Magnesium is mainly used to treat pregnancy-induced hypertension, usually in the form of magnesium sulfate (MgSO₄) [1]. The use of magnesium to treat acute myocardial infarction, asthma, arrhythmia, and pheochromocytoma; and to induce postoperative analgesia; is increasing [2]. Despite the fact that magnesium is relatively safe hemodynamically [2,3], magnesium use is associated with serious complications in terms of neuromuscular blockade (NMB), including respiratory arrest [4]. Magnesium enhances the action of nondepolarizing neuromuscular blocking agents (NMBAs) [5]. The effective dose (ED)₅₀ of pipecuronium changed in inverse proportion to the increase in magnesium concentration in animal studies [6]. Magnesium shortened the time to NMB onset mediated by rocuronium at an effective concentration (EC) of effective concentration (EC)₉₅ [7]. In addition, magnesium prolonged the duration of action of rocuronium [8,9]. A study using isolated phrenic nerve diaphragm preparations from the mouse found that a higher dose of sugammadex was required to reverse rocuronium-induced NMB when the magnesium concentration was increased from 3.50 to 4.07 mM [10]. Recurarization can occur when magnesium is added after sugammadex reversal [11], but recurarization was reduced when the sugammadex dose was increased [10]. Previous studies led us to expect that the rocuronium effect would increase and the sugammadex-mediated reversal effect would decrease when magnesium was administered [8–10]. However, it remained unclear whether these effects would change in proportion to the magnesium concentration.

The normal level of magnesium in the human body is 0.7–1.1 mM and the therapeutic concentration used to treat pre-eclampsia is 2–4 mM [1]. The patellar reflex is inhibited by 3.5–5 mM magnesium and respiratory paralysis develops at magnesium levels greater than 6 mM [4]. In pre-eclampsia patients, the methods used to monitor the side-effects of magnesium therapy include measurement of blood magnesium concentration, the patellar reflex, and various clinical features [12].

We planned the present study to explore whether increased concentrations of MgSO₄, from the reference level (1 mM), which is normal in the human body to 2, 3, and 4 mM, proportionally enhanced NMB caused by rocuronium and hindered sugammadex-induced NMB reversal.

After approval was granted by Asan hospital Institutional Review Board for Animal Studies (no. 2016-13-067), 40 male Sprague-Dawley rats (219 [13] g, 43 [6] days of age) were randomly allocated to four groups with different magnesium concentrations (1, 2, 3, and 4 mM, each n = 10). The rats were anesthetized with 30 mg/kg intraperitoneal Zoletil^® (Virbac^®, France) and euthanized. The thoracic cages were excised en bloc and isolated left phrenic nerve hemi-diaphragms prepared. Each specimen was placed in a 75 ml organ bath containing Krebs buffer solution (pH 7.4: NaCl 118 mM, KCl 5.0 mM, CaCl₂ 2.5 mM, NaHCO₃ 30 mM, KH₂PO₄ 1.0 mM, MgSO₄ 1.0 mM, and glucose 11.4 mM) [13]. The bath was held at 35°C via outer warm water circulation and 95% oxygen and 5% carbon dioxide were continuously bubbled into the Krebs solution. The train-of-four (TOF) and twitch height responses were recorded mechanomyographically. The rib of each specimen was fixed to point downward, and the tendinous portion upward; the specimen was connected to a force displacement transducer^® (Grass FT03^®, Grass Instruments^®, USA) and preloaded to 2 g. A bipolar platinum electrode was connected to the phrenic nerve and stimulated with TOF pulses (2 Hz supramaximal stimulation) 0.2 ms in length (pulse duration) using a Grass S88 stimulator^® and an SIU5 stimulus isolation unit^® (Grass Instruments^®, USA). TOF stimulation was continued at intervals of 20 s until the end of the study. Muscle contraction was measured using the force displacement transducer and digitalized with the aid of the Powerlab^® acquisition system (Millar^®, USA). All signals were digitally saved using charting software (Lab Chart^®, ADinstruments^®, USA). After 30 min of stabilization, the magnesium concentrations were set to 2, 3, or 4 mM using MgSO₄ in the 2, 3, and 4 mM groups. MgSO₄ was not added to the 1 mM group because the magnesium concentration of the Krebs solution was 1 mM. After additional stabilization over 20 min, the T1 and train-of-four ratio (TOFR) after MgSO₄ administration were compared to the pre-administration values within each group. Then, the first dose of rocuronium (200 μg) was added to the organ bath. The T1 heights just before administration of rocuronium were considered the control values for each group. The rocuronium dose was increased in 100 μg increments until T1 height was completely suppressed in each group. Prior to every addition of rocuronium, we waited for 10 min or observed similar T1 height responses on at least three consecutive measurements. We then calculated the EC₅₀, EC₉₀, and EC₉₅ of each group.

Ten minutes after the T1 height was completely suppressed, equimolar doses of sugammadex (3.57-fold the total weight [mg] of rocuronium added) were in turn added. T1 height and TOFR recovery were observed for 30 min after administration of sugammadex; and we measured the following variables: maximum T1 height; the recovery index (the time to recovery from 25% to 75% of the control value of T1 height); the time taken for T1 height to attain 95% of the control T1 height (the 95% T1 time); the time taken for the TOFR to attain 0.9 (the TOFR 0.9 time); and the number of samples that failed to attain TOFR 0.9 and 95% of the control T1 height in each group.

For thirty minutes of initial stabilization in each group when the magnesium level was not increased, the T1 height and TOFR did not differ among the groups. However, after MgSO₄ administration and an additional 20 min of stabilization, T1 height exhibited significant decline in the 2, 3, and 4 mM magnesium groups (P = 0.002, 0.001, and < 0.001 respectively), but the TOFR did not (Table 1). As the concentration of MgSO₄ increased, the EC₅₀, EC₉₀, and EC95 significantly decreased (all groups P < 0.001), but pairwise comparisons showed no difference between the 3 and 4 mM magnesium groups (Table 2).

The primary outcome, the TOFR 0.9 time, did not differ among the groups after bolus sugammadex administration. T1 height recovery (maximum T1 height, recovery index, and the 95% T1 time), and the number of samples that failed to reach TOFR 0.9 or T1 height 95% of control did not differ among the groups after bolus sugammadex administration (Table 3).

In this study, TOFR 0.9 times after equimolar sugammadex administration were not affected by an increase in magnesium concentration. However, we recorded T1 height depression but no TOFR depression in the 2, 3, and 4 mM MgSO₄ groups compared to before magnesium addition. In addition, the EC₅₀, EC₉₀, and EC₉₅ decreased as the magnesium concentration increased from 1 to 4 mM, but we found no statistically significant change between 3 and 4 mM MgSO₄.

Magnesium induces NMB by reducing prejunctional acetylcholine (Ach) release, decreasing the excitability of the junctional nicotinic Ach receptor (nAchR) and the muscle membrane. Magnesium acts principally at the prejunctional membrane by competitively antagonizing calcium influx [14]. As magnesium reduces all four of the twitch heights of the TOF, magnesium has only a minor effect on TOF fading [15]. Therefore, magnesium causes T1 height depression, but not any significant change in the TOFR.

The effect of magnesium are dependent on its concentration [14,16]; NMB is enhanced when NMBAs are concurrently administered. When the magnesium concentration increased from 1 to 3 mM, the EC₅₀, EC₉₀, and EC₉₅ of rocuronium decreased by about 22–28% for each 1 mM increase (Table 2). In a pipecuronium study, when the magnesium concentration was increased to 1.2, 2, 2.75, and 3.5 mM, the ED₅₀ of pipecuronium continuously fell, but no change was found between 2 and 2.75 mM in the rat [6]. The atracurium requirement was not reduced, even though the magnesium infusion rate was increased during total intravenous anesthesia with propofol, in humans [17]. It is unclear whether magnesium concentrations within a certain range may exhibit a ceiling-like interaction with rocuronium and other nondepolarizing NMBAs [6,17]. Further study is needed.

Sugammadex effectively reverses rocuronium-induced NMB in patients undergoing magnesium therapy [15]. Elevated magnesium concentrations increase the levels of sugammadex required for reversal, and recurarization after injection of magnesium was reduced when the sugammadex dose was increased [10]. Therefore, we expected that addition of equimolar doses of sugammadex to baths that had received rocuronium in the presence of different concentrations of magnesium might be associated with different extents of recovery. However, neither the T1 height nor TOFR recovery differed among groups after equimolar doses of sugammadex were added. At the TOFR 0.9, 70% of postsynaptic receptors were still occupied by NMBAs [18,19]. This phenomenon may explain NMB recurrence after the administration of NMB-promoting agents such as magnesium [20]. Recurarization after magnesium administration may reflect insufficient dosing by sugammadex and residual NMBA. Rocuronium was completely removed in the present study because no pharmacokinetic variation was evident in the organ bath tests. Thus, changes in magnesium concentration (from 1 to 4 mM) did not seem to affect recovery mediated by sugammadex. This result is consistent with that of a rabbit study in which sugammadex at doses of more than 4 mg/kg was associated with similar recoveries among magnesium-infused and non-infused groups of rabbits [21]. It is likely that the dose of sugammadex is more important than the magnesium concentration in terms of clinically acceptable limits; neuromuscular transmission monitoring is essential to ensure that sufficient sugammadex is administered.

A TOFR of 0.9 is considered safe in terms of extubation and recovery [22]; we used the TOFR 0.9 time as the primary outcome. The primary outcome, the TOFR 0.9 time, did not differ among the groups after bolus sugammadex administration. In an earlier study of sugammadex reversal, a lower dose of sugammadex (≤ 4 mg/kg) after NMB induced by rocuronium of 1.2 mg/kg allowed recovery of T1 height faster than TOFR that is also similarly seen during spontaneous recovery and cholinesterase reversal, but the opposite result was reported when highdose sugammadex (> 8 mg/kg) was given [23]. We suspect that small amounts of free NMBAs not removed by sugammadex may have a greater affinity for postsynaptic nAchR than presynaptic nAchR early in the recovery period [24,25]. Thus, TOFR recovery may precede T1 height recovery in the early recovery period. Sufficient T1 recovery follows within minutes. A previous study also posed the question of whether this feature of sugammadex reversal was unchanged in the presence of medication that affects nAchR such as magnesium or an aminoglycoside [23]. T1 height recovery (the T1 95% time) was faster than TOFR recovery (the TOFR 0.9 time) in the present study (Table 3). This suggests that an equimolar dose of sugammadex may recover T1 height earlier than TOFR when a rocuronium-induced NMB is in play in the presence of magnesium.

A limitation of this study is that the recovery conditions were not identical among the groups. The rocuronium dose was the highest and the time available for observation prior to sugammadex administration the longest in the 1 mM magnesium group than the other groups. This may affect the recovery results. A further study may be needed to more accurately evaluate the effect of magnesium concentration on the required sugammadex level. With the presence of different concentrations of MgSO₄, the addition of increments of sugammadex followed by single-bolus administration of rocuronium would be useful to observe recovery. The second limitation is that half of the 1 mM group failed to reach T1 height 95% that of the control. Even though it is not statistically significant, some experimental errors may have occurred.

In conclusion, increases in the MgSO₄ concentration in isolated phrenic nerve hemi-diaphragmatic preparations from the rat did not affect TOFR recovery after addition of equimolar doses of sugammadex, but the NMB induced by rocuronium was enhanced in proportion to the magnesium concentration from 1 to 3 mM.