Lithium Toxicity from the Addition of an ACE Inhibitor with an Unexpected Type I Brugada Pattern ECG: Case Files of the Medical Toxicology Fellowship at the University of California, San Diego

Case Presentation

A 48-year-old woman with a history of bipolar disorder, a hiatal hernia, hypercholesterolemia, hypertension, obesity, and type II diabetes was referred to the emergency department (ED) by her psychiatrist with the concern for lithium toxicity. The patient’s medications included atorvastatin, bupropion, duloxetine, insulin glargine, lisinopril, lithium carbonate, metformin, and ranitidine. The atorvastatin, bupropion, duloxetine, insulin glargine, metformin, and ranitidine were medications she had been on longstanding.

Approximately 1 month prior to ED presentation, the patient started taking immediate-release lithium carbonate 900 mg every evening for bipolar disorder. Twelve days after beginning lithium, laboratory testing revealed a normal blood urea nitrogen (BUN) of 10 mg/dL, a normal creatinine (Cr) of 0.87 mg/dL, and a therapeutic blood lithium concentration of 0.8 mmol/L (reference range 0.6–1.2 mmol/L). The estimated glomerular filtration rate (eGFR) at that time was > 60 mL/min.

Twenty-one days after lithium initiation, the patient was prescribed lisinopril 40 mg once per day for hypertension. Several days after beginning lisinopril, the patient developed confusion, nausea with occasional vomiting, and tremors. She reported occasionally dropping objects, having difficulty controlling her legs while driving, and falling once. She called her psychiatrist, to whom she described her symptoms, which she reported as worsening. She had also developed dizziness and what she described as “shakiness”. Her psychiatrist referred her for evaluation and she presented to the ED approximately 2 weeks following the addition of lisinopril. The patient described being compliant with her medications, denied any intentional self-harm, and denied any history of diarrhea. She also denied using any other medications including the use of any non-steroidal anti-inflammatory drugs (NSAIDs). She denied the regular use of ethanol.

Initial vital signs in the ED were temperature 98.2 °F, pulse 86/min, blood pressure 117/60 mmHg, and respiratory rate 20/min. The patient was alert and oriented, but required redirection at times in answering questions. She was noted to have dry mucous membranes and tremors in her hands. Her strength was normal and symmetrical and she was not hyperreflexic and did not exhibit clonus. She was able to perform finger-to-nose, heel to shin, and rapid alternating movements normally but she felt too unsteady to ambulate.

Initial laboratory testing revealed the following: sodium 127 mmol/L, potassium 4.4 mmol/L, BUN 28 mg/dL, Cr 2.2 mg/dL (eGFR rate 24 mL/min), glucose 119 mg/dL, and lithium 2.0 mmol/L. Her last dose of lithium was ingested approximately 12 h prior. Ethanol was not measurable in her blood.

Why Did This Patient Develop Lithium Toxicity?

There was a clear temporal relationship between initiating the angiotensin converting enzyme (ACE) inhibitor, lisinopril, and the subsequent onset of acute kidney injury and lithium toxicity. It was also thought that nausea and vomiting from the lithium toxicity itself contributed to volume depletion and acute kidney injury. At ED presentation, the patient’s creatinine was elevated reflecting decreased glomerular filtration. Lithium is not metabolized and is excreted almost entirely by the kidney, making plasma lithium concentrations exquisitely sensitive to physiological factors that affect renal function [1]. Lithium also has a relatively narrow therapeutic index, so small decreases in renal function and corresponding small increases in serum lithium concentrations can cause lithium toxicity as it did in this case.

Is There an Established Connection Between the Initiation of an ACE Inhibitor in Patients Taking Lithium and the Development of Lithium Toxicity?

Yes; reports began to appear soon after the introduction of ACE inhibitors in the 1980s [1]. A retrospective case-control study analyzed 20 patients who had been on lithium therapy and were subsequently begun on an ACE inhibitor. Four of the 20 patients developed symptoms of lithium toxicity [2]. A retrospective nested case-control study of elderly patients, thought to be at particularly high risk for lithium toxicity, demonstrated increased risk of lithium toxicity within a month of initiating an ACE inhibitor [3]. Additionally, multiple case reports and series have described the development of lithium toxicity in patients treated with both lithium and an ACE inhibitor [4–13]. In the majority of cases, patients were on chronic lithium therapy and developed lithium toxicity within 3 to 5 weeks after initiation of the ACE inhibitor [4, 6, 8, 9, 11]. In reports where lithium toxicity occurred at significantly longer time periods after ACE inhibitor initiation, the ACE inhibitor alone did not appear to be the explanation. In these cases, volume depletion from decreased fluid intake or significant diarrhea was present [13, 14]. A small prospective study addressed the potential for ACE inhibitors to produce lithium toxicity. In nine healthy volunteers administered low doses of lithium, no statistically significant changes in lithium concentrations were found following the addition of a low dose of the ACE inhibitor enalapril. However, one subject did experience a 31% increase in their serum lithium level after the ACE inhibitor was added [15]. The authors appropriately noted that a limitation was the low dosages of both lithium and ACE inhibitor.

What Is the Proposed Mechanism Responsible for the Potential Drug Interaction Between Lithium and ACE Inhibitors?

Multiple mechanisms may contribute to this drug interaction. One mechanism is simply that ACE inhibitors, by their mechanism of action, may decrease the GFR. The intraglomerular pressure, and consequently the GFR, is regulated by vasomotor tone of the afferent (pre-glomerular) and the efferent (post-glomerular) arterioles. Angiotensin II causes vasoconstriction of the efferent arterioles, serving to maintain intraglomerular pressure. ACE inhibitors prevent the conversion of angiotensin I to angiotensin II and consequently prevent this vasoconstriction. The resulting decrease in intraglomerular pressure leads to a decreased GFR. As a consequence of the decreased GFR, serum creatinine may increase by as much as 30% following initiation of an ACE inhibitor [16]. Typically, this effect is seen within 3–5 days of initiation and stabilizes within a week [17]. This effect is generally not concerning, and physicians are warned not to stop the ACE inhibitor if this occurs [16]. The decrease in intraglomerular pressure is beneficial in certain situations, such as diabetic renal disease, in which a high intraglomerular pressure is associated with progression of renal disease. However, in the case of lithium, given its narrow therapeutic index and dependence on renal clearance, even small, normally inconsequential decreases in renal function such as from an ACE inhibitor could lead to toxicity. In the majority of published cases of lithium toxicity attributed to the addition of an ACE inhibitor, the serum creatinine is in fact elevated [4, 6, 8, 9, 11, 12]. However, in most cases where the serum creatinine is detailed, as in the case we describe, the creatinine is elevated greater than 30% above baseline, suggesting another mechanism may be involved [4, 6, 8, 11, 12]. In a retrospective study by Juurlink et al., renal function was unfortunately not documented [3]. There are several conditions where initiation of an ACE inhibitor can lead to an exaggerated rise in creatinine, beyond the 30% rise that may be expected even with therapeutic dosing. Many of these are conditions where there is reduced blood pressure in the afferent arterioles, such as hypotension, hypovolemia, and renal artery stenosis [16, 18, 19].

In the case we describe, our impression was that the developing lithium toxicity itself led to nausea and vomiting, subsequent hypovolemia, and may have contributed to the exaggerated creatinine elevation, beyond what may have occurred from the ACE inhibitor alone. Although nausea and vomiting are well-described manifestations of lithium toxicity, many of the prior reports depicting the interaction between lithium and ACE inhibitors do not describe this as a contributing factor [4, 6, 8, 9, 11, 12]. Some authors have suggested that lithium itself, via effects on volume and sodium homeostasis, may contribute to the drug interaction [20].

A mechanism beyond a direct effect on GFR may include increases in reabsorption of lithium in both the proximal and distal renal tubule as a result of decreased aldosterone, the latter caused by the ACE inhibitor [12]. Such a mechanism would be supported by the retrospective study by Finley et al. who detailed a reduction of lithium clearance after the addition of an ACE inhibitor and that only 25% of this clearance reduction could be attributed to a rise in serum creatinine [2].

Given the known potential interaction between lithium and ACE inhibitors, some authors have recommended to avoid the combination of both medications if possible [2]. If both drugs are used, close monitoring is recommended [3, 12].

Can Lithium Itself Cause Renal Dysfunction?

There remains some debate about this topic, but the authors of a recent review suggest that the predominant view is that progressive chronic kidney disease, and in certain cases, end-stage renal disease can occur with long-term lithium use. The mechanism appears to be chronic tubulointerstitial nephritis and it generally only occurs in patients who have been on lithium for many years, which our patient had not [21].

What Are Other Common Etiologies of Why Patients on Therapeutic Lithium Develop Toxicity?

Etiologies include any clinical condition, such as hypovolemia, that decreases GFR. Drugs other than ACE inhibitors that may decrease GFR and are implicated in precipitating lithium toxicity include angiotensin II receptor blockers, diuretics, and non-steroidal anti-inflammatory drugs [1]. In the previously cited study by Juurlink et al., a dramatically increased risk of lithium toxicity was also found within a month of initiating a loop diuretic [3]. Further, given the biochemical similarity of lithium to sodium, enhanced reabsorption of lithium from the proximal renal tubule can occur from volume depletion and thiazide diuretics, even in the absence of a decreased GFR [1].

Should This Patient Undergo Extracorporeal Treatment?

Lithium is considered highly dialyzable and ECTR, such as by hemodialysis, can reduce lithium concentrations from the blood at a rate exceeding normal renal clearance several-fold. The Extracorporeal Treatments in Poisoning Workgroup is an international panel of experts who publish evidence-based recommendations regarding ECTR for poisonings. In 2015, the group published recommendations for when to utilize ECTR for lithium [22]. The group suggested that the clinical decision should take into account the following factors: lithium concentration, renal function, pattern of lithium toxicity, the patient’s clinical status, and the availability of ECTR. Recommendations are provided combining the strength of the recommendation (numerical) followed by the quality of evidence (alphabetical). ECTR is recommended for severe lithium poisoning (1D: strong recommendation with very low quality of evidence). Specifically, ECTR is given a 1D recommendation if renal function is impaired (eGFR < 45 mL/min) and the lithium concentration is greater than 4.0 mmol/L, or in the presence of a decreased level of consciousness, seizures, or life-threatening dysrhythmias irrespective of the lithium concentration. Further, ECTR is suggested with a 2D level of recommendation (weak recommendation with very low quality of evidence) if the lithium concentration is greater than 5.0 mmol/L, if significant confusion is present, or if the expected time to obtain a lithium concentration less than 1.0 mmol/L with optimal management is greater than 36 h [22]. We opted not to perform ECTR in the patient reported here, which is supported by the above recommendations.

Case Continuation

The patient was admitted to the hospital and both the lithium and ACE inhibitor were stopped. The patient was intravenously hydrated with the goal of maintaining normal urine output.

Is There Anything Else that Can Be Done to Help Eliminate Lithium?

The administration of sodium polystyrene sulfonate (SPS) can be considered. Sodium polystyrene sulfonate is a cation-exchange resin commonly used in the treatment for hyperkalemia. It has been well demonstrated that SPS can adsorb to lithium in vitro, and in both animal and human models of oral lithium ingestion SPS administration reduced absorption. Beyond prevention of absorption, however, in other animal models, enteral SPS administration reduced lithium concentrations in the setting of chronic lithium use and after intravenous administration of lithium [23–28]. Efficacy of removal was greater at higher doses of SPS and with multiple-doses [29]. These latter studies suggest that SPS can accelerate removal of the drug from the blood back into the intestine (“gut dialysis”) [30]. A retrospective case series of patients with chronic lithium toxicity suggested that compared to patients who did not receive SPS, the half-life of lithium was reduced significantly (20.5 versus 43.2 h). In that study, each dose utilized was 30 g and a range of one to four doses were administrated [30]. The current utility of SPS for lithium toxicity remains unclear and, as the authors of the retrospective series point out, a prospective study is warranted. Increasing recognition of the rare, but potentially fatal gastrointestinal injury caused by SPS administration should be taken into consideration given the current absence of proven outcome improvement for use with lithium toxicity [31].

Case Continuation

The patient was administered one dose of SPS with the goal of accelerated removal from the blood. SPS dosing was not repeated, and it was unclear if it had any clinical effect.

An electrocardiogram (ECG) was performed as part of the routine management of a lithium-poisoned patient. The ECG was performed multiple hours after the patient had presented to the ED and after hydration had been begun. Near the time of the ECG, the repeat sodium and lithium concentrations were 132 mmol/L and 1.9 mmol/L, respectively.

What ECG Changes Have Been Attributed to Lithium?

A recent article systematically reviewed this topic and does an excellent job summarizing the findings [32]. A wide variety of ECG changes due to lithium have been described. The most common ECG finding is T-wave inversion and the second most common is sinus bradycardia. Typically, the T-wave inversions and sinus bradycardia occur at therapeutic lithium concentrations and are clinically insignificant. Many of the other reported ECG findings are often or exclusively described in the setting of elevated lithium concentrations and include sinoatrial blocks, PR prolongation, QT prolongation/dispersion, and ventricular tachydysrhythmias [32].

The ECG Performed in the ED Is Depicted in Fig. 1. What Does the ECG Demonstrate?

Fig. 1.

ECG performed on hospital day 1 that demonstrates a type 1 Brugada pattern

The ECG demonstrates a type I Brugada pattern, characterized by coved ST segment elevation > 2 mm in > 1 lead among the right precordial leads (V1 and V2) followed by a negative T wave [33]. The ECG finding was completely unexpected to the providers, prompting significant discussion and further reading regarding its significance.

What Is the Brugada Syndrome?

The Brugada syndrome is an inherited sodium channelopathy in patients with structurally normal hearts who have a high incidence of sudden death in early adulthood from polymorphic ventricular tachycardia and/or ventricular fibrillation. It is estimated that the Brugada syndrome is responsible for at least 4% of all sudden deaths and at least 20% of sudden deaths in patients with structurally normal hearts. The mean age of sudden death is 41 [34]. Electrocardiographically, Brugada syndrome is characterized by distinct ST segment elevation in the right precordial leads that appears to be linked to reduced inward sodium current from dysfunctional sodium channels. ECG manifestations are often dynamic or concealed and may be unmasked by various conditions. Currently, two ECG repolarization patterns in the right precordial leads are recognized, including the diagnostic type I pattern described above. The type II pattern has a saddleback pattern [35]. According to an expert consensus published in 2013, Brugada syndrome is diagnosed in patients who have a type I Brugada pattern ECG that occurs spontaneously or after provocative drug testing with intravenous administration of a class I antidysrhythmic drug. Those patients with a type II Brugada ECG pattern can be diagnosed with Brugada syndrome if intravenous administration of a class I antidysrhythmic drug induces a type I Brugada pattern ECG [36]. A prior consensus statement from 2005 that remains widely cited emphasized for diagnosis the presence of a type I Brugada pattern ECG in conjunction with one of the following: documented ventricular fibrillation or polymorphic ventricular tachycardia, a family history of sudden cardiac death at < 45 years of age, coved-type ECGs in family members, inducibility of ventricular tachycardia with programmed electrical stimulation, syncope, or nocturnal agonal respiration [34].

Case Continuation

An ECG from 2 years prior was obtained that displayed an incomplete right bundle branch block with no other abnormalities. The patient did not have a history of syncope or any known episodes of ventricular dysrhythmias, nor was she aware of any family history of the same or sudden cardiac death. However, the patient’s aunt on her father’s side and a female cousin both had pacemakers placed although no further information on the indications for pacemaker insertion was available.

Have Brugada ECG Patterns Been Previously Attributed to Lithium?

Yes. Darbar et al. in 2005 were the first to describe the association of lithium with Brugada ECG patterns. In their report, they describe two patients in whom lithium appeared to unmask the presence of the Brugada syndrome. They describe a 26-year-old patient with a history of recurrent syncope who had a baseline normal ECG prior to lithium initiation. Two months after beginning lithium, a type I Brugada ECG pattern was identified in the setting of a therapeutic lithium concentration. Electrical stimulation at electrophysiology precipitated ventricular fibrillation that was successfully defibrillated. An internal defibrillator was placed and lithium was discontinued. Subsequent ECGs demonstrated a type II Brugada ECG pattern. The other patient described was a 65-year-old on longstanding lithium (nearly 20 years) who also had a history of recurrent syncope. A family history revealed multiple family members having sudden cardiac death. A type I Brugada ECG pattern was identified after a syncopal episode. Programmed electrical stimulation induced non-sustained ventricular tachycardia. An internal defibrillator was placed and lithium was continued. During an episode of lithium toxicity, a type I Brugada ECG pattern was again identified that resolved when his lithium dose was decreased [37].

What Have Reports Subsequent to the One by Darbar et al. Detailed?

There have been multiple case reports of Brugada pattern ECGs associated with the use of lithium since the initial report by Darbar et al [38–45]. In many cases, similar to the one we present, a type I Brugada pattern was recognized in a patient on lithium in whom there is no personal or family history of syncope or cardiac arrest, and the type I Brugada ECG findings resolved with discontinuation of the lithium [39, 41–44]. The type I Brugada pattern has been recognized in both the setting of lithium toxicity [41–44] and therapeutic lithium concentrations [39, 40, 45, 46]. Only two of these reports describe a history of syncope or cardiac arrest. One report details a patient with a history of recurrent syncope for 8 years while on lithium, who presented following a syncopal episode and was found to have a Brugada pattern ECG. Provocative drug testing was subsequently done that was consistent with Brugada syndrome [46]. In the other case, a patient presented after resuscitation from cardiac arrest during which he had a wide-complex tachycardia. During hospitalization, a type I Brugada pattern was identified [40]. In the above two cases, automatic internal cardiac defibrillators were placed [40, 46]. Genetic testing can be done to investigate if Brugada syndrome is present and was confirmed in at least one patient who had a Brugada pattern ECG caused by lithium [42].

What Is the Mechanism Proposed for How Lithium Can Unmask Brugada Pattern ECGs?

Darbar et al. demonstrated with an in vitro model that lithium blocks cardiac sodium channels in a dose-dependent manner and at concentrations well below therapeutic levels [37, 47]. By blocking cardiac sodium channels, lithium may unmask sodium channel dysfunction similar to the diagnostic provocation done with intravenous administration sodium channel blocking agents (type I antidysrhthmics). This sodium channel dysfunction may manifest as a Brugada pattern ECG and/or more seriously as a life-threatening ventricular dysrhythmia.

What Is Brugada Phenocopy?

Brugada phenocopy is an emerging concept used to describe conditions that induce Brugada-like ECG manifestations in patients who do not have actual congenital Brugada syndrome. Such conditions include certain metabolic conditions, myocardial ischemia, pericardial disease, and pulmonary embolism [48]. Hyponatremia is one of the metabolic conditions described to cause Brugada phenocopy [49–52]. It is thought that diminution of inward sodium current is essential to produce the Brugada pattern ECGs and that with hyponatremia the reversible ECG changes are produced by a diminished transmembrane sodium concentration gradient [49]. Our patient presented with hyponatremia (sodium of 127 mmol/L) that was attributed predominantly to acute kidney injury. The initial ECG was performed near the time the sodium was 132 mmol/L. In published reports of hyponatremia-induced Brugada phenocopy, the serum sodium is typically lower (< 113 mmol/L), although in one case, it was as high as 121 mmol/L [49–52].

Case Continuation

The type I Brugada pattern in our patient (see Fig. 2) persisted the following day despite her sodium being in the normal range (136 mmol/L), proving that mild hyponatremia was not the cause. Her lithium concentration at that point was still elevated at 1.6 mmol/L.

Fig. 2.

Four chronological ECG’s throughout hospital stay with associated sodium and lithium levels (mmol/L)

Does Our Patient Have Brugada Syndrome or Brugada Phenocopy?

According to Baranchuk et al. who initially proposed the concept of Brugada phenocopy, our patient may in fact have actual Brugada syndrome. Given the mechanistic similarity of lithium to the diagnostic agents used in provocative testing, lithium may have in fact unmasked rather than mimicked Brugada syndrome [48].

What Further Testing Could Be Done to Confirm if the Patient Has Brugada Syndrome?

Provocative testing as described above with intravenous administration of a type I antidysrhythmic such as ajmaline, flecainide, or procainamide by a cardiac electrophysiologist could be performed. Additionally, genetic testing could be performed.

Case Continuation

Intravenous hydration was continued. The patient’s renal function rapidly improved, her lithium concentrations decreased and her symptoms of lithium toxicity progressively resolved. Daily ECG’s were performed. On hospital day 3, the ECG demonstrated a type II Brugada pattern at which point her sodium was 137 mmol/L and lithium was 1.1 mmol/L. (see Fig. 2) At hospital discharge 3 days after admission, the patient’s BUN was 7 mg/dL, Cr was 0.91 mg/dL, and her lithium concentration was 0.8 mmol/L. She was discharged without a blood pressure medication and in replacement of lithium she was placed on lamotrigine and ziprasidone. At discharge, the patient was referred for outpatient follow-up with a cardiac electrophysiologist.

Have Lamotrigine or Ziprasidone Been Associated with Brugada Pattern ECGs?

Lamotrigine is a known sodium-channel blocker [53] and Brugada pattern ECGs have been described in patients with actual Brugada syndrome and in a patient with Brugada phenocopy [54–56]. Intriguing in these reports is the potential effect of lamotrigine on provocative testing. In one case, a patient had a type II Brugada ECG in the setting of lamotrigine toxicity and had a false positive provocative test with procainamide (positive test in the presence of lamotrigine and negative test in its absence) [54]. In another case, provocative testing with ajmaline was performed in a patient ultimately diagnosed with true Brugada syndrome (patient and family with history of syncope, confirmatory genetic testing) in the presence and absence of lamotrigine. In the presence of lamotrigine, provocative testing was proarrhythmogenic [56]. We are not aware of ziprasidone being associated with a Brugada pattern ECG. Lamotrigine and other sodium channel blocking drugs would not appear to be an ideal psychoactive drugs for our patient if she were to have true Brugada syndrome.

Case Conclusion

Subsequent to being discharged, the patient has had a myocardial perfusion scan that demonstrated no prior myocardial infarctions, nor inducible ischemia. It was performed as part of pre-operative clearance for a sleeve gastrectomy. The patient has remained on lamotrigine and has had neither syncopal episodes nor any other episodes suggesting having had a ventricular dysrhythmia. An ECG performed 9 months after hospital discharge did not demonstrate a Brugada pattern, although there was some slight coving of the ST segment in V1. Losartan, an angiotensin receptor blocker, was begun for hypertension and no significant adverse effect on renal function has occurred. The patient was evaluated by an electrophysiologist, who discussed options including provocative testing. The patient agreed to having an echocardiogram and cardiac magnetic resonance imaging to evaluate for structural heart disease and genetic testing for Brugada syndrome. The plan was to consider genetic testing of family members based on the results found on our patient. However, the patient moved away and to date has not had any of this additional testing.

Sources of Funding

None

Compliance with ethical standards

Conflicts of Interest

None