P&T News: April 1995, Vol. 15, No. 10

Neuroleptic Malignant Syndrome

Daniel A. Smith, Pharm. D.
Peer Review Status: Internally Peer Reviewed by Peggy C. Nopoulos, M.D., Assistant Professor, Department of Psychiatry


Neuroleptic malignant syndrome (NMS) is a rare, but potentially, life-threatening adverse effect associated with antipsychotic drugs and other drugs that affect dopaminergic neurotransmission. Since NMS occurs very infrequently, it is extremely difficult to characterize it objectively under controlled conditions. This review will attempt to differentiate and assess the clinical significance of well established and less well established features of NMS.

Etiology and Incidence
Neuroleptic malignant syndrome has been associated with all dopamine blocking drugs, including the pharmacologically distinct antipsychotics clozapine and risperidone.1-3 Other dopamine antagonists such as the antiemetics metoclopramide, prochlorperazine, promethazine, and droperidol also cause NMS. l Abrupt withdrawal of dopaminergic drugs may also produced an NMS-like condition.

Neuroleptic malignant syndrome is very rare.5-7 Estimates of the frequency of NMS in prospective studies range from 0.07% to 2.2%.8-10 Analysis of NMS across studies suggests a frequency of approximately 0.2%.1 This range in frequencies may be due to differences in diagnostic criteria for NMS from center to center.4, 7

Clinical Features and Laboratory Abnormalities
Most patients with NMS exhibit muscle rigidity, hyperpyrexia, altered consciousness, and autonomic instability. The rigidity seen in NMS is often referred to as Lead pipe rigidity because of the extreme nature of the reaction. In other cases akinesia, dyskinesia, waxy flexibility, and cogwheeling may occur instead of the classic rigidity.12 The fever seen in NMS usually exceeds 38 degrees C and sometimes exceeds 41 degrees C.1,12 Mental status changes associated with NMS may include stupor, coma, delirium, or catatonia.1 Autonomic instability associated with NMS usually includes tachycardia and alterations in blood pressure. Respiratory distress may accompany these signs.1 Many atypical cases have also been reported, often lacking one or more of these four classic signs.1, 4, 12, 13 Variability in severity and diagnostic inconsistency among centers may account for this.

Extreme rigidity leading to muscle necrosis often contributes to elevations in creatine kinase (CK), lactic dehydrogenase (LDH), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) seen in NMS.1 Creatine kinase elevations, sometimes 100 to 200 times normal, are seen in most, but not all patients with NMS.12 Rhabdomyolysis and myoglobinuria may result and lead to renal failure. White blood cell count elevations between 10,000 to 20,000/mm3 are also common.14

Clinical Course
Reviews of case reports suggest that most cases of NMS occur within the first one to two weeks following antipsychotic initiation or dose increase, although NMS may occur at any time during treatment.1, 14, 15

Neuroleptic malignant syndrome is a self-limiting condition once the offending agent has been discontinued. It will resolve with supportive care in one to two weeks unless the patient has been receiving a depot antipsychotic in which case it may take up to one month for NMS to resolve.l4, 16

Complications
Untreated or unrecognized cases of NMS may result in numerous complications (see Table 1).4, 13, 14 Despite this, a review of previously published case reports identified long-term sequelae in only four of 120 cases.15 This perspective is confirmed by two long-term follow-up case series involving 16 patients.l7, l8 No significant longterm sequelae were identified in these series.

Table 1. Complications of NMS 4, 13, 14

Medical Complications

Causes of Death

Aspiration pneumonia
Renal failure
Cardiac arrest
Seizures
Sepsis
Pulmonary embolism
Pulmonary edema
Rhabdomyolysis
Respiratory failure
Cardiorespiratory arrest
Pneumonia
Pulmonary embolism
Sepsis
Hepatorenal failure
Disseminated intravascular coagulation

 

Estimates of mortality resulting from NMS typically range from 10 to 20%.19 Most studies of this issue have been retrospective reviews of previously published case reports. In a review by Shalev et al, 202 cases of NMS meeting operational diagnostic criteria were identified between 1959 and 1987.19 The mortality among these patients was 18.8%. When stratified by date of publication, the mortality rates were 27.7% before 1980, 22.6% from 1980 to 1983, and 11.6% from 1984 to 1987, a statistically significant trend. Predictors of mortality noted in this report are myoglobinuria and rhabdomyolysis. Mortality was 47 % and 56 % in these patients, respectively.

No cases of mortality were reported from pooled data from five reports of 54 prospectively-evaluated NMS cases.5, 8, 10, 18, 20 These data suggest that retrospective reviews of case reports overestimate true NMS mortality.

Differential Diagnosis
Neuroleptic malignant syndrome is a diagnosis of exclusion.1, 12, 13 Numerous other disorders that may mimic NMS in part or whole must be first ruled out (Table 2). Among these diagnoses, acute lethal catatonia is of particular interest because it is indistinguishable from NMS.1, 12, 13 Before antipsychotics became available, acute lethal catatonia was diagnosed more frequently than it is today.13 Some have suggested that cases of acute lethal catatonia have been misdiagnosed as NMS since antipsychotics have become available. Others maintain that antipsychotics are the primary treatment for acute lethal catatonia,21 although this is also disputed. 1,12

The serotonin-syndrome is another important diagnosis to consider in suspected cases of NMS. The serotoninsyndrome is a potentially fatal result from the combination of a monoamine oxidase inhibitor with a serotonin reuptake inhibitor (SSRI).22 Features of the serotonin-syndrome are similar to NMS and include excitement, diaphoresis, rigidity, hyperthermia, tachycardia, and hypertension. Depending on the circumstances of the case involved, the necessary washout period between use of an monoamine oxidase inhibitor and an SSRI may range from one to more than five weeks. It is during this washout period that a misdiagnosis of NMS would be most likely, particularly when a patient is seeing more than one physician and complete records are not available.

Table 2. Differential Diagnosis of NMS(1, 12, 13)

  • Infections
  • Tumors
  • Trauma
  • Seizures
  • Acute Lethal Catatonia
  • Malignant Hyperthermia
  • Serotonin Syndrome
  • Anticholinergic Delirium
  • Severe Parkinsonism
  • Heat Stroke
  • Dystonic Reactions
Various sets of operational criteria to aid in the diagnosis of NMS have been published.11 Diagnostic criteria for NMS are useful primarily as a research tool to help assure consistent diagnoses of NMS across studies, a recurrent problem in much of the current literature. Some may find these criteria useful to clarify or solidify the diagnosis of NMS.

Risk Factors
Assessing the clinical importance of the many proposed risk factors for NMS is difficult because of the conflicting and uncontrolled nature of the NMS literature. In many cases, it is not apparent whether the risk factors are truly related to NMS or simply a confounding variable resulting from some other aspect of the case. Most data suggest the risk of NMS may be minimized through the use of conservative dosing strategies involving a single antipsychotic.

Antipsychotic dose and psychomotor agitation are intimately linked as NMS risk factors. Patients with NMS tend to receive higher antipsychotic doses, a faster antipsychotic titration rate, and a greater number of intramuscular antipsychotic injections compared to controls.6 Exhaustion, dehydration, and psychomotor agitation have also been linked to NMS. 1, 13 In one case series, 18 of 24 patients were agitated before developing NMS.20 In a case-control study, psychomotor agitation was more common in the NMS patients compared to controls.6 Since agitated patients are more likely to receive a large antipsychotic dose, it is impossible to differentiate the risk associated with each of these factors. It may be that either factor alone, or in combination, predisposes one to NMS. In spite of these data, NMS may occur at low doses in non-agitated patients.1

In two case control studies, antipsychotic potency was not a risk factor for NMS.5, 6 Some reviews have suggested that high potency antipsychotics, and haloperidol in particular, may be risk factors.l, l3, 14 In one review, haloperidol accounted for 57% of published cases, while chlorpromazine accounted for 24%.14 These results may be partially explained by the fact that haloperidol is one of the most frequently prescribed antipsychotics. In addition, rapid titration to high dose is easier with high potency antipsychotics because of their lack of cardiovascular adverse effects relative to low potency antipsychotics.

One case control study has found a link between NMS and the depot antipsychotic fluphenazine decanoate.5 However, every patient in this study who received fluphenazine decanoate was also receiving supplementary oral antipsychotics. In addition, a case control study6 and a review23 of the depot antipsychotic literature have failed to confirm this relationship.

Although greater than 5096 of reported NMS cases involve concomitant psychotropic drugs, it is not clear that these drugs increase the risk of NMS.1s5-6 The combination of lithium with an antipsychotic has been widely associated with NMS,1, 13 but two case control studies have failed to confirm this association.5, 6 Keck et al have suggested that the use of lithium is simply an indirect reflection of psychomotor agitation resulting from acute mania.6

Some reviews suggest that NMS is twice as common in males.13 Others suggest that antipsychotics are used differently between genders.1 Men may be more likely to receive a higher antipsychotic dose if they are perceived as more threatening by their caregivers. Neuroleptic malignant syndrome has been reported in all age groups and does not differentiate between NMS patients and controls.l, 6 There are two case reports that suggest that predisposition to NMS may include a genetic component.24, 25 Antipsychotics should be administered cautiously in patients with a family history of NMS.

High environmental temperatures have been suggested as a risk factor for NMS.26 However, there is little objective evidence that NMS is caused by high environmental temperatures since NMS has been reported in all extremes of temperature and climate.1

Despite the remarkable clinical similarity between NMS and malignant hyperthermia, there is little evidence that these two disorders are related or that patients with NMS are also at risk for malignant hyperthermia. In a review of 48 patients with NMS who also received electroconvulsive therapy (which includes anesthesia), no cases of malignant hyperthermia occurred.27 In studies using the halothane-caffeine contracture tests, an in vitro test used to identify patients who are susceptible to malignant hyperthermia, patients with a history of NMS were considered not susceptible to malignant hyperthermia.28 Neuroleptic malignant syndrome may occur in any patient exposed to an antipsychotic, regardless of diagnosis.1 In one case series, only one of 20 NMS patients was schizophrenic, and 14 of them had an affective illness.16 In another series, 11 of 12 cases of NMS were in schizophrenics.5 Patients with a history of NMS are at increased risk for NMS. This issue is discussed in detail below under "Antipsychotic Rechallenge. "

Pathogenesis
There are numerous theories that attempt to explain the mechanism by which NMS occurs.1, 2, 4. 12, 13 None of these theories satisfactorily explains why NMS only occurs in some patients or why NMS does not always recur even if the patient is re-exposed to the same antipsychotic that originally produced the NMS.29 Despite this, most theories attribute the majority of the clinical features of NMS to dopamine blockade since dopamine antagonists cause NMS and dopamine agonists are used to treat NMS.1, 2, 4, 12, 13

The tremor and rigidity seen in NMS have been linked to nigrostriatal dopamine blockade, an extension of the parkinsonian adverse effects seen at therapeutic doses of antipsychotics.12, 13 Extreme rigidity may contribute to hyperthermia, muscle damage, elevated creatine kinase, and rhabdomyolysis. Peripherally, antipsychotics also affect intracellular calcium transport resulting in an increased calcium concentration and altered muscle fiber contractility. This effect is reversed by dantrolene.13

Central dopamine blockade in the hypothalamus may result in impaired temperature regulation by altering the "set points of body temperature maintained in the hypothalamus. In addition, the heat dissipating mechanisms affected by the hypothalamus through the autonomic nervous system (e.g., shivering, sweating, and peripheral vasoconstriction or vasodilatation) may be disrupted.12

Treatment
Perhaps the most baffling mystery of NMS is that of rational treatment. Since NMS occurs so rarely, it is impossible to perform a study under ideal circumstances. Less objective forms of research are used instead. A common method used to study NMS is an analysis of previously published case reports referred to as a case report analysis.14-16, 19, 27, 30, 31 In a case report analysis, the authors attempt to identify a comprehensive collection of NMS cases These cases of NMS are analyzed and classified in an attempt to characterize various aspects of NMS.16, 19, 27, 31

Case report analyses are subject to biases as a result of the research design. A significant problem is the lack of diagnostic consistency among reports. One study used operational diagnostic criteria to identify NMS cases,l9 while another study included cases if they were Consistent with NMS."16 The criteria for diagnosing NMS were not stated in other reports.27, 3l

Another difficulty associated with evaluating case report analyses is the lack of randomly allocated treatments or specific treatment protocols. A number of limitations of samples derived from case reports have been identified:16 1) more severe cases may be more likely to receive specific treatments over supportive care; 2) cases with a good outcome to specific treatment may be reported more frequently than cases with a poor outcome; 3) more severe cases may be reported more frequently; 4) "classic" cases or, conversely, more atypical cases may be reported more frequently; 5) year of treatment may affect outcome since mortality has declined over the past decade; and 6) treatment sites vary from case to case. Because of all of these variables, the study populations in case report analyses can hardly be considered "normal. "

Despite each of these caveats, many clinicians will choose to use a specific treatment such as bromocriptine or dantrolene because drugs have a low risk of significant toxicity compared to the NMS, and they may be beneficial. Furthermore, the possibility of an unfavorable outcome and legal liability if treatment is not provided may influence many clinicians to use a specific treatment.

Supportive Care. Discontinuation of the causative agent is the single most important step in treating NMS.1 While success has been reported with continued treatment of psychosis with an antipsychotic in some cases,32 most other reports suggest this may be associated with a poorer outcome.8, 27, 33 Other support should be administered as needed. Fluid replacement, fever reduction, and support of cardiac, respiratory, and renal function are commonly necessary.

Pharmacologic Treatments.There are two common pharmacologic approaches used to treat NMS. Dopaminergic drugs such as bromocriptine or amantadine are used to counteract the dopamine blockade that presumably leads to NMS. Dantrolene is used to treat malignant hyperthermia, a condition that bears striking clinical resemblance to NMS. As a skeletal muscle relaxant, dantrolene acts to decrease rigidity, and possibly fever.

Rosenberg and Green compared supportive care, bromocriptine therapy, and dantrolene therapy in a case report analysis of 67 NMS patients identified between 1977 and 1987.16 Cases were included if they specified age, sex, and diagnosis; signs and symptoms of NMS; mode of treatment; and either response time (time until the patient's symptoms improved) or resolution time (time until all symptoms disappeared). Sixty-five of the cases were attributed to an antipsychotic or metoclopramide. The remaining two cases were related to abrupt dopamine agonist withdrawal. The mean response time was 6.8 days for supportive care alone, 1.0 day for bromocriptine, and 1.7 days for dantrolene (p less than 0.0005 for both treatments compared to supportive care alone). The mean resolution time was 15.8 days for supportive care alone, 9.9 days for bromocriptine, and 9.0 days for dantrolene (p less than 0.1 , a nonsignificant trend for each of the specific treatments compared to supportive care alone).

In a review published by Sakkas et al, a total of 734 cases of NMS were identified who were treated with supportive care alone or with a specific treatment or combination of specific treatments of bromocriptine, dantrolene, or amantadine (Table 3).31 The authors of this review attempted to "identify every published case on NMS" for inclusion in their case report analysis. The diagnostic criteria used to identify cases of NMS were not reported. The authors failed to find a significant benefit from the combination of specific treatments compared to any single specific treatment alone. Among the three outcomes measures, recurrence of NMS following discontinuation of a specific treatment is the most useful because these are the only cases that include an objective measure of the specific treatments' effects (i.e., recurrence of fever, rigidity, etc.). The clinician's opinion as an outcomes measure is less useful because of the uncontrolled biases in each original case report. The comparison of mortality rate between supportive care alone and a specific treatment is complicated because this study included every known case of NMS. Many of the historical cases were almost certainly treated differently than modem cases. In addition, since supportive care has probably improved over the years since NMS was first reported, inclusion of older cases in the analysis would tend to make supportive care alone appear less effective than specific treatments that have only been used in recent years.

Table 3. Results from Sakkas et al, 1993(31)

Treatment

Improvement Rate (Clinician's Opinion)

Relapse Rate Following Removal of Specific Treatment

Mortality Rate

Amantadine Alone

Bromocriptine Alone

Dantrolene Alone

Supportive Care Alone

63% (n=19)

94% (n=54)

79% (n=53)

Not Applicable

29% (n=7)

18% (n=17)

6% (n=35)

Not Applicable

6% (n=17)*

8% (n=51)**

9% (n=58)**

21% (n=438)

_________
*p = 0.23 compared to supportive care alone
**p less than 0.05 compared to supportive care alone

In the Shalev et al case report analysis, mortality was no different between supportive care (7/52, 13.5%) and any specific treatment of bromocriptine, dantrolene, or amantadine, alone or in combination (4/43, 9.3%).}9 These data suggest that specific treatments have no effect on mortality. Unfortunately, the small number of actual fatalities and the limitations of the case report analysis study design make any firm conclusions impossible.

There is one study that contradicts the notion that specific treatments for NMS are more effective than supportive care alone.34 In this analysis of 20 consecutive referrals, the decision to use supportive care alone was based on the prescriber's familiarity with the specific treatment literature. There was no randomization, blinding, or treatment protocol. The mean duration of NMS symptoms in the supportive care only group (n = 12) was 6.8 days compared to 9.9 days for the eight patients who received bromocriptine and/or dantrolene (p less than 0.05). There was a trend for more severe medical illness among the patients who received a specific treatment. These results led the authors to suggest that specific treatments may actually worsen outcome. However, in another interpretation, one might surmise that the patients who received bromocriptine and/or dantrolene received them specifically because they were more ill or because they had a more severe case of NMS.

Electroconvulsive Therapy (ECT). Two case report analyses studying the efficacy of ECT for NMS have been published.27, 35 ECT has been used to treat NMS because of the clinical similarities between NMS and acute lethal catatonia and because of the anecdotal reports of the dramatic effect of ECT for acute lethal catatonia. l In the Mann et al case report analysis, ECT was considered effective in 20 of 27 cases and partially effective in three cases. Two patients in this series developed serious cardiovascular complications during ECT, cardiac35 arrest in one and ventricular fibrillation in the other. It may be that during ECT, NMS patients comprise a minority who are predisposed to serious cardiovascular complications, a rare occurrence, typically. Conversely, it is possible that these patients were reported primarily because of the severity of their complications, a bias of case report literature. In the Davis et al case report analysis, 24 of 29 cases improved with ECT.27 Three of these patients improved despite continued antipsychotic treatment. Of the five nonresponders, all continued to receive antipsychotics. Of interest, the patients' psychiatric illness frequently improved along with their NMS symptoms during ECT.

Treatment Recommendations. Bromocriptine and dantrolene are the two most widely studied specific treatments for NMS (Table 4). Choice of a specific drug should be based on the patient. Bromocriptine is an inexpensive oral drug, which by nature of its dopaminergic pharmacology, directly opposes the effect of antipsychotics. However, this same effect may also worsen a patient's psychosis. Dantrolene, while inexpensive orally, is extremely expensive as an injection. Dantrolene injection should be reserved for patients who cannot receive oral drugs, and they should be switched to oral dantrolene as soon as is reasonable. The most serious adverse effect of dantrolene is severe hepatotoxicity that occurs following prolonged exposure to high doses; this is rarely a problem in NMS. There is no evidence that there is any difference in efficacy between oral bromocriptine and oral or intramuscular dantrolene.l6, 31 Likewise, there is no evidence that combining two or more specific treatments improves response.31 Combinations are only recommended if a single agent has failed.

Table 4. Comparison of Bromocriptine and Dantrolene

Drug

Mechanism of Action

Dosage

Average Wholesale Price*

Adverse Effects

Bromocriptine

Dopamine Agonist

7.5 to 40 mg/day orally in divided doses

$1.44/2.5 mg tablet
$2.19/5 mg capsule

Psychosis, hypotension, nausea

Oral Dantrolene

Muscle Relaxant

4 to 8 mg/kg/day in divided doses

$0.67/25 mg tablet
$1.44/50 mg tablet
$1.25/100 mg tablet

Hepatotoxicity

Intravenous Dantrolene

Muscle Relaxant

2 to 3 mg/kg/day initially; maximum 10 mg/kg/day

$55.21/20 mg vial

Hepatoxicity

____________
* Acquisition cost based on Average Wholesale Price in Drug Topics Red Book 1994. Price to the patient may vary depending on individual pharmacy pricing practices.

Caroff and Mann suggest ECT as a treatment alternative in cases where differentiation between acute lethal catatonia and NMS is difficult, in cases in which pharmacologic treatment has failed, and in resolved cases of NMS in which antipsychotic rechallenge is inadvisable despite continued psychosis.1

Antipsychotic Rechallenge
Data from reviews of previously published case reports suggest that patients with a history of NMS have a 30 to 50% risk of recurrence of NMS following antipsychotic rechallenge.30, 36-38 In one review, the length of time between the resolution of the symptoms of NMS and antipsychotic rechallenge was related to recurrence.38 The NMS recurrence rate was 63 % (7/11 patients) if the antipsychotic was reintroduced within 5 days of the resolution of the initial episode of NMS. The recurrence rate dropped to 30% (10/33 patients) if more than five days elapsed. Of interest, four of seven cases were successfully rechallenged using the same antipsychotic that initially caused the NMS. Likewise, Susman and Addonizio reported a higher recurrence rate if rechallenge occurred before complete resolution of NMS symptoms.37 Caroff and Mann found that the NMS recurrence rate dropped from 30% to 15% if a low potency antipsychotic was used to rechallenge.30 However, this may be due to the use of lower equivalent doses with low potency antipsychotics compared to high potency antipsychotics because of cardiovascular side effects associated with low potency antipsychotics.

In one series, 13 of 15 patients were successfully rechallenged with an antipsychotic.29 However, five patients (3396) experienced recurrence on the first rechallenge. Rechallenge was successful in every case in which two weeks elapsed following the episode of NMS and rechallenge. Likewise, successful rechallenge was more likely if lower dose was used at rechallenge. However, one patient was successfully rechallenged on the same regimen initially associated the episode of NMS. The effect of antipsychotic potency on NMS recurrence, could not be evaluated due to the small patient population.

Rechallenge Recommendations. Many NMS patients will continue to require an antipsychotic. The following recommendations may help prevent recurrence of NMS.29, 30, 36-38

1. Reassess the indication for the antipsychotic.

2. Wait two weeks after resolution of NMS before rechallenge.

3. Rechallenge with a different chemical class antipsychotic and/or a different potency.

4. Use the lowest dose possible. Titrate slowly.

5. Consider alternative treatments such as such as benzodiazepines for agitation. Benzodiazepines may be effective alone for agitation, or, if given in combination with an antipsychotic, they will allow for a lower antipsychotic dose.

6. Avoid the long-acting depot antipsychotics haloperidol decanoate and fluphenazine decanoate since patients with a history of NMS are at higher risk for NMS in the future and because an episode of NMS may last up to one month when associated with these dosage forms.

Summary
Neuroleptic malignant syndrome is a rare but serious adverse effect associated with dopamine blocking drugs. The classic features of NMS include extreme rigidity, fever, autonomic instability and mental status changes. The most important risk factors associated with NMS are a high antipsychotic dose, rapid dose titration and psychomotor agitation. Early recognition and treatment of NMS will minimize complications. Supportive care combined with immediate discontinuation of the causative agent is the primary treatment of NMS. In addition, specific drug treatments such as bromocriptine or dantrolene are frequently used. Whenever possible, it is important to allow a period of two weeks after an episode of NMS has completely resolved before reinitiating antipsychotic treatment. Use of a different antipsychotic may minimize the risk of recurrence of NMS.

References

1. Med Clin North Am 1993;77:185-202.
2. Pharmacotherapy 1993;13:510-4.
3. Lancet 1994;344:1228.
4. Gen Pharmacol 1990;21:367-86.
5. Am J Psychiatry 1990; 147:1149-55.
6. Arch Gen Psychiatry 1989;46:914-8.
7. Psychiatry Res 1992;44:251-6.
8. Am J Psychiatry 1988; 145:517-8.
9. Br J Psychiatry 1992;161:254-7.
10.J Clin Psychiatry 1989;50:352-5.
11. J Clin Psychiatry 1992;53:56-62.
12.Med Clin North Am 1993;77:477-92.
13.ProgNeurobiol 1991;36:425-36.
14.Biol Psychiatry 1987;22:1004-1020.
15.Acta Psychiatr Scand 1986;73:337-47.
16.ArchInternMed 1989;149:1927-31.
17.J Clin Psychiatry 1988;49:154-6.
18.Hosp Community Psychiatry 1991;42:197-9.
19.J Clin Psychiatry 1989;50:18-25.
20.Am J Psychiatry 1989;146:717-25.
21.Am J Psychiatry 1989; 146:324-8.
22.J Clin Psychopharmacol 1990;10:48-50.
23.J Clin Psychiatry 1992;53:426-33.
24. Pharmacopsychiatry 1987;20: 168-79.
25.Br J Psychiatry 1991; 158: 850-3.
26.Am J Psychiatry 1988; 145: 110-1.
27. Convulsive 7her 1991;7: 111-20.
28.Aaa Anaesthesiol Scand 1989;33:676-80.
29.Br J Psychiatry 1991; 159:709-12.
30. Psychopharmacol Bull 1988;24:25-9.
31. Psychiatr Ann 1991;21: 157-64.
32.Br J Psychiatry 1989;154: 1024.
33.Am J Psychiatry 1991;148:880-2.
34.J Clin Psychiatry 1989;50:295-8.
35. Convulsive Therapy 1990;6:239-7.
36.J Clin Psychopharmacol 1986;6:257-73.
37.J Nerv Ment Dis 1988; 176:23440.
38.Drug Intell Clin Pharm 1988;22:475-9.

Next Page | Title Page