Early Paralysis for the Management of ARDS

Mechanisms by Which NMBAs Improve Survival in ARDS

ARDS continues to be a major source of morbidity and mortality. It affects approximately 190,000 patients annually in the United States and is associated with a mortality estimated at >40%.¹⁰ ARDS is an inflammatory disease process characterized by diffuse bilateral pulmonary edema, decreased lung compliance, and hypoxemia.⁶ The only well-established therapy for ARDS is the supportive use of lung-protective mechanical ventilation strategies aimed at minimizing further lung damage by limiting tidal volumes and decreasing mean airway pressures.⁹ Until recently, there has not been any specific pharmacologic therapy shown to improve mortality or outcomes in ARDS. In a randomized controlled trial of 340 subjects with severe ARDS, however, Papazian et al in the ACURASYS trial¹⁸ reported that the use of NMBAs early in the course of treatment decreased the duration of mechanical ventilation and improved survival. The major causes of death in patients with ARDS are severe hypoxemia and multi-system organ failure, secondary to infection, sepsis, hemodynamic instability, and ventilator-induced lung injury.⁶ It is important to examine how the use of NMBAs can influence this pathophysiology.

Reduced pulmonary compliance and diffusion limitations secondary to edema and inflammation at the alveolar-capillary level, which result in hypoxemia and hypercarbia, are hallmarks of ARDS. Without paralysis, hypoxemia and hypercarbia stimulate respiratory drive, which can lead to increased tidal volumes, active exhalation, and patient-ventilator asynchrony.⁶ With improved patient-ventilator synchrony with the administration of NMBAs, tidal volumes can be tightly regulated, thus decreasing the barotrauma and volutrauma caused by overdistention of alveoli from high tidal volume ventilation. Limiting active exhalation with paralysis also allows for better control over PEEP, thereby reducing atelectotrauma or injury due to repetitive opening and closing of lung units. Patient-ventilator asynchrony can be caused by several different mechanisms, including ineffective inspiratory effort during the exhalation cycle, double-triggering (breath-stacking), inappropriate cycling, and aborted inspiration.⁸ These asynchronies occur when patients are more interactive with the ventilator and oftentimes air-hungry, a common scenario for a patient with ARDS. The result of asynchrony is ineffective ventilation and increased airway pressures, both of which are detrimental to a lung-protective mechanical ventilation strategy. In an observational study, Blanch et al⁸ measured an asynchrony index in 50 subjects mechanically ventilated on various modes of mechanical ventilation. The asynchrony index was based on the number of asynchronous events a subject had with the ventilator in a given hour. They reported that an asynchrony index of <10% (compared with those >10%) was associated with decreased ICU and hospital mortality and a trend toward shorter duration of mechanical ventilation.⁸ NMBAs can eliminate this asynchrony by removing patient effort and allowing the ventilator to control the triggering and cycling of breaths.

Severe hypoxemia contributes to the mortality associated with ARDS. NMBAs improve oxygenation in ARDS. In a randomized controlled trial of 56 subjects with ARDS, Gainnier et al¹⁷ examined the effects of NMBAs on oxygenation parameters. Compared with placebo, subjects who received NMBAs for 48 h early in the course of ARDS had significantly improved P_aO2/F_IO2 after 48 h. This effect persisted for several days beyond discontinuation of the NMBA. The NMBA group also had faster weaning of PEEP, indicating improved lung compliance at the end of the 5-d study period.¹⁷ A separate randomized trial by Forel et al²¹ reported similar improvements in oxygenation with the use of NMBAs in early ARDS.²¹ The mechanism for improved oxygenation remains unclear. It may be a result of improved ventilation-perfusion relationships that occur with the initiation of muscle paralysis and the ability to maintain consistent levels of PEEP throughout all lung fields.⁶ There is some evidence that muscle paralysis or even deep sedation with a Ramsay score of <6 reduces the oxygen consumption by ventilatory muscles.^22–24 The reduction in oxygen consumption might decrease demand for oxygen delivery and carbon dioxide removal, ultimately resulting in decreased cardiac output and decreased pulmonary blood flow, which then may minimize the edema seen at the alveolar-capillary level in ARDS. What is interesting to note is that for both the Gainnier et al¹⁷ and Forel et al²¹ trials, the improvement in oxygenation persisted throughout the 5-d study period, well beyond the 48 h of muscle paralysis,^17,21 which suggests that the reduction in oxygen consumption cannot be the primary mechanism of improved oxygenation.

Another mechanism in which NMBAs could improve survival in ARDS may involve an anti-inflammatory effect. The lungs of a patient with ARDS are subject to overinflation and overstretching by mechanical ventilation, which may produce biotrauma, a regional and systemic inflammatory response that may generate or amplify multi-system organ failure. Causes of biotrauma include repetitive opening and closing of atelectatic lung units, surfactant alterations, loss of alveolar-capillary barrier, and bacterial translocation.²⁵ In animal models of ARDS, Imai et al²⁶ found that high tidal volume low PEEP mechanical ventilation can increase inflammatory markers and rates of apoptosis in the kidney and small intestine, which suggested a multi-system organ response to varying strategies of mechanical ventilation. The trial by Forel et al²¹ of NMBAs in ARDS evaluated inflammatory cytokine levels in the serum and bronchoalveolar lavage samples of subjects randomized to NMBA versus placebo. They reported that bronchoalveolar lavage levels of interleukin (IL)-8, IL-6, and IL-1B and serum levels of IL-6 and IL-1B were lower in the NMBA group at 48 h after randomization.²¹

One explanation of the efficacy of NMBAs in decreasing this inflammatory response is that paralysis induces greater homogeneity of the distribution of PEEP and tidal volume. This would reduce de-recruitment of dependent lung areas and minimize the repetitive opening and closing of atelectatic lung units and perhaps limit inflammatory mediator release normally associated with de-recruitment and atelectasis.²¹ Cisatracurium may also have a primary receptor-mediated anti-inflammatory effect on cells.⁶ The clinical importance of this effect, however, remains unclear at this point.

Outcomes of NMBA Administration in Early ARDS

So the question remains: If NMBAs can decrease inflammation, improve oxygenation, and improve patient-ventilator asynchrony in patients with ARDS, does this translate to improved outcomes in the ICU? To date, Papazian et al¹⁸ have performed the largest randomized controlled study, the ACURASYS trial, to help answer this question. They conducted a multi-center, placebo-controlled, double-blind study of 340 subjects with ARDS (defined as a P_aO2/F_IO2 of <150 with a PEEP of ≥5 cm H₂O and a tidal volume of 6–8 mL/kg ideal body weight) to determine whether a short period of treatment with cisatracurium early in the course of severe ARDS would improve clinical outcomes. The baseline characteristics between the treatment and control arms were similar with the exception that the cisatracurium group had a lower baseline mean P_aO2/F_IO2 ratio compared with placebo, 106 versus 115 (P = .03), respectively. The primary end point was 90-d mortality. Using the Cox multivariate regression model and adjusting for baseline P_aO2/F_IO2, plateau pressure, and Simplified Acute Physiology Score II, they found that the hazard ratio for death at 90 d in the cisatracurium group, as compared with placebo, was 0.68 (95% CI 0.48–0.98, P = .04). In crude analysis, the 90-d mortality rate in the cisatracurium group was 31% (95% CI 25–38%) compared with 40% in the placebo group (95% CI 33–48%), P = .08. The improved survival in the cisatracurium group was limited to the two thirds of subjects presenting with a P_aO2/F_IO2 of <120.

Secondary outcomes evaluated were ventilator days, ICU stay, number of organ or system failures, barotrauma, ICU-acquired paresis, and muscle strength as evaluated by the Medical Research Council scale. The cisatracurium group had more ventilator-free days than the placebo group during the first 28 and 90 d (10.6 d vs 8.5 d, P = .04, 53.1 d vs 44.6 d, P = .03) and more days spent outside of the ICU within the first 90 d compared with placebo (47.7 vs 39.5, P = .03). The cisatracurium group had a decrease in the number of days without organ failure compared with placebo when assessing coagulation abnormalities, hepatic failure, and renal failure, but these differences were not statistically significant. Pneumothorax occurred in a larger percentage in the placebo group (11% vs 4% in the cisatracurium group, P = .04). Before the development of pneumothorax in either study group, none of the subjects had an elevated plateau pressure necessitating changes in management. The incidence of ICU-acquired paresis as evaluated by the Medical Research Council scale on day 28 or at time of ICU discharge did not differ significantly between the 2 study groups.

Criticisms of ACURASYS Trial

Although the results of the ACURASYS trial are very encouraging, we must consider that, although the trial included 340 subjects, it is statistically under-powered to evaluate mortality. According to the authors, 885 subjects should be needed to achieve 80% statistical power.¹⁸ In addition, the survival benefit was limited to those subjects with a P_aO2/F_IO2 of <120, further limiting statistical strength. The study also has not been reproduced in other multi-center randomized trials and has not been attempted with other NMBAs besides cisatracurium.

Critics of the ACURASYS trial argue that if one of the major benefits of NMBAs is the improvement in patient-ventilator synchrony, then it would have been advantageous for the study to monitor and assess patient-ventilator interactions to determine a causal relationship. Hence, the relationship remains speculative. One can argue that the incidence of patient-ventilator asynchrony may not be as high as expected in the ARDS population.²⁷ Epstein²⁸ reported that only 30% of ARDS subjects suffered from ineffective ventilator triggering compared with subjects with respiratory failure from other mechanisms.^27,28 If the incidence of asynchrony is low, then intervention with NMBAs may not be advantageous. If asynchrony is problematic in a patient with ARDS, then other non-pharmacologic interventions might effectively improve patient-ventilator dynamics.

For example, Chanques et al,²⁹ evaluating mechanically ventilated subjects in the ICU, reported that 26% exhibited severe breath-stacking asynchrony with a mean asynchrony index of 44%. Management options included no intervention, increasing sedation/analgesia, or making a ventilator adjustment. Compared with baseline, the decrease in asynchrony index was greater after adjusting ventilator settings than after increasing sedation/analgesia (P < .001) or deciding to tolerate asynchrony (P < .001). The transition to pressure-support ventilation and increased inspiratory time were independently associated with a reduction in asynchrony index.²⁹ In the ACURAYS trial, all subjects were mechanically ventilated in a volume assist-control mode, and this mode was not adjusted for the sake of blinding. All subjects were sedated to a Ramsay score of 6 (no response to glabelar tap) before the initiation of treatment. If there were sustained episodes of elevated end-inspiratory plateau pressure > 32 cm H₂O, which can occur during episodes of asynchrony, then tidal volume and PEEP levels were adjusted, sedation was increased, and open-label cisatracurium was bolused, but the mode of ventilation was never changed.¹⁸ It is conceivable that inadequate management of asynchrony in the placebo group could have predisposed to worse outcomes.

Another concern in the ACURASYS trial is that depth of neuromuscular blockade was not monitored. The use of a train-of-four peripheral nerve simulator is generally considered a standard of care for managing and titrating the effect of neuromuscular blockade. Due to obvious blinding considerations, monitoring of train-of-four was not done in this study. As a result, adequacy and duration of neuromuscular blockade in the treatment group can be called into question. There are reports of tachyphylaxis and resistance to neuromuscular blockade in critically ill patients.^30,31 Thus, it is possible that the benefits observed in the cisatracurium group were independent of neuromuscular blockade.

Risks Associated With the Use of NMBAs

There are inherent risks to the use of NMBAs in patients in the ICU. Unrecognized ventilator disconnections can quickly lead to hypoxemia and hypercarbia and cause cardiopulmonary collapse. Inadequate sedation and analgesia in a paralyzed patient can cause extreme psychologic distress. As an example, Nelson et al³² investigated the relationship between the use of NMBAs during acute lung injury and the quality of life of survivors in a retrospective study of 24 subjects questioned after treatment in the ICU. Post-traumatic stress disorder symptoms were positively correlated with days of sedation and days of NMBA use but not with severity of illness. The inhibition of a cough reflex can lead to poor secretion clearance and mucus plugging that can cause increased airway pressures and hypoxemia if not effectively addressed. Patients are more at risk for corneal abrasion and ulceration given their inability to blink; thus, artificial tears should be applied intermittently, and eyes should be taped shut to prevent drying. During the period of paralysis, neurologic evaluation is not possible; therefore, recognition of neurologic deterioration during this time will be delayed.

Prolonged paralysis after discontinuation of NMBAs, defined as an increase in the time to recovery of 50 to 100% longer than predicted by pharmacologic parameters, is a problem encountered specifically after prolonged infusions on NMBAs and more commonly reported in the steroid-based NMBAs, including pancuronium, vecuronium, and rocuronium.^33–35 This is primarily due to the accumulation of NMBAs or their metabolites.³⁶ Steroid-based NMBAs undergo extensive hepatic metabolism, producing active drug metabolites, some of which can be 80% as potent as the parent compound. These active drug metabolites are renally excreted; thus, accumulation of parent drug and active metabolites is enhanced in the setting of hepatic and/or renal dysfunction.³⁶ Drug-drug interactions can also potentiate depth of motor blockade and may prolong recovery. Drugs that are implicated include local anesthetics, aminoglycoside antimicrobials, calcium channel blockers, magnesium, diuretics, lithium, and antiarrhythmics, including procainamide and quinidine.³⁶

Whether NMBAs contribute to ICU-acquired weakness is still under debate. What is known is that critical illness polyneuropathy and critical illness myopathy are responsible for significant morbidity in critically ill patients. Research for the last 2 decades has shown that critical illness polyneuropathy and myopathy can affect up to 50% of the ICU population. They present as limb and respiratory muscle weakness and, even with improvement, can cause varying degrees of disability after discharge from the hospital, from subtle weakness with limitation in exercise tolerance to profound disability.^37,38 Sepsis, systemic inflammatory response syndrome, and multi-organ failure are risk factors for critical illness polyneuropathy and myopathy. A systematic review of published work showed evidence of critical illness polyneuropathy and myopathy in 46% (95% CI 43–49) of adult ICU subjects who had lengthy mechanical ventilation, sepsis, or multi-organ failure.³⁹ Risk factors have been discussed in the literature. In a prospective cohort study of mechanically ventilated ICU subjects, De Jonghe et al⁴⁰ identified independent risk factors for the development of ICU-acquired weakness, which include female sex, multiple organ dysfunctions, duration of mechanical ventilation, and administration of corticosteroids. Duration of vasopressor support, duration of ICU stay, hyperglycemia, low serum albumin, and neurological failure have also been identified as risk factors.^41,42 With respect to NMBAs, the literature is contradictory. In a cohort study of 73 septic subjects with multi-organ dysfunction syndrome requiring mechanical ventilation, Garnacho-Montero et al⁴³ reported that 63% of subjects developed critical illness polyneuropathy as diagnosed by electrophysiologic studies. Of the subjects diagnosed with critical illness polyneuropathy, 18% of them received NMBAs during their ICU stay. After multivariate analysis, they determined that the use of NMBAs was independently associated with the development of critical illness polyneuropathy.⁴³ Griffiths and Hall⁴⁴ reported that simultaneous use of NMBAs and corticosteroids could be associated with muscle weakness, whereas NMBA use alone was not identified as an independent risk factor. Other studies, however, have not been able to support the relationship between NMBAs and development of critical illness polyneuropathy and myopathy.^{18,40,42,45,46} What we have learned from the literature is that immobility has profound effects on skeletal muscle and is a risk factor for muscle weakness during critical illness.³⁸ The use of NMBAs can certainly increase the duration of immobility, but the evidence to implicate it as an independent risk factor for prolonged disability is not definitive.

The Appropriate Use of NMBAs in ARDS Patients

The choice of NMBA and how it is utilized is critical. The ideal neuromuscular blocker would have a titratable effect, rapid onset and offset of paralysis to allow for neurological assessments, no adverse hemodynamics or physiologic adverse effects, hepatic and renal function-independent elimination, inactive metabolites, and reasonable cost. The available NMBAs in the United States are pancuronium, vecuronium, rocuronium, atracurium, and cisatracurium (Table 1). Since the combination of long-term high-dose corticosteroids and prolonged NMBAs, particularly the steroidal NMBAs, could increase the likelihood of developing an acute myopathy,^34,44 attempts should be made to avoid the steroidal NMBA (vecuronium, rocuronium, pancuronium) infusions. Vecuronium has active metabolites. Pancuronium, vecuronium, and rocuronium undergo hepatic metabolism and renal elimination to varying degrees. Both atracurium and pancuronium have adverse effects. Both atracurium and cisatracurium undergo Hoffman elimination, an organ-independent spontaneous degradation, which is temperature- and plasma pH-dependent. Cisatracurium is free of active metabolites and adverse effects, making it an appealing NMBA. To date, no study has definitively compared the use of different NMBAs in ARDS subjects.

In addition to the choice of NMBA, the method of utilization of NMBA can impact outcomes (Table 2). The use of NMBAs in ARDS should be rare, but when used, attempts should be made to minimize their dose and duration.⁴⁷ Initiation and duration should be limited to the first 48 h in severe ARDS with discontinuation as soon as possible. In an effort to limit NMBA dosing, depth of blockade should be guided by peripheral neuromuscular monitoring. The train-of-four, obtained from the peripheral neuromuscular monitor, goal is to achieve and maintain 1 to 2 twitches. In a prospective, randomized single-blind trial of 77 critically ill subjects receiving continuous NMBA, dosing guided by clinical response versus peripheral neuromuscular monitoring resulted in less drug per hour and total drug used. Subjects who underwent peripheral neuromuscular monitoring recovered neuromuscular function (relative risk 1.89) and spontaneous ventilation (relative risk 2.27, 95% CI of 1.23–4.21, P = .019) faster than control subjects.⁴⁸ When NMBAs are used, ensure adequate sedation and analgesia before neuromuscular blockade, frequent turning and pressure point padding, elevation of the head of bed >30 degrees, suctioning to clear secretions, close supervision and avoidance of ventilator disconnections, and application of eye care to avoid corneal abrasions (Table 3).