To Name :
To Email :
From Name :
From Email :
Comments :

Cases that Test Your Skills

Depression, medication, and ‘bad blood’

Clinicians find 2 antidepressants that reduce Mr. G’s chronic depression. Unfortunately, each medication decreases his WBCs. What would you do?

Vol. 6, No. 5 / May 2007

CASE: Sad and suicidal

Mr. G, age 44, has chronic depression with suicidality. At presentation he says he has felt sad and suicidal for 2 weeks. He also has no appetite and trouble sleeping at night.

Mr. G’s depression has left him unable to work and has led to 4 hospitalizations over 10 years. He first attempted suicide in 1984 after his ex-wife took their child and left him. He endorses no suicide plan and has been sober for 7 years after 12-plus years of alcohol abuse, but says he has been tempted lately to resume drinking.

The patient was taking an antidepressant but stopped while at a homeless shelter, where he had been staying for several weeks. For more than 20 years, he also has been taking phenytoin, 300 mg/d, and phenobarbital, 30 mg bid, for a seizure disorder.

Mr. G is admitted with a working diagnosis of recurrent major depressive disorder. White blood cell count (WBC) at admission is 5.12×109/L and neutrophils are 3.6×109/L—both low-normal readings. Other laboratory results are normal.

We continue phenytoin and phenobarbital at the same dosages and start the selective serotonin reuptake inhibitor (SSRI) citalopram, 20 mg/d, which interacts minimally with both anticonvulsants.

After 2 weeks, Mr. G’s seizures are well controlled and he is tolerating citalopram, but his depressive symptoms have not improved. We cross-taper citalopram to prevent SSRI-induced discontinuation syndrome and start the dopamine and norepinephrine reuptake inhibitor bupropion, 75 mg bid. We titrate bupropion over 2 weeks to 150 mg each morning and 300 mg at bedtime, and watch Mr. G closely for seizures. Although his seizure history contraindicates bupropion use, we think he can tolerate the medication because his seizure disorder is well controlled.

Mr. G’s affect, appetite, and energy are improving with bupropion, but a routine complete blood count (CBC) 5 days after the medication is started reveals leukopenia (WBC 3.04×109/L) without neutropenia (neutrophils 1.9×109/L). Repeat blood tests 18 and 32 days after the first blood draw show continued low WBC. The gastrointestinal medicine team tests Mr. G’s liver function but finds no abnormalities.

The author’s observations

Mr. G’s low WBC and neutrophil counts coincided with bupropion use, suggesting medication-induced leukopenia. Phenytoin can cause neutropenia and other adverse hematologic effects,1 but the patient had been using phenytoin and phenobarbital for years with no adverse reactions.

A medical cause also is unlikely. Mr. G’s liver function is normal, and he shows no other signs or symptoms of a medical problem. Bone marrow biopsy and immunologic workup could rule out cancer, but the timing of Mr. G’s abnormal blood readings strongly suggests bupropion intolerance.

TREATMENT: Other medications

We immediately stop bupropion, start the serotonin-norepinephrine reuptake inhibitor (SNRI) venlafaxine at 37.5 mg bid, and titrate it over 5 days to 225 mg/d. Blood draws 3 and 5 days after bupropion discontinuation show slight increases in WBC.

Eleven days after venlafaxine is started, Mr. G’s WBC and neutrophils are normal. However, he has become increasingly irritable and volatile, often arguing with a staff nurse and other patients. We cross-taper venlafaxine over 5 days, start the SSRI sertraline at 50 mg/d, and titrate sertraline over 1 week to 150 mg/d. Mr. G’s irritability and depressive symptoms improve at the latter dosage.

Because Mr. G developed neutropenia while taking a medication not associated with this adverse effect, we start watching his WBC counts more closely than usual. WBC is 4.58×109/L 8 days after sertraline is started but falls to 3.4×109/L after another 8 days, with neutrophils at 1.5×109/L for both readings (Table).

We add lithium, 300 mg bid, to increase Mr. G’s neutrophils and augment sertraline’s antidepressant effects. Four days later, WBC is 5.8×109/L with neutrophils at 4.2×109/L.

We stop lithium briefly to see if WBC remains normal. After 3 days, WBC drops to 3.25×109/L with neutrophils at 1.5×109/L. We restart lithium, 300 mg/d, and Mr. G’s WBC increases to 4.18×109/L 4 days later, with neutrophils at 2.1×109/L.


Mr G’s white blood cell (WBC) and neutrophil counts (NC)*
while taking bupropion and sertraline


When measurements were taken



None for several weeks

Baseline, first hospital admission



Bupropion, 75 mg bid

5 days after starting bupropion



Bupropion, 450 mg/d total

23 days after starting bupropion



Bupropion, 450 mg/d total

2 weeks after previous test



Sertraline, 150 mg/d

8 days after starting sertraline (titration period)



Sertraline, 150 mg/d

16 days after starting sertraline



Sertraline, 150 mg/d, and lithium, 300 mg bid

4 days after lithium augmentation



None for 3 months

Baseline, second hospital admission



Sertraline, 150 mg/d

12 days after restarting sertraline


Not available

* Normal WBC values: 4.5 to 11×109/L; normal neutrophil values: 1.5 to 8×109/L

The authors’ observations

For Mr. G, both bupropion and sertraline appear to have caused neutropenia on separate occasions.

To our knowledge, bupropion-induced leukopenia or neutropenia have not been reported in the literature. Neutropenia—a rare adverse effect of antidepressants2—and leukopenia were seen during bupropion’s pre-marketing trials but were not definitely attributed to the drug.1 According to pre- and post-marketing data, leukopenia was “infrequently” reported among 5,100 subjects who received bupropion.3

To our knowledge, sertraline-induced neutropenia has not been reported in nongeriatric patients, although sertraline-induced neutropenia4 and agranulocytosis5 have been reported in patients age >65. The Committee on Safety of Medicine in the United Kingdom has received 2 other reports of neutropenia and 1 report of leukopenia with sertraline.5

In one clinical trial, 2 of 1,304 patients taking unknown dosages of sertraline had low neutrophils (<15% of WBC). Incidence of abnormal hematologic readings did not differ significantly between the sertraline and placebo groups (data on file, Pfizer).

Medication is the second most common cause of acquired neutropenia, with infection being most common.6 By definition, drug-induced neutropenia occurs within 4 weeks after starting the drug and usually resolves within 30 days after stopping it.

Neutropenia is an idiosyncratic reaction unrelated to pharmacologic action. Although overall neutropenia incidence is unknown, reported incidence of the rare, more severe agranulocytosis ranges from approximately 1 to 10 cases per million people annually, and medications have been implicated in 70% of these cases.6 Conversely, only 2 of 97 incidental neutropenia cases studied by Lima et al7 were medication-induced.

Drug-induced neutropenia can result from immune-mediated destruction of neutrophils by circulating antibodies or from direct toxic effects upon marrow granulocyte precursors. Whereas immune-mediated onset is acute and explosive, toxic effect is insidious (months to years) and asymptomatic.8 Clozapine is thought to deliver a direct toxic effect, whereas the thyroid-regulating drug propylthiouracil generates anti-neutrophil antibodies.9

Mr. G’s acute onset (within 5 to 16 days of starting bupropion or sertraline) and prompt return of neutropenia after stopping lithium suggest acute immune-mediated circulating neutrophil destruction.

Treating leukopenia

After 4 failed or intolerable antidepressant trials, lithium augmentation seemed reasonable and ultimately improved Mr. G’s neutrophil count and his mood.

Lithium has helped resolve clozapine-induced neutropenia in case reports.10-12 Well-controlled studies, however, have followed only patients with antineoplastic, drug-induced neutropenia.1

By acting on cyclic nucleotides, lithium prompts colony-stimulating factor production, which in turn stimulates neutrophil production by pluripotent stem cells. As with Mr. G, patients reach neutrophilia 3 to 7 days after starting lithium.

If the patient cannot tolerate lithium, try switching antidepressants or using growth factors to increase neutrophils.

Switching antidepressants.The SSRIs escitalopram or paroxetine, or the SNRI duloxetine are effective and do not necessarily cause neutropenia. Start at below-normal dosages to gauge tolerability, then titrate to normal dosages. Avoid tricyclics, which pose a higher risk of neutropenia than other antidepressant classes.

Case reports13,14 associate fluoxetine and mirtazapine with neutropenia. The patient who received mirtazapine, 30 mg/d, later responded well to sertraline, 50 mg/d.13

If the new antidepressant is ineffective, consider adding the mood-stabilizing anticonvulsant lamotrigine, 12.5 mg/d. Increase lamotrigine to 25 mg/d after 1 week, then titrate by 25 mg weekly to 100 to 400 mg/d depending on efficacy and tolerability.

Although lamotrigine has been associated with neutropenia in case reports,1 it is safer than other anticonvulsants. Carbamazepine, oxcarbazepine, and valproic acid can cause blood dyscrasias, which can lead to serious infection, abnormal bleeding, or other complications.

Using growth factors.Although their efficacy is not proven, growth factors are minimally toxic and might have helped Mr. G. Granulocyte colony-stimulating factor and granulocyte macrophage colony-stimulating factor resolved neutropenia in uncontrolled studies, but results of one randomized controlled trial were equivocal.8

TESTING: CT findings

Approximately 2 months after admission—shortly after a blood draw shows normal WBC and neutrophils—Mr. G complains of dizziness. He says he accidentally hit his head against a side table.

We order a full neurologic workup to check for traumatic brain injury or brain damage caused by long-term alcohol abuse:

  • Head CT shows evidence of previous cerebrovascular infarcts in the bilateral frontal and cerebellar lobes and basal ganglia.
  • MRI shows atrophied mammillary bodies, fornix, and corpus callosum.
  • Magnetic resonance angiography reveals small cerebral vessel disease.

These findings and subsequent neuropsychiatric test results suggest an organic cause of depression, likely secondary to 12 years of alcohol abuse. In light of this new information, we change Mr. G’s diagnosis to mood disorder with depressive features secondary to a general medical condition.

FOLLOW-UP: Awaiting discharge

After 3 months of continuous hospitalization, Mr. G has become euthymic and nonsuicidal, though at times oversensitive and combative. We transfer him to an assisted-living center and continue sertraline, 150 mg/d; phenytoin, 300 mg/d; phenobarbital, 30 mg bid; lithium, 300 mg/d; and trazodone, 50 mg at night as needed for insomnia.

We also place Mr. G in a day treatment program for mentally ill chemical abusers. A psychiatrist sees him every 2 weeks, and staff supervise him daily.

The authors’ observations

Mr. G’s extended hospital stay allowed us to closely observe him and offered ready access to laboratory facilities while we cross-tapered medications. In outpatient treatment, however, a serious and life-threatening medication-induced complication could easily be missed.

If economically feasible, take CBCs for all patients before prescribing any medication that could cause neutropenia, such as an antidepressant or mood stabilizer. Make sure geriatric or medically ill patients have had a CBC ≤3 months before presentation and are seeing a primary care physician as needed. Order follow-up CBC for these patients 1 month after presentation, then every 6 months if CBC is normal.

For medically healthy outpatients, be sure CBC has been checked ≤6 months before presentation. Monitor CBC and urge the patient to see a primary care doctor if infection symptoms emerge. Watch for gingivitis, tooth abscess, and other oral cavity infections—which often are overlooked—and sore throat or fever.

Also check electrolytes and screen for SSRI-induced hyponatremia at baseline for all at-risk patients.

Stop the offending drug when WBC reaches <2×109/L or with absolute neutrophil count (ANC) <1.5×109/L, then take a peripheral smear to confirm neutropenia. If the patient is asymptomatic, check ANC 2 to 3 times weekly, particularly if he or she recently had an infection or started a medication that can cause neutropenia. Neutropenia should resolve within 6 to 8 weeks of stopping the offending drug.

If neutropenia persists, order bone marrow biopsy in collaboration with an internist or hematologist to test for cancer. If the biopsy is negative, test for:

  • HIV infection
  • antinuclear antibodies to check for collagen vascular disease
  • antineutrophil antibody to rule out immune neutropenia
  • serum folate and B12 deficiency secondary to low WBC.

Also perform an immunoglobulins/immune evaluation to check for defects in cellular or humoral immunity, and bone marrow culture to test for infection.8

FOLLOW-UP: Stressor and relapse

Seven months later, Mr. G is readmitted for depression. Three months earlier, he had stopped all medications and resumed drinking after a family member died. WBC at admission is 3.70×109/L

Did you miss this content?
Psychosis and catatonia after dancing with a dangerous partner