Results of the search

See Figure 1.

Figure 1

---

---

Study flow diagram.

There were three new included studies (Beyer‐Westendorf 2017; Boehler 2014; Spirkoska 2015).

Included studies

Three additional studies were included in this third update (Beyer‐Westendorf 2017; Boehler 2014; Spirkoska 2015), giving a total of 33 studies involving 7296 participants (Andreozzi 1996; Anonymous 1970; Archer 1977; Belcaro 1989; Belcaro 1990; Belcaro 1999; Belcaro 2011; Beyer‐Westendorf 2017; Boehler 2014; Cosmi 2012; Decousus 2010b; De Sanctis 2001; Ferrari 1992; Gorski 2005; Holzgreve 1989; Incandela 2001; Katzenschlager 2003; Koshkin 2001; Kuhlwein 1985; Lozano 2003; Marchiori 2002; Marshall 2001; Messa 1997; Nocker 1991; Nusser 1991; Pinto 1992; Rathbun 2012; Spirkoska 2015; Stenox Group 2003; Titon 1994; Uncu 2009; Vesalio Group 2005; Winter 1986). Nine studies reported data for 50 participants or fewer, 13 trials for 50 to 100 participants, and 11 studies for 100 participants or more.

Interventions and comparisons varied greatly among the studies. Nine trials included a topical treatment group (Belcaro 2011; De Sanctis 2001; Gorski 2005; Holzgreve 1989; Incandela 2001; Katzenschlager 2003; Nocker 1991; Pinto 1992; Winter 1986); three used a surgical treatment group (Belcaro 1989; Belcaro 1999; Lozano 2003); 14 used LMWH (Belcaro 1989; Belcaro 1990; Belcaro 1999; Cosmi 2012; Gorski 2005; Katzenschlager 2003; Lozano 2003; Marchiori 2002; Rathbun 2012; Spirkoska 2015; Stenox Group 2003; Titon 1994; Uncu 2009; Vesalio Group 2005); six used NSAIDs (Anonymous 1970; Ferrari 1992; Nusser 1991; Rathbun 2012; Stenox Group 2003; Titon 1994); two used fondaparinux (Beyer‐Westendorf 2017; Decousus 2010b); one used rivaroxaban (Beyer‐Westendorf 2017), and nine studies used another oral (Archer 1977; Belcaro 1989; Belcaro 1999; Koshkin 2001; Kuhlwein 1985; Messa 1997), intramuscular (Andreozzi 1996), intravenous (Marshall 2001), or non‐pharmacological (Boehler 2014) treatment.

Excluded studies

Two additional studies were excluded in this update (Ng 2010; Supe 2013) and one study which had previously been in the Studies awaiting classification section (Bijuan 2003) was also excluded making a total of 37 excluded studies. The reasons for exclusion are listed in the Characteristics of excluded studies table. Twenty‐one studies included a mixed population and it was not possible to extract data separately for ST (Allegra 1981; Annoni 1991; Argenteri 1983; Bagliani 1983; Becherucci 2000; Bergqvist 1990; Bracale 1996; Bruni 1979; Della Marchina 1989; Luttichau 1989; Mari 1982; Marsala 1985; Mauro 1992; Paciaroni 1982; Porters 1981; Pozza 1980; Seccia 1989; Seghezzi 1972; Seligman 1969; Stolle 1986; Tomamichel 1983). In one study it was not possible to extract outcome data separately for the two study treatment groups (Agus 1993). Four studies included people without a diagnosis of ST of the legs (Bernicot 1980; Gandhi 1984; Resta 1967; van Cauwenberge 1972), and two studies included people with DVT (Di Perri 1986; Rea 1981). In one study, it was unclear whether the study was randomised or not (Giorgetti 1990). Six studies included people with ST of the arm (Gouping 2003; Mehta 1975; Ng 2010; Rozsos 1994; Supe 2013; van der Knaap 1988), and in one study, the evaluated outcomes were not among those considered in the present review (Ibanez‐Bermudez 1996). For one study, we were unable to retrieve sufficient information to judge eligibility fully (Bijuan 2003).

Allocation

Thirteen studies adequately generated the randomisation sequence (Belcaro 2011; Beyer‐Westendorf 2017; Cosmi 2012; Decousus 2010b; Gorski 2005; Katzenschlager 2003; Marchiori 2002; Marshall 2001; Messa 1997; Rathbun 2012; Spirkoska 2015; Vesalio Group 2005; Winter 1986), one was not adequate (Uncu 2009), and the remaining 18 studies were unclear. Eighteen studies adequately concealed allocation (Belcaro 2011; Beyer‐Westendorf 2017; Cosmi 2012; Decousus 2010b; Marshall 2001; Rathbun 2012; Spirkoska 2015; Stenox Group 2003), and the remaining 25 studies were unclear.

Blinding

Ten studies had a double‐blinded design (Cosmi 2012; Decousus 2010b; De Sanctis 2001; Incandela 2001; Marshall 2001; Nusser 1991; Pinto 1992; Rathbun 2012; Stenox Group 2003; Vesalio Group 2005), and in nine studies it was unclear whether blinding was attempted (Anonymous 1970; Archer 1977; Ferrari 1992; Koshkin 2001; Kuhlwein 1985; Marchiori 2002; Nocker 1991; Spirkoska 2015; Winter 1986). The remaining 14 studies did not attempt to blind the assessment of the outcomes or did not report whether blinding was used.

Incomplete outcome data

Seven studies performed the analysis according to the ITT principle (Beyer‐Westendorf 2017; Decousus 2010b; De Sanctis 2001; Kuhlwein 1985; Marchiori 2002; Messa 1997; Nocker 1991; Vesalio Group 2005); in nine this was unclear, while in the remaining studies the percentage of participants randomised and subsequently excluded from the analysis ranged from 2% to 33% (Anonymous 1970; Archer 1977; Belcaro 1989; Belcaro 1999; Belcaro 2011; Boehler 2014; Cosmi 2012; Ferrari 1992; Gorski 2005; Holzgreve 1989; Katzenschlager 2003; Lozano 2003; Marshall 2001; Spirkoska 2015; Stenox Group 2003; Titon 1994). In Beyer‐Westendorf 2017 the primary analysis was originally planned as ITT analysis, but later modified into a per‐protocol analysis. Although 7.6% (36/471) participants randomised were excluded from the primary analysis authors report results also according to the ITT principle.

Selective reporting

Most studies were free of selective reporting except Anonymous 1970; Belcaro 2011; Lozano 2003; Spirkoska 2015; and Uncu 2009 (all high risk), which did not provide data on some of the specified outcomes and Winter 1986 (unclear risk), which was reported as an abstract only and did not prespecify outcomes.

Fondaparinux

The CALISTO study, a large double‐blinded, placebo‐controlled RCT, evaluated a prophylactic dose (2.5 mg subcutaneously (sc) once daily) of fondaparinux given for 45 days (Decousus 2010b). The primary efficacy outcome of this RCT (i.e. composite of death from any cause, symptomatic PE, symptomatic DVT, or symptomatic extension to the saphenofemoral junction or symptomatic recurrence of ST up to day 47) was reduced by 85% by fondaparinux (RR 0.15, 95% CI 0.08 to 0.26) with a number needed to treat for an additional beneficial outcome (NNTB) of 20. The incidence of each component of the primary efficacy outcome was significantly reduced in the fondaparinux group compared with the placebo group except for the incidence of PE and of death, which did not differ significantly between the two groups (Analysis 1.1; Analysis 1.6). The risk of the composite of symptomatic DVT or PE was reduced by 85% with fondaparinux compared with placebo (RR 0.15, 95% CI 0.04 to 0.50) with a NNTB of 88. Fondaparinux was associated with lower rates of extension (RR 0.08, 95% CI 0.03 to 0.22) and recurrence of ST (RR 0.21, 95% CI 0.08 to 0.54). By day 47, major bleeding had occurred in one participant (0.1%) in each group (RR 0.99, 95% CI 0.06 to 15.86; P = 1.00). The rate of clinically relevant non‐major, minor, and total bleeding; arterial thromboembolic complications; and adverse effects of treatment did not differ significantly between the two groups.

In the SURPRISE study, 472 people with ST and one or more risk factors for thromboembolic complications (older than 65 years, male sex, previous ST or DVT or PE, active cancer or history of cancer, autoimmune disease, or involvement of non‐varicose veins) were randomised to fondaparinux (2.5 mg sc once daily) or the oral direct factor Xa inhibitor rivaroxaban (10 mg once daily) (Beyer‐Westendorf 2017). In the per‐protocol analysis, the incidence of the primary efficacy outcome (i.e. composite of symptomatic DVT or PE, progression or recurrence of ST, and all‐cause mortality at 45 days) was comparable in the rivaroxaban and fondaparinux groups at day 45 (3% with rivaroxaban versus 2% with fondaparinux; HR 1.9, 95% CI 0.6 to 6.4) and at 90 days, (7% with rivaroxaban versus 7% with fondaparinux; HR 1.1, 95% CI 0.5 to 2.2).

There were similar results when the analysis was performed according to the ITT principle. Fondaparinux was associated with a non‐statistically significant reduction of symptomatic VTE, DVT, recurrence of ST, mortality, clinically relevant non‐major bleeding, serious adverse events, or adverse effects of treatment compared with rivaroxaban (Analysis 2.2; Analysis 2.3; Analysis 2.5; Analysis 2.6; Analysis 2.8; Analysis 2.10; Analysis 2.11). There were no cases of PE, extension of ST or major bleeding in either treatment arm.

Low molecular weight heparin and unfractionated heparin

Fourteen studies included a LMWH group (Belcaro 1989; Belcaro 1990; Belcaro 1999; Cosmi 2012; Gorski 2005; Katzenschlager 2003; Lozano 2003; Marchiori 2002; Rathbun 2012; Spirkoska 2015; Stenox Group 2003; Titon 1994; Uncu 2009; Vesalio Group 2005).

Although not statistically significant, the incidence of VTE tended to be lower with both prophylactic and therapeutic LMWH compared with placebo shortly after treatment (prophylactic: RR 0.25, 95% CI 0.03 to 2.24; therapeutic: RR 0.26, 95% CI 0.03 to 2.33). However, at the end of the three‐month follow‐up, this difference was even less evident suggesting a catch‐up phenomenon (Analysis 3.2; Analysis 4.2) (Stenox Group 2003). Prophylactic and therapeutic LMWH given for eight to 12 days significantly reduced ST extension or recurrence, or both, compared with placebo (prophylactic: RR 0.44, 95% CI 0.26 to 0.74; therapeutic: RR 0.46, 95% CI 0.27 to 0.77). There were no episodes of major bleeding or heparin‐induced thrombocytopenia (HIT) in any treatment group (Stenox Group 2003).

Combined therapy with LMWH plus elastic compression stockings seemed to reduce the incidence of VTE and ST extension or recurrence, or both, compared with elastic stockings alone (VTE: RR 0.08, 95% CI 0.00 to 1.38; ST extension or recurrence, or both: RR 0.08, 95% CI 0.01 to 0.59), although the former difference was not statistically significant (Belcaro 1999). This study did not provide data on safety outcomes.

Two studies randomised participants to topical treatment with heparin spray gel or LMWH (Gorski 2005; Katzenschlager 2003). There was a non‐significant decrease in DVT with LMWH (RR 0.30, 95% CI 0.03 to 2.70) and relief of local symptoms of ST was similar between both treatments at 21 days (Gorski 2005).

One study evaluated LMWH versus surgical treatment (saphenofemoral disconnection) (Lozano 2003). There was a comparable reduction of VTE events and a similar safety profile in the two study groups. There were numerically more cases of ST extension or recurrence with LMWH, although numbers were low and differences were not statistically significant (RR 3.00, 95% CI 0.33 to 27.23).

Three studies evaluated LMWH versus NSAIDs (Rathbun 2012; Stenox Group 2003; Titon 1994). Compared to NSAIDs, both fixed‐dose LMWH and weight‐adjusted LMWH seemed to have a similar effect on VTE (RR 0.93, 95% CI 0.24 to 3.63) and ST recurrence (RR 1.01, 95% CI 0.58 to 1.78). In the study by Rathbun 2012, there was one case of ST progression into the posterior tibial veins and one symptomatic PE, both in the LMWH group. Rathbun 2012 was not pooled with the other two studies for the outcome VTE since the administration of therapeutic LMWH in any person with thrombus progression during follow‐up could have introduced significant confounding. In Stenox Group 2003, which used placebo as a control group, an indirect comparison between prophylactic LMWH and NSAIDs suggested a non‐statistically significant reduction in VTE at the end of treatment (RR 0.45, 95% CI 0.04 to 4.89). There were no major bleeding events or HIT in either group. Rathbun 2012 reported two episodes of cutaneous rash with LMWH. In addition, there was a significant reduction in pain with both LMWH and NSAIDs with no differences between the groups.

One study compared LMWH alone versus LMWH combined with the anti‐inflammatory agent acemetacin (Uncu 2009). There were no cases of VTE, extension of ST, or major bleeding in either group. The effects on signs and symptoms of ST were not statistically different (Analysis 14.4; Analysis 14.5; Analysis 14.6; Analysis 14.7).

Three studies compared different regimens of LMWH head‐to‐head but without using a placebo or inactive control group (Cosmi 2012; Spirkoska 2015; Vesalio Group 2005). In Cosmi 2012 (the STEFLUX study), the incidence of symptomatic VTE at the end of treatment and at the three‐month follow‐up was not different in the 30‐day intermediate‐dose LMWH and 30‐day prophylactic dose LMWH groups (Analysis 5.1; Analysis 5.2), and it was lower in the 30‐day intermediate‐dose LMWH compared with the 10‐day intermediate‐dose LMWH (VTE end‐of‐treatment: 0.46% with 30‐day intermediate‐dose LMWH versus 4.72% with 10‐day intermediate‐dose LMWH, RR 0.10, 95% CI 0.01 to 0.75; VTE 3‐month follow‐up: 1.82% with 30‐day intermediate‐dose LMWH versus 5.19% with 10‐day intermediate‐dose LMWH, RR 0.35, 95% CI 0.11 to 1.09; Analysis 7.1; Analysis 7.2). At the three‐month follow‐up, symptomatic PE had occurred in none of the participants of the 30‐day intermediate‐dose group and in one participant of both the 10‐day intermediate‐dose and 30‐day prophylactic‐dose LMWH groups. The incidence of symptomatic DVT did not differ between the three groups (Analysis 5.4; Analysis 6.4; Analysis 7.4). ST extension at three months was significantly reduced by the 30‐day intermediate‐dose LMWH in comparison to both the 30‐day prophylactic (2.28% with 30‐day intermediate‐dose versus 8.29% with 30‐day prophylactic dose LMWH; RR 0.28, 95% CI 0.10 to 0.73) and 10‐day intermediate dose LMWH (10.38%, RR 0.22, 95% CI 0.08 to 0.57), while recurrence of ST was similar in the 30‐day intermediate‐dose LMWH and the other two groups (Analysis 5.6; Analysis 7.6). There were no cases of major bleeding or HIT. The intensity of local symptoms evaluated by visual analogue scales (VAS) was comparable in the 30‐day intermediate‐dose, 10‐day intermediate‐dose, and the 30‐day prophylactic‐dose LMWH at the start of treatment (5.0 with 30‐day intermediate‐dose, 5.1 with 10‐day intermediate‐dose, 5.1 with 30‐day prophylactic‐dose; P = 0.97) as well as at the end of treatment (0.7 with 30‐day intermediate‐dose, 0.6 with 10‐day intermediate‐dose, 0.8 with 30‐day prophylactic‐dose; P = 0.10) and at three months (0.2 with 30‐day intermediate‐dose, 0.3 with 10‐day intermediate‐dose, 0.4 with 30‐day prophylactic‐dose; P = 0.47). Allergic reactions occurred in 0.4% with 30‐day intermediate‐dose, 1.4% with 10‐day intermediate‐dose, and 0% with 30‐day prophylactic‐dose.

In Spirkoska 2015, one participant in the intermediate‐dose LMWH group developed a symptomatic PE versus none in the prophylactic‐dose LMWH group. Asymptomatic ST progression into DVT occurred in one participant in each group. There were no major bleeding events. The authors reported a regression of the thrombus at the end of the study period in 66% of participants receiving the prophylactic‐dose LMWH compared to 80% in the intermediate‐dose LMWH and a complete thrombus resolution with recanalisation in three participants (9.7%) in the prophylactic‐dose LMWH and six (19.4%) participants in the intermediate‐dose LMWH. These differences were not statistically significant.

In the Vesalio Group 2005, one month of weight‐adjusted full therapeutic dose of LMWH or fixed prophylactic‐dose LMWH led to a similar reduction in ST extension or recurrence, or VTE (RR 1.20, 95% CI 0.42 to 3.40) over a three‐month follow‐up. In the prophylactic‐dose LMWH group most of the VTE events (77%) occurred while participants were still on treatment, whereas only 33% of participants on therapeutic‐dose LMWH developed VTE during LMWH administration. This advantage was lost after drug discontinuation with no difference at the end of the study period. There was no major bleeding or HIT during the study. Local symptoms and signs regressed faster with therapeutic dose LMWH although the difference was not statistically significant.

Two studies used sc UFH at prophylactic doses as the comparator treatment (Belcaro 1999; Marchiori 2002). Relative to elastic stockings alone, prophylactic sc UFH plus elastic stockings was associated with a statistically non‐significant lower VTE rate (RR 0.08, 95% CI 0.00 to 1.47) and a 83% reduction in ST extension or recurrence (RR 0.17, 95% CI 0.04 to 0.72) (Belcaro 1999). One study compared high‐ versus low‐dose sc UFH. There was a non‐significant 83% reduction in VTE (RR 0.17, 95% CI 0.02 to 1.30) and a 27% (RR 0.73, 95% CI 0.34 to 1.55) reduced rate of ST extension or recurrence among participants treated with high‐dose UFH (Marchiori 2002). There were no episodes of major bleeding or HIT in either study group.

Two studies evaluated sc heparin calcium (Belcaro 1989; Belcaro 1990). The combination of elastic stockings plus heparin calcium did not significantly improve local symptoms and signs compared with elastic stockings alone. Treatment with heparin calcium was correlated with a faster reduction of the analogue score and the area at maximum temperature than with defibrotide, although the difference was not significant. There were no adverse effects.

Non‐steroidal anti‐inflammatory drugs

Six studies included an NSAID group (Anonymous 1970; Ferrari 1992; Nusser 1991; Rathbun 2012; Stenox Group 2003; Titon 1994). Of these, two compared NSAIDs with placebo (Anonymous 1970; Stenox Group 2003), three NSAID with LMWH (Rathbun 2012; Stenox Group 2003; Titon 1994), and two randomised participants to two different NSAIDs (Ferrari 1992; Nusser 1991). The trials comparing NSAIDs versus LMWH have been discussed previously (Rathbun 2012; Stenox Group 2003; Titon 1994).

NSAIDs significantly reduced the risk of ST extension or recurrence, or both, by 54% compared with placebo (RR 0.46, 95% CI 0.27 to 0.78) (Stenox Group 2003). However, there were no differences in the incidence of VTE or in the resolution of local symptoms and signs. While there were no major bleeding episodes recorded in any of the NSAID or placebo groups, indomethacin tended to increase the rate of adverse effects compared with placebo (RR 2.60, 95% CI 0.95 to 7.08) (Anonymous 1970).

In one study, oral acemetacin led to a better resolution of the local clinical picture than diclofenac (Nusser 1991). Another trial compared nimesulide with diclofenac sodium (Ferrari 1992). Local symptoms were similarly improved by both treatments. In the group of participants randomised to nimesulide, there was a lower incidence of gastric pain episodes (RR 0.25, 95% CI 0.03 to 2.08) although this difference was not statistically significant (Ferrari 1992).

Topical treatment

Nine studies included a topical treatment group (Belcaro 2011; De Sanctis 2001; Gorski 2005; Holzgreve 1989; Incandela 2001; Katzenschlager 2003; Nocker 1991; Pinto 1992; Winter 1986). The comparison of heparin spray gel versus LMWH has been discussed earlier (see Low molecular weight heparin and unfractionated heparin section; Gorski 2005; Katzenschlager 2003).

Belcaro 2011 randomised participants to three doses of heparin spray gel versus placebo for seven to 14 days. After one week, there was a significant reduction in pain assessed by the VAS with heparin spray gel (‐93.13% reduction with heparin spray gel versus ‐61.35% reduction with placebo; P < 0.0001), a lower erythema extension (‐92% with heparin spray gel versus ‐26% with placebo; P < 0.012), and thrombus length (‐40.81 with heparin spray gel versus ‐4.22 with placebo; P < 0.0001). There were no adverse events or drug‐related reactions reported.

One study randomised participants to receive topical methylthioadenosine or placebo (Pinto 1992). Methylthioadenosine was associated with a non‐significant reduction in local signs and symptoms relative to placebo.

A significant improvement in the local symptomatology was observed with diclofenac gel (Nocker 1991) and essaven gel (De Sanctis 2001; Incandela 2001) compared with placebo.

Holzgreve 1989 and Winter 1986 compared two different types of gel. Holzgreve 1989 evaluated diclofenac gel versus etofenak gel and showed a comparable efficacy profile of the two topical medications. Winter 1986 compared diclofenac gel and heparin gel and found a better efficacy with diclofenac gel.

None of the studies evaluating a topical treatment reported data on VTE or ST recurrence.

Surgery

Three studies included a surgical treatment (Belcaro 1989; Belcaro 1999; Lozano 2003). One study compared surgery (saphenofemoral disconnection) with LMWH (see Low molecular weight heparin and unfractionated heparin section; Lozano 2003). The remaining two studies compared surgery combined with elastic stockings with elastic stockings alone (Belcaro 1989; Belcaro 1999).

One trial found that thrombectomy plus elastic stockings with or without venoruton led to an improvement of the local clinical signs and a greater reduction in the number of veins with ST compared with elastic compression bandages alone (Belcaro 1989). There were no cases of DVT in either group. One trial found that ligation of the vein plus elastic stockings was associated with a non‐significant reduction in VTE events (RR 0.33, 95% CI 0.07 to 1.60) and ST recurrence and extension (RR 0.46, 95% CI 0.18 to 1.15) relative to the control treatment (Belcaro 1999).

Compared with elastic stockings alone, venous stripping plus elastic stockings decreased the risk of ST extension and recurrence (RR 0.09, 95% CI 0.01 to 0.64) and seemed to be associated with a lower, non‐significant, incidence of VTE (RR 0.37, 95% CI 0.08 to 1.78) (Belcaro 1999).

Other

Nine studies evaluated an oral (Archer 1977; Belcaro 1989; Belcaro 1999; Koshkin 2001; Kuhlwein 1985; Messa 1997), intramuscular (Andreozzi 1996), intravenous (Marshall 2001), or non‐pharmacological (Boehler 2014) treatment. Beyer‐Westendorf 2017 compared oral rivaroxaban with fondaparinux and has been presented above (see Fondaparinux section).

Compared with placebo, oral vasotonin was associated with a higher proportion of participants who were cured or improved (Kuhlwein 1985). The criteria to determine the response to study treatment were not described. Vasotonin seemed to be well tolerated, with one case of poor tolerability among participants treated with vasotonin (3%) versus five cases (13%) in the placebo arm (RR 0.20, 95% CI 0.02 to 1.63).

The combination of venoruton, thrombectomy, and elastic stockings versus elastic stockings alone has been discussed above (see Low molecular weight heparin and unfractionated heparin section; Belcaro 1989). In the same trial, venoruton combined with elastic stockings led to an improvement of local symptoms compared with elastic stockings alone.

One study evaluating oral heparansulphate versus oral sulodexide suggested a greater decrease in local pain, itching, and redness in participants receiving oral heparansulphate than in the group receiving sulodexide (Messa 1997).

Compared with placebo, oxyphenbutazone reduced local tenderness four‐fold and halved the intensity of pain and erythema (Archer 1977).

One study evaluated oral vitamin K antagonists in combination with elastic stockings, which suggested a non‐significant reduction in VTE events (RR 0.08, 95% CI 0.00 to 1.47) and ST extension or recurrence (RR 0.42, 95% CI 0.16 to 1.13) with vitamin K antagonists plus elastic stockings compared with elastic stockings alone (Belcaro 1999).

Two studies addressed the use of enzyme therapy versus placebo (Koshkin 2001; Marshall 2001). Enzyme treatment seemed to improve local symptoms although the criteria to evaluate the response to study treatment were not reported.

One trial assessed the efficacy of three doses of desmin (Andreozzi 1996). There was a better control of local symptoms with higher doses of desmin without any increase in the risk of adverse events.

One study evaluated LMWH plus compression stockings versus LMWH alone (Boehler 2014). At the three‐week follow‐up, there was no difference between the two groups with regard to pain intensity (mean difference ‐0.15 cm, 95% CI ‐0.95 to 0.65), skin erythema (‐6.54 cm², 95% CI ‐17.94 to 4.86), or quality of life evaluated through the 36‐item Short Form (SF‐36) Physical score (mean difference 2.92, 95% CI ‐1.04 to 6.88) and SF‐36 Mental score (‐2.98, 95% CI ‐7.30 to 1.34). There was no DVT or HIT.