Optimal Dose of Neuroleptic in Acute Schizophrenia A Controlled
Study of the Neuroleptic Threshold and Higher Haloperidol Joseph
P.
McEvoy, MD;
Gerard E.
\s=b\ After individual determination of neuroleptic threshold (NT) doses of haloperidol, 106 patients with schizophrenia or schizoaffective disorder (Research Diagnostic Criteria) were treated openly at such doses (mean, 3.7\m=+-\2.3 mg/d) for 2 weeks. Ten responding patients were discharged and unavailable for
follow-up or refused subsequent randomization, and one nonresponding patient refused randomization. The remaining 95 responding or nonresponding patients were then randomly assigned, double-blind, to a dosage of haloperidol two to 10 times higher (mean, 11.6 \m=+-\4.7mg/d) or to a continuing NT dosage (mean, 3.4\m=+-\2.3mg/d) for another 2 weeks. Of the 58 patients exposed only to NT dosages of haloperidol, 72% clinically recovered within the 5-week trial. Higher dosages given to 47 patients did not lead to greater improvement in measures of psychosis, but did produce slightly greater declines in measures of hostility. Higher dosages did regularly lead to significant increases in distressing extrapyramidal side effects. (Arch Gen Psychiatry. 1991;48:739-745) dosage of The optimalbenefit while
a drug for an individual patient should be selected with the intent of maximizing therapeutic minimizing unwanted side ef¬ fects. When Gardos et al1 reviewed the available doseresponse studies for neuroleptic drugs in 1973, which was 20 years after these agents had been introduced, they con¬ cluded that, "Optimal dosage is a concept frequently re¬ ferred to but never clearly defined." Fortunately, more in¬ formation is now available. In 1980, Appleton and Davis2 reviewed the results of 65 studies in which chlorpromazine was compared with placebo in the treatment of acute schizophrenia. Among 26 studies in which the dosage was 300 mg/d or less, chlorpromazine proved superior to pla¬ cebo in 17 (65%). Among 13 studies in which the dosage was 301 to 500 mg/d, chlorpromazine proved superior to
placebo in 11 (85%). Chlorpromazine proved superior to Accepted
publication July 20, 1990. Department of Psychiatry, Duke University Medical Center, John Umstead Hospital, Butner, NC (Dr McEvoy); and Department of Psychiatry, University of Pittsburgh (Pa), Western Psychiatric Institute and Clinic (Mr Hogarty and Dr Steingard). Reprint requests to Department of Psychiatry, Duke University Medical Center, John Umstead Hospital, Butner, NC 27509 (Dr McEvoy). From the
for
Hogarty, MSW;
Sandra
Steingard,
Dose
MD
placebo in all 26 studies (100%) in which the dosage was greater than 500 mg/d. In 1988, Baldessarini et al3 con¬ cluded from a review of the 19 most methodologically sound (double-blind, random assignment) trials contrast¬ ing two or more neuroleptic dosages in acutely psychotic patients that "a dose-benefit relationship for neuroleptic treatment is probably to be found between the daily equiv¬ alent of 100 and 700 mg of chlorpromazine, and higher doses are not only unlikely to be more effective but may yield inferior average benefits as well as increased risk of side effects." These authors estimated the median effec¬ tive dosage for antipsychotic benefit in acutely psychotic patients to be in the range of 350 mg of chlorpromazine
equivalents daily.
Thus, a tentative conclusion from available data is that small percentage of patients will respond to 100 mg of chlorpromazine daily, more to 200 mg/d, more to 300 mg/d, and so forth, until essentially all those patients who will respond do so by 500 to 700 mg/d. An argument can be made, based on this conclusion, to treat all acutely psy¬ chotic patients with 500 to 700 mg of chlorpromazine (or 10 to 14 mg of haloperidol or fluphenazine) daily, since such dosages will cover the minimum needed for essen¬ tially all potentially responsive patients.4 However, the dose-response relationship for extrapyramidal side effects (EPSEs) appears to overlap closely that for antipsychotic benefit, with an estimated median effective dosage be¬ tween 200 and 400 mg of chlorpromazine equivalents daily.5 A substantial proportion of patients will, perhaps needlessly, suffer discomfort when treated with haloperi¬ dol dosages of 10 to 14 mg/d.6 In 1961, Haase7 hypothesized that the lowest neurolep¬ tic dosages at which individual patients develop slight in¬ creases in rigidity (he referred to this first appearance of an increase in rigidity as "crossing the neuroleptic thresh¬ old") are also the lowest dosages at which these patients attain maximum antipsychotic benefit. He argued that no greater therapeutic benefit accrues over time to patients prescribed higher dosages, only increments in unwanted side effects, and that below the neuroleptic threshold (NT) little or no benefit occurs. Simply stated, those few pa¬ tients who become stiff at 100 mg of chlorpromazine daily are the same few patients who will respond therapeutically to that dosage, and so on for the patients initially a
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Michigan User on 06/18/2015
developing rigidity
with each
progressive dosage incre¬ hypothesis has not, unfortunately, been well tested in prospective controlled trials. Only two ment. Haase's NT
small studies, both over 20 years old, appear in the lit¬ one an open trial involving 18 patients8 and the other a prospective, controlled trial involving 19 schizo¬ phrenic patients.9 In general, therapeutic equivalency be¬ tween NT and standard dosage was observed in these studies, but the NT dosage was predictably lower. Other investigators appeared to assume that Haase was suggest¬ ing a linear relationship between increasing EPSEs and clinical benefit.10 To the contrary, Haase found no con¬ tribution for increasingly severe side effects in the deter¬ mination of optimal dosage. Haase and Janssen11 believed that NT dosages could be established clinically in more than 90% of patients and that these dosages ranged most commonly between 3 and 6 mg of haloperidol daily. In this report, we present the results of a large prospec¬ tive comparison of the effects of empirically determined individual NT dosages of haloperidol and more "stan¬ dard" dosages of haloperidol among acutely ill schizo¬ phrenic patients admitted to a contemporary, brief-stay, inpatient facility. The study was initiated in mid-1986 and concluded 3 years later. The investigation sought to pro¬ vide answers to clinical questions that remain central to the treatment of acutely ill schizophrenic patients: Can a sim¬ ple but reliable and operationally defined clinical criterion indicate a low dosage at which maximum antipsychotic benefit is derived with minimal associated EPSEs? Will "nonresponding" patients after 3 weeks of low NT dosage treatment ultimately "respond" if given a further 2-week treatment exposure at the low NT dosage? Or do these initial nonresponders to the NT dosage improve only if given higher dosages? Do subsequent higher dosages, over time, result in greater improvement among early re¬ sponders compared with a continuing NT dosage? And, finally, if further improvement does follow assignment to a higher dosage, is the "benefit" offset by the greater "cost" of increased EPSEs? erature:
been prescribed for at least 3 weeks (oral neuroleptics) or for 2 months (depot haloperidol or fluphenazine). Medication noncompliance was determined by history obtained from the patient, family members, and outpatient treatment staff, as well as by the absence of EPSEs other than tardive dyskinesia.13 Exclusion criteria included the presence of an organic mental syndrome, epilepsy, or mental retardation; drug or alcohol abuse that contributed to the presentation of the index episode (eg, intoxication or withdrawal in proximity to the time of admission); or a history of neuroleptic treatment resistance as evidenced by an intractable course with severe and persistent positive psy¬ chotic features despite maintenance antipsychotic medication. Of the 106 patients, 81 were diagnosed as having schizophrenic disorder, 10 as having schizoaffective disorder-depressed, and 15 as having schizoaffective disorder-manic. For 30% of patients, the study provided the first exposure to neuroleptic medications. Fifteen of these 32 first-treated patients had been continuously psychotic for more than 6 months before admission. Among eight first-exposure patients who had been continuously psychotic for less than 1 month, six had showed a deterioration of function or intermittent psychotic symptoms for many months or years be¬ fore admission.
Study Design
For diagnostic purposes, patients typically remained drug free for 2 or 3 days after admission. On day 1 of the protocol, 2 mg of haloperidol was prescribed, and the daily dose was subse¬ quently increased by 2 mg every other day until the neuroleptic threshold was crossed (ie, rigidity increased from baseline) or a dosage of 10 mg/d was reached. (Only three of the original 117 compliant patients faded to show an increase in rigidity by 10 mg/d.) Once the NT was reached within the first 10 to 12 days of treatment (and most often by the end of 1 week), the dosage was fixed and the patient was treated openly for 14 days. If ex¬ cessive rigidity was present at a dosage of 2 mg of haloperidol daily, the dosage was reduced to 1 or 0.5 mg/d to yield only min¬ imal or mild rigidity. At day 24, all 95 continuing patients were randomly assigned, on a double-blind basis, to further treatment with haloperidol for an additional 2 weeks, either continuing at their NT dosage or at a dosage two to 10 times higher (Fig 1). If a patient's NT dosage was 0.5 or 1 mg/d, the high-dosage equivalent for patients ran¬ domized to this condition (n 5) was 5 mg/d; if 2 or 4 mg/d (n 32), the high dosage was 10 mg/d; and if 6, 8, or 10 mg/d (n 10), the high dosage was 20 mg/d. The average±SD high dosage tested between days 24 and 38 was 11.6±4.7 mg/d vs 3.4±2.3 mg/d for those continuing at the NT dosage. For respond¬ ers at day 24, hospital policy frequently necessitated discharge, as is often typical among contemporary, brief-stay, inpatient units. These patients completed the controlled phase of the study as outpatients, with pill counts and plasma haloperidol level monitoring used to assess compliance. Inspection of avadable haloperidol plasma levels between days 24 and 38 revealed no declines in haloperidol levels among patients assigned to NT dos¬ ages, and only one patient assigned to a higher dosage failed to show the expected rise. With the use of our previously published criteria for anticholinergic dosage selection, biperiden, 2 mg up to four times daily, was available on a regular basis for the treatment of EPSEs, but only after the patient's NT dosage had been identified. Lorazepam, 2 mg,15 or diphenhydramine, 50 mg, up to four times daily, also were avail¬ able as needed throughout the protocol for treatment of agitation. (As will be shown, these supplements were not used more fre¬ quently by NT responders during the open phase or by NT recip¬ ients during the controlled phase.) Determinations of NT dosage and EPSE assessments were postponed for at least 24 hours after a dose of lorazepam or cüpheiihychairiine. No other psychoactive medication (eg, hthium carbonate) was permitted. =
=
=
PATIENTS AND METHODS
Patient One hundred
Population
twenty consecutively admitted patients with
schizophrenia or schizoaffective disorder (mainly schizophrenia) diagnosed by Research Diagnostic Criteria who initially met study criteria were identified for further study and provided writ¬ ten informed consent once study procedures were reviewed and explained. (Fourteen of these patients were removed from the protocol shortly after admission: three did not develop rigidity within the haloperidol dosage range and time allowed, three were medication noncompliant, five patients with schizoaffective disorder required appropriate supplemental medication, and three patients were administratively terminated [one because of
transfer, one because of medical ülness, and one because of with¬ drawal of consent]). Of the 106 patients in the protocol, 57 were male and 49 were female; they ranged in age from 18 to 57 years
(mean±SD age, 31.5±9.5 years). All patients were rated at least "moderately ill" on the Clinical Global Impression (CGI) global severity item, and as a group they were
in the moderately severe range of psychopathology 16-item Brief Psychiatric Rating Scale [BPRS] total
(mean±SD
40.1±6.1, and mean CGI global severity score was 5.0±0.7 at baseline).12 All but three were admitted to the Inpa¬ tient Schizophrenia Unit of the Western Psychiatric Institute and Clinic, Pittsburgh, Pa. Three patients refused hospitalization but completed the study as outpatients. Patients were medication noncompliant before admission, or antipsychotic drugs had not score was
Identification of the NT Dosage Each patient was repeatedly examined over several days before the initiation of haloperidol treatment to establish a baseline level of muscle tone. Patients then were examined dady during the 10-
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Michigan User on 06/18/2015
Higher Dose
(Mean, 11.6*4.7 mg/d)
Treatment at
Identification of
Neuroleptic
Threshold Doses
Neuroleptic
Threshold Doses
(Mean, 3.7±2.3 mg/d)
Neuroleptic Threshold Dose (Mean, 3.4±2.3 mg/d)
Fig 1.—Study design.
12-day haloperidol threshold-dose finding period to identify clinically appreciable increase in rigidity. Examinations were routinely made during morning ward rounds, between 9 and 10 am daily. A simple standard examination was used, for which high in¬ terrater agreement has been reported.16 With the patient stand¬ ing and passively cooperating, the examiner flexed and extended each of the patient's arms in turn at the elbow and the wrist. Patients who had difficulty relaxing were asked to perform and concentrate on rapid-alternating movements with the other arm. The operational criteria for the NT focused on an increase in cog¬ wheel rigidity from baseline. In nearly all cases, a clear increase in palpable rigidity occurred during the dose-finding phase. For those patients in whom change was initially slight, 2 days of fur¬ ther observation at the same dosage was used to determine whether more striking rigidity developed. In those cases that re¬ mained equivocal, the haloperidol dosage was increased to the next level to elicit more clearly perceptible rigidity. However, in 17 patients this resulted in excessive rigidity and dosage reduc¬ tion back to the previous level. As noted above, in three patients an increase in rigidity was not induced within the dosage and time limitations of the protocol, and they were dropped from the study. to a
Assessments
Psychopathology ratings were completed by two of us (J.P.M. and S.S.) at major assessment points (days 0, 7, 10, 17, 24, 31, and 38) and included the 16-item BPRS, the CGI of Improvement and Severity of Illness, and two global judgments of overall ther¬ apeutic efficacy and side effects.12 Ratings were based on daily observation of the patient during the preceding 2 days as well as on a detailed interview on the scheduled assessment day. Anal¬
yses included the BPRS total score, factor scores, and items, as well as the global judgments. In addition, the physician completed EPSE and anticholinergic side effect forms at major assessment points. These ratings were also made after standard physical examinations and questioning of each patient. Assessment of EPSEs included three items ad¬ dressing bradykinesia-rigidity (reported and observed rigidity, as well as gait and posture), two items addressing tremor (re¬ ported and observed), and two items addressing akathisia (re¬ ported and observed). Each item was scored on a six-point, interval-defined scale that ranged from 0 (absent) to 5 (severe). The anticholinergic side effect form included questions regarding subjective dry mouth, blurred vision, cardiac awareness, hot and dry skin, urinary retention, and constipation. Physical exami¬ nations were also made for dry mouth, dilated pupils, tachy¬ cardia, and hot and dry skin. These measures were scored on a six-point scale that ranged from 0 (absent) to 5 (severe). A research nurse completed the Wing negative symptom scale (flatness or incongruity of affect, poverty of speech, slowness of movement, underactivity, social withdrawal, and amotivation).17 These ratings were also based on daily clinical observa-
tions of the
patient, summed over the previous 2 days, as well detailed interview on the scheduled assessment day. Fi¬ nally, those patients who were well enough to attempt a selfreport of perceived medication effects (on repeated occasions) completed the six-factor Medication Response Questionnaire.18 as on a
Plan of Analysis
At days 24 and 38, patients were judged to be responders if their BPRS total score was less than or equal to 32 (16 items), with all psychosis items (conceptual disorganization, hallucinatory be¬ havior, hostility, suspiciousness, and unusual thought content) rated "mild" or less; their CGI global severity item was rated "mild" or less; and their CGI global change item was rated at least
"moderately improved."
Baseline scores on measures of psychopathology and side ef¬ fects for responding and nonresponding patients at day 24 were compared by analyses of variance. Responding and nonrespond¬ ing patients at day 24 were further contrasted by analyses of covariance on day 10 and 24 assessments, with baseline scores as covariates. In addition, outcome on the various measures of psychopa¬ thology and side effects at day 38 was assessed by a two-factor analysis of variance representing change scores between days 24 and 38. The factors accommodated responder status at day 24 (yes/no), dosage assignment (NT/high), and their interactions. (Analyses of covariance were initially considered but not used in that the initial level differences at day 24 attributable to responder
were obviously not random.) Analyses regarding the use of as-needed medications ad¬ dressed two questions: (1) whether these medications were used more frequently by responders or nonresponders during the first 24 days of open treatment at NT haloperidol dosages and (2) whether they were used more frequently by patients assigned to NT or higher haloperidol dosages during the double-blind dos¬ age comparison (days 24 through 38). Groups were first com¬ pared as to whether different proportions of patients required as-needed medications. Subsequently, the groups were com¬ pared as to the mean number of days a medication was given and the mean total number of doses given, corrected for total days at risk. Analyses included 102 patients; four patients in whom daily lorazepam doses were used to diminish catatonic immo¬ bility were excluded.19
status
RESULTS The mean haloperidol dosage at which the 106 patients crossed the NT was 3.7±2.3 mg/d. Patients who had previously been exposed to neuroleptic treatment crossed the NT at a higher haloperidol dosage (4.3±2.4 mg/d) than did those patients taking a neuroleptic for the first time (2.1 ±1.1 mg/d; t -6.44, P-c.001). The number of patients crossing the NT at each of the available dosages is illustrated in Fig 2. Ninety-five of the 106 patients entered the randomized trial extending between days 24 and 38. Of the 11 remaining patients,
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Michigan User on 06/18/2015
significantly =
2.—Distribution of daily neuroleptic threshold haloperidol doses. Numbers at the tops of bars reflect all patients (N 706); num¬ bers in parentheses, patients first exposed to neuroleptics within this study (n 32). The median effective dose for haloperidolinduced increases in rigidity was 2 mg/d for patients exposed for the first time and 4 mg/d for previously exposed patients.
Fig
=
=
five met responder criteria and were discharged before random¬ ization at day 24 but failed to return for follow-up, and six patients
(five responders and one nonresponder) refused randomization
on
Fig 3.—Psychopathology ratings during open treatment at neuro¬ leptic threshold doses. Open symbols represent responders by day 24 (n 57); closed symbols, nonresponders by day 24 (n 49). As¬ terisk indicates responders were significantly more improved than nonresponders, . (analysis of covariance, baseline levels as covariates). BPRS indicates Brief Psychiatric Rating Scale; CCI, Clin¬ ical Global
day 24 because of the possibility of higher-dosage treatment. Outcome of the Initial
=
=
Impression.
22 R
Open Trial
At day 24, after a minimum of 2 weeks at fixed NT dosages during the open phase of treatment, 57 (54%) of the 106 patients were rated as responders. The responder criterion appears to reflect readiness for discharge; the correlation between the study assessment day that a patient was first judged to be a responder and the day of discharge was .70 (P
Study of the Neuroleptic Threshold and Higher Haloperidol Joseph
P.
McEvoy, MD;
Gerard E.
\s=b\ After individual determination of neuroleptic threshold (NT) doses of haloperidol, 106 patients with schizophrenia or schizoaffective disorder (Research Diagnostic Criteria) were treated openly at such doses (mean, 3.7\m=+-\2.3 mg/d) for 2 weeks. Ten responding patients were discharged and unavailable for
follow-up or refused subsequent randomization, and one nonresponding patient refused randomization. The remaining 95 responding or nonresponding patients were then randomly assigned, double-blind, to a dosage of haloperidol two to 10 times higher (mean, 11.6 \m=+-\4.7mg/d) or to a continuing NT dosage (mean, 3.4\m=+-\2.3mg/d) for another 2 weeks. Of the 58 patients exposed only to NT dosages of haloperidol, 72% clinically recovered within the 5-week trial. Higher dosages given to 47 patients did not lead to greater improvement in measures of psychosis, but did produce slightly greater declines in measures of hostility. Higher dosages did regularly lead to significant increases in distressing extrapyramidal side effects. (Arch Gen Psychiatry. 1991;48:739-745) dosage of The optimalbenefit while
a drug for an individual patient should be selected with the intent of maximizing therapeutic minimizing unwanted side ef¬ fects. When Gardos et al1 reviewed the available doseresponse studies for neuroleptic drugs in 1973, which was 20 years after these agents had been introduced, they con¬ cluded that, "Optimal dosage is a concept frequently re¬ ferred to but never clearly defined." Fortunately, more in¬ formation is now available. In 1980, Appleton and Davis2 reviewed the results of 65 studies in which chlorpromazine was compared with placebo in the treatment of acute schizophrenia. Among 26 studies in which the dosage was 300 mg/d or less, chlorpromazine proved superior to pla¬ cebo in 17 (65%). Among 13 studies in which the dosage was 301 to 500 mg/d, chlorpromazine proved superior to
placebo in 11 (85%). Chlorpromazine proved superior to Accepted
publication July 20, 1990. Department of Psychiatry, Duke University Medical Center, John Umstead Hospital, Butner, NC (Dr McEvoy); and Department of Psychiatry, University of Pittsburgh (Pa), Western Psychiatric Institute and Clinic (Mr Hogarty and Dr Steingard). Reprint requests to Department of Psychiatry, Duke University Medical Center, John Umstead Hospital, Butner, NC 27509 (Dr McEvoy). From the
for
Hogarty, MSW;
Sandra
Steingard,
Dose
MD
placebo in all 26 studies (100%) in which the dosage was greater than 500 mg/d. In 1988, Baldessarini et al3 con¬ cluded from a review of the 19 most methodologically sound (double-blind, random assignment) trials contrast¬ ing two or more neuroleptic dosages in acutely psychotic patients that "a dose-benefit relationship for neuroleptic treatment is probably to be found between the daily equiv¬ alent of 100 and 700 mg of chlorpromazine, and higher doses are not only unlikely to be more effective but may yield inferior average benefits as well as increased risk of side effects." These authors estimated the median effec¬ tive dosage for antipsychotic benefit in acutely psychotic patients to be in the range of 350 mg of chlorpromazine
equivalents daily.
Thus, a tentative conclusion from available data is that small percentage of patients will respond to 100 mg of chlorpromazine daily, more to 200 mg/d, more to 300 mg/d, and so forth, until essentially all those patients who will respond do so by 500 to 700 mg/d. An argument can be made, based on this conclusion, to treat all acutely psy¬ chotic patients with 500 to 700 mg of chlorpromazine (or 10 to 14 mg of haloperidol or fluphenazine) daily, since such dosages will cover the minimum needed for essen¬ tially all potentially responsive patients.4 However, the dose-response relationship for extrapyramidal side effects (EPSEs) appears to overlap closely that for antipsychotic benefit, with an estimated median effective dosage be¬ tween 200 and 400 mg of chlorpromazine equivalents daily.5 A substantial proportion of patients will, perhaps needlessly, suffer discomfort when treated with haloperi¬ dol dosages of 10 to 14 mg/d.6 In 1961, Haase7 hypothesized that the lowest neurolep¬ tic dosages at which individual patients develop slight in¬ creases in rigidity (he referred to this first appearance of an increase in rigidity as "crossing the neuroleptic thresh¬ old") are also the lowest dosages at which these patients attain maximum antipsychotic benefit. He argued that no greater therapeutic benefit accrues over time to patients prescribed higher dosages, only increments in unwanted side effects, and that below the neuroleptic threshold (NT) little or no benefit occurs. Simply stated, those few pa¬ tients who become stiff at 100 mg of chlorpromazine daily are the same few patients who will respond therapeutically to that dosage, and so on for the patients initially a
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Michigan User on 06/18/2015
developing rigidity
with each
progressive dosage incre¬ hypothesis has not, unfortunately, been well tested in prospective controlled trials. Only two ment. Haase's NT
small studies, both over 20 years old, appear in the lit¬ one an open trial involving 18 patients8 and the other a prospective, controlled trial involving 19 schizo¬ phrenic patients.9 In general, therapeutic equivalency be¬ tween NT and standard dosage was observed in these studies, but the NT dosage was predictably lower. Other investigators appeared to assume that Haase was suggest¬ ing a linear relationship between increasing EPSEs and clinical benefit.10 To the contrary, Haase found no con¬ tribution for increasingly severe side effects in the deter¬ mination of optimal dosage. Haase and Janssen11 believed that NT dosages could be established clinically in more than 90% of patients and that these dosages ranged most commonly between 3 and 6 mg of haloperidol daily. In this report, we present the results of a large prospec¬ tive comparison of the effects of empirically determined individual NT dosages of haloperidol and more "stan¬ dard" dosages of haloperidol among acutely ill schizo¬ phrenic patients admitted to a contemporary, brief-stay, inpatient facility. The study was initiated in mid-1986 and concluded 3 years later. The investigation sought to pro¬ vide answers to clinical questions that remain central to the treatment of acutely ill schizophrenic patients: Can a sim¬ ple but reliable and operationally defined clinical criterion indicate a low dosage at which maximum antipsychotic benefit is derived with minimal associated EPSEs? Will "nonresponding" patients after 3 weeks of low NT dosage treatment ultimately "respond" if given a further 2-week treatment exposure at the low NT dosage? Or do these initial nonresponders to the NT dosage improve only if given higher dosages? Do subsequent higher dosages, over time, result in greater improvement among early re¬ sponders compared with a continuing NT dosage? And, finally, if further improvement does follow assignment to a higher dosage, is the "benefit" offset by the greater "cost" of increased EPSEs? erature:
been prescribed for at least 3 weeks (oral neuroleptics) or for 2 months (depot haloperidol or fluphenazine). Medication noncompliance was determined by history obtained from the patient, family members, and outpatient treatment staff, as well as by the absence of EPSEs other than tardive dyskinesia.13 Exclusion criteria included the presence of an organic mental syndrome, epilepsy, or mental retardation; drug or alcohol abuse that contributed to the presentation of the index episode (eg, intoxication or withdrawal in proximity to the time of admission); or a history of neuroleptic treatment resistance as evidenced by an intractable course with severe and persistent positive psy¬ chotic features despite maintenance antipsychotic medication. Of the 106 patients, 81 were diagnosed as having schizophrenic disorder, 10 as having schizoaffective disorder-depressed, and 15 as having schizoaffective disorder-manic. For 30% of patients, the study provided the first exposure to neuroleptic medications. Fifteen of these 32 first-treated patients had been continuously psychotic for more than 6 months before admission. Among eight first-exposure patients who had been continuously psychotic for less than 1 month, six had showed a deterioration of function or intermittent psychotic symptoms for many months or years be¬ fore admission.
Study Design
For diagnostic purposes, patients typically remained drug free for 2 or 3 days after admission. On day 1 of the protocol, 2 mg of haloperidol was prescribed, and the daily dose was subse¬ quently increased by 2 mg every other day until the neuroleptic threshold was crossed (ie, rigidity increased from baseline) or a dosage of 10 mg/d was reached. (Only three of the original 117 compliant patients faded to show an increase in rigidity by 10 mg/d.) Once the NT was reached within the first 10 to 12 days of treatment (and most often by the end of 1 week), the dosage was fixed and the patient was treated openly for 14 days. If ex¬ cessive rigidity was present at a dosage of 2 mg of haloperidol daily, the dosage was reduced to 1 or 0.5 mg/d to yield only min¬ imal or mild rigidity. At day 24, all 95 continuing patients were randomly assigned, on a double-blind basis, to further treatment with haloperidol for an additional 2 weeks, either continuing at their NT dosage or at a dosage two to 10 times higher (Fig 1). If a patient's NT dosage was 0.5 or 1 mg/d, the high-dosage equivalent for patients ran¬ domized to this condition (n 5) was 5 mg/d; if 2 or 4 mg/d (n 32), the high dosage was 10 mg/d; and if 6, 8, or 10 mg/d (n 10), the high dosage was 20 mg/d. The average±SD high dosage tested between days 24 and 38 was 11.6±4.7 mg/d vs 3.4±2.3 mg/d for those continuing at the NT dosage. For respond¬ ers at day 24, hospital policy frequently necessitated discharge, as is often typical among contemporary, brief-stay, inpatient units. These patients completed the controlled phase of the study as outpatients, with pill counts and plasma haloperidol level monitoring used to assess compliance. Inspection of avadable haloperidol plasma levels between days 24 and 38 revealed no declines in haloperidol levels among patients assigned to NT dos¬ ages, and only one patient assigned to a higher dosage failed to show the expected rise. With the use of our previously published criteria for anticholinergic dosage selection, biperiden, 2 mg up to four times daily, was available on a regular basis for the treatment of EPSEs, but only after the patient's NT dosage had been identified. Lorazepam, 2 mg,15 or diphenhydramine, 50 mg, up to four times daily, also were avail¬ able as needed throughout the protocol for treatment of agitation. (As will be shown, these supplements were not used more fre¬ quently by NT responders during the open phase or by NT recip¬ ients during the controlled phase.) Determinations of NT dosage and EPSE assessments were postponed for at least 24 hours after a dose of lorazepam or cüpheiihychairiine. No other psychoactive medication (eg, hthium carbonate) was permitted. =
=
=
PATIENTS AND METHODS
Patient One hundred
Population
twenty consecutively admitted patients with
schizophrenia or schizoaffective disorder (mainly schizophrenia) diagnosed by Research Diagnostic Criteria who initially met study criteria were identified for further study and provided writ¬ ten informed consent once study procedures were reviewed and explained. (Fourteen of these patients were removed from the protocol shortly after admission: three did not develop rigidity within the haloperidol dosage range and time allowed, three were medication noncompliant, five patients with schizoaffective disorder required appropriate supplemental medication, and three patients were administratively terminated [one because of
transfer, one because of medical ülness, and one because of with¬ drawal of consent]). Of the 106 patients in the protocol, 57 were male and 49 were female; they ranged in age from 18 to 57 years
(mean±SD age, 31.5±9.5 years). All patients were rated at least "moderately ill" on the Clinical Global Impression (CGI) global severity item, and as a group they were
in the moderately severe range of psychopathology 16-item Brief Psychiatric Rating Scale [BPRS] total
(mean±SD
40.1±6.1, and mean CGI global severity score was 5.0±0.7 at baseline).12 All but three were admitted to the Inpa¬ tient Schizophrenia Unit of the Western Psychiatric Institute and Clinic, Pittsburgh, Pa. Three patients refused hospitalization but completed the study as outpatients. Patients were medication noncompliant before admission, or antipsychotic drugs had not score was
Identification of the NT Dosage Each patient was repeatedly examined over several days before the initiation of haloperidol treatment to establish a baseline level of muscle tone. Patients then were examined dady during the 10-
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Michigan User on 06/18/2015
Higher Dose
(Mean, 11.6*4.7 mg/d)
Treatment at
Identification of
Neuroleptic
Threshold Doses
Neuroleptic
Threshold Doses
(Mean, 3.7±2.3 mg/d)
Neuroleptic Threshold Dose (Mean, 3.4±2.3 mg/d)
Fig 1.—Study design.
12-day haloperidol threshold-dose finding period to identify clinically appreciable increase in rigidity. Examinations were routinely made during morning ward rounds, between 9 and 10 am daily. A simple standard examination was used, for which high in¬ terrater agreement has been reported.16 With the patient stand¬ ing and passively cooperating, the examiner flexed and extended each of the patient's arms in turn at the elbow and the wrist. Patients who had difficulty relaxing were asked to perform and concentrate on rapid-alternating movements with the other arm. The operational criteria for the NT focused on an increase in cog¬ wheel rigidity from baseline. In nearly all cases, a clear increase in palpable rigidity occurred during the dose-finding phase. For those patients in whom change was initially slight, 2 days of fur¬ ther observation at the same dosage was used to determine whether more striking rigidity developed. In those cases that re¬ mained equivocal, the haloperidol dosage was increased to the next level to elicit more clearly perceptible rigidity. However, in 17 patients this resulted in excessive rigidity and dosage reduc¬ tion back to the previous level. As noted above, in three patients an increase in rigidity was not induced within the dosage and time limitations of the protocol, and they were dropped from the study. to a
Assessments
Psychopathology ratings were completed by two of us (J.P.M. and S.S.) at major assessment points (days 0, 7, 10, 17, 24, 31, and 38) and included the 16-item BPRS, the CGI of Improvement and Severity of Illness, and two global judgments of overall ther¬ apeutic efficacy and side effects.12 Ratings were based on daily observation of the patient during the preceding 2 days as well as on a detailed interview on the scheduled assessment day. Anal¬
yses included the BPRS total score, factor scores, and items, as well as the global judgments. In addition, the physician completed EPSE and anticholinergic side effect forms at major assessment points. These ratings were also made after standard physical examinations and questioning of each patient. Assessment of EPSEs included three items ad¬ dressing bradykinesia-rigidity (reported and observed rigidity, as well as gait and posture), two items addressing tremor (re¬ ported and observed), and two items addressing akathisia (re¬ ported and observed). Each item was scored on a six-point, interval-defined scale that ranged from 0 (absent) to 5 (severe). The anticholinergic side effect form included questions regarding subjective dry mouth, blurred vision, cardiac awareness, hot and dry skin, urinary retention, and constipation. Physical exami¬ nations were also made for dry mouth, dilated pupils, tachy¬ cardia, and hot and dry skin. These measures were scored on a six-point scale that ranged from 0 (absent) to 5 (severe). A research nurse completed the Wing negative symptom scale (flatness or incongruity of affect, poverty of speech, slowness of movement, underactivity, social withdrawal, and amotivation).17 These ratings were also based on daily clinical observa-
tions of the
patient, summed over the previous 2 days, as well detailed interview on the scheduled assessment day. Fi¬ nally, those patients who were well enough to attempt a selfreport of perceived medication effects (on repeated occasions) completed the six-factor Medication Response Questionnaire.18 as on a
Plan of Analysis
At days 24 and 38, patients were judged to be responders if their BPRS total score was less than or equal to 32 (16 items), with all psychosis items (conceptual disorganization, hallucinatory be¬ havior, hostility, suspiciousness, and unusual thought content) rated "mild" or less; their CGI global severity item was rated "mild" or less; and their CGI global change item was rated at least
"moderately improved."
Baseline scores on measures of psychopathology and side ef¬ fects for responding and nonresponding patients at day 24 were compared by analyses of variance. Responding and nonrespond¬ ing patients at day 24 were further contrasted by analyses of covariance on day 10 and 24 assessments, with baseline scores as covariates. In addition, outcome on the various measures of psychopa¬ thology and side effects at day 38 was assessed by a two-factor analysis of variance representing change scores between days 24 and 38. The factors accommodated responder status at day 24 (yes/no), dosage assignment (NT/high), and their interactions. (Analyses of covariance were initially considered but not used in that the initial level differences at day 24 attributable to responder
were obviously not random.) Analyses regarding the use of as-needed medications ad¬ dressed two questions: (1) whether these medications were used more frequently by responders or nonresponders during the first 24 days of open treatment at NT haloperidol dosages and (2) whether they were used more frequently by patients assigned to NT or higher haloperidol dosages during the double-blind dos¬ age comparison (days 24 through 38). Groups were first com¬ pared as to whether different proportions of patients required as-needed medications. Subsequently, the groups were com¬ pared as to the mean number of days a medication was given and the mean total number of doses given, corrected for total days at risk. Analyses included 102 patients; four patients in whom daily lorazepam doses were used to diminish catatonic immo¬ bility were excluded.19
status
RESULTS The mean haloperidol dosage at which the 106 patients crossed the NT was 3.7±2.3 mg/d. Patients who had previously been exposed to neuroleptic treatment crossed the NT at a higher haloperidol dosage (4.3±2.4 mg/d) than did those patients taking a neuroleptic for the first time (2.1 ±1.1 mg/d; t -6.44, P-c.001). The number of patients crossing the NT at each of the available dosages is illustrated in Fig 2. Ninety-five of the 106 patients entered the randomized trial extending between days 24 and 38. Of the 11 remaining patients,
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Michigan User on 06/18/2015
significantly =
2.—Distribution of daily neuroleptic threshold haloperidol doses. Numbers at the tops of bars reflect all patients (N 706); num¬ bers in parentheses, patients first exposed to neuroleptics within this study (n 32). The median effective dose for haloperidolinduced increases in rigidity was 2 mg/d for patients exposed for the first time and 4 mg/d for previously exposed patients.
Fig
=
=
five met responder criteria and were discharged before random¬ ization at day 24 but failed to return for follow-up, and six patients
(five responders and one nonresponder) refused randomization
on
Fig 3.—Psychopathology ratings during open treatment at neuro¬ leptic threshold doses. Open symbols represent responders by day 24 (n 57); closed symbols, nonresponders by day 24 (n 49). As¬ terisk indicates responders were significantly more improved than nonresponders, . (analysis of covariance, baseline levels as covariates). BPRS indicates Brief Psychiatric Rating Scale; CCI, Clin¬ ical Global
day 24 because of the possibility of higher-dosage treatment. Outcome of the Initial
=
=
Impression.
22 R
Open Trial
At day 24, after a minimum of 2 weeks at fixed NT dosages during the open phase of treatment, 57 (54%) of the 106 patients were rated as responders. The responder criterion appears to reflect readiness for discharge; the correlation between the study assessment day that a patient was first judged to be a responder and the day of discharge was .70 (P
