Massive metoprolol overdose requiring ECMO: brief review of the evidence behind recommended treatments

Abstract

A man in his late 60s developed shock after ingesting 7500 mg of metoprolol tartrate that was refractory to all medical treatment including hyperinsulinaemic euglycaemia, intravenous lipid emulsion and dialysis, eventually needing rescue extracorporeal membrane oxygenation. A brief review of the recommended treatments in beta-blocker overdose is therefore warranted.

Background

Beta-blocker (BB) overdose, whether intentional or unintentional, is not uncommon. Glucagon, intravenous calcium, inotropes, hyperinsulinaemic euglycaemia (HIE), intravenous lipid emulsion (ILE) therapy and dialysis are current recommendations in the treatment of BB overdoses. Evidence supporting the efficacy of glucagon and calcium supplementation in massive BB overdoses is sparse. Evidence for HIE and ILE have been mixed. Also, dialysis has been shown to remove water-soluble BBs like atenolol. Metoprolol is unique pertaining to its pharmacokinetic properties in comparison with other BBs. Here, we present a 69-year-old man who developed shock after ingesting 7500 mg of metoprolol tartrate that was refractory to all medical treatments including HIE, ILE and dialysis, eventually needing rescue veno-arterial extracorporeal membrane oxygenation (VA-ECMO).

Case presentation

A man in his late 60s presented with a history significant for anxiety, bipolar depression and suicide attempts. He was brought in by ambulance at 06:30 after an overdose of his home medications, which included 100 tablets of 75 mg metoprolol tartrate and possibly lorazepam and hydroxyzine. He also has a history of coronary artery disease with elective coronary bypass grafting in 2015 (normal ejection fraction (EF) reported at the time), hypertension and hyperlipidaemia. His family found him unresponsive at home in the early hours of the morning of admission with empty medication bottles of the above medications and called Emergency Medical Services (EMS). The timing of alleged overdose is unclear but was thought to be between 6 and 8 hours prior to admission. The amount of metoprolol tablets consumed was deduced from when it was refilled to the date of overdose. Lorazepam and hydroxyzine prescriptions were on as-needed basis and the amount taken is entirely unclear, although the refills were marked for every 2 weeks and the most recent one was more than 2 months ago. He has no history of alcohol abuse or recreational use of any drugs.

Investigations

ECG showed normal sinus rhythm with normal QRS and QT intervals. His non-invasive cuffed blood pressure (BP) recorded was 63/57 mm Hg, heart rate (HR) 73 beats/min, temperature 35.6°C, respiratory rate 10/min and SpO₂ 70%. His White cell count on admission was 17.6/mm³ with no left shift, haemoglobin (Hb) 142 g/L and platelets 242×10⁹/L. His sodium was 139 mmol/L, potassium 5.4 mmol/L, chloride 108 mmol/L, bicarbonate 23 mmol/L, blood glucose 192 mg/dL, non-ionised calcium 6.9 mg/dL and normal liver transaminases. Urinalysis was normal with no evidence of infection. His serum alcohol, acetaminophen and salicylate levels were undetectable. Total creatine kinase was normal as were his cardiac enzymes which included creatine kinase muscle bound and troponin I. Chest X-ray (CXR) post-intubation showed adequate tube positioning with no pneumonic consolidation or suspicious infiltrate. Urine drug screen was positive only for benzodiazepines. Coagulation profile was normal. His post-intubation blood gas showed normocapnia with normal pH. Lactic acid was 1.4 mmol/L. He was transferred to the medical intensive care unit (ICU) for further management.

Treatment

When he arrived in the emergency department, his Glasgow coma score (GCS) was 7/15. He was minimally responsive, and his breathing appeared shallow. He was emergently intubated for airway protection. Central (right internal jugular (IJ)) and arterial lines (right femoral) were inserted. He was given 2 L of intravenous crystalloid and started on low-dose norepinephrine (NE). His ventilator settings included a tidal volume of 600 mL (8 cc/kg), Positive end expiratory pressure (PEEP) e 8 mm Hg, respiratory rate 18/min and FiO₂ 100%. His SpO₂ improved to 100% and so did his BP to 114/53 mm Hg, with heart rate 67 beats/min. In the ICU, his FiO₂ was titrated down to 50% and sedation was prescribed on as-needed basis to maintain an Richmond agitation sedation scale (RASS) of 0. However, a few hours later, patient’s BP decreased with a mean arterial pressure (MAP) of around 60 and HR down to beats 44/min, still in normal sinus rhythm. Bedside echocardiogram (ECHO) showed a significant drop in EF to <20%. NE was titrated up to 0.1 mcg/kg/min to maintain a HR of greater than 60/min and MAP greater than 65. One milligram of atropine was given to no effect. He was given two boluses of 5 mg intravenous glucagon, in response to which his HR increased to 60 beats/min with no improvement in his BP. He was started on glucagon infusion at 1 mg/hour that was titrated up to 5 mg/hour to sustain the response. After about 12 hours on glucagon infusion, his MAP was close to 60 mm Hg and his HR 46 beats/min and still in sinus rhythm. Glucagon was discontinued, NE infusion was titrated up (2 mcg/kg/min) and 1 g of calcium chloride was given. His BP responded and MAP increased to 68 mm Hg. This response was sustained for about 5 hours before two further doses were given at similar intervals. With his poor EF he was subsequently started on epinephrine (EPI) infusion (0.2 mcg/kg/min) after being on max dose NE (3 mcg/kg/min). With his MAPs in the low 60s, he developed acute kidney injury associated with metabolic acidosis. His lactate increased to 4.6 mmol/L. He was started on a bicarbonate infusion and nephrology consulted for impending dialysis. His EPI infusion was now increased to 1 mcg/kg/min.

Hyperinsulinaemic euglycaemia (HIE) was started with insulin infusion starting at 1 unit/kg/hour with 20% dextrose immediately. This was titrated up to 5 units/kg/hour carefully monitoring his blood sugar. No hypoglycaemic episodes occurred. After about 2.5 hours with no immediate effect from HIE and further deterioration, 20% intravenous lipid emulsion (ILE) was initiated. With both HIE and ILE running for over 8 hours, patient’s haemodynamics continued to deteriorate with the HR dropping to 26 beats/min. He was found to be in complete heart block, prompting placement of a temporary percutaneous pacemaker via his left femoral vein under fluoroscopy. The pacemaker rate was set to 80 beats/min with good capture. The patient’s MAP, however, remained tenuous between 55 and 58 mm Hg requiring the addition of vasopressin (VASO 0.04 mcg/kg/min) to keep his MAP close to 65 mm Hg. His urine output continued to drift down, and his creatinine peaked at 3.9 mg/dL. His O₂ requirements increased to 100% and his CXR showed bilateral pulmonary infiltrates suggestive of pulmonary oedema. His ventilator settings were adjusted to a PEEP of 10, with caution being given to pressor requirements. Continuous renal replacement therapy (CRRT) was started for treatment of volume overload with net negative 50–100 cc/hour and also hopes of dialysing metoprolol. His repeat blood gases despite CRRT for more than 12 hours showed a pH 7.10, Pco₂ 63, PO₂ 36 on 100% FiO₂ and maximum pressor/inotropes (NE 3 mcg/kg/min, EPI 2 mcg/kg/min and VASO 0.04 mcg/kg/min). The patient was drowsy but was able to open his eyes and squeeze his fingers when asked to. Neurologically being intact, and after discussion with his family members, VA-ECMO using centrifugal pump (Maquet, Hirrlingen, Germany) was initiated at the bedside after consulting cardiothoracic surgery. Ten thousand units of heparin was given during cannulation. Right femoral arterial cannulation (via right arterial line) with 18Fr (length 15 cm), right femoral venous cannula 25Fr (length 38 cm), uncoated circuit tubing and a polymethylpentene membrane (Maquet, Hirrlingen, Germany) oxygenator were used. Flows were gradually increased to 4 L/min. Within 30 min, his MAP was close to 70 mm Hg. His SpO₂ was 100% on pulse oximetry. Heparin infusion was started to maintain activated clotting time close to 200. Because of the lack of overnight perfusionists at our hospital, he was transferred to our main campus on ECMO.

Outcome and follow-up

At the tertiary facility, good flows were maintained at 5.54 L/min, MAP 78 mm Hg, FiO₂ 100% and sweep 3 L/min. His tidal volume was set to ultraprotective settings at 2–4 cc/kg. Arterial blood gases showed a pH 7.36, Pco₂ 32 mm Hg, PO₂ 329 mm Hg and HCO₃ 18 mmol/L on 80% FiO₂. Poor circulation was noted in the right extremity that required a distal perfusion catheter (DPC). With steady MAPs maintained on VA-ECMO, his pressor/inotropes doses significantly reduced by greater than 50% in less than 24 hours. CRRT was resumed but was stopped the following day because the patient’s urine output improved to 50–75 cc/hour. Day 3 VA-ECMO was switched to veno-venous extracorporeal membrane oxygenation (VV-ECMO) while on minimal pressor (NE 0.02, EPI 0.08, VASO 0.02) using the left IJ vein. The placement was confirmed with a transoesophageal ECHO and in the process his EF was noted to be 60%. (Estimated Glomerular Filtration Rate (EGFR) was 85 mL/min/1.73 m² with an improved creatinine of 1.16 mg/dL. By day 7, his VV-ECMO was discontinued and he remained off pressor. However, he remained on mechanical ventilation because of hyperagitated delirium and recurrent mucus plugging requiring aggressive pulmonary toilet and bronchoscopies. He failed multiple Spontaneous breathing trials (SBT) trials but was eventually extubated to 6 L nasal cannula after being on the ventilator for 15 days. He was subsequently transferred to inpatient rehabilitation and then to inpatient psychiatry on room air 6 days after extubation. A timeline of the treatments instituted during patient’s hospital stay is depicted in figure 1.

Figure 1.

Timelines of various treatments instituted during patient’s hospital stay. CRRT, continuous renal replacement therapy; HIE, hyperinsulinaemic euglycaemia; ILE, intravenous lipid emulsion; VA-ECMO, venoarterial extracorporeal membrane oxygenation; VV-ECMO, veno-venous extracorporeal membrane oxygenation.

Discussion

BB overdose, whether intentional or unintentional, is not uncommon.¹ Massive overdoses of any BB are sometimes difficult to treat and may even be resistant to the current recommended treatments. Much of the treatments are focused on either promoting the cyclic-AMP (c-AMP) pathway or expediting its elimination from the tissues. Beta-adrenergic receptors (β-ARs) in the heart are predominantly β1 receptors located in the contracting myocytes and nodal tissues. These β-ARs are coupled to a Gs-protein that when activated by NE or EPI activate adenyl cyclase to form cAMP from ATP. c-AMP then activates c-AMP-dependent protein kinases that phosphorylates L-type calcium channels causing calcium to enter the cell. This further causes release of calcium from the endoplasmic reticulum into the cell binding to troponin C, increasing inotropy and chronotropy. BB essentially blocks the action of NE and EPI. Depending on receptor selectivity, BBs are further divided as cardioselective and non-cardioselective² (see table 1). Cardioselective BBs predominantly act on β1 receptors. Non-cardioselective BBs have effect on β1 and β2 receptors, and some have effect on α1 receptors in addition to β1 and β2. β2 and α1 receptors are located in the vascular smooth muscle. β2 receptor stimulation causes vasodilation and α1 causes vasoconstriction. Significant cardioselective BB overdoses in general therefore decrease HR, cardiac output (CO), cardiac index (CI), stroke volume (SV) and some non-cardioselective BB overdose (such as carvedilol) decrease systemic vascular resistance mainly because of its α1 blocking property. Our patient consumed 7500 g of a cardioselective BB. While there is a robust understanding of the underlying pathophysiological basis of BB overdoses, evidence supporting the efficacy of glucagon, calcium supplementation. Pressor/Inotropes, HIE and ILE in the treatment of massive BB overdose is unclear. A brief review of the recommended treatments in BB overdose is warranted.

Table 1.

Pharmacokinetic properties of commonly used beta-blockers prescribed in the USA^{2 20}

Beta-blocker	Adrenoreceptor	T½ (hours)	Lipid solubility	Primary metabolism	Dialysable	Vasodilator effect
Atenolol	β1	3–4	Low	Renal	Yes	No
Bisoprolol	β1	14–22	Moderate	Renal/hepatic	Yes	No
Carvedilol	α1, β1 and β2	6–8	Moderate	Hepatic	No	Yes
Esmolol	β1	9 min	Low	Erythrocyte esterase	Yes	No
Labetalol	α1, β1 and β2	6–8	Low	Renal	No	Yes
Metoprolol	β1	3–4	Moderate	Hepatic	No	No
Propranolol	β1 and β2	3–5	High	Hepatic	No	No
Sotalol	β1 and β2	7–15	Low	Renal	Yes	No

Glucagon had a modest effect on HR but not MAP in our patient. The initial response prompted a continuous infusion. Glucagon exerts its effect by promoting the c-AMP pathway by activating adenyl cyclase enzyme independent of the beta blockade. It is important to note that to date there are no human studies or trials to validate its efficacy. Three animal studies reported increase in HR but no effect in MAP or CO.^3–5 In a small sample size of 39 patients suspected of BB toxicity, Catlin et al noted that only slightly more than half (56%) even had a response to glucagon.⁶ In the responders, the HR improved by a mean of 12 bpm and 11 mm Hg in MAP.⁶ Glucagon also comes in 1 mg ampules, so maintaining a high-dose infusion up to 5–10 mg/hour may be problematic depending on the resource and availability of the drug. After failing to achieve a good BP in our patient, intravenous calcium was administered and did improve MAP in our patient. In an interesting animal study of 14 rats, 4 were given high-dose intravenous calcium, 6 normal dose and 4 low dose, followed by BB infusion.⁷ The effect of BB on the myocardium was lowest in the high-dose group. In another animal study, 11 dogs were given propranolol infusion resulting in drop in MAP, calcium improved CI and SV in all of them with no effect on AV conduction.⁸ Several case reports have documented haemodynamic stability and survival following calcium administration in BB toxicity.^9–12 Based on the c-AMP pathway, it makes sense to administer calcium, but this resulted only in a brief response despite the recommended 1 g of calcium chloride administered three times. It is difficult to ascertain whether chronotropic or inotropic effects predominate in BB overdose. This is important because despite maintaining a MAP of close to 65 with the addition of epinephrine, our patient still became bradycardic, progressing to a junctional rhythm and subsequently a third-degree heart block prompting an emergent temporary transvenous pacemaker insertion. While chronotropic effects were managed with a temporary pacemaker, the patient’s MAP remained tenuous.

HIE showed great promise with good feasibility and safety profile after failing standard therapy in the treatment of cardiotoxicant overdoses.^{13 14} Cardiac myocytes usually use free fatty acids as substrates for fuel instead of glucose in shock-like states. With high doses of insulin, glucose along with lactate and oxygen are transported into myocardial cells, increasing inotropy. Animal studies have been convincing in the use of HIE in BB toxicity.^{15 16} However, high-quality clinical human trials have been lacking. Large case series from a poison centre in Minnesota showed good survival with HIE use in BB and CCB overdoses. The author, however, cited several limitations including the nature of these studies without a control arm, in accurate or incomplete data and, significant degree of selection and unmeasured bias leading to lack of definitive conclusion. Notably in one small case series of 20 patients with overdose (8 CCB, 5 BB, 7 CCB+BB) which did include a control arm, there was no difference in the outcome when HIE was compared with conventional vasopressors/inotropes.¹⁷ Another small case series in 14 patients with BB and CCB overdose, only slightly more than half the patients responded to HIE.¹⁸ With good physiological rationale, the administration of HIE appears to be safe in a monitored environment, although its true clinical impact is largely undetermined at this present time. After failing to respond to HIE, 20% ILE was started. ILE has a ‘lipid sink’ to extract drugs from tissue cells that has been proven effective with many lipophilic drug poisonings.¹⁹ Metoprolol is moderately lipophilic as opposed to propranolol which is the most lipophilic BB²⁰ (see table 1). In a study involving 20 rabbits infused with propranolol, a trend towards higher MAP was noted.²¹ However, in a similar study of 20 rabbits using metoprolol infusion, ILE failed to augment CO, MAP or HR.²² In a retrospective human case series published in 2017 involving 36 patients (10 were pure BB poisoning), ILE failed to augment MAP or HR without achieving clinical or statistical significance.²³ Similarly, in another very small case series of two patients with BB poisoning, no response was noted with ILE and both patients died shortly after ILE administration.²⁴ The author concluded that the cardiac arrests had a temporal relationship with the administration, although no specific cause was determined. Complications associated with ILE are pancreatitis from high lipids, acute respiratory distress syndrome and elevated liver transaminases, which our patient did not experience. Overall, the use of ILE should surely be considered in strongly lipophilic BBs, but the success associated with it may not be apparent in less lipophilic BBs.

Our patient soon progressed to maximum doses of NE, EPI and VASO after a failure of HIE and ILE. Evidence to support a single or any pressor/inotrope combination in severe BB overdoses is still lacking. In multiple case reports that depicted the use of EPI, NE, dopamine, dobutamine, isoprenaline, ephedrine, there was no drug that clearly showed superiority one over the other.^{25 26} Most of the haemodynamic instability was achieved from a combination of these drugs guided by Swan-Ganz placement which our patient did not have. In fact two small case series of 15 and 12 patients each showed no difference with vasopressors/inotropes compared with HIE.^{9 27} Despite being on maximum doses of NE, EPI and VASO, our patient failed to achieve the target MAP. This can be explained by the supersaturated state of the B-AR from the BB overdose. Finally, with worsening haemodynamics, our patient’s urine output dropped with the development of acute kidney injury, metabolic acidosis and subsequently volume overload that prompted CRRT. Among BBs, atenolol and sotalol are water soluble and renally excreted. Toxicities from these drugs have been successfully treated with CRRT and intermittent haemodialysis in published case reports with notable decrease in drug half-life.^28–30 Evidence to support use of dialysis in non-water-soluble BBs such as metoprolol is lacking.²⁰ One study suggested the possibility that metoprolol may have dialysis potential comparable to atenolol.³¹ While no expert consensus for the management of BB exists, reputed online resources such as upToDate^R recommend dialysis for only severe water-soluble BB overdoses and not for metoprolol.³² With a tenuous MAP in our patient, haemodialysis was not an option.

Our final option was ECMO. ECMO in drug overdoses boasts an impressive 80% survival rate based on the review of Toxicology Investigators Consortium.³³ VA-ECMO not only provides haemodynamic support perfusing all the vital organs but also provides time for the body to metabolise the drug. In a group of 62 patients with cardio-toxicants, Masson et al reported 3 out 3 cardiac arrest patients survived with ECMO as opposed to 0 out of 7 without ECMO, and 9 out of 11 patients with severe shock survived as opposed to 23 out of 41.³⁴ Majority of the patients have a relatively short time on the ECMO circuit and would be ideal candidates without being exposed to potential complications. ECMO is, therefore, indeed a life-saving intervention in this cohort of patients. While our patient did not have a major complication, he did have limb ischaemia on the side of arterial cannulation which was promptly identified and corrected with a DPC with no major sequelae. This complication is not uncommon, and many facilities insert a DPC pre-emptively to avoid limb ischaemia.³⁵ Our patient was on VA-ECMO for 3 days before his haemodynamics were restored and later downgraded to VV-ECMO.

Patient’s perspective.

“I am so thankful to the physicians for saving my life even though I did not want to at the time. It was impulsive and I am glad the physicians gave me every opportunity to survive this unfortunate ordeal.”

Learning points.

• Massive beta-blocker (BB) overdoses may be refractory to treatments with glucagon, calcium supplementation, vasopressors/inotropes and hyperinsulinaemic euglycaemia.

• Intravenous lipid emulsion works best for highly lipophilic BBs.

• Dialysis is an option but works best for water-soluble BBs like atenolol.

• We recommend that physicians must be cognisant of the pharmacokinetic properties of different BBs and understand their treatment limitations and strengths based on the BB ingested.

• Patients with massive BB overdose particularly metoprolol who show resistance to medical treatment must be referred early to a tertiary centre with extracorporeal membrane oxygenation capabilities.