Validation of a Clinical Global Impression Scale for Aggression (CGI-A) in a sample of 558 psychiatric patients

Article Outline

• Abstract
• 1. Introduction
  • 1.1. Aims of the study
• 2. Methods
  • 2.1. Context and sample
  • 2.2. Assessments and measures
  • 2.3. Data analysis
• 3. Results
  • 3.1. PANSS-EC and CGI-A scores
  • 3.2. Validity of CGI-A
• 4. Discussion
  • 4.1. Key findings
  • 4.2. Strengths
  • 4.3. Limitations
• 5. Conclusion
• Role of funding source
• Contributors
• Conflict of interest
• Acknowledgements
• References
• Copyright

Abstract 

Objective

Clinical management of aggression depends on the availability of easily administrable measurements allowing reliable evaluation. The present study's aim is to validate a Clinical Global Impression-Severity of Aggression scale (CGI-A).

Method

558 inpatients with psychiatric disorders and an agitated–aggressive syndrome at baseline were continuously assessed over 5 days using CGI-A and the Positive and Negative Syndrome Scale-Excited Component (PANSS-EC). Equipercentile linking, correlation analyses and linear regression were applied.

Results

Relationship between CGI-A and PANSS-EC total score was found to be linear. On a 5-level CGI-A scale, values of 1 to 5 points were found to correspond to PANSS-EC scores of 12.2, 16.7, 21.3, 25.8, and 30.4, respectively (average increase: 4.6). All findings remained stable when only data from patients with schizophrenia spectrum disorders were analyzed.

Conclusions

The CGI-A is proposed as a quickly administrable scale for the assessment of patients' aggressiveness.

Keywords: Aggression, Psychosis, Equipercentile linking, Validity

Back to Article Outline

1. Introduction 

Agitation and aggression due to a primary psychiatric disturbance are common phenomena in psychiatric patients (Allen et al., 2001). These syndromes constitute a serious problem and pose danger to health care professionals, patients, and other persons involved. They require immediate non-pharmacological and pharmacological interventions, sometimes including involuntary treatment, the use of restraints, or seclusion. Expert consensus guidelines state that quick assessment of the situation plays a critical role in the treatment of acute aggression (Allen et al., 2005).

As psychiatric interviews and patient self-ratings may cause exacerbation of agitation and aggression, there is a need for measurement tools which can be administered easily and are based on clinicians' observation only. These tools may enable clinicians and researchers to safely assess severely aggressive or agitated patients.

Such an instrument, based on clinicians' observation only, is the Positive and Negative Syndrome Scale-Excited Component (PANSS-EC) for the assessment of agitation (Kay et al., 1987). The PANSS-EC has been applied in trials evaluating acute agitation and aggression in psychiatric patients (Meehan et al., 2002, Baker et al., 2003, Barzman et al., 2006, Currier et al., 2006, Pascual et al., 2006, Pascual et al., 2007, San et al., 2006, Zhong et al., 2007), especially in schizophrenia spectrum disorders (Breier et al., 2002, Baker et al., 2003, Wright et al., 2003, Turczynski et al., 2004, San et al., 2006), in bipolar mania (Baker et al., 2003, San et al., 2006), in adolescents with bipolar disorder and disruptive behaviour disorders (Barzman et al., 2006), and in dementia (Meehan et al., 2002, Zhong et al., 2007).

Recently, the Clinical Global Impressions (CGI) score was used to measure aggression in patients with psychiatric disorders by means of the CGI-Aggression (CGI-A; Frazier et al., 1999, Sharif et al., 2000, Sival et al., 2004, MacMillan et al., 2006). The CGI-A was derived from the CGI-Severity of Illness (CGI-S) scale, which was designed as an overall measure of illness severity in psychiatric disorders (Guy, 1976), and has been adapted to measure severity of illness in specific disorders, e.g. schizophrenia (Haro et al., 2003, Haro et al., 2007), major depression (Salzmann and Robin, 1995), bipolar disorder (Spearing et al., 1997), or in various specific syndromes [e.g. suicidality (Lindenmayer et al., 2003)]. The CGI-A is particularly promising for the assessment of aggression because it is quickly administered and also based on clinicians' observation only. However, the CGI-A has never been validated, and its relation to the PANSS-EC is unclear.

1.1. Aims of the study 

Validation of the Clinical Global Impression Scale for Aggression (CGI-A) in patients with different psychiatric disorders by comparison of CGI-A ratings with ratings of the Positive and Negative Syndrome Scale-Excited Component (PANSS-EC) in an unselected sample of 558 patients that were acutely admitted because of an agitated–aggressive syndrome.

Back to Article Outline

2. Methods 

2.1. Context and sample 

Data were acquired in the IMPULSE study (Immediate Aggression Control, Antipsychotics and Tranquillization: An Observational Study), a clinical post-marketing trial supported by Eli Lilly Deutschland GmbH. The study's aim was to assess pharmacological treatment in an unselected sample with psychiatric emergencies related to different psychiatric disorders. 558 patients were included between July 2003 and October 2004 in 102 acute psychiatric inpatient units. The study was conducted in accordance with all local and national regulations.

Patients meeting the following criteria were included: newly admitted patients presenting with agitation with or without aggression, requiring antipsychotic treatment, and age ≥18 years. No further inclusion or exclusion criteria were pre-defined to preserve the naturalistic character of the trial. Because this was an observational study, German law required no patient consent at the time the study was conducted. This recruitment strategy assured the evaluation of an unselected sample.

2.2. Assessments and measures 

Assessments were carried out at baseline and at days 1, 2, 3, 4, and 5. Agitation was pre-defined according to Lindenmayer (2000) as a state of poorly organized and aimless psychomotor activity including symptoms such as motor restlessness, heightened responsivity to external or internal stimuli, irritability, and/or decreased sleep. Aggression included intimidating behavior, aggression to property, demeaning or hostile verbal behavior, and aggression to persons (Lindenmayer, 2000).

Diagnoses were documented according to the International Classification of Diseases, 10th revision (WHO, 1992). At all visits, the following instruments were completed: The Excited Component (PANSS-EC; Kay and Sevy, 1990, Breier et al., 2002, Meehan et al., 2002) of the Positive and Negative Syndrome Scale (PANSS; Kay et al., 1987), comprising the items “poor impulse control”, “tension”, “hostility”, “uncooperativeness”, and “excitement” and rated on a seven point Likert scale from “absent” to “extremely severe” (score range from 5 to 35 points). To ensure the validity of PANSS-EC ratings, definitions of each PANSS-EC item were provided on the same sheet where the respective ratings had to be documented. The Clinical Global Impression-Severity of Illness Scale (CGI-S) (Guy, 1976) was originally designed as an overall measure of severity of a patient's illness. In the current study, the original 7-point CGI-S (categories: 1=normal, not at all ill; 2=borderline ill; 3=mildly ill; 4=moderately ill; 5=markedly ill; 6=severely ill; 7=extremely ill) was adapted to account for the global impression of aggression severity (CGI-A). To increase rating simplicity, a 5-point Likert scale (categories: 1=no, 2=slight, 3=moderate, 4=severe, and 5=aggressive behavior present) was used.

2.3. Data analysis 

Visits from all assessments were pooled to include a broad span of rating values, to avoid floor and ceiling effects, and to obtain more generalizable results applicable to patients with varying severity of aggression. For all 558 patients, 3310 measurements for PANSS-EC and CGI-A were available.

Because the PANSS-EC was derived from the PANSS, which itself was only validated for the assessment of patients with schizophrenia spectrum disorders (Kay et al., 1987), all analyses were repeated for patients with these diagnoses [schizophrenia (n=232; mainly of the paranoid subtype (n=202)), schizotypal disorder (n=2), persistent delusional disorder (n=9), acute and transient psychotic disorders (n=18), and schizoaffective disorder (n=69)]. For this second analysis, 330 (59.1%) patients with a total of 1964 measurements remained.

All analyses were carried out using SAS version 9.1 (SAS Institute Inc., Cary, NY, USA). Correlation and regression analyses (linear mixed models) as well as equipercentile linking of the CGI-A and the PANSS-EC items were conducted to examine the scale's diagnostic validity. Equipercentile linking has recently been used to link the Brief Psychiatric Rating Scale (BPRS) (Leucht et al., 2005a) and the PANSS (Leucht et al., 2005b, Leucht et al., 2006) to CGI ratings.

Back to Article Outline

3. Results 

A total of 558 patients were enrolled at 102 participating centres. Of the 558 patients, 523 (93.7%) were included at psychiatric hospitals, 13 (2.3%) at a forensic hospital, and for 22 (3.9%) this information was missing. Most of the 558 patients were male (63.3%) with a median age of 38 years (mean 40.3, SD 16.2).

Most patients were diagnosed with schizophrenia spectrum disorders (n=330, 59.1%), followed by substance use, mood, and personality disorders (see Table 1). With regard to the type of psychiatric emergency, 73.1% (n=408) of the patients were acutely agitated, 70.4% were at risk to harm others (n=393), and 47.3% were at risk of self-harm (n=264). 238 patients (42.7%) were treated following a compulsory admission.

Table 1. Correlation of the CGI-A and PANSS-EC scales for diagnostic subgroups

243 patients (43.5%) entered into the study without pharmacological treatment. Of those being previously treated, 214 (38.4%) had received antipsychotics and 106 (19.0%) benzodiazepines; 67 patients (12.0%) were previously treated with mood stabilizers and 64 (11.5%) with antidepressants. Of the complete sample, 225 (40.4%) were treated with antipsychotic monotherapy during the study. 389 patients (69.7%) received concomitant treatment with benzodiazepines, mostly lorazepam (n=248, 44.4%) or diazepam (n=175, 31.4%). The most frequently used antipsychotics were olanzapine (n=390, 69.9%), haloperidol (n=132, 23.7%), risperidone (n=72, 12.9%), and quetiapine (n=38, 6.8%; multiple nominations were possible).

Almost all patients (97.4%) completed the 5-day study period; 1.2% of patients were lost to follow-up, 0.6% discontinued due to adverse events, and 0.8% discontinued due to other reasons.

3.1. PANSS-EC and CGI-A scores 

For all patients (n=558), the mean PANSS-EC total score (SD) decreased progressively from 25.1 points (SD 5.9) to 12.0 points (SD 5.9) at the last day (day 5). For the entire population at baseline, 17.7% displayed overt aggressive behaviour (CGI-A score=5), 26.2% were severely aggressive (CGI-A score=4) and 30.8% were moderately aggressive (CGI-A score=3). Overall, 74.7% of patients were found to have a CGI-Aggression score in the range of 3 to 5 points. The mean CGI-A at baseline was 3.3 points (SD 1.2). Over the 5-day observation period, the percentage of patients with at least moderate to high levels of aggression (CGI-A scores of 3 to 5 points) decreased progressively. At last observation (day 5), only 0.9% of the patients displayed aggressive behaviour, 1.4% were severely aggressive, and 8.6% moderately aggressive. At endpoint, 10.9% had a CGI-A score in the range of 3 to 5 points. The percentage of patients considered to be not aggressive (CGI-A score=1) was 7.9% at baseline and increased consistently over the observation period to 55.0% at day 5. The mean CGI-A at day 5 was 1.6 points (SD 0.8).

3.2. Validity of CGI-A 

In the complete sample, Spearman's correlation coefficient between CGI-A and the PANSS-EC total score was r=0.83 (95%-CI 0.82–0.84; cf. Table 1). Correlations for the PANSS-EC items varied between 0.73 for item 28 (poor impulse control) and 0.81 for item 7 (hostility). For patients with schizophrenia spectrum disorders, Spearman's rho between CGI-A and PANSS-EC total score was r=0.85 (95%-CI 0.84–0.85), correlations for the PANSS-EC items varied between 0.74 for item 28 (poor impulse control) and 0.82 for item 7 (hostility). Pearson's correlation coefficients were not different.

The relationship between CGI-A and PANSS-EC total score was found to be linear, with an average increase of 4.6 points (95%-CI 4.4–4.7) on the PANSS-EC for each additional CGI-A point for all patients (cf. Fig. 1).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Distribution of the PANSS-EC total scores corresponding to CGI-A values for all patients (unadjusted data). Box=25% and 75% quartiles, line=median, whiskers=minimum and maximum values.

The average increase was 4.7 points (95%-CI 4.5–4.8) for patients with schizophrenia spectrum disorders. In a linear mixed model, the CGI-A score explained 68.7% of the variance of the PANSS-EC total score for all patients and 69.5% for patients with schizophrenia spectrum disorders. The same model resulted in expected PANSS-EC total scores for no aggression (CGI-A=1) of 12.2 points (95%-CI 11.9–12.5) for all patients and 12.3 points (95%-CI 11.9–12.7) for patients with schizophrenia spectrum disorders. This score increased linearly to 30.4 points (95%-CI 30.0–30.8) for the highest value–aggressive behaviour present (CGI-A=5) for all patients and 30.9 points (95%-CI 30.4–31.4) for patients with schizophrenia spectrum disorders.

The equipercentile linking method yielded similar results with PANSS-EC total scores of 10, 17, 23, 28, and 32 corresponding to the 5 CGI-A categories for all patients and 9, 17, 23, 28, and 32 for patients with schizophrenia spectrum disorders.

Back to Article Outline

4. Discussion 

Management of aggression in schizophrenia spectrum disorders as well as in other psychiatric disorders remains a clinical challenge. Because rapid assessment of the severity of aggression is crucial for initial choice and further adjustment of an optimal anti-aggressive treatment including psychopharmacotherapy, easily administrable assessments with high face-validity and concurrent validity with other, well established instruments are needed. The CGI-A, a Likert scale to measure aggression tailored after the well-known CGI-S, has recently been applied in a number of trials, but has not yet been validated against other established rating scales.

In the present study, we used data from an observational study on the effectiveness and tolerability of psychopharmacological treatment of agitation and aggression. The PANSS-EC, a repeatedly applied measure derived from the PANSS scale, and the CGI-A were available as measures of aggression.

4.1. Key findings 

A mean baseline PANSS-EC total score of 25.1 and a mean CGI-A of 3.3 with 74.7% of the patients being at least moderately to severely aggressive (CGI-A≥3) correspond to a sample of severely ill patients. PANSS-EC total scores, reported in previous publications, were generally lower (Breier et al., 2002, Wright et al., 2003, Currier et al., 2004). Only Turczynski et al., 2004, San et al., 2006 reported comparably high PANSS-EC levels. At the end of the observation period, the mean PANSS-EC total score and mean CGI-A were low. Therefore, a large range of PANSS-EC and CGI-A values was available for analyses.

The correlation between CGI-A and PANSS-EC total scores was high (r=0.83). This was also observed for all single items of the PANSS-EC with correlations from r=0.72 to r=0.81. In comparison, Leucht and colleagues found a medium to high correlation (Spearman's rho: 0.56 to 0.73) between CGI-S and the PANSS total score (Leucht et al., 2005b) in 4091 patients with schizophrenia. The lower correlation reported by Leucht and colleagues might be caused by higher variance of PANSS scores due to the use of the 30-item full version of the PANSS and because data from seven clinical trials were pooled for analysis leading to a higher sample heterogeneity.

CGI-A and PANSS-EC were found to be linearly related, with a PANSS-EC of 12.2 corresponding to a CGI-A of 1, an average increase of 4.6 PANSS-EC points per CGI-A point, and a PANSS-EC score of 30.4 corresponding to CGI-A=5. Using equipercentile linking, similar results were obtained. About 70% of the variance of the PANSS-EC score could be explained by CGI-A.

The application of linear regression for the comparison of rating scales such as CGI-A and PANSS-EC has been critically discussed. It has been argued that, using linear regression, one scale would be treated as independent variable measured without error, and the other as the dependent variable measured with error. However, since values from both scales are obtained by psychopathological ratings and, thus, represent measurements with error, the application of linear regression is considered to be conceptually problematic (Leucht et al., 2005a, Leucht et al., 2005b, Leucht et al., 2006). On the other hand, it may be argued that linear regression provides an estimate for a dependent variable given the independent variable, and can – from a mathematical point of view – be applied even if both variables were measured with error. Additionally, it takes into account the correlation between the two variables and provides measurements regarding uncertainty, e.g. confidence intervals or p-values. Equipercentile linking, a method using both of the assessed scales symmetrically, has been proposed as a technique avoiding the conceptual problems of linear regression. However, this method doesn't take into account the correlation between the scales. It only considers different means, variances, skewness and higher degrees of momentum for each endpoint, but not for the joint distribution of both endpoints. Additionally, it does not provide any measurement regarding uncertainty. Considering the advantages and drawbacks of both linear regression and equipercentile linking, both methods to assess the association between CGI-A and PANSS-EC were used in this study.

In the current study a 5-level CGI-A was used, omitting the differentiation between “no illness”/“borderline ill” as well as between “markedly ill” and “severely ill” compared to the original version of the 7-level CGI-S (Guy, 1976). This may reduce the degree to which different levels of aggression can be discriminated. On the other hand, it increases rating simplicity. As several recent publications applied the 7-level version of the CGI-A (Frazier et al., 1999, Sival et al., 2004, MacMillan et al., 2006), PANSS-EC values corresponding to the 7-level CGI-A are needed for conversion between scales and comparison of different studies using either PANSS-EC or CGI-A. Values for the 7-level CGI-A version could be calculated from our data under the assumption that (i) the minimum PANSS-EC value, corresponding to CGI-A=1, and the maximum PANSS-EC value, corresponding to CGI-A=5/7, may be held fixed and that (ii) the relation between PANSS-EC and a 7-level CGI-A remains linear. Under these assumptions, an increase of 3.0 PANSS-EC points per 7-level CGI-A category was calculated. For all patients, PANSS-EC total scores of 12.2, 15.2, 18.3, 21.3, 24.3, 27.4, and 30.4 points would correspond to the 7 categories on the CGI-A. For patients with schizophrenia spectrum disorders, CGI-A categories would be equivalent to 12.3, 15.4, 18.5, 21.6, 24.7, 27.8, and 30.9 points on the PANSS-EC total score with an increase of 3.1 points per CGI-A category.

Because the PANSS was originally designed for the assessment of psychopathology in schizophrenia (Kay et al., 1987), and the PANSS-EC represents a subset of this scale, all analyses were repeated for the patient subgroup with schizophrenia spectrum disorder. All findings remained stable when only these patients were analyzed. Therefore, the CGI-A is applicable not only for patients with schizophrenia spectrum disorders, but for all psychiatric patients with an agitated and aggressive syndrome independently of primary diagnosis.

4.2. Strengths 

To the authors' knowledge, this is the first study validating CGI-A against an established measure of agitation and aggression. A large sample of 558 patients with an agitated and aggressive syndrome was followed daily from baseline to day 5 after initiation of anti-aggressive treatment. A high range of PANSS-EC and CGI-A scores could be obtained due to the fact that data were utilized from a fairly unselected sample with high initial and low endpoint severity scores of the scales under study. Subsequently ceiling and floor effects were unlikely to produce limited generalizability of the results. Because of methodological concerns brought forward against linear regression and equipercentile linking, both methods were used in the current analyses and gave similar results. Additionally, results are generalizable not only to patients with schizophrenia, but are applicable to a broad range of psychiatric patients treated for agitation and aggression.

4.3. Limitations 

Because of the observational nature of the current study, there are a number of methodological limitations: A large number of raters and study centres (n=102) were participating, raters were not blind to therapy and previous measurements, and there was no assessment of inter-rater reliability. Further, a 5-level CGI-A version was applied in the current study, whereas the 7-level CGI-A has been more commonly used in previous publications. Although the linear correlation between PANSS-EC and CGI-A allowed for the transfer of our results to the 7-level CGI-A, the direct validation of the latter version appears mandatory.

Back to Article Outline

5. Conclusion 

In the management of agitation and aggression, quickly and reliably administrable assessments are needed. The PANSS-EC is a subset of operationalized items from the PANSS scale and is a well-known rating-scale for agitation, excitement and aggression in schizophrenia spectrum disorders. The present results suggest a strong linear correlation between PANSS-EC and CGI-A. CGI-A scores 1 through 5 were equivalent to PANSS-EC total scores of 12.2, 16.7, 21.3, 25.8, and 30.4, respectively. On a 7-level CGI-A, values 1 through 7 corresponded to 12.2, 15.2, 18.3, 21.3, 24.3, 27.4, and 30.4 PANSS-EC points. These results allow for an easier clinical interpretation of PANSS-EC scores, and enable comparisons between studies where only one of both scales was used. Furthermore, this means that the CGI-A, a quickly administrable physician-rated measure with high face-validity relying on clinical experience, can be substituted for the PANSS-EC without losing relevant information. This further encourages application of the CGI-A as an instrument for rapid evaluation of behavioural emergencies that is not limited to patients with schizophrenia spectrum disorders, but can be used for all patients with an agitated and aggressive syndrome.

Back to Article Outline

Role of funding source 

The study was supported by Eli Lilly Company, Germany.

Back to Article Outline

Contributors 

Authors DN and HPH designed the study and wrote the protocol. Authors CGH and ML managed the literature searches and analyses. Authors AS and BGS undertook the statistical analyses, and author CGH wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript.

Back to Article Outline

Conflict of interest 

Drs Lambert, Naber, and Schimmelmann have received educational grants from Eli Lilly Company and furthermore Drs Huber, Lambert, Naber, and Schimmelmann have received grants from Eli Lilly Company for other studies and honorariums for lectures. None of the authors have more substantial associations with Eli Lilly Company, Germany. Drs Schacht, Hundemer, and Wagner are employees of the Medical Neuroscience Department, Lilly Deutschland GmbH.

Back to Article Outline

Acknowledgements 

The authors want to express their gratitude to the participating physicians.

Back to Article Outline

References 

1. Allen MH, Currier GW, Hughes DH, Reyes-Harde M, Docherty JP. The expert consensus guideline series. In: Postgraduate medicine. 2001;p. 1–88
   • View In Article
2. Allen MH, Currier GW, Carpenter D, Ross RW, Docherty JP. Treatment of behavioral emergencies 2005. J. Psychiatr. Pract. 2005;11(Suppl 1):5–108
   • View In Article
   • MEDLINE
   • CrossRef
3. Baker RW, Kinon BJ, Maguire GA, Liu H, Hill AL. Effectiveness of rapid initial dose escalation of up to forty milligrams per day of oral olanzapine in acute agitation. J. Clin. Psychopharmacol. 2003;23:342–348
   • View In Article
   • MEDLINE
   • CrossRef
4. Barzman DH, DelBello MP, Adler CM, Stanford KE, Strakowski SM. The efficacy and tolerability of quetiapine versus divalproex for the treatment of impulsivity and reactive aggression in adolescents with co-occurring bipolar disorder and disruptive behavior disorder(s). J. Child Adolesc. Psychopharmacol. 2006;16:665–670
   • View In Article
   • MEDLINE
   • CrossRef
5. Breier A, Meehan K, Birkett M, David S, Ferchland I, Sutton V, et al. A double-blind, placebo-controlled dose–response comparison of intramuscular olanzapine and haloperidol in the treatment of acute agitation in schizophrenia. Arch Gen Psychiatry. 2002;59:441–448
   • View In Article
   • MEDLINE
   • CrossRef
6. Currier GW, Chou JC, Feifel D, Bossie CA, Turkoz I, Mahmoud RA, et al. Acute treatment of psychotic agitation: a randomized comparison of oral treatment with risperidone and lorazepam versus intramuscular treatment with haloperidol and lorazepam. J. Clin. Psychiatry. 2004;65:386–394
   • View In Article
   • MEDLINE
   • CrossRef
7. Currier GW, Trenton AJ, Walsh PG, van Wijngaarden E. A pilot, open-label safety study of quetiapine for treatment of moderate psychotic agitation in the emergency setting. J. Psychiatr. Pract. 2006;12:223–228
   • View In Article
   • MEDLINE
   • CrossRef
8. Frazier JA, Meyer MC, Biederman J, Wozniak J, Wilens TE, Spencer TJ, et al. Risperidone treatment for juvenile bipolar disorder: a retrospective chart review. J. Am. Acad. Child Adolesc. Psych. 1999;38:960–965
   • View In Article
9. Guy W. Clinical global impression. ECDEU Assessment Manual for Psychopharmacology, revised. In: Guy, W. Rockville, MI, National Insitute of Mental Health. 1976;p. 217–222
   • View In Article
10. Haro JM, Kamath SA, Ochoa S, Novick D, Rele K, Fargas A, et al. The Clinical Global Impression-Schizophrenia scale: a simple instrument to measure the diversity of symptoms present in schizophrenia. Acta Psychiatr. Scand. 2003;16–23Suppl
    • View In Article
11. Haro JM, Ochoa S, Gervin M, Mavreas V, Jones P. Assessment of remission in schizophrenia with the CGI and CGI-SCH scales. Acta Psychiatr. Scand. 2007;115:163–164author reply 164
    • View In Article
    • MEDLINE
    • CrossRef
12. Kay SR, Sevy S. Pyramidical model of schizophrenia. Schizophr. Bull. 1990;16:537–545
    • View In Article
    • MEDLINE
13. Kay SR, Fiszbein A, Opler LA. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 1987;13:261–276
    • View In Article
    • MEDLINE
14. Leucht S, Kane JM, Kissling W, Hamann J, Etschel E, Engel R. Clinical implications of Brief Psychiatric Rating Scale scores. Br. J. Psychiatry. 2005;187:366–371
    • View In Article
    • MEDLINE
    • CrossRef
15. Leucht S, Kane JM, Kissling W, Hamann J, Etschel E, Engel RR. What does the PANSS mean?. Schizophr. Res. 2005;79:231–238
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (120 KB)
    • CrossRef
16. Leucht S, Kane JM, Etschel E, Kissling W, Hamann J, Engel RR. Linking the PANSS, BPRS, and CGI: clinical implications. Neuropsychopharmacology. 2006;31:2318–2325
    • View In Article
    • MEDLINE
17. Lindenmayer JP. The pathophysiology of agitation. J. Clin. Psychiatry. 2000;61(Suppl 14):5–10
    • View In Article
18. Lindenmayer JP, Czobor P, Alphs L, Nathan AM, Anand R, Islam Z, et al. The InterSePT scale for suicidal thinking reliability and validity. Schizophr. Res. 2003;63:161–170
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (130 KB)
    • CrossRef
19. MacMillan CM, Korndorfer SR, Rao S, Fleisher CA, Mezzacappa E, Gonzalez-Heydrich J. A comparison of divalproex and oxcarbazepine in aggressive youth with bipolar disorder. J. Psychiatr. Pract. 2006;12:214–222
    • View In Article
    • MEDLINE
    • CrossRef
20. Meehan KM, Wang H, David SR, Nisivoccia JR, Jones B, Beasley CM, et al. Comparison of rapidly acting intramuscular olanzapine, lorazepam, and placebo: a double-blind, randomized study in acutely agitated patients with dementia. Neuropsychopharmacology. 2002;26:494–504
    • View In Article
    • MEDLINE
    • CrossRef
21. Pascual JC, Madre M, Puigdemont D, Oller S, Corripio I, Diaz A, et al. A naturalistic study: 100 consecutive episodes of acute agitation in a psychiatric emergency department. Actas Esp. Psiquiatr. 2006;34:239–244
    • View In Article
    • MEDLINE
22. Pascual J, Perez V, Martin J, Safont G, Puigdemont D, Alvarez E. Olanzapine orally-disintegrating tablet in severe psychotic agitation: a naturalistic study. Actas Esp. Psiquiatr. 2007;35:47–51
    • View In Article
    • MEDLINE
23. Salzmann E, Robin JL. Multicentric double-blind study comparing efficacy and safety of minaprine and imipramine in dysthymic disorders. Neuropsychobiology. 1995;31:68–75
    • View In Article
    • MEDLINE
    • CrossRef
24. San L, Arranz B, Querejeta I, Barrio S, De la Gandara J, Perez V. A naturalistic multicenter study of intramuscular olanzapine in the treatment of acutely agitated manic or schizophrenic patients. Eur. Psychiatr. 2006;21:539–543
    • View In Article
25. Sharif ZA, Raza A, Ratakonda SS. Comparative efficacy of risperidone and clozapine in the treatment of patients with refractory schizophrenia or schizoaffective disorder: a retrospective analysis. J. Clin. Psychiatry. 2000;61:498–504
    • View In Article
    • MEDLINE
    • CrossRef
26. Sival RC, Duivenvoorden HJ, Jansen PA, Haffmans PM, Duursma SA, Eikelenboom P. Sodium valproate in aggressive behaviour in dementia: a twelve-week open label follow-up study. Int. J. Geriatr. Psychiatry. 2004;19:305–312
    • View In Article
    • MEDLINE
    • CrossRef
27. Spearing MK, Post RM, Leverich GS, Brandt D, Nolen W. Modification of the Clinical Global Impressions (CGI) Scale for use in bipolar illness (BP): the CGI-BP. Psychiatry Res. 1997;73:159–171
    • View In Article
    • Abstract
    • Full-Text PDF (1361 KB)
    • CrossRef
28. Turczynski J, Bidzan L, Staszewska-Malys E. Olanzapine in the treatment of agitation in hospitalized patients with schizophrenia and schizoaffective and schizofreniform disorders. Med. Sci. Monit. 2004;10:PI74–PI80
    • View In Article
    • MEDLINE
29. WHO . ICD-10: The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines. Geneva: World Health Organisation (WHO); 1992;
    • View In Article
30. Wright P, Meehan K, Birkett M, Lindborg SR, Taylor CC, Morris P, et al. A comparison of the efficacy and safety of olanzapine versus haloperidol during transition from intramuscular to oral therapy. Clin. Ther. 2003;25:1420–1428
    • View In Article
    • Abstract
    • Full-Text PDF (677 KB)
    • CrossRef
31. Zhong KX, Tariot PN, Mintzer J, Minkwitz MC, Devine NA. Quetiapine to treat agitation in dementia: a randomized, double-blind, placebo-controlled study. Curr. Alzheimer Res. 2007;4:81–93
    • View In Article
    • MEDLINE
    • CrossRef