When you are planning to start a patient on medications, what labs should you order at baseline, and what should you order over time? This is quite a different question from whether to do a set of labs at the start of treatment to screen for disease etiology, and its an area where I believe we should be much more proactive.

Note: Before you start any medication on a woman of child-bearing age, you should order a urine pregnancy test just in case!


Specific Serotonin Reuptake Inhibitors (SSRIs). No lab monitoring required. However, you may want to order the appropriate labs if patients experience any of these three recently reported medical complications of SSRIs.

1. Bleeding. Most commonly manifesting as GI bleeding (Meijer W. Arch Internal Medicine 2004;164:2367-2370), the bleeding risk of SSRIs is not thought to be caused by platelet dysfunction, but is probably a direct effect of serotonergic stimulation. This is such a rare side effect that routine monitoring of the hematocrit is not indicated, but order a CBC if a patient presents with symptoms suggestive of blood loss.

2. Hyponatremia. Significant SSRI-induced hyponatremia (below 130) is rare, and it is most likely to occur in patients over 65 within 30 days of starting the SSRI (Consult Pharm 2000;15:160-77. http://www.ascp.com/publications/tcp/20 00/feb/cr-hypo.shtml). Again, it is too rare a side effect to merit routine Na monitoring, but consider ordering an electrolyte panel if an elderly patient recently started on SSRIs reports fatigue, dizziness, or cramping.

3. Osteoporosis. Two recent studies suggest that SSRI use decreases bone density in the elderly, placing them at greater risk for osteoporosis. The effect was not huge, but was large enough for the authors to recommend that elderly patients on SSRIs should have routine bone density screenings (see this months Research Updates for more details and references).

Effexor XR (venlafaxine XR). Patients should have their blood pressure checked periodically after starting or increasing the dose of Effexor XR. The risk of hypretension is dose-dependent, so monitoring should be more vigilant at doses of 225 mg or higher.

Cymbalta (duloxetine). Since Cymbalta causes elevation of alanine transaminase (ALT) in 1% of patients, check ALT at some point after the patient starts it.

Tricyclics. In patients with pre-existing cardiac disease, order an ECG both before starting a tricyclic and after reaching a therapeutic dose. Some authorities recommend a screening ECG in any patient above 40 or 50, regardless of cardiac history. There is some evidence endorsing the value of monitoring the serum level of nortriptyline, with a therapeutic window of 50-150 ng/mL correlating with the best antidepressant results.

MAOIs. Phenelzine (Nardil) has been reported to cause liver failure in case reports (Gomez-Gil et al., Annals Internal Medicine 1996;124:692-693), so some clinicians recommend monitoring LFTs after starting it.


The importance of the metabolic syndrome has been pounded into our brains as a result of the marketing wars between the different makers of atypical antipsychotics. To review: Metabolic syndrome it is defined as a combination of abdominal obesity, elevated fasting plasma glucose levels, elevated triglyceride levels, low HDL cholesterol levels, and hypertension.

Several antipsychotics appear to cause the metabolic syndrome, though there is disagreement about which ones lead to significant risk. TCPR has had the displeasure of reviewing this very complicated literature in past issues, and based on this we divide up the commonly used antipsychotics into two categories: metabolically dirty vs. metabolically clean. Another good source of recommendations came from the Mt. Sinai group (Marder et al., Am J Psych 2004; 161:1334-1349).

Metabolically dirty antipsychotics include: Zyprexa (olanzapine), clozapine, Risperdal (risperidone), Seroquel (quetiapine), chlorpromazine, and thioridazine.

Metabolically clean (or at least cleaner) antipsychotics: Abilify (aripiprazole), Geodon (ziprasidone), haloperidol, Trilafon (perphenazine).

Here are our monitoring recommendations for these two different categories:

Dirty antipsychotics. Weight. Determine BMI (body mass index, defined as weight divided by height) at baseline, once a month for the first three months, then every three months. Glucose. 1. Baseline fasting glucose (below 100 is normal, 100-125 is pre-diabetes, above 126 is diabetes). If your patient can’t manage to get to the lab before eating, order an HbA1c, which is a measure of long term glucose control. 2. Follow-up fasting glucose 4 months after starting med and then yearly, unless patients are gaining weight: if so, continue Q 4 mo. monitoring. Ask patients about polyuria or polydipsia to monitor for diabetes. Lipids. Baseline fasting lipid panel: total cholesterol, low-density lipoprotein (LDL) and HDL cholesterol, and triglyceride levels. Check lipids again 3 months later, then every 2 years; refer to PCP if LDL is higher than 130 mg/dl.

Clean antipsychotics. Weight. Baseline, 6 months, then yearly. Glucose. Baseline glucose (fasting not necessary); then yearly. Lipids. Baseline fasting lipid panel every 2 years.

ECG Monitoring

Mellaril (thioridazine), Serentil (mesoridazine, no longer available in U.S.), and Orap (pimozide) should not be prescribed for anyone with known heart disease. Geodon can be prescribed in patients with heart disease, but you should get a baseline ECG, and get follow-up ECG. In patients with no cardiac history, no screening ECG is required.


Patients on Risperdal and most first generation antipsychotics should be asked screening questions about symptoms of elevated prolactin. For women, ask about changes in menstruation or libido, and whether they have noticed a milk discharge from breasts. For men, ask about libido and sexual dysfunction. Order prolactin levels only if screening questions indicate possible hyperprolactinemia.

Mood Stabilizers

See the Chart for recommendations on monitoring mood stabilizers.