Introduction

Tubocurarine was introduced into anesthetic practice 70 years ago [1]. However, guidelines regarding dosage, monitoring of depth of block, and antagonism of the drug’s residual effects were slow to emerge. Perhaps the best clinical advice available in the first decade after tubocurarine’s introduction was contained in a 1953 article by Morris et al. [2]. To paraphrase, “At the end of surgery small doses of an anti-cholinesterase should be administered until ventilatory exchange seemed improved. Additional doses administered at 5 min intervals should be given until there is no detectable change for the better.” It gradually became obvious that clinicians needed a better way of monitoring the perioperative effects of neuromuscular blocking agents (NMBAs). In 1958, Christie and Churchill-Davidson [3] suggested that the indirectly evoked mechanical response to nerve stimulation might prove to be a practical clinical tool in the diagnosis of prolonged apnea after the use of muscle relaxants, and described a small battery-powered peripheral nerve stimulator (PNS) that they used for this purpose.

It was not until the mid-1960s that the effort to move PNS devices into the operating room gained some critical momentum. In 1965 two commercially available units were described [4, 5] and an editorial [6] opined, “The only satisfactory method of determining the degree of neuromuscular block is to stimulate a motor nerve with an electric current and observe the contraction of the muscles innervated by that nerve.” However, these early PNS devices did not incorporate modern parameters for testing the adequacy of neuromuscular function, such as train-of-four (TOF) stimulation [7]. Thus, it took more than another decade before the true incidence of postoperative residual neuromuscular block (PORB) began to emerge. In a now-classic article, Viby-Mogensen et al. [8] reported that 42 % of patients studied on arrival in their postanesthesia care units (PACU) had a TOF ratio less than 0.70, despite routine administration of neostigmine at the end of surgery. These findings were subsequently confirmed by other investigators [9, 10]. In these early studies, all patients received traditional long-acting relaxants. By the late 1980s, articles appeared that suggested that the incidence of PORB was considerably less with atracurium or vecuronium [11–14]. It was hypothesized, as intermediate-acting agents supplanted pancuronium and d-tubocurarine, that PORB would recede as a clinical problem. Unfortunately, these predictions did not materialize.

The reported incidence of PORB differs considerably among investigators, as different criteria may be used to define recovery (a TOF ratio less than 0.70, 0.80, or 0.90; TOF measured at the time of tracheal extubation or on arrival in the PACU; spontaneous recovery or anticholinesterase-assisted antagonism). Nevertheless, it is possible that as many as 20 % of patients arrive in the PACU with TOF ratios of 0.70 or less following the administration of an intermediate-duration NMBA, and twice as many arrive with TOF ratios less than 0.90 [15–23, 24•]. This is certainly true when pharmacologic reversal of neuromuscular block is omitted. The pertinent question becomes, does the presence of undetected residual weakness have clinical consequences? In 1989 Miller noted [25], “There are, in fact, no outcome data assessing the role of residual paralysis… when the occasional case of airway obstruction or inability to dispose of vomitus or oropharyngeal secretions occurs in the recovery room.” When another editorial in 2003 [26] suggested that “it is time to move from discussion to action and introduce objective neuromuscular monitoring in all operating rooms, not just those occupied by researchers and aficionados of muscle relaxants,” there was considerable backlash. One letter as a response to the editorial opined, “The editorial invokes ‘evidence-based practice,’ but the ‘evidence’ lies outside the realm of practice. Evidence should have clinical consequence. The newest mantra of medicine was ignored: outcome-based” [27]. These authors made a reasonable point—outcome data were missing. It is actually only in the past decade or so that investigators have produced convincing clinical outcome data substantiating the premise that PORB in the PACU is not always benign. This review is an attempt to collate these data and comment on their clinical and patient-safety significance.

However, despite the paucity of clinical outcome studies prior to the last decade, there was ample evidence to suggest that even modest to minor degrees of residual neuromuscular weakness might have untoward postoperative consequences. Most of these observations came from volunteer studies in which signs and symptoms of residual block are correlated with the TOF ratio at the adductor pollicis muscle. The latter parameter is a surrogate marker for PORB [28, 29]. An analysis of these surrogate data is, we believe, a prerequisite for placing available clinical outcomes in the proper context.

The TOF Ratio as a Surrogate Marker of PORB: A Brief Review

In 1975 Ali et al. [30] examined the respiratory correlates of the TOF ratio in an article that remains a classic in the neuromuscular literature. In eight volunteers, control vital capacity, peak expiratory flow rate, and the maximum negative inspiratory force were recorded. These tests stood as measures of mechanical respiratory reserve. Each volunteer was then given small incremental doses of d-tubocurarine until the TOF ratio, measured mechanomyographically, was 0.60 or less. The tests of respiratory function were then repeated at this level of block and at frequent intervals as spontaneous recovery progressed. At a TOF ratio of 0.60, vital capacity and peak expiratory flow rate were still 90 % or greater of the control values (see Table 1). On the basis of these, Ali et al. suggested that a TOF ratio of 0.60 or greater “would indicate that respiratory muscle function should be more than adequate for providing sustained spontaneous ventilation and pulmonary toilet.” Ali later revised this value upward to 0.70, because he determined that a TOF ratio of 0.70 or greater “reliably indicates the recovery of single twitch to control height and a sustained response to tetanic stimulation at 50 Hz for 5 s.” In addition, Ali et al. [31, 32] concluded that a TOF ratio of 0.70 correlated well with clinical signs of recovery in healthy patients. As a result of these seminal studies, and for almost the next two decades, a TOF ratio of 0.70 or greater was generally accepted as defining “adequate recovery” from nondepolarizing neuromuscular block. Sharpe et al. [33] in 1990 replicated the study of Ali et al. and concluded that a TOF ratio as low as 0.6, recorded from the hypothenar muscles in unanesthetized subjects, was consistent with near-normal respiratory function (i.e., the forced vital capacity and the forced expiratory volume in 1 s returned to 95 % of the control values, although the negative inspiratory force was still decreased by 28 %). A similar study by El Mikatti et al. [34] in 1995 reported essentially identical findings.

Table 1 The train-of-four (TOF) ratio as a measure of mechanical respiratory reserve
Full size table

However, by the mid 1990s the adequacy of a TOF ratio of 0.70 as a threshold of satisfactory neuromuscular recovery was beginning to be questioned. Isono et al. [35] demonstrated in awake volunteers that at a TOF ratio of 0.80 (a level at which there was minimal effect on hand grip strength) the electromyographic responses of the suprahyoid muscles were markedly depressed. They concluded that even this modest degree of peripheral block impaired the swallowing mechanism and significantly increased the risk of pulmonary aspiration. Eriksson et al. [36, 37] observed that a partial nondepolarizing neuromuscular block (TOF ratio of 0.70) impaired the hypoxic ventilatory regulation, suggesting an effect of vecuronium on carotid body chemosensitivity. The Stockholm group [38, 39] was also able to demonstrate that partial paralysis caused pharyngeal dysfunction and increased the risk of aspiration at TOF ratios less than 0.90 (measured mechanomyographically at the adductor pollicis muscle).

At about the same time, Kopman et al. [40] attempted to correlate clinical signs and symptoms associated with mild nondepolarizing block (TOF ratios of 0.70–0.90) in ten healthy volunteers who received a mivacurium infusion. No volunteer required intervention in order to maintain a patent airway, and peripheral arterial oxygen saturation while breathing room air was 96 % or greater at all times. However, a TOF ratio of 0.70 was accompanied by significant signs and symptoms of weakness. None of the volunteers considered themselves even remotely “street ready” at this level of neuromuscular block. Grip strength was decreased to 57 ± 11 % of the control level, and only one volunteer could retain a wooden tongue depressor between his teeth when the investigator attempted to remove it (using only two fingers on the blade). Several volunteers could not sip water through a straw because they could not make a tight seal around it with their lips. Most volunteers had some difficulty in swallowing, and all experienced diplopia and difficulty in following a moving object with their eyes at this level of block. Interestingly, all volunteers could sustain a 5-s head lift once the TOF ratio was at least 0.75. The tongue depressor test, conversely, was not passed until the TOF ratio had on average returned to 0.85. As the level of block receded, so did symptoms, but even at a TOF ratio of 0.90, grip strength was depressed by an average of 16 %, and all volunteers still reported visual disturbances. Eikermann et al. [41] in a study of 12 volunteers maintained at a TOF ratio of 0.83 ± 0.06 observed that seven of their subjects described difficulty in swallowing, and five subjects were unable to maintain a seal around a mouthpiece used to measure respiratory function.

More recently, Heier et al. [42••] performed a study that in large measure duplicated the earlier protocol of Kopman [40]. However, they induced a deeper level of block (a TOF ratio of 0.40) in their unanesthetized volunteers, and measured their vital capacity and forced vital capacity. At an average TOF ratio of 0.42, the mean handgrip strength was approximately 20 % of the baseline value, and the vital capacity was decreased by 26 ± 7 % from the control value. Responses to clinical tests of muscle function, however, differed greatly among volunteers at comparable TOF ratios. One volunteer lost the ability to clench the teeth at a TOF ratio of 0.84, but three others retained this ability at a TOF ratio of 0.30. None of the volunteers had significant clinical residual effects of neuromuscular block at a normalized acceleromyographic TOF ratio greater than 0.90.

On the basis of these recent data, there is now a general consensus that adequate recovery from nondepolarizing neuromuscular block is not assured until the TOF ratio (measured by mechanomyography or electromyography) at the adductor pollicis muscle is equal to or greater than 0.90. When acceleromyography (AMG) is used, some investigators [43] suggest that a value of 1.00 should be used; however, normalized AMG values of 0.90 are equally acceptable [44, 45].

PORB: Contributing Factors

Although the benchmark for defining satisfactory recovery from nondepolarizing neuromuscular block (a TOF ratio of 0.90 or greater) is now well established, most clinicians are ill equipped to measure this parameter. Traditional bedside or clinical tests of recovery such as the 5-s head lift or tidal volume are grossly inadequate for the task, and have been thoroughly discredited [9, 36, 46]. In addition, clinicians should be able to assess neuromuscular function prior to a patient’s emergence from anesthesia in order to determine the timing of tracheal extubation. An awake and cooperative patient should not be required to evaluate adequacy of recovery.

Although conventional PNS are useful intraoperatively to assess the depth of moderate or deep neuromuscular block (i.e., for measuring the TOF count), subjective (visual or tactile) detection of fade becomes very uncertain at TOF ratios greater than 0.40 [47]. Double-burst stimulation is somewhat more sensitive to tactile and visual assessment [48], but even with this stimulation pattern, fade will be missed 50 % of the time at TOF ratios of approximately 0.60 [49]. Sadly, recent surveys suggest that even this basic monitor is far from being used universally [50–52]. Quantitative or objective monitors that can accurately measure the TOF ratio in real time are now commercially available. Unfortunately, these devices appear to be available to no more than 10 % of clinicians, and this figure may be overly optimistic [43, 53].

Other factors probably play a role in the continued occurrence of PORB as well. A high level of unwarranted optimism exists with regard to the speed with which NMBAs of intermediate duration undergo spontaneous recovery. Reliance on the time interval since the last administered dose of relaxant to determine the need for or the dose of pharmacologic reversal is unreliable. Despite claims of 30–40-min clinical duration [54], 2 h after administration of a single intubating dose of vecuronium, TOF ratios less than 0.50 are not rare [55], and 30 % of patients will not have a achieved a TOF ratio of 0.90 [20].

Anticholinesterase-induced reversal of intermediate-duration NMBAs also has limitations, and may contribute to the occurrence of residual weakness. In 1995 Beemer et al. [56] reported that in practical terms, the maximum depth of neuromuscular block that can be antagonized effectively by anticholinesterases approximately corresponds to the reappearance of the fourth response to TOF stimulation (threshold TOF count of 4). However, it should be noted that he viewed a TOF ratio of 0.70 or greater as representing satisfactory recovery. Attempted reversal at a TOF count of 1–2 may restore the TOF ratio to 0.4–0.6 relatively quickly (5–10 min). However, this rapid initial pharmacologic recovery is followed by a prolonged plateau phase during which spontaneous elimination of the neuromuscular blocking drug becomes the major determinant of recovery [57]. Thus, we may have an extended “window of vulnerability” in which subjective detection of residual block is unlikely but return of adequate neuromuscular function has not yet occurred (TOF ratio of 0.40–0.90) [58].

It is not surprising, therefore, that large numbers of patients still arrive in the PACU with undetected residual neuromuscular block. However, as noted earlier, the exact incidence of PORB is unclear, as the reported frequency is highly dependent on the details of an investigator’s protocol. Even if we take a conservative approximation and assume that only 15–20 % of patients arrive in the PACU with TOF ratios less than 0.90, this still represents a very large number of patients at risk. The latest data available from the US Department of Health and Human Services (Centers for Disease Control and Prevention) [59] report an estimated 53.3 million surgical and nonsurgical procedures performed in the USA. If 25 % of these cases require general anesthesia and the use of NMBAs, then more than one million patients may be at risk of residual neuromuscular weakness. Hence, the question: What is the evidence that undetected residual block has adverse clinical consequences?

Clinical Outcome Studies

In the decade following the availability of nondepolarizing relaxants of intermediate duration (atracurium and vecuronium), evidence began to appear that strongly suggested that the choice of neuromuscular blocker administered had practical consequences. Ballantyne and Chang [60] in a retrospective study of 270 patients who received NMBAs compared the recovery times (PACU admission-to-discharge duration) for traditional long-acting agents (pancuronium or d-tubocurarine) with those for short- and intermediate-duration agents (mivacurium, atracurium, and vecuronium). They found that the 75 patients receiving pancuronium and d-tubocurarine took an average of 56 min longer to recover than the 195 patients who received atracurium, vecuronium, or mivacurium. Several years later, Murphy et al. [61] confirmed these findings. They compared recovery times (pancuronium vs rocuronium) in 70 randomly selected patients. Forty percent of patients in the pancuronium group had TOF ratios less than 0.7 on arrival in the PACU, compared with only 6 % of patients in the rocuronium group (P < 0.001). Significant delays in meeting discharge criteria were observed in the pancuronium group.

A recent study by Butterly et al. [62] using only agents of intermediate duration confirmed that the length of stay in the PACU was significantly longer for patients with TOF ratios less than 0.9, compared with patients with adequate recovery of neuromuscular transmission (defined as TOF ratio greater than 0.90). Thus, it is not the agent per se that is important in determining total recovery time in the PACU; it is the extent of residual block (the degree of recovery) that the patient manifests on arrival in the PACU that is of consequence.

Hypoxia is also commoner in patients exhibiting PORB on arrival in the PACU. Bissinger et al. [63] compared patients who had received pancuronium or vecuronium, and described a 20 % frequency of PORB in the pancuronium group, but only 7 % in those patients given vecuronium. For a recovery threshold of a TOF ratio greater than 0.70, the incidence of residual weakness in the PACU was 45 % in patients receiving pancuronium, and 8 % in patients receiving vecuronium. The incidence of hypoxia (defined as a peripheral oxygen saturation greater than 5 % below preanesthetic levels) was six times higher in the pancuronium group (60 vs 10 %). Murphy et al. [64] subsequently confirmed that small differences in residual block have consequences. They studied 185 patients who were randomized to intraoperative acceleromyographic monitoring (AMG group) or qualitative TOF monitoring (PNS group). The tracheas of PNS patients were extubated when standard criteria were met and no fade (tactile/visual) was observed during TOF stimulation. AMG patients had a TOF ratio greater than 0.80 as an additional tracheal extubation criterion. On arrival in the PACU, TOF ratios of both groups were measured with AMG. Thirty percent of the patients in the PNS group arrived in the PACU with TOF ratios less than 0.90, compared with only 5 % in the AMG group. In the AMG group, none of the patients exhibited arterial saturations below 90 % during transport or arrival in the PACU. In the PNS group, the incidence of desaturation during transport and in the PACU was 21 and 10 %, respectively. A more recent study by Sauer et al. [65••] provides further evidence that residual block of even very modest degree can have adverse clinical implications. A total of 114 patients were randomized to receive pharmacologic reversal with either neostigmine (0.02 mg/kg) or placebo at the end of surgery. Both patient groups received rocuronium (0.60 mg/kg) to facilitate tracheal intubation, plus small maintenance doses to keep the TOF count at 2 intraoperatively. Patients had AMG monitoring on one arm and conventional PNS devices on the contralateral arm. Recovery was allowed to continue spontaneously until no visual/tactile fade to TOF stimulation was detected. In the neostigmine group, objective measurement of the TOF ratio was then begun, neostigmine was administered, and tracheal extubation was delayed until the AMG-measured TOF ratio was 1.0. In the placebo group, tracheal extubation was performed at the clinician’s discretion; the median TOF ratio at extubation was 0.70 (0.46–0.90). On arrival of patients in the PACU there was no difference (14 vs 16 % each group) in the incidence of severe hypoxia (peripheral oxygen saturation below 90 %). However, 75 % of the patients in the neostigmine group had oxygen saturation greater than 93 %, but only 49 % of the placebo group patients achieved this level of oxygenation (P < 0.02). It is possible that the postoperative hypoxia observed in this study may have had nothing to do with residual neuromuscular weakness, since at a TOF ratio of 0.70, most measures of mechanical respiratory reserve (with the possible exception of maximum negative inspiratory force) have returned to more than 90 % of the control value [28]. However, as reported previously, the difference in oxygenation may be due to the direct effect of nondepolarizing blockers on carotid body chemosensitivity [34].

In a very large prospective study in 1992, Pedersen et al. [66] demonstrated a relationship between anesthesia involving pancuronium and postoperative pulmonary complications (POPC)—especially when pancuronium was administered during long surgical procedures. A more widely cited study by Berg et al. [67] drew similar conclusions 5 years later. A total of 691 adult patients undergoing general anesthesia were randomized to receive pancuronium, atracurium, or vecuronium. On arrival of patients in the PACU, TOF ratios were measured mechanomyographically. The incidence of residual block, defined as a TOF ratio less than 0.7, was significantly higher in the pancuronium group (26 %) than in the atracurium/vecuronium groups (5.3 %). In the pancuronium group, significantly more patients with residual block developed POPC (16.9 %) as compared with patients admitted to the PACU without residual block (4.8 %). In the atracurium/vecuronium groups, the incidence of POPC was not significantly different in patients with (4.2 %) or without (5.4 %) residual block. Multiple regression analysis indicated that a TOF ratio less than 0.7 following the use of pancuronium was a potential risk factor for development of POPC.

Perhaps the most convincing outcome study regarding the clinical implications of PORB, by Murphy et al. [68], appeared in 2008. The aim of their investigation was to quantify the extent of neuromuscular block in patients with signs or symptoms of critical respiratory events (CRE) in the PACU. A CRE was defined as fitting one or more of the following criteria: (1) upper airway obstruction requiring an intervention; (2) mild–moderate hypoxemia [O2 saturation (SpO2) of 93–90 %] on 3 L/min nasal cannula O2 that was not improved after active interventions (such as increasing O2 flows to more than 3 L/min, application of high-flow face mask O2, verbal requests to breathe deeply, tactile stimulation); (3) severe hypoxemia (SpO2 < 90 %) on 3 L/min nasal cannula O2 that was not improved after active interventions; (4) signs of respiratory distress or impending ventilatory failure (respiratory rate of more than 20 breaths per minute, accessory muscle use, tracheal tug); (5) inability to breathe deeply when instructed by the PACU nurse; (6) patient complaining of symptoms of respiratory or upper airway muscle weakness (difficulty breathing, swallowing, or speaking); (7) patient requiring tracheal reintubation in the PACU; and (8) clinical evidence or suspicion of pulmonary aspiration after tracheal extubation. Murphy et al. collected data over a 1-year period. PACU nurses identified patients with evidence of a predefined CRE during the first 15 min of PACU admission. TOF ratios were immediately quantified in these patients using AMG (cases group). TOF data were also collected in a control group that consisted of patients undergoing a general anesthetic during the same period who were matched with the case group by age, sex, and surgical procedure. Of the 7,489 patients who received a general anesthetic during the 1-year period, 61 developed a CRE. Forty-two of these patients were matched with controls and constituted the study group for statistical analysis. The mean (±standard deviation) TOF ratio was 0.62 (±0.20) in the case group, with 74 % of patients having TOF ratios less than 0.70. In contrast, the mean TOF ratio in the matched controls was 0.98 (±0.07), and no matched control patients had TOF ratios less than 0.70. Murphy et al. concluded that “a high incidence of severe residual blockade was observed in patients with CREs, which was absent in control patients without CREs. These findings suggest that incomplete neuromuscular recovery is an important contributing factor in the development of adverse respiratory events in the PACU.”

However, these results need to be placed in clinical context. Assume that the incidence of PORB (a TOF ratio less than 0.90) on arrival of patients in the PACU is 20 %. Thus, over 1,400 of the patients of Murphy et al. likely had some degree of PORB but did not suffer a noticeable adverse respiratory event. Stated differently, the data of Murphy et al. suggest that PORB is associated with a frequency of short-term CREs of perhaps only 4 or 5 %, and the incidence of actual long-term morbidity is likely to be even less than that. Unfortunately, the data of Murphy et al. also demonstrate that not all patients may be so lucky. Eight of the 7,459 patients (0.1 %) studied by Murphy et al. required emergent tracheal reintubation in the PACU owing to respiratory insufficiency. Although the risk of this complication is low, few clinicians would argue that this complication, however rare, is trivial. Extrapolation of these data nationally may suggest that over 20,000 patients may undergo emergency tracheal reintubation in the PACU owing to residual neuromuscular weakness (0.1 % of 20 million annual surgical procedures in the USA). Nevertheless, most patients seem to tolerate (or at least survive) residual block of modest extent. It is perhaps for this reason that many clinicians fail to appreciate that residual block may have (significant) adverse clinical consequences.

An ambitious case–control study involving 869,483 patients undergoing anesthesia between 1995–1997 in the Netherlands should give us pause [69]. “Cases” were patients who either remained comatose or died during anesthesia or within 24 h of undergoing anesthesia. “Controls” were patients who neither remained comatose nor died during anesthesia or within 24 h of undergoing anesthesia. Data were collected by means of a questionnaire, the anesthesia and recovery form. Odds ratios were calculated for risk factors, adjusted for confounders. The authors reported that reversal of muscle relaxants at the end of anesthesia (odds ratio, 0.10) had a major positive effect that was associated with reduced perioperative mortality within 48 h after surgery and anesthesia. The study is important because it suggests that anesthetic management can dramatically reduce perioperative mortality. If one reads between the lines, this study implies that undetected PORB may have deadly, if rare, consequences.

On a less somber note, there are other reasons to avoid even modest levels of residual block. As numerous volunteer studies have indicated, even TOF ratios as high as 0.90 do not guarantee the absence of symptoms of residual weakness [38, 40]. Patient satisfaction as well as safety should be the clinician’s goal. Subjective symptoms such as “general weakness,” blurry vision, difficulty in speaking, and difficulty in swallowing are common in the PACU. Recent evidence confirms that as the TOF ratio recovers, these symptoms diminish and the “quality” of the PACU experience improves [70•].

Conclusion

It has been three decades since Viby-Mogensen et al. [8] alerted the anesthesia community to the fact that undetected PORB in the PACU was not a rare occurrence. Nevertheless, clinicians do not seem to have internalized this truism. How else are we to explain the fact that (1) neuromuscular monitors are so sporadically used in day-to-day practice [46–49] and (2) recent editorials [71–73] and review articles [74, 75••, 76••, 77••, 78] continue to appear warning that PORB really does have clinical consequences?