A tale of two patients


Journal of Practical Psychiatry and Behavioral Health, May 1999, 160-164

In this month's column, I focus on two frequent clinical dilemmas: 1) the patient who does not optimally benefit from a drug even at its maximum recommended dose, and 2) the patient who responds at a dose below the usually effective dose. I present two real life case vignettes illustrating these situations and discuss them in the context of the basic principles of clinical psychopharmacology reviewed in earlier columns.

Case 1
Ms. A is a 34-year-old female with schizoaffective disorder. Her psychotic and manic symptoms had been successfully treated with haloperidol, 4 mg/day, and carbamazepine, 1000 mg/day, for several years; however, she then developed a depressive episode, which the clinician decided to treat with the selective serotonin reuptake inhibitor (SSRI), sertraline. He followed the dosage recommendation in the package insert and started the patient at 50 mg/day. Because of a lack of response, he increased the dose at biweekly intervals until the maximum recommended daily dose of 200 mg/day was reached. After 4 weeks on this dose, the patient had experienced only modest improvement. What would you do at this point if you were the treating physician?

Case 2
Ms. G is a 67-year-old female who presented with her second major depressive episode, melancholic subtype. She was not psychotic or suicidal, and had no obvious psychosocial stressor. Her first depressive episode, which had occurred 10 years earlier, had responded well to nortriptyline, 100 mg/day. Her mother and a sister have also had recurrent depressive episodes. The patient's medical history was unremarkable, and she was on no other medications. The physician decided to treat her with the SSRI, sertraline. Given the patient's age, he initiated treatment at 25 mg/day with the intention of increasing the dose. At her next visit 1 week later, Ms. G was already reporting some improvement. For that reason, the physician elected not to increase the dose but to see her back in 1 week. At that visit, she was essentially in remission. This patient is therefore the mirror image of the patient in Case 1, remitting despite being on half of the usually effective dose of this antidepressant. Given this scenario, do you think this patient needs to stay on the antidepressant?

Discussion of Case 1

Many physicians would simply conclude that Ms. A had a condition that was not responsive to sertraline and would switch to a different antidepressant, perhaps even another SSRI. In fact, switching from one SSRI to another in the case of treatment failure is a common strategy in clinical practice. However, there are no rigorous clinical trial data to support this practice, which would seem to make sense only for pharmacokinetic reasons. So what would you do at this point?

In this case, the patient's physician ordered therapeutic drug monitoring (TDM) of both sertraline and carbamazepine. The results are shown in Table 1.

Table 1 - Plasma levels of sertraline and
carbamazepine in Case 1
  Drug Daily dose Time after dose Level 
carbamazepine 1000 mg 12 hours 8.0 mcg/ml
sertraline 200 mg 12 hours 17.0 ng/ml

The rationale for using TDM with carbamazepine is probably clear to most readers. Carbamazepine is generally acknowledged to have a therapeutic range and the level in Table 1 is within that range. However, some readers may wonder why one would do TDM with sertraline and what the result means. This is due to the common misperception that SSRIs do not have a therapeutic plasma level range.

I have addressed this issue from a theoretical standpoint in previous columns.1-6 Here I will use the two cases presented above to illustrate the clinical applications of the principles enumerated in those earlier columns.

Figure 1. Equation 1: Relationship of Pharmacodynamics, Pharmacokinetics, and Biological Variance in Determining Overall Result of Drug Treatment.
Clinical response = Affinity for the site of action X Drug concentration at the site of action X Underlying biology of the patient
  • absorption
  • distribution
  • metabolism
  • elimination
  • diagnosis
  • genetics
  • gender
  • age
  • organ function
  • environment

Regular readers of this column know that the effect of any drug is determined by the three variables in equation 1 (Figure 1). The first variable in the equation determines what the drug is capable of doing. The second variable determines what the drug will do. The third variable determines whether a given patient will be more or less sensitive to the drug's effect in comparison to the usual patient in the drug's registration trials.

Regular readers also know that the second variable in equation 1 is determined by EQUATION 2:

Drug level =  dosing rate

Thus, any drug that has a usually effective dose must have a usually effective concentration range since the dose determines the range. Using equation 2, the clinician can determine the patient's clearance by knowing the drug level and the dosing rate. Thus, the most basic goal of TDM is to determine the patient's ability to clear the drug in question and then to determine whether the patient has achieved a level consistent with the range usually achieved on the usually effective dose. If the patient is below that level, he or she has not been given an adequate trial of that medication.

Viewed from the perspective of these two equations, clinical trials are essentially population pharmacokinetic studies. They are used to determine the usually effective dose, which in turn determines the concentration range of the drug that is safe, well tolerated, and effective in the usual patient enrolled in the clinical trial.

However, patients in clinical practice can be quite different from those enrolled in a clinical trial. For example, it is doubtful that any patient has been on carbamazepine during any efficacy trial of sertraline (or any other SSRI for that matter). Yet, the patient in Case 1 was. That is important because carbamazepine induces the cytochrome P450 enzyme CYP 3A, which is partly responsible for the oxidative metabolism of sertraline. As a result, this patient was likely to have a more rapid clearance of sertraline than would usually be the case. In fact, TDM was done to assess that possibility and to determine whether the patient had achieved the concentration usually associated with optimal antidepressant response.

The mean 12-hour post-dose plasma level of sertraline is approximately 30 ng/ml on 50 mg/day and goes up proportionally (i.e., linear pharmacokinetics) as the dose goes up.7 However, despite receiving 200 mg/day (i.e., the maximum recommended daily dose), Ms. A, the patient in Case 1, developed a plasma sertraline level below the average typically achieved on 50 mg/day (i.e., the usually
effective, minimum dose).8, 9

Based on equation 2, there are two reasons this patient could have developed such a low level: 1) the dosing rate was too low or 2) clearance was too high. It is important to remember that the dosing rate is what the patient is actually taking rather than what the doctor is prescribing. The clinician must consider the possibility that this patient was not being compliant with the prescribed dose. However, the TDM with carbamazepine suggests that she was compliant, first, because her plasma carbamazepine levels were in the range that would be expected for the dose prescribed (i.e., 1000 mg/day), and second, because this carbamazepine plasma level was comparable to levels measured in this patient on the same dose on a yearly basis for the past 3 years. Thus, the TDM results with carbamazepine indicate that this patient was taking her car-bamazepine as prescribed. It seems unlikely (although not impossible) that this patient would be compliant with her carbamazepine but not with her sertraline. Moreover, on direct questioning, the patient indicated that she had been taking her sertraline as prescribed.

For all these reasons, the TDM results suggest that this patient's clearance of sertraline is over four times faster than the usual patient in the clinical trials of sertraline - which is consistent with her being on the CYP 3A inducer, carbamazepine. This suggests that the patient has not received an adequate trial of sertraline despite being on the usual maximum recommended dose. The TDM results in this case indicate that this patient would need to be on a dose in excess of 300 mg/day of sertraline to adequately test its antidepressant effectiveness. Thus, TDM could be used to justify the use of such a dose in this patient.

Alternatively, these TDM results could indicate to the physician that the patient should be switched to another SSRI which is not principally dependent on a CYP enzyme that is inducible by carbamazepine. The SSRIs paroxetine or fluoxetine would fit this requirement. Such a switch could be a more cost-effective strategy than using > 300 mg/day of sertraline.

Paroxetine has several advantages over fluoxetine in this case. First, as a result of its active metabolite, fluoxetine has a long half-life.10 As discussed in earlier columns, the long half-life means that the usual patient will take from several weeks up to even a couple of months to reach steady state as well as to wash out the drug. Since this patient has a schizoaffective disorder, it is possible that any SSRI could induce rapid cycling. The long half-life of fluoxetine raises the possibility of a significant delay in the onset of such rapid cycling. The more delayed the onset of an effect, the more difficult it is to determine a cause and effect relationship. Moreover, the effects of fluoxetine will persist for a protracted period of time after its discontinuation.11-13 Another advantage of paroxetine is that it inhibits fewer CYP enzymes than fluoxetine. For these reasons, fluoxetine should be reserved for situations in which the advantages of its long half-life outweigh these two disadvantages.

As discussed in earlier columns, paroxetine at low levels is principally metabolized by CYP 2D6, which it also inhibits.14,15 Moreover, CYP 2D6 is not inducible by carbamazepine. Therefore, on paroxetine, 20 mg/day, this patient might achieve paroxetine plasma levels near those that are typically seen in patients on this dose and thus respond. The principal reservation with this recommendation is that, at higher levels, paroxetine is likely metabolized by CYP 3A so that concomitant treatment with carbamazepine might still interfere with the patient reaching effective levels of paroxetine. This can be assessed by doing TDM with paroxetine just as was done with sertraline. In the event that this proves to be the case and the patient does not optimally respond, the physician could then try fluoxetine.

If the clinician switched this patient from sertraline to paroxetine without understanding this rationale and saw a response, he or she might erroneously conclude that paroxetine and sertraline work through different mechanisms in different patients. There is no compelling evidence to support that position, although there are some pharmacodynamic differences between the drugs (i.e., effects on CYP 2D6, nitric oxide, and the muscarinic acetylcholine receptor). The available data suggest that all the SSRIs work through the same mechanism of action, inhibition of the serotonin uptake pump. However, the usually effective antidepressant dose of paroxetine could work in a patient who did not benefit from the maximum recommended dose of sertraline for a pharmacokinetic rather than a pharmacodynamic reason. This would be a case where the potential therapeutic advantage of paroxetine could outweigh the disadvantages
posed by its substantial inhibition of CYP 2D6.

This case also demonstrates that sertraline can be the victim of a CYP-enzyme-mediated drug-drug interaction even though it is less likely than most other SSRIs to cause such an interaction.12,13

Discussion of Case 2

The physician in this case wondered whether his patient, Ms. G, had had a drug specific response to sertraline or a "placebo" response (i.e., a response to good clinical management). This is a reasonable question. As discussed in earlier columns,1,16 the results from clinical trials suggest that only one out of two responders in the drug treatment arm are responding specifically to the drug rather than to the good clinical management that is inherent in such a study.17

The clinician suspected that the patient had had a "placebo" response for two reasons. First, she got better on half the usually effective dose (i.e., 25 rather than 50 mg/day). Second, she was in full remission after only 2 weeks of treatment. Many clinicians might not want to look too closely at such success, but this clinician wondered whether it would be better to stop sertraline if the
patient did not need it.

There were also several clinical reasons to suspect that the response was drug specific. First, Ms. G had had a previous episode. Second, she had a positive family history for major depression. Both of these facts decrease the likelihood of a transient depressive episode or a "placebo" response.18

However, the clinician was still bothered by the fact that the patient had responded to such a low dose of sertraline. For that reason, TDM was done, the results of which are shown in Table 2.

Table 2 - Plasma levels of sertraline in Case 2
  Drug Daily dose Time after dose Level 
sertraline 25 mg 12 hours 28 ng/ml

This patient needed only half the usually effective antidepressant dose of sertraline to achieve a sertraline plasma level near the average level achieved in patients who received 50 mg/day in the fixed dose efficacy trials of sertraline. In other words, this patient cleared sertraline half as fast as the usual patient in such trials.

Case 2 represents the flip side of Case 1, with this patient having a clearance that was 16 times slower than the patient in Case 1. The reason for the slower clearance in the patient in Case 2 was not established. Possible reasons include age-related decline in cardiac, liver, and/or renal function (i.e., the third variable in equation 1). Regardless of the reason, the TDM results in Case 2 are consistent with the contention that this patient did respond specifically to sertraline-the fact that she was on a dose lower than what is usually effective was misleading. Based on these findings, the clinician decided to maintain the patient on sertraline for a full course of acute and maintenance treatment.

The reader might wonder whether the difference between 17 and 28 ng/ml is physiologically or clinically meaningful. The difference is small (11 ng/ml to be precise). However, like all SSRIs, sertraline was developed to be a potent and selective serotonin reuptake inhibitor. In fact, it is one of the most potent and selective.19,20 Potency is variable 1 in equation 1. The more potent the drug, the smaller the amount that is needed to produce an effect. Amount is variable 2 in equation 1. That is the essence of the relationship between pharmacodynamics and pharmacokinetics. That is also the reason why clinical trials are population pharmacokinetic studies. Clinical trials determine the dose that is needed in the usual patient in the clinical trial to achieve the concentration needed to engage the right mechanism of action to the right degree to achieve the desired physiological/clinical effect. Clinicians must take into account differences in their patients that can shift the dose-response curve. That is variable 3 in equation 1. This column presented two patients who were at either end of the spectrum in terms of clearance as it affects the dose-response curve. While 11 ng/ml may seem like a small amount, this difference occurred at the steep ascending portion of the sertraline concentration-response curve in terms of serotonin uptake inhibition, as discussed in my last column.1

Was the laboratory helpful in these two cases?

Yes and no. The laboratory was helpful in providing the TDM results. The laboratory was not helpful in terms of providing the necessary information to interpret the results.

In Case 2, the laboratory provided only the assay result without any information about what it might mean. In Case 1, the laboratory provided the following information as well as the result:

The average peak plasma levels following single doses of sertraline are as follows:
Dose: 50   100 200 300 400
Level: 9.5 16 56 78 88
The average peaks plasma levels following daily doses of sertraline are as follows:
Dose: 50 100 150 200  
Level: 32 54 144 190  
Peak levels occur 4.5 to 8.4 hours after a dose of sertraline (i.e., Time maximum or Tmax). The desmethyl metabolite of sertraline is inactive and is not measured in this assay.

Although this is a reasonable amount of information, little of it is relevant and some of it may even be unintentionally confusing. It is useful to know that the level does not include a value for desmethylsertraline. This metabolite is substantially less active than the parent drug in terms of the ability to inhibit serotonin uptake and thus probably does not contribute to the antidepressant efficacy of the parent drug in contrast to the desmethyl metabolite of fluoxetine, norfluoxetine.19,20 The rest of the information tells us quite a bit about the linearity of sertraline pharmacokinetics following single and multiple doses, but does not tell us about the relationship, if any, between these levels and antidepressant efficacy. Some might say there is no relationship but the basic principles of pharmacology, as expressed in equation 1, say that there must be a relationship.

Another problem is that the laboratory provided information based on peak levels (i.e., 4.5 to 8.4 hours after a dose of sertraline) rather than levels at 12 hours after the dose, which was the time this patient's plasma level was sampled. Twelve hours or longer is the interval typically chosen for TDM because these levels are postabsorption and postdistribution and are thus more reproducible than peak levels. Such levels also correlate better with the concentration in deep compartments such as the brain where the drug is presumably working. Plasma drug levels are a surrogate marker or measure of the levels of the drug in the brain.1

The information provided on the laboratory report was initially more confusing than illuminating to the patient's physician. It seemed to suggest that the patient's level was 10 times lower than expected (i.e., 17 versus 190 ng/ml). Peak levels, of course, will be higher than postabsorption, postdistribution levels. Had the clinician not realized the difference, he might have come to one of several erroneous conclusions: 1) that the levels are unreliable, 2) that the patient had to be noncompliant to have such low levels, or 3) that the clearance of sertraline in this patient was ten times faster than in the usual patient. The third conclusion could have led to prescribing a much higher dose than was needed to achieve potentially therapeutic levels.

The point is that reference laboratories may be able to provide reliable and accurate measurements of psychiatric drug levels but may not be helpful when it comes to the proper interpretation of these results. Clinicians cannot simply trust the laboratory on this issue. Instead, they must know enough to be able to properly interpret the results and their implications relative to the patient's care.


These cases are anecdotal and do not prove that TDM with sertraline (or any other SSRI) helps with patient care. Such proof would require prospective, controlled, blinded, and adequately powered research studies. However, such research is quite expensive and is unlikely ever to be conducted for several reasons. TDM research is somewhat of an orphan area that falls between the cracks of the interests of the two major sources of funding for psychiatric drug research: the pharmaceutical industry and the federal government. Industry sponsored research is quite reasonably focused on supporting either the drug's registration or marketing. TDM research is generally perceived as not fitting either objective and hence is not likely to be funded by the manufacturer. Federal funding must be divided between basic and applied research and is too limited to support more than a small number of clinical studies.

The dearth of clinically useful applied research is not unique to TDM with newer antidepressants. Physicians often do not have studies that match the specifics of their case. For example, the patients enrolled in antidepressant trials are a quite rarefied subset of the patients treated by psychiatrists. The practitioner therefore must frequently extrapolate from the available research data to his or her patient based on pharmacologic principles and clinical judgment.

The goal of this column was to use real life cases to illustrate one way in which TDM can be used with newer drugs to potentially improve the care of patients. As with many of my columns, this one was written in response to my contact with physicians across the country. Many do not appear to be aware that one can measure the drug levels of commonly used psychiatric medications-yet they seem quite receptive to the approach taken in these two cases when they learn about it.

I leave it to the readers to determine for themselves whether the course of action taken in these cases makes clinical sense to them. While sertraline was the drug used here, it could have been a number of other psychiatric medications for which the usually effective dose has been rigorously established. Moreover, the use of TDM is not restricted to improving efficacy. For example, the column, "A Message from Titanic," illustrated how TDM can be used to detect the persistent risk of a drug-drug interaction after a long-lived drug had been discontinued.


  1. Preskorn SH, Finding the signal through the noise: The use of surrogate markers. J Pract Psychiatry Behav Health 1999;5:104-109
  2. Preskorn SH, Have you phenotyped your patient lately? J Pract Psychiatry Behav Health 1996;2(2):115-7
  3. Preskorn SH, Why did Terry fall off the dose-response curve? J Pract Psychiatry Behav Health 1998;4:363-7
  4. Preskorn SH, To monitor or not to monitor? J Pract Psychiatry Behav Health 1996;2:172-5
  5. Preskorn SH, To monitor or not to monitor, II: The glass is more than half full. J Pract Psychiatry Behav Health 1996;2:307-10
  6. Preskorn SH, If lack of concentration didn’t cause the fall, what did? J Pract Psychiatry Behav Health 1996;2:364-7
  7. Ronfeld R, Tremaine L, Wilner K, Pharmacokinetics of sertraline and its n-demethyl metabolite in elderly and young male and female volunteers. Clin Pharmacokinet 1997;32:22-30
  8. Fabre LF, Abuzzahab FS, Amin M, et al, Sertraline safety and efficacy in major depression: a double-blind fixed-dose comparison with placebo. Biol Psychiatry. 1995;38:592-602
  9. Preskorn SH, Lane RM, Sertraline 50 mg daily: The optimal dose in the treatment of depression. Intl Clin Psychopharmacol. 1995;10:129-141
  10. Preskorn SH, Shad MU, Alderman J, Lane R, Fluoxetine: Age and dose dependent pharmacokinetics and CYP 2C19 inhibition. Am Soc Clin Pharmacol Ther. 1998;63:166. Abstract
  11. Preskorn SH, A message from Titanic. J Prac Psych Behav Hlth. 1998;4:236-242
  12. Preskorn SH, Clinically relevant pharmacology of selective serotonin reuptake inhibitors. An overview with emphasis on pharmacokinetics and effects on oxidative drug metabolism. Clin Pharmacokinet 1997; 32:1-21
  13. Richelson E, Pharmacokinetic drug interactions of new-age antidepressants: A review for the practitioner on effects on the metabolism of other drugs. Mayo Clin Proc 1997;72:835-47
  14. Harvey AT, Preskorn SH, Cytochrome P450 enzymes: interpretation of their interactions with selective serotonin reuptake inhibitors. Part I. J Clin Psychopharmacol. 1996;16:273-285
  15. Harvey AT, Preskorn SH, Cytochrome P450 enzymes: interpretation of their interactions with selective serotonin reuptake inhibitors. Part II. J Clin Psychopharmacol. 1996;16:345-355
  16. Preskorn SH, A dangerous idea. J Pract Psychiatry Behav Health 1996;2:231-4
  17. Janicak PG, Davis JM, Preskorn SH, Ayd FJ Jr, Principles and Practice of Psychopharmacotherapy. 2nd ed. Baltimore, Md: Lippincott, Williams & Wilkins; 1997
  18. Brown WA, Dornseif BE, Wernicke JF, Placebo response in depression: A search for predictors. Psychiatry Res. 1988;26: 259-264
  19. Bolden-Watson C, Richelson E, Blockade by newly-developed antidepressants of biogenic amine uptake into rat brain synaptosomes. Life Sci. 1993;52:1023-1029
  20. Hyttel J, Comparative pharmacology of selective serotonin reuptake inhibitors (SSRIs). Nord J Psychiatry. 1993;47 (suppl 30):5-12

Copyright and Disclaimer

©2010, Sheldon H. Preskorn, M.D.
site design by CyberKansas Technologies
Questions or Comments about the site?