Δευτέρα 10 Φεβρουαρίου 2020

Perioperatively Acquired Weakness

ORIGINAL RESEARCH ARTICLES: ORIGINAL CLINICAL RESEARCH REPORT

Perioperatively Acquired Weakness

Lachmann, Gunnar MD^*,†; Mörgeli, Rudolf MD^*; Kuenz, Sophia MD^*; Piper, Sophie K. PhD^†,‡; Spies, Claudia MD^*; Kurpanik, Maryam^*; Weber-Carstens, Steffen MD^*,†; Wollersheim, Tobias MD^*,†; on behalf of the BIOCOG Consortium

Author Information

Anesthesia & Analgesia: February 2020 - Volume 130 - Issue 2 - p 341-351

doi: 10.1213/ANE.0000000000004068

FREE
SDC

Abstract

BACKGROUND:

Skeletal muscle failure in critical illness (intensive care unit-acquired weakness) is a well-known complication developing early during intensive care unit stay. However, muscle weakness during the perioperative setting has not yet been investigated.

METHODS:

We performed a subgroup investigation of a prospective observational trial to investigate perioperative muscle weakness. Eighty-nine patients aged 65 years or older were assessed for handgrip strength preoperatively, on the first postoperative day, at intensive care unit discharge, at hospital discharge, and at 3-month follow-up. Functional status was evaluated perioperatively via Barthel index, instrumental activities of daily living, Timed Up and Go test, and functional independence measure. After exclusion of patients with intensive care unit-acquired weakness or intensive care unit stay of ≥72 hours, 59 patients were included into our analyses. Of these, 14 patients had additional pulmonary function tests preoperatively and on postoperative day 1. Blood glucose was measured intraoperatively every 20 minutes.

RESULTS:

Handgrip strength significantly decreased after surgery on postoperative day 1 by 16.4% (P < .001). Postoperative pulmonary function significantly decreased by 13.1% for vital capacity (P = .022) and 12.6% for forced expiratory volume in 1 second (P = .001) on postoperative day 1. Handgrip strength remained significantly reduced at hospital discharge (P = .016) and at the 3-month follow-up (P = .012). Perioperative glucose levels showed no statistically significant impact on muscle weakness. Instrumental activities of daily living (P < .001) and functional independence measure (P < .001) were decreased at hospital discharge, while instrumental activities of daily living remained decreased at the 3-month follow-up (P = .026) compared to preoperative assessments.

CONCLUSIONS:

Perioperatively acquired weakness occurred, indicated by a postoperatively decreased handgrip strength, decreased respiratory muscle function, and impaired functional status, which partly remained up to 3 months.

KEY POINTS

Question: Does a clinically measurable “perioperatively acquired weakness” occur?
Findings: Postoperative muscle strength, pulmonary function, and functional status significantly decreased after surgery.
Meaning: Perioperatively acquired weakness is a relevant perioperative comorbidity. A standardized assessment should identify patients with perioperatively acquired weakness and enable initiation of an adequate follow-up care.

Intensive care unit-acquired weakness is a common complication in intensive care resulting in increased intensive care unit and hospital length of stay, increased morbidity and mortality, and an impaired long-term outcome.^1–3 Survivors suffer from significant physical impairments, delayed return to the workforce, and substantially reduced quality of life.⁴^,⁵ Importantly, handgrip strength was independently associated with poor hospital outcome.² Intensive care unit-acquired weakness develops early during the intensive care unit stay and is frequently seen in mechanically ventilated patients suffering from sepsis and multiple organ failure.^6–9 Risk factors of intensive care unit-acquired weakness include sedation, immobilization, and hyperglycemia, whereas the impact of neuromuscular blocking agents and corticosteroids remains controversial.⁹ Patients undergoing surgical procedures in general anesthesia are exposed to a similar setting, including sedation, immobilization, mechanical ventilation, and glycemic variability, as well as a severe surgically induced inflammatory response.¹⁰ Dysglycemia is associated with muscle degradation and impaired outcome in the intensive care unit setting.^11–13 Crucially, weakness and induction of muscle protein degradation of the diaphragm were shown to take place after only 2 hours of mechanical ventilation.¹⁴ We hypothesize that, due to the similarities in setting and muscular pathology, a relevant muscular functional deficit may develop in the perioperative period and not exclusively in the intensive care unit setting.

Weakness is known as a negative predictor for outcome, leading to insufficient mobility and impaired respiratory function³ and possibly hindering postoperative physiotherapeutic interventions. Early mobility and physiotherapy are established preventive measures to reduce postoperative complications.^15–17 Weakness and the inability to partake in preventive interventions render patients vulnerable for postoperative complications, such as hospital-acquired pneumonia. As a result, a perioperatively acquired weakness may create the basis for short- and long-term postoperative sequelae¹⁸ both for intensive care unit and for non-intensive care unit surgical patients. However, the impact of surgical procedures on muscle weakness has not yet been investigated.

Our primary objective was to determine whether clinically measurable peripheral weakness or functional decline occur during the perioperative period. Secondary objectives included the evaluation of respiratory muscle function in the perioperative period, as well as any association between perioperative muscle weakness and surgical time, intraoperative glucose metabolism, or inflammatory markers.

METHODS

Study Participants and Data Acquisition

Table 1. - Timeline of Measurements

	Preoperative	Postoperative Day 1	Intensive Care Unit Discharge	Hospital Discharge	3-mo Follow-up
Handgrip strength test	X	X	(X)	X	X
Medical Research Council score	X	X	(X)		X
Barthel index	X			X	X
Instrumental activities of daily living	X			X	X
Timed Up and Go test	X			X	X
Functional independence measure	X			X
Pulmonary function test	X	X

, measured in all patients; (X), measured in only ICU patients.

This is a subgroup investigation of the prospective, observational Biomarker Development for Postoperative Cognitive Impairment in the Elderly study (www.biocog.eu), which aims to identify neuroimaging and molecular biomarkers for postoperative delirium and postoperative cognitive deficits.¹⁹ Inclusion criteria were age ≥65 years and elective surgery with at least 60 minutes expected surgical time. Although Biomarker Development for Postoperative Cognitive Impairment in the Elderly is a large multicenter study, this subgroup investigation included only patients from the Charité – Universitätsmedizin Berlin. A subgroup of 89 patients with a clinical indication for the placement of an arterial catheter during anesthesia was evaluated according to our study protocol (Table 1). All patients received guideline-based anesthesiological and surgical treatment according to our standard operating procedures.²⁰ In order to avoid bias due to postoperative intensive care unit complications, we focused on patients who proceeded to a normal ward after the recovery room stay, as well as patients with short intensive care unit stays, intermediate care unit or postanesthesia care unit for observation purposes. Therefore, we excluded all patients with a postoperative intensive care unit stay of ≥72 hours or with positive intensive care unit-acquired weakness diagnosis (Medical Research Council score ≤48, according to the study by De Jonghe et al¹), as well as non-intensive care unit patients who were subsequently transferred to an intensive care unit before discharge. For the purposes of this analysis, the term “ICU patients” includes those treated in an intensive care unit, intermediate care unit, or postanesthesia care unit. The institutional review board provided ethics approval (Ethikkommission der Charité – Universitätsmedizin Berlin, EA2/092/14). This clinical trial meets the requirements set out by the International Council for Harmonisation Good Clinical Practice Guidelines and Declaration of Helsinki. Written informed consent was obtained from all patients. The trial was registered before patient enrolment at clinicaltrials.gov (NCT02265263; principal investigator, Prof. Claudia Spies, date of registration: September 23, 2014).

Peripheral Muscle Strength

Handgrip strength was measured via dynamometer (SAEHAN SH5002; SAEHAN Corp, Changwon, Korea), using the best of 3 trials of the dominant hand, and the data were calculated as percentages of patients’ expected age- and gender-matched standard values.²¹ Muscle strength was measured using the Medical Research Council score for muscle strength according to the study by De Jonghe et al,¹ assessing 12 predefined muscle groups (3 muscle groups in each of the upper and lower limbs). If an Medical Research Council score could not be evaluated at single muscle groups due to clinical reasons, results of the contralateral muscle group were used to calculate the sum score.²² For the purpose of a daily longitudinal observation, a subset of 7 patients received daily measurements of handgrip strength and muscle strength scoring from the day before surgery until postoperative day 7. For correlation analyses, decreases in handgrip strength and muscle strength by Medical Research Council score were defined as relative change between preoperative and first postoperative age- and gender-adjusted measurements. Results of muscle strength are shown as mean Medical Research Council of the sum Medical Research Council score for all patients.

Functional Measurements

Timed Up and Go test,²³ Barthel index,²⁴ instrumental activities of daily living,²⁵ and functional independence measure²⁶ questionnaires were used for assessment of functional status. For correlation analyses, a decrease or increase in the scores was defined as a relative change between preoperative and postoperative values at hospital discharge.

Respiratory Muscle Function

A subset of 14 patients received pulmonary function tests by spirometry (Pneumotrac Model 6800; Vitalograph, Ennis, Ireland). Vital capacity and forced expiratory volume in 1 second (FEV1) were measured and presented as percentages of patients’ expected age- and gender-matched standard values. For correlation analyses, vital capacity and FEV1 decreases were defined as relative change between preoperative- and postoperative age- and gender-adjusted measurements.

Intraoperative Glucose Metabolism

Intraoperative arterial blood gas analyses, including blood glucose measurement, were performed by ABL800 FLEX Radiometer (Radiometer Medical ApS, Brønshøj, Denmark) immediately after the establishment of an arterial catheter before each surgical procedure and then regularly every 20 minutes until the end of anesthesia. If it did not exceed 60 minutes, the first glucose measurement after extubation was considered in the analysis. Intraoperative mean, minimum, and maximum blood glucose, as well as blood glucose variability given by the SD, were determined. Dysglycemia was defined according to current recommendations²⁷: severe hypoglycemia (<40 mg/dL), mild hypoglycemia (levels between 40 and 79 mg/dL), mild hyperglycemia (levels between 150 and 179 mg/dL), and severe hyperglycemia (≥180 mg/dL).

Perioperative Inflammation

By using chart reviews of blood samples measured within clinical routine, we reviewed C-reactive protein (CRP) values and leukocyte concentrations, both preoperatively and on postoperative day 1, in order to determine whether a perioperative immune response is associated with perioperative weakness.

Statistical Analysis

Data were expressed according to their scaling as median (with the limits of the interquartile range) or frequencies (%). Since the majority of the outcome variables were not normally distributed, only nonparametric testing was performed. Nonparametric Friedman test was used for global testing of handgrip strength, Medical Research Council score, and the functional scores over all measured time points. In a second step, a post hoc analysis of our data used Wilcoxon signed-rank tests for pairwise comparison of different time points (longitudinal analysis) within 1 patient group. Moreover, nonparametric Mann–Whitney U test and Kruskal–Wallis test were used for cross-sectional comparisons between different patient subgroups (total intravenous versus volatile anesthesia, rocuronium versus cisatracurium, epidural catheter versus no epidural catheter, intraabdominal/intrathoracic versus other surgeries, comparisons among different subspecialties). Nonparametric Spearman’s correlations were performed to assess correlations between duration of surgery, glucose metabolism and inflammatory response on handgrip strength, a Medical Research Council score, and vital capacity and FEV1 decreases. A 2-tailed P < .05 was considered statistically significant. All P values constitute exploratory data analysis without adjustment for multiple comparisons. Since several tests were performed within the investigation, adjustment of multiple testing within the several outcomes analyzed does not ensure that the overall type 1 error is 5% or less. All P values of statistical tests are to be understood as exploratory and interpreted with caution and with no confirmatory generalization of the results.²⁸^,²⁹

No sample size calculation was done a priori for this exploratory subgroup analysis, but a sample size of 60 would have 80% power to detect a standardized effect size of 0.368 using a paired t-test with a 0.050 two-sided significance level (nQuery Advisor 7.0; Statistical Solutions Ltd, Cork, Ireland). When (very conservatively) assuming that SD of the change in handgrip strength of perioperatively acquired weakness patients would be as high as in patients suffering from intensive care unit-acquired weakness (Patsaki et al³⁰ report an average decrease of 38% with 24% SD), this effect size estimate would correspond to a detectable difference in handgrip strength of 0.088 (9%), which is about the 10% change that we would consider clinically relevant. Numerical calculations were performed with IBM SPSS Statistics (IBM Corp, Armonk, NY), version 25.

RESULTS

Study Population and Patient Characteristics

Table 2. - Patient Characteristics, Intraoperative Data, and Outcome Parameters

	Measured (n = 89)	Analyzed (n = 59)	Analyzed 3-mo Follow-up (n = 23)
Age (y)	74 (69–77)	74 (69–77)	72 (69–76)
Male gender, n (%)	44 (49.4)	33 (55.9)	12 (52.2)
Body mass index (kg/m²)	26.5 (22.6–29.6)	26.8 (24.0–30.1)	29.8 (26.4–32.1)
Diabetes, n (%)	28 (31.5)	22 (37.3)	6 (26.1)
ASA physical status II/III (n)	42/47	30/29	12/11
Hypertension, n (%)	71 (79.8)	45 (76.3)	19 (82.6)
Diabetes, n (%)	28 (31.5)	22 (37.3)	6 (26.1)
Coronary and/or chronic heart disease, n (%)	30 (33.7)	17 (28.8)	6 (26.1)
Renal insufficiency, n (%)	22 (24.7)	16 (27.1)	4 (17.4)
Liver cirrhosis, n (%)	6 (6.7)	5 (8.5)	1 (4.3)
Cancer, n (%)	62 (69.7)	40 (67.8)	14 (60.9)
Type of surgery, n (%)^a
Intrathoracic	9 (10.1)	7 (11.9)	4 (17.4)
Intraabdominal	63 (70.8)	38 (64.4)	13 (56.5)
Other	17 (19.1)	14 (23.7)	6 (26.1)
Duration of surgery (min)	219 (131–360)	166 (116–256)	166 (108–290)
Total intravenous/volatile anesthesia (n)	16/73	11/48	5/18
Rocuronium/cisatracurium for intraoperative muscular blocking (n)	59/30	39/20	15/8
Intraoperative glucose maximum (mg/dL)	153 (133–184)	149 (127–183)	149 (121–181)
Intraoperative glucose mean (mg/dL)	131 (117–155)	129 (115–158)	129 (112–151)
Pain numeric rating scale on postoperative day 1	0 (0–3)	0 (0–3)	0 (0–5)
Epidural catheter, n (%)	42 (47.2)	27 (45.8)	10 (43.5)
Intensive care unit duration (h)	23 (0–103)	4 (0–23)	18 [0–26)
Hospital length of stay (d)^b	9 (6–20)	8 (5–10)	7 (5–9)
In-hospital death, n (%)	12 (13.5)	1 (1.7)	0

ata are presented in frequencies (%) or median (interquartile range). Patient characteristics were recorded upon enrollment.

bbreviation: ASA, American Society of Anesthesiologists.

See Supplemental Digital Content 6, Table 2, http://links.lww.com/AA/C741, for full list of types of surgeries.

Calculated after exclusion of the deceased patients.

Figure 1.:
Consolidated Standards of Reporting Trials diagram. As a part of the BIOCOG project, a subgroup of 89 patients received tight intraoperative blood glucose measurement, as well as a preoperative and postoperative weakness assessment. The term “ICU patients” includes those treated in an ICU, IMC or PACU. BIOCOG indicates Biomarker Development for Postoperative Cognitive Impairment in the Elderly; ICU, intensive care unit; ICUAW, ICU-acquired weakness; IMC, intermediate care unit; MRC, Medical Research Council; PACU, postanesthesia care unit.

From our subgroup investigation cohort of 89 patients, 9 patients were diagnosed with intensive care unit-acquired weakness at intensive care unit discharge, 19 additional patients had an extensive intensive care unit stay (≥72 hours), and 2 non-intensive care unit patients had to be subsequently admitted to the intensive care unit during the postoperative period. Therefore, the final analysis included a total of 59 patients (Figure 1). All patients were placed under general anesthesia. With 1 exception, all patients were extubated immediately after surgery. The patient who remained intubated (due to hypothermia) was extubated in the intensive care unit 5 hours after the end of surgery. In addition, a single patient required neostigmine for reversal of a residual muscular block, while no other patients required any reversal agents. Patient characteristics and outcome parameters are shown in Table 2, including data from the original 89 patients for transparency. The 3-month follow-up rate was only 39.0%.

Peripheral Muscle Function

Preoperatively, the handgrip strength of our patients corresponded to their expected normal values, adjusted for age and gender 97.9% (84.3%–114.6%) (expected values: 32.2 kg [23.5–39.1 kg]; measured values: 29.0 kg [21.5–38.5 kg]). The preoperative median Medical Research Council score was 60 (59–60), ie, no preoperative weakness was observed.

Considering all measured time points, handgrip strength showed significant differences (Friedman test, P = .006), whereas no differences occurred for Medical Research Council scores. The first postoperative handgrip strength was significantly decreased for all patients (median decrease 16.4% [0.0%–22.7%]; Wilcoxon test, P < .001). Handgrip strength remained reduced at intensive care unit discharge (median decrease compared to baseline 18.9% [1.4%–27.9%]; Wilcoxon test, P < .001). At hospital discharge, handgrip strength was still significantly reduced compared to preoperative levels (median decrease, 3.6% [2.1%–11.7%]; Wilcoxon test, P = .016). At the 3-month follow-up, levels of handgrip strength (median decrease, 11.1% [0.0%–20.5%]; Wilcoxon test, P = .012) were significantly lower compared to preoperative measurement. Trends of handgrip strength and Medical Research Council score are shown in Figure 2. Absolute values of handgrip strength are shown in Supplemental Digital Content 5, Table 1, http://links.lww.com/AA/C740.

Figure 2.:
(Left) Handgrip strength left over different time points. Fifty-one patients were measured preoperatively. Of these, 47 patients were postoperatively measured on POD1, 29 patients at ICU discharge, 45 patients at hospital discharge, and 23 patients at the 3-month follow-up. Box plots with additional mean (scattered line). (Right) MRC score over all time points. Fifty-two patients were measured preoperatively. Of these, 50 patients were additionally measured on POD1, 27 patients at ICU discharge, and 12 patients at the 3-month follow-up. Box plots with additional mean (scattered line), Friedman test over all time points with pairwise post hoc comparison using Wilcoxon signed-rank tests. ICU indicates intensive care unit; MRC, Medical Research Council; POD, postoperative day.

In the 7 patients with daily handgrip strength measurements, from postoperative day 1 until postoperative day 7, peak of weakness was seen on postoperative day 1, with a slight recovery until postoperative day 7 (see Supplemental Digital Content 1, Figure 1, http://links.lww.com/AA/C736). Similarly, the 7-patient group with daily Medical Research Council scores is shown in Supplemental Digital Content 2, Figure 2, http://links.lww.com/AA/C737. Overall, the duration of surgery correlated weakly with decreases of handgrip strength (r = 0.294; P = .053). Neither mode of general anesthesia (total intravenous versus volatile anesthesia) nor choice of intraoperative muscular blocking agents (rocuronium versus cisatracurium) had an influence on handgrip strength (Mann-Whitney U test, P = .490 and P = .562) or Medical Research Council decrease (Mann-Whitney U test, P = .602 and P = .181).

To avoid selection bias, course of handgrip strength was calculated for all 89 measured patients, including intensive care unit-acquired weakness and intensive care unit patients (stay ≥72 hours), showing similar results (see Supplemental Digital Content 3, Figure 3, http://links.lww.com/AA/C738). In addition, we found a significant correlation of surgical time and a decrease in handgrip strength (r = 0.331; P = .006) in this cohort. To analyze the influence of pain, we correlated the numeric rating scale at postoperative day 1 with handgrip strength decrease and found no correlation (r = 0.003; P = .985). Furthermore, no significant difference in handgrip strength decrease was found between patients with and without epidural catheter (Mann-Whitney U test, P = .424). Of note, no significant differences in handgrip strength were seen between different specialties (Kruskal–Wallis test, P = .118; see Supplemental Digital Content 6, Table 2, http://links.lww.com/AA/C741) and also not between intraabdominal/intrathoracic and other types of surgery (Mann–Whitney U test, P = .808).

Respiratory Muscle Function

Basic patient characteristics and outcome parameters of this cohort are shown in Supplemental Digital Content 7, Table 3, http://links.lww.com/AA/C742. The preoperative pulmonary test results of our patients were low compared to their expected age- and gender-adjusted values, with vital capacity at 87.5% (66.5%–98.3%; expected value, 3.7 L [2.4–4.1 L]; measured value, 2.7 L [2.1–3.5 L], P = .03) and FEV1 at 85.0% (79.0%–104.0%; expected value, 2.7 L [1.9–3.0 L], measured value, 2.1 L [1.7–2.4 L]; Wilcoxon test, P = .095).

Figure 3.:
VC and FEV1 preoperatively and postoperatively. Fourteen patients were assessed preoperatively. Of them, 12 patients were additionally measured on POD1. Box plots with additional mean (scattered line), Wilcoxon signed-rank test. FEV1 indicates forced expiratory volume in 1 second; POD, postoperative day; VC, vital capacity.

After surgery, vital capacity (median decrease 13.1% [2.7%–21.3%]; Wilcoxon test, P = .022) and FEV1 (median decrease, 12.6% [10.7%–21.7%]; Wilcoxon test, P = .001) decreased significantly (Figure 3), whereas the FEV1/vital capacity ratio remained unchanged (preoperatively, 76.0% [69.3%–86.8%]; postoperatively, 77.0% [68.4%–82.6%]; Wilcoxon test, P = .470). Duration of surgery correlated significantly with FEV1 decrease (r = 0.583; P = .029), but no correlation with vital capacity (r = 0.031; P = .917) could be observed. Neither mode of general anesthesia (total intravenous versus volatile anesthesia) nor choice of intraoperative muscular blocking agents (rocuronium versus cisatracurium) were significantly associated with vital capacity (Mann–Whitney U test, P = .659 and P = .659) or FEV1 (P = .198 and P = 1.000) decrease.