Routine use of LMWH prophylaxis is associated with a lower incidence of venous thromboembolic events following an ankle fracture

Open AccessPublished:November 15, 2021DOI:https://doi.org/10.1016/j.injury.2021.11.028

      Highlights

      • VTE was diagnosed in 1.7% of the patients with an operatively treated ankle fracture.
      • Routine use of LMWH was associated with a lower incidence of VTE.
      • VTE was most common during the first week among patients without prescribed LMWH.
      • The onset was later among patients with LMWH prophylaxis who still suffered a VTE.

      Abstract

      Background

      Venous thromboembolic events (VTE) are well-known and serious complications following a trauma to the lower extremities. There is an ongoing debate on the benefit of low-molecular-weight heparin (LMWH) as prophylaxis following ankle fracture treatment. We examined the association between the incidence of VTE and the use of LMWH-prophylaxis following an ankle fracture, as well as factors affecting the risk of VTE.

      Methods

      In this retrospective cohort study, data on ankle fractures and fracture treatment from the Swedish Fracture Register was linked to data from the Swedish National Patient Register and the Swedish Prescribed Drug Register. Patients with VTE and patients who received LMWH prophylaxis were identified. The treating orthopedic departments were sent a questionnaire about their guidelines regarding the use of LMWH prophylaxis.

      Results

      222 cases of diagnosed VTE were identified among 14,954 ankle fractures. Orthopedic departments with higher-than-average use of LMWH prophylaxis among non-operatively treated ankle fractures had a lower incidence of VTE (OR 0.60, CI 0.39–0.92). Among operatively treated patients, departments with a guideline for the routine use of LMWH prophylaxis also had lower incidence of VTE (OR 0.56, CI 0.37–0.86). A later onset of VTE was seen among patients prescribed LMWH prophylaxis, with a mean of 56 days to onset (CI 44–67), compared to 39 days (CI 33–45) in patients without prescribed prophylaxis. During the first two weeks following injury, there was only one case of VTE in patients with prescribed LMWH, compared to 39 cases of VTE among patients without prescribed prophylaxis.

      Conclusions

      Routine use of LMWH in patients with operatively treated ankle fractures was associated with a lower incidence of VTE. A more frequent use of LMWH among patients with non-operatively treated ankle fractures were associated with a lower incidence of VTE. The onset occurred later among patients with LMWH-prophylaxis who still suffered a VTE.
      DVT (Deep vein thrombosis), LMWH (Low-molecular-weight heparin), PE (Pulmonary embolism), SFR (Swedish Fracture Register), VTE (Venous thromboembolic event)

      Keywords

      Background

      Venous thromboembolic event (VTE) is a multifactorial disease and a result of the interaction between genetic and environmental risk factors. Obesity, surgery, immobilization, trauma or fracture, cancer, oral contraception, increasing age and atherosclerosis are known or suggested risk factors [
      • Heit J.A.
      Epidemiology of venous thromboembolism.
      ,
      • Cushman M.
      Epidemiology and risk factors for venus thrombosis.
      ]. A substantial mortality rate is associated with VTE, especially from pulmonary embolism (PE), even if mortality is affected by increasing age and comorbid conditions [
      • Tagalakis V.
      • Patenaude V.
      • Kahn S.R.
      • Suissa S.
      Incidence of and mortality from venous thromboembolism in a real-world population: the Q-VTE Study Cohort.
      ]. Other common and serious complications are post-thrombotic syndrome and recurrence of VTE [
      • Leizorovicz A.
      Long-term consequences of deep vein thrombosis.
      ].
      Ankle fractures are among the most common fractures, with an incidence of about 170 cases per 100,000 person-years [
      • Elsoe R.
      • Ostgaard S.E.
      • Larsen P.
      Population-based epidemiology of 9767 ankle fractures.
      ,
      • Juto H.
      • Nilsson H.
      • Morberg P.
      Epidemiology of Adult Ankle Fractures: 1756 cases identified in Norrbotten County during 2009-2013 and classified according fot AO/OTA.
      ]. About half of them are treated operatively []. Meta-analysis shows an incidence of 2–3% of symptomatic VTE in patients with immobilization of the lower limb without use of prophylaxis [
      • Chapelle C.
      • Rosencher N.
      • P Jacques Zufferey
      • Mismetti P.
      • Cucherat M.
      • Laporte S.
      • Group Meta-Embol
      Prevention of venous thromboembolic events with low-molecular-weight heparin in the non-major orthopaedic setting: meta-analysis of randomized controlled trials.
      ,
      • Hickey B.A.
      • Watson U.
      • Cleves A.
      • Alikhan R.
      • Pugh N.
      • Nokes L.
      • Perera A.
      Does thromboprophylaxis reduce symptomatic venous thromboembolism in patients with below knee cast treatment for foot and ankle trauma? A systematic review and meta-analysis.
      ,
      • Zee A.A.
      • van Lieshout K.
      • van der Heide M.
      • Janssen L.
      • Janzing H.M.
      Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
      ]. The number of asymptomatic VTEs is considerably higher than symptomatic VTEs. Lapidus et al. found 28% of deep vein thrombosis (DVT) by phlebography after operatively treated ankle fractures [
      • Lapidus L.J.
      • Ponzer S.
      • Elvin A.
      • Levander C.
      • Lärfars G.
      • Rosfors S.
      • de Bri E.
      Prolonged thromboprophylaxis with Dalteparin during immobilization after ankle fracture surgery: a randomized placebo-controlled, double-blind study.
      ]. The Cochrane Vascular Group conclude that moderate-quality evidence showed that use of low-molecular-weight heparin (LMWH) reduced the incidence of DVT in outpatients with lower limb immobilization [
      • Zee A.A.
      • van Lieshout K.
      • van der Heide M.
      • Janssen L.
      • Janzing H.M.
      Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
      ].
      Our objectives were to determine the incidence of diagnosed VTE among both operatively and non-operatively treated ankle fractures registered in the Swedish Fracture Register (SFR). We investigated the use of LMWH prophylaxis among the orthopedic departments affiliated to the SFR and how different use affects the incidence of VTE. We also examined factors affecting the risk of VTE.

      Methods

       Study design, participants, and data source

      In this retrospective cohort study, we included all patients 18 years of age or older with an ankle fracture registered in the SFR and treated at one of the 33 orthopedic departments in Sweden that reported to the SFR between January 1, 2012 and June 30, 2016.
      These patients were, via their unique personal identification number, linked to the Swedish National Patient Register and the Swedish Prescribed Drug Register, both administered by the Swedish National Board of Health and Welfare.
      The Swedish Fracture Register (SFR) has been prospectively, non-mandatorily collecting data on all orthopedic fractures since 2012. Data on injury mechanism, fracture classification, method of treatment and patient-reported outcome are registered in the SFR and has been more thoroughly described by Wennergren et al. [
      • Wennergren D.
      • Ekholm C.
      • Sandelin A.
      • Möller M.
      The Swedish fracture register: 103,000 fractures registered.
      ,
      • Wennergren D.
      • Möller M.
      Implementation of the Swedish fracture register.
      ]. The accuracy of the classification of registered fractures in the SFR has been examined in several earlier studies and the accuracy of ankle fracture classification was reported to be substantial [
      • Knutsson S.B.
      • Wennergren D.
      • Bojan A.
      • Ekelund J.
      • Möller M.
      Femoral fracture classification in the Swedish Fracture Register - a validity study.
      ,
      • Wennergren D.
      • Stjernström S.
      • Möller M.
      • Sundfeldt M.
      • Ekholm C.
      Validity of humerus fracture classification in the Swedish fracture register.
      ,
      • Wennergren D.
      • Ekholm C.
      • Sundfeldt M.
      • Karlsson J.
      • Bhandari M.
      • Möller M.
      High reliability in classification of tibia fractures in the Swedish Fracture Register.
      ,

      Juto H., Möller M., Wennergren D., Edin K., Apelqvist I., Morberg P. Substantial accuracy of fracture classification in the Swedish Fracture Register: evaluation of AO/OTA-classification in 152 ankle fractures. 2016;47:2579–83. 10.1016/j.injury.2016.05.028.

      ].
      The data in this study includes the first years of the SFR and the coverage and completeness rose as more departments joined the register. The estimated completeness of registered ankle fractures in the SFR for the first year of the study, 2012, was 5% and for the final year, 2016, it was 41% (Appendix 1). The mean completeness of the contributing departments for all fractures of the extremities for 2018 was 70% [

      Socialstyrelsen [The National Board of Health and Welfare]. Täckningsgrader 2019, jämförelser mellan nationella kvalitetsregister och hälsodataregistren. [Coverage 2019 – comparison in national registers of health and medical care] available at https://www.socialstyrelsen.se/globalassets/sharepoint-dokument/artikelkatalog/ovrigt/2019-12-6489-tabeller.xlsx (Accessed 6 September 2020).

      ].
      The diagnosis retrieved from the Swedish National Patient Register was from the period between 2011 and 2017. The SNPR started collecting data on inpatient discharge diagnoses in 1964. The register is mandatory for both privately and publicly funded health care. The coverage of inpatient discharge diagnosis is about 99% with a positive predictive value generally between 85 and 95% depending on the diagnosis. Recording of diagnoses for outpatients began in 2001 and the missing data of the main diagnosis is reported to be around 3% [
      • Ludvigsson J.F.
      • Andersson E.
      • Ekbom A.
      • Feychting M.
      • Kim J.L.
      • Reuterwall C.
      • et al.
      External review and validation of the Swedish national inpatient register.
      ,

      Socialstyrelsen [The National Board of Health and Welfare]. Bortfall och kvalitet i patientregistret [Loss and quality of the swedish national patient register]. Available at: https://www.socialstyrelsen.se/statistik-och-data/register/alla-register/patientregistret/bortfall-och-kvalitet/ (Accessed 7 August 2020).

      ].
      All medical prescriptions in Sweden have been registered in the Swedish Prescribed Drug Register since 2005, as well as information about the date it was dispensed at the pharmacy. Recording of data for all prescribed drugs is mandatory and automated. The loss of data on prescribed drugs is estimated as negligible. However, data on over-the-counter drugs and drugs used at hospitals is not included in the register. Prescribed drugs constitute 84%, and hospital use is 3% of defined daily doses of the total quantity of drugs sold in Sweden [
      • Wettermark B.
      • Hammar N.
      • Fored C.M.
      • Leimanis A.
      • Otterblad Olausson P.
      • Bergman I.
      • et al.
      The new Swedish Prescribed Drug Register-opportunities for pharmacoepidemiological research and experience from the first six months.
      ,

      Socialstyrelsen [The National Board of Health and Welfare]. Bortfall och kvalitet om läkemedelsregistret [Loss and quality of the swedish prescribed drug register]. Available at: https://www.socialstyrelsen.se/statistik-och-data/register/alla-register/lakemedelsregistret/bortfall-och-kvalitet/ (Accessed August 7, 2020).

      ].
      Information on the guidelines used in each department regarding the use of LMWH prophylaxis was collected from the orthopedic surgeon in charge of the registrations in the SFR at each affiliated department.
      The STROBE cohort reporting guidelines were followed [
      • von Elm E.
      • Altman D.G.
      • Egger M.
      • Pocock S.J.
      • Gøtzche P.C.
      • Vandenbroucke J.P.
      The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies.
      ].

       Variables

      Data from the SFR on patients (age and sex), injury (date and type), fracture (side, classification and open or closed) and the treatment of the fracture (date, operative/non-operative) were included in this study.
      Information on all prescriptions of anticoagulants (ACT-code: B01) came from the Swedish Prescribed Drug Register. LMWH was used both as prophylaxis and as treatment for a VTE. All prescriptions of anticoagulant drugs were examined, and the use of LMWH prophylaxis was defined by the dose, the describing text of the prescriptions and the time in relation to the injury and to an eventual VTE.
      Information on cardiovascular therapy (ACT-code: C01–09), antidiabetic therapy (ATC-code: A10) and cancer (ICD-10: C00–97) was used to identify potential confounders.
      Some orthopedic departments in Sweden, instead of writing prescriptions, routinely distribute syringes of LMWH to admitted patients when discharged from the ward. This could introduce bias to the study. Therefore, all orthopedic departments which register data on adult ankle fractures in the SFR were sent questionnaires about the use of LMWH prophylaxis. They were asked both if LMWH as prophylaxis was routinely used on all patients with operatively treated ankle fractures and if the guidelines were in written form. In cases of routine use of LMWH, a follow-up question was whether they mainly wrote prescriptions or distributed the syringes to the patients when discharged from the ward.

       Outcome

      We used diagnosed or/and treated VTE as outcome. This was defined as a patient included in the cohort of ankle fractures identified in the SFR and with and ICD-10 diagnosis code in the SNPR for VTE (I 80.1–82.9 and/or I 26.0–9) and a diagnosis within six months from the date of the fracture. When the information on the prescriptions for an anticoagulant drug clearly stated treatment of VTE, these were also included. PE (I26.0–9) within six months was also used in the same way as an outcome.

       Statistical analyses

      The main assumption for the data analysis was that departments have different traditions that could be used to compare the effect of the LMWH prophylaxis. The data was analysed according to the treatment method, since we made the assessment that operatively treated patients were mainly admitted to hospital and were therefore affected by a possibility of staff distributing LMWH prophylaxis when discharged from the ward. We assessed that this risk for the non-operative treated patients was minimal. Hence, non-operatively treated cases were compared solely on the prescription rate of LMWH prophylaxis from the department treating the fracture. A mean value of the prescription was calculated, and departments were grouped according to lower or higher than average. The incidence of both all VTEs (PE, DVT and unspecified VTE) and only PE in non-operative treated ankle fractures was compared between departments with a higher-than-average use of LMWH and departments with a lower-than-average use.
      Comparing operatively treated ankle fractures was more complex because of a tradition in some departments of distributing syringes from the ward when discharging patients. Therefore, we used data from the orthopedic departments’ prescription rate as well as the departments’ guidelines in the analysis. Two departments that reported only non-routine use of LMWH prophylaxis were still categorized as routine users because of a very high prescription rate (87%). In all other cases, responses from the departments were in concurrence with the data from the Swedish Prescribed Drug Register on their prescription rate. The authors all agreed on the conclusion that comparisons according to departmental guidelines gave the most valid and accurate result among the operatively treated patients. The incidence of VTE and PE were compared between departments with guidelines stating routine use of LMWH prophylaxis to those without routine use. Cases with missing data on treatment and among operatively treated cases, on departments’ guidelines were presented separately. To minimize the risk of bias, directed acyclic graphs, as described by von Elm et al., was used when designing the model [
      • Shrier I.
      • Platt R.W.
      Reducing bias through directed acyclic graphs.
      ].
      A logistic regression model was used for analysis of variables affecting the odds of VTE, for operative and non-operative treated cases. Variables included for each model are declared in each table. Backward Stepwise method was used in the analysis. The time between the injury and the VTE was compared in patients with and without prescription of LMWH. The study size was limited by the amount of data available in the SFR and the requirement of being able to carry out a six-month follow-up from the national prescription database.
      IBM SPSS Statistics version 24 was used for the statistical calculations. 95% confidence intervals were used and achieved from the custom table module and the cross-table module in SPSS. A p-value of less than 0.05 was considered significant in the logistic regression.

      Results

       Patients and descriptive data

      From the SFR, 14,954 ankle fractures in patients 18 years of age or older were available and included in the study. 797 (5.3%) of the injuries were registered as high energy. Bilateral ankle fractures were seen in 34 (0.2%) patients, 266 (1.8%) of the fractures were open and 544 (3.6%) of the patients had multiple fractures registered, such as both ankle and wrist or hip fractures. The mean age was 55 years (range 18–104). Of the patients included, 41% had at least one cardiovascular medication, 9% had an antidiabetic drug and 1.1% had a cancer diagnosis (Table 1).
      Table 1Demography, characteristics and amount VTE according to treatment and use of LMWH. Values are proportions in percent (95% CI) unless otherwise specified.
      All included casesNon-operative treatmentOperative treatmentUnknown treatment
      All non-operatively treated casesDepartments with lower than mean precriptionsaDepartments with higher than mean precriptionsaAll operatively treated casesLMWH use after risk evaluationbRoutinely use LMWHbUnknown guidelineb
      n = 14,954n = 7217n = 3487n = 3730n = 7310n = 1143n = 5460n = 707n = 427
      Contributing departments, nc3333171633822329
      Women60 (59–60)62 (60–63)63 (61–64)61 (59–62)58 (56–59)57 (54–60)58 (56–59)58 (54–61)60 (55–64)
      Age, mean in years55 (55–56)57 (57–57)58 (57–59)56 (56–57)54 (53–54)54 (53–55)53 (53–54)55 (54–56)54 (52–56)
      AO/OTA type A24 (24–25)41 (39–42)39 (38–41)42 (40–43)9 (8–10)10 (8–12)9 (8–10)7 (5–9)18 (14–21)
      AO/OTA type B63 (62–64)55 (54–56)56 (54–57)54(53–56)71 (70–72)69 (67–72)71 (70–73)72 (68–75)64 (59–68)
      AO/OTA type C12 (11–12)3 (3–4)4 (3–4)3 (2–4)20 (19–20)20 (18–22)19 (18–20)20 (18–24)18 (14–21)
      Open fractures1.8 (1.6–2.0)0.4 (0.2–0.5)0.3 (0.2–0.6)0.4 (0.2–0.6)3.1 (2.7–3.5)3.1 (2.3–4.3)3.1 (2.7–3.6)3.0 (1.9–4.4)3.3 (1.9–5.3)
      Cardiovaskular drug41 (41–42)44 (42–45)45 (44–47)42 (40–44)39 (38–40)41 (38–43)39 (38–40)39 (36–43)41 (36–46)
      Diabetes9 (9–10)10 (10–11)11 (10–12)10 (9–11)8 (8–9)9 (7–11)8 (8–9)8 (6–10)11 (8–14)
      Malignancy1.1 (1.0–1.3)0.2 (0.1–0.3)0.2 (0.1–0.4)0.2 (0.1–0.3)2.0 (1.7–2.4)2.1 (1.4–3.1)2.0 (1.7–2.4)2.3 (1.4–3.6)1.2 (0.4–2.5)
      LMWH prescription36 (35–36)13 (12–14)8 (7–9)18 (17–19)58 (56–59)20 (17–22)66 (64–67)56 (53–60)46 (42–51)
      Distribution LMWH typed25/10/1/19/3/0/04/2/1/115/3/0/039/17/2/17/9/2/143/20/2/155/1/1/035/8/3/1
      Length prophylaxis, mean in days16 (15–16)19 (18–20)21 (18–25)18 (17–19)15 (14–15)23 (21–26)14 (14–14)14 (13–16)18 (16–20)
      VTE1.5 (1.3–1.7)1.3 (1.0–1.5)1.6 (1.2–2.1)0.9 (0.7–1.3)1.7 (1.4–2.0)2.6 (1.8–3.7)1.5 (1.2–1.8)2.0 (1.1–3.2)1.4 (0.6–2.9)
      PE0.2 (0.1–0.3)0.2 (0.1–0.3)0.1 (0.1–0.3)0.2 (0.1–0.4)0.2 (0.1–0.3)0.3 (0.1–0.8)0.1 (0.1–0.3)0.3 (0.1–0.9)0.5 (0.1–1.5)
      VTE among non-prescriptione1.4 (1.2–1.7)1.2 (1.0–1.5)1.6 (1.2–2.0)0.9 (0.6–1.3)1.8 (1.4–2.3)2.7 (1.8–3.9)1.6 (1.1–2.2)0.3 (0.0–1.5)0.9 (0.2–2.8)
      VTE among prescriptione1.6 (1.3–2.0)1.5 (0.9–2.4)2.3 (1.0–4.6)1.2 (0.6–2.3)1.6 (1.3–2.1)2.2 (0.9–4.8)1.4 (1.1–1.9)3.3 (1.8–5.4)2.0 (0.7–4.7)
      a Cases at departments grouped by the departments prescription rate in accordance with the total mean of 13%.
      b Cases at departments grouped according to guideline of LMWH use.
      c Number of different departments contribution with cases to the subgroup.
      d Distribution of drug and dosage of LMWH (Dalteparin 5000 IU/Tinzaparin 4500 IU/Enoxaparin 4000 IU/other) counted for the whole group of patients included.
      e VTE among cases with or without prescription of LMWH.

       Departments’ guidelines

      We received responses from 30 out of 33 affiliated departments. Of the 30 orthopedic departments, 10 (33%) responded that they did not have any guidelines of routine use of LMWH as VTE prophylaxis on operative treated patients with ankle fractures. Twelve departments (40%) routinely prescribed LMWH and 8 departments (27%) routinely distributed the syringes of LMWH for the rest of the treatment period when the patients were discharged.

       Use of LMWH

      Among the non-operative treated patients, 13% were prescribed LMWH. This group had higher comorbidity with significantly higher prevalence of diabetes, cardiovascular disease, and malignancy. Instead, when comparing lower than average LMWH-using departments to higher than average using departments there was less of a difference in comorbidity. Of the operative treated patients, 58% were prescribed LMWH when all departments were included in the analysis.
      Dalteparin 5000 IU was the most common LMWH prophylaxis prescribed in the study and was used in 69% of the cases. It was mainly used by 17 of the contributing departments. Tinzaparin 4500 IU was used in 27% and was mainly used by eleven of the departments. Enoxaparin 4000 IU was used in 3% of the cases and mainly used by two departments. Other doses of those drugs constituted the last part of less than 2%. The mean period with prescribed LMWH prophylaxis was 16 days. The most common length of the prescriptions was 10 days, and this was the case in 61% of the prescriptions. Among the operatively treated cases the after risk evaluation using departments had a longer mean duration of the prescribed LMWH-prophylaxis than routinely using departments. In non-operatively treated cases, the same was seen among the lower than average LMWH-using departments. Distribution of type, dosage, and mean duration of the prescribed LMWH among different groups used in this study are shown in Table 1.

       Venous thromboembolism

      In the study, 222 cases (1.5%) of VTE were identified within six months of the fracture. Among those, 188 were identified from the ICD-10 classification and 34 from the information on the prescription label. Of the 222 cases of VTE, 136 were diagnosis with deep vein thrombosis, 64 were an unspecified VTE and 28 were pulmonary embolism. Six of the cases with PE also had a diagnosis of DVT or unspecified VTE. In the operatively treated patients, there were 125 cases of VTE (1.7%), 91 cases (1.3%) in non-operatively treated patients, and 6 (1.4%) cases in patients with an unknown treatment.
      VTE was most common during the first week among patients who had not been prescribed LMWH and the third week among patients prescribed LMWH. There was only one case of VTE within the first 2 weeks detected among patients who were prescribed LMWH compared with 39 case among patients without LMWH-prescription (Fig. 1). In patients prescribed LMWH, the VTE had a significantly later onset, with a mean of 56 (CI 44–67) days, compared to 39 (CI 33–45) days in patients without prophylaxis (Table 2).
      Fig. 1
      Fig. 1Number of VTE-cases weeks after injury in patients with or without prescription of LMWH-prophylaxis. All departments, 9622 cases without and 5332 cases with LMWH-prescription were included.
      Table 2Mean time in days (CI 95%) between injury and VTE in cases with or without prescription of LMWH-prophylaxis.
      Non LMWH-prophylaxisLMWH-prophylaxis
      DaysNDaysN
      Non-operative treatment39 (31–47)7778 (46–111)14
      Operative treatment40 (29–51)5650 (37–64)69
      Operative treatment (excluding 11 departments)a32 (22–42)3847 (37–57)47
      Totalb39 (33–45)13556 (44–67)87
      a 8 departments which routinely use LMWH-prophylaxis from wards and 3 departments with unknown guidelines were excluded.
      b Also including unknown treatment.
      Among operatively treated patients there was a lower incidence of VTE treated in departments with a guideline of routine use of LMWH prophylaxis. In these departments, 1.5% of the patients had a VTE compared to 2.6% among patients at departments with risk evaluation before use of LMWH (OR 0.56, CI 0.37–0.86).
      When comparing non-operatively treated patients in higher-than-average LMWH-using departments to those in lower than average, the incidence of VTE was lower 0.9% compared to 1.6% (OR 0.60, CI 0.39–0.92).

       Regression models

      In the final logistic regression model for the incidence of VTE among non-operatively treated ankle fracture patients, three variables remained significant. A cancer diagnosis increased the odds of getting VTE by 11.6 times. Every year of age increased the odds of getting a VTE by 2%. Being treated at a department with a higher than average use of LMWH decreased the odds of VTE by 40%. No other explanatory variable included in the preliminary model was significant (Table 3).
      Table 3Logistic regression with variables affecting the odds of VTE among non-operatively treated ankle fractures.
      OR95% CIp-value
      Increasing age (per year)1.021.00–1.030.007
      Malignancy diagnosis11.62.5–540.002
      Treatment at department with higher than mean use of LMWH0.600.39–0.920.02
      Non-significant variables in the model were: sex, cardiovascular drug, diabetes mellitus, AO/OTA type, complex injury (multiple/open/bilateral fracture), other anticoagulant before injury, prescription of LMWH prophylaxis and length LMHW prophylaxis.
      The final logistic regression model for the incidence of VTE among operative treated ankle fractures contained two significant variables. Treatment for cardiovascular disease increased the odds of getting VTE by 78% and being treated at a department with a guideline of routine use of LMWH decreased the odds by 44%. No other explanatory variable included in the preliminary model was significant (Table 4).
      Table 4Logistic regression with variables affecting odds of VTE among operatively treated ankle fractures.
      OR95% CIp-value
      Cardiovascular drug prescription1.781.22–2.590.000
      Treated at department with routinely use of LMWH0.560.37–0.860.008
      Non-significant variables in the model were: age, sex, malignancy diagnosis, diabetes mellitus, AO/OTA type, complex injury (multiple/open/bilateral fracture), other anticoagulant before injury, prescription of LMWH prophylaxis and length LMHW prophylaxis.

      Discussion

      In this study we observed a higher incidence of VTE among patients treated for both non-operatively and operatively treated ankle fractures in departments with a low use of LMWH prophylaxis.
      The use of LMWH prophylaxis in patients with non-operatively treated ankle fractures is still controversial, even though it results in a greater reduction of incidence of VTE compared to operatively treated ankle and foot fractures [
      • Zee A.A.
      • van Lieshout K.
      • van der Heide M.
      • Janssen L.
      • Janzing H.M.
      Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
      ]. In our study, only 13% of non-operative treated patients were prescribed LMWH prophylaxis. Those who received prophylaxis had a higher number of comorbidities compared to patients without prophylaxis. Hence, it appears as if LMWH prophylaxis is mainly used following a risk assessment.
      In the non-prophylactic group, there were more AO/OTA type A fractures. A type A fracture is not always treated with a cast and therefore probably at less risk of VTE. We do not see any effect in the logistic regression that a non-operatively treated AO/OTA type B or C would increase the risk of a VTE compared to an AO/OTA type A.
      The difference in use of LMWH between the higher and lower than mean using departments was only about 10% (8% vs 18%) but this still gave an OR 0.60. This was a surprisingly substantial reduction of incidence. It can be compared with an OR 0.31 shown by Zee et al. [
      • Zee A.A.
      • van Lieshout K.
      • van der Heide M.
      • Janssen L.
      • Janzing H.M.
      Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
      ] between intervention and control groups but in both symptomatic and non-symptomatic DVTs. Our result showed a substantial risk reduction with only a slightly higher use of LMWH. Our interpretation is that a wider use of prophylaxis in non-operative treated patients with ankle fractures and risk factors of VTE have an effect and is plausible. This is also in concordance with the result shown by Zee et al. [
      • Zee A.A.
      • van Lieshout K.
      • van der Heide M.
      • Janssen L.
      • Janzing H.M.
      Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
      ].
      LMWH prophylaxis prescribed to patients with operatively treated ankle fractures is less controversial. All departments in the study used it to some extent, either as a routine or after a risk evaluation of the patients. One of the main research questions in this study was if the routine use of LMWH prophylaxis would result in a reduction in the amount of VTEs compared to the use of a risk evaluation in operatively treated patients with ankle fractures. One third of the departments in the study responded that they use individual risk evaluations to make the decision of whether prophylaxis should be given. About one fifth of the operative treated patients in the study were treated at these departments and about one fifth were prescribed LMWH prophylaxis. When comparing these departments to departments routinely using LMWH, there was a substantial reduction of incidence of VTE with an odd ratio of 0.56. Zee et al. [
      • Zee A.A.
      • van Lieshout K.
      • van der Heide M.
      • Janssen L.
      • Janzing H.M.
      Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
      ] showed about the same odds ratio (0.54) among patients with prophylaxis on operatively treated ankle fractures, but in both symptomatic and non-symptomatic DVTs.
      Departments with no guidelines of routine use of LMWH prescribed to a higher degree extended time prophylaxis than departments with such guidelines. Hence, less use of prophylaxis seems to be correlated with a longer duration. This strengthens the conclusion that there were mainly patients with risk factors or perceived risk factors for VTE who were prescribed LMWH in these departments. This is also seen to some extent among the non-operative treated cases. This might affect the result of the study but in that case, it would make it more difficult to show a difference between the groups.
      Another important finding of the study was that patients who received prophylaxis but still developed VTE had a later onset of the VTE. Interestingly, there was only one VTE in the first two weeks in this group. Lapidus et al. [
      • Lapidus L.J.
      • Ponzer S.
      • Elvin A.
      • Levander C.
      • Lärfars G.
      • Rosfors S.
      • de Bri E.
      Prolonged thromboprophylaxis with Dalteparin during immobilization after ankle fracture surgery: a randomized placebo-controlled, double-blind study.
      ] compared 1 week and 6 weeks of prophylaxis in a randomised controlled trial and could not show any significant difference in asymptomatic DVT incidence. With 228 patients, this study might have been underpowered to show a difference. In the logistic regression of the present, retrospective study, the extended LMWH prophylaxis did not affect the incidence of VTE. This was not expected either. In this context, long-term prophylaxis should be seen as a sign of the treated patient having risk factors of VTE. When interpreting the data of this study and especially Fig. 1, the question still emerges whether prophylaxis after an ankle fracture should be longer than ten days.
      The main strength of the study is the extent of the variables included in the data from the SFR and by combining the other registers with high quality data. This makes it possible to detect quite rare but very serious events like VTE in everyday orthopedic settings. The large number of contributing departments in the SFR strengthen the external validity and gives the possibility of examining different treatment traditions. Previously published meta-analyses show significant effects but have included different diagnoses and different treatment methods. The studies included in this study have a more defined study group but often the sample size is too small to be able to show a significant effect. For example, Selby et al. halted the recruitment to their study of clinically important VTE among operatively treated lower leg fractures when it became obvious that they would not be able to show a significant difference even if they doubled their initial sample size [
      • Selby R.
      • Geerts W.H.
      • Kreder H.J.
      • Crowther M.A.
      • Kaus L.
      • Sealey F.
      d-KAF (Dalteparin in Knee-to-Ankle Fracture) Investigators. A double-blind, randomized controlled trial of the prevention of clinically important venous thromboembolism after isolated lower leg fractures.
      ].
      There are obvious limitations to the current study. The study design is a non-randomized retrospective study. One major weakness is the tradition in some departments of giving patients syringes of LMWH when the patient is discharged. Therefore, parts of our data are less valid. Furthermore, we did not have data on known risk factors like obesity, physical status, family history of VTE or oral contraception. Also, early post-injury mobilization and weight-bearing could also influence the rate of VTE. The risk of these factors covary with the pattern of different treatment traditions of the different departments. This must be considered as very low and therefore we believe that the risk of bias this could introduce is minor. Another limitation, as this was a register study, was we had no information on how the diagnosis of VTE in each case was carried out. We used registered diagnosis as stated in Appendix 2 from the Swedish National Patient Register. It is mandatory to report all diagnoses in the register by both public and private health care. The recommendation in Sweden is the use of a combination of clinical assessment, using Wells score and D-dimer, in the diagnosis of a suspected DVT and if these are positive, further investigation with for example duplex ultrasound [
      Socialstyrelsen [The National Board of Health and Welfare]
      Socialstyrelsens riktlinjer för vård av blodpropp/venös tromboembolism 2004 [Guidelines of treatment of venous thromoboembolism 2004].
      ]. We believe that the quality of the included data of diagnosed VTE is quite high, especially considering that we used both information from registered diagnosis as well as prescribed treatments to find cases of VTE. Another limitation was the low coverage of the first years of the SFR and a risk of attrition bias. However, there are still many contributing departments in the study and therefore we assess this risk of bias as low.
      The use of LMWH can produce adverse effects that must be considered. In this study we have not investigated the complications following use of LMWH. However, bleeding and heparin-induced thrombocytopenia are well known and serious but rare complications [
      • Eck R.J.
      • Bult W.
      • Wetterslev J.
      • Gans R.O.B.
      • Meijer K.
      • Keus F.
      • van der Horst I.C.C.
      Intermediate dose low-molecular-weight heparin for thrombosis prophylaxis: systematic review with meta-analysis and trial sequential analysis.
      ,
      • Ahmed I.
      • Majeed A.
      • Powell R.
      Heparin induced thrombocytopenia: diagnosis and management update.
      ]. A doubling of minor adverse effects between patients with LMWH prophylaxis and those without has been reported and Hickey et al. showed that one major bleeding occurred in every ten avoided symptomatic DVTs [
      • Hickey B.A.
      • Watson U.
      • Cleves A.
      • Alikhan R.
      • Pugh N.
      • Nokes L.
      • Perera A.
      Does thromboprophylaxis reduce symptomatic venous thromboembolism in patients with below knee cast treatment for foot and ankle trauma? A systematic review and meta-analysis.
      ,
      • Zee A.A.
      • van Lieshout K.
      • van der Heide M.
      • Janssen L.
      • Janzing H.M.
      Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
      ]. Even if we show, in concurrence with earlier studies, that there is a substantial reduction of VTE among patients with prophylaxis there is still a question as to risk and cost versus benefits. The reduction of diagnosed VTE by half but from a rather low level against the very rare but serious complications of the use of LMWH. A register-based randomized trial maybe would be needed to get enough precision and power.

      Conclusion

      Routine use of LMWH in patients with operatively treated ankle fractures was associated with a lower incidence of VTE. A more frequent use of LMWH among patients with non-operatively treated ankle fractures was associated with a lower incidence of VTE. The onset occurred later among patients with LMWH-prophylaxis who still suffered a VTE.

      Ethics

      The study was approved by the Regional Ethic Committee at Umeå University (d-nr: 2016/230–31).

      CRediT authorship contribution statement

      Hans Juto: Methodology, Formal analysis, Data curation, Investigation, Writing – original draft, Writing – review & editing. Magnus Hultin: Methodology, Formal analysis, Data curation, Writing – review & editing. Michael Möller: Methodology, Writing – review & editing. Per Morberg: Methodology, Supervision, Funding acquisition, Writing – review & editing.

      Acknowledgement

      The study was funded by the Department of Research and Innovation at Region Norrbotten, Sweden.

      Appendix. Supplementary materials

      Appendices

      Appendix 1Amount of ankle fractures in Sweden year by year and completeness in the SFR, calculated on an incidence of 170 ankle fractures per 100,000 person-years.
      Year20122013201420152016201720182019
      Adult population7,629,7847,694,3997,764,0877,827,9557,920,7628,000,6618,076,8248,149,100
      Estimated amount ankle fractures13,73413,85013,97514,09014,25714,40114,53814,668
      Registered ankle fractures6593115383350825866653067077055
      Completeness5%22%27%36%41%45%46%48%
      No of contributing departments516242937404044
      Appendix 2Used ICD-10 diagnosis as definition of VTE in the study.
      I80.1 Phlebitis and thrombophlebitis of femoral vein
      I80.2 Phlebitis and thrombophlebitis of other and unspecified deep vessels of lower extremities
      I80.3 Phlebitis and thrombophlebitis of lower extremities, unspecified
      I80.8 Phlebitis and thrombophlebitis of other sites
      I80.9 Phlebitis and thrombophlebitis unspecified site
      I26.0 Pulmonary embolism with acute cor pulmonale
      I26.9 Pulmonary embolism without acute cor pulmonale

      References

        • Heit J.A.
        Epidemiology of venous thromboembolism.
        Nat Rev Cardiol. 2015; 12: 464-474https://doi.org/10.1038/nrcardio.2015.83
        • Cushman M.
        Epidemiology and risk factors for venus thrombosis.
        Semin Hematol. 2007; 44: 62-69
        • Tagalakis V.
        • Patenaude V.
        • Kahn S.R.
        • Suissa S.
        Incidence of and mortality from venous thromboembolism in a real-world population: the Q-VTE Study Cohort.
        Am J Med. 2013; 126 (832.e13-21)https://doi.org/10.1016/j.amjmed.2013.02.024
        • Leizorovicz A.
        Long-term consequences of deep vein thrombosis.
        Haemostasis. 1998; 28 (Suppl 3): 1-7
        • Elsoe R.
        • Ostgaard S.E.
        • Larsen P.
        Population-based epidemiology of 9767 ankle fractures.
        Foot Ankle Surg. 2018; 24: 34-39https://doi.org/10.1016/j.fas.2016.11.002
        • Juto H.
        • Nilsson H.
        • Morberg P.
        Epidemiology of Adult Ankle Fractures: 1756 cases identified in Norrbotten County during 2009-2013 and classified according fot AO/OTA.
        BMC Musculoskelet Disord. 2018; 19: 441https://doi.org/10.1186/s12891-018-2326-x
      1. The Swedish Fracture Register: Annual Report 2019 (in Swedish). Available at: https://registercentrum.blob.core.windows.net/sfr/r/VGR0050_SFR_-rsrapport-2019_Digital-1-uppslag-BJl2qw9v38.pdf.

        • Chapelle C.
        • Rosencher N.
        • P Jacques Zufferey
        • Mismetti P.
        • Cucherat M.
        • Laporte S.
        • Group Meta-Embol
        Prevention of venous thromboembolic events with low-molecular-weight heparin in the non-major orthopaedic setting: meta-analysis of randomized controlled trials.
        Arthroscopy. 2014; 30: 987-996https://doi.org/10.1016/j.arthro.2014.03.009
        • Hickey B.A.
        • Watson U.
        • Cleves A.
        • Alikhan R.
        • Pugh N.
        • Nokes L.
        • Perera A.
        Does thromboprophylaxis reduce symptomatic venous thromboembolism in patients with below knee cast treatment for foot and ankle trauma? A systematic review and meta-analysis.
        Foot Ankle Surg. 2018; 24: 19-27https://doi.org/10.1016/j.fas.2016.06.005
        • Zee A.A.
        • van Lieshout K.
        • van der Heide M.
        • Janssen L.
        • Janzing H.M.
        Low molecular weight heparin for prevention of venous thromboembolism in patients with lower-limb immobilization.
        Cochrane Database Syst. 2017; 8 (pub4)CD006681https://doi.org/10.1002/14651858.CD006681
        • Lapidus L.J.
        • Ponzer S.
        • Elvin A.
        • Levander C.
        • Lärfars G.
        • Rosfors S.
        • de Bri E.
        Prolonged thromboprophylaxis with Dalteparin during immobilization after ankle fracture surgery: a randomized placebo-controlled, double-blind study.
        Acta Orthop. 2007; 78: 528-535
        • Wennergren D.
        • Ekholm C.
        • Sandelin A.
        • Möller M.
        The Swedish fracture register: 103,000 fractures registered.
        BMC Musculoskelet Disord. 2015; 16: 338https://doi.org/10.1186/s12891-015-0795-8
        • Wennergren D.
        • Möller M.
        Implementation of the Swedish fracture register.
        Unfallchirurg. 2018; 121: 949-955https://doi.org/10.1007/s00113-018-0538-z
        • Knutsson S.B.
        • Wennergren D.
        • Bojan A.
        • Ekelund J.
        • Möller M.
        Femoral fracture classification in the Swedish Fracture Register - a validity study.
        BMC Musculoskelet Disord. 2019; 20 (May 8): 197https://doi.org/10.1186/s12891-019-2579-z
        • Wennergren D.
        • Stjernström S.
        • Möller M.
        • Sundfeldt M.
        • Ekholm C.
        Validity of humerus fracture classification in the Swedish fracture register.
        BMC Musculoskelet Disord. 2017; 18 (Jun 10): 251https://doi.org/10.1186/s12891-017-1612-3
        • Wennergren D.
        • Ekholm C.
        • Sundfeldt M.
        • Karlsson J.
        • Bhandari M.
        • Möller M.
        High reliability in classification of tibia fractures in the Swedish Fracture Register.
        Injury. 2016; 47 (Feb): 478-482https://doi.org/10.1016/j.injury.2015.11.002
      2. Juto H., Möller M., Wennergren D., Edin K., Apelqvist I., Morberg P. Substantial accuracy of fracture classification in the Swedish Fracture Register: evaluation of AO/OTA-classification in 152 ankle fractures. 2016;47:2579–83. 10.1016/j.injury.2016.05.028.

      3. Socialstyrelsen [The National Board of Health and Welfare]. Täckningsgrader 2019, jämförelser mellan nationella kvalitetsregister och hälsodataregistren. [Coverage 2019 – comparison in national registers of health and medical care] available at https://www.socialstyrelsen.se/globalassets/sharepoint-dokument/artikelkatalog/ovrigt/2019-12-6489-tabeller.xlsx (Accessed 6 September 2020).

        • Ludvigsson J.F.
        • Andersson E.
        • Ekbom A.
        • Feychting M.
        • Kim J.L.
        • Reuterwall C.
        • et al.
        External review and validation of the Swedish national inpatient register.
        BMC Public Health. 2011; 11: 450https://doi.org/10.1186/1471-2458-11-450
      4. Socialstyrelsen [The National Board of Health and Welfare]. Bortfall och kvalitet i patientregistret [Loss and quality of the swedish national patient register]. Available at: https://www.socialstyrelsen.se/statistik-och-data/register/alla-register/patientregistret/bortfall-och-kvalitet/ (Accessed 7 August 2020).

        • Wettermark B.
        • Hammar N.
        • Fored C.M.
        • Leimanis A.
        • Otterblad Olausson P.
        • Bergman I.
        • et al.
        The new Swedish Prescribed Drug Register-opportunities for pharmacoepidemiological research and experience from the first six months.
        Pharmacoepidemiol Drug Saf. 2007; 16: 726-735https://doi.org/10.1002/pds.1294
      5. Socialstyrelsen [The National Board of Health and Welfare]. Bortfall och kvalitet om läkemedelsregistret [Loss and quality of the swedish prescribed drug register]. Available at: https://www.socialstyrelsen.se/statistik-och-data/register/alla-register/lakemedelsregistret/bortfall-och-kvalitet/ (Accessed August 7, 2020).

        • von Elm E.
        • Altman D.G.
        • Egger M.
        • Pocock S.J.
        • Gøtzche P.C.
        • Vandenbroucke J.P.
        The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies.
        J Clin Epidemiol. 2008; 61: 344-349https://doi.org/10.1016/j.jclinepi.2007.11.008
        • Shrier I.
        • Platt R.W.
        Reducing bias through directed acyclic graphs.
        BMC Med Res Methodol. 2008; 8: 70https://doi.org/10.1186/1471-2288-8-70
        • Selby R.
        • Geerts W.H.
        • Kreder H.J.
        • Crowther M.A.
        • Kaus L.
        • Sealey F.
        d-KAF (Dalteparin in Knee-to-Ankle Fracture) Investigators. A double-blind, randomized controlled trial of the prevention of clinically important venous thromboembolism after isolated lower leg fractures.
        J Orthop Trauma. 2015; 29: 224-230https://doi.org/10.1097/BOT.0000000000000250
        • Socialstyrelsen [The National Board of Health and Welfare]
        Socialstyrelsens riktlinjer för vård av blodpropp/venös tromboembolism 2004 [Guidelines of treatment of venous thromoboembolism 2004].
        Socialstyrelsens riktlinjer för vård av blodpropp/venös tromboembolism 2004 [Guidelines of treatment of venous thromboembolism 2004]. Socialstyrelsen, Stockholm2004
        • Eck R.J.
        • Bult W.
        • Wetterslev J.
        • Gans R.O.B.
        • Meijer K.
        • Keus F.
        • van der Horst I.C.C.
        Intermediate dose low-molecular-weight heparin for thrombosis prophylaxis: systematic review with meta-analysis and trial sequential analysis.
        Semin Thromb Hemost. 2019; 45: 810-824https://doi.org/10.1055/s-0039-1696965
        • Ahmed I.
        • Majeed A.
        • Powell R.
        Heparin induced thrombocytopenia: diagnosis and management update.
        Postgrad Med J. 2007; 83: 575-582https://doi.org/10.1136/pgmj.2007.059188