalization for cellulitis and acute DVT. Chronic DVTs on the other hand appear to be a relatively common substrate for cellulitis. Overall the use of CUS had minimal impact on patient management and the routine use of CUS to assess for DVT in hospitalized patients with cellulitis appears to be an example of overuse. Conflict of Interest Statement The authors have no conflict of interest related to this report. Acknowledgements The authors gratefully thank Dr. Amy Justice for her assistance and review of the manuscript. References [1] HCUP Databases. Healthcare Cost and Utilization Project (HCUP). 2013. www.hcup-us. ahrq.gov/nisoverview.jsp. Accessed March 1,2014.
[2] Falagas ME, Vergidis PI. Narrative review: diseases that masquerade as infectious cellulitis. Ann Intern Med 2005;142(1):47–55. [3] Stevens DL, Eron LL. Cellulitis and soft-tissue infections. Ann Intern Med Jan 6 2009; 150(1):ITC11. [4] Gunderson CG, Chang JJ. Risk of deep vein thrombosis in patients with cellulitis and erysipelas: a systematic review and meta-analysis. Thromb Res Sep 2013;132(3): 336–40. [5] Jenkins TC, Sabel AL, Sarcone EE, Price CS, Mehler PS, Burman WJ. Skin and softtissue infections requiring hospitalization at an academic medical center: opportunities for antimicrobial stewardship. Clin Infect Dis 2010;51(8):895–903. [6] Maze MJ, Pithie A, Dawes T, Chambers ST. An audit of venous duplex ultrasonography in patients with lower limb cellulitis. N Z Med J 2011;124(1329):53–6. [7] Swartz MN. Clinical practice. Cellulitis. N Engl J Med 2004;350(9):904–12. [8] Wells PS, Anderson DR, Rodger M, Forgie M, Kearon C, Dreyer J, et al. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med 2003; 349(13):1227–35. [9] Kearon C, Kahn SR, Agnelli G, Goldhaber S, Raskob GE, Comerota AJ. Antithrombotic therapy for venous thromboembolic disease: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. (8th Edition) Chest Jun 2008;133(6 Suppl.):454S–545S. [10] Stevens DL, Bisno AL, Chambers HF, Everett ED, Dellinger P, Goldstein EJ, et al. Practice guidelines for the diagnosis and management of skin and soft-tissue infections. Clin Infect Dis 2005;41(10):1373–406. [11] Maze MJ, Skea S, Pithie A, Metcalf S, Pearson JF, Chambers ST. Prevalence of concurrent deep vein thrombosis in patients with lower limb cellulitis: a prospective cohort study. BMC Infect Dis 2013;13:141. [12] Lewis SD, Peter GS, Gomez-Marin O, Bisno AL. Risk factors for recurrent lower extremity cellulitis in a U.S. Veterans Medical Center population. Am J Med Sci Dec 2006;332(6):304–7. [13] Karppelin M, Siljander T, Vuopio-Varkila J, Kere J, Huhtala H, Vuento R, et al. Factors predisposing to acute and recurrent bacterial non-necrotizing cellulitis in hospitalized patients: a prospective case–control study. Clin Microbiol Infect Jun 2010; 16(6):729–34. [14] Katz MH. Overuse of health care: where are the data? Arch Intern Med Jan 23 2012; 172(2):178. [15] Berwick DM, Hackbarth AD. Eliminating waste in US health care. JAMA 2012; 307(14):1513–6. [16] Qaseem A, Alguire P, Dallas P, Feinberg LE, Fitzgerald FT, Horwitch C, et al. Appropriate use of screening and diagnostic tests to foster high-value, cost-conscious care. Ann Intern Med 2012;156(2):147–9. [17] Cassel CK, Guest JA. Choosing wisely: helping physicians and patients make smart decisions about their care. JAMA 2012;307(17):1801–2. [18] Garner S, Littlejohns P. Disinvestment from low value clinical interventions: NICEly done? BMJ 2011;343:d4519. [19] Korenstein D, Falk R, Howell EA, Bishop T, Keyhani S. Overuse of health care services in the United States: an understudied problem. Arch Intern Med 2012;172(2): 171–8. [20] Goodacre S, Sutton AJ, Sampson FC. Meta-analysis: The value of clinical assessment in the diagnosis of deep venous thrombosis. Ann Intern Med 2005; 143(2):129–39. [21] Hirsch DR, Ingenito EP, Goldhaber SZ. Prevalence of deep venous thrombosis among patients in medical intensive care. JAMA 1995;274(4):335–7. [22] Dennis M, Mordi N, Graham C, Sandercock P. The timing, extent, progression and regression of deep vein thrombosis in immobile stroke patients: observational data from the CLOTS multicenter randomized trials. J Thromb Haemost Nov 2011; 9(11):2193–200. [23] Zubrow MT, Urie J, Jurkovitz C, Jiang X, Bowen JR, DiSabatino A, et al. Asymptomatic deep vein thrombosis in patients undergoing screening duplex ultrasonography. J Hosp Med Jan 2014;9(1):19–22. [24] Ten Cate-Hoek AJ, Prins MH. Management studies using a combination of D-dimer test result and clinical probability to rule out venous thromboembolism: a systematic review. J Thromb Haemost Nov 2005;3(11):2465–70.