Recognition and treatment of serotonin syndrome

Case description

Mr J.W., an 80-year-old man with a history of depression, was admitted to hospital with pneumonia. His condition deteriorated and he was sent to the intensive care unit (ICU) and placed on mechanical ventilation for several days. During his stay in the ICU, his medications were reviewed. His use of fluoxetine, which he’d taken for depression for almost 10 years, was discontinued. Following discharge from the ICU, his mood was noted to be low by the medical staff. Approximately 1 week after discontinuation of the fluoxetine, he started using 20 mg of paroxetine daily.

Within 24 hours of starting paroxetine, Mr J.W. was found to be confused and agitated with periods of unresponsiveness. Vital signs revealed a temperature of 38.5°C and a pulse of 115 beats per minute. A neurological examination revealed myoclonus in all limbs with any stimulation. Assessment by a family physician with training in care of the elderly identified serotonin syndrome as the probable cause of his presentation.

The paroxetine was discontinued and the patient was given intravenous fluids to decrease the risk of renal failure. Mr J.W. initially received 2 mg of lorazepam intravenously, then received 3, 1-mg doses of lorazepam every 4 hours, resulting in decreased tachycardia, hypertonicity, and clonus. He was discharged from hospital without antidepressants, and he planned to follow up with his family doctor several weeks after hospitalization in order to have his mood reviewed.

Clinical presentation

The original and revised criteria5,6 for diagnosis of serotonin syndrome emphasize the mental status changes and clinical findings typically observed in more severe cases. As a result, these criteria have been criticized for not highlighting the wider spectrum of toxic side effects that can be seen with this disorder. For this reason, some authors prefer to use the term serotonin toxicity to include findings of earlier or milder cases.7

It is important to note that symptoms of serotonin syndrome usually present within 6 to 8 hours of initiating or increasing serotonergic medications.7,8 The onset tends to be more acute than in a condition such as neuroleptic malignant syndrome, which shares some other features with serotonin toxicity.

Serotonin syndrome usually manifests with autonomic changes, mental status changes, and neurological findings. Mild, moderate, and severe signs and symptoms of serotonin syndrome are summarized in Table 2.4 There are no data on the percentage of cases in the mild, moderate, or severe categories, in part because of the low rates of diagnosis of mild cases.

Medication factor

Various medications are associated with serotonin syndrome.9–15 Medications that affect any of the steps in serotonin metabolism or regulation can provoke toxicity. Antidepressants, particularly selective serotonin reuptake inhibitors (SSRIs), are among the most commonly implicated drugs. Drug interactions can cause severe toxicity and can involve common over-the-counter medications. Drugs with long half-lives can interact several weeks after discontinuation; fluoxetine interactions have been reported up to 5 weeks after discontinuation.16 Even single doses of SSRIs have been implicated in serotonin syndrome. It is relevant to note that serotonin syndrome is not considered to be an idiosyncratic drug reaction but is a predictable response to elevated levels of serotonin.11 Table 39,11–15 provides a list of important drugs and drug interactions and the mechanism of increasing serotonin levels.

Overdosing with single agents such as SSRIs does not often cause severe toxicity. Approximately 15% of SSRI overdoses result in moderate symptoms. Severe cases appear to be more likely after drug interactions, particularly monoamine oxidase inhibitors (MAOIs) interacting with other antidepressants or with serotonin releasers, such as amphetamines.11 Patient factors associated with developing toxicity are not clearly identified in the literature but might relate to factors affecting the pharmacokinetics of offending agents (eg, renal and hepatic failure). Patients who are genetically deficient in the cytochrome P450 2D6 enzyme (8% of whites) are more susceptible if they are taking drugs such as venlafaxine, paroxetine, tricyclics, dextromethorphan, and methadone.17

Diagnosis

A thorough history of current and recent medication use is important, as is ruling out the use of illicit drugs and dietary supplements. Clinicians should also consider pupil size and reactivity, skin colour, presence of diaphoresis, dryness of oral mucosa, and presence or absence of bowel sounds. Diagnosis is clinical and investigations are mostly done to identify sequelae of severe toxicity, including renal function, creatine kinase, myoglobin, and clotting parameters.18

As previously mentioned, the original criteria for diagnosing serotonin syndrome describe only severe cases. To help delineate the spectrum of illness, researchers in Australia reviewed more than 2000 cases of self-poisoning with serotonergic drugs and related the clinical findings to a diagnosis of serotonin syndrome made by a clinical toxicologist. From this information, a clinical decision tree was developed. This clinical approach has better sensitivity and specificity than older criteria (sensitivity 84%, specificity 97%).7 These decision rules are found in Table 4.7

Several conditions should be considered when evaluating patients with confusion and neurological changes. Neuroleptic malignant syndrome (NMS) bears some resemblance to serotonin syndrome, with similar symptoms of fever, mental status changes, and altered muscle tone. However, patients with NMS are usually akinetic with rigidity, have decreased levels of consciousness, and are more likely to have mutism rather than rambling speech, which is associated with serotonin toxicity. More important, the onset of NMS is slow, developing over days rather than hours.19

One cause of delirium with neurological hyperexcitability is metabolite-mediated opiate toxicity, which can occur with rapid escalation of opioid doses, high opioid doses, or low opioid doses in patients with renal failure. Opiate toxicity is most likely caused by meperidine and morphine but can be seen with all opioid agents.20

Anticholinergic toxicity is differentiated by presence of skin colour changes (red as a beet), dry mouth (dry as a bone), and constipation or absence of bowel sounds .4,19

Treatment

Prompt recognition of toxicity and discontinuation of offending medications are most important. Many cases of the syndrome are self-limiting if medications are stopped early. The mortality of severe serotonin syndrome is estimated to range from 2% to 12%. The Toxic Exposure Surveillance System in the United States reported 93 deaths due to serotonin syndrome in 2002.1

Treatment is guided by the severity of presentation. There are currently no randomized placebo-controlled trials to guide pharmacologic treatment of serotonin syndrome, although case reports and series suggest several medications that are beneficial in managing the condition.

Supportive care, including intravenous fluids, is indicated in patients with vital sign abnormalities. Neurological symptoms, including serious myoclonus and hyperreflexia, are sometimes treated with benzodiazepines, but there is little evidence for this approach. Hyperthermia should be aggressively managed with external cooling, hydration, and benzodiazepines (eg, diazepam, lorazepam). Patients with a temperature higher than 41°C should be intubated with induced neuromuscular paralysis. There is a limited role for traditional antipyretics, as the mechanism of serotonin syndrome is due to muscle tone rather than central thermoregulation. Patient agitation secondary to delirium should be managed using nonpharmacologic strategies and benzodiazepines, when appropriate. Physical restraints should be avoided, as they can increase hyperthermia, lactic acidosis, and rhabdomyolysis (level II evidence).

The antihistamine cyproheptadine, which is also a 5-HT_2A inhibitor, should be considered in moderate cases and is recommended in severe cases, despite a lack of randomized controlled trial evidence (level II evidence). It is available only as an oral preparation; the initial dose is 12 mg; the dosage is then adjusted to 2 mg every 2 hours until symptoms improve.21,22

Some antipsychotics have 5-HT_2A antagonist effects and are sometimes used (level II evidence). Sublingual olanzapine and intramuscular chlorpromazine (50 to 100 mg) are options.23,24 Chlorpromazine can cause serious hypotension and should be avoided in severe cases with shock.

Caution is required in using antipsychotics for treating serotonin syndrome, as NMS can be misdiagnosed as serotonin syndrome. Also, case reports have implicated antipsychotics as precipitants of serotonin syndrome in individuals receiving concomitant serotonergic medications.1,25

Dantrolene, a skeletal muscle relaxant used for treatment of NMS, has been reported to improve symptoms of serotonin syndrome in a case series26; however, it has also been implicated in the development of serotonin toxicity and is not generally recommended.4,11 Propranolol, which has 5-HT_1A antagonist activity and a long half-life, can potentiate hypotension and make improvement in tachycardia a less effective strategy for monitoring response to treatment.4,22

Mood management

No articles were found that assessed risk in depressed patients needing ongoing mood treatment after resolution of serotonin syndrome. Given that it is a predictable reaction, however, reinitiation after adequate washout should not result in increased risk, especially if an agent with less serotonergic activity is chosen (eg, bupropion, mirtazapine).10

All patients newly started on antidepressant medications should be assessed for suicidal thoughts. A recent Canadian study27 reviewed coroners’ reports of suicides and found that patients treated with SSRIs were 5 times more likely to have committed suicide within the first month of treatment than patients treated with other antidepressants. The absolute increase in risk was small (29/100 000 for SSRIs versus 6.2/100 000 for other anti-depressants). It is very important to note that patients who were not treated at all were at the highest risk.

This recommendation of increased monitoring of patients newly started on antidepressants is echoed in the guidelines on depression by the Canadian Coalition for Seniors’ Mental Health. These guidelines also suggest monitoring for serotonin-related side effects and suicidal thoughts early in treatment. Although a challenge in office practice, the guidelines suggest patients be seen weekly for at least several weeks after initiating treatment.28,29

Levels of evidence

Level I: At least one properly conducted randomized controlled trial, systematic review, or meta-analysis

Level II: Other comparison trials, non-randomized, cohort, case-control, or epidemiologic studies, and preferably more than one study

Level III: Expert opinion or consensus statements