Sage Journals: Discover world-class research

Abstract

Background: Personality disorder comorbidity has been extensively studied in young adult populations, to a lesser extent in elderly populations, and not at all in an Australian population. This study examines PD comorbidity over the life span 18-100.

Aim: The object of this study was to examine the interactions of comorbid personality disorder and age on outcome of Axis I disorders.

Method: A total of 238 consecutive consenting eligible psychiatric inpatients were assessed on admission, prior to discharge, and after 6 and 12 months as regards symptoms, function, well-being, relapse and readmission rates and social supports. Outcomes were compared for young (18–40 years old), middle-aged (41–64) and old (65+) patients.

Results: Patients improved over time symptomatically and functionally. Across all age groups patients with comorbid personality disorder had worse outcomes than those without, but improved though never to the same extent. Personality disorder was associated with increased rates of relapse and readmission in the whole sample and in the older group, but not increased length of stay. Severity of personality disorder was associated with poorer outcome.

Conclusion: Personality disorder adversely affects outcomes, particularly for younger (and older) patients with psychiatric disorders independently of diagnosis and other factors.

Keywords

personality disorders older adults comorbidity outcome

Comorbid personality disorder (PD) occurs in up to 70% of inpatients [1–4] and in 30–60% of psychiatric outpatients [5–7]. Rates may be lower in older patients [8] and in the community where the prevalence of PDs has been estimated to be between 6 and 10% [9].

Reported effects of comorbid PD on the outcome of Axis I disorder are varied. The severity of the personality dysfunction may be more influential on outcome than the existence of PDs per se, emphasizing the need to assess the severity as well as the presence of personality disorder [10]. Young adults have been found to have more complex PDs (abnormal traits across many dimensions) than older adults and might be expected to have worse outcomes [4]. Also, although treatments for Axis I disorders, e.g. for depression, are efficacious in people with PD, those with PD have more severe symptoms at the beginning of treatment and this difference persists so that they still have a higher level of symptoms than equivalent patients without PD at the completion of treatment.

While several reviews have concluded that the presence of PD adversely affects outcomes of Axis I disorders, the majority of these focused on anxiety and depression [11]. Part of the difference between outcomes for PD and non-PD patients may result from increased dropout rates in the PD group and poor interpersonal interactions in therapy [12] and their need for longer, more intense multimodal therapies [13].

Other studies report no difference or differential effects depending on the treatment [14]. Hirschfeld [15] did not find that personality disorders impeded treatment for depression but treatment of the depression could lead to improvement in the PD. Mulder reported no overall effect of PD on depression outcome except in patients with Cluster B PD who responded less well to tricyclics [16]. Increased lengths of stay in patients comorbid for PD have been reported by some [17] but not others [18]. Increased rates of readmission have been reported [19].

The interaction of age and PD on Axis I disorder outcome has received less attention. Agbayewa found that young PD inpatients had longer episodes of their Axis I disorder and poorer family relationships whereas the older inpatients had more severe episodes [20]. In a study of late life depression, people with PD were more likely to be admitted than those without PD [21]. PD may affect the frequency and mode of expression of Axis I disorders [22]. Reasons for these differences include different ways of measuring outcome, different populations, differing severity of the PD, different types of PD, different Axis I disorders, and defined age groups. All studies so far have examined discreet age cohorts and none has yet compared outcomes across age groups.

The aim of this paper was to examine the impact of comorbid PD on the outcome of a range of Axis I disorders in a sample of psychiatric inpatients aged 18 to 100. Our hypotheses were that PD comorbidity would be detrimental to outcome as regards symptoms, function and well-being, and that this effect would be more pronounced in the young. Additionally, we hypothesized that PD would be associated with an increased length of stay and greater rate of relapse and readmission.

Method

Sample

Participants were 238 consecutive consenting competent adult psychiatric inpatients (89 male, 149 female) admitted to an acute adult psychiatric unit or a psychogeriatric unit in one of two public hospitals. Adults aged less than 65 years (n = 134, range 18 to 64 years) were patients admitted to a 24 bed adult psychiatric unit located in Westmead Hospital, Sydney. Older participants (n = 104; age range 65 to 100 years) came from either of the eight bed psychogeriatric units located at Westmead Hospital or Prince of Wales Hospitals in Sydney where as a rule all older adults are admitted to the psychogeriatric units. The Westmead Unit is in a middle-class/working-class area and the Prince of Wales Unit caters for a more socio-economically diverse population with more patients from professional backgrounds. As equal numbers of patients came from each psychogeriatric unit there was a wide socio-economic distribution of patients. Admissions to public hospital psychiatric units are determined by geographical catchment area and are free. Private facilities in both areas admit patients of all ages and were not included in this study. For the purposes of this study participants were categorized into 3 age groups, young (18–40), middle-aged (41–64), and old (65+ years).

Participants were excluded from the study if they had a Mini-Mental State Examination (MMSE) score of less than 24, or showed significant cognitive dysfunction on the General Practitioner Cognitive Assessment instrument (GPCOG), Frontal Assessment Battery (FAB) or clock drawing tests, suffered from severe physical illness, were not proficient in English, were respite patients or had a length of stay (LOS) less than 3 days (details and references follow).

Measures

Psychiatric diagnoses were established using the Structured Clinical Interview for DSM-IV Axis I disorders [23]. The SCID-I has demonstrated reliability in adults [24] and has been used in older adult studies [25]. Cognitive evaluation included the Mini-Mental State Examination [26], the Frontal Assessment Battery [27], GPCOG [28], and the 10-point clock test [29]. Presence of a personality disorder was determined using the Structured Clinical Interview for DSM IV Axis II (SCID-II) Personality Disorders Personality Questionnaire [23], a semi-structured diagnostic interview for assessing the ten DSM IV (American Psychiatric Association, 1994) Axis II personality disorders, as well as depressive and passive aggressive personality disorders. Categorical diagnoses of specific personality disorders required participants to satisfy a certain number of criteria at an agreed level of severity from the 119 questions. The PDs are grouped into three clusters, Cluster A (‘odd, eccentric’) comprises paranoid, schizoid, and schizotypal; Cluster B (‘dramatic, erratic’) comprises antisocial, borderline, histrionic and narcissistic; and Cluster C (‘anxious’) comprises obsessive–compulsive, avoidant and dependent personality disorders. This questionnaire has demonstrated reliability and construct validity against DSM-IV criteria [30]. It includes all of the diagnostic criteria aligned with each PD type, ensuring no features are missed, as happens with a clinical diagnosis and accounts for the higher rates of PD when an instrument is used [31]. This instrument defines cases as subjects who fulfil sufficient criteria at sufficient intensity for a specific PD or PDs. Assessment of inter-rater reliability, performed between the two raters on 20 patients, was high: r = 0.85.

Outcomes were determined using measures of symptom severity, the Symptom Checklist-90 Global Symptom Index (SCLGSI)[32], function, the Global Assessment of Function (GAF)[33], physical and emotional morbidity, the SF-12 [34], and well-being, the Personal Well-being Index (PWI) [35]. The Sarason Social Support Questionnaire was used to measure the number of social supports (SSQ-N) as well as satisfaction with social supports (SSQ-S) [36]. The expanded version of the Brief Psychiatric Rating Scale [37,38] was used to assess psychiatric symptoms. Coping behaviour was measured using the brief COPE [39]. Relapse was defined as meeting criteria for Axis I disorder after at least eight weeks of recovery and readmission as admission to a psychiatric facility within the 12 month period after discharge.

Protocol

After complete description of the study to the subjects, written informed consent was obtained. Participants were assessed four times, shortly after admission and prior to discharge using structured interviews conducted by an experienced psychiatrist, and again at 6 and 12 months. The follow-up interviews were conducted by a research psychologist who was blind to all previous data. On admission participants were screened to make sure they met the inclusion criteria, the cognitive battery was completed and the SCID-I was administered to establish Axis I diagnoses. In order to avoid state effects on personality, the SCID-II was administered shortly before discharge when the Axis I symptoms had largely resolved, by the same psychiatrist. Relapse and readmission data were gathered from the subjects at the 6 and 12 month follow-up interviews and from hospital records.

Analyses

The statistical program SPSS 17 (Chicago, IL) was used to analyse the data. Two-sided tests with a significance level of 5% were used throughout. The distributions of categorical variables by age group were compared using chi-squared tests. Repeated measures analysis of variance was used to assess the joint effects of PD, age group and time (baseline, discharge, follow up) on the symptom and function scores. Linear regression models which adjusted for the effect of baseline scores were used to quantify the difference between PD and non-PD groups in key outcome variables at discharge and follow up. Multiple linear regression analysis was used to determine the relative contribution of significant variables to symptomatic outcome, functional outcome and readmission.

Ethics

The protocol was approved by the ethics committees of Westmead and Prince of Wales Hospitals and Sydney University.

Results

Demographics and relevant clinical baseline data are shown in Table 1. There was a high prevalence of PD across the spectrum of Axis I disorders with 67.2% of all patients being so diagnosed. When this was analysed by age, younger patients were far more likely to have complex PDs (39% versus 10% of old) while older patients were more likely to be free of PD (40% versus 26% of young). Of 116 patients with a primary Axis I diagnosis of mood disorders, 71 (61.2%) had a PD. Corresponding rates were: for schizophrenia 31/55 (56.4%), drug and alcohol abuse 11/13 (84.6%), anxiety disorders 4/4 (100%), eating disorders 13/14 (92.9%), and other disorders 30/36 (83.3%).

Table 1.

Demographic and clinical details of sample

		Young (N = 76)	Middle-aged (N = 57)	Old (N = 98)	χ²	df	p
Age	Range	17–40	41–64	65–99
	Mean	27.49	52.33	76.23
	SD	6.63	6.81	7.44
Gender	Male	21 (27.6%)	27 (47.2%)	40 (39.8%)
	Female	55 (72.4%)	30 (52.6%)	58 (60.2%)	5.97	2	0.057
First Language	English	54 (71.1%)	45 (78.9%)	80 (80.4%)
	Other	22 (28.9%)	12 (21.1%)	18 (19.6%)	1.9	2	0.316
Marital status	Married	22 (28.9%)	24 (42.1%)	37 (38.2%)
	Separated	5 (6.6%)	19 (33.3%)	17 (16.7%)
	Never married	49 (64.5%)	9 (15.8%)	10 (10.8%)
	Widowed	0 (0%)	5 (8.8%)	34 (34.3%)	97.7	6	0.000
Education	Primary	0 (0%)	6 (10.5%)	27 (28.4%)
	Junior High	21 (27.6%)	26 (45.6%)	53 (53.9%)
	Senior High/Matriculation	15 (19.7%)	4 (7.0%)	5 (4.9%)
	TAFE/trade	11 (14.5%)	11 (19.3%)	6 (5.9%)
	University	29 (38.2%)	10 (17.5%)	7 (6.9%)	72.68	8	0.000
Living arrangements	Alone	11 (14.5%)	17 (29.8%)	45 (45.9%)
	Family	62 (81.5%)	38 (66.6%)	45 (45.9%)
	Supported care	1 (1.3%)	2 (3.5%)	6 (6.1%)
	Friends/flatmate	2 (2.6%)	0 (0%)	2 (2.0%)	23.8	4	0.000
Employment	Full-time	16 (21.1%)	13 (22.8%)	2 (2.0%)
	Part-time	14 (18.4%)	5 (8.8%)	3 (2.9%)
	Not working	46 (60.5%)	39 (68.4%)	93 (95.1%)	36.3	3	0.000
Psychiatric history	No	10 (13.2%)	11 (19.3%)	27 (26.5%)
	Yes, recent (12 months)	18 (23.7%)	9 (15.8%)	24 (24.5%)
	Chronic (>5 years)	48 (63.2%)	37 (64.9%)	47 (49.0%)	7.81	4	0.114
MMSE	Mean score	30	30	28.5
GPCOG	Mean score	9	9	7.9
FAB	Mean (SD)	16.61 (1.69)	14.74 (3.19)	14 (3.19)
SF-12 physical	Mean (SD)	49.1 (13.5)	43.0 (14.7)	44.3 (12.6)
SSQ-N	Mean (SD) number of social supports	3.23 (2.37)	2.19 (2.35)	2.00 (2.00)
SSQ-S	Mean (SD) satisfaction with social supports	4.59 (1.35)	4.24 (1.60)	4.09 (1.83)

TAFE, Technical and Further Education Commission; MMSE, Mini Mental State Examination; GPCOG, General practitioner assessment of cognition; FAB, frontal assessment battery at bedside; SF-12 physical, assessment of physical morbidity; SSQ-N, social support questionnaire-number of social supports; SSQ-S, social support questionnaire-satisfaction with social supports.

Symptoms

Symptom outcome by age

Scores on all outcome measures improved over time for patients in each age group whether or not they had a PD (see Table 2) and all age groups improved at the same rate.

Table 2.

Scores for all subjects irrespective of PD status, comparing rates of change for three age groups (young, middle, old)

	Admission	Discharge	Follow up	F	F
Measure	Mean	SD	Mean	SD	Mean	SD	time	sig	Time* age	P value
GAF young	37.64	16.24	58.55	17.2	64.1	17.12
GAF mid	34.76	14.58	55.34	19.7	56.89	18.24
GAF old	35.37	17.90	62.2	19.0	63.89	18.48	221.8	0.000	1.83	0.123
SCLGSI young	1.59	0.79	1.1	0.92	0.90	0.86
SCLGSI mid	1.33	0.71	0.99	0.74	0.92	0.65
SCLGSI old	1.14	0.54	0.63	0.55	0.54	0.49	110.9	0.000	2.069	0.089
PWI young	34.80	12.84	40.72	13.35	45.04	14.71
PWI mid	33.24	13.73	38.31	14.14	39.64	14.1
PWI old	38.16	12.65	45.63	11.24	45.26	13.49	45.54	0.000	1.803	0.146
SF-12m young	30.65	12.65	38.90	10.83	40.0	10.45
SF-12m mid	30.07	10.3	37.36	10.38	38.6	9.63
SF-12m old	31.85	12.1	40.73	11.07	41.62	11.73	82.4	0.000	0.184	0.938

GAF, Global Assessment of Functioning; SCLGSI, Symptom Check List-90 Global Symptom Index; PWI, Personal Wellbeing Index; SF-12m, measure of mental morbidity.

Symptom outcome by PD

Over time both comorbid and non-comorbid groups improved in response to treatment as demonstrated by comparison of discharge and follow-up scores after controlling for admission scores (Tables 3 and 4). In absolute terms, for the whole group, PD patients fared worse over time on the SCLGSI than those without PD (F (2,454) = 4.08, p < 0.05) and after taking baseline scores into account, discharge and follow-up scores were worse in the PD group (at discharge, β = 0.186, t = 4.23, p < 0.001; at follow up, β = 0.277, t = 5.03, p < 0.001).

Table 3.

Admission, discharge and follow-up scores of key variables for three age groups for Personality Disorder (PD) and no Personality Disorder (No PD) groups

		Admission	Discharge	Follow up
		PD	No PD	PD	No PD	PD	No PD	F
Measure		Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Time*PD	P value
GAF	Young	38.19	15.18	36.05	16.95	56.61	16.95	64.1	17.13	61.03	18.1	72.65	10.12	3.92	0.022
GAF	Middle	33.95	14.02	36.93	16.3	51.55	17.22	65.46	22.79	53.7	17.31	65.4	18.50	2.22	0.120
GAF	Old	31.26	16.24	41.22	18.72	57.84	17.82	68.42	19.13	57.87	17.12	72.47	17.1	0.719	0.487
SCLGSI	Young	1.69	0.79	1.33	0.77	1.29	0.93	0.472	0.55	1.12	0.89	0.283	0.318	5.19	0.009
SCLGSI	Middle	1.58	0.64	0.688	0.42	1.21	0.72	0.40	0.40	1.05	0.65	0.59	0.54	3.45	0.042
SCLGSI	Old	1.18	0.55	1.08	0.50	0.75	0.59	0.455	0.46	0.718	0.50	0.298	0.37	4.114	0.018
SF-12m	Young	29.53	9.37	33.84	13.99	36.4	10.49	46.04	8.52	37.88	10.87	46.06	5.99	1.769	0.176
SF-12m	Middle	28.24	10.1	34.94	9.5	35.2	9.91	43.12	9.62	36.07	8.53	45.33	9.38	0.376	0.675
SF-12m	Old	31.47	11.89	32.39	12.54	39.41	10.6	42.60	11.56	38.13	11.09	46.59	10.92	4.56	0.014
PWI	Young	33.44	12.16	38.75	14.25	38.48	12.51	47.14	13.94	42.11	14.87	53.5	10.62	1.405	0.247
PWI	Middle	30.98	13.61	39.84	12.29	35.61	14.41	46.23	10.1	36.21	13.7	49.69	10.18	0.752	0.433
PWI	Old	37.01	13.33	39.95	11.46	44.12	11.64	47.98	10.3	40.96	13.87	51.94	9.76	6.41	0.004

GAF, global assessment of functioning for three age groups; SCLGSI, Symptom Check List-90 Global Symptom Index; SF-12m Short Form health survey, Mental subscale); PWI, Personal Wellbeing Index.

Table 4.

Mean difference and standard error in scores between PD and non-PD groups adjusted for the effect of baseline scores by age group and overall

Dependent variable	PD-noPD	SE	P value
DSCLGSI
Young	0.531	0.144	0.000
Middle	0.058	0.163	0.724
Old	0.220	0.091	0.017
Total	0.298	0.070	0.000
FSCLGSI
Young	0.663	0.177	0.000
Middle	0.033	0.207	0.874
Old	0.390	0.089	0.000
Total	0.412	0.082	0.000
DGAF
Young	−8.29	4.15	0.049
Middle	−13.45	5.11	0.011
Old	−6.54	3.76	0.085
Total	−9.01	2.40	0.000
FGAF
Young	−13.17	4.18	0.003
Middle	−10.98	5.30	0.043
Old	−11.2	3.46	0.002
Total	−11.818	2.33	0.000
DSF-12m
Young	−7.87	2.42	0.002
Middle	−4.79	2.48	0.059
Old	−2.08	1.85	0.262
Total	−4.49	1.24	0.000
FSF-12m
Young	−7.46	2.56	0.005
Middle	−7.28	2.64	0.008
Old	−8.23	2.19	0.000
Total	−7.84	1.38	0.000
DSF-12p
Young	−5.419	1.62	0.001
Middle	0.716	1.19	0.550
Old	−0.021	1.54	0.989
Total	−0.886	.924	0.338
FSF-12p
Young	−5.83	2.13	0.008
Middle	2.67	2.73	0.333
Old	0.885	2.06	0.668
Total	−.226	1.32	0.864
DPWI
Young	−4.38	2.09	0.040
Middle	−3.62	2.38	0.136
Old	−1.98	1.69	0.246
Total	−3.34	1.16	0.004
FPWI
Young	−10.00	3.80	0.011
Middle	−10.11	4.03	0.016
Old	−9.83	2.47	0.000
Total	−9.75	1.84	0.000

DSCLGSI Symptom Check List General Symptom Index at discharge, FSCLGSI, at follow up; DGAF Global Assessment of Function at discharge, FGAF, at follow up; DSF-12m, Short Form Health Survey, mental subscale at discharge, FSF-12m, mental subscale at follow up; DSF-12p, Short Form Health Survey, physical subscale at discharge; FSF-12p, physical subscale at follow up; DPWI, Personal Wellbeing Index at discharge, FPWI, at follow up.

Symptom outcome by age and PD

In all three age groups at all time points symptoms were more severe and function was lower in the personality disordered group. In the young and the old, symptoms as measured by the SCLGSI improved significantly less over time in PD than in non-PD patients. The middle aged group with PD also improved less over time but this was not as significant. When the interaction of PD, time and age was included in the analysis it was found to be non-significant for all measures except SCLGSI (F (4,446) = 4.53, p < 0.005).

Function

Function by age

Function, measured using the GAF, improved in the whole cohort over time (mean = 35 on admission to 58 on discharge to 63 at follow up) and there was no significant difference between age groups (F (4,452) = 1.83, p = 0.123).

Function by PD

Taking the whole group, PD patients’ function improved significantly less than the non-comorbid group (F (2,454) = 5.37, p < 0.05).

Function by PD and age

In the young and the old, rate of change was less, and discharge and follow up scores were worse for PD than non-PD groups. In the middle age group, rate of change over time was not significantly different but the PD group showed significantly worse function at discharge and follow up (Tables 3 and 4).

Morbidity and well-being

On the SF-12 mental sub-scale all subjects improved over time from a baseline mean score of 31 to 38 on discharge to 40 on follow up and with no statistical difference between age groups (for the SF-12m, F (4,452) = 13.2, p = 0.938). Likewise for the PWI, mean scores improved from 35 to 40 on discharge which was maintained at follow up. There was no difference between age groups (F (4,420) = 1.8, p = 0.146). When comparing PD and non-PD groups well-being was significantly poorer in the PD group (for SF-12m, F (2,454) = 5.53, p < 0.05; and for PWI, F (2,422) = 6.43, p < 0.005).

Age, PD and morbidity

In the young, change over time was not significantly different between PD and non-PD groups but the PD group had significantly worse discharge and follow up scores. In the middle age group change over time was no different for PD and non-PD groups but the outcome was worse in the PD group at follow up for both measures. In the elderly, there was less improvement for the PD group in the SF-12 mental subscale and PWI and both were worse at follow up.

Severity of PD

Comparisons were made between simple (1–2 PDs) and complex (PD diagnoses in all three clusters) PDs on all outcome measures and in all but physical well-being there was a gradient with no PD having better outcomes than simple PD and complex (qua severe) PD having the worst outcome (Figure 1).

Figure 1.

Change over time for SCLGSI of total population comparing no PD (N = 77), simple PD (N = 99) and complex PD (N = 53). Time 1, admission; time 2, discharge; time 3, 6 months. Significant difference between test conditions (F (4,452) = 3.44, p < 0.05).

Social supports

Satisfaction with social supports was significantly worse in the PD groups in all age groups (young age group (t = 2.2, df = 72, p < 0.05), middle age group (t = 3.5, df = 55, p < 0.001), old age group (t = 2.4, df = 102, p < 0.05)); the number of social supports was also significantly lower in young (t = 2.03, df = 72, p < 0.05) and older (t = 2.6, df = 102, p < 0.05) PD patients.

Relapse and readmission

Unexpectedly, the presence of a personality disorder made no difference to the length of stay for the sample as a whole as well as each age group independently. There were higher rates of relapse (47% versus 26%, p < 0.005) and readmission in the PD group (33% versus 19.2%, p < 0.05) as a whole than in non-PD patients. Readmission rates for PD patients were higher for each age group but this was only significant for older patients (t = − 2.90, df = 102, p (2-tailed) < 0.05).

Confounding variables

We considered the influence of age, gender, education, primary Axis I diagnosis, baseline BPRS, PD, physical comorbidity, chronicity, social supports and coping strategies on the findings.

A regression was performed with these as independent variables and follow up SCLGSI as the dependent variable. Only PD complexity predicted (worse) outcome (F (2,212) = 24.7, p < 0.001). When this was repeated with follow up GAF as the dependent variable, PD (F (2,212) = 14.8, p < 0.001), Axis I diagnosis (F (2,212) = 3.6, p < 0.05), and less education (F (4,212) = 3.02, p < 0.02) predicted worse outcome. With readmission as dependent variable only chronicity (B = 1.35, df = 1, p = 0.006) and Axis I diagnosis (B = 0.994, df = 1, p = 0.016) predicted readmission.

Regressions were repeated for each of the three age groups. Symptomatic outcome (SCLGSI) at follow up was predicted only by the presence of a PD in young (F (1,64) = 7.76, p < 0.05) and elderly (F (1,83) = 14.8, p < 0.005) but not in the middle group. Functional outcome was predicted by chronicity in the young (F (2,64) = 3.4, p < 0.05), and middle (F (2,41) = 4.9, p < 0.05) and schizophrenia in the old (F (2,83) = 4.5, p < 0.02), and by PD in young (F (1,64) = 3.5, p < 0.05) and old (F(1,83) = 18.7, p < 0.000) age groups. In the elderly PD was a significant factor predicting worse outcome on all three measures of outcome. Data are available on request.

The Axis I diagnosis of schizophrenia and to a lesser extent depression predicted readmission in all age groups (B = 1.01, df = 1, p < 0.02); chronicity predicted readmission in young (B = − 21, df = 1, p < 0.000) and old (B = − 2.74, df = 1, p < 0.03); and PD predicted readmission in the elderly (B = − 1.6, df = 1, p < 0.05).

Discussion

We confirm the literature that comorbid personality disorder is common in the inpatient population though slightly less so in the older group [2]. Comorbidity is particularly common in certain groups such as those with drug and alcohol abuse, and eating disorders; however, it is seen in all diagnoses and across the demographic spectrum. Severe personality disorders were more common in the young group, and the old group had a higher proportion of patients without comorbid personality disorders. Outcomes for the middle group were less affected by the presence of the PD. This lends weight to reports that personality disorder tend to ‘mellow’ with age and that the old are less likely to meet criteria for more than one personality disorder [40] although this has not been confirmed by others [41], though how older age is defined varies, often as low as fifty [42].

In all three age groups symptoms and function were significantly worse in the PD groups at all time points, confirming our hypothesis. Moreover, the more severe the PD, the worse the outcome. Importantly, this did not mean that PD patients were less responsive to treatment. In some instances responses paralleled the non-PD group, as evidenced by change scores, but they remained more symptomatic and never reached the level of recovery of the non-PD group. These effects might explain the conflicting results of different studies, some reporting no effect of PD and others reporting a negative effect. However, when taking account of the baseline scores there was still a significant difference between the comorbid and the non-comorbid groups on several measures, especially post discharge. In addition, when other possible determinants of outcome were entered into a regression analysis PD was found to be the most significant determinant of outcome, with Axis I and chronicity making lesser contributions.

Our outcome results suggest that there is a more significant effect of personality disorder on symptomatic and functional outcomes in the young and to a lesser extent in the older groups than in the middle age group. Our findings fit with descriptions of waning of effects of personality in middle age [1] and a slight re-emergence in old age [43]. Perhaps the answer is that patients with personality disorders have more resources in middle age but start to lose them as they become older. While the number of social supports as measured by the SSQ-N in young and old comorbid patients was lower in those with PD than in non-PD patients, numbers were similar in the middle aged.

We confirmed that the negative effect of PD in all age groups was more pronounced following discharge, when previously non-significant differences became significant [44]. This is also seen in our relapse and readmission data where all PD patients, especially older PD patients, were significantly more likely to relapse and be readmitted to hospital. That Molinari and Marmion [45] found no difference in relapse rates between PD and non-PD elderly patients might be explained by their focus purely on affective disorders.

These effects were seen over a wide variety of diagnoses, not just the mood disorders. The absence of significant social support particularly in the young and the elderly PD groups may account for some of their deterioration post discharge where the presence of personality disorders certainly has a significant effect [44]. Our young and elderly PD groups had fewer social supports and less satisfaction with social supports than the middle group.

Contrary to our hypothesis, length of stay was not affected by the presence of PD. This is likely due to other factors such as the preference of the treating consultant, comorbid physical illness, severity of Axis I disorder, and accommodation availability.

Strengths of this study are that it was prospective, longitudinal, incorporated a battery of structured interviews, and included a moderately large patient population. Limitations include bias in the number of patients who were excluded or who refused, and by there being more women in the sample. Reasons for admission may differ by age, the young being more likely to be admitted purely for psychiatric disorders and the elderly for a combination of psychiatric and physical morbidity. Also, patients may have been readmitted to other facilities, but this is unlikely as public patients are only admitted to the hospital in their own area and all patients were interviewed regarding readmission at 6 and 12 months. Medication variables were not controlled for and scales were devised for young adults and have not been validated in the elderly. Finally, the same interviewer made both Axis I and II diagnoses. However, structured interviews reduced the risk of bias, and subsequent interviews were completed by blind raters. It could not be discounted that on discharge when the SCID-II was completed that some residual symptoms of the Axis I disorder could still be influencing the diagnosis of PD. However, it was felt to be minimal, moreover research has found that a diagnosis of PD even in the presence of Axis I pathology has significance [46].

Footnotes

Acknowledgements

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

References

1. Engels

Duijsens

Haringsma

van Putten

. Personality disorders in the elderly compared to four younger age groups: a cross-sectional study of community residents and mental health patients. J Personal Disord 2003; 17:447–459.

2. Stevenson

Boyce

Brodaty

Datyner

. The effect of age on prevalence, type and diagnosis of personality disorder in psychiatric inpatients. Int J Geriatr Psychiatry 2010. doi: 10.1002/gps.2645.

3. Corruble

Ginestet

Guelfi

. Comorbidity of personality disorders and unipolar major depression: a review. J Affect Disord 1996; 37:157–170.

4. Greenberg

Craighead

Evans

Craighead

. An investigation of the effects of comorbid Axis II pathology on outcome of inpatient treatment for unipolar depression. Journal of Psychopathology and Behavioral Assessment 1995; 17:305–321.

5. Keown

Holloway

Kuipers

. The prevalence of personality disorders, psychotic disorders and affective disorders amongst the patients seen by a community mental health team in London. SocPsychiatry Psychiatr Epidemiol 2002; 37:225–229.

6. Coolidge

Segal

Hook

Stewart

. Personality disorders and coping among anxious older adults. J Anxiety Disord 2000; 14:157–172.

7. Zimmerman

Rothschild

Chelminski

. The prevalence of DSM-IV personality disorders in psychiatric outpatients. Am J Psychiatry 2005; 162:1911–1918.

8. Abrams

Bromberg

. Personality disorders in the elderly: a flagging field of inquiry. Int J Geriatr Psychiatry 2006; 21: 1013–1017.

9. Jackson

Burgess

. Personality disorders in the community: a report from the Australian National Survey of Mental Health and Wellbeing. Soc Psychiatr Epidemiol 2000; 35:531–538.

10.

10. Tyrer

Gunderson

Lyons

Tohen

. Extent of comorbidity between mental state and personality disorders. J Personal Disord 1997; 11:242–259.

11.

11. Reich

Vasile

. Effect of personality disorders on the treatment outcome of Axis I conditions: an update. J Nerv Ment Dis 1993; 181:475–484.

12.

12. Reich

. The effect of Axis II disorders on the outcome of treatment of anxiety and unipolar depressive disorders: a review. J Personal Disord 2003; 17:387–405.

13.

13. Enns

Swenson

McIntyre

Swinson

Kennedy

, CANMAT Depression Work Group. Clinical guidelines for the treatment of depressive disorders. VII. Comorbidity. Can J Psychiatry 2001; 46:S77–90.

14.

14. Berlim

Turecki

. Definition, assessment, and staging of treatment-resistant refractory major depression: a review of current concepts and methods. Can J Psychiatry 2007; 52:46–54.

15.

15. Hirschfeld

. Personality disorders and depression: comorbidity. Depress Anxiety 1999; 10:142–146.

16.

16. Mulder

. Depression and personality disorder. Curr Psychiatry Rep 2004; 6:51–57.

17.

17. Smith

Deutsch

Schwartz

Terkelsen

. The role of personality in the treatment of schizophrenic and schizoaffective disorder inpatients: a pilot study. Bull Menninger Clin 1993; 57: 88–99.

18.

18. Wancata

Benda

Windhaber

Nowotny

. Does psychiatric comorbidity increase the length of stay in general hospitals? Gen Hosp Psychiatry 2001; 23:8–14.

19.

19. Donat

. Personality traits and psychiatric rehospitalization: a two-year follow up. J Pers Assess 1997; 68:703–711.

20.

20. Agbayewa

. Occurrence and effects of personality disorders in depression: are they the same in the old and young? Can J Psychiatry 1996; 41:223–226.

21.

21. Rao

. Does personality disorder influence the likelihood of in-patient admission in late-life depression? Int J Geriatr Psychiatry 2003; 18:960–961.

22.

22. Sadavoy

. The effect of personality disorder on Axis I disorders in the elderly. Handbook of counselling and psychotherapy with older adults. Hoboken, NJ: Wiley, 1999: 397–413.

23.

23. First

Spitzer

Gibbon

Williams

. Structured Clinical Interview for DSM-IV Personality Disorders, (SCID-II). Washington, DC: American Psychiatric Press, 1997.

24.

24. Zanarini

Skodol

Bender

. The Collaborative Longitudinal Personality Disorders Study: reliability of Axis I and II diagnoses. J Personal Disord 2000; 14:291–299.

25.

25. Thompson

Gallagher

Czirr

. Personality disorder and outcome in the treatment of late-life depression. J Geriatr Psychiatry 1988; 21:133–153.

26.

26. Folstein

Folstein

McHugh

. ‘Mini-mental state’. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12:189–198.

27.

27. Dubois

Slachevsky

Litvan

Pillon

. The FAB: a Frontal Assessment Battery at bedside. Neurology 2000; 55:1621–1626.

28.

28. Brodaty

Pond

Kemp

. The GPCOG: a new screening test for dementia designed for general practice. J Am Geriatr Soc 2002; 50:530–534.

29.

29. Sunderland

Hill

Mellow

. Clock drawing in Alzheimer's disease: a novel measure of dementia severity. J Am Geriatr Soc 1989; 37:725–729.

30.

30. Piedmont

Sherman

Williams

. A first look at the Structured Clinical Interview for DSM-IV Personality Disorders Screening Questionnaire: more than just a screener? Measurement and Evaluation in Counseling and Development 2003; 36:150–160.

31.

31. Fogel

Westlake

. Personality disorder diagnoses and age in inpatients with major depression. J Clin Psychiatry 1990; 51: 232–235.

32.

32. Vallejo

Jordan

Diaz

Comeche

Ortega

. Psychological assessment via the Internet: a reliability and validity study of online (versus paper-and-pencil) versions of the General Health Questionnaire-28 (GHQ-28) and the Symptoms Check-List-90-Revised (SLR-90-R). J Med Internet Res 2007; 9:1–10.

33.

33. Endicott

Spitzer

Fleiss

Cohen

. The Global Assessment Scale: a procedure for measuring overall severity of psychiatric disturbance. Arch Gen Psychiatry 1976; 33:766–771.

34.

34. Jenkinson

Layte

Jenkinson

. A shorter form health survey: can the SF-12 replicate results from the SF-36 in longitudinal studies? J Public Health Med 1997; 19:179–186.

35.

35. International Wellbeing Group. Personal Wellbeing Index, fourth. Melbourne: Australian Centre on Quality of Life, Deakin University; 2006.

36.

36. Sarason

Sarason

Shearin

Pierce

. A brief measure of social support: practical and theoretical implications. J Soc Pers Rels 1987; 4:497–510.

37.

37. Lukoff

Liberman

Nuechterlein

. Symptom monitoring in the rehabilitation of schizophrenic patients. Schizophr Bull 1986; 12:578–602.

38.

38. Dingemans

Linszen

Lenior

Smeets

. Component structure of the expanded Brief Psychiatric Rating Scale (BPRS-E). Psychopharmacology 1995; 122:263–267.

39.

39. Carver

. You want to measure coping but your protocol's too long: Consider the Brief COPE. Int J Behav Med 1997; 4:92–100.

40.

40. Molinari

Marmion

. Personality disorders in geropsychiatric outpatients. Psychol Rep 1993; 73:256–258.

41.

41. Condello

Padoani

Uguzzoni

Caon

De Leo

. Personality disorders and self-perceived quality of life in an elderly psychiatric outpatient population. Psychopathology 2003; 36:78–83.

42.

42. Alnaes

Torgersen

. DSM-III symptom disorders (Axis I) and personality disorders (Axis II) in an outpatient population. Acta Psychiatr Scand 1988; 78:348–355.

43.

43. Reich

Nduaguba

Yates

. Age and sex distribution of DSM-III personality cluster traits in a community population. Compr Psychiatry 1988; 29:298–303.

44.

44. Morse

Lynch

. A preliminary investigation of self-reported personality disorders in late life: prevalence, predictors of depressive severity, and clinical correlates. Aging Ment Health 2004; 8:307–315.

45.

45. Molinari