Nortriptyline in the Treatment of ADHD: A Chart Review of S8 Cases

TIMOTHY E. WILENS, M.D., JOSEPH BIEDERMAN, M.D., DAVID E. GEIST, RONALD STEINGARD, M.D.,AND THOMAS SPENCER, M.D.

Abstract. Objective: The potential benefit of the tricyclic antidepressant medication, nortriptyline (NT), inthe treatment of children and adolescents with attention deficit hyperactivity disorder (ADHD) was evaluated .Method : A systematic search was conducted from a computerized data base of all clinic patients in an outpatientpediatric psychopharmacology unit treated with NT for ADHD. The records of the 58 subjects identified (37children and 21 adolescents) were reviewed for overall response, dose, serum levels, and adverse effects. Results:Ninety-seven percent of the identified subjects had failed to respond to an average of four previous medicat iontrials, 84% had at least one comorbid diagnosis with ADHD, and 47% were receiving at least one concurrentmedication. NT doses ranged from 0.4 to 4.5 mglkg (X± SD = 2.0 ± 1.0 mglkg) and subjects received NT from0.4 to 57.9 months (11.9 ± 14.0 months). Overall , 76% of subjects were considered to have a moderate to markedimprovement by an independent rater, which was corroborated by their clinicians . There was no association betweenresponse and age, rate of comorbidity, number of previous medication trials, or concurrent pharmacotherapy (allp NS). Although there were no overall differences in serum NT levels between responders and nonresponders,significantly more patients within the suggested therapeuti c range in adults of 50 to 150 ng/ml were classified as"markedly improved" than those outside this range (68% versus 35%, p < 0.03). Mild adverse effect s werereported in 20 subjects (34%). Conclusions: These findings suggest that NT may be an effective well-toleratedagent for ADHD children and adolescents. Additional controlled investigations utilizing NT for ADHD should beundertaken J. Am. Acad. Child Adolesc. Psychiatry, 1993, 32, 2:343- 349. Key Words: nortriptyline, serumnortriptyline levels, attention deficit hyperactivity disorder , comorbidity .

Attention-deficit hyperactivity disorder (ADHD) is a het­erogeneous disorder of unknown etiology . It is a major clini­cal and public health problem in the United States in termsof morbidity and disability of children, adolescents, andadults (Anderson et aI., 1987). There is increasing recogni­tion that there is considerable comorbidity with ADHD,including oppositional and conduct disorders as well as de­pressive and anxiety disorders (Biederman et al., 1991).These findings suggest that subgroups of ADHD may havediffering clinical courses and perhaps different pharmaco­logical responses (Biederman, 1991). Although the DSM­III-R definitions of ADHD anchor this report, some of thestudies described below use other means of classification,including DSM-III attention-deficit disorder, and dimen­sional descriptions of clinical syndromes. For simplicity ofexposition, the names of DSM-III-R categories will be usedgenerically unless otherwise specified .

Although the psycho stimulants have traditionally beenthe medication of choice in the treatment of ADHD, approxi­mately 10% (Elia et aI., 1991) to 40% (Barkley , 1977; Kleinet aI., 1980; Wilens and Biederman, 1992) of ADHD chil­dren and adolescents who are treated with stimulants arenonresponders or cannot tolerate the associated adverse ef-

Accepted July 23, 1992.From The Pediatric Psychopharmacology Unit, Consolidated De­

partments of Psychiatry, Massachusetts General Hospital, and Har­vard Medical School, Boston, MA.

This study was supported by USPHS (NIMH) grant MH-41314(JB).

Reprint requests to Dr. T.E. Wi/ens, Pediatric PsychopharmacologyUnit, ACC 725, Massachusetts General Hospital, Boston, MA 02114.

0890-8567/93/3202--o343$03.0010©1993 by the American Acad­emy of Child and Adolescent Psychiatry .

J.Am.Acad. Child Adolesc. Psychiatry, 32:2, March 1993

fects . Additionally, stimulants are short-acting agents , oftenrequiring multiple daily dosing, which can lead to noncom­pliance, social embarrassment of having to take medicine inschool, and marked symptom rebound (Wilens and Bieder­man, 1992). These findings highlight the need for alternativesafe and effective therapeutic agents in the treatment of thisdisorder , particularly in comorbid and stimulant-refractorycases.

Within the past two decades the tricyclic antidepressants(TCA) have been used increasingly as an alternative treat­ment to the stimulants for ADHD (Biederman et al., 1989a,1989b; Biederman, 1991). The TCAs offer several advan­tages over the stimulants in the treatment of ADHD. TCAshave a long half-life permitting flexible dosing, minimal riskfor abuse or dependence, and may have beneficial effects incomorbid anxiety and depressi ve conditions within ADHD(Biederman et aI., 1992). However, TCAs also are associ­ated with important adverse effects. For example, they re­quire careful cardiovascular monitoring and may beassociated with withdrawal adverse effects when abruptlydiscontinued. Efficacy of the TCAs for ADHD has beenreported for amitriptyline (Kraft et aI., 1966; Krawkowski,1965), imipramine (Gross, 1973; Rapoport, 1965; Rapoportet al., 1974; Werry , 1980; Winsberg et aI., 1972), desipra­mine (DMI) (Biederman et aI., 1986, 1989a, 1989b; Garfin­kel et aI., 1983; Gastfriend et al., 1984), and clornipramine(Garfinkel et al., 1983). Open investigations with TCAs forADHD have generally shown 80% to 90% response rates(Biederman et aI., 1986; Gastfriend et aI., 1984; Rapoport,1965; Saul, 1985), whereas the response rate ranges from48 % (Rapoport et aI., 1974 ) to 68% (Biederman et al.,1989a) in controlled studies. Hence, for children who do notrespond to stimulants, TCAs are thought to be an appropriate

343

WILENS ET AL.

drug of second choice (Biederman, 1991; Pliszka, 1987).The secondary amine TCAs, DMI and nortriptyline (NT),

have been preferred in clinical practice due to their morebenign adverse effects profile (Meltzer, 1987). While DMIhas been investigated in the treatment of children and adoles­cents with ADHD (Biederman et aI., 1986, 1989b; Garfinkelet aI., 1983; Gastfriend et aI., 1984), surprisingly, NT hasreceived little attention.

In an initial open trial of NT in 16 autistic inpatients withhyperactivity, Kurtis (1966) found NT to be "moderately"effective in treating the behavioral problems of these sub­jects. In a case review of 118 children and adolescents with"prominent hyperactive symptoms," Watter and Dreifuss(1973) found NT to be 55% effective in treating ADHDsymptoms. However, the study population was poorly de­fined and included a large proportion of children with "di­verse neurological disorders." Similarly, Saul (1985)described a 90% response rate with NT in 60 stimulantrefractory patients with ADHD. However, in that retrospec­tive report, the study population was diverse with a paucityof information on diagnostic criteria, outcome assessmentmeasures, weight-corrected dosing, and NT levels.

Since the TCAs amitriptyline, imipramine, clomipramine,and DMI have been shown to be effective in the treatmentof ADHD, and there has been initial evidence of efficacy ofNT in pediatric populations with ADHD-like symptoms, wespeculated that NT would also be effective in the treatmentof ADHD in children and adolescents. To this end, we retro­spectively evaluated the efficacy of NT in a large populationof ADHD children and adolescents.

Method

This was a systematic retrospective chart review derivedfrom an electronic data base of all children and adolescentsreceiving routine clinical care in a specialized Pediatric Psy­chopharmacology Unit who were being treated with anydosage of NT for ADHD for at least 1 week. Excluded werepatients with tic disorders (N = 9) and mental retardation(N = 2). Records of identified subjects (N = 58) werereviewed for diagnosis(es), previous medication trials, NTdose, serum NT concentration, adverse effects, and responseto NT treatment. All diagnoses were made by the treatingchild psychiatrists using DSM-III-R (American PsychiatricAssociation, 1987) criteria based on interviews with theparents and child, and upon review of school and psycholog­ical testing information. In addition to the clinical evalua­tion, the diagnosis of ADHD was confirmed at the time ofintake using the module of ADHD from the Kiddie-Schedulefor Affective Disorders and Schizophrenia for School-AgeChildren-Epidemiologic Version (Orvaschel, 1985). This in­strument has been extensively used by our group over thepast 5 years in family-genetic studies of psychiatrically andpediatrically referred children and adolescents with ADHD(Biederman et aI., 1992b).

Extent of response of ADHD symptoms to NT treatmentwas determined by an independent board certified child psy­chiatrist by review of the clinical notes, and verified withthe treating child psychiatrist. To rate level of improvementin ADHD. symptoms, a global scale termed Clinician's

344

Global Improvement Rating (CGIR) was used. This was anordinal rating system from 1 to 3 in which a rating of 1denoted worsened or unchanged symptoms, a rating of 2denoted mild to moderate improvement, and a rating of 3denoted marked improvement.

Serum NT levels were available on 45 patients. Venousblood samples were drawn 10 to 14 hours after the last doseof NT after a minimum of 5 days on a constant dose forsteady-state serum level determination as previously re­ported with DMI (Biederman et al., 1989a, 1989b). NTserum concentrations were determined by liquid chromotog­raphy (Puopolo and Flood, 1987) in the clinical chemistrylaboratories at the Massachusetts General Hospital. ECGswere obtained on all patients at baseline and while on NTby a standard 12-lead resting ECG procedure in the supineposition.

Continuous data were analyzed by unpaired, Student's t­test and ANOVA as indicated. Categorical data were ana­lyzed by chi-squared analysis. Statistical significance wasdefined at the 0.05 level (p > 0.05, NS) and all tests weretwo-tailed. Data are expressed as X ± SD unless otherwisestated.

Results

The electronic search identified 37 7- to 12-year-old chil­dren (10.3 ± 1.6 years) and 21 13- to 18-year-old adoles­cents (15.1 ± 1.9 years). Patients received NT for a meanof 11.9 ± 14.0 months (range = 0.4 to 57.9 months). Fifty­four percent of patients received NT treatment for at least 6months and 34% for more than 1 year.

There were no significant differences in gender represen­tation between children and adolescents (X2 = 1.0, p NS).Eighty-four percent (N = 49) of the total sample of ADHDsubjects had at least one comorbid diagnosis; of those, 48%(N = 28) had one, 28% (N = 16) had two, and 9% (N =5) had three comorbid diagnoses (Table 1). There were nosignificant differences in rates of comorbidity with ADHDbetween children and adolescents (X2 = 5.9, p NS).

All but two subjects (97%) had been previously treatedwith other pharmacological interventions. In these cases, thetreating clinician indicated in the record that these patientsdid not adequately respond or could not tolerate the previousmedication(s). Thus, these cases were considered treatment­resistant. The two subjects in whom NT was the first medica­tion used included two adolescent boys (14 and 15 yearsold) with a comorbid mood disorder in one and comorbidmood, anxiety, and oppositional defiant disorders in theother. Of the total sample (N = 58), 19% (N = 11) had oneprevious medication trial, 24% (N = 14) had two trials, 21%(N = 12) had three trials, 19% (N = 11) had four trials, 9%(N = 5) had six trials, and 5% (N = 3) had seven or moreprevious trials. Of the 56 subjects with previous unsuccess­ful medication trials, 60% (N = 35) received methylpheni­date (0.5 to 2 mg/kg) and 76% (N = 44) DMI (1.0 to 5.5mg/kg). There were no differences in the number of previoustreatments between children (4.4 ± 1.9 trials) and adoles­cents (3.2 ± 1.3 trials; X2 = 8.6, p NS). Likewise, therewere no significant differences in the types of previous med­ication treatments between children and adolescents (X2 <

J.Am. Acad. Child Adolesc. Psychiatry, 32:2, March 1993

3.6, p NS). Concurrent medications were used with NT in27 subjects (47%) and included stimulants (N = 10), lithiumcarbonate (N = 9), clonidine (N = 8), neuroleptics (N = 5),anticonvulsants (N = 3), and clonazepam (N = 1). Elevensubjects (19%) received two concurrent medications withNT.

Overall, 76% of subjects responded to NT treatment withmoderate to marked improvement (Table 2); of those, 48%had a marked response. There was a trend , albeit not signifi­cant , for an increased overall response (CGIR 2 and 3) inadolescents (19/21 , 90%) comp ared with children (25/37,68%) treated with NT (X2 = 3.8, P NS). There were nogender differences in response to NT (X2 = 0.7, P NS).Seventy-two percent (N = 25) of patients who had a previ­ous methylphenidate trial (N = 35) had a positi ve responseto NT. Of the patients who had a previous DMI trial (N =44),70% (N = 31) had a positive response to NT, of whom45% (N = 20) had a marked response to treatment. Ofthe patients who had previous treatments of both DMI andmethylphenidate (N = 26), 65% (N = 17) had a positiveresponse to NT. As expected, subjects with the poorest re­sponse had the shortest treatment duration (2 .7 ± 2.6months ) compared with those showing moderate improve­ment (5.9 ± 5.5 months), and those with a marked sustainedimprovement and the longest duration of treatment (20.1 ±16.1 months).

T ABLE 1. Clinical Characteristics of Subjects (N = 58)

Age (years) (X ± SD) 12.1 ± 2.9N

MalesComorb id Diagnoses"

Mood disorderOppositional defiant disorderConduct disorderAnxiety disorderDevelopmental learn ing disorderPervasive developmental disorderObsessive compulsive disorder

Previous Medication trials"StimulantTricyclic antidepressantClonidineOther (benzodiazepine,

neurolepti c, anticonvulsant)

a Categories are not mutually exclusive.

5049341855541

56354417

29

%

86845931

99972

97607629

50

NORTRIPTYLINE FOR ADHD

There were no differences in the number of comorbiddiagnoses between the poor, moderate, and marked respond­ers (1.9 ± 0.9,2.6 ± 0.8, and 2.3 ± 0.8 comorbid diagnoses,respectively, X2 = 7.2, P NS). Additionally, having either aspecific or no comorbid disorder with ADHD did not influ­ence the response to NT treatment (all X2 < 4.1, all ps NS).Of the nine patients with ADHD without comorbid diagnosis(of whom eight had failed trials of both stimulants andTCAs), 56% (N = 5) had a moderate or marked response.It is noteworthy that all ADHD patients with condu ct disor­der (N = 5) responded positi vely to NT.

Those subjects in the poorest response group tended tohave more prior medication treatments (4.7 ± 1.9 trials)compared with those having a moderate (3.8 ± 1.5 trials)and marked (3.6 ± 1.8 trials) response (X2 = 9.4, p NS).There were no associations between specific types of previ­ous unsuccessful medication trials and response (all X2 <3.0, ps NS). There were no significant differences in re­sponse rate between subjects receiving (N = 27) and notreceiving (N = 31) concurrent medications with NT (X2 =

0.3, p NS). Further analysis by specific types of adjuncti veagents also failed to reveal significant diffe rences (all X2 <3.4, P NS).

Patients received daily NT doses of 73.6 ± 33.1 mg(range = 20 to 200 mg) and weight-corrected doses of 1.94± 0.99 mg/kg (range = 0.4 to 4.5 mg/kg). There were nodifferences between responders (N = 44) and nonresponders(N = 13) in either daily NT doses (74.8 ± 31.9 versus 70.0± 38.0 mg, respecti vely, t = 0.5, p NS) or in weight­corrected doses (1.9 ± 0.9 versus 2.1 ± 1.2 mglkg , t = 0.7,P NS). There were no differences in NT doses among sub­jects with the poorest response (treatment duration = 2.7months; daily NT dose = 2.13 ± 0.34 mg/kg), moderateresponse (treatment durat ion = 5.9 months; daily NT dose= 1.60 ± 0.24 mg/kg ), and those with a marked response(treatment duration = 20.1 months; daily NT dose = 2.20± 0.17 mg/kg, F = 1.6, P NS).

There were no significant overall differences in serumNT levels in responders compared with nonresponders (96.3± 51.6 versus 83.4 ± 43.1 ng/ml , t = 0.6, p NS). In orderto further evalu ate the association between response and NT,finding s were analyzed after stratification of the sampleinto those whose NT levels fell within and outside of the .suggested therapeutic range in adult mood disorders of bloodNT levels (50 to 150 ng/ml) (Asberg et aI., 1971). Thisanalysis revealed that significantly more patients within this

T AB LE 2. Effect of Comorbidity on Global Response" of Treatment-Resistant ADHD Patients on Nortriptyline

Comorbid Diagnosis"

All ADHD+ Mood disorder+ Oppositional defiant disorder+ Conduct disorder

N

5834185

N

4429145

All ModerateResponders Respond ers

% N %

76 16 2875 11 3278 4 22

100 3 00

MarkedResponders

N %

28 4818 5310 562 40

a Ratings were determined by the Clinician's Global Improvement Rating of ADHD symptom s.b Diagnostic categories were not mutuall y exclusive.

J.Am.Acad. Child Adolesc. Psychiatry,32:2, March 1993 345

WILENS ET AL.

range were classified as "markedly improved" than thoseoutside of this range (68% versus 35%; X2 = 7.0, p < 0.03).

Adverse effects attributable to NT treatment were gener­ally mild and occurred in 34% (N = 20) of patients. Specificside effects of treatment included lethargy and gastrointesti­nal distress (N = 5 each), weight gain (N = 4), symptomaticorthostatic hypotension (N = 3), insomnia and increasedagitation (N = 2 each). Adverse effects were severe enoughto warrant discontinuation of medication in a 9-year-old boywho became agitated on the medication. ECG follow-upassessment did not reveal clinically significant conductionabnormalities in any patient.

Discussion

The current retrospective analysis of our clinic's entireexperience of NT for ADHD identified a group of highlytreatment-resistant ADHD patients, of whom 76% re­sponded to NT treatment with 48% of the patients having amarked response. Neither the presence of comorbidity withADHD, number or type of previous medication trials, norconcurrent pharmacotherapy affected outcome. There was atrend for those patients with NT serum levels in the 50to 150 ng/ml range to have a more favorable response totreatment. Among previous nonresponders to DMI, 70% hada positive response to NT treatment.

The current findings are consistent with other reports ofNT treatment in ADHD (Kurtis, 1966; Saul, 1985; Watterand Dreifuss, 1973). The current findings of a marked re­sponse in 48% of children and adolescents with NT aresimilar to findings reported by Watter and Dreifuss (1973),who found a 55% response rate in a chart analysis of 118NT-treated hyperactive children (ages 2 to 15). Differencesbetween the current investigation and the study by Watterand Dreifuss (1973) are notable for a diverse patient popula­tion which included patients with neurological disorders,mental retardation, and learning handicaps and did not reportthe diagnostic criteria used for the definition of ADHD. Inthe current study the diagnosis of ADHD was DSM-III-R­based, and patients with mental retardation and tic disorderswere excluded. Additionally, the current study used a higherdaily NT dose range (20 to 200 mg versus 20 to 75 mg) andmonitored serum NT levels. Despite these methodologicaldifferences, our findings mirrored the results of Watter andDreiffus (1973) and support the notion of a favorable re­sponse to NT in a comorbid population of ADHD childrenand adolescents, despite previous stimulant failures.

The findings in the current study also are consistent witha more recent report by Saul (1985), who found that 90%of children receiving NT for ADHD responded positivelyas assessed by teachers after 6 weeks of treatment. Similarto the clinical characteristics of our sample, in that openstudy of 60 ADHD patients, 50% had comorbid depressionand 50% had failed prior stimulant trials (Saul, 1985). Fol­lOW-Up assessment measures were not described in detail;and although clinical response was recorded at 6 weeks, nooverall duration of effect was reported. The current findings,combined with two previous reports of NT for ADHD (Saul,1985; Watter and Dreifuss, 1973), suggest that NT appearsto be effective in treating ADHD in approximately 68% of

346

a combined total of 236 children and adolescents studied.The response to NT found in this report (76%) is also

consistent with reports of ADHD treatment with otherTCAs, including imipramine (Rapoport, 1965; Winsberg etaI., 1972), amitriptyline (Kraft et aI., 1966; Krawkowski,1965), chlomipramine (Garfinkel et aI., 1983), and DMIBiederman et aI., 1986; Garfinkel et aI., 1983; Gastfriend etaI., 1984). This is particularly notable considering the highrate of comorbidity (84%) and treatment-resistant status(97%) in the current study. Based on studies of DMI inwhich there was an approximate 20% difference in responserate between open (83 to 90%) (Biederman et aI., 1986;Gastfriend et aI., 1984) and controlled investigations (68%)(Biederman et aI., 1989a), it is possible to speculate that asimilar difference would be found in a controlled study ofNT.

It has been suggested that the response to treatment withTCAs may be short-lived (Quinn and Rapoport, 1975; Rapo­port et aI., 1974). In the current study, the response to NTwas generally noted by clinicians within 2 to 4 weeks of theinitiation of NT and was sustained in responders for a meanof 15 months. These findings are consistent with reports ofsustained effects of DMI in ADHD (Biederman et aI., 1986;Gastfriend et aI., 1984). While these findings suggest thatresponse to NT treatment may be sustained, this effect maybe due to the use of robust doses. For example, consideringthe approximate two-fold potency equivalency of NT com­pared with other TCAs (Meltzer, 1987), the mean weight­corrected NT daily dose of 1.9 mg/kg used in our series ishigher than the estimated weight-corrected imipramine doseof 2.9 mg/kg used in studies reporting diminished efficacyover time (Quinn and Rapoport, 1975; Rapoport et aI.,1974). While in clinical practice the lowest effective doseneeds to be individually identified, our findings indicate thatrelatively high doses of NT may be needed in most patientsto achieve and maintain efficacy of this treatment over time.

A serum NT concentration range of 50 to 100 ng/mlhas been suggested for the treatment of juvenile depression(Geller et aI., 1985b, 1986, 1992) and that of 50 to 150 ng/ml for the treatment of adultdepression (Asberg et aI., 1971).Our preliminary data suggest that a similar window may beoperant in ADHD as indicated by our findings that signifi­cantly more patients within the 50 to 150 ng/ml range wereclassified as "markedly improved" than those outside ofthis range (68% versus 35%). Although these pilot data maybe a useful dosing guide in the treatment of children andadolescents with ADHD, adequate prospective controlledstudies evaluating the relationships of response to NT areneeded before clinical recommendations concerning optimalNT dosages in the treatment of ADHD can be delineated.

In our series 97% of the patients received other pharmaco­therapy prior to NT treatment including stimulants and DMIwithout adequate success, and our findings seem to suggestthat NT may be helpful in treatment-resistant ADHD pa­tients. However, the basis for our definition of treatment­resistance was limited by the available information in themedical record. While lack of adequate definition of treat­ment-resistance does not detract from the positive responseto NT observed in our series, our findings cannot fully

J.Am.Acad.ChildAdolesc.Psychiatry,32:2,March1993

answer whether NT is effective in the treatment of ADHDchildren in general, or only in those with treatment-resistantADHD. To address this important clinical issue, future stud­ies evaluating response to NT should include ADHDpatients with and without a history of prior treatment­resistance using a more rigorous definition of treatment­resistance than the one used in our series.

It is of note that 70% ofthose patients who had failed aprevious trial of DMI responded favorably to NT. While thereasons for these findings remain unknown, if replicated inprospective comparative studies, they may suggest that someADHD patients respond to different TCAs. That NT wasfound effective in these treatment-resistant patients does notindicate that NT is more effective than stimulants, DMI, orother treatments. Our data seem to suggest that NT may behelpful in ADHD patients when there is a poor response orintolerable adverse effects to other treatments.

We found an 84% rate of comorbidity with ADHD in oursample, which is generally higher than reported with ADHD(Biederman et aI., 1991). It is possible that the treatment­resistant status of the ADHD children and adolescents in thisreport may be associated with this high rate of comorbidity.Conversely, comorbid presentations may be associated witha high rate of treatment failure (Biederman, 1991; Pliszka,1987). The favorable response rate reported despite this highrate of comorbidity indicates that NT may be helpful incomplex, comorbid ADHD cases. While the response rateof NT treatment in nine ADHD children without comorbid­ity was only 56%, it is important to reiterate that thesechildren failed to adequately respond to prior medications.Since approximately one-third of all ADHD patients areresistant to tradition pharmacological treatments (Biedermanet aI., 1989a; Wilens and Biederman, 1992), in our seriesmore than half of patients responded to NT providing newhope for patients and families for the treatment of a pre­viously unresponsive disorder.

Adverse effects were reported in 20 NT treated subjects(34%) and were generally mild, similar to other reports withNT (Watter and Dreifuss, 1973; Geller et aI., 1985a, 1985c,1989, 1992) and imipramine (Preskorn et aI., 1983; Puig­Antich, 1987). Moreover, since many of these patients alsoreceived concurrent medication, it is unclear whether theadverse effects were due to NT or other medications. Therewere no clinically significant electrocardiographic abnor­malities found. The most common side effects included leth­argy and weight gain. Our findings of individuals withclinically significant weight gain is discrepant with that ofa recent short-term study of NT in children (Geller et aI.,1992). However, if replicated in controlled investigations,this outcome could be of interest as weight loss has beenassociated with stimulants (Spencer et al., 1992; Wilens andBiederman, 1992) and TCAs in children (Geller et aI., 1992;Spencer et aI., 1992).

NT is the major metabolite of amitriptyline metabolism.The mechanism of action of TCAs on ADHD is not wellknown. NT may ameliorate ADHD symptoms by its effecton synaptic norepinephrine reuptake blockade with subse­quent alterations in the noradrenergic and dopaminergicpathways, both of which have been implicated in the patho-

J.Am.Acad. Child Adolesc. Psychiatry, 32:2, March 1993

NORTRIPTYLINE FOR ADHD

genesis of ADHD (Meltzer, 1987; Zametkin and Rapoport,1987).

In our series, 47% of subjects received concurrent medica­tions. This reflects, in part, the complexity and severityof the clinical picture of our patients. While the use ofmonotherapy is always preferable, complex highly comorbidcases often require combined pharmacotherapy. The use ofconcurrent medications raises issues of drug-drug interac­tions. However, the combinations used in our series arenot known for important adverse interaction when used incombination with TCAs (Physicians Desk Reference, 1991).

The findings discussed in this report should be viewedagainst their methodological limitations. The data presentedare retrospective and, as such, are subject to observer bias.However, the magnitude of improvement is consistent withother open reports of DMI (Biederman et al., 1986; Gast­friend et aI., 1984), imipramine (Rapoport, 1965), and NT(Saul, 1985; Watter and Dreifuss, 1973) for ADHD. It isnoteworthy that the sample we report represents the entireexperience with NT for ADHD in our clinic. Psychiatricdiagnoses were made by the treating child psychiatrists us­ing DSM-III-R criteria based on interviews with the parentsand child, and upon review of school and psychologicaltesting information and not by structured interviews. Whilethe use of structured diagnostic interviews would have beenmore informative for the evaluation of comorbid disordersthan the clinical assessment, we were limited by the avail­able information in the patient's medical record. Neverthe­less, this shortcoming does not invalidate the findings oncomorbidity reported here since the clinical assessment mostlikely led to underestimation of comorbid disorders. Thehigh rate of comorbidity in these cases may reflect ascertain­ment bias since patients referred to a specialized clinic maybe more refractory to treatment.

In our series, NT was used in cases deemed by the as­sessing clinician not to have adequately responded to previ­ous medication trials. Unfortunately, the availableinformation in the patients' medical record did not permitmore precision regarding the adequacy in dose or durationof previous medication efforts.

In a small number of cases (N = 2) in which NT wasdiscontinued within 1 to 2 weeks of its initiation, the shortduration of treatment may not have been long enough toadequately evaluate a treatment response. Fifty-nine percentof patients had a comorbid mood disorder. Since NT is anantidepressant, it is possible that the therapeutic effects ofthe medicine are accounted for by its antidepressant and notanti-ADHD effects. However, the assessment of responsewas based on the symptoms of ADHD and not that of theconcomitant mood disorder. Moreover, while NT is an effec­tive antidepressant for adult patients, recent controlled stud­ies in prepubertal and adolescent samples of depressedpatients have failed to show its efficacy in juvenile depres­sion (Geller et aI., 1990, 1992). More work needs to be doneto disentangle the anti-ADHD and antidepressant effects ofNT in ADHD samples with and without comorbid depres­sion. This is of timely interest as attention to comorbidityin the treatment of ADHD is considered of utmost scientificimportance as reflected in a recent NIMH initiative (RFA

347

WILENS ET AL.

MH-92-03). Fifty percent of the subjects were treated withadditional medications which may have added to an im­proved treatment response attributed to NT. Although itremains to be determined whether NT alone will be affectivein ADHD, it is noteworthy that a separate analysis compar­ing patients with and without concurrent pharmacotherapy(including specific agents) did not show differences in re­sponse to ADHD. Finally, the relationship of ADHD re­sponse and serum NT levels may have been biased asclinicians may have used the perceived "therapeutic win­dow" of 50 to 150 ng/ml in titrating NT for optimal re­sponse.

Despite methodological limitations, the present findingssuggest that NT may be an effective and well-tolerated agentfor some ADHD patients. These findings support the needfor additional controlled investigations of the efficacy andsafety of NT in ADHD, as well as comparisons with moretraditional medications such as the stimulants and otherTCAs used for ADHD. In light of concerns regarding poten­tial cardiac adverse effects of TCAs in children and adoles­cents (Winsberg, 1975), and, in particular, sudden deathsreported during DMI treatment (The Medical Letter, 1990),further evaluations of NT-associated ECG changes are nec­essary to assess the risks and benefits of this treatment.

ReferencesAmerican Psychiatric Association (1987), Diagnostic and Statistical

Manual of Mental Disorders, 3rd edition-revised (DSM-lll-R).Washington, DC: American Psychiatric Association.

Anderson, J. c., Williams, S., McGee, R. & Silva, P. A. (1987) ,DSM-lll disorders in preadolescent children. Arch. Gen. Psychiatry ,44:69-76.

Asberg, M., Cronholm, B., Sjoqvist, F. & Tuck, D. (1971), Relation­ship between plasma level and therapeutic effect of nortriptyline.Br. Med. J., 3:331-334.

Barkley, R. A. (1977), A review of stimulant drug research withhyperactive children. J. Child Psychol. Psychiat., 18:137-165.

Biederman, J. (1991), Psychopharmacology. In: Textbook of Childand Adolescent Psychiatry. ed. J. M. Weiner. Washington, DC:American Psychiatric Press, pp. 545-570.

Biederman, J., Gastfriend, D. R. & Jellinek, M. S. (1986), Desipraminein the treatment of children with attention deficit disorder. J. Clin.Psychopharmacol, 6:359-363 .

Biederman, J., Newcom, J. & Sprich, S. (1991), Comorbidity of atten­tion deficit hyperactivity disorder with conduct, depressive, anxiety ,and other disorders. Am. J. Psychiatry, 148:564-577.

Biederman, J., Baldessarini, R. J., Wright, V., Knee, D. & Harmatz,J. S. (1989a), A double-blind placebo controlled study of desipra­mine in the treatment of ADD: I. Efficacy. J. Am. Acad. ChildAdolesc. Psychiatry , 28:777-784.

Biederman, J., Baldessarini , R. J. , Wright, V., Knee, D., Harmatz, J.S. & Goldblatt, A. (1989b) , A double-blind placebo controlled studyof desipramine in the treatment of ADD : IT Serum drug levels andcardiovascular findings . J. Am. Acad. Child Adoles c. Psychiatry,28:903-911.

Biederman, J., Baldessarini, R., Wright , V., Keenan, K. & Faraone,S. (in press), A double-blind placebo controlled study of desipra­mine in the treatment of attention deficit disorder: III. Lack ofimpact of comorbidity and family history factors on clinical re­sponse. J. Am. Acad. Child Adolesc. Psychiatry,

Biederman, J., Faraone, S. V., Keenan, K., Benjamin, J. & Krifcher,B. (1992b), Further evidence for family-genetic risk factors in atten­tion deficit disorder: Patterns of comorbidity in probands and rela­tives in psychiatrically and pediatrically referred samples. Arch.Gen. Psychiatry, 49:728-738.

Elia, J., Borcherding, B. G., Rapoport, J. L. & Keysor, C. S. (1991) ,

348

Methylphenidate and dextroamphetamine treatments of hyperactiv­ity: Are there true nonresponders? Psychiatry Res., 36:141-155.

Garfinkel, B. D., Wender, P. H., Sloman, L. & O'Neill, I. (1983),Tricyclic antidepressants and methylphenidate treatment of atten­tion deficit disorder in children. J. Am. Acad. Child Adolesc. Psychi­atry, 2:343-348.

Gastfriend, D. R., Biederman, J. & Jellinek, M. S. (1984), Desipraminein the treatment of adolescents with attention deficit disorder. Am.J. Psychiatry, 141:906-908.

Geller, B., Cooper, T. B. & Chestnut, E. C. (1985a), Serial monitoringand achievement of steady state nortriptyline plasma levels in de­pressed children and adolescents. J. Clin . Psychopharmacol.,5:213-216.

Geller, B., Cooper, T. B., Chestnut, E. c., Anker, J. A., Price , D. T.& Yates, E. (1985b ), Child and adolescent nortriptyline single dosekinetics predict steady state plasma levels and suggested dose:Preliminary data . J. Clin. Psychopharmacol., 5: 154-158.

Geller, B., Cooper, T. B., Farooki, F. Q. & Chestnut, E. C. (1985c),Dose and plasma levels of nortriptyline and chlorpromazine indelusionally depressed adolescents and of nortriptyline in nondelu­sionally depressed adolescents. Am. J. Psychiatry, 142:336-338.

Geller, B., Cooper, T. B., Chestnut, E. C., Anker, J. A. & Schluchter,M. D. (1986), Preliminary data on the relationship between nortrip­tyline plasma level and response in depressed children. Am. J.Psychiatry , 143:1283-1286.

Geller, B., Cooper, T. B., McCombs, H. G., Graham, D. & Wells, J.(1989), Double-blind placebo-controlled study of nortriptyline indepressed children using a "fixed plasma level" design. Psycho­pharmacol. Bull., 25:101-108 .

Geller, B., Cooper, T. B., Graham , D. L., Marsteller, F. A. & Bryant,D. M. (1990), Double-blind placebo-controlled study of nortripty­line in depressed adolescents using a "fixed plasma level" design.Psychopharmacol. Bull., 26:85-90.

Geller, B., Cooper, T. B., Graham, D. L., Fetner, H. H., Marsteller,F. A. & Wells , J. M. (1992) , Pharmacokinetically designed double­blind placebo-controlled study of nortriptyline in 6- to 12-year­olds with major depressive disorder. J. Am. Acad. Child Adolesc.Psychiatry, 31:34-44.

Gross , M. (1973), Imipramine in the treatment of minimal brain dys­function in children. Psychosomatics, 14:283-285 .

Klein , D. F., Gittelman, R., Quitkin, F. & Rifkin, A. (1980), Diagnosisand drug treatment of childhood disorders. In: Diagnosis and DrugTreatment of Psychiatric Disorders: Adults and Children. eds. D.F. Klein, R. Gittelman, F. Quitkin and A. Rifkin . Baltimore: Wil­liams and Wilkins, pp. 590-775.

Kraft, I. A., Ardali, c., Duffy, J., Hart, J. & Pearce, P. R. (1966),Use of amitriptyline in childhood behavioral disturbances. Int. J.Psychiatry , 2:611--614.

Krawkowski, A. J. (1965), Amitriptyline in treatment of hyperkineticchildren. Psychosomatics, 6:611--614.

Kurtis, L. B. (1966), Clinical study of the response to nortriptyline onautistic children. Int. J. Neuropsychiatry, 2:298-301.

Meltzer, H. Y. (1987), Psychopharmacology. The Third Generationof Progress. New York: Raven Press .

Orvaschel, H. (1985), Psychiatric interviews suitable for use in re­search with children and adolescents. Psychopharmacol. Bull.,21:737- 748.

Physicians Desk Reference (1991). Oradell, NJ: Medical EconomicsCompany.

Pliszka, S. R. (1987), Tricy clic antidepressants in the treatment ofchildren with attention deficit disorder. J. Am. Acad. Child Adolesc.Psychiatry, 26:127-132.

Preskom, S. H., Weller, E. B., Weller, R. A. & Glotzbach, E. (1983),Plasma levels of imipramine and adverse effects in children. Am.J. Psychiatry, 140:1332-1335.

Puig-Antich, J., Perel, J. M., Lupatkin, W., et al. (1987), Imipraminein prepubertal major depressives disorders. Arch. Gen. Psychiatry,44:81-89.

Puopolo, P. R. & Flood, J. G. (1987); Detection of interference bycyclobenzaprine in liquid chromatographic assays of tricyclic anti­depressants. Clin.Chem., 33:819-820.

Quinn, P. O. & Rapoport, J.L. (1975), One year follow-up of hyperac-

J.Am. Acad. Child Adolesc. Psychiatry, 32:2,March 1993

tive boys treated with imipramine or methylphen idate. Am. J. Psy­chiatry, 132:241-245.

Rapoport, J. (1965), Childhood behavior and learning problems treatedwith imipramine. Int. J. Neuropsychiatry, 1:635- 642.

Rapoport, J., Quinn, P.O., Bradbard, G., Riddle, K. D. & Brooks, E.(1974), Imipramine and methylphenidate treatments of hyperactiveboys. Arch. Gen. Psychiatry, 30:789-793.

Saul, R. C. (1985), Nortriptyline in attention deficit disorder. Clin.Neuropharma col., 8:382-384.

Spencer, T., Biederman, J., Wright, V. & Danon, M. (1992), Growthdeficits in children treated with desipramine: a controlled study. J.Am. Acad. Child Adolesc. Psychiatry, 31:235-243.

The Medical Letter (1990), Sudden death in children treated with atricyclic antidepressant. The Medical Letter, 32:53.

Watter, N. & Dreifuss, F. E. (1973) , Modification of hyperkinetic

J.Am .Acad. Child Adolesc. Psychiatry,32:2, March 1993

NORTRIPTYLINE FOR ADHD

behavior by nortriptyline. Virginia Medical Monthly, 100:123-126.Werry, J. (1980), Imipramine and methylphenidate in hyperactive

child ren. J. Child Psycho/. Psychiatry, 21:27-35.Wilen s, T. E. & Biederman, J. (1992), The stimulants. In: The Psychi­

atric Clinics of North America, ed. D. Shafer. Philadelph ia: W. B.Saunde rs, pp. 191-222.

Winsberg, B. G. (1975), Imipramine and electrocardiographic abnor­malities in hyperactive children. Am.·J. Psychiatry, 132:542-547.

Winsberg, B. G., Bialer, I., Kupietz, S. & Tobi as, J. (1972), Effectsof imipramine and dextroamphetamine on behavior of neuropsychi­atrically impaired children. Am. J. Psychiatry, 128:1425-1431.

Zametkin, A. J. & Rapoport , J. L. (1987), Neurobiology of attentiondeficit disorder with hyperactivity: where have we come in 50years ? J. Am. Acad. Child Adolesc. Psychiatry, 26:676-686.

349