Arrhythmias and Palpitations
Arrhythmias may occur as a consequence of treatment with all antipsychotics, TCAs, some SSRIs, lithium, stimulants, and anticholinergic drugs, or in the setting of toxicity states. Clinicians should be familiar with cumulative risk factors for prolonged ventricular repolarization (the QTc interval; see Table 7–3) or ventricular depolarization (in the case of TCAs), and the relevance for baseline electrocardiographic monitoring (e.g., in adults with known or suspected cardiac disease, before starting a TCA, lithium, or in some instances an SGA). Palpitations (the subjective awareness of heartbeats) are usually benign in the absence of underlying structural heart disease, but carry a wide differential diagnosis that involves factors unrelated to psychotropic drugs.
The arrhythmogenic potential of some psychotropic drugs is long established and often poses a significant deterrent to prescribing otherwise effective medications. Examples of these drugs include TCAs by virtue of their anticholinergic effects, α-adrenergic blockade, and quinidine-like effects from the blockade of fast sodium channels in myocardial cells.
Table 7–1 summarizes known relationships between electrocardiographic changes and common psychotropic medications. Some experts advise obtaining a baseline ECG in adults over age 40 (in addition to those with a history of cardiac disease) before beginning a TCA, lithium, and some SGAs (such as those with a greater potential for QTc prolongation, as described in Table 7–2).
Most SGAs carry risks for both tachycardia (presumably due to anticholinergic effects) and orthostatic hypotension (probably due to α1-adrenergic blockade). Proper management involves measurement of heart rate and blood pressure, including orthostatic measurements, and gradual dosage increases when necessary. If a β-blocker is being used to treat akathisia or tremor, monitoring heart rate and blood pressure is especially important to assure no exacerbation of hypotension from α1-adrenergic blockade.
Palpitations refer to the awareness or subjective experience of irregular heartbeats. They may or may not reflect actual ectopic beats. Atrial premature complexes (APCs) or premature ventricular complexes (PVCs) that are isolated, intermittent, and arise spontaneously are usually benign and common occurrences in healthy people, unless the contractions are accompanied by other cardiovascular signs (such as chest pain, dizziness, or syncope) or arise in the setting of structural heart disease.
Psychiatrists who evaluate palpitations should review all of a patient’s medications (both psychiatric and nonpsychiatric) to identify drugs that may cause tachycardia (e.g., anticholinergics), prolong the QT interval (see Tables 7–2 and 7–3), or pharmacokinetically inhibit the metabolism of anticholinergic or QT-prolonging drugs (see Tables 2–6 and 2–7 in Chapter 2, “Pharmacokinetics, Pharmacodynamics, and Pharmacogenomics”). Pulse rate and regularity should be measured and the heart auscultated to discern premature beats, particularly if occurring as couplets or triplets. Obtaining an electrocardiogram (ECG) is appropriate to assure normal intervals and identify APCs or PVCs. Runs of bigeminy or trigeminy warrant more extensive studies (e.g., Holter monitoring) and referral to a cardiologist. Echocardiography may be indicated for patients with a murmur who complain of palpitations. Iatrogenically, stimulants may cause sinus or supraventricular tachyarrhythmias but not APCs or PVCs. Rare reports of new-onset PVCs have been described with the use of modafinil (Oskooilar 2005). Clinicians obviously should recognize and consider nonpsychotropic drug causes of a rapid or irregular heart rhythm, including the effects of thyroid hormone, inhaled beta agonists, antihypertensive agents, caffeine, and nicotine, as well as the potential contribution of hyperthyroidism, electrolyte abnormalities, and anxiety, among other possible etiologies.
Electrocardiographic changes and concerns
Anticonvulsants and lithium
QRS prolongation; heart block and ventricular arrhythmias possible in overdose.
Tachycardia or bradycardia.
Rare associations reported with lamotrigine overdose and Brugada patterna on ECG, presumably via the effect of lamotrigine on sodium channels, as well as QRS widening; routine ECG monitoring not indicated.
Reversible T-wave changes, sinus bradycardia, sick sinus syndrome, heart block, case reports of Brugada patterna on ECG; controversial case reports of QTc prolongation.
Tachycardia, premature beats, nonspecific ST-T changes, QRS prolongation in overdose.
Possible association with relatively minor increase in heart rate (<5 beats per minute); no other known electrocardiographic abnormalities.
May increase heart rate due to increased noradrenergic tone.
May decrease heart rate; rare increase in QTc intervals with escitalopram, fluoxetine, fluvoxamine, and sertraline; more likely increase in QTc with citalopram (dose related) or in the setting of overdoses or drug interactions.
No known effects on any ECG parameters when dosed up to 80 mg/day (Edwards et al. 2013).
No QTc > 10 msec when dosed up to 40 mg/day.
Tachycardia possible due to vagolytic effects; quinidine-like effects possible (i.e., blockade of fast sodium channels causes prolonged depolarization with decreased myocardial contractility, leading to PR prolongation, QRS widening, right axis deviation, and bradycardia or heart block in overdose); TCAs can inhibit cardiac hERG K+ channel currents, prolonging action potentials and increasing the risk for QT prolongation.
Sinus tachycardia, QTc prolongation (particularly with pimozide, thioridazine, and intravenous haloperidol), ventricular arrhythmias.
Tachycardia (lurasidone, risperidone), first-degree AV block (lurasidone), QTc prolongation (ziprasidone, iloperidone, and potentially others in dose-related fashion).
Anxiolytics or sedative-hypnotics
Rare reports of QTc prolongation with lorazepam in patients with underlying arrhythmia.
No clinically significant QTc prolongation at 12 times the maximal recommended dose.
Miscellaneous CNS agents
Mean 4.5 msec (24-mg single dose) (90% CI=2.4–6.5 msec) increase in QTc interval (healthy volunteers).
Mean 6.7 msec (40-mg dose) (healthy volunteers); for CYP2D6 PMs, maximally dosed (80 mg/day) ΔQTc=11.7 msec.
Note. AV=atrioventricular; CI=confidence interval; CNS=central nervous system; ECG = electrocardiogram; FGA=first-generation antipsychotic; hERG = human ether-à-go-go-related gene (which codes for the alpha subunit of the K+ ion channel in cardiac myocytes); SGA=second-generation antipsychotic; SNRI=serotonin-norepinephrine reuptake inhibitor; SSRI=selective serotonin reuptake inhibitor; TCA=tricyclic antidepressant.
Psychiatrists (and other health professionals) often rely on computer-read interpretations of ECG parameters, including QTc intervals. However, one must bear in mind the presence of factors that can interfere with accurate reading of QTc intervals (whether computerized or manual), including wide QRS complexes, pacemakers, and rapid heart rates.
Sudden Cardiac Death
Clinicians should be aware that all antipsychotics and psychostimulants carry a small but statistically significantly increased risk for sudden cardiac death.
Findings in FDA registration trials
Likelihood > placebo (OR, 95% CI)a
No known QTc prolongation.
Manufacturer’s product information: at dosages of 120 or 600 mg/day, “no patients experienced QTc increases >60 msec from baseline, nor did any patient experience a QTc >500 msec”; 5.1-msec and 4.5-msec increases from baseline at 40 mg/day and 120 mg/day doses (Meltzer et al. 2011).
–0.10 (–0.21 to 0.01)
–4.2-msec decrease from baseline (Marder et al. 2003).
0.01 (–0.13 to 0.15)
12.3-msec increase from baseline (8-mg dose); no subjects had a change exceeding 60 msec or QTc >500 msec (manufacturer’s product information).
0.05 (–0.18 to 0.26)
4.7-msec increase from baseline (Glassman and Bigger 2001).
0.11 (0.03 to 0.19)
14.5-msec increase from baseline (Glassman and Bigger 2001).
0.17 (0.06 to 0.29)
6.8-msec increase from baseline (Glassman and Bigger 2001).
0.22 (0.11 to 0.31)
11.6-msec increase from baseline (Glassman and Bigger 2001).
0.25 (0.15 to 0.36)
2- to 5-msec increase from baseline as compared with placebo; no observations of QTc ≥500 msec (manufacturer’s product information).
0.30 (–0.04 to 0.65)
~9-msec QTc prolongation at 12 mg twice daily; ~19 msec if coadministered with CYP2D6 or CYP 3A4 inhibitors (manufacturer’s product information).
0.34 (0.22 to 0.46)
20.3-msec increase from baseline (Glassman and Bigger 2001).
0.41 (0.31 to 0.51)
35.8-msec increase from baseline (Glassman and Bigger 2001).
~5- to 8-msec increase from baseline (34 mg/day); sporadic observations of QTc (Fridericia-corrected) > 500 msec and increases > 60 msec from baseline, but no reports of TdP (manufacturer’s product information)
Note. CI=confidence interval; FDA=U.S. Food and Drug Administration; FGA=first-generation antipsychotic; msec=milliseconds; OR=odds ratio; SGA = second-generation antipsychotic; TdP=torsades de pointes.
bHigh doses of haloperidol or intravenous haloperidol may carry a greater risk for QTc prolongation than oral or intramuscular haloperidol.
Antifungal agents (e.g., ketoconazole, fluconazole)
Certain antihistamines (astemizole and terfenadine onlya)
Certain antibiotics (including azithromycin, ciprofloxacin, clarithromycin, erythromycin, gemifloxacin, levofloxacin, moxifloxacin, norfloxacin, ofloxacin, rufloxacin) and antimalarials (e.g., chloroquine, mefloquine, halofantrine)
Certain antiemetics (e.g., ondansetron, prochlorperazine)
Certain antifungal agents (e.g., ketoconazole)
Certain antineoplastics (e.g., anthracyclines, 5-fluorouracil, alkylating agents)
Certain SSRIs at high doses (e.g., citalopram and escitalopram), possibly fluoxetine
Class I antiarrhythmics (e.g., quinidine, disopyramide, procainamide) and Class II antiarrhythmics (e.g., amiodarone, sotalol, dofetilide)
Congenital QT prolongation syndrome (arising in ~1/5,000 live births)
Congestive heart failure
Left ventricular dysfunction or ventricular arrhythmias
TCAs, maprotiline, trazodone
Note. SSRI=selective serotonin reuptake inhibitor; TCA=tricyclic antidepressant.
aBoth agents have been withdrawn from the U.S. market; noted for historical purposes.
Sudden cardiac death due to an arrhythmia is a particular concern in a number of clinical contexts. All antipsychotics, both FGAs and SGAs, may carry an increased (approximately twofold), dose-related risk for sudden cardiac death due to presumptive ventricular arrhythmias (Ray et al. 2009). Additionally, anticonvulsant drugs that block sodium channels (e.g., carbamazepine, gabapentin) have been linked with an approximately three- to sixfold increased risk for sudden cardiac death in epilepsy patients (Bardai et al. 2015). During treatment with antipsychotics, Glassman and Bigger (2001) identified 10–15 events per 10,000 patient years, also noting an approximate twofold increased risk for sudden death during antipsychotic therapy as compared to the general population. In addition, case reports exist of sudden death among children <age 14 after exposure to TCAs, particularly desipramine, often in the absence of a preexisting history of cardiac arrhythmia. Although TCAs are currently seldom used in children, appropriate precautions before their initiation in pediatric populations should involve a baseline ECG and identification of any underlying heart disease.
For many decades, antipsychotics have been known to have the potential for prolonging the time for ventricular repolarization (i.e., the QT interval on an ECG, corrected for heart rate [QTc*]), with an associated potential for causing TdP.
When Pfizer sought its indications for ziprasidone from the FDA, a comparative study was conducted to evaluate QTc prolongation with ziprasidone relative to several other conventional antipsychotics or SGAs, in the presence or absence of CYP inhibitors (see Glassman and Bigger 2001). Importantly, this is the sole study that provides direct comparative data across several antipsychotics (ziprasidone, risperidone, olanzapine, quetiapine, thioridazine [■], and haloperidol) under controlled conditions. Phenothiazines (including thioridazine [■] and mesoridazine [■]) are often considered to pose the greatest risk among antipsychotics for QTc prolongation. Antipsychotic-associated QTc prolongation may in general be a dose-related phenomenon (Reilly et al. 2000).
Table 7–3 provides a summary of information from both that Pfizer-sponsored study and from available FDA registration trial data of QTc effects with FGAs and SGAs, although absolute differences cannot easily be construed about relative QTc effects across agents due to between-subject differences in baseline risk factors for QTc duration. One must also consider that the paucity of head-to-head comparisons among antipsychotics makes it difficult to draw robust inferences about the “relative” safety of one agent over another with respect to the risk for repolarization abnormalities (although a meta-analysis of seven SGAs in schizophrenia concluded that aripiprazole appears to have the lowest overall risk for QTc prolongation [Chung and Chua 2011]).
The relationship between QTc prolongation and TdP is complex, impacted by multiple factors, and often patient-specific and context-dependent. For example, the clinical significance of a medication-associated QTc increase by 10 msec is likely different if a baseline QTc rises from 490 to 500 msec than from 380 to 390, or in a patient whose baseline QTc is in the higher versus lower range of normal. Importantly, the biological “propensity” of any single individual for QT prolongation and TdP is a key moderator of overall cardiac risk when deciding on the safety versus benefits of a medication that has the potential to prolong the QT interval. This point is illustrated by the example of database reports linking short-term use of azithromycin with sudden cardiac death, prompting an FDA warning about its use in patients with pre-existing or concomitant risk factors for QT prolongation (e.g., those described in Table 7–3) (www.fda.gov/drugs/drugsafety/ucm341822.htm).
The aforementioned Pfizer comparative study of antipsychotic effects on QTc intervals identified marked QTc prolongation with thioridazine, prompting imposition of a boxed warning (■) with thioridazine (see Table 1–6 in Chapter 1, “The Psychiatrist as Physician”) and a physician notification letter warning of the risk for QTc prolongation and TdP. Mesoridazine received a similar boxed warning (■) shortly thereafter. The FDA traditionally categorizes increasing degrees of clinically meaningful risk associated with QTc prolongation from baseline as follows: ≤5 msec as “probably no concern,” 6–10 msec as “increasing concern,” 11–20 msec as “uncertain concern,” and >20 msec as “definite concern” (see also U.S. Food and Drug Administration 2005).
Generally, in patients with elevated QTc (i.e., ≥450 msec in men and ≥470 msec in women) and psychotic features, it is advisable to minimize or avoid exposure to all antipsychotics unless the perceived benefit outweighs the potential risk. In such patients, cautious observation with serial ECG monitoring may be warranted, as well as the elimination or correction of other potential causes of QTc prolongation (see Table 7–3). QTc intervals ≥500 msec pose a substantial hazard for developing potentially fatal ventricular arrhythmias (notably, TdP) and generally signal the need to eliminate antipsychotic drugs. In some instances, however, the clinical necessity of antipsychotic medication for patients with significant baseline QTc prolongation may warrant recommendations for cardiovascular implantable electronic devices such as permanent pacemakers, implantable cardioverter defibrillators, and cardiac resynchronization therapy devices (for further discussion of the psychiatrist’s role in collaborative decision making alongside cardiologists in psychiatric patients at risk for TdP, see Brojmohun et al. 2013).
From a practical standpoint, it is often useful to compare current ECGs to prior ECGs in order to determine whether suspected abnormalities represent interval changes. New findings may warrant drug cessation or consultation with a cardiologist, whereas stable features may pose lesser concern for iatrogenic risks. Clinicians also should be mindful of cumulative risk factors for arrhythmias in a given patient, such as electrolyte abnormalities, hypothyroidism, advanced age, effects of alcohol, and the synergistic arrhythmogenic potential of concomitant drugs (e.g., fluoroquinolone antibiotics, anticholinergics, and other drugs noted in Table 7–3).
Some authors recommend that because of the potential for sudden cardiac death with all antipsychotics, clinicians should determine, before initiating an antipsychotic, whether a patient has had syncope, has relatives with known congenital QT prolongation syndrome, or has relatives who experienced sudden death at an early age, and should obtain a baseline ECG in older adults or those with a history of known cardiac disease (Glassman and Bigger 2001). The American Heart Association (AHA) also advises obtaining a baseline ECG (mainly to assure the absence of QT prolongation) before beginning a TCA or phenothiazine in children or adolescents (Gutgesell et al. 1999), with repeat assessment after steady-state dosing is achieved. Adult women generally may have a greater risk than men for developing TdP from medications that prolong QTc.
A summary of recommendations to aid decision making about antipsychotic use in the context of QTc prolongation is provided in Table 7–4.
• Measure baseline QTc; serial QTc monitoring as clinically appropriate (e.g., after dosage increases or addition of other medications that can prolong QTc or inhibit CYP450 2D6).
• Avoid low-potency FGAs, IV haloperidol, ziprasidone, and possibly iloperidone.
• Favor atypical antipsychotics with low risk for QTc prolongation (e.g., aripiprazole, lurasidone).
• For IV haloperidol in hospital settings: obtain baseline and at least daily QTc, Mg+, K+, and PO4–, and continuous ECG monitoring for those with baseline QTc>500 msec, significant risk factors, or high-dose requirements (e.g., total dose>2 mg/day).
• If QTc increases to >500 msec during treatment: check and replete K+, Mg+, PO4–; review overall medication regimen to identify other agents that may increase QTc; consider holding antipsychotic until QTc falls <500 msec; consider alternative agents (e.g., nonantipsychotic drugs for agitation, such as divalproex or benzodiazepines).
Note. ECG=electrocardiographic; FGA=first-generation antipsychotic; IV = intravenous.
TCAs may be arrhythmogenic by virtue of their anticholinergic and quinidine-like effects. Orthostatic hypotension may result from α1-adrenergic blockade with tertiary amine TCAs (e.g., amitriptyline, imipramine). SSRIs generally lack effects on cardiac conduction with the apparent exception of citalopram and its enantiomer escitalopram; in August 2011, the FDA issued an alert that dosages of citalopram >40 mg/day may prolong QT intervals and advised against the use of citalopram in patients with congenital long QT syndrome or other conditions associated with risk of QT prolongation, if possible. The maximum recommended dose of citalopram was changed to 20 mg/day for patients older than 60 years. Specifically, the FDA noted dose-related QTc increases with citalopram of 8.5, 12.6, and 18.5 msec at respective doses of 20, 30, and 40 mg/day, while escitalopram-related QTc increases were 4.5, 6.6, and 10.7 msec at respective doses of 10, 20, and 30 mg/day (www.fda.gov/Drugs/DrugSafety/ucm297391.htm). Accordingly, caution is recommended when using citalopram and possibly escitalopram in patients with other risk factors for QT prolongation (see Table 7–3). Citalopram also has been associated with significantly greater QT prolongation in overdoses as compared with changes seen in fluoxetine, fluvoxamine, paroxetine, or sertraline (Isbister et al. 2004). It is advisable to obtain a baseline ECG in patients taking ≥40 mg/day of citalopram and in those with risk factors for prolonged QTc.
In 2013, the FDA revised the package insert language for fluoxetine, stating that it should be used with caution in patients with congenital long QT syndrome, a previous history of QT prolongation, a family history of long QT syndrome or sudden cardiac death, and other conditions that predispose to QT prolongation, including CYP450 2D6 inhibitors or PM genotypes.
MAOIs lack anticholinergic adverse effects but may cause orthostatic hypotension, bradycardia, and shortened PR and QTc intervals on ECG (McGrath et al. 1987).
Psychostimulants, as well as atomoxetine, can mildly raise heart rate and blood pressure, but there is no compelling evidence that they in-crease QTc. There are rare reports in the literature of PVCs or ventricular or supraventricular tachycardias, as well as ST elevation and MI in overdose. Controversy remains about the necessity of routine ECG screening before the initiation of stimulant therapy (e.g., for attention-deficit/hyperactivity disorder). Stimulants can slightly raise heart rate and blood pressure, but the risk for sudden cardiac death from stimulants appears to be no higher than the background rate seen in the general population. Nonetheless, an AHA Scientific Statement suggests that in children and adolescents, it is “reasonable” to obtain a prestimulant ECG to identify underlying risk factors for sudden cardiac death, such as hypertrophic cardiomyopathies, long QT syndrome, and preexcitation or reentrant arrhythmias such as Wolff-Parkinson-White syndrome (Vetter et al. 2008). No similar AHA recommendation exists regarding baseline ECG screening in adults. More important than whether or not one obtains an ECG is the process by which a clinician evaluates cardiovascular risk before starting any sympathomimetic agent. An appropriate history would include assessment of the following:
A history of fainting or dizziness
Chest pain or shortness of breath on exertion
History of heart murmur or known arrhythmia
Family history of unexplained or sudden death before age 35
Family history of cardiac arrhythmias
Decisions to obtain a baseline ECG or consultation with a cardiologist before stimulant initiation in adults are usually best determined in case-by-case fashion depending on an individual’s history and suspected cardiac risk factors.
Other Cardiac Disturbances
Cardiomyopathies have been reported to occur during treatment with clozapine and risperidone. Myocarditis may occur with clozapine (■), and pericarditis is a rare adverse effect of gabapentin.
Elderly patients with underlying cerebrovascular or cardiovascular disease may be at greater risk for cerebrovascular events from use of SGAs, although causal links remain controversial. Clinicians should recognize the presence of baseline vascular disease when formulating risk-benefit treatment decisions.
Four controlled treatment studies of risperidone for dementia-related psychosis from 1999 to 2003 identified a significantly increased risk for transient ischemic attacks or cerebrovascular accidents with risperidone (4%) compared with placebo (2%) (Wooltorton 2002), prompting a subsequent FDA public health advisory linking the use of SGAs with an increased risk for cerebrovascular accidents. Subsequent case-control studies that failed to replicate these observations have pointed to possible confounding factors, such as vascular dementia, hypertension, and other noniatrogenic risk factors for transient ischemic attacks or cerebrovascular accidents that may have been overrepresented among antipsychotic recipients. At the same time, because SGAs demonstrate low efficacy but high dropout rates in dementia-related psychosis, enthusiasm for their use is modest. The absence of highly effective remedies for dementia-related psychosis requires thoughtful risk-benefit analyses when considering treatments, in addition to close monitoring for adverse neurological or cardiovascular changes.
Clinicians should be aware of psychotropic drugs that may be associated with peripheral edema and assure the absence of other etiologies (i.e., cardiogenic factors, hepatic dysfunction, lymphatic obstruction, nephrotic syndrome, hypothyroidism) by history taking, physical examination, and appropriate laboratory measures (i.e., thyroid-stimulating hormone, serum protein, liver enzymes, urinalysis). After determining the absence of other noniatrogenic causes, diuresis may be advisable as an initial step in management.
Peripheral edema has a wide differential diagnosis that can include many drug-induced etiologies. It may result from a number of drug classes, including vasodilators (e.g., β-blockers and α1-adrenergic receptor antagonists), calcium channel blockers, NSAIDs, estrogen-containing compounds, and thiazolidinediones. The mechanisms by which some psychotropic agents may cause peripheral edema are not well understood, although it is thought that extravasation from capillary beds in the lower extremities may result from vasodilation caused by α1-blocking drugs, such as antipsychotics. Edema caused by some GABAergic anticonvulsants (e.g., divalproex, gabapentin, tiagabine) has been hypothesized to result from direct GABAergic effects on peripheral vascular resistance.
Case reports of lithium-associated peripheral edema date to the early 1970s and are thought to reflect redistribution of sodium from the intracellular to the extracellular compartment (water follows sodium) or alterations in renal tubular absorption of sodium. The phenomenon does not appear related to serum lithium levels or dosing toxicity and may variably resolve either spontaneously, via dosage reductions, or through coadministration of a potassium-sparing diuretic such as amiloride at a dosage of 5 mg bid.
Manufacturers’ package insert information identifies peripheral edema as an uncommon side effect associated with olanzapine (3%) and divalproex (1%–5%), in the latter often arising during long-term therapy. In FDA registration trials of pimavanserin for psychosis associated with Parkinson’s disease, peripheral edema was more common with active drug (7%) than with placebo (2%). Edema has also been reported in postmarketing studies during treatment with quetiapine, ziprasidone, risperidone, gabapentin, escitalopram, trazodone, and mirtazapine. Case reports also have described peripheral edema arising during cotherapy with divalproex plus risperidone or divalproex plus quetiapine.
The evaluation of peripheral edema involves the following considerations, which can greatly facilitate communication exchange and collaboration with primary care physicians when appropriate.
Is the edema unilateral or bilateral? Unilateral edema in a lower extremity suggests an etiology that is less likely pharmacological than structural-anatomical (e.g., deep venous thrombosis, pelvic malignancies, lymphedema) or infectious (e.g., cellulitis).
Is there a suggestion of cardiopulmonary (rather than medication-induced) origin? Edema caused by congestive heart failure or pulmonary hypertension will typically present with physical examination findings involving the lungs (e.g., wheezing or rales), heart (e.g., presence of an S3 gallop), head and neck (e.g., jugular venous distension), or abdomen (e.g., hepatomegaly in right-sided heart failure). The presence of jugular venous distension suggests either a cardiac or pulmonary origin or the possibility of acute renal failure.
Is the onset acute or chronic? Edema caused by medications typically has a relatively acute onset in close proximity to the initiation of a suspected causal medication. By contrast, chronic lower-extremity edema in adults over age 50 is most often caused by venous insufficiency.
Are electrolyte abnormalities present? Hyponatremia may lead to peripheral edema due to fluid overload (e.g., water intoxication from polydipsia).
Is an eating disorder present? Rebound fluid retention can occur from alternating patterns of dehydration, vomiting, and laxative or diuretic abuse.
Are serum protein and albumin normal? Peripheral edema that is likely related to medications typically involves normal serum protein levels. Other medical etiologies of peripheral edema with normal serum protein levels include severe hypothyroidism, lymphedema, and angioedema (see the section “Allergic Reactions and Angioedema” in Chapter 20, “Systemic Reactions”). The presence of low serum protein and albumin raises the suspicion of hepatic or renal disease or of severe malnutrition.
Is proteinuria present? Low serum protein or albumin levels with proteinuria point to nephrotic syndrome.
Are liver enzymes normal? In the absence of proteinuria, abnormal liver enzymes suggest primary hepatic disease (e.g., cirrhosis).
Table 7–5 summarizes key points for psychiatrists in the assessment and management of peripheral edema.
• Determine the timing, onset, location, and extent of edema. Clarify if patient has a history of trauma, surgery, or predisposing medical conditions (e.g., past malignancy or liver disease) that may indicate etiologies other than drug-induced edema.
• By physical examination, determine location of edema (lower extremities [ankle or pretibial], upper extremities, sacral regions when recumbent), bilaterality, and degree of pitting. The presence of ulcerations or warmth may suggest cellulitis. Determine normal bilateral pedal and popliteal pulses. Palpable calf cords or tenderness, calf warmth, or pain on dorsiflexion may suggest deep venous thrombosis. Affirm normal cardiac and breath sounds by auscultation, and the absence of jugular venous distension. Physical examination may also help to assure the absence of ascites or hepatomegaly.
• Ensure that serum protein, liver enzymes, and electrolytes are normal; determine the absence of proteinuria (by urinalysis).
• Consider diuresing normokalemic and normonatremic patients with a brief (e.g., 5-day) course of furosemide or spironolactone.
• Track progress of diuresis by measuring baseline and daily weights.
• Obtain primary medical consultation if edema persists or if drug discontinuation and diuresis fail to resolve the problem, or when indicated when laboratory abnormalities or physical examination findings suggest a noniatrogenic underlying process.
If the absence of other medical etiologies has been established, the clinician can treat drug-induced edema—often a benign phenomenon—either conservatively (i.e., by leg elevation or compression stockings) or with the short-term use of a diuretic. Pitting pretibial edema caused by antipsychotics or anticonvulsants may respond to a brief (e.g., 3- to 5-day) course of furosemide 10–40 mg/day or spironolactone 50–100 mg/day [maximum 100 mg qid]). Generally, monitoring serum potassium levels is unnecessary in a normokalemic individual who begins a short course of furosemide, although this may be warranted during longer-term therapy. Thiazide diuretics, such as hydrochlorothiazide 25 mg/day, also may be used but tend to produce a less robust diuresis than does furosemide.
In the case of lithium-associated peripheral edema, the potassium-sparing diuretic amiloride (e.g., 5 mg bid) is preferred to diuretics that act on the distal convoluted tubule, such as hydrochlorothiazide, because amiloride is less likely to raise serum lithium levels. Loop-acting diuretics (e.g., furosemide) also have not been shown to increase serum lithium levels (Jefferson and Kalin 1979). If electing to diurese peripheral edema with a thiazide diuretic, the clinician might choose to reduce a standing lithium dosage or to monitor serum lithium levels more closely depending on the existing lithium dosage and the duration of diuretic therapy.
Referral to a primary care doctor should occur when causes other than drug-induced edema are suspected, laboratory abnormalities or physical examination findings suggest causes other than drug-induced bilateral edema, or edema persists despite diuresis.
The clinician should keep in mind that cessation of a likely causal agent typically leads to resolution of drug-induced edema, and the decision to discontinue a presumed offending agent (rather than manage the side effect of edema) depends on the persistence of the problem, recurrence after response to acute treatment of the edema, and availability of viable alternative primary psychotropic compounds.
Patients who receive psychotropic drugs that can alter blood pressure should have their blood pressure monitored on a regular basis by the prescriber. Sympathomimetic (e.g., noradrenergic) agents reported to cause hypertension should not be administered in patients with unstable hypertension. Present recommendations are not to use sublingual or oral nifedipine for probable drug-induced hypertensive crises, but rather to obtain blood pressure measurements for suspected hypertension-induced signs (e.g., headache) and to seek medical attention as appropriate.
Catecholaminergic agents, including psychostimulants and noradrenergic agents (e.g., SNRIs, TCAs, atomoxetine, and bupropion), carry the risk of raising blood pressure and increasing heart rate via their sympathomimetic effects. In premarketing studies of venlafaxine XR (across dosages from 75 to 375 mg/day), sustained hypertension (defined as a diastolic blood pressure >90 mm Hg and ≥10 mm Hg above baseline, on three consecutive occasions) occurred in 3% of patients with major depression, with incidence rates of <1% in patients with anxiety disorders (i.e., panic disorder, generalized anxiety disorder, or social anxiety disorder). Incidence rates of sustained diastolic hypertension rose proportionally with increasing dosages, reaching 13% at dosages >300 mg/day. Among patients treated with venlafaxine XR across diagnoses in FDA registration trials, 1% of patients had a ≥20 mm Hg rise in systolic blood pressure (manufacturer’s package insert, Wyeth Pharmaceuticals). The manufacturer of venlafaxine advises against initiating therapy in patients with poorly controlled preexisting hypertension, and either decreasing the dosage or discontinuing therapy in patients with new-onset hypertension arising during treatment with venlafaxine.
Rates of sustained diastolic hypertension with desvenlafaxine (0.7%–1.3%) or venlafaxine XR (0.5%–3%) are lower than the dose-related rates seen with venlafaxine IR (3%–13%). The SNRI duloxetine was associated with trivial changes in diastolic blood pressure (<1 mm Hg) across its indications at dosages ≤120 mg/day. The time course for developing hypertension that is likely attributable to an SNRI has not been identified from clinical trials, but hypertension may be dose dependent and more likely to arise sooner rather than later after treatment initiation. In normotensive patients who are stably dosed on an SNRI but who later develop elevated blood pressure, investigation and treatment of other causes of hypertension may be warranted before it is advisable or necessary to discontinue a long-standing SNRI.
As reported in manufacturers’ prescribing information, the incidence of hypertension with bupropion is <1%, which is no different from the incidence with placebo. In FDA registration trials, hypertension occurred in only 1% of subjects taking mirtazapine, despite its noradrenergic properties, and hypertension occurred less frequently than in subjects taking TCAs (Watanabe et al. 2010).
Hypertensive crises are of well-known concern during pharmacotherapy with MAOIs. The calcium channel blocker nifedipine, and the nonselective α-adrenergic antagonist phentolamine, are traditionally the antihypertensive agents used in monitored settings to manage MAOI-associated hypertension. A contentious literature has emerged on the pros and cons of instructing MAOI recipients to carry 10-mg nifedipine capsules to swallow, bite and swallow, or place sublingually in the event of developing a headache that could be a manifestation of a hypertensive crisis. A 1996 review challenged the safety and efficacy of this practice, citing an increased risk for adverse cardiovascular or cerebrovascular consequences—with little evidence of benefit—from the practice of routinely instructing patients to take nifedipine capsules outside of a monitored setting if headaches occur during MAOI therapy (Grossman et al. 1996). Beta-blockers, with the exception of labetolol (which also blocks α-adrenergic receptors), are generally not advisable for treating MAOI-associated hypertensive crises because their blockade of β2-adrenergic receptors allows unopposed α1-adrenergic vasoconstriction, potentially making hypertension worse.
Another issue regarding cardiovascular safety with MAOIs involves the question about the necessity of discontinuing their use before surgery (or electroconvulsive therapy) in which general anesthesia is used. The manufacturers of tranylcypromine, phenelzine, isocarboxazid, and transdermal selegiline caution against the use of MAOIs during surgery that involves general anesthesia and recommend discontinuing an MAOI at least 10 days before undergoing elective surgery with general anesthesia. However, contemporary studies report no differences in blood pressure and heart rate attributable to the continued use of MAOIs during surgery involving general anesthesia, and most experts believe that discontinuing MAOIs preoperatively is not routinely necessary (El-Ganzouri et al. 1985). Nevertheless, cardiovascular safety may be jeopardized by the use of injectable sympathomimetic vasoconstrictors such as epinephrine during surgical procedures in patients taking MAOIs.
Finally, intravenous administration of ketamine exerts a CNS sympathomimetic effect and has been reported to increase systolic blood pressure by about 10–20 mm Hg transiently (up to 4 hours) in about one-third of patients during infusions for treatment of major depression.
Myocarditis and Cardiomyopathy
Clinicians should recognize that myocarditis is a rare adverse effect with clozapine and should be alert to its presenting signs (e.g., fever, chest pain, shortness of breath), particularly during the first 1–2 months after treatment initiation. ECG and measurement of serum troponin or creatine kinase-MB and CRP should be obtained, along with consultation from a cardiologist in the setting of laboratory abnormalities. Clozapine should be discontinued and not reintroduced when myocarditis occurs.
Clozapine carries a known, rare risk for myocarditis (■), which arises through poorly understood mechanisms that may include clozapine-induced release of inflammatory cytokines, hypercatecholaminemia, and type I IgE–mediated acute hypersensitivity reactions (Merrill et al. 2006). Rapid clozapine titration also has been a factor implicated in case reports of myocarditis. Eosinophilia appears not to be a useful parameter in diagnosing clozapine-induced myocarditis, although elevated CRP (>100 mg/dL) has been reported even when troponin levels are normal. The majority of cases become manifest within the first 4–8 weeks of treatment initiation, and patients can present with a variety of signs and symptoms, including fever, chest pain, shortness of breath, tachycardia, and leukocytosis. Importantly, 20% of individuals may develop a benign and transient fever during clozapine initiation, making it important for clinicians to pay close attention to other systemic features that could signal the presence of myocarditis, NMS, or other correlates of fever during clozapine therapy—particularly given the rarity of myocarditis (with a reported incidence of 0.015%–0.188%) (Merrill et al. 2006). In their systematic review of the phenomenon, Merrill et al. (2006) advised that for clozapine recipients who develop chest pain, fever, dyspnea, or flulike symptoms, clinicians should consider obtaining an ECG (observing for T-wave changes or ST segment elevation) and serum troponin or creatine kinase-MB. Consultation with a cardiologist is advisable in the presence of abnormal parameters.
Management of clozapine-induced myocarditis hinges on discontinuation of clozapine, as well as the possible use of β-blockers (e.g., metoprolol), ACE inhibitors, and diuretics; the role of corticosteroids appears more controversial (Merrill et al. 2006). Also rarely, clozapine use has been associated with development of pericarditis, dilated cardiomyopathy, or congestive heart failure—the last-mentioned manifested by exertional dyspnea, orthopnea, dizziness, and fatigue; management involves drug cessation and serial echocardiographic monitoring.
Rechallenge with clozapine after resolution of myocarditis is highly controversial. Most authorities discourage even its contemplation, and there exist only a handful of case reports in the literature describing successful rechallenge. Manu and colleagues (2012) identified only four such cases, three of which involved successful rechallenge (ranging from 2–104 weeks after initial clozapine discontinuation) with serial troponins (e.g., three times per week) and frequent ECGs. Those authors concluded that insufficient safety data exist to justify reexposure to clozapine after myocarditis.
Case reports also have described myocarditis occurring during treatment with quetiapine, potentially as a drug hypersensitivity reaction associated with eosinophilia, leukopenia, and thrombocytopenia.
Clinicians should be aware of drugs that can cause orthostatic hypotension, typically resulting from α1-adrenergic blockade. Orthostatic hypotension may not necessarily be dose related and may represent a persistent phenomenon during continued administration of a causal agent. Blood pressure and heart rate should be measured while the patient is sitting and standing, several minutes apart, to evaluate the presence and extent of autonomic changes. Increased oral hydration and salt intake may not reliably counteract pharmacological orthostatic hypotension, but avoiding dehydration may help to diminish exacerbation of the phenomenon. Patients should be advised to get up slowly from sitting or supine positions. Treatment with fludrocortisone, the vasoconstrictor midodrine, or the parasympathomimetic agent pyridostigmine may sometimes be appropriate antihypotensives. Significant orthostatic hypotension, particularly in patients with low resting blood pressure or cardiac disease, may warrant cessation of the causal agent.
Orthostatic hypotension may be associated with a wide range of psychotropic agents that cause vasodilation by blocking peripheral vascular α1-adrenoreceptors. These agents include TCAs, MAOIs (e.g., tranylcypromine at dosages >30 mg/day), mirtazapine, atomoxetine, duloxetine, venlafaxine, desvenlafaxine, and SGAs, as well as cannabinoids. Antihypertensive α1-blocking drugs such as prazosin (used sometimes to counteract nightmares; see the section “Nightmares and Vivid Dreams” in Chapter 19, “Sleep Disturbances”) and terazosin (used sometimes to counteract hyperhidrosis; see the section “Hyperhidrosis” in Chapter 8, “Dermatological System”) may cause orthostatic hypotension and syncope, particularly after a first dose. Patients who report dizziness upon standing or sitting after rising from a supine position should be examined for postural hypotension and tachycardia. Clinicians should consider other nonpharmacological etiologies when relevant, such as dehydration, prolonged bed rest, dysautonomias, anemia, and Parkinson’s disease. The use of nonpsychotropic vasodilators or diuretics also may play contributing roles. The differential diagnosis of orthostatic hypotension also includes vasovagal syncope, as might occur when sudden-onset light-headedness includes visual, auditory, and other sensory abnormalities. Conservative management of pharmacologically induced orthostatic hypotension involves counseling patients to rise slowly from seated or supine positions, avoid dehydration, and assure the safe use of any concomitant drugs that may further cause anticholinergic, vasodilatory, arrhythmogenic, or extrapyramidal adverse effects.
Occasionally, orthostatic hypotension warrants pharmacological treatment. Use of antihypotensive drugs generally has been confined to the treatment of dysautonomias or cardiovascular dysregulation in critical care settings, and none has formally been studied to counteract iatrogenic orthostatic hypotension. The α1 agonist midodrine (dosed typically at 5 mg three times daily, t½=~3–4 hours) is a generally safe and well-tolerated vasopressor in people without significant heart or kidney disease, posing few side effects of its own (headache, flushing, dry mouth). Fludrocortisone, a synthetic mineralocorticoid and plasma volume expander, is dosed initially at 0.1–0.2 mg/day (with a usual maximum of 0.4–0.6 mg/day); patients must be monitored for hypokalemia and supine hypertension (the latter especially when fludrocortisone is added to midodrine). Finally, the cholinesterase inhibitor pyridostigmine (dosed initially as 30 mg twice daily and increased to a maximum of 60 mg three times daily [or as a 180 mg/day slow-release tablet]) yields a modest pressor effect by enhancing ganglionic neuro-transmission through the baroreflex pathway; its effect can be safely amplified by combining it with low-dose (e.g., 5 mg/day) midodrine.
The symbol ■ is used in this chapter to indicate that the FDA has issued a boxed warning for a prescription medication that may cause serious adverse effects.