Summary

Clinical Pharmacokinetics 9: 555-563 (1984)

0312-5963/84/ II 00-0555/$04.50/0 © ADIS Press Limited All rights reserved.

Two Prospective Dosing Methods for Nortriptyline

Paul J. Perry, Jerry L. Browne, Bruce Alexander, Ming T. Tsuang, Arnold D. Sherman and Frederick J. Dunner Division of Clinical Pharmacy, College of Pharmacy, Department of Psychiatry, College of Medicine, The University of Iowa, Iowa City, and Department of Psychiatry, Iowa City Veterans Administration Hospital, Iowa City

This study compared two prospective pharmacokinetic dosing methods to predict steady­state concentrations of nortriptyline. One method required multiple determinations of the nortriptyline plasma concentration to estimate the drug's steady-state concentration. The second method required a single nortriptyline concentration drawn at a fixed time. pre! erably 36 hours, following a nortriptyline test dose. The 36-hour nortriptyline plasma con­centrations (NTP 36h) were substituted into the straight-line equation of q~ = 17.2 + 3.74 (NTP 36h). where q~ is the average steady-state concentration for a 100 mglday dose of nortriptyline.

No differences were noted between the observed steady-state nortriptyline concentration of 121 ± 19 nglml. the 36-hour single-point prediction mean concentration of 121 ± 21 nglml. or the multiple-point prediction mean concentration of 122 ± 19 nglml. Because of the similar findings between the two methods. the clinical advantages and disadvantages of each kinetic approach are discussed to put these prospective dosing protocols into their proper perspective.

It is generally recommended that to maximise the probability of a positive therapeutic response to the tricyclic antidepressant nortriptyline, a dose producing a steady-state nortriptyline plasma con­centration of 50 to 150 ng/ml should be admini­stered to the patient (Risch et aI., 1981). Numerous investigators have described differing mathemati­cal methods to prospectively predict nortriptyline steady-state plasma concentrations following the administration of an initial test dose. Alexander­son (1972) demonstrated that it was possible to predict individual patient steady-state nortripty­line concentrations by calculating the nortriptyline clearance from 5 or more plasma concentrations. A single nortriptyline plasma concentration drawn at either 24, 48 or 72 hours following a 100mg nor-

triptyline test dose has been shown to correlate sig­nificantly with steady-state concentrations follow­ing 100 mg/day nortriptyline maintenance doses (Braithwaite et al., 1978).

Cooper and Simpson (1978) have also demon­strated a significant correlation between a single 24-hour nortriptyline concentration and steady-state nortriptyline concentrations. Similar correlations of steady-state tricyclic antidepressant concentra­tions with single plasma concentrations following a test dose have also been established for imipra­mine and desipramine by Potter et a1. (1980), and for amitriptyline by Madakasira et a1. (1982).

The validity of utilising as few as 2 plasma es­timations of nortriptyline following a test dose to prospectively predict steady-state nortriptyline

Two Prospective Dosing Methods for Nortriptyline

concentrations has recently been described (Browne et aI., 1983). A correlation coefficient of 0.94 (p < 0.01) was observed between the predicted and observed nortriptyline concentration. Thus, these authors postulated that the multiple-point predic­tive method could produce a greater degree of ac­curacy and sensitivity in predicting steady-state plasma concentrations than the single-point pre­dictive method.

It was the aim of this investigation to determine whether multiple-point methods of predicting steady-state concentrations produced any greater degree of predictive accuracy and sensitivity than the single-point methods.

Materials and Methods

The study involved 27 patients (11 males, 16 females), ranging in age from 25 to 76 years, ad­mitted to the psychiatry service of two university­affiliated hospitals. All subjects met the DSM III criteria for the diagnosis of major depressive dis­order (American Psychiatric Association, 1980). Exclusion criteria included patients with signifi­cant renal, hepatic and/or cardiovascular disease as well as mentally retarded or court-committed patients. Subjects were required to be drug stable but not drug free for inclusion in the study to pre­vent any potential metabolic interactions that might alter the nortriptyline concentrations. Each indi­vidual gave informed consent for participation in the study.

The study required the administration of a 100mg test dose of nortriptyline at approximately 8pm to each subject. Post-test dose blood samples (10ml) were drawn by intravenous puncture at ap­proximately 0, 12, 24 and 36 hours. A final steady­state nortriptyline plasma concentration was drawn at a time greater than 5 times the estimated plasma nortriptyline half-life following the administration of a fixed daily dose. After analysis of the plasma samples, patients were started on nortriptyline doses that were predicted to achieve therapeutic plasma concentrations between 50 and 150 ng/ml.

The blood samples were collected in polyure­thane 10mi syringes via antecubital puncture and

556

then transferred to heparinised Vacutainer tubes, with the stoppers removed, and sealed with Para­film®. At no time did the sample come in contact with the stoppers of the Vacutainer tubes. The plasma was separated by centrifugation at 1500 rpm for 20 minutes and stored at SOc. For each sample a I ml aliquot of plasma was analysed in triplicate (Cooper et aI., 1976). The instrument used for ana­lysis consisted of a Shimadzu gas chromatograph GC-7 Aj7 AG Series equipped with a flame ther­moionic detector (FTD7) ionisation nitrogen phos­phorus detector. This method has a coefficient of variation of 5% at a level of 5 ng/ml.

Multiple-Point Prediction Method

In order to predict the nortriptyline dose ne­cessary to have the patient's plasma concentrations fall within the therapeutic range, a standard phar­macokinetic method was employed (Gibaldi and Perrier, 1975). This technique required that the nortriptyline half-life and in tum the nortriptyline terminal phase elimination constant be established for each individual patient. The accumulation fac­tor (R) was calculated by:

R= ) (css)tTD

(Eq. 1),

where K is the elimination rate constant, and T is the dosing interval. R in actuality is simply the ratio of the steady-state nortriptyline concentration at time t for the test dose if given as a steady-state maintenance dose [(CSS)tTD] to the nortriptyline concentration at time t following the single test dose [CtTD]. Rearranging equation 1,

(Eq. 2).

For predicting other steady-state nortriptyline plasma concentrations, the following equation was utilised:

(CSS) = (DM) (CSS)ITD

I TD (Eq. 3),

where (CSS)t is the steady-state plasma nortriptyline concentration at time t, DM is the maintenance dose, TD is the test dose, and (CSS)tTD is the steady-

Two Prospective Dosing Methods lor Nortriptyline 557

Table I. Observed and normalised steady-state plasma nortriptyline concentrations

Patient Nortriptyline Observed steady-state Adjusted steady-state dose nortriptyline nortriptyline (mg/day) cone. (ng/ml) cone. (ng/ml)a

50 107 214

2 50 113 216

3 100 116 116

4 100 139 139

5 125 101 81

6 100 147 147

7 100 146 146

8 100 133 133

9 50 91 182

10 100 141 141

11 100 141 141

12 50 62 124

a The steady-state nortriptyline concentrations were retrospectively normalised to a 100 mg/day maintenance dose.

state plasma nortriptyline concentration produced by the test dose at time t. 15 patients were dosed according to this method.

For those patients who had previously been re­ceiving nortriptyline or amitriptyline prior to entry into the study, equation 4 was applied to equations 1 and 2 to take into account residual drug in the body prior to the test dose:

(Eq.4),

where C1TD is the concentration of drug in the plasma resulting specifically from the test dose of drug administered, and is equal to the total amount of drug in the plasma minus the amount of drug present from doses administered prior to the test dose. C1 is the plasma concentration at the various sampling intervals, with t equal to either 12,24 or 36 hours. Co is the plasma concentration just prior to administration of the test dose, with t being the time between collection of C1 and Co. This pro­cedure subtracts the amount of drug present in each sample which is not a result of the test dose.

Single-Point Prediction Method

A mathematical relationship between drug con­centration in serum or plasma at steady-state and a single drug concentration at some time after a test dose has previously been described (Slattery et aI., 1980). The derivation suggests that a reciprocal relationship exists between the maintenance dose and test dose drug concentration, whereas a direct proportional relationship exists between the mean steady-state drug concentration and the test dose drug concentration. In order to examine the latter mathematical relationship, the steady-state nor­triptyline plasma concentrations at II hours after administration were retrospectively adjusted to a 100 mg/day maintenance dose for the first 12 patients entered in the study. The normalised nor­triptyline concentrations were subjected to regres­sion analysis against test dose concentrations at 12, 18,24,30,36,42 and 48 hours. These nortriptyline concentrations were derived by regression analysis of the original data points. The linear regression

Two Prospective Dosing Methods for Nortriptyline

equation and correlation coefficient were deter­mined for each of these comparisons. A total of 12 patients were subjected to this analysis.

Bivariate linear regression analysis of the ex­ponential nortriptyline elimination phase curves was utilised to calculate the 12-, 18-, 24-, 30-, 36-, 42-, and 48-hour time intervals for each of the initial 12 patients treated with nortriptyline. From these data, each individual time interval for the 12 patients was subjected to bivariate linear regres­sion analysis against the observed steady-state plasma nortriptyline concentration corrected for a 100 mg/day maintenance dose.

Subsequently, 15 additional patients were pro­spectively dosed to compare the single- and mul­tiple-point prospective dosing equations. The ob­served plasma nortriptyline concentration was compared with the predicted plasma nortriptyline concentrations calculated from the single-point and multiple-point prediction methods utilising mul­tiple analysis of variance (MANOY A). Significance was defined as an a of 0.05 for all tests. In cases where the F ratio was found to be significant, the t-test for multiple comparisons was used to deter­mine which specific means differed significantly (McNemar, 1962).

558

Results

Following the administration of the test dose, the first 12 patients were administered mainten­ance doses which, according to the multiple-point prospective predictive method, placed the patients within the nortriptyline 'therapeutic window' of 50 to 150 ng/mt. The observed nortriptyline concen­trations for these patients ranged from 62 to 147 ng/ml (mean 120 ± 26 ng/ml) with the doses rang­ing from 50 to 125 mg/day. The steady-state nor­triptyline concentrations were then adjusted to a 100 mg/day dose. These data are summarised in table I. The regression equations derived for the steady-state nortriptyline concentrations versus the single-dose nortriptyline concentrations at the 6-hour time intervals are shown in table II.

The 12-,24- and 36-hour steady-state prediction equations shown in table II were prospectively tested. Since the equation gives the steady-state concentration for a 100 mg/day dose, the mainten­ance dose was adjusted proportionately such that the prediction was within the therapeutic range. The single-point method predictions were compared with the multiple-point method predictions.

The multiple analysis of variance comparing the

Table II. Regression equations for normalised steady-state nortriptyline plasma concentrations versus single-dose nortriptyline con­centrations calculated at 6-hour intervals

Time after Linear regression test dose equation· (h)

12 C~~ = 20.2 + 2.07x

18 C~~ = 14.5 + 2.50x

24 q~ = 15.0 + 2.87x

30 q~ = 15.5 + 3.30x

36 q.~ = 17.2 + 3.74x

42 q.~ = 21.0 + 4.20x

48 q~ = 25.8 + 4.64x

Correlation coefficient

0.83

0.87

0.89

0.92

0.94

0.95

0.95

p value

< 0.001

< 0.001

< 0.001

< 0.001

< 0.001

< 0.001

< 0.001

a C~~ is the steady-state nortriptyline concentration for a 100 mg/day maintenance dose; x is the plasma nortriptyline concentration at either 12, 18, 24, 30, 36, 42, or 48 hours following a 100mg test dose for the corresponding equations.

Two Prospective Dosing Methods for Nortriptyline

12-hour single-point equation predictions with the multiple-point method predictions and the ob­served nortriptyline steady-state concentrations for 14 patients demonstrated a significant difference between the measurements (F = 6.86; p < 0.01). [Only 14 patients were utilised in this analysis be­cause the fifteenth patient's 12-hour level was not available.] The predicted steady-state nortriptyline plasma concentrations for the multiple-point method were similar to the observed nortriptyline concentrations. The mean observed plasma nor­triptyline concentration was 123 ± 19 ng/ml, which was identical to the mean predicted plasma nor­triptyline concentration for the multiple-point method of 123 ± 19 ng/ml. However, when the predicted steady-state nortriptyline mean plasma concentrations for the 12-hour single point method were compared with the observed nortriptyline plasma concentrations, a significant difference was observed (p < 0.01). The mean predicted plasma nortriptyline steady-state concentration for the 12-hour single-point method was 144 ± 30 ng/ml.

The multiple analysis of variance comparing the 24-hour single-point equation predictions with the multiple-point method predictions and the ob­served nortriptyline concentrations also demon­strated a significant difference between the meas­urements (F = 6.20; p < Om). The multiple-point method predictions were similar to the observed predictions for the 15 patients tested. The mean observed plasma nortriptyline concentration of 121 ± 19 ng/ml was nearly identical to the mean pre­dicted plasma nortriptyline concentration of 122 ± 19 ng/ml for the multiple-point method. How­ever, when the predicted steady-state nortriptyline plasma concentrations for the 24-hour single-point method were compared with the observed nortrip­tyline plasma concentrations, a significant differ­ence was observed (p < 0.01). The mean predicted plasma nortriptyline concentration for the 24-hour single-point method of 130 ± 18 ng/ml differed significantly from the observed plasma nortripty­line values.

Finally, the multiple analysis of variance com­paring the 36-hour single-point equation predic­tions with the multiple-point method predictions

559

and the observed nortriptyline concentrations for 15 patients demonstrated no difference between the predictions (F = 0.04; p > 0.05). The mean pre­dicted plasma nortriptyline concentration of 121 ± 20 ng/ml for the 36-hour single-point method was nearly identical to the 3-point method mean prediction of 122 ± 19 ng/ml and the observed plasma nortriptyline mean of 121 ± 19 ng/ml. These data are presented in table III.

Discussion

To derive the single-point equations for pre­dicting the steady-state nortriptyline concentra­tions, the original steady-state nortriptyline plasma concentrations had to represent a fixed mainten­ance dose of 100 mg/day of nortriptyline. As can be seen in table I, only half of the patients were receiving 100 mg/day maintenance doses. Thus, the steady-state concentrations had to be adjusted up­ward or downward to a 100 mg/day dose. We felt justified in this methodological approach since pharmacokinetically there is no reason to expect the adjusted steady-state concentrations to be sig­nificantly different even if all 12 of the patients were taking 100 mg/day maintenance doses since, as equation 5 (below) shows, the steady-state plasma concentration (0S

) changes directly with the change in the maintenance dose (DM):

ess = (F) (DM) (eL) (r)

(Eq. 5),

where F is the fraction of the dose absorbed, eL the clearance and T the dosing interval; it is as­sumed that these 3 terms remain constant with a change in dose. Kragh-Sorensen and Larsen (1980) have clinicaUy demonstrated the validity of this as­sumption of a proportional relationship between nortriptyline dose and steady-state plasma nortrip­tyline concentrations within a plasma level range of 20 to 296 ng/ml. Also, blindly dosing a patient with a 100 mg/day dose could have resulted in ex­cessive adverse effects in the 5 patients receiving the 50 mg/day doses, a subtherapeutic nortripty­line concentration in 1 patient, and quite probably, a longer hospitalisation in these 6 patients.

Two Prospective Dosing Methods for Nortriptyline 560

Table III. Observed versus predicted nortriptyline plasma concentrations for the prospective single- and multiple-point prediction methods

Patient 36-hour plasma Maintenance Observed Predicted steady-state conc. (ng/mll nortriptyline conc. dose steady-state conc. (ng/mll (mg/dayl (ng/mll

single-point multiple-point method" method

13 42 50 89 87 88

14 25 100 104 111 104

15 35b 100 127 148 150

16 38 75 124 119 121

17 39 75 125 122 117

18 17 125 109 101 99

19 20b 125 121 115 127

20 35b 100 155 148 149

21 24 100 92 107 102

22 42 75 125 131 133

23 54 50 108 110 112

24 39 75 126 122 124

25 23 150 147 155 145

26 21 100 124 96 114

27 34b 100 145 144 141

a Steady-state value determined by substituting the 36-hour plasma nortriptyline 100mg test dose concentration (x) into the equation y = 17.2 + 3.74x, where y is the steady-state plasma nortriptyline concentration following a 100mg maintenance dose and proportioned according to the administered daily dose.

b These 4 plasma samples were drawn at 37 hours, necessitating the 36-hour plasma concentrations to be estimated by regression analysis.

The prediction of steady-state drug concentra­tions from a single test dose and multiple concen­tration measurements is based on calculating the accumulation factor (R) as shown in equation 1. This equation estimates the accumulation ratio for a drug given as an intravenous bolus and obeying l-compartment distributional kinetics for both minimum and maximum drug concentrations. However, nortriptyline in this case is administered orally, and at least in some patients, demonstrates 2-compartment distributional kinetics. Calcula­tions of R are acceptable for 2-compartment dis­tributional models for calculating only minimum

drug concentrations where steady-state concentra­tions in the elimination phase are being sought. Utilising this method, Browne et al. (1983) have demonstrated a highly significant correlation be­tween predicted and observed nortriptyline con­centrations. Thus, nortriptyline prospective pre­dictive pharmacokinetic problems can be treated as a I-compartment open linear model.

The results of the study indicate that the mul­tiple-point pharmacokinetic model produces nor­triptyline predictions that are similar to those pro­duced by a single-point pharmacokinetic model that requires a 36-hour nortriptyline plasma sample fol-

Two Prospective Dosing Methods for Nortriptyline

lowing a lOOmg test dose. Since both methods are equivalent in their accuracy and sensitivity the ad­vantages and disadvantages of the two methods de­serve discussion.

Multiple-Point Method

The advantages of the mUltiple-point method include the knowledge of each patient's elimina­tion half-life, and the flexibility of the sampling times and the test dose. A patient can have the nortriptyline concentrations drawn at convenient times, provided there is (preferably) a 24-hour timespan between the 2 samples. Dawling et al. (1980) described a single-dose nortriptyline phar­macokinetic dosing schedule for elderly patients (> 69 years of age) utilising a 50mg test dose. Their study intimates that elderly patients ought to be given smaller nortriptyline test doses. Of the 3 patients (16, 20, 24) who were 65 years of age or greater in the present study, none encountered any notable adverse effects from the 100mg test dose nor any extraordinary variance in the predicted nortriptyline concentrations from the observed concentrations. However, the test dose is flexible with the multiple-point method, whereas this is not the case with the single-point prediction method. We have encountered patients who, because of their past antidepressant drug history, could not reason­ably be expected to tolerate a Ioomg test dose. Thus, flexibility is the greatest benefit of utilising the multiple-point method.

The disadvantages of the multiple-point method are: firstly, a hand-held calculator or minicomputer with linear regression capability is required to quickly solve the algebraic equations; secondly, the extra expense of an additional plasma nortriptyline measurement; and finally, it may be more difficult with some outpatients to obtain 2 or more blood samples than with mpatients. The calculator/com­puter disadvantage should decrease as their use in­creases in the future. Since tricyclic antidepressant analysis generally costs approximately US $40 per sample, this may present more of a problem to some patients.

The inpatient versus outpatient disadvantage is,

561

in our opinion, a patient-specific problem rather than a general problem.

Single-Point Method

The advantages of the single-point method in­clude the cost savings and mathematical simplic­ity. Nortriptyline dosing nomograms can be con­structed based on the single-point equations shown in table II. A nomogram based on the 36-hour single-point equation is presented in figure 1. The individual linear equations for the daily mainten­ance doses shown in figure 1 were calculated by multiplying the right side of the 36-hour equation by the constant (daily maintenance dose/IOOmg/ day). Thus the equation for the 50 mg/day main­tenance dose would be q~ = (17.2 + 3.74x) (50 mg/day/lOOmg/day) or q~ + 8.6 + 1.87x. Main­tenance dose equations were calculated over a 50 to 150 mg/day dose range in 25mg increments. To utilise the nomogram, the user locates the patient's 36-hour nortriptyline test dose concentration on the x-axis and then from this point visually scans the y-axis to determine which intersecting nortripty­line maintenance dose falls within the therapeutic window. As an example, a 30 ngiml nortriptyline test dose concentration predicts that 50, 75, and 100 mg/day maintenance doses will produce thera­peutic levels of65, 97, and 129 ng/ml, respectively.

The disadvantages of the single-point equation are the inflexibility of the test dose sampling times and choice of tricyclic antidepressant, as well as ignorance of the patient's half-life. The plasma sampling problem is critical, since if the 36-hour single-point sample or the II-hour steady-state sample is missed, there is no alternative procedure. With the multiple-point sampling method, plasma sampling errors are easily overcome by remem­bering that only 2 samples are necessary, preferably drawn at least 24 hours apart. The single-point method is nortriptyline-specific, whereas the mul­tiple-point method can be applied to all of the tri­cyclic antidepressants. Thus, dosing nomograms need to be derived for the other antidepressants. Finally, one other caution is necessary concerning the single point method. DeVane (1980) has noted

Two Prospective Dosing Methods for Nortriptyline

that the relationship between a tricyclic antidepres­sant test dose serum concentration and a steady­state concentration is linear over only a limited range of (css)?D values. Thus as in the case of this dosing nomogram, maintenance dose predictions outside the experimentally tested range should not be made. Thus the acceptable tested dosage pre­diction range for the nomogram is only 50 to 150 mg/day.

As regards the individual half-life value, since this is unknown, it is probably not reasonable to draw a steady-state concentration until at least 8 days after the maintenance dose has been started. This recommendation is based upon the fact that in this population of 27 patients, the half-life val­ues ranged from 12.6 to 40.9 hours, with a mean nortriptyline half-life of 26.5 ± 6.7 hours. Thus,

200 1~0 ~g/d

175 ./

./ ./

'" .; ./

150 , .; .;

f ./ ./ -.

./ Cl S 125 '" <.i ./

./

562

steady-state concentrations can be ascertained sooner with the multiple-point method.

Finally, both methods are equivalent regarding speed of implementation in that the test dose must be followed by a 36-hour drug-free period. The mUltiple-point method can be implemented over a shorter period but this has been shown to slightly reduce the sensitivity and accuracy of the method - although to a statistically insignificant degree (Browne et a\., 1983).

Therapeutic Implications

After considering the advantages and disadvan­tages of the two predictive methods, it is recom­mended that clinicians determine which method satisfies the needs of their practice. We recom-

125 mg/d 100 mg/d ./

./ ./

75 mg/d

/./ /.

/ . / .

/' 50 mg/d /.

co ./ /. 0 ./ ..,.--U

'" /././. ell ..,.--.S 100 ./ ..,.--~ ./ ..,.--Q.

~ ..,.--0 /' ..,.--co ..,.--~ 75 /. ..,.--f! ./' ..,.--II> ::.. ---"0 ---<'II ~ 50 ..,.--II> -..,.--::; 0 -';' - 25

15 20 25 30 35 40 45 50 55

36-hour single-point nortriptyline cone. (ng/ml)

Fig. 1. Nortriptyline single-point dosing nomogram. From the patient's 36-hour nortriptyline test dose concentration on the x-axis, the maintenance dose which will produce steady-state concentrations within the therapeutic window (50 to 150 nglml) can be de­termined.

Two Prospective Dosing Methods for NortriPtyline

mend that either one or both methods be utilised since both possess significant advantages over em­pirical dosing, which normally requires 2 to 4 weeks for titration and evaluation of clinical response in order to determine the therapeutic dose. It is as­sumed that either ofthese methods is considerably faster. Also, it has been estimated that 70% of patients taking tricyclic antidepressants fail to take 25 to 50% of their prescribed dose (Kessler, 1978). Since 40% of patients receiving tricyclics have plasma concentrations outside the therapeutic range, it is apparent that patient compliance can be improved by the utilisation of the prospective dosing procedures described herein.

References

Alexanderson, B.: Pharmacokinetics of nortriptyline in man after single and multiple oral doses. The predictability of steady state plasma concentrations from single dose plasma level data. European Journal of Clinical Pharmacology 4: 82-91 (1972).

American Psychiatric Association: Affective disorders; in Diag­nostic and Statistical Manual of Mental Disorders, pp. 205-224 (American Psychiatric Association, Washington DC 1980).

Braithwaite, R.; Montgomery, S. and Dawling, S.: Nortriptyline in depressed patients with high plasma levels. Clinical Pharmacology and Therapeutics 23: 303-308 (1978).

Browne, J.L.; Perry, P.J.; Alexander, B.; Sherman, A.D.; Tsuang, M.T.: Dunner, FJ. and Pfohl, B.M.: Pharmacokinetic pro­tocol for predicting plasma nortriptyline levels. Journal of Clinical Psychopharmacology 3: 351-355 (1983).

Cooper. T.B.; Allen, D. and Simpson, G.M.: A sensitive method for the determination of amitriptyline and nortriptyline in human plasma. Psychopharmacology Communications 2: 105-116 (1976).

Cooper, T.B. and Simpson, G.M.: Prediction of individual dosage of nortriptyline. American Journal of Psychiatry 135: 330-335 (1978).

563

Dawling, S.: Crome, P.; Braithwaite, R.A. and Lewis, R.R.: Nor­triptyline therapy in elderly patients: Dosage prediction after single dose pharmacokinetic study. European Journal of Clinical Pharmacology 18: 147-150 (1980).

DeVane, c.L.: Tricyclic antidepressants; in Evans et al. (Eds) Ap­plied Pharmacokinetics. Principles of Therapeutic Drug Monitoring, pp.549-585 (Applied Therapeutics, San Fran­cisco 1980).

Gibaldi, M. and Perrier, D.: Pharmacokinetics, p. liS (Marcel Dekker, New York 1975).

Kessler, K.A.: Tricyclic antidepressants: Mode of action and clinical use; in Lipton et at. (Eds) Psychopharmacology: A Generation of Progress, pp. 1289-1302 (Raven Press, New York 1978).

Kragh-Sorensen, P. and L<\rsen, N.-E.: Factors influencing nor­triptyline steady state kinetics: Plasma and saliva levels. Clinical Pharmacology and Therapeutics 28: 796-803 (1980).

Madakasira, S.; Preskorn, S.H.: Weller, R. and Pardo, M.: Single dose prediction of steady state plasma levels of amitriptyline. Journal of Clinical Psychopharmacology 2: 136-139 (1982).

McNemar, Q.: Psychological Statistics, pp. 294-303 (John Wiley and Sons, New York 1962).

Potter, W.Z.: Zavadil, A.P.; Kopin, I.J. and Goodwin, F.K.: Single­dose kinetics predict steady-state concentrations of imipra­mine and desipramine. Archives of General Psychiatry 37: 314-320 (1980).

Risch, S.c.; Kalin, N.H.; Janowsky, D.S. and Huey, L.Y.: Indi­cations and guidelines for plasma tricyclic antidepressant con­centration monitoring. Journal of Clinical Psychopharmacol­ogy I: 59-63 (1981).

Slattery, J.T.; Gibaldi, M. and Koup, J.R.: Prediction of main­tenance dose required to attain a desired drug concentration at steady state from a single determination 'of concentration" after an initial dose. Clinical Pharmacokinetics 5: 377-385 (1980).

Address for correspondence and reprints: Dr Paul J. Perry, Col­lege of Pharmacy, University of Iowa, Iowa City. IA 52242 (USA).