Prevalence of Personality Disorders Among Clients in Treatment for Addiction

Louise Nadeau, PhD1, Michel Landry, PhD2, Stéphane Racine, MPs3


This study determined the prevalence of personality disorders among clients in treatment for addiction; this prevalence was compared with those found in similar studies and in clinical samples of individuals suffering from other Axis I disorders. Our sample comprised 255 subjects. The first edition of the Millon Clinical Multiaxial Inventory (MCMI) was used. Only 11.8% of the subjects did not score over 84 on any of the 11 Axis II scales. Over one-half had a score of 84 or higher on the passive-aggressive and dependent-personality scales. The mean number of scales in the 84+ category was 2.68. Comparisons show that this sample was more severe in most cases.

(Can J Psychiatry 1999;44:592–596)

Key Words: personality disorder, substance disorder, Millon Clinical Multiaxial Inventory, treatment, gender

Personality disorders (PDs) among clients with substance disorders (SDs) have been less studied than have Axis I disorders, but instruments that are valid and inexpensive to administer, such as the Millon Clinical Multiaxial Inventory (MCMI) (1) used in the present study, have allowed their prevalence to be documented.

Clinical studies have found that the proportion of individuals presenting both SDs and PDs varied from 53% to 100% (2–10). The most frequently diagnosed disorders are antisocial, borderline, narcissistic, and dependent personality disorders. The simultaneous presence of Axis I and II disorders increases the severity of all disorders (11,12). In cases of Axis II comorbidity, addiction can be successfully controlled in specialized treatment centres. However, the needs of these clients often exceed the capabilities of the services normally provided (13,14). There is an increased likelihood of premature termination and reduced effectiveness of treatment (15–17). For those who complete treatment, remission is comparable to that observed among those without PDs, although the level of psychological distress remains high (18).

Objectives

This study determines the prevalence of PDs among individuals admitted to public treatment centres in Quebec and compares these results with findings of similar studies and with those of clinical samples suffering from other Axis I disorders.

Method

Our sample comprised 255 clients—182 men and 73 women—in 8 SD treatment centres in Quebec. The subjects were francophone and aged 18 years or over. The mean age was 34.5 years: 12% aged 18–24 years; 41% aged 25–34 years; 34% aged 35–44 years; and 13% aged 45 years or older. Testing took place after 7 days of treatment in order to exclude individuals suffering from acute withdrawal. Participation was voluntary, and all subjects signed a consent form.

We compared our results with those of 5 studies (4,8,19–21) that had used the MCMI-I (Table 1).

Table 1. Comparison of 5 studies that used the Millon Clinical Multiaxial Inventory (MCMI-I)


Author

N

Diagnosis at admission

Type of treatment

Time of testing

Ness and others (20)

30

Opiate abuse or dependence

After 48–72 hours of abstinence

Brown (19)

50

All substance dependencies

12-step treatment from 21–30 days detoxification

After 21–30 days of treatment

Craig and others (4)

86

Opiate abuse or dependence

Detoxification

After detoxification

Craig and others (4)

107

Cocaine abuse

Detoxification

After detoxification

Marsh and others (8)

159

Opiate abuse or dependence

Methadone treatment

After admission to treatment

Craig and others (21)

106

Alcohol abuse or dependence

Craig and others (21)

100

Opiate abuse or dependence


To better understand the specificity of SD, our results were also compared with those of 3 clinical samples from Quebec. The first comprised 180 subjects treated for erectile or orgasmic disorders. The other 2 samples included individuals treated for sexual impulse problems at a medicolegal clinic: one comprised 44 men who had committed a rape or had intrusive rape fantasies, and in the other were 87 men who had committed pedophilic acts. All subjects were francophone. All subjects signed a consent form.

We used the first edition of the MCMI (1), a self-report inventory asking 175 true or false questions. Our results focused on the 11 personality scales (PSs) described in Table 2. The MCMI was translated into French (22). Based on psychometric qualities (that is, internal consistency, test– retest reliability, and discriminant validity), this translation is reliable and valid (23). Millon’s taxonomy parallels that of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III).

Table 2. Millon Clinical Multiaxial Inventory base rate scores


 

By scale


Highest score


Average score


 

Base rate score > 84


Base rate score > 74


       

Axis II

%

Rank

%

Rank

%

Rank

%

Rank

No score +

11.8

1.2

0

11.8

Schizoid

33.3

4

52.5

5

3.1

8

70.61

5

Avoidant

48.6

3

68.6

3

14.5

3

81.05

2

Dependent

52.9

2

69.4

2

21.2

1

75.85

3

Histrionic

10.6

9

24.7

8

2.7

9

58.63

9

Narcissistic

12.5

6

20.8

10

4.3

5

56.58

10

Antisocial

12.5

6

26.3

7

3.5

7

60.90

8

Compulsive

0.0

11

0.4

11

0.0

11

39.66

11

Passive-aggressive

56.9

1

76.5

1

20.8

2

82.30

1

Schizotypal

7.1

10

22.0

9

0.4

10

65.85

7

Borderline

21.2

5

55.3

4

3.9

6

75.45

4

Paranoid

12.5

6

31.4

6

4.7

4

69.05

6

Mixed

9.0


Base rate (BR) scores were computed for each scale. A BR score of 74 or more (74+) means that the subject shows signs of the disorder; a score of 84 or more (84+) signals prominence of a disorder. Fifteen tests were rejected because of the MCMI validity index (1 or more on the Y scale); 9 were rejected because of exaggeration of symptoms (sum of scales 1 to 8 greater than 164); 6 were rejected because of missing data; 3 were rejected because the subjects were under age 18 years. In the other clinical samples, 25 tests (8%) were rejected.

Statistical tests designed to compare means and proportions were used. The types of tests used are specified below. Differences between genders were calculated when the number of subjects allowed it.

Results

The first column of Table 2 shows the results by scale. Subjects are classified in this category for each scale on which their score is 84+. Only 11.8% did not score over 84 on any scale, over one-half scored on the passive-aggressive (56.9%) and dependent (52.9%) PS, and 0% scored on the compulsive PS. For the 74 cutoff point, the third column of Table 2 shows 98.8% of subjects to be above this threshold. Over 50% scored on either the passive-aggressive, dependent, avoidant, borderline, or schizoid PS. None scored more that 74 on the compulsive PS. Significantly more women than men scored 84+ on the histrionic, schizotypal, borderline, and paranoid PS. Using the 74 cutoff point, significant gender differences remained only for the histrionic and borderline PS.

The fifth column of Table 2 presents the highest score obtained on a scale. Each subject is assigned to a single category. If the same score is obtained on 2 or more scales, the mixed category is used. Over one-fifth of subjects obtained their highest score on the passive-aggressive (20.8%) or the dependent (21.2%) PS. Comparisons by gender were performed using chi-square or Fisher’s test when more than 20% of the cells contained less than 5 elements. Women were underrepresented on the passive-aggressive scale.

In the seventh column of Table 2, mean scores are presented for each scale. Mean scores for the avoidant, dependent, passive-aggressive, and borderline PSs are all in the 74+ range. As shown in Table 3, comparisons by gender were made using t-tests. Women had a higher mean score than men on the borderline PS.

Table 3. Mean score on personality scales by gender (substance abuse disorder, Quebec sample)


 

Mean score (SD)


Scale

Male

Female

Total

Schizoid

70.37 (23.44)

71.19 (25.59)

70.61 (24.03)

Avoidant

81.04 (21.11)

81.10 (23.99)

81.05 (21.92)

Dependent

75.34 (24.66)

77.15 (25.83)

75.85 (24.97)

Histrionic

59.82 (19.23)

55.64 (28.70)

58.63 (22.38)

Narcissistic

56.95 (21.29)

55.66 (24.92)

56.58 (22.35)

Antisocial

61.68 (22.10)

58.96 (21.99)

60.90 (22.06)

Compulsive

39.18 (18.02)

40.88 (21.85)

39.66 (19.16)

Passive-aggressive

82.67 (23.36)

81.38 (25.30)

82.30 (23.89)

Schizotypal

65.60 (11.16)

66.48 (14.14)

65.85 (12.07)

Borderlinea

73.57 (12.19)

80.12 (16.40)

75.45 (13.81)

Paranoid

68.11 (12.96)

71.40 (16.31)

69.05 (14.05)


aP < 0.01.

For each subject, we summed the number of scales in the 84+ category. The mean number of scales was 2.68, with 71% of subjects scoring more than 84 on 1 scale. For the cutoff score of 74, the mean number of scales was 4.48, with 95% of subjects scoring at least 1 scale. The Mann-Whitney U nonparametric test was used to make comparisons by gender, since this variable is not continuous. On average, women scored 84+ on 3.08 scales, compared with 2.52 scales for men (U = 5614.5, P = 0.05). There was no statistically significant difference for the 74 cutoff point.

Our results were compared with those of 8 samples mentioned in Table 1. t-Tests showed that for 6 scales—schizoid, avoidant, dependent, passive-aggressive, schizotypal, and borderline—our sample had the highest scores and, in most cases, they differed significantly from those of other samples. Our sample had the lowest scores for the narcissistic and compulsive scales. On 3 scales—histrionic, antisocial, and paranoid—our sample was close to the midpoint of the distribution.

The mean BR scores and standard deviations of our sample were also compared with those of 3 other Quebec clinical samples. ANOVA and a Student-Newman-Keuls test were used. Our sample obtained higher scores on several scales (Table 4). These results also indicate that these 4 samples present different Axis II profiles.

Table 4. Comparison of Millon Clinical Multiaxial Inventory scale score between 4 samples


 

Alcohol and drug abuse

Sexual
dysfunction


Rape


Pedophilia


Total

Axis II

(n = 255)

(n = 173)

(n = 34)

(n = 79)

(n = 541)

Schizoid

70.61

45.95–

61.50

61.77–

60.86

Avoidant

81.05

53.01–

70.50–

72.62–

70.19

Dependent

75.85

57.40–

67.15

76.38

69.48

Histrionic

58.63

55.08–

56.15

52.65

56.46

Narcissistic

56.58

63.35+

60.41

57.53

59.12

Antisocial

60.90

66.55+

67.38

59.56

62.92

Compulsive

39.66

60.14+

51.29+

51.63+

48.69

Passive-aggressive

82.30

51.65–

63.94–

63.76–

68.64

Schizotypal

65.85

58.21–

62.09

65.29

63.09

Borderline

75.45

62.66–

67.12–

69.96–

70.04

Paranoid

69.05

69.46

67.32

69.68

69.16


+This sample mean is significantly higher than the drug alcohol abuse sample.
–This sample mean is significantly lower than the drug alcohol abuse sample.

Discussion

The MCMI is not a diagnostic instrument (24). It measures whether certain features have attained clinical significance. Our own test–retest reliability study (23) has shown that the intensity of symptoms decreases after a few weeks of treatment. All authors except Millon have reported similar results (25). Toxic effects of substances probably decrease after a few weeks of treatment (26).

Our results reveal relatively severe personality disorganization in the upper range, compared with other studies with addicted clients. These results confirm that no single PD is typical of all addicts (27). However, these comparisons are subject to limitations. The length of time between admission to treatment and test completion may have varied from one study to another. Particular psychosocial characteristics of the different samples may have influenced the results.

The comparison with 3 Quebec clinical samples confirms the more severe Axis II features of our sample. These findings are consistent with the results of a previous study (28) in 3 rehabilitation centres using a validated French version of the Addiction Severity Index (29), which found greater deterioration of the psychological, family, and legal spheres in Quebec samples than in those from outside Quebec. The consistency confirms the validity of the 2 instruments and also hints about the links between Axis I and II disorders.

Results concerning the passive-aggressive PS of the MCMI-I should be interpreted with caution. Several clinicians noted, as we did, that its high scores suggest an overlap with SD. In the DSM-IV, the passive-aggressive PD has been relegated to an annex because of doubts as to its specificity. Millon excluded this scale from the MCMI-III.

Results according to gender show that PDs are more prevalent among women, no matter how the results are analyzed. Numerous studies have demonstrated similar results (30). Women, however, have as good a prognosis as men and, in some cases, a better one (31). The specific effect of PD on treatment outcome in reference to gender, however, is unknown.

SDs and Axis II disorders must be treated concurrently (10,32,33). Those in charge of treatment programs should examine how best to adapt their services to deal with Axis II comorbidity.


Clinical Implications

Limitations

Acknowledgements

This research was supported by a grant from the Conseil québécois de la recherche sociale to the team Recherche et Intervention sur les Substances psychoactives – Québec (RISQ).

References

1. Millon T. Millon Clinical Multiaxial Inventory. Minneapolis (MN): National Computer Systems; 1983.

2. Calsyn DA, Saxon AJ. Personality disorder subtypes among cocaine and opioid addicts using the Millon Clinical Multiaxial Inventory. International Journal of the Addictions 1990;25:1037–49.

3. Craig RJ. A psychometric study of the prevalence of DSM-III personality disorders among treated opiate addicts. International Journal of the Addictions 1988;23:115–24.

4. Craig RJ, Olson RE. MCMI comparisons of cocaine abusers and heroin addicts. J Clin Psychol 1990;46:230–7.

5. Dougherty RJ, Lesswing NJ. Inpatient cocaine abusers: an analysis of psychological and demographic variables. J Subst Abuse Treat 1989;6:45–7.

6. Khantzian EJ, Treece C. DSM-III psychiatric diagnosis of narcotic addicts. Arch Gen Psychiatry 1985;42:1067–71.

7. Koenigsberg HW, Kaplan RD, Gilmore MM, Cooper AM. The relationship between syndrome and personality disorder in DSM-III: experience with 2,462 patients. Am J Psychiatry 1985;142:207–12.

8. Marsch DT, Stile SA, Stoughton NL, Trout-Landen BL. Psychopathology of opiate addiction: comparative data from the MMPI and MCMI. Am J Drug Alcohol Abuse 1988;14:17–27.

9. Weiss RD, Mirin SM, Michael JL, Sollogub AC. Psychopathology in chronic cocaine abusers. Am J Drug Alcohol Abuse 1986;121:17–29.

10. Matano RA, Locke KD, Schwartz K. MCMI personality subtypes for male and female alcoholics. J Pers Assess 1994;63:250–64.

11. DeJong CA, Van Den Brink W, Harteveld FM, Van Der Wielen EG. Personality disorders in alcoholics and drug addicts. Compr Psychiatry 1993;34:87–94.

12. Sheenan MF. Dual diagnosis. Special issue: contemporary topics in drug dependence and alcoholism. Psychiatr Q 1993;64:107–34.

13. Clopton JR, Weddige RL, Contreras SA, Fliszar GM, Arredondo R. Treatment outcome for substance misuse patients with personality disorder. International Journal of the Addictions 1993; 28:1147-1153.

14. Boudreault L, Duhamel D, Maltais K, Marceau JP. Les troubles de la personnalité dans le contexte de l’intervention en toxicomanie. In: Brisson P, editor. L’usage des drogues et la toxicomanie. Volume II. Montréal: Gaëtan Morin; 1994. p 437–57.

15. Reich JH, Green AI. Effect of personality disorders on outcome of treatment. J Nerv Ment Dis 1991;179:74–82.

16. Reich JH, Vasile EG. Effect of personality disorders on the treatment of Axis I conditions: an update. J Nerv Ment Dis 1993;181:475–84.

17. Nace EP, Davis CW. Treatment outcome in substance-abusing patients with a personality disorder. Am J Addict 1993;2:26–33.

18. Penick EC, Nickel EJ, Cantrell PF, Powell BJ, Read MR, Thomas MM. The emerging concept of dual diagnosis: an overview and implications. [Special issue: Managing the dually diagnosed patient: current issues and clinical approaches.] Journal of Chemical Dependency Treatment 1990;3:1–27.

19. Brown HP. Substance abuse and the disorders of the self: examining the relationship. J Psychoactive Drugs 1992;21:139–44.

20. Ness R, Handelsman L, Aronson MJ, Hershkowitz A, Kanof PD. The acute effects of rapid medical detoxification upon dysphoria and other psychopathology experienced by heroin abusers. J Nerv Ment Dis 1994;182:353–9.

21. Craig RJ, Verinis JS, Vexler S, Personality characteristics of drug addicts and alcoholics on the Millon Clinical Multiaxial Inventory. J Pers Assess 1985;49:156–60.

22. D’Elia A, Lagier PM. Inventaire clinique multiaxial de Millon. Montréal: Bureau d’intervention psychosociale; 1986.

23. Landry M, Nadeau L, Racine S. La prévalence des troubles de la personnalité dans huit centres de réadaptation du Québec. Cahier de recherche. Montréal: Recherche et Intervention sur les Substances psychoactives/Québec; 1996.

24. Perry JC. Problems and considerations in the valid assessment of personality disorders. Am J Psychiatry 1992;149:1645–53.

25. Zimmerman M. Diagnosing personality disorders: a review of issues and research methods. Arch Gen Psychiatry 1994;51:225–45.

26. Schuckit MA, Monteiro MS. Alcoholism, anxiety and depression. British Journal of Addiction 1983;83:1373–80.

27. Craig RJ. The role of personality in understanding substance abuse. Alcoholism Treatment Quarterly 1995;13:17–27.

28. Guyon L, Landry M. L’abus de substances psychoactives, un problème parmi d’autres? Portrait d’une population en traitement. Psychotropes, revue internationale des toxicomanies 1996;1:61–81.

29. McLellan AT, Luborsky L, O’Brien CP. An improved evaluation instrument for substance abuse patients. J Nerv Ment Disorder 1981;168:26–33

30. Nadeau L. Les femmes et l’alcool: état de la question. In: Brisson P, editor. L’usage des drogues et la toxicomanie. Volume II. Montréal: Gaëtan Morin; 1994. p 231–54.

31. Lightfoot L, Adrian M, Leigh G, Thompson J. Women’s use of and views on substance abuse treatment. In: Adrian M, Lundy C, Eliany M, editors. Women’s use of alcohol, tobacco and other drugs in Canada/Les Canadiennes et l’usage d’alcool, de tabac et d’autres drogues. Toronto: Addiction Research Foundation; 1996. p 228–46.

32. Klein RH, Orleans JF, Soulé CR. The Axis II group: treating severely characterologically disturbed patients. Int J Group Psychother 1991;41:97–115.

33. Southwick SM, Satel SL. Exploring the meanings of substance abuse: an important dimension of early work with borderline patients. Am J Psychother 1990;44:61–7.


Résumé

Cette étude a déterminé la prévalence des troubles de personnalité parmi les clients en traitement pour toxicomanie ; cette prévalence a été comparée à celles constatées dans des études semblables et dans des échantillons cliniques de personnes souffrant d’autres troubles de l’axe I. Notre échantillon se composait de 255 sujets. La première édition de l’Inventaire clinique multiaxial de Millon (MCMI) a été utilisée. Seulement 11,8 % des sujets n’ont pas obtenu de résultat supérieur à 84 à aucune des 11 échelles de l’axe II. Le nombre moyen d’échelles dans la catégorie des 84 et plus était 2,68. Les comparaisons indiquent que cet échantillon était plus gravement malade dans la plupart des cas.


Manuscript received April 1998, revised, and accepted November 1998.

1Associate professor, Département de psychologie, Université de Montréal, Montreal, Quebec.

2Director of Professional Services, Centre Dollard-Cormier, Montreal, Quebec.

3Scientist, Recherche et Intervention sur les Substances psychoactives - Québec (RISQ), Montreal, Quebec.

Address for correspondence: Dr L Nadeau, Département de psychologie, Université de Montréal, CP 6128, succ « Centre-ville », Montreal, QC  H3C 3J7

email: louise.nadeau.2@umontreal.ca

Can J Psychiatry, Vol 44, August 1999