Association of early mobility with the incidence of deep-vein thrombosis and mortality among critically ill patients: a post hoc analysis of PREVENT trial

Immobility in patients admitted in the intensive care unit (ICU) is common due to the nature of critical illness and the interventions provided for its management, including sedatives, narcotics, and/or paralytic agents. Immobility increases the risk of multiple adversities that include muscle atrophy and weakness [1], insulin resistance, pressure injury [2], pneumonia [3], and venous thromboembolism (VTE) [4]. These adversities may be induced by multiple mechanisms such as reduced muscle loading and enhanced overproduction of reactive oxygen and nitrogen species [5], and may persist for several months and years after ICU discharge [6, 7]. Early mobility has been recommended for ICU patients [8], due to improvement in various outcomes [9‐16]. A systematic review of 8 randomized controlled trials and 10 observational studies that evaluated the association between mobility with or without concomitant thromboprophylaxis and VTE in adult hospitalized patients observed variable mobility interventions across studies, and found mixed results on the association of mobility and VTE [17]. The heterogeneity of studies and the variable definition of mobility made it impossible to quantify any therapeutic ambulation dose [17]. The review concluded that mobility alone should not be considered as an adequate modality for thromboprophylaxis [17]. Another systematic review and meta-analysis of 39 randomized controlled trials, most had significant risks of bias, found that early mobility in critically ill patients was associated with lower incidence of deep-vein thrombosis (7 studies, 730 patients, risk ratio 0.16; 95% confidence interval (CI) 0.06, 0.47; I² 0%), but not with mortality [18].

The role of early mobility for VTE prevention is not well studied. This study aimed to assess the association of early mobility with the incidence of proximal lower-limb deep-vein thrombosis and 90-day mortality of critically ill patients.

The mobility practices of the participating centers during the trial periods are presented in Additional file 1: Table S2. Most centers had a protocol for early mobility during the trial with a physiotherapist providing treatments to a median of 10 patients on a median of 5 days per week.

In this cohort of critically ill patients who were expected to stay in the ICU for > 72 h, a minority of patients received high early mobility levels in the first 3 days. Additionally, mobility remained limited throughout the ICU stay. Higher mobility was associated with lower mortality rates, and lower duration of organ support, but with similar incident proximal lower-limb deep-vein thrombosis and pulmonary embolism.

Critical illness is associated with immobility, often confounded by pre-ICU underlying limitation to mobility, cardiopulmonary instability and ICU therapies such as sedation and neuromuscular blockers. In the current study, we found that mobility more than range of motion was provided to only 441/1708 (25.8%) patients during the first 3 days of ICU stay. This finding is not uncommonly observed in ICUs [21, 22]. A point prevalence study of patients with acute respiratory failure in 42 ICUs at 17 US hospitals found that mobility (range of motion and above) from any healthcare provider was provided on 501/770 (65%) patient-days and by physical or occupational therapists on 247/770 (32%) patient-days [21]. Mobility was performed less for mechanically ventilated patients (48% versus 26% for other patients, p < 0.001) [21].

Several studies have shown that mobility in critically ill patients was associated with better clinical outcomes, including lower rates of deep-vein thrombosis, ventilator-associated pneumonia, and pressure injury, increased muscle strength and shorter duration of mechanical ventilation, and ICU and hospital stay [18, 23, 24]. In a pre-post quasi-experimental study in trauma patients (1,044 patients in the pre- and 1,132 patients in the post-intervention cohort), an early mobility program was associated with lower rate of deep-vein thrombosis (6.7% versus 10.9%, p < 0.01; adjusted relative risk 0.67, 95% CI 0.50, 0.90) [25]. Another pre-post quasi-experimental study evaluated a structured progressive mobility protocol in trauma and burns patients (184 patients in the pre- and 159 patients in post-intervention cohort) observed lower rate VTE in postintervention cohort (7.5% versus 21%, p = 0.0004) [14]. A systemic review and meta-analysis also showed that mobility reduced the risk of deep-vein thrombosis (risk ratio 0.16; 95% CI 0.06, 0.47) [18]. This was based on 7 randomized controlled trials, mostly with small sample sizes (total of 730 patients) and significant biases [18]. Additionally, the studies used different definitions of mobility [17, 18]. In the current study, we found that higher early mobility levels were associated with similar incident lower-limb deep-vein thrombosis and pulmonary embolism compared with lower mobility level (range of motion) in patients expected to stay > 72 h in the ICU. As all study patients were on pharmacologic prophylaxis by study design, it is possible that early mobility has little impact on VTE prevention in such a setting, especially that mobilization was probably relatively short in duration for each session relative to the rest of the time during the day spent in the ICU. We also found that higher mobility levels were associated with lower mortality and less organ support. A meta-analysis of randomized and controlled trials found that mobility had no impact on short- and long-term mortality in ICU patients but led to more days alive and out of hospital to day 180 [24]. More recently, The Treatment of Mechanically Ventilated Adults with Early Activity and Mobilization (TEAM) trial randomized 750 adult patients on mechanical ventilation and found that an increase in early active mobilization (sedation minimization and daily physiotherapy) did not result in a significantly greater number of days alive and out of hospital compared with the usual level of mobilization in the ICU [26]. In this trial, both groups received active physiotherapy; the number of days per patient when physiotherapy assessment occurred was 0.94 ± 0.11 in the early active mobilization group and 0.81 ± 0.24 in the usual care group [26]. The limited separation in treatment exposure between the two groups and the active mobility program in the control arm, which do not reflect mobility practices in most ICUs, keep this question open [21, 27].

The strengths of the study include that data came from a multicenter randomized controlled trial wherein twice-weekly structured surveillance for lower-limb deep-vein thrombosis was performed in a large number of patients, thus reducing ascertainment bias. Our study characterized important associations between higher versus lower early mobility levels and important clinical outcomes besides mortality. VTE is rarely studied as an outcome of mobility. Our review of clinicaltrials.gov database showed that none of the 10 ongoing randomized controlled trials that were evaluating mobility in critically ill patients had VTE as a pre-specified outcome (Additional file 1: Table S5). Our study also has limitations. First, we did not have detailed description of mobility, such as duration, intensity, whether being active or passive, and contraindications, including the practices of sedation and neuromuscular blockade. Second, we could not analyze patients according to individual mobility level, given the low number of patients who had higher mobility levels. Third, although our findings indicated that early mobility was associated with lower mortality in ICU patients, we cannot establish a causal relationship between mobility and mortality. The observed association between early mobility and mortality and the lack of association between early mobility and lower-limb deep-venous thrombosis might be related to differences across patient groups in baseline characteristics including VTE risk factors and care processes, and other unmeasured confounders that were not accounted for in our multivariable analysis models. It is likely that the illness severity of the patients who had a higher mortality rate led them to being less mobile.

Only a small proportion of critically ill patients had a high level of mobility early in their ICU stay. There was no association between early mobility levels and lower-limb deep-vein thrombosis. Although early mobility level was independently associated with lower mortality risk, this association does not establish causality and randomized controlled trials are required to assess whether and to what extent this association is modifiable.