Association of Linezolid With Risk of Serotonin Syndrome in Patients Receiving Antidepressants

Design

Retrospective, population-based, cohort study

N= 1,134

Objective

To examine the incidence of serotonin syndrome in patients receiving oral linezolid and how concomitant antidepressant treatment changes this risk

Study Groups

Antidepressants (n= 215)

No antidepressants (n= 919)

Inclusion Criteria

Exclusion Criteria

Methods

This was a retrospective analysis of the Ontario Drug Benefit database, which captures outpatient prescription medications dispensed to persons older than 65 years old in Ontario, Canada. Patients who were prescribed linezolid were separated into those who were also taking concurrent antidepressants or not. 

The antidepressants included (1) selective serotonin reuptake inhibitors (SSRIs): paroxetine, fluvoxamine, fluoxetine, sertraline, citalopram, and escitalopram; (2) serotonin-norepinephrine reuptake inhibitors (SNRIs): venlafaxine, desvenlafaxine, and duloxetine; (3) tricyclic antidepressants: clomipramine, amitriptyline, desipramine, imipramine, nortriptyline, protriptyline, doxepin, and trimipramine; (4) monoamine oxidase (MAO)-inhibitors: isocarboxazid, phenelzine, selegiline, tranylcypromine, and moclobemide; (5) norepinephrine and dopamine reuptake inhibitors: bupropion; and (6) other: trazodone, mirtazapine, buspirone, amoxapine, maprotiline, and nefazodone.

Data was also recorded on the use of other serotonergic medications: lithium, 5HT3 receptor antagonists (dolasetron, granisetron, ondansetron, and palonosetron), metoclopramide, methylphenidate, dexmethylphenidate, fentanyl, meperidine, methadone, oxycodone, tramadol, triptans (almotriptan, eletriptan, frovatriptan, naratriptan, rizatriptan, sumatriptan, and zolmitriptan), ergot derivatives (dihydroergotamine, ergotamine, and methylergonovine), dextromethorphan, chlorpheniramine, cyclobenzaprine, carbamazepine, valproic acid, tryptophan, dextroamphetamine, rasagiline, and ritonavir.

Duration

Outcome Measures

Primary: clinically significant serotonin syndrome based on a physician's diagnosis, Sternbach criteria, or the Hunter Serotonin Toxicity Criteria within 30 days of starting oral linezolid treatment

Secondary: altered mental status, hospitalization, or death within 30 days of starting linezolid

Baseline Characteristics

Antidepressants (n= 215)

No antidepressants (n= 919)

Age, years

66-69

70-79

≥80

22.3%

38.6%

39.1%

19.3%

42.4%

38.3%

Female

Mean days of linezolid treatment

Number of concurrent antidepressants

Taking other serotonergic drugs

Results

There were fewer than 6 cases of serotonin syndrome reported overall. The exact numbers were not reported per the data policy of the database due to potentially identifiable patient information. Based on fewer than 6 events, the possible risk difference for serotonin syndrome ranged from -0.5% to 2.3%. The number of serotonin syndrome cases were fewer in the antidepressant group.

Using propensity-score matching, the adjusted risk of serotonin syndrome was lower in the antidepressant group (adjusted risk difference, -1.2%; 95% confidence interval, -2.9% to 0.5%; p= 0.50). Per this data, the worst-case scenario would be a 0.5% increase in the risk of serotonin syndrome due to antidepressants, which is equivalent to a number needed to harm of 200.

Secondary endpoints within 30 days of linezolid

Altered mental status/confusion

Hospitalization

Death from any cause

11.6%

42.8%

7.4%

6.9%

43.3%

4.8%

0.02

0.93

0.13

Adverse Events

Study Author Conclusions

In this cohort study, concomitant antidepressants did not significantly increase the risk of serotonin syndrome in patients taking linezolid. Within the 95% CI, the worst-case scenario would be a 0.5% increase in the risk of serotonin syndrome due to antidepressants, which would be a number needed to harm of 200. As such, it is likely safe to prescribe linezolid in patients taking antidepressants, and antidepressants should not be an absolute contraindication for linezolid.

InpharmD Researcher Critique

Inherent limitations of this retrospective, database analysis include the possibility of confounding variables and incomplete data sets. For the primary endpoint, serotonin syndrome was diagnosed retrospectively based on a constellation of diagnosis codes, which did not match perfectly with the diagnostic criteria. Additionally, the small number of events seen limit the analysis and accuracy of the results.

Incidence of Serotonin Syndrome With Combined Use of Linezolid and Serotonin Reuptake Inhibitors Compared With Linezolid Monotherapy

Design

Retrospective, single-center, case-control study

N= 348

Objective

To compare the incidence of serotonin syndrome when linezolid was administered alone and in combination with selective serotonin reuptake inhibitors (SSRIs) or serotonin-norepinephrine reuptake inhibitors (SNRIs)

Study Groups

Combination therapy (n= 87)

Linezolid monotherapy (n= 261)

Inclusion Criteria

Adult inpatients who received at least 1 dose of linezolid while admitted; use of an SSRI or SNRI during linezolid or within 14 days before linezolid

Exclusion Criteria

Methods

This was a retrospective study from a single academic medical center in Iowa. Patient records were screened for an SSRI or SNRI administration during or within the 14 days preceding treatment with linezolid. Any patient with at least 1 inpatient administration of an SSRI or SNRI simultaneously with or up to 2 weeks before an inpatient dose of linezolid was included in the combination therapy group.

Patients who did not receive an inpatient dose of an SSRI or SNRI with or within the 2 weeks before linezolid were eligible for inclusion in the linezolid monotherapy group as long as the patient did not also have any outpatient administrations of an SSRI or SNRI in the 2 weeks before linezolid, as determined by medication histories before admission and discharge medication lists.

Duration

Outcome Measures

Diagnosis of serotonin syndrome, including the presence of at least 3 Sternbach criteria or 1 Hunter criterion

Baseline Characteristics

Combination therapy (n= 87)

Linezolid monotherapy (n= 261)

Age, years

Male

Median length of stay, days

ICU admission

Median ICU length of stay, days

60.9%

2

59.4%

2

0.801

0.587

Median linezolid doses

Received other serotonergic agents*

Tricyclic antidepressants

Other antidepressants

Analgesics

Antiemetics

Migraine therapy

94.3%

3.4%

37.9%

65.5%

67%

3.4%

88.1%

5.4%

22.6%

55.9%

57.1%

0

0.104

0.578

0.005

0.117

0.050

0.015

Of the patients who received a concurrent SSRI or SNRI, citalopram was the most common agent (41.4%).

Additionally, 23% of patients receiving combination therapy had the SSRI or SNRI discontinued before linezolid use

ICU: intensive care unit

*Agents other than SSRIs or SNRIs

Results

Combination therapy (n= 87)

Linezolid monotherapy (n= 261)

p-value

Possible serotonin syndrome diagnosis

Met at least 3 Sternbach criteria

Met at least 1 Hunter criterion

12 (13.8%)

1 (1.1%)

35 (13.4%)

5 (1.9%)

0.928

0.533

Documented serotonin syndrome diagnosis

No significant difference was seen in the incidence of serotonin syndrome between the combination therapy and monotherapy groups (relative risk, 3.00; 95% confidence interval, 0.19 to 47.45).

Adverse Events

Study Author Conclusions

InpharmD Researcher Critique

Linezolid and Serotonergic Drug Interactions: a Retrospective Survey

Design

Retrospective chart review

N= 72

Objective

Study Groups

Concomitant linezolid and SSRI (n= 52)

Linezolid and SSRI within 14 days (n= 20)

Inclusion Criteria

Received linezolid and an SSRI or venlafaxine either concomitantly or within 14 days of each other

Exclusion Criteria

N/A

Methods

Information was collected from inpatient medical records. Sternbach criteria and Boyer algorithm for diagnosis of serotonin syndrome were used to identify indicative clinical features. Patients with a suspected diagnosis were stratified by low or high probability of having experienced serotonin syndrome based on clinical features. 

Low probability was described as meeting at least 3 Sternbach criteria, any Boyer criteria, and symptoms did not progress with continued therapy of if concomitant therapy was stopped but symptoms did not resolve. High probability was described as meeting three or more Sternbach criteria or any Boyer criteria with no clear alternative explanations or reversal of symptoms with discontinuation of therapy. 

Duration

April 2000 - November 2004

Outcome Measures

Probability of serotonin syndrome

Baseline Characteristics

Results

Endpoint

Concomitant linezolid and SSRI (n= 52)

Linezolid and SSRI within 14 days (n= 20)

Probability of serotonin syndrome

High

Low

2 (4%)

1 (2%) 

0

1 (5%) 

Two patients had a high probability of serotonin syndrome. Only one patient had serotonin syndrome listed as a diagnosis on their medical records. This patient was a 30-year-old female who received concomitant linezolid and sertraline, as well as trazodone and fentanyl. The patient was documented as having four Sternbach clinical criteria: agitation (anxiety), myoclonus, diaphoresis, and shivering, as well as two Boyer criteria: spontaneous clonus and inducible clonus with either agitation or diaphoresis. 

The second patient with high probability of serotonin syndrome did not have a diagnosis documented in the medical record. This 81-year-old female received concomitant therapy with linezolid and venlafaxine, followed by citalopram. Six Sternbach clinical criteria were met: mental status changes, agitation, myoclonus, hyperreflexia, tremor, and incoordination and four Boyer criteria were met: tremor and hyperreflexia, spontaneous clonus, muscle rigidity with a body temperature of 138􏰁 °C and ocular or inducible clonus, and ocular clonus with either agitation or diaphoresis. 

Adverse Events

N/A

Study Author Conclusions

InpharmD Researcher Critique

Many of the patients (65 of 72) in this study received a serotonergic medication other than linezolid or an SSRI, increasing the risk for signs and symptoms of serotonin syndrome. The accuracy of information assessed by this study was limited to what was documented in patient medical records and clinical judgment of healthcare providers present when symptoms occurred.