Acute microcirculatory effects of medium frequency versus high frequency neuromuscular electrical stimulation in critically ill patients - a pilot study

Patient selection

Two distinct groups of patients were enrolled in the study. Patients eligible for the first group (referred from this point on as the ‘acute group’) had to meet the diagnostic criteria for the systemic inflammatory response syndrome (SIRS) or sepsis of any severity [17] for a minimum of three and a maximum of five days at the day of the session. These time constraints were selected because they represent the minimum time period in which ICUAW can develop [18]. Exclusion criteria were age under 18 years, pregnancy, pre-existing neuromuscular disease, conditions that limit lower limb mobility (for example, burns, fractures), pacemakers, and presence of edema or subcutaneous fat that interfered with the application of NMES or with the assessment of the microcirculation.

The second group consisted of patients who had developed clinically manifest ICUAW. These will subsequently be referred to as the ‘ICUAW group’. The diagnosis of ICUAW was made on clinical grounds after the patients had been awaken from sedation and their orientation and cooperation had been ascertained, following a protocol described previously [11, 19].

Study design

This was a prospective, randomized pilot study. The study protocol was approved by the Scientific Committee of Evangelismos Hospital. Written informed consent to participate in the study was provided by all patients or their next of kin.

In all patients, demographic and clinical data were recorded, as were medications, sedative agents and ventilation mode used on the day of the NMES session. Vital signs, arterial oxygenation, central venous oxygen saturation and lactate levels were obtained before and after the NMES session. Creatine kinase (CK) levels were monitored for four days following the application of NMES.

NMES application and evaluation of muscle contractions

NMES was applied to both lower limbs of the patients. An electronic stimulator (Cefar Rehab Pro, Cefar Medical AB, Malmö, Sweden) was used for the application. After appropriate skin preparation, rectangular adhesive electrodes (9 × 5 cm) were placed over the motor points of the vastus lateralis, vastus medialis and peroneus longus muscles.

Patients were randomly assigned to one of two NMES protocols. The HF (high frequency) protocol consisted of symmetric, biphasic, trapezoid pulses at 75 Hz, with pulse duration of 400 μs, duty cycle of 5 seconds on and 21 seconds off, ramp-up time of 1.5 seconds and ramp-down time of 0.8 seconds. The MF (medium frequency) protocol consisted of symmetric, biphasic, trapezoid pulses at 45 Hz, with pulse duration of 400 μs, duty cycle of 5 seconds on and 12 seconds off, ramp-up time of 1.5 seconds and ramp-down time of 0.8 seconds. The duration of each session was 30 minutes for both protocols, with an additional 5-minute warm-up and a 5-minute recovery phase using 10 Hz current of 400 μs pulse duration.

The specific waveforms were selected because they have been shown to maximize force and delay fatigue, compared with other forms such as the sinusoidal [20, 21]. Frequencies and duty cycles were combined based on results from a study which specifically investigated optimal combinations of NMES settings [22].

Current intensity was continually adjusted during each session. It has been known that when a muscle group is stimulated with constant intensity, the generated torque gradually decreases [22, 23]. In a set of preliminary trials, we verified that the need to increase the administered intensities is also true for the two NMES protocols under investigation (data not shown). As a result, in alert patients, current intensity was initially set to the maximal tolerated level, and was increased by 10% (or less, if the patient felt discomfort) every three minutes. In sedated patients, after the warm-up, the intensity that produced a maximal response was determined, and the 60% of that value was used as the starting point. In this case also current intensity was increased by 10% every three minutes.

Muscle contractions of the quadriceps muscles were graded by one of the researchers, who had no knowledge of which protocol was being used. A score was given according to the following scale: 0 = no contraction, 1 = palpable contraction, 2 = visible contraction, 3 = slight knee extension against gravity, 4 = full knee extension [24].

Near-infrared spectroscopy and the vascular occlusion technique

NIRS is used to estimate the concentration of specified chromophores in a given tissue volume. In clinical settings, the parameter of interest is the degree of hemoglobin oxygenation, termed tissue oxygen saturation (StO₂). This quantity is evaluated at the microcirculatory compartment, according to principles described previously [25, 26].

We used a continuous-wave spectrometer (Hutchinson InSpectra Model 650, Hutchinson Technology, Hutchinson, MN, USA). This device emits light in four wavelengths and uses a wide gap second-derivative method to estimate StO₂ with a measurement depth of 15 mm [27].

NIRS measurements were continuously recorded at two distinct sites. A probe was placed over the thenar eminence in order to capture the effects of NMES on the systemic microcirculation. In addition, the vascular occlusion technique (VOT) was applied in order to reveal features of the microcirculation that the simple monitoring cannot provide. The VOT consists of applying a standardized ischemic stimulus to the upper extremity by inflating a pneumatic cuff to 50 mmHg above the systolic pressure and maintaining it for three minutes. After that the cuff is released and perfusion is reestablished.

During the ischemic phase, the StO₂ signal decreases. Its slope, expressed in % units per minute (downslope, %/minute), has been shown to be a proxy for local muscle O₂ consumption rate [28]. During the reperfusion phase, the StO₂ signal rapidly increases, usually to a value higher than the initial one, and then gradually returns to baseline. The initial increase of the signal, expressed in % units per second (upslope, %/second), is an index of endothelial reactivity, while the duration of the hyperemic phase reflects vascular reserve [29].

A second NIRS probe was placed over the vastus lateralis muscle, in a randomized manner with regards to side. The parameters that were evaluated were the StO₂ desaturation slope (ΔStO_2slope), which was calculated as the average of the slopes of the StO₂ signal in each contraction and which serves as an index of muscle O₂ consumption [30, 31], the StO₂ minimum amplitude (ΔStO_2min), which is the difference between the minimum StO₂ value during a contraction and baseline StO₂ and is an index of O₂ demand relative to O₂ supply [31, 32], and the StO₂ maximum amplitude (ΔStO_2max), which is the difference between the maximum StO₂ value during the recovery period and baseline StO₂, with higher values indicating increased O₂ supply relative to O₂ demand [32].

All NIRS measurements were stored and analyzed off-line using the proprietary software Hutchinson InSpectra Analysis v. 2.4 (Hutchinson Technology, Hutchinson, MN, USA), running in Matlab v. 7.0 (The MathWorks Inc., Natick, MA, USA).

Statistical analysis

Continuous variables are expressed as mean ± standard deviation. Ordinal variables are expressed as median and interquartile range (IQR). Comparisons were made using Mann–Whitney U-test for independent samples and Wilcoxon’s signed rank test for dependent samples. Correlations were estimated using Spearman’s rank correlation coefficient. Categorical variables were compared with Fisher’s exact test. P values < 0.05 were considered statistically significant.

Linear regression models were used to evaluate the impact of predictor variables on continuous outcome variables. Model comparison was made using Akaike’s Information Criterion (AIC) to evaluate which predictors were more closely associated with the outcome [33, 34].

All analyses were performed using IBM SPSS 20 for Windows (IBM Corp., Armonk, NY, USA).

Systemic microcirculation

The parameters of the systemic microcirculation before and after the NMES session for the acute and the ICUAW groups are shown in Table 4. In both patient groups, StO₂ upslope and time to baseline were modified by the application of NMES in a statistically significant degree, while this was not the case for the StO₂ downslope. The combined effects of NMES protocol and patient group in the measured microcirculatory parameters are shown In Figures 1, 2, 3 and 4.

	Acute group (n = 19)			ICUAW group (n = 12)
Microcirculatory parameter	Before	After	P	Before	After	P
Tissue oxygenation (StO2)	78 ± 11	77 ± 11	0.2	76 ± 8	75 ± 8	0.4
StO2 downslope (%/minute)	8.6 ± 3.4	9.1 ± 3.1	0.4	9.4 ± 2.6	10.3 ± 3.6	0.1
StO2 upslope (%/second)	2.4 ± 1.4	2.8 ± 1.5	0.03	2.9 ± 1.3	3.6 ± 1.8	0.04
Time to baseline (seconds)	160 ± 48	144 ± 47	0.003	162 ± 47	147 ± 34	0.04

Current intensity and muscle contractions

The current intensity applied at the start of the NMES session was 57 ± 22 mA in the HF group and 63 ± 24 mA in the MF group (P = 0.4). By the end of the session the intensities had been raised to 87 ± 37 mA and 99 ± 36 mA, respectively (P = 0.3). The increase in current intensity that was achieved during the NMES session was 32 ± 18 mA with the HF protocol and 38 ± 17 mA with the MF protocol (P = 0.35 for the comparison of the two protocols). The muscle contractions that were elicited from the quadriceps muscle had a median rating of 2 and IQR of 2 with the HF protocol, and also a median rating of 2 and IQR of 2 with the MF protocol (P = 0.6 for the comparison of the two protocols).

When the two groups of patients were compared, patients in the acute group and those in the ICUAW group reached similar current intensities, with increases in intensity of 32 ± 18 mA and 40 ± 16 mA, respectively (P = 0.2 for the comparison). On the other hand, and in line with their diagnosis, patients in the ICUAW group developed less intense muscle contractions in the lower limbs than patients in the acute group for comparable current intensities: median 1.5 with IQR 2 versus median 3 with IQR 2 (P = 0.002).

NIRS measurements of the vastus lateralis

Of the 31 patients included in the study, only in seventeen were NIRS readings from the vastus lateralis technically feasible. In the rest, no NIRS signal could be obtained, despite the fact that during the design of the study we had anticipated the known factors that interfere with the NIRS readings. Patients with and without NIRS readings did not differ with respect to severity of illness (APACHE II score 19 ± 5 versus 17 ± 3, P = 0.5 and SOFA score 7 ± 2 versus 6 ± 3, P = 0.3) and presence of ICUAW (7/17 versus 5/14, P = 0.5). Patients with NIRS readings, however, had somewhat stronger muscle contractions with median 2 and IQR 2 versus median 1 and IQR 2 (P = 0.08). They also differed in their response in the thenar NIRS measurements: the StO₂ upslope increased by 0.7 ± 0.9 versus 0.4 ± 0.8%/second, while time to baseline decreased by 18 ± 22 versus 14 ± 22 seconds.

In the patients with NIRS measurements (10 in the HF protocol and 7 in the MF protocol), the effects of NMES on the microcirculation of the exercising muscle did not differ between the two protocols. Baseline StO₂ values in the vastus lateralis were 48 ± 19 units versus 51 ± 17 units, respectively (P = 0.7). ΔStO_2slope was 61 ± 6%/minute in the HF protocol and 69 ± 23%/minute in the MF protocol (P = 0.5). ΔStO_2min was 5 ± 4 units with the HF protocol and 7 ± 4 units with the MF protocol (P = 0.3). ΔStO_2max was 6 ± 7 units with the HF protocol and 5 ± 4 units with the MF protocol (P = 0.6).

All three measured microcirculatory parameters in the exercising muscle differed between the acute and the ICUAW group of patients. ΔStO_2slope was 85 ± 27%/minute in the acute group and 42 ± 13%/minute in the ICUAW group (P = 0.04). ΔStO_2min was 8 ± 4 units in the acute group and 3 ± 2 units in the ICUAW group (P = 0.005). ΔStO_2max was 8 ± 6 units in the acute group and 3 ± 2 units in the ICUAW group (P = 0.04).

The strength of the muscle contractions correlated well with the microcirculatory parameters (ΔStO_2slope: r_s = 0.7 with P = 0.05; ΔStO_2min: r_s = 0.8 with P < 0.001; ΔStO_2max: r_s = 0.6 with P = 0.01).