Introduction

Physiotherapy is regarded as a standard therapy in intensive care units, and patients are regularly treated with active exercises or passive range of motion (PROM). A multicenter prospective observational study in five Australian tertiary hospitals has demonstrated that the incidence of adverse events during physiotherapy intervention is low, leading the authors to postulate in conclusion that physiotherapy in an intensive care unit (ICU) is safe [1]. Furthermore, patients of a multidisciplinary internal medicine ICU, who received physiotherapy showed a significant reduction in ventilatory dependency and had a significant reduction in their length of stay compared to a group who received standard nursing care [2]. Another study has shown that interruption of sedation in combination with physical and occupational therapies in the earliest days of critical illness results in better functional outcome at hospital discharge, a shorter duration of delirium, and more ventilator-free days compared with standard care [3]. However, there is uncertainty about the effect of physiotherapy on patients with acute cerebral diseases treated on ICU. Very early rehabilitation might lead to an improvement in the outcome of patients with acute neurological deficits, as could be shown for patients following stroke [4, 5]. On the other hand, increased intracerebral pressure (ICP) has been observed during several procedures, such as endotracheal suctioning, bronchoscopy or intramuscular injections [6–8]. In the majority of cases, ICP elevation was only transient, but in some patients, a critical long-term rise of ICP has been observed [6]. ICP increases significantly in all patients during nursing care activities involving moving the patient in the bed, such as turning the body, PROM, or rotation of the head, [9]. The fear of high peaks of ICP often results in an empirically “minimal touch therapy” in patients with severely elevated ICP. An attempt is made to avoid manipulation by nursing care activities or therapeutic procedures. Although physiotherapy is frequently applied in neuro-critical care units, little literature exists on this issue. So far, there have been only two studies, with limited numbers of patients, dealing with the effect of physiotherapy on ICP and cerebral perfusion pressure (CPP) [10, 11].

Methods

All patients in our ICU receive physiotherapy once a day. An observational study was performed on patients who received intracranial pressure measurements (Neurovent-P, Raumedic AG, Münchberg, Germany) and who were treated with PROM on our 12-bed neuro-ICU (tertiary hospital, Department of Neurology and Department of Neurosurgery, Klinikum Kassel, Germany). We prepared our publication according to the Strengthening the Reporting of Observational Studies in Epidemiology statement [12]. Between March 2010 and October 2011, a total number of 1,069 patients were treated in our ICU. 419 of them required mechanical ventilation. 88 patients were monitored with an intraparenchymal cerebral pressure measurement. 84 of them were enroled into this study. All these patients were intubated on a respirator. Four patients could not be included because of a temporary lack of personnel in the Department of Physiotherapy. ICP monitoring had been performed due to the necessity for monitoring patients with reduced consciousness. The indication for monitoring ICP was made by the senior neurosurgeon-in-charge, and this study had no influence on this decision.

Intracerebral pressure, CPP, mean arterial pressure (MAP), and heart rate (HR) were recorded continuously every minute, over 15 min before therapy, during therapy, and 15 min after treatment with PROM (Infinity® Delta, Lübeck, Dräger). Physiotherapy was performed by continuous PROM for a period of 26 min, according to the standard protocol for physiotherapy in our ICU. All joints of the upper and lower extremities were moved according to their physiological mobility over the greatest available range of motion. PROM also included hip rotation for 1 min. Patients were placed in a supine position with 30° head-up position before the start of the therapy. Ventilation was either spontaneous via tracheostoma, or with an assisted or a controlled mode. Depth of sedation was measured before each treatment using the Ramsay Scale (RS), and a subjective evaluation of the degree of sedation (low, medium, or deep) was made by the senior neuro-intensivist. The presence of external ventricular drainage might have an impact on ICP and was, therefore, also documented. The requirement of Osmo-therapy in the 24 h before treatment was also listed.

For the purpose of analysis, the data treatment units were divided into two groups: patients with a mean ICP <15 mmHg (Group 1) and patients with a mean ICP ≥15 mmHg before physiotherapy (Group 2). Data were analyzed using SPSS for Windows version 13.0 (SPSS Inc.). All values are expressed as mean ± standard deviation. Continuously measured variables included ICP, CPP, HR, and MAP for all three phases of the procedure (before, during, and after treatment). Differences between these phases were assessed using the Friedmann test and post-hoc paired Wilcoxon tests. A p value of p < 0.05 was considered significant. All p values were corrected for multiple testings in the number of tests using the Bonferroni method.

Results

The study consisted of 84 patients (f:m = 1:1) with a mean age of 55 ± 15 years. A total number of 298 treatment units were evaluated. The most frequent primary diagnoses were subarachnoidal hemorrhage (37 %), traumatic brain injury (26 %), and intracerebral hemorrhage (21 %). The baseline characteristics of the patients are presented in Table 1.

Table 1 Clinical characteristics of the patients
Full size table

Group 1 consisted of 224 (75.2 %) and Group 2 of 74 (24.8 %) different treatment units. Baseline parameters for these groups are shown in Table 2. 72 % of our patients received analgosedation during physiotherapy (69 % of Group 1 and 81 % of Group 2), and 98 % of our patients were ventilated mechanically. Only 2 % could breathe spontaneously via a tube. Overall, the mean baseline ICP was 11.5 ± 5.1 mmHg. ICP decreased significantly during therapy to 10.5 ± 5.3 mmHg (p = 2.0e–10) and then returned to a baseline level of 11.5 ± 5.5 mmHg, with no significant difference between mean ICP before and after treatment (p = 0.652). In Group 1, a baseline level of mean ICP of 9.4 ± 3.7 mmHg was observed, reducing significantly by 0.7 mmHg (p = 3.8e–6) during therapy. Patients with a mean ICP ≥15 mmHg before the procedure (Group 2) also showed a significant drop of ICP from a baseline level of 18.1 ± 2.7 mmHg to 16.1 ± 3.9 mmHg during therapy (p = 3.7e–6), a reduction of 2 mmHg. However, in contrast to Group 1, ICP reduction persisted and was significantly lower after the procedure (16.7 ± 4.5 mmHg) compared to the baseline level mmHg (p = 0.002) (Table 3; Fig. 1). Overall, there were only 15 treatment units (5 %) with an elevated ICP of over 20 mmHg before therapy.

Table 2 Baseline parameters of 289 treatment units
Full size table
Table 3 Wilcoxon test with Bonferroni correction (α = 0.05/3 = 0.016) for ICP, CPP, and HR
Full size table
Fig. 1
figure 1

Box plots showing mean values for ICP in mmHg 15 min before, during physiotherapy, and 15 min after physiotherapy with PROM. There was a significant decrease of ICP during therapy in both Group 1 and Group 2

Full size image

Mean CPP before therapy was 82.7 ± 14.2 mmHg (overall), 84.3 ± 14.1 mmHg (Group 1), and 77.8 ± 13.4 mmHg (Group 2). A post-hoc Wilcoxon test with Bonferroni correction showed a significant rise of 1.1 mmHg during therapy (p = 0.003) overall, and also in Group 1 (rise of 1.2 mmHg), but not in Group 2. No significant difference was observed between mean CPP before and after treatment overall or in Group 1 and Group 2 (Table 3; Fig. 2). Baseline level of MAP before treatment was 94.1 mmHg overall (Group 1 = 93.7 mmHg, Group 2 = 95.2 mmHg). There were no significant changes of MAP when comparing the time points before, during, and after treatment. HR before treatment was 79/min overall, 77/min in Group 1, and 83/min in Group 2. Mean HR decreased during and after therapy in all groups (Table 3).

Fig. 2
figure 2

Box plots showing mean values for CPP in mmHg before, during, and after physiotherapy with PROM. A significant increase in CPP during therapy was seen only in Group 1

Full size image

Overall, a significant decrease of ICP could be observed in 53.4 % (n = 159) of all treatment units during PROM, whereas 22.1 % (n = 66) did not show any changes. Significant rise of mean ICP could be observed in 24.5 % (n = 73) of all treatment units during PROM for a mean ICP elevation of 1.6 ± 1.5 mmHg. Comparing treatment unit with rise of ICP during therapy with treatment units of normal or decreased ICP with regard to gender, diagnosis, Ramsay score, analgosedation, amount of analgosedation, presence of external ventricular drainage, position of the patient before therapy, ventilation mode, the presence of Rotorest Kinetic Treatment Table, and Osmotherapy within 24 h before treatment revealed no significant differences between these groups. There were no therapy-related complications that led to termination of the treatment. Treatment for increased ICP was not required during or after any treatment units.

Discussion

With 84 patients and 289 treatment units, the present study is the largest observation dealing with physiotherapy and its effect on ICP and CPP. Critically ill patients in ICU are regularly treated with physiotherapeutic intervention. This is all the more important as physiotherapy is one of the pillars in the treatment of motor deficits. It is astonishing to note that research has paid so little attention to this topic [13]. So far there have been only a few other studies on changes of cerebral hemodynamic during physiotherapy [10, 11, 14]. In 1996, a publication with only 12 patients and 20 treatment units, including only six patients who needed mechanical ventilation, found no significant effects of PROM on MAP, ICP, HR or CPP [11]. Compared to the present study, the mean duration of PROM was much lower (26 vs 7 min). Unfortunately, patients with elevated ICP of more than 20 cm of water before physiotherapy were excluded in this study [11]. This group of patients is of paramount interest, and the question of whether physiotherapy is safe and feasible in patients with severe neurological diseases and elevated ICP remained unresolved in this study.

Another investigation involving 65 patients in a neurosurgical ICU postulated that physiotherapy can be used safely in patients with normal or increased ICP. However, a majority of the patients were awake and could participate in active exercises [10]. Only 12 patients showed an elevated ICP >15 mmHg before treatment with PROM. The length of stay in ICU was, on average, about 2 days and only 35 of 65 patients were mechanically ventilated, suggesting that the severity of illness of their patients was only moderate. In contrast to the previous studies [10, 11], 98 % our patients were ventilated mechanically. GCS on admission was low, Ramsay score was high and most of the patients had analgosedation due to the severity of their neurological disease. Therefore physiotherapy could only be done by PROM. A comparison of mean ICP before, during, and after PROM for all patients showed that it decreased significantly during therapy before returning to baseline levels after treatment. A significant reduction of mean ICP could also be observed in the group of patients with elevated ICP before physiotherapy (Group 2). Surprisingly, only this group showed a persistent reduction of ICP after physiotherapy over the investigation period of 15 min. However, the amount of ICP reduction was only mild (1 mmHg in Group 1, 2 mmHg in Group 2). Interestingly, the effect of ICP reduction during physiotherapy was also observed by Brimioulle et al. who found a decrease of mean ICP of 1 mmHg during treatment with PROM in supine position, with a simultaneous rise of CPP of 3 mmHg. Furthermore, patients with high ICP showed a decrease of 2 mmHg and an increase of CPP of 7 mmHg. This was also true for conscious patients with elevated ICP who could perform active exercises with upper and lower extremities [10]. However, we found a significant rise of CPP only in Group 1 and only during therapy. Mean MAP did not show any significant changes, and therefore had no impact on CPP. The pathophysiology of ICP reduction during physiotherapy remains unclear, and publications concerning changes of cerebral hemodynamics are rare in the literature. Increased cerebral blood flow velocities during passive and active exercises have been demonstrated in healthy individuals using transcranial Doppler ultrasound of the middle cerebral artery [15, 16]. It also remains unclear whether these results can be compared to patients with severe cerebral diseases and elevated ICP. It is possible that movements of the extremities might lead to an improved drainage of the cerebral venous system and, therefore, to a reduction in ICP. With regard to the 24.5 % of treatment units that showed a mild ICP elevation during therapy, we were unable to find any influencing factors, such as analgosedation, Ramsay score or external ventricular drainage. However, none of these cases required an ICP therapy.

The above results show that early treatment with PROM is safe and feasible. It does not have any substantial effect on ICP or CPP. This was true both for the patients with normal ICP and the patients with elevated ICP before treatment. These results are of paramount importance for neuro-ICU, as the necessary early physiotherapeutic treatment of patients with severe cerebral disease can be applied without fear of harming the patient. However, there are some limitations to this study. First of all, the group of patients with elevated ICP of more than 20 mmHg was small. Furthermore, ICP measurement does not mean a better outcome for patients. However, this aspect was not the aim of this observation. Further randomized studies are required to show the benefit of early physiotherapy with PROM on the clinical outcome parameters of patients treated on ICU. Some issues still remain unresolved. Obviously, we must take into account the time limitation of our treatment units. Might physiotherapy for more than 26 min have a distinct impact on stronger and long-lasting ICP reduction? Moreover, further studies on cerebral hemodynamic parameters such as cerebral blood flow or brain tissue oxygen concentration would be of interest to understand cerebral pathophysiological changes during physiotherapy. Moreover, further studies are needed to evaluate if there are differences between several diagnoses such as SAH, trauma or ICH. The problem with heterogeneous diagnosis could well be overcome with a multicenter study.