Abstract
Background Skeletal muscle wasting and weakness are common in patients with sepsis in the intensive care unit, although less is known about deficits in diaphragm and limb muscles when mechanical ventilation also is required.
Objective The objective of this study was to concurrently investigate relative differences in both thickness and strength of respiratory and peripheral muscles during routine care.
Design A prospective, cross-sectional study of 16 alert patients with sepsis and 16 people who were healthy (control group) was used.
Methods Assessment was made of the diaphragm, upper arm, forearm, and thigh muscle thicknesses with the use of ultrasound; respiratory muscle strength by means of maximal inspiratory pressure; and isometric handgrip, elbow flexion, and knee extension forces with the use of portable dynamometry. To describe relative changes, data also were normalized to fat-free body mass (FFM) measured by bioelectrical impedance spectroscopy.
Results Patients (9 men, 7 women; mean age=62 years, SD=17) were assessed after a median of 16 days (interquartile range=11–29) of intensive care unit admission. Patients' diaphragm thickness did not differ from that of the control group, even for a given FFM. When normalized to FFM, only the difference in patients' mid-thigh muscle size significantly deviated from that of the control group. Within the patient sample, all peripheral muscle groups were thinner compared with the diaphragm. Patients were significantly weaker than were the control group participants in all muscle groups, including for a given FFM. Within the critically ill group, limb weakness was greater than the already-significant respiratory muscle weakness.
Limitations Volitional strength tests were applied such that successive measurements from earlier in the course of illness could not be reliably obtained.
Conclusions When measured at bedside, survivors of sepsis and a period of mechanical ventilation may have respiratory muscle weakness without remarkable diaphragm wasting. Furthermore, deficits in peripheral muscle strength and size may exceed those in the diaphragm.
Skeletal muscle dysfunction affects both respiratory and peripheral muscles in patients who are critically ill. Although insults that either precipitate or are associated with critical illness–acquired neuromuscular disorders are multifactorial, patients with sepsis-related multiple organ failure who have required mechanical ventilation may be particularly at risk of muscle wasting and weakness.1–5 In the clinical setting, the combination of these and other factors associated with routine care may have unequal effects on different muscles. This conclusion is primarily based on the following observations.
First, differences between respiratory and peripheral muscle groups have been described in several chronic cardiorespiratory conditions, including chronic obstructive pulmonary disease and chronic heart failure.6–8 Changes may relate to the pathophysiology of each condition, but systemic inflammation may contribute to some common mechanisms of damage.8,9 Similarly, a unique pattern of dysfunction may exist as the result of the pathophysiology of sepsis, which includes cytokine responses that can amplify inflammatory- and immobility-induced10 changes in muscle and have secondary effects that propagate damage.11 Furthermore, not only does sepsis-induced myopathy affect both respiratory and peripheral muscles, but the mechanism by which failure of the neuromuscular organ system occurs appears to be different from and predominant over other cachectic states such as nutritional deprivation or disuse atrophy that may coexist in critical illness.12 For respiratory muscles, this mechanism may be confounded by the addition of mechanical ventilation, which itself can result in relatively isolated diaphragm dysfunction.13,14 Yet, the paradigm of additive effects of sepsis and mechanical ventilation on the diaphragm is contentious, from a limited number of apparently conflicting laboratory investigations.15–17
Second, a systematic review of observational studies that measured an aspect of respiratory and peripheral muscle function in adults who were critically ill suggested that there may be a relative preservation of respiratory muscle function in comparison with peripheral groups in patients with either intensive care unit (ICU)–acquired weakness syndromes or sepsis and mechanical ventilation.18 However, this suggestion was based on observations from electrophysiological testing and muscle biopsy, which have not been confirmed at bedside with clinical tests. Methodological limitations of existing studies also included inconsistent sampling strategies, poor definitions of the timing of measurements, and either the absence of or inadequate descriptions of a control group.
Furthermore, although respiratory muscle weakness has been related to limb weakness19 and patterns of weakness have been described in peripheral muscle groups,20,21 no study has specifically sought to compare respiratory and peripheral muscles or to report relative deficits in terms of muscle wasting. The comparison of respiratory and nonrespiratory muscle group strength with respect to wasting is important because muscle-specific force (being force per area of muscle) may be altered in the absence of atrophy, just as atrophy may occur without a reduction in specific force (as occurs with disuse or immobility). Reliable measurements of muscle thickness22 (as an indicator of wasting) can be obtained with diagnostic ultrasound, with techniques described to image the diaphragm23,24 and peripheral groups in the critically ill.25–27 Furthermore, if there is a relationship between weakness and wasting measured at bedside, then there may be potential to further develop imaging methods for the early identification of weakness in patients who cannot complete volitional strength tests.
Finally, understanding the unique elements of muscle dysfunction is important to inform targeted rehabilitation programs in the ICU. Specifically, it is important that there is a sound rationale for strength training in selected respiratory and non-respiratory muscle groups,1,28–30 in addition to more generalized early mobilization.31 This is of particular importance for patients with sepsis, which is a risk factor for ICU and respiratory muscle weakness19,32 that may translate to poor long-term functional outcomes33 and health-related quality of life.34
The primary objectives of this study were to determine muscle strength and size (as a measure of wasting) in the respiratory and limb muscles of a sample of critically ill survivors of sepsis as compared with healthy control patients and to determine relative effects in the degree of dysfunction between muscle groups.
Method
A prospective, cross-sectional design was used with a case-controlled element. Patients admitted to a single tertiary ICU were prospectively screened from November 2010 through December 2011. Consecutive patients who were ≥18 years of age and required >12 hours of invasive mechanical ventilation were identified. Patients were excluded if they met any of the following criteria: open cardiac, thoracic, or upper abdominal surgery with an anticipated hospital admission of <14 days; pre-existing or acute central nervous system, motor neuron, peripheral nerve, or neuromuscular junction disorder that results in neuromuscular weakness (other than ICU-acquired weakness); conditions known to have a differential effect on respiratory and peripheral muscles such as chronic obstructive pulmonary disease (GOLD scale, moderate or severe),35 chronic heart failure (New York Heart Association class III or IV),36 or cystic fibrosis; admission related to self-harm, cognitive/intellectual impairments that precluded informed consent, or psychiatric admission; alcohol-related admission, liver failure, or dependence; morbid obesity; cancer or metastatic disease; pre-existing chronic systemic inflammatory condition or current immunosuppression; prolonged hospitalization in the previous 3 months; anticipated hospital length of stay of ≤5 days; non–English-speaking; pregnancy; abnormal chest wall or thoracic spine anatomy; anorexia; and severe burns. The remaining patients were monitored while mechanically ventilated for sepsis, according to the consensus definition.37
Patients who were septic, required mechanical ventilation for ≥5 days, and were able to provide informed consent were approached to participate in the study. This patient selection was done once all the following criteria were met: currently admitted to the ICU; no palliative care order; no subsequent identification of an exclusion criterion; >2 weeks after open upper abdominal or thoracic surgery; no hemodyamic, respiratory, or mobility restriction that precluded body weight measurement; minimal ventilation support sufficient for respiratory muscle measurements (fraction of inspired oxygen ≤40%, positive end-expiratory pressure [PEEP] ≤7.5 cm H2O, pressure support ≤12 cm H2O); sufficient wakefulness (determined as a Richmond Agitation and Sedation Scale score of −1 to +1 and ≥14 on the Glasgow Coma Scale)38; and no renal dialysis at the time of measurements or in the previous 24 hours. These milestones were used in an attempt to standardize the timing of the clinical outcome assessments of muscle strength, thickness, and body composition. The criteria were selected to ensure that these measures could be performed safely at the earliest point at which acceptable reliability and validity could be expected.
Volunteers who were healthy and did not meet the same relevant exclusion criteria were matched 1:1 to participants who were critically ill for sex and age (within 2 years), which is an approach that has been used in previous studies of patients who are critically ill.18 All participants provided informed consent.
Clinical Status
Data, including severity of illness, clinical care (duration of mechanical ventilation, administration of medications and nutrition), and attentiveness (Attention Screening Examination [ASE]),39 were collected from the patients. The minimum target ASE score that has been used to identify patients who have suitable attention to adhere to volitional measures of muscle strength40 and for the absence of inattention39 has been described as ≥8 of 10. However, ASE scores were not used to prospectively make a formal diagnosis of delirium or alter the timing of volitional strength testing. Furthermore, a venous blood sample was collected from all participants for the measurement of selected circulating cytokines (tumor necrosis factor-α; interleukin [IL]-6, IL-10, IL-18; and monocyte chemotactic protein-1) by enzyme-linked immunosorbent assays according to manufacturer protocols (human tumor necrosis factor-α, IL-6, IL-10, and monocyte chemotactic protein-1 [DuoSet ELISA development system, R&D Systems Inc, Minneapolis, Minnesota]; human IL-18 [platinum ELISA, BMS267/2, e-Bioscience, San Diego, California]). All participants also had their pre-existing comorbidity and activities of daily living independence rated using the Charlson Comobidity Index,41 Functional Comorbidity Index,42 and Katz Index.43 The following clinical measures were completed within 1 day by a single investigator, with each parameter measured in triplicate. Individual components were omitted if contraindicated.
Body Composition Analysis
Body weight was measured by means of a scale system appropriate to each participant's mobility, and supine body length was measured with the use of a tape measure to the nearest 0.1 cm. Fat-free body mass (FFM) was approximated with the use of a tetrapolar bioelectrical impedance spectroscopy technique as previously described and expressed as a percentage of body weight (SFB7, ImpediMed Ltd, Pinkenba, Queensland, Australia).44 Briefly, electrodes were placed on the dorsal hand and foot unilaterally, and participants rested supine with the medial surfaces of the limbs abducted, resting away from and not touching the body. Data then were uploaded to the Bioimp software (v. 5.3.1.1, Impedimed Ltd).
Diaphragm Thickness
The thickness of the right hemidiaphragm at the zone of apposition in the midaxillary line of the ninth intercostal space was used as a measure respiratory muscle size, as previously described.22 Accordingly, the diaphragm was imaged with participants semirecumbent at 45 degrees and at an end-expiratory lung volume during unsupported tidal breathing with the use of a linear-array ultrasound transducer (75L38EA with the DP-6,600, Shenzhen Mindray Bio-medical Electronics Co Ltd, Shenzhen, China). Diaphragm thickness was measured between the superficial edge of the diaphragmatic pleura and the peritoneum, where the limiting membranes were closest to parallel.
Peripheral Muscle Thicknesses
Peripheral muscle size was approximated by measurement of the thickness of the anterior mid–upper arm, mid-forearm, and mid-thigh musculature, with the use of the same ultrasound unit as previously described.22 Accordingly, participants were positioned supine, and thicknesses were measured between the most superficial aspects of the fat-muscle interfaces and limiting tissues (ie, humerus, interosseous membrane, and femur).
Peripheral Muscle Strength
Peak isometric handgrip, elbow flexion, and knee extension forces were evaluated in modified recumbent positions with the use of portable dynamometers, as previously described (Jamar hydraulic hand dynamometer, Sammons Preston Rolyan, Bollingbrook, Illinois; Lafayette manual muscle test system model 01163, Lafayette Instrument, Lafayette, Indiana).45 The Medical Research Council sum score also was determined by grading 3 upper limb and 3 lower limb groups bilaterally to ascertain whether patients met the ICU-acquired weakness criteria32 score of <48 of 60.
Inspiratory Muscle Strength
Maximal inspiratory pressure (MIP) was measured as the maximum pressure sustained for 1 second, with the participant in a semirecumbent posture at 45 degrees (MicroRPM, Micro Medical Ltd, Kent, United Kingdom). Participants with a normal upper airway were coached to inspire from residual lung volume through a flanged mouthpiece (MTH6400, Micro Medical Ltd); a unidirectional valve method that allowed expiration but closed with inspiratory flow was used for uninterrupted 20- to 25-second periods in patients who were tracheostomized.46
Data Analysis
Right-sided measurements were averaged with the exception of MIP, in which the best effort was selected. The reproducibility of the triplicate measurements of muscle strength and thicknesses variables was determined by calculation of the intraclass correlation coefficient (ICC). Continuous variables are expressed as either mean (standard deviation) or median (interquartile range), as appropriate for the distribution. Parametric statistics were used to analyze primary outcomes once it was determined that assumptions were met (PASW v. 17.0.2, SPSS Inc, Chicago, Illinois). Thus, an independent-samples t test was used to evaluate differences between groups, and correlations were associated by Pearson r with expression as r2. Patients' muscle thickness and strength data also were expressed as z scores, with reference values obtained from the control group for within-group analysis by repeated-measures analysis of variance. A sample of 14 patients was determined to be required a priori to detect a large difference (f=0.4) between the thickness and strength of 4 muscle groups by repeated-measures analysis of variance, assuming ρ=.3, α=.05, and β=1 minus 0.2.47 Participants who had missing data for individual variables are identified in the results and were excluded from relevant analyses. A level of P<.05 was considered statistically significant.
Role of the Funding Source
Financial support for equipment and laboratory materials was provided by Flinders Medical Centre Critical Care Medicine research funds and by Flinders Medical Centre Foundation Seeding Grant 2010.
Results
Patient Flow and Characteristics
The flow of screened patients is depicted in Figure 1. A total of 1,700 patients were screened to ascertain that they were ≥18 years of age and required ≥12 hours of invasive mechanical ventilation; of the people who were ascertained, 562 were excluded. Of 43 eligible patients (septic and required mechanical ventilation for at least 5 consecutive days), clinical measures were unable to be completed before ICU discharge in 27 people. Of the 27 patients in whom clinical measures were not able to be completed, 9 were within 2 weeks of abdominal or thoracic surgery, 4 died, 3 were unable to give informed consent because of delirium, 2 were found to have metastatic disease, 1 had a brain abscess, 1 subsequently had a stroke, 2 required renal dialysis, 2 were missed as discharged from the ICU early on the day they met eligibility criteria, 2 had restricted mobility precluding body-weight measurement, and 1 declined consent. Therefore, 16 patients who were critically ill with sepsis (9 men, 7 women) and 16 people who were healthy (control group) were studied. All participants had lived in their own home before hospital admission and were independent in activities of daily living according to the Katz Index,43 with the exception of 1 person who resided in a low-level care facility and had assistance with bathing. All patients survived their initial ICU admission. Four patients had ICU readmissions, and 2 people died during their subsequent hospital stay. Only 3 of 14 survivors who were previously from home returned home directly from the acute care hospital (Tab. 1).
Flowchart of study enrollment of participants who were critically ill. ICU=intensive care unit.
Participant Characteristics (N=16)a
Patient characteristics are outlined in Tables 1 and 2. A tracheostomy was performed in half of the sample, which was still present in 5 patients at the time of clinical assessments, which were completed at a median of 16 days (interquartile range=11–29) of ICU admission. At that time, all patients were classified as alert and calm on the Richmond Agitation and Sedation Scale and had a Glasgow Coma Scale score of 15 of 15. Furthermore, 15 of 16 patients achieved an ASE score of ≥8 of 10. Other physiological characteristics of the critically ill sample had acceptable reference ranges for clinical measurement (results not shown).
Characteristics of Participants Who Were Critically Ill and Those Who Were Healthy (Control Group)a
In terms of physical therapy, 13 patients' ability was limited to limb exercises in bed (assisted or against gravity), with some of these patients also sufficiently stable for a fully assisted transfer to sit out of bed in an appropriately supportive chair. Because of the criteria for clinical measures, patients were generally measured at the earliest time that they were able to actively and safely engage in mobilization with physical therapists (ie, beyond sitting on the edge of the bed). Thus, physical therapy management before this time had been restricted to basic bed exercises and more passive interventions. Only 3 participants were capable of performing a standing transfer to a chair, with some level of physical assistance from up to 2 people. In each case, this transfer was performed as part of body-weight measurement and was the first time that participants had been able to do such a transfer. Patients did not receive physical therapy on the day of clinical assessments before measurements.
Equivalence of Patients and Controls
Patients who were critically ill had a significantly greater degree of comorbidity compared with the control group according to both the Charlson Comorbidity Index41 and Functional Comorbidity Index42 scores (Tab. 2). Patients also had greater circulating levels of IL-6 compared with the control group (Tab. 2). In a comparison of anthropometric characteristics, there was no difference in supine body length, body weight, percentage of predicted body weight,48 or body mass index between patients and controls. Body composition analysis was completed in 15 patients because of a technical difficulty with the bioelectrical impedance spectroscopy device on 1 occasion. There was a strong trend toward lower FFM in patients compared with control participants; yet, when expressed as a percentage of body weight (%FFM), patients had a significantly lower FFM than did control participants (Tab. 2).
Because the groups were incomparable on FFM and several measures of both strength and bulk were linearly correlated with this variable (eFigs. 1 and 2), muscle data also were normalized to FFM to facilitate description of relative changes. Peripheral muscle forces tended to be correlated with the thickness of their corresponding muscle group in the critically ill group; thus, strength measurements were normalized to muscle size (Fig. 2).
Plots of (A) diaphragm thickness versus maximal inspiratory pressure, (B) mid–upper arm muscle thickness versus elbow flexion force, (C) mid-forearm muscle thickness versus handgrip force, and (D) mid-thigh muscle thickness versus knee extension force. ▲=patient in critically ill group, ○=participant in control (healthy) group.
Respiratory and Peripheral Muscle Measurements
There was excellent reproducibility of the triplicate measurements of diaphragm thickness (ICC=.938), peripheral muscle thicknesses (ICC≥.976), MIP (ICC=.904, n=13), and peripheral muscle forces (ICC≥.833) in the patients that was comparable to that of the control group (results not shown). The MIP testing was omitted in 1 patient with a bronchopulmonary fistula and in another patient in whom technically satisfactory measurements were prevented by facial muscle weakness and inadequate mouth seal. One other patient performed only 1 MIP maneuver, with this result taken as the “best effort,” and was included in the main analysis.
Table 3 summarizes the primary outcome measures. There was no difference in diaphragm thickness between the patients who were critically ill with sepsis and control participants, even for a given FFM; however, upper arm, forearm, and thigh muscle bulk was lower in the patients. Yet, corrected for FFM, there was only significantly less thigh muscle size in patients compared with the control group.
Thickness and Strength of Respiratory and Limb Muscles in Study Groupsa
Patients who were critically ill with sepsis were significantly weaker than were the control participants, as determined by the measures of respiratory and limb muscle strength (P≤.001). Even when MIP, handgrip, elbow flexion, and knee extension forces were expressed relative to FFM and muscle size, differences between study groups remained (P≤.001 for all muscles).
Comparisons Between Muscle Groups
To determine the degree of dysfunction between muscle groups in patients who were critically ill, differences in z scores were analyzed. Compared with the diaphragm, there was a significant difference in the thickness (P≤.001, Fig. 3A) and thickness/FFM (P≤.015, Fig. 3B) of all peripheral muscles. Furthermore, the thinning of the thigh musculature was significantly more marked than in the upper arm and appeared greater than in the forearm.
Diaphragm and limb muscle thicknesses of patients who were critically ill expressed as z scores. The z scores were normally distributed, and data were analyzed by repeated-measures analysis of variance. Box plots represent medians, interquartile ranges, and both 10th and 90th percentiles. Patient data are expressed relative to data for the control (healthy) group; thus, the reference line at zero indicates the score in which there would be no deviation from normative values. Data in graph A are measurements of thickness, in which ○ represents an outlier who was a 76-year-old man with sepsis. Data in graph B are measurements of thickness/fat-free mass (FFM), in which ○ represents an outlier who was a 28-year-old woman with sepsis.
Similarly, there was significantly less strength (P≤.02, Fig. 4A) and strength/FFM (P≤.001, Fig. 4B) in the peripheral muscles compared with the respiratory muscles. However, when differences between muscle groups were analyzed with strength expressed relative to thickness, weakness appeared most marked in the upper limb, particularly proximally in the upper arm (Fig. 4C). Finally, the pattern of muscle weakness was not affected by a clinical diagnosis of ICU-acquired weakness (results not shown). Only 8 participants (50% of the sample) met the criteria for ICU-acquired weakness, defined as a Medical Research Council sum score of <48 of 60 (sample mean=48, interquartile range=42–54).
Maximal inspiratory pressure (MIP) and peripheral muscle forces of patients who were critically ill expressed as z scores. The z scores were normally distributed, and data were analyzed by repeated-measures analysis of variance. Box plots represent medians, interquartile ranges, and both 10th and 90th percentiles. Patient data are expressed relative to data for the control (healthy) group; thus, the reference line at zero indicates the score in which there would be no deviation from normative values. Data in graph A are measurements of strength, in which ○ represents outliers who were an 81-year-old man with sepsis and a 24-year-old man with septic shock. Data in graph B are measurements of strength/fat-free mass (FFM), in which ○ represents outliers who were the same 81-year-old man with sepsis and a 72-year-old man with septic shock. Data in graph C are measurements of strength/thickness, in which ○ represents an outlier who was the same 81-year-old man with sepsis as in the analysis of graphs A and B.
Discussion
This study sampled a very ill group of patients with sepsis who required mechanical ventilation and other care that included sedation and vasopressor support. Through the use of readily available clinical assessments and a pragmatic approach, this study suggests that by 2 weeks of ICU admission, muscles of different functionality may not be equally affected by a combination of insults that occur with critical illness. Furthermore, the pattern of muscle wasting as measured by muscle thicknesses differed from that of muscle weakness. This finding is akin to reductions in muscle mass and specific force in various animal models of sepsis.11
With regard to muscle wasting, there was no difference in diaphragm thickness between the patients and the control participants. Previous studies in the ICU have suggested both relatively isolated diaphragm dysfunction13,14,49 and relative preservation of the diaphragm compared with other muscles,18,50 which may depend on the cohort studied. Rapid atrophy (a reduction in cross-sectional area) has been observed in type I and II fibers of inactive diaphragms from organ donors who were brain dead and mechanically ventilated in comparison with surgical control patients13,51 but without such alterations in the pectoralis major13 or biceps brachii49 muscles. These findings appear to reflect the effect of controlled mechanical ventilation in patients without respiratory drive or neurotrophic influences. Atrophy may still occur in the diaphragm of other mechanically ventilated patients who have died in the ICU,52 but increased levels of myogenic regulatory factors with lower markers of proteolysis may indicate compensatory mechanisms to maintain diaphragm mass as compared with the quadriceps muscles of the same patients.50 However, these findings could reflect an extreme loss of function and failure of the muscular organ system in fatal circumstances.
Only 2 studies have longitudinally tracked changes in diaphragm thickness on ultrasound with mechanical ventilation. The first study reported on a small heterogeneous sample (N=7) under pressure-regulated volume control ventilation (VT=400 mL, PEEP=5 cm H2O) in which an average reduction in thickness of 6% per day was observed23; yet, the median duration of ventilation was limited to 5 days (range=4–9), there was no control group, and a limitation of diaphragm ultrasound in such longitudinal studies of mechanically ventilated patients is the unknown effect of PEEP on resting thickness. Conversely, diaphragm thickness was reported to be unchanged for up to 14 days after baseline in patients who were mechanically ventilated at enrollment for acute respiratory failure24; yet, only 4 of 12 patients with sequential measurements were included at the day-14 analysis, there was no control group, and the total duration of mechanical ventilation was not described. A limitation of the present study is that longitudinal measurements were not obtained because it was not the aim to track the rapidity or greatest magnitude of diaphragm damage or the time course of recovery, which remains to be determined in ventilated patients with sepsis.
Our diaphragm thickness findings must be considered in light of several limitations because it could appear that they contradict previous observations. First, as identified, few clinical ultrasound studies have attempted to quantify the morphometry of the diaphragm with mechanical ventilation,23,34 and studies in patients with sepsis are lacking. Therefore, existing data for the power calculation for this variable were limited. Nonetheless, it appears that a large difference in diaphragm thickness may not be expected and that this component of analysis may have been underpowered. In addition, because of limitations of the technique and ICU population, it was not possible to standardize absolute lung volume among patients. Thus, our results represent a clinical pattern at a single point during recovery from sepsis when patients are starting to actively mobilize out of bed, a median of 256 hours of mechanical ventilation over 16 days of ICU admission.
The present study also demonstrated that peripheral muscle wasting was most marked in the mid-thigh musculature. The quadriceps has been the most extensively studied peripheral muscle group in patients who are critically ill, with measurements of muscle layer thickness on ultrasound in observational and interventional studies.25,26,30,53 The greatest magnitude of wasting appears to occur in patients with sepsis and multiple failed organs1,27 and at a faster rate in the first week of illness.26 There also may be a trend toward larger relative losses of muscle mass in patients with greater muscle thicknesses on study recruitment compared with patients with less muscle mass.53 Accordingly, the quadriceps muscles are a considerable protein reservoir that may be a target of increased proteolysis and preferential type II fiber atrophy.2
With regard to muscle weakness, our patients were significantly weaker in all muscle groups compared with the control participants, even when normalized to FFM or muscle size. This finding indicates that factors other than wasting account for strength loss and that the measurement of muscle size on ultrasound may not be an appropriate surrogate measure of strength. In addition, because the relationships between strength and muscle mass and size differed between both patients and control participants and the diaphragm and peripheral muscle groups (Fig. 2 and eFig. 1), we suggest that factors accounting for variability in strength are different in patients compared with healthy people and in the diaphragm compared with the limb muscles. These factors may include the severity of illness or sepsis and markers of systemic inflammation4 such as plasma IL-6, which has been associated with myosin loss.54 Future studies should be adequately powered and designed to investigate unexplained variances in muscle strength and include factors other than muscle size and mass.
We also report that patients' respiratory muscle weakness was less marked than was their peripheral muscle weakness. Although there are limitations to the use of volitional tests of muscle strength, attempts were made to minimize these limitations by assessing patients who were alert and attentive. Furthermore, the level of measurement error and reliability was as expected and comparable to other studies. Still, few data exist in combined models of sepsis and mechanical ventilation to explain this result.15–17 It may be that there is a balance between the prevention of ventilator-induced diaphragm dysfunction by maintaining some diaphragm activity and the protective use of mechanical ventilation to avoid load and systemic inflammatory–induced injury amid the increased metabolic demands of sepsis. For example, with controlled mechanical ventilation, as workload is taken over by the ventilator, a relative energy substrate over supply may occur in diaphragm fibers,49 as may a reduction in blood flow that is not seen in peripheral muscles.55 Conversely, with sepsis, respiratory muscles are depleted of energy-rich phosphates56 despite a relative preservation of blood flow by comparison with peripheral muscles.57 This balance may vary throughout the course of illness and recovery and with reloading.
Considerations for Clinical Practice
Patients who are critically ill with sepsis who require invasive mechanical ventilation may develop a combined generalized myopathy and ventilator-induced diaphragm dysfunction that may be recognized as a failure to wean or as ICU-acquired weakness.32 However, if respiratory muscle weakness is suspected, the potential for worse-than-expected peripheral neuromuscular deficits also should be considered, prompting regular standardized physical examination of muscle function at the bedside. Furthermore, the inability of patients to perform well on volitional tests of strength is important because they require the same motivation, effort, and coordination to cough to clear secretions and move themselves to improve their independence. Only about 20% of our patients, however, were able to return to their preadmission residence on discharge from their acute hospital stay without any additional rehabilitation or care supports, which suggests that weakness detected in the ICU could be a predictor of patients' discharge disposition or level of independence and requires further research. Finally, therapies that prevent or rehabilitate both a loss of muscle mass and muscle strength, potentially focusing on peripheral muscles, should be investigated, with recognition that the processes that regulate weakness and atrophy may not be the same.
Limitations
In addition to limitations already outlined, the findings of this study should be considered in light of its cross-sectional design. Furthermore, the interpretation of diaphragm findings is somewhat limited in that the duration that various ventilation modes were used was not recorded. Still, patients in the present study were treated with individualized weaning strategies aimed at minimal use of controlled mechanical ventilation and early progression to triggered modes of ventilatory support before either extubation or tracheostomy with intermittent spontaneous breathing trials, which is when patients in the present study were assessed. However, the degree of diaphragm activation may be more important than the mode, with the optimal level of activation required to mitigate ventilator-induced diaphragm dysfunction ultimately unknown, in addition to a lack of feasible and reliable methods for monitoring diaphragm activity at bedside.
For the case-controlled component of this study, a healthy ambulatory group was recruited. Although the groups were matched on some characteristics, they were incomparable on comorbidities or a pre-existing illness trajectory. Alternatively, larger normative datasets could have been used, although no reference sets of sufficient size or demographic spread could be located for the comparison of the range of outcomes that have been further modified for application to the critically ill cohort in the present study. Still, the quality of this study might have been improved had the 1:1 patient-to-control participant ratio been increased. Otherwise, a control group of patients without sepsis who also fail to meet exclusion criteria and require prolonged mechanical ventilation would allow investigation of the specific effects of sepsis; however, from our recruitment experience, there were few such patients.
Finally, as the pathophysiology of sepsis-related skeletal muscle dysfunction appears dominant and different from immobilization-induced disuse atrophy, we did not try to control for the duration of immobility. Rather, our results reflect both sepsis-induced and immobility-induced changes because our criteria for clinical assessments meant that patients were recruited almost as soon as they were able to start actively participating in mobility therapy.
In conclusion, muscle wasting and weakness may occur unevenly in patients recovering from sepsis and an extended period of mechanical ventilation. In particular, in this cohort of patients, the diaphragm was relatively well preserved amid more excessive deficits in the peripheral muscle groups. However, these findings may not be the case in the presence of other comorbid conditions6–8 or at other stages of illness and recovery. Further insight will be gained by future research with more objective measures of muscle atrophy (computerized tomography, biopsy, more complex body composition analysis) but particularly nonvolitional muscle force (phrenic, ulnar, and femoral nerve stimulation), which then should be studied for correlation with more patient-centric functional measures.
The Bottom Line
What do we already know about this topic?
Muscle weakness and wasting are common complications of critical care that may occur most severely in patients who are admitted with sepsis and require mechanical ventilation.
What new information does this study offer?
Different muscle groups may not be affected to the same extent. This study suggests that there may be a relative preservation of respiratory muscle function compared with the quadriceps femoris muscle group.
If you're a patient or a caregiver, what might these findings mean for you?
While recovering from a critical illness, patients may have difficulty with tasks such as sit-to-stand, and interventions, such as early rehabilitation with a focus on functional lower limb strength training, are likely to be most beneficial.
Footnotes
Both authors provided concept/idea/research design and data analysis. Dr Baldwin provided writing, data collection, and project management. Professor Bersten provided fund procurement, facilities/equipment, institutional liaisons, and consultation (including review of manuscript before submission).
This study protocol conformed to the principles of the Declaration of Helsinki and was approved by the Institutional Human Research Ethics Committee of Southern Adelaide Health Service/Flinders University (No. 395/10).
This research, in part, was presented at: Proceedings of the ANZICS/ACCCN Intensive Care ASM; October 17–19, 2013; Adelaide, Australia.
Financial support for equipment and laboratory materials was provided by Flinders Medical Centre Critical Care Medicine research funds and by Flinders Medical Centre Foundation Seeding Grant 2010.
- Received February 11, 2013.
- Accepted August 30, 2013.
- © 2014 American Physical Therapy Association
References
Issue highlights
