728
BIOL PSYCHIATRY 1990;28:728-732
Tryptophan Treatment of Aggressive Psychiatric Inpatients Jan Volavka, Martha Crowner, David Brizer, Antonio Convit, Herman Van Praag, and Raymond F. Suckow
This double-blind, piaceSo-controlled study tested the effectiveness of tryptophan (TRP) in the treatment of aggressive psychiatric inpatiems. After a baseline observation period of I month, patients were"randomly assigned to treatment either with TRP (up to 6 g/day) or wi.thplacebo. There were 10 subjects in each treatment group. These treatments were administeredfor 25-35 days, after which the patients were observed for I month. Throughout this study, patients were receiving other medications. Injections of antipsychotics and sedat'ves were administered as needed to control agitated or violent behavior, blood levels of TRP and other large neutral amino acids were obtained repeatedly, and ratios between TRP and other amino acids were co,nputed. These analyses confirmed signi3~cant increases of TRP ratios in TRP-treated patients. TRP treatment had no effect on the number of violent incidents, but it significantly reduced the need for injections cf antipsychotics and sedatives. The study thus provided indirect support for beneficial effects of TRP in aggressive psychiatric inpatients.
introoucfion Aggressive behavior seems to be associated with a disturbance of central serotonergic function ~n patients with personality disorders (Brown 1979; Coccaro et al 1989), violent offender_~ and impulsive fire setters (Linnoi!a et al 1983; Virkkunen et al 1989), persons with 47,XYY syndrome (Bioulac et al 1980), and in alcoholics (Branchey et al 1984). A decrease of serotonin availability has been infernal fro~ decreased !eve!s of 5hydroxyindoleacetic acid (serotonin's primary mehabolite) in the cerebrospinal fluid, or the decreased levels of ~'ptophar~ (TRP) in peripheral blood. TRP is the precursor of serotonin. The firs' step in the conversion of TRf ~ tc seretonin is effectcd by TRP hydroxylase. As this cnzy:~ is unsaturated, inc~as~d TgP availability will lead to increased serotonin synthesis. Elevating the levels of TltP by increasing its dietary intake migh~ therefore result in a redu.:tion of aggressive behavior. In the one available placebo-
From the Nathar, S. Kline Institute for Psychiatric Research, Orangeburg, New York (J.V., A.C., R.F.S.); the New York University Medical Center, New York, New York ($.V., M.C., A.C.); the Manhattan Psychiatric Center, New York, New York (J.V.. M.C., A.C.); the Western lnsii~ute of Neuropsychiatry, Salt L&~.eCity, Utah (D.B.); the Albert Einstein College of Medicine/Montefiore Medical Ceatet', Bronx, New York (H.V.P.): and the New Yoik State Psychiatric Institute, New York, New York (R.F S.). Ad,~ro_ss~print reque~t.s to Jan Velavka0 M.D., Nathan S. Kline Institute for Psychiatric Research, Orangeburg, New York 10967. Received October 20, 1989; revised Febraary !'7, 1990. © 1990 Society of Biological Psychiatry
0006-3223/90/$03.50
Tryptophan Treatment of Aggression
BIOL PSYCHIATRY 1990;28:728-732
729
con~ol!ed TRP treatment study of aggression, a grout of !2 violent schizt, phrenic patients received either 4 or 8 g of TRP a day for 4 weeks. A reduction of incidents was noted in patients on either dosage regimen (Morand et al 1983). In a single case open study, the combination of TRP and trazodone reduced agg:~ss~ve behavior in a retarded man (O'Neii et al 1986). TRP competes with other large ne~li.ral amino ac;.ds (LNAA) (leucine, isoleucine, valine, phenylalanine, and tyrosine) for a carrier that transports these amino acids to "~'-,,~,brain (Fernstrom and Wurtman 1q72). Therefore, the plasma ratio of TRP to these competing LNAA is more important titan the absolute TRP plasma concentration for the. TRP influx in the brain. We report a study of TRP effects on aggression using the TRP ratios. Methods Twenty consenting psychiatric inpatients (12 men, 8 women) aged 19--5~ were the subjects. Each patient had at least two docun~.er:e:! incidents of violent, agitated, or fearinducing behavior within the mo~tb preceding the study. All of the patients had a history of persistent aggressive behavior extending for months or years prior to the onset of the study. Their primary DSM-III-R diagnoses were schizophrenia (n = 10), bipolar disorders (n = 3), personality disorder (2 borderline, 1 antisocial, and 1 organic personality disorder), 1 atypical psychosis, I impulse control disorder (not specified), and 1 mental retardation. Seven of the patients gave a history of arrest for violent crime, and the records documented a history of alcohol abuse in 5 cases. The patients were receiving various nonexperimental antipsychotic medications prior to the onset of experimental treatment: these regularly administered medications were not changed (in type or dose) for 1 month prior to experimental treatment, and then throughout the study (except in 2 patients). As usual in violent patients, certain injectable medications were used on an as-needed (p.r.n.) basis throughout the study for the control of agitation and violent behavior. These p.r.n. injections varied among patients. ~aloperidol, sodium amytal, chlorpromazine- stelazine, prolixine hydrochloride, or lorazepam were used for p.r.n, injections. For the purpose of data analysis in this study, each p.r.n, injection was treated as a discrete event (independent of the type or dose of medication it delivered). At the end of the baseline observation period of I month, subjects were randomly assigned to treatment with placebo or with TRP. Chocolate bars cont~:ined either 9.5 g of chocolate and 0.5 g L-TRP (for active treatment), or 10 g of chocolate (placebo). These placebo and TRP bars could not be distinguished from each other by appearance or taste. In 16 of the .20 patients, the medication schedule was as follows: two chocolate ba_rs on day 1, six n n d~y 'z ~hen 12 bars dai~y for 3 weeks, then 6 bars for 1 day, 'and finally two bars of chocolate on the last day of treatment. In the first 4 patients (2 TFLP, 2 placebo subjects), the period of dose escalation (from 0 to 12 bars) and dose taper (from I2 bars to (") were longer, varying between 2 d~:ys and 1 week. The reason for these variations of escalation and taper rates was cur initial uncertainty about the ~ssibi!ity of side effects and withdrawal effects of TRP. The chocolate bars were administered in divided dt~.~es (three times a day) in 16 of the 20 patients. In the first 4 patients, the total daily dose was divided so that the chocolate was administered four times a day. As raentioned above, the standing orders for regular (nonexpe-imental) medications remained anchanged. This was true also for p.r.n, injections. Hc-wever, the decisien to admini.~ter each p.r.n.
730
BIOL PSYCHIATRY 1990;28:728-7~;2
J. Volavka et al
injection depended on the judgment of ward personnel (within the limits prescribed by the doctor's standing order for the p.r.n, medication). Blood samples for TRP and other LNAA were ~"awn (7 AM and 9 AM) on the following 4 days: the day before the start of e~perimental treatment (baseline 1), the day treatment started (baseline 2), the day marking l full week on maximal dose, the first day of dose decrement, and 8 days after the end of experimental treatmer~t. Violent events were followed during the baseline period, t~eatment period, and for a posttreatment follow-up period of l month. We noted agitation, verbal as~caalt, assault on self, assault on others, and property assault. These events were abstracted from patients' progress notes and other docuraentation generated on the ward. In our experience, this method yields about twice as many incidents than the officially filed incident reports. Brief Psychi:~c Rating Scale (BPRS) (Guy 1976) and a simple scale for sedation (available on ~qucst) were administered weekly throughout the study, which was double-blind. Plasma amino acids other than TRP were assayed by a modified procedure of Bidlingmeyer et al (1984) and further described in detail by Richardson et al (1988). Results The effects of TRP enrichment on the ratio of TRP to the competing LNAA (see above) is shown in Table 1. (The original average values of TRP and other LNAA are available on request.) Treatments were administered at 7 AM, after the ~rst blood sample was drawtl. I r...r,sc r e s u l t s in{ii~;a{.c m a t ul~ztary c m l~mn~;nL a ~ m n ~ , , u . u y i n ~ 7 ~ s ~ d .i..tu~_TP~- i'aU~@
2 hr after TRP administration (compared with placebo), but the effect iargely dissipated at some point between 2 and 12 hr later. The 7 AM (pretreatment) ratios showed no significant difference between the placebo aad TRP groups, and no sir,niiicant e~¢ects of time (two-way ANOVA). TRP treatment had no effect on the number of violent incidents. The number of violent incidents per week in the placebo group (including verbal assault, assault on self, assault on others or on property) averaged over the 4-~eek pretreatment period and over subjects was 0.98. Duriqg treatment it was 0.97, and during posttreatment follow-up (5 weeks) it was 0.44. The analogous averages for ~he TRP group were 1.50 before, 1.53 during, and 0.30 after treatment. Similar lack of TRP effect wa~; observed for the BPRS factor "hostility-suspiciousness." Ratings on the sedation scale did not discriminate between TRP and placebo groups. \ As mentioned earlier, the p.r.n, medication order remained c~nstant throughou~ 0,~r~ study for individual patients, but the decision to actually admimster the injection was made by the staff depending on the patient's behavior. These p.r,a, injections were given to cope with patients' agitation and threate'~ing or violent beha~,ior. As no st~ff members were aware of the patients' group membership (placebo or TPV), the number of p.r.n. :~njections ~{:ministered was considered a valid outcome variable reflecting the stafi:'s judgment of 0atient behavior. The average numbers of p.r~a, injections are in Table 2. In c~rder to test the hypothesis that TRP treatment reduced the need for o.r.n, injections, we had to account for the pretreatment differences in the p.r.n, injections b~ween the placebo and the TRP group. This was accomplished by an analysis of covariance which showed a significant difference (f = 5,61; df = i,19; p = 0.03) between t~le number of p.r.n, injections administered (per day) duriiig the TRP and placebo treatments. An analogous analysis of covariance was performed for the posttreatmevt follow-up period. The difference between the two treatment g~'oups was no longer significant, indicating
Tryptophan Treatment of Aggression
BIOL PSYCHIATRY 1990;28:728-732
73 i
Table 1. TRP Ratios (Multiplied by 100)° Baseline 1 Treatment Placebo TRP
Baseline 2
Max. Dose
Dose Red.
Follow-up
7 AM.
9 AM
7 AM
9 AM
7 AM
9 AM
7 AM
9 AM
7 AM
9 AM
22 21
20 18
20 20
21
20 25
22 6~
24 23
23 39 t'
22 21
23 22
29 b
aEach ratio was computed by dividing TRP value by the sum of the values of the other LNAA (leucine, isoleucine, valine, phenylalanine, and tyrosine). Each ratio in Table i represents an average for 10 patients. TRP (or placebo) was administered immediately after the 7AM blood sample was d~wn. ~" bp < 0.05 (TRP vs placebo).
that the reduction of p.r.n, injections associated with TRP treatment failed to persist beyond the treatment period. As mentioned above, the neuroleptic medication for 2 of the 20 patients changed during the study. We have therefore repeated the analyses for p.r.n, injections after the exclusion of these 2 patients. The reduction of p.r.n, injections associated with TRP treatment remained statistically significant, and remained significant during the follow-up period. Finally, we should peint out that the treatment had no apparent side effects.
Discussion Our results failed to support the principal hypothesis that TRP would reduce aggressive behavior; the general decrease of assaultive behaviors under both treatment conditions (TRP and placebo) was probabb attributable to environmental effects. However, TRP treatment has apparently reduced the need for p.r.n, injectable medication. It is possible that this reduction reflected the therapeutic effect of TRP on aggressive behavior. Furthermore, one can speculate that the use of p.r.n, medication may have obs:ured the differences in aggressive behavior between the placebo and TRP treatment groups. There is a possibility that the ward staff wa3 very ju0Zcious in the use of p.r.n, inj~.ctions, administering ~hem as soon as the patient showed any sign of impending violenct:. If this happened, the p.r.n, preventative use would mask the difference between TRP and placebo treatment effects, on overt violence. The generalizability of our results is limited by the small number of subjec~:~ and the sho~ duration of the trial; the latter deficiency is par',icularly damaging i~ v~,w of the fact that violent incidents were rare. la conclusion, our study ~ra-.,'-:led indirect support for the TRP beneficial effects on _,ggressive behavior in psychiatric patients. Future studies using higher doses of TRP may show mo:'c pronounced effects. However, sach studies must wait until the re,'e~t reports (Blevins et al 1989) of seriou~ adverse effects of TRP (eosinophil'~a-myalgia syndrome) can be clarified.
Table 2. ?.:':_-'=~."Number of p.r.n, lnjectio,t~ per Da) Treatment group Placebo TRP
Baseline
~x~rimcnta~ treatment
Follow-up
0.107 0. ! 48
0.077 0.0.57"
0.061 0.075
Op = 0.03 (TRP vs placebo, ANCOVA).
732
BIOLPSYCHIATRY 1990;28:728-732
J. Valavka et al
Note Added i~ Print Recent evidence strongly suggests that the eosinophilia-myalgia syndrome associated with the ingestion of TRP was caused by a co~taminant or an alteration in a subset of TRP manufactured by a single company (Slutsker et al 1990). Nestle Research Laboratoriesprovided the TRP and placebo chocolates. Dr. Peter Leathwood (Nestle Resea.,eh) was vet:},he!p_~d and made insightful comments. Dr. Richard J. Wurtman (Massachusetts Institute of Technology) gave us hnportant advice in the planning stage of this study.
References Bidlingmeyer BA, Cohen SA, ]'arvin TL (1984): Rapid analysis ot: amino acids using pre-column derivitizafion. J Chromatogr 336:93-104. Bioulac B, Benezich M, Rena~ld B, Noel B, Roche D (1980): Serotonergic function in the 47,XYY syndrome. Biol Psychiatly 15:917-923. Blevins WL, Hertzman P, Ting M, et al t!989) Eo~inophilia-myalgia syndrome--New Mexico. JAM4 262:3116. Bre,achey L, Branchey M, Shaw S, Lieber CS (1984): Depression, suicide, and aggression in alcoholics and their relat2onship to plasma amino acids. Psychiatry Res 12:219-226. Brown GL, Goodwin I-L=K,Ballenger JC, Goyer PF, Major LF (1979): Aggression in humans correlate~ with cerebrospinal fluid amine metabolites. Psychiatry Res 1:131-139. Coccaro EF, Siever i.J, Klar HM, et al (1989): Serotonergic studies in patients with affective and person~,iity disorders. Awch Gen Psychiatry 46:587-599. Fe~stmm JD, Wurtrnan R2: (1977.): Brain serotorfin content: Physiological regulation by plasma neutra'l amino acids. Science 178:414--416. Guy W (1976): ECDEU A.:ses;ment Manual for Psychopharmacoiogy. Rockville, MD: National Institute of Mental Health. Linnoi!a M~ Virk!canen M, Scheinin M, Nuutila A, Rimon R, Goodwin FK (1983): Low cerebrospinal fluid 5-hydroxyindoleacetic acid concentration differentiates impulsive from nonimpu!aive violent behavior. Life Sci 33:2609-2614, Moraad C, Young SN, Ervin FR (1983): Clinical response of aggressive schizophrenics to o,al tryptophan, t~iol Psychiatr,y 18:575-578. O~Neil M, Page N, ,~dkins WN (1986): Tryptophan-trazodone treatment of aggressive behavior. Lancet ii:859-860. Richardson MA, Suckow R, Whittaker R, et al (1988): Phenylalanine, phenylethylamine and tardive dyskinesia in psychiatric patients .;a trace amines: Comparative and clinical neurobiology. Boulton AA, Jurio AV, Downer RGM (eds). Clifton, NJ: Humana Press, pp. 409--422. Slutsker L, Hoesly FC, Miller L, Williams LP, Watson JC (1990): Eosinophiiia-myalgia syndrome associated with exposure to tryptophan from a single manufacturer. JAMA 264:213-217. Virkkunen M, De Jong J, Bartko J, Goodwin FK, Linnoila M (1989): ]~e!ationsh;p of psychobiological variables to recidivism in violent offenders and impulsive fire setters. Arch Gen Psychiatry 46:600-603.
BIOL PSYCHIATRY 1990;28:728-732
Tryptophan Treatment of Aggressive Psychiatric Inpatients Jan Volavka, Martha Crowner, David Brizer, Antonio Convit, Herman Van Praag, and Raymond F. Suckow
This double-blind, piaceSo-controlled study tested the effectiveness of tryptophan (TRP) in the treatment of aggressive psychiatric inpatiems. After a baseline observation period of I month, patients were"randomly assigned to treatment either with TRP (up to 6 g/day) or wi.thplacebo. There were 10 subjects in each treatment group. These treatments were administeredfor 25-35 days, after which the patients were observed for I month. Throughout this study, patients were receiving other medications. Injections of antipsychotics and sedat'ves were administered as needed to control agitated or violent behavior, blood levels of TRP and other large neutral amino acids were obtained repeatedly, and ratios between TRP and other amino acids were co,nputed. These analyses confirmed signi3~cant increases of TRP ratios in TRP-treated patients. TRP treatment had no effect on the number of violent incidents, but it significantly reduced the need for injections cf antipsychotics and sedatives. The study thus provided indirect support for beneficial effects of TRP in aggressive psychiatric inpatients.
introoucfion Aggressive behavior seems to be associated with a disturbance of central serotonergic function ~n patients with personality disorders (Brown 1979; Coccaro et al 1989), violent offender_~ and impulsive fire setters (Linnoi!a et al 1983; Virkkunen et al 1989), persons with 47,XYY syndrome (Bioulac et al 1980), and in alcoholics (Branchey et al 1984). A decrease of serotonin availability has been infernal fro~ decreased !eve!s of 5hydroxyindoleacetic acid (serotonin's primary mehabolite) in the cerebrospinal fluid, or the decreased levels of ~'ptophar~ (TRP) in peripheral blood. TRP is the precursor of serotonin. The firs' step in the conversion of TRf ~ tc seretonin is effectcd by TRP hydroxylase. As this cnzy:~ is unsaturated, inc~as~d TgP availability will lead to increased serotonin synthesis. Elevating the levels of TltP by increasing its dietary intake migh~ therefore result in a redu.:tion of aggressive behavior. In the one available placebo-
From the Nathar, S. Kline Institute for Psychiatric Research, Orangeburg, New York (J.V., A.C., R.F.S.); the New York University Medical Center, New York, New York ($.V., M.C., A.C.); the Manhattan Psychiatric Center, New York, New York (J.V.. M.C., A.C.); the Western lnsii~ute of Neuropsychiatry, Salt L&~.eCity, Utah (D.B.); the Albert Einstein College of Medicine/Montefiore Medical Ceatet', Bronx, New York (H.V.P.): and the New Yoik State Psychiatric Institute, New York, New York (R.F S.). Ad,~ro_ss~print reque~t.s to Jan Velavka0 M.D., Nathan S. Kline Institute for Psychiatric Research, Orangeburg, New York 10967. Received October 20, 1989; revised Febraary !'7, 1990. © 1990 Society of Biological Psychiatry
0006-3223/90/$03.50
Tryptophan Treatment of Aggression
BIOL PSYCHIATRY 1990;28:728-732
729
con~ol!ed TRP treatment study of aggression, a grout of !2 violent schizt, phrenic patients received either 4 or 8 g of TRP a day for 4 weeks. A reduction of incidents was noted in patients on either dosage regimen (Morand et al 1983). In a single case open study, the combination of TRP and trazodone reduced agg:~ss~ve behavior in a retarded man (O'Neii et al 1986). TRP competes with other large ne~li.ral amino ac;.ds (LNAA) (leucine, isoleucine, valine, phenylalanine, and tyrosine) for a carrier that transports these amino acids to "~'-,,~,brain (Fernstrom and Wurtman 1q72). Therefore, the plasma ratio of TRP to these competing LNAA is more important titan the absolute TRP plasma concentration for the. TRP influx in the brain. We report a study of TRP effects on aggression using the TRP ratios. Methods Twenty consenting psychiatric inpatients (12 men, 8 women) aged 19--5~ were the subjects. Each patient had at least two docun~.er:e:! incidents of violent, agitated, or fearinducing behavior within the mo~tb preceding the study. All of the patients had a history of persistent aggressive behavior extending for months or years prior to the onset of the study. Their primary DSM-III-R diagnoses were schizophrenia (n = 10), bipolar disorders (n = 3), personality disorder (2 borderline, 1 antisocial, and 1 organic personality disorder), 1 atypical psychosis, I impulse control disorder (not specified), and 1 mental retardation. Seven of the patients gave a history of arrest for violent crime, and the records documented a history of alcohol abuse in 5 cases. The patients were receiving various nonexperimental antipsychotic medications prior to the onset of experimental treatment: these regularly administered medications were not changed (in type or dose) for 1 month prior to experimental treatment, and then throughout the study (except in 2 patients). As usual in violent patients, certain injectable medications were used on an as-needed (p.r.n.) basis throughout the study for the control of agitation and violent behavior. These p.r.n. injections varied among patients. ~aloperidol, sodium amytal, chlorpromazine- stelazine, prolixine hydrochloride, or lorazepam were used for p.r.n, injections. For the purpose of data analysis in this study, each p.r.n, injection was treated as a discrete event (independent of the type or dose of medication it delivered). At the end of the baseline observation period of I month, subjects were randomly assigned to treatment with placebo or with TRP. Chocolate bars cont~:ined either 9.5 g of chocolate and 0.5 g L-TRP (for active treatment), or 10 g of chocolate (placebo). These placebo and TRP bars could not be distinguished from each other by appearance or taste. In 16 of the .20 patients, the medication schedule was as follows: two chocolate ba_rs on day 1, six n n d~y 'z ~hen 12 bars dai~y for 3 weeks, then 6 bars for 1 day, 'and finally two bars of chocolate on the last day of treatment. In the first 4 patients (2 TFLP, 2 placebo subjects), the period of dose escalation (from 0 to 12 bars) and dose taper (from I2 bars to (") were longer, varying between 2 d~:ys and 1 week. The reason for these variations of escalation and taper rates was cur initial uncertainty about the ~ssibi!ity of side effects and withdrawal effects of TRP. The chocolate bars were administered in divided dt~.~es (three times a day) in 16 of the 20 patients. In the first 4 patients, the total daily dose was divided so that the chocolate was administered four times a day. As raentioned above, the standing orders for regular (nonexpe-imental) medications remained anchanged. This was true also for p.r.n, injections. Hc-wever, the decisien to admini.~ter each p.r.n.
730
BIOL PSYCHIATRY 1990;28:728-7~;2
J. Volavka et al
injection depended on the judgment of ward personnel (within the limits prescribed by the doctor's standing order for the p.r.n, medication). Blood samples for TRP and other LNAA were ~"awn (7 AM and 9 AM) on the following 4 days: the day before the start of e~perimental treatment (baseline 1), the day treatment started (baseline 2), the day marking l full week on maximal dose, the first day of dose decrement, and 8 days after the end of experimental treatmer~t. Violent events were followed during the baseline period, t~eatment period, and for a posttreatment follow-up period of l month. We noted agitation, verbal as~caalt, assault on self, assault on others, and property assault. These events were abstracted from patients' progress notes and other docuraentation generated on the ward. In our experience, this method yields about twice as many incidents than the officially filed incident reports. Brief Psychi:~c Rating Scale (BPRS) (Guy 1976) and a simple scale for sedation (available on ~qucst) were administered weekly throughout the study, which was double-blind. Plasma amino acids other than TRP were assayed by a modified procedure of Bidlingmeyer et al (1984) and further described in detail by Richardson et al (1988). Results The effects of TRP enrichment on the ratio of TRP to the competing LNAA (see above) is shown in Table 1. (The original average values of TRP and other LNAA are available on request.) Treatments were administered at 7 AM, after the ~rst blood sample was drawtl. I r...r,sc r e s u l t s in{ii~;a{.c m a t ul~ztary c m l~mn~;nL a ~ m n ~ , , u . u y i n ~ 7 ~ s ~ d .i..tu~_TP~- i'aU~@
2 hr after TRP administration (compared with placebo), but the effect iargely dissipated at some point between 2 and 12 hr later. The 7 AM (pretreatment) ratios showed no significant difference between the placebo aad TRP groups, and no sir,niiicant e~¢ects of time (two-way ANOVA). TRP treatment had no effect on the number of violent incidents. The number of violent incidents per week in the placebo group (including verbal assault, assault on self, assault on others or on property) averaged over the 4-~eek pretreatment period and over subjects was 0.98. Duriqg treatment it was 0.97, and during posttreatment follow-up (5 weeks) it was 0.44. The analogous averages for ~he TRP group were 1.50 before, 1.53 during, and 0.30 after treatment. Similar lack of TRP effect wa~; observed for the BPRS factor "hostility-suspiciousness." Ratings on the sedation scale did not discriminate between TRP and placebo groups. \ As mentioned earlier, the p.r.n, medication order remained c~nstant throughou~ 0,~r~ study for individual patients, but the decision to actually admimster the injection was made by the staff depending on the patient's behavior. These p.r,a, injections were given to cope with patients' agitation and threate'~ing or violent beha~,ior. As no st~ff members were aware of the patients' group membership (placebo or TPV), the number of p.r.n. :~njections ~{:ministered was considered a valid outcome variable reflecting the stafi:'s judgment of 0atient behavior. The average numbers of p.r~a, injections are in Table 2. In c~rder to test the hypothesis that TRP treatment reduced the need for o.r.n, injections, we had to account for the pretreatment differences in the p.r.n, injections b~ween the placebo and the TRP group. This was accomplished by an analysis of covariance which showed a significant difference (f = 5,61; df = i,19; p = 0.03) between t~le number of p.r.n, injections administered (per day) duriiig the TRP and placebo treatments. An analogous analysis of covariance was performed for the posttreatmevt follow-up period. The difference between the two treatment g~'oups was no longer significant, indicating
Tryptophan Treatment of Aggression
BIOL PSYCHIATRY 1990;28:728-732
73 i
Table 1. TRP Ratios (Multiplied by 100)° Baseline 1 Treatment Placebo TRP
Baseline 2
Max. Dose
Dose Red.
Follow-up
7 AM.
9 AM
7 AM
9 AM
7 AM
9 AM
7 AM
9 AM
7 AM
9 AM
22 21
20 18
20 20
21
20 25
22 6~
24 23
23 39 t'
22 21
23 22
29 b
aEach ratio was computed by dividing TRP value by the sum of the values of the other LNAA (leucine, isoleucine, valine, phenylalanine, and tyrosine). Each ratio in Table i represents an average for 10 patients. TRP (or placebo) was administered immediately after the 7AM blood sample was d~wn. ~" bp < 0.05 (TRP vs placebo).
that the reduction of p.r.n, injections associated with TRP treatment failed to persist beyond the treatment period. As mentioned above, the neuroleptic medication for 2 of the 20 patients changed during the study. We have therefore repeated the analyses for p.r.n, injections after the exclusion of these 2 patients. The reduction of p.r.n, injections associated with TRP treatment remained statistically significant, and remained significant during the follow-up period. Finally, we should peint out that the treatment had no apparent side effects.
Discussion Our results failed to support the principal hypothesis that TRP would reduce aggressive behavior; the general decrease of assaultive behaviors under both treatment conditions (TRP and placebo) was probabb attributable to environmental effects. However, TRP treatment has apparently reduced the need for p.r.n, injectable medication. It is possible that this reduction reflected the therapeutic effect of TRP on aggressive behavior. Furthermore, one can speculate that the use of p.r.n, medication may have obs:ured the differences in aggressive behavior between the placebo and TRP treatment groups. There is a possibility that the ward staff wa3 very ju0Zcious in the use of p.r.n, inj~.ctions, administering ~hem as soon as the patient showed any sign of impending violenct:. If this happened, the p.r.n, preventative use would mask the difference between TRP and placebo treatment effects, on overt violence. The generalizability of our results is limited by the small number of subjec~:~ and the sho~ duration of the trial; the latter deficiency is par',icularly damaging i~ v~,w of the fact that violent incidents were rare. la conclusion, our study ~ra-.,'-:led indirect support for the TRP beneficial effects on _,ggressive behavior in psychiatric patients. Future studies using higher doses of TRP may show mo:'c pronounced effects. However, sach studies must wait until the re,'e~t reports (Blevins et al 1989) of seriou~ adverse effects of TRP (eosinophil'~a-myalgia syndrome) can be clarified.
Table 2. ?.:':_-'=~."Number of p.r.n, lnjectio,t~ per Da) Treatment group Placebo TRP
Baseline
~x~rimcnta~ treatment
Follow-up
0.107 0. ! 48
0.077 0.0.57"
0.061 0.075
Op = 0.03 (TRP vs placebo, ANCOVA).
732
BIOLPSYCHIATRY 1990;28:728-732
J. Valavka et al
Note Added i~ Print Recent evidence strongly suggests that the eosinophilia-myalgia syndrome associated with the ingestion of TRP was caused by a co~taminant or an alteration in a subset of TRP manufactured by a single company (Slutsker et al 1990). Nestle Research Laboratoriesprovided the TRP and placebo chocolates. Dr. Peter Leathwood (Nestle Resea.,eh) was vet:},he!p_~d and made insightful comments. Dr. Richard J. Wurtman (Massachusetts Institute of Technology) gave us hnportant advice in the planning stage of this study.
References Bidlingmeyer BA, Cohen SA, ]'arvin TL (1984): Rapid analysis ot: amino acids using pre-column derivitizafion. J Chromatogr 336:93-104. Bioulac B, Benezich M, Rena~ld B, Noel B, Roche D (1980): Serotonergic function in the 47,XYY syndrome. Biol Psychiatly 15:917-923. Blevins WL, Hertzman P, Ting M, et al t!989) Eo~inophilia-myalgia syndrome--New Mexico. JAM4 262:3116. Bre,achey L, Branchey M, Shaw S, Lieber CS (1984): Depression, suicide, and aggression in alcoholics and their relat2onship to plasma amino acids. Psychiatry Res 12:219-226. Brown GL, Goodwin I-L=K,Ballenger JC, Goyer PF, Major LF (1979): Aggression in humans correlate~ with cerebrospinal fluid amine metabolites. Psychiatry Res 1:131-139. Coccaro EF, Siever i.J, Klar HM, et al (1989): Serotonergic studies in patients with affective and person~,iity disorders. Awch Gen Psychiatry 46:587-599. Fe~stmm JD, Wurtrnan R2: (1977.): Brain serotorfin content: Physiological regulation by plasma neutra'l amino acids. Science 178:414--416. Guy W (1976): ECDEU A.:ses;ment Manual for Psychopharmacoiogy. Rockville, MD: National Institute of Mental Health. Linnoi!a M~ Virk!canen M, Scheinin M, Nuutila A, Rimon R, Goodwin FK (1983): Low cerebrospinal fluid 5-hydroxyindoleacetic acid concentration differentiates impulsive from nonimpu!aive violent behavior. Life Sci 33:2609-2614, Moraad C, Young SN, Ervin FR (1983): Clinical response of aggressive schizophrenics to o,al tryptophan, t~iol Psychiatr,y 18:575-578. O~Neil M, Page N, ,~dkins WN (1986): Tryptophan-trazodone treatment of aggressive behavior. Lancet ii:859-860. Richardson MA, Suckow R, Whittaker R, et al (1988): Phenylalanine, phenylethylamine and tardive dyskinesia in psychiatric patients .;a trace amines: Comparative and clinical neurobiology. Boulton AA, Jurio AV, Downer RGM (eds). Clifton, NJ: Humana Press, pp. 409--422. Slutsker L, Hoesly FC, Miller L, Williams LP, Watson JC (1990): Eosinophiiia-myalgia syndrome associated with exposure to tryptophan from a single manufacturer. JAMA 264:213-217. Virkkunen M, De Jong J, Bartko J, Goodwin FK, Linnoila M (1989): ]~e!ationsh;p of psychobiological variables to recidivism in violent offenders and impulsive fire setters. Arch Gen Psychiatry 46:600-603.
