Accepted Manuscript

Interview-based ratings of DSM-IV Axis II/DSM-5 section II

personality disorder symptoms in consecutively admitted
insomnia patients: A comparison study with consecutively
admitted psychotherapy patients matched on age and gender

Antonella Somma, Sara Marelli, Laura E. Giarolli, Cesare Maffei,

Luigi Ferini-Strambi, Andrea Fossati

PII: S0010-440X(18)30159-7
DOI: doi:10.1016/j.comppsych.2018.09.005
Reference: YCOMP 52033
To appear in: Comprehensive Psychiatry

Please cite this article as: Antonella Somma, Sara Marelli, Laura E. Giarolli, Cesare
Maffei, Luigi Ferini-Strambi, Andrea Fossati , Interview-based ratings of DSM-IV Axis
II/DSM-5 section II personality disorder symptoms in consecutively admitted insomnia
patients: A comparison study with consecutively admitted psychotherapy patients matched
on age and gender. Ycomp (2018), doi:10.1016/j.comppsych.2018.09.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 1

Interview-Based Ratings of DSM-IV Axis II/DSM-5 Section II Personality Disorder Symptoms in

Consecutively Admitted Insomnia Patients: A Comparison Study with Consecutively Admitted

Psychotherapy Patients Matched on Age and Gender.

Antonella Somma a, c, Sara Marelli a, b, Laura E. Giarolli b, c, Cesare Maffei a, c,

Luigi Ferini-Strambi a, b, and Andrea Fossati a, c

PT
RI
a: Faculty of Psychology, Vita-Salute San Raffaele University, Milan, Italy

SC
b: Sleep Disorder Unit, San Raffaele Hospital, Milan, Italy

c: Clinical Psychology and Psychotherapy Unit, San Raffaele Hospital, Milan, Italy
NU
MA
ED
PT
CE
AC

Correspondence concerning this article should be sent to Andrea Fossati, Clinical Psychology and

Psychotherapy Unit, San Raffaele Turro Hospital, via Stamira d’Ancona 20, 20127, Milan, Italy.

Telephone: +39 02 26433241; fax: +39 02 26433408; e-mail: [email protected]

 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 2

Abstract

Background. Selected personality features may represent important predisposing as well as

perpetuating factors for insomnia, and previous studies stressed the importance to assess personality

disorders in insomnia patients.

Methods. In order to evaluate the relationships between DSM-IV axis II/DSM-5 Section II

Personality Disorders (PDs) and insomnia, a sample of 171 consecutively admitted insomnia

PT
patients and a sample of 171 psychotherapy patients, matched on age and gender were administered

RI
the Italian translation of the Structured Clinical Interview for DSM-IV Axis II Personality

SC
Disorders, Version 2.0 (SCID-II). Among insomnia patients, 52.0% (n = 89) received at least one

DSM-IV axis II/DSM-5 Section II PD diagnosis according to SCID-II assessment.

NU
Results. Any PD base rate estimate in our insomnia patient sample was significantly and markedly

higher than the median and mean base rate estimates for any PD in the general population. Within-
MA

group analyses showed that Narcissistic, Not otherwise specified PD, Histrionic PD, and Borderline

PD represented the most frequently diagnosed-both dimensionally and categorically-DSM-IV axis

ED

II/DSM-5 Section II PD features in our insomnia patient sample. When continuously-scored PDs

were considered, insomnia patients showed a significantly lower number of Paranoid and
PT

Borderline PD features than psychotherapy patients; however, the corresponding effect size
CE

estimates suggested that these differences were modest. None of the categorically-scored PDs

significantly differentiated insomnia patients from psychotherapy patients.

AC

Conclusions. As a whole, our findings seemed to suggest that personality dysfunction may play a

role in insomnia, while stressing the need for a dimensional approach to the assessment of

maladaptive personality traits even in insomnia patients.

Keywords: insomnia; personality disorders; controlled study; psychotherapy patients; dimensional

models.
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 3

1 Introduction

Insomnia is one of the most common sleep disorders, affecting 9–10% of the general

population in the United States [1], which may have a negative impact on psychological wellbeing

and quality of life [2] and may even predispose to other disorders, e.g., anxiety, depression,

substance use disorders [3]. Insomnia is defined as an ongoing difficulty in getting to sleep, staying

asleep, waking up too early or bad quality of sleep, accompanied by significantly impaired daytime

PT
functioning [4]. A model describing how (relatively stable) predispositions, precipitants and

RI
perpetuating factors interact to induce and maintain insomnia may help explaining the

SC
pathophysiology of many insomnia features [1; 5-6]. For instance, van de Laar and colleagues [1]

acutely pointed out that selected personality features may represent important predisposing and/or
NU
perpetuating factors for insomnia. Notably, high levels of neuroticism – i.e., the tendency to

experience a wide range of negative emotions [7] – were shown to increase the mood depressing
MA

effect of experimentally induced insomnia [8].

Based on these considerations, it was not surprising to observe a substantial body of

ED

literature that tried to assess the relationships between both general and pathological personality

traits, and insomnia. In their extensive review of the literature on this topic, van de Laar and
PT

colleagues [1] reported that in general insomnia patients showed elevations on measures of
CE

neuroticism and perfectionism personality traits. Recently, Dørheim and colleagues [9] documented

in a sample of 3752 pregnant women that neuroticism and agreeableness were associated with
AC

insomnia in pregnancy, although no personality trait predicted post-partum insomnia. However,

they pointed out that the exact causal relations remain controversial. Indeed, certain personality

features could play a predisposing role towards insomnia; however, the same features could be a

consequence of the negative impact of insomnia on daytime functioning [1]. Moreover, there are

several different theories to explain personality but – unfortunately – there is no unifying concept

[1]. To complicate this issue further, the exclusion of psychiatric disorders in an insomnia group

‘normalizes’ personality measurements [1]. These considerations led van de Laar and colleagues [1]
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 4

to stress the importance to assess personality disorders – as well as other psychiatric disorder – in

future studies [1].

Based on extant meta-analyses of the published literature on sleep disorders and mental

disorders [10-11], a total of 10 studies were carried out on sleep disorders in personality disorders

(PDs); however, they focused almost exclusively (n = 9; 90.0%) on Borderline PD. Moreover, they

were small-sample studies; for instance, the five studies on sleep disorders in Borderline PD that

PT
were included in Baglioni and colleagues’ [11] meta-analyses provided data only on a total of 89

RI
subjects with Borderline PD and 85 control subjects. Interestingly, Baglioni and colleagues [11]

SC
reported that Borderline PD was characterized by reduced sleep continuity and depth (i.e., shorter

duration of sleep Stage 1, and longer duration of sleep Stage 2 and slow wave sleep) and heightened
NU
rapid-eye movement (REM) sleep pressure (defined by shorter REM latency, increased REM

density, and longer duration of REM sleep), although these findings were definitively non-specific.
MA

Studies specifically examining PDs and chronic insomnia suggested that Avoidant,

Dependent, and Obsessive–Compulsive PDs (i.e., DSM-IV axis II/DSM-5 Section II Cluster C PDs;
ED

[12-13]) and sub-threshold features of these disorders may co-occur with chronic insomnia and poor

sleep quality more than other PD clusters [14-16], even in adult subjects with chronic insomnia and
PT

hypnotic medication dependence [17]. Some studies suggested poor sleep quality even among
CE

Cluster A PD subjects [18]. Notwithstanding these relevant findings suggesting that insomnia

patients may show elevations on PD symptoms that are characterized by perfectionism (i.e.,
AC

Obsessive-Compulsive PD) and/or neuroticism (e.g., Avoidant PD), the majority of these studies

relied exclusively on self-report measures to assess PDs, with no control sample.

Although previous studies showed the importance of the relationship between insomnia and

PDs, research on the prevalence of PDs (including Cluster B PDs) in patients with insomnia, is still

limited [17]. Assessing PDs in patients suffering from insomnia may be helpful in tailoring standard

treatments to the needs of patients suffering from both insomnia and PDs; moreover, the value of

identifying PD features could be the development of preventive sleep education [e.g., 17].
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 5

1.1 The Present Study: Aims and Hypotheses

Starting from these considerations, we aimed at testing the following research questions in a

sample of consecutively admitted insomnia patients: (1) Do insomnia patients show PD base rate

estimates that are significantly larger than the median/mean PD base rate estimates that were

reported for the general population in epidemiology studies?; (2) Are there selected DSM-IV axis

II/DSM-5 Section II PD that are significantly more frequent than other PDs in insomnia patients?;

PT
(3) Do insomnia patients show significant differences on PD frequency estimates when compared to

RI
a sample of psychotherapy patients of the same size who were matched on age and gender?

SC
Based on previous findings, we expected: (1) Insomnia patient sample PD base rate

estimates to be significantly higher than those reported for the general population [1; 9-11]; (2)
NU
Features of Cluster C and A personality disorders to be associated with insomnia [14-18]; (3)

Insomnia patients to show significantly lower rates of overall pathological personality traits than
MA

psychotherapy patients. The present research contributes to the field representing the first attempt at

systematically evaluating (e.g., using a semi-structured interview and a comparison group) the base
ED

rates of PDs in insomnia patients, representing the first step in assessing the need for developing

therapeutic techniques and coping strategies for individuals suffering from both insomnia and PDs.
PT

2 Methods
CE

2.1 Participants

The current study was based on insomnia participants (n = 171) and age- and sex-matched
AC

psychotherapy patient participants (n = 171). Insomnia participants were consecutively admitted to

the Sleep Disorders Center of the San Raffaele Turro Hospital, after an initial assessment conducted

by a neurologist expert in sleep medicine. All participants with insomnia disorder met the clinical

criteria for chronic insomnia disorder according to International Classification of Sleep Disorders

(ICSD-3; [4]). To improve the accuracy of insomnia diagnosis, all participants completed the Italian

translations of the Pittsburg Sleep Quality Index [PSQI; 19] and the Insomnia Severity Index [ISI;

20]; mean values were 11.91 (SD = 3.07; Cronbach’s  = .83) and 18.11 (SD = 4.05; Cronbach’s 
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 6

= .79) for PSQI and ISI, respectively. All participants underwent two consecutive nights of

polysomnographic (PSG) recording to ensure correct diagnostic classification. All patients spent an

adaptation night before the PSG recording. Lights-out time was based on the individual’s usual

bedtime and ranged between 11:00 and 11:30 pm. The following signals were recorded:

electroencephalogram (six channels, including C3 or C4 and O1 or O2, referred to the contralateral

mastoid); electrooculogram; electromyography (EMG) of the submentalis muscle; EMG of the right

PT
and left tibialis anterior muscles; electrocardiogram (one derivation) according to the American

RI
Academy of Sleep Medicine [21] scoring criteria. The sleep respiratory pattern of each patient was

SC
monitored using oral and nasal airflow thermistors and/or nasal pressure cannula, thoracic and

abdominal respiratory effort strain gauge, and by monitoring oxygen saturation (pulse-oximetry).
NU
Sleep stages were scored following standard criteria [e.g., 22] on 30-s epochs using the sleep

analysis software Hypnolab 1.2 (SWS Soft, Italy). Inclusion criteria were as follows: the absence of
MA

dementia, psychiatric disorders different from personality disorders, other sleep disorders and

neurological comorbidities. In addition, patients had to be drug-free for at least 2 months before
ED

their inclusion in the study or had a stable therapy for 6 months.

Ninety-six (56.1%) participants were female and 75 (43.9%) were male; participants’ mean
PT

age was 40.88 years, SD = 10.48 years (range: 18-67 years). Subsequently, we relied on ‘e1071’ R
CE

package [23] to systematically match the insomnia patient sample with a large psychotherapy

patient sample (N = 507) based on the composition of age and gender in the insomnia participant
AC

sample. All psychotherapy participants were admitted to the Clinical Psychology and

Psychotherapy Unit in order to receive psychotherapy treatment for interpersonal difficulties and/or

problems with behavior and emotional regulation on a strictly voluntary basis. Potential participants

were screened for the following exclusionary criteria: (1) age less than 18 years; (2) IQ less than 80;

(3) diagnosis of schizophrenia, schizoaffective disorder, schizophreniform disorder, or delusional

disorder according to DSM-IV diagnostic criteria; (4) diagnosis of dementia or organic mental
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 7

disorder according to DSM-IV diagnostic criteria; and (5) education level lower than elementary

school. All participants in the current research passed this screening procedure.

2.2 Procedure

All participants volunteered to take part in the study after being presented with a detailed

description of all aspects of the study and signed a written informed consent. All participants were

at least 18 years old when the study was carried out. Participants were administered all measures as

PT
part of their routine clinical assessment; IRB approval was obtained for all aspects of this study.

RI
Participants were treated in accordance with the Ethical Principles of Psychologists and Code of

SC
Conduct, and none of them received an incentive, either directly or indirectly for participating. The

SCID-II was administered by expert (Doctor of Psychology) clinicians who received extensive
NU
training in SCID-II assessment; clinicians were kept blind to the aims of the study and to the

subject’ study group.

MA

2.3 Measures

2.3.1 Structured Clinical Interview for DSM-IV Axis II Personality Disorders, Version 2.0 [24]. The
ED

SCID-II is a 140-item semi-structured interview designed to provide both a categorical and

dimensional (i.e., number of criteria) assessment of DSM-IV Personality Disorders (PDs). The
PT

SCID-II was preceded by administration of its self-report screening questionnaire (PQ). The
CE

interrater reliability and internal consistency of the Italian translation of the SCID-II in clinical

participants were assessed in a previous study [25]. Although the SCID II was developed to asses
AC

DSM-IV axis II PDs, we considered it as a measure of DSM-5 Section II PDs because the PD

diagnostic criteria in the DSM-5 Section II were retained with no changes from the DSM-IV axis II

(APA, 2013). Thus, in the present study, we considered only the 10 PD diagnoses that were retained

in DSM-5 Section II [13].

Although the DSM-5 Section II retained a categorical model of PD assessment, an

impressive amount of taxometric data documented that PDs are dimensional in nature [26-27], and

that the use of arbitrary diagnostic thresholds may result in severe loss of clinical (and research)
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 8

information [28]. Thus, in the present study we relied both on continuously-scored (i.e., number of

symptoms) and categorically-assessed PDs for within-group and between-group comparisons.

Clinicians who administered the interviews were kept blind to the aim of the study, as well

as to the subject’s condition (insomnia vs. psychotherapy group). Because six Doctor of Psychology

(PsyD) expert clinicians trained in administering the SCID-II participated in the present study, we

used a pairwise interview design in order assess the inter-rater reliability of the SCID-II diagnoses.

PT
In the present study, the interrater reliability of SCID-II diagnoses was assessed on the first 60

RI
(17.5%) consecutively admitted participants. For each of the first 60 participants, two interviewers

SC
were randomly extracted and assigned the role of interviewer and independent rater, respectively;

each clinical psychologist acted the same number of times as interviewer or independent rater. In
NU
the present study, the ICC values for the individual SCID-II PD dimensional counts ranged from .83

(Antisocial PD) to .99 (Obsessive-Compulsive PD), median ICC value = .92, SD = .05, all ps <.001.
MA

The chance-corrected agreement (i.e., Cohen κ coefficient value) on any PD diagnosis was .92, p

<.001, whereas a Cohen κ value for the individual SCID-II PD diagnosis ranged from .83
ED

(Antisocial PD) to .99 (Obsessive-Compulsive PD), median ICC value = .92, SD = .05, all ps <.001.

Cohen κ value for SCID-II Mixed PD diagnosis was .86, p <.001.

PT

2.4 Data analysis

CE

Shapiro-Wilk and Kolmogorov-Smirnoff tests were used for checking normality of the

continuously-scored DSM-IV axis II/DSM-5 Section II PDs scores. If the assumption of normality is
AC

violated, non-parametric statistics will be employed. Mann-Whitney U tests were used to compare

continuously-scored DSM-IV Axis II/DSM-5 Section II PD means between insomnia patients and

psychotherapy patients; Rosenthal’s [29] r was used as an effect size measure for Mann-Whitney

test. The nominal significance level (i.e., p <.05) for Mann-Whitney U tests was corrected

according to the Bonferroni procedure (i.e., p <.0045). Following Newcombe’s [30] indications, we

relied on relative risk (RR) to evaluate both significance (i.e., p <.0042) and effect size of the

differences in PD proportions that were observed between insomnia patients and psychotherapy
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 9

patients. Within each sample, repeated measure Dunn-Bonferroni post-hoc contrasts were used to

compare continuously-scored DSM-IV Axis II/DSM-5 Section II PD means and categorically-

scored DSM-IV Axis II/DSM-5 Section II PDs frequencies. Friedman repeated-measure ANOVA

followed by Dunn-Bonferroni post-hoc comparisons were run to assess mean rank differences in

continuously-scored PDs within both insomnia patients and psychotherapy patients; since a total of

20 ANOVAs were performed, the Bonferroni nominal significance level (i.e. p <.05) was corrected

PT
according to the Bonferroni procedure and set at p <.0025. Cochran Q values were used to evaluate

RI
the existence of significant differences among PD base rate estimates in both samples. Binomial test

SC
with two-tailed p-value was used to carry out statistical comparisons between PD base rate

estimates in our samples and corresponding median/mean prevalence estimates for the general
NU
population. The relationships between the polysomnographic data, PSQI scores, ISI scores, and

continuously scored PDs was assessed using Spearman r correlation coefficient ().
MA

3 Results

Demographic characteristics of the insomnia participant sample and age- and sex-matched
ED

psychotherapy patient participants are listed in Table 1. Continuously-scored DSM-IV axis II/DSM-

5 Section II PDs were not normally distributed in both insomnia patient sample, Shapiro-Wilks
PT

statistic median value = .60, min. = .07 (Antisocial PD), max. (overall number of PD criteria) = .96,
CE

all ps <.001, and psychotherapy patient sample, Shapiro-Wilks statistic median value = .68, min. =

.21 (Schizotypal PD), max. (overall number of PD criteria) = .95, all ps <.001. Similar findings
AC

were observed when we relied on Kolmogorov-Smirnoff test for assessing PD score normal

distribution. Thus, we relied on nonparametric statistics for all comparisons.

When we considered the relationships between the polysomnographic data, PSQI scores, ISI

scores, and continuously scored PDs, as a whole, no significant association was observed. In our

insomnia patient sample, PSG data and continuously scored PDs were not significantly associated,

median  = .02, SD = .07. Similarly, no significant association was observed between DSM-IV Axis
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 10

II/DSM-5 Section II PD continuous ratings and PSQI scores (median  = .04, SD = .11), and ISI

scores (median  = .01, SD = .06).

3.1 Median/Mean Personality Disorder Base Rate Comparisons: Insomnia Patient vs. General

Population Epidemiological Studies

Comparisons between PD base rate estimates in our insomnia patient sample and

corresponding median/mean prevalence estimates for the general population [31] are summarized in

PT
Table 2. Interestingly, the base rate estimate for any PD diagnosis in our insomnia patients was

RI
significantly lower than both Torgersen’s [31] median value (65.6%; z = -3.74, two-tailed p <.001)

SC
and mean value (64.4%; z = -3.39, two-tailed p <.001) for psychiatric samples.

Of course, the base rate estimate (72.5%) for any PD diagnosis in our psychotherapy patient
NU
sample was significantly higher than median (z = 24.47, two-tailed p <.001) and mean (z = 25.70,

two-tailed p <.001) base rate estimates for the general population [31]. When we compared the any
MA

PD base rate estimate in our psychotherapy patient sample with the median value for psychiatric

samples [31], we observed no significant difference, z = 1.90, two-tailed p >.05; however, our base
ED

rate estimate was significantly higher than the mean base rate estimate for any PD diagnosis in
PT

psychiatric samples, z = 2.21, two-tailed p <.05.

3.2 Frequency of DSM-IV axis II/DSM-5 Section II PDs in the Insomnia Patient Sample:
CE

Within Group Comparisons and Comparisons with the Age-and-Gender-Matched

Psychotherapy Patient Sample

AC

Considering the individual DSM-IV axis II/DSM-5 Section II PDs and the number of DSM-

IV axis II/DSM-5 Section II PD criteria in the insomnia patient sample and in the psychotherapy

patient sample, the descriptive statistics, within-group comparisons, and between-group

comparisons are summarized in Table 3. For ease of presentation, the DSM-IV axis II/DSM-5

Section II PDs are listed in DSM-5 Section II order. In both samples, the sum of the frequency of

subjects who received specific PD diagnoses exceeded the frequency of subjects who received at

least one PD diagnosis because of multiple PD diagnoses. Means and frequencies with different
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 11

superscripts were significantly different (i.e., p <.05) in repeated measure Dunn-Bonferroni post-

hoc contrasts. For instance, Narcissistic PD showed a significantly higher mean number of PD traits

than all other PDs, whereas the mean number of Paranoid PD and Schizoid PD was not significantly

different.

Friedman repeated-measure ANOVA results showed that the ten continuously-scored PDs

had significant mean rank differences within both insomnia patients, H(9) = 415.41, p <.001,

PT
2Friedman = .27, and psychotherapy patients, H(9) = 309.80, p <.001, 2Friedman = .20. As it can be

RI
seen in Table 3, Dunn-Bonferroni post-hoc comparisons indicated that continuously-scored

SC
Narcissistic PD showed a significantly higher mean rank value than all other continuously scored

PDs in both samples. Similar considerations held for categorically-rated DSM-IV axis II/DSM-5
NU
Section II PD diagnoses. Cochran Q values suggested the existence of significant differences

among PD base rate estimates both in the insomnia patient sample, Q = 149.47, p <.001, I2 = 93.3,
MA

and in the psychotherapy patient sample, Q = 158.79, p <.001, I2 = 93.7.

4 Discussion
ED

Confirming and extending previous evidence [1; 10-11], our findings seemed to suggest that

personality dysfunction, at least as it was operationalized in the DSM-5 Section II, may play a role
PT

in insomnia, although it is still controversial if it represents a predisposing factor a perpetuating

CE

factor, or simply a consequence of the impact of insomnia on daily life [1; 5]. It should be observed

that our data were based on expert clinician’s ratings of PDs who administered the SCID-II blind to
AC

study aims and subject status (i.e., insomnia patient vs. psychotherapy patient). Moreover, we relied

on a psychotherapy patient control group of the same size, whose members were carefully matched

on age and gender. Finally, to our knowledge our study represented the first attempt at considering

PD prevalence rates in insomnia patients also in comparison with epidemiological data [31].

4.1 Do insomnia patients show PD base rate estimates that are significantly larger than the

median/mean PD base rate estimates that were reported for the general population in

epidemiology studies?
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 12

According to our findings more than 50% of the insomnia patients received at least one

DSM-IV axis II/DSM-5 Section II PD diagnosis; moreover, our insomnia patients met on average

roughly six PD criteria. It should be observed that in our study dysfunctional personality traits were

extensively assessed by expert clinicians using a semi-structured interview (i.e., the SCID-II) who

were blind to patient’s group membership, rather than relying on patients’ self-reports.

In our study, the any PD base rate estimate in our insomnia patient sample was significantly

PT
and markedly higher than the median (proportion difference = 40.0%) and mean (proportion

RI
difference = 41.0%) base rate estimates for any PD in the general population [31]. When compared

SC
to epidemiological mean/median prevalence data for psychiatric samples [31], our insomnia

patients showed significantly lower rates of overall dysfunctional personality traits, although the
NU
differences in proportions were in the 12%-14% range, suggesting small differences. Interestingly

our psychotherapy control group was at least as characterized by dysfunctions in personality as, not
MA

to say more, than the average (i.e., mean/median) psychiatric sample in international clinical

populations [33].
ED

Thus, in terms of general amount of personality dysfunction, our data suggest that insomnia

patients are sharply different from the general population, while manifesting a significantly, albeit
PT

slightly lower overall amount of maladaptive personality traits than psychotherapy/psychiatric

CE

patients. Consistent with previous studies [1; 10-11], our findings indicate the importance of

assessing personality and its pathology in insomnia patients.

AC

4.2 Are there selected DSM-IV axis II/DSM-5 Section II PD that are significantly more

frequent than other PDs in insomnia patients?

Within-group analyses showed that Narcissistic, Not otherwise specified/with other

specification PD, Histrionic PD, and Borderline PD represented the most frequently diagnosed -

both dimensionally and categorically - DSM-IV axis II/DSM-5 Section II PD features in our

insomnia patient sample; rather, Paranoid, Schizoid, Schizotypal, and Antisocial PD features were

diagnosed significantly less than most of the other DSM-IV axis II/DSM-5 Section II PDs. These
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 13

findings seemed to be at least partially consistent with previous reports [14-16; 18] based on self-

report measures which documented that specific PDs have different base rate estimates in insomnia

patients. We are aware that our data are somewhat at variance with the findings of self-report

studies, which documented and excess of Cluster C PDs in insomnia patients [14-17]. Although

sampling differences may partially explain these differences, it is known that relying on self-reports

and interview-based measures to assess DSM-IV axis II/DSM-5 Section II PDs is likely to result in

PT
very different estimates, mostly because of poor convergent validity of the PD instruments [34].

RI
Although no “gold standard” has been identified for PD assessment, semi-structured interviews

SC
represent a major resource for PD evaluation [34].

Interestingly, both Borderline and Narcissistic PDs have been consistently linked to emotion
NU
dysregulation [35-36], which represents a dysfunctional personality features that has been

empirically linked to insomnia [37]. Moreover, previous studies [e.g., 38], showed the existence of
MA

a relationship between specific PDs characterized by high levels of antagonism (i.e., behaviors that

put the individual at odds with other people; 13) and insomnia, also among forensic psychiatric
ED

inpatients [e.g., 38-39]. Finally, Sabouri and colleagues [40] examined the associations between

dark triad traits (i.e., narcissism, Machiavellianism, and psychopathy), and sleep disturbances in a
PT

sample of young adults, and found that Machiavellianism and psychopathy were associated with
CE

sleep disturbances. As a whole, these findings seemed to stress the importance of studying the

relationship between antagonistic personality traits and insomnia, while suggesting the potential
AC

usefulness of adopting a dimensional perspective on personality assessment [28].

4.3 Do insomnia patients show significant differences on PD frequency estimates when

compared to a sample of psychotherapy patients of the same size who were matched on age

and gender?

In our study, the overall amount of personality pathology, both continuously-scored (i.e., n.

of PD symptoms) and categorically-scored (i.e., any PD diagnosis base rate), was significantly

larger in the psychotherapy patient control group than in the insomnia patient sample; however, the
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 14

size of these differences may be considered small by conventional standards [32-33]. Interestingly,

the insomnia patient pattern of within-group differences among both continuously-scored and

categorically-assessed PDs was reproduced with minimal differences also in our psychotherapy

patient sample. Although this result may have been influenced by sampling participants in both

groups from the same hospital, it seems to indicate that insomnia patients are pretty similar to

psychotherapy patients in terms of the shape of the maladaptive personality traits profile.

PT
As a whole, between-group contrasts on individual PDs seemed to support the hypothesis

RI
that insomnia patients were simultaneously similar to psychotherapy patients and different from

SC
general population data [31] only on selected PDs. When continuously-scored PDs were

considered, insomnia patients showed a significantly lower number of Paranoid and Borderline PD
NU
features than psychotherapy patients; however, the corresponding effect size estimates suggested

that these differences were modest [32]. Interestingly, none of the categorically-scored PDs
MA

significantly differentiated insomnia patients from psychotherapy patients. Although this finding

may reflect the loss of statistical power that occurs when continuous variables are discretized [41],
ED

it seems also to indicate the frequency of extreme scorers (i.e., subjects scoring above the PD

diagnostic threshold) may not be different in insomnia patients and psychotherapy patients for all
PT

DSM-IV axis II/DSM-5 Section II PDs, at least when they were assessed using the SCID-II.
CE

4.4 Conclusions

Only the frequency of categorically-scored DSM-IV axis II/DSM-5 Section II Borderline,

AC

Histrionic, and Narcissistic PDs significantly discriminated insomnia patients from general

population base rate estimates [31]; at the same time, these PDs showed approximately the same

base rates in insomnia patients and psychotherapy patients. In our opinion, this finding seemed to

confirm and extend previous evidence on the association between Borderline PD and sleep

disorders [10-11], suggesting that Borderline, Histrionic and Narcissistic PDs may represent core

personality problems in insomnia patients, making them akin to a sample of psychiatry patients

voluntarily seeking for psychotherapy treatment because of personality problems (i.e., our
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 15

psychotherapy patient sample). Of course, we do not mean to say that personality disorders always

occurs in insomnia patients or that it has a causal role in insomnia disorders. Rather, we suggest that

maladaptive personality traits should be routinely assessed and taken into account in insomnia

patient treatment, independent from whether it has a predisposing or perpetuating role in insomnia

[1]. Indeed, it would be consistent with integrated theories and models about the etiology and

pathophysiology of insomnia [e.g., 6]. For instance, the hyperarousal model of insomnia

PT
emphasizes an interplay between psychological (e.g., personality traits) and physiological factors in

RI
the etiology and perpetuation of chronic insomnia [e.g., 6]. Moreover, these aspects may be

SC
particularly relevant considering that previous studies found an association between personality

characteristics and insomnia treatment outcomes [e.g., 1]. Future studies may examine the effect of
NU
tailored treatment strategies insomnia patients, for instance taking into account the role of PD

comorbidities and/or adjusting for personality characteristics.

MA

Finally, our data seemed to be in line with studies suggesting the need for adopting

dimensional models to correctly assess dysfunction in personality [28]. Indeed, a number of

ED

significant, non-negligible associations were observed among continuously-scored PDs both in

insomnia patients and psychotherapy patients (although qualitative differences between the two
PT

groups were observed in these correlation patterns), suggesting that DSM-IV axis II/DSM-5 Section
CE

II PDs represent poorly separated constructs [28]. Moreover, in terms of overall personality

pathology prevalence, insomnia patients seemed to be closer to psychotherapy/psychiatry samples

AC

than to general population sample, while being significantly different even from the former. This

finding is highly consistent with a dimensional distribution of maladaptive personality traits, which

could be hardly described in terms of DSM-5 Section II PDs [28]. Notably, an Alternative Model of

Personality Disorder was provided in DSM-5 Section III [13], along with traditional PD symptom

criteria listed in DSM-5 Section II. A core component of the DSM-5 AMPD is an empirically based

model of maladaptive personality domain and traits which represents dysfunctional variants of the
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 16

Five Factor Model personality dimensions [13]. This system of dysfunctional traits may be used in

future studies to dimensionally assess dysfunction in personality in insomnia samples.

4.5 Limitations

Of course, we feel that the results of our study should be considered in the light of several

limitations. Our study was based in a cross-sectional design; indeed, longitudinal studies are needed

to clarify if personality features actually represent risk factors for insomnia, or if they rather

PT
represent long-term behavior modifications due to the impact of insomnia on daily life. Although

RI
we relied on consecutive admissions to select our participants, our data should be considered based

SC
on convenience study groups rather than on actually representative samples.

In the present study, we did not find any significant association between PSG data and
NU
continuously scored PDs. In their meta-analysis, Baglioni and colleagues [11] observed that sleep

continuity, sleep depth, and REM sleep disturbances may be associated with Borderline PD. In a
MA

sense, our findings extend previous studies [for quantitative reviews, see 10-11] on the relationships

between PSG data and PDs.

ED

Although we assessed PDs in insomnia patients using a semi-structured interview with

sound inter-rater reliability data, we were not able to assess the insomnia base rate among
PT

personality-disordered psychotherapy control patients. Moreover, we were not able to assess sleep
CE

quality and quantity among participants undergoing psychotherapy; these aspects represent major

limitations of our study, and they stress the need of further studies on this topic. When compared to
AC

Torgersen’s [31] data, our psychotherapy control patients showed a number of differences in base

rate estimates for individual PDs in international clinical populations. Although other explanations

may be possible, e.g., heterogeneity in PD base rate estimates [31], we feel that the main reason for

discrepant findings was the fact that our control groups was not composed by general psychiatry

patients; rather, it included only patients who voluntarily asked for psychotherapy treatment

because of personality problems.

 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 17

References

1. van de Laar M, Verbeek I, Pevernagie D, Aldenkamp A, Overeem S. The role of personality

traits in insomnia. Sleep Med Rev;2010;14:61-8.

2. Zammit GK, Weiner J, Damato N, Sillup GP, McMillan CA. Quality of life in people with

insomnia. Sleep 1999;22:379–85

3. Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders: an

PT
opportunity for prevention? J Am Med Assoc 1989;262:1479–84.

RI
4. American Academy of Sleep Medicine. International Classification of Sleep Disorders:

SC
Diagnostic and Coding Manual, 3rd edn. American Academy of Sleep Medicine:

Westchester, IL, 2014.

NU
5. Spielman AJ, Glovinsky PB. Introduction: the varied nature of insomnia. In: Hauri PJ, editor.

Case studies in insomnia. New York: Plenum; 1991. pp. 1–15.

MA

6. Riemann D, Spiegelhalder K, Feige B, Voderholzer U, Berger M, Perlis M, Nissen C. The

hyperarousal model of insomnia: A review of the concept and its evidence. Sleep Med Rev
ED

2010;14:19-31.

7. John, O. P., & Srivastava, S. (1999). The Big Five Trait taxonomy: History, measurement, and
PT

theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and
CE

research. New York, NY, US. pp. 102-138.

8. Blagrove M, Akehurst L. Personality and the modulation of effects of sleep loss on mood and
AC

cognition. Pers Individ Dif 2001;30:819–28.

9. Dørheim S, Garthus-Niegel S, Bjorvatn B, Eberhard-Gran M. Personality and Perinatal Maternal

Insomnia: A Study Across Childbirth. Behav Sleep Med 2014;14:34-48.

10. Benca RM, Obermeyer WH, Thisted RA, Gillin JC. Sleep and psychiatric disorders. A meta-

analysis. Arch Gen Psychiatry 1992;49:651–68.

11. Baglioni C, Nanovska S, Regen W, Spiegelhalder K, Feige B, Nissen C et al. Sleep and mental

disorders: A meta-analysis of polysomnographic research. Psychol Bull 2016;142:969-990.

 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 18

12. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders

4th edition: DSM-IV. American Psychiatric Press Inc, Washington, DC.

13. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders

5th edition: DSM 5. American Psychiatric Association, Arlington VA.

14. Tan TL, Kales JD, Kales A, Soldatos CR, Bixler EO. Biopsychobehavioral correlates of

insomnia. IV: diagnosis based on DSM-III. Am J Psychiatry 1984;141:357–62.

PT
15. Mahendran R, Subramaniam M, Chan YH. Psychiatric morbidity in patients referred to an

RI
insomnia clinic. Singapore Med J 2007;48:163–5.

SC
16. Schramm E, Hohagen F, Kappler C, Grasshoff U, Berger M. Mental comorbidity of chronic

insomnia in general practice attenders using DSM-III-R. Acta Psych Scand 1995;91:10–7.
NU
17. Ruiter M, Lichstein K, Nau S, Geyer J. Personality disorder features and insomnia status

amongst hypnotic-dependent adults. Sleep Med 2012;13:1122-9.

MA

18. Atalay H. Comorbidity of insomnia detected by Pittsburgh Sleep Quality Index with anxiety,

depression, and personality disorders. Isr J Psychiatry Relat Sci 2011;48:54–9.

ED

19. Curcio G, Tempesta D, Scarlata S, Marzano C, Moroni F, Rossini PM, ... et al. Validity of the

Italian version of the Pittsburgh sleep quality index (PSQI). Neurol Sci 2013;34:511-9.
PT

20. Castronovo V, Galbiati A, Marelli S, Brombin C, Cugnata F, Giarolli L, et al. Validation study
CE

of the Italian version of the Insomnia Severity Index (ISI). Neurol Sci 2016;37:1517-24.

21. AASM (American Academy of Sleep Medicine). International classification of sleep disorders.
AC

3nd ed.; 2014 (ICSD-3). Westchester, IL.

22. Iber C, Ancoli-Israel S, Chesson A, Quan S. The AASM manual for the scoring of sleep and

associated events: rules, terminology and technical specifications. American Academy of Sleep

Medicine, Westchester; 2007

23. Meyer D, Dimitriadou E, Hornik K, Weingessel A, Leisch F (2017). e1071: Misc Functions

of the Department of Statistics (E1071), TU Wien.

 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 19

24. First MB, Gibbon M, Spitzer RL, Williams JBW, Benjamin L. Structured clinical interview for

DSM-IV axis II personality disorders. New York, NY: Biometric Research Department; 1994

25. Maffei C, Fossati A, Agostoni I, Barraco A, Bagnato M, Deborah D et al. Interrater Reliability

and Internal Consistency of the Structured Clinical Interview for DSM-IV Axis II Personality

Disorders (SCID-II), Version 2.0. J Pers Disord. 1997;11:279-84.

26. Haslam N, Holland E, Kuppens P. Categories versus dimensions in personality and

PT
psychopathology: a quantitative review of taxometric research. Psychol Med 2012;42:903-20.

RI
27. Coghill D, Sonuga-Barke E. Annual Research Review: Categories versus dimensions in the

SC
classification and conceptualisation of child and adolescent mental disorders - implications of

recent empirical study. J Child Psychol Psychiatry 2012;53:469-489.

NU
28. Clark LA. Assessment and diagnosis of personality disorder: Perennial issues and an emerging

reconceptualization. Annu Rev Psychol 2007;58:227-57.

MA

29. Rosenthal R, 1991. Meta-analytic procedures for social research (revised). Newbury Park,

CA: Sage.
ED

30. Newcombe R. A deficiency of the odds ratio as a measure of effect size. Stat Med.

2006;25:4235-40.
PT

31. Torgersen S. Prevalence, sociodemographics, and functional impairment. In: American

CE

Psychiatric Publishing Textbook of Personality Disorders, JM. Oldham, AE. Skodol, DS. Bender,

American Psychiatric Publication; 2014.

AC

32. Cohen J. Statistical power analysis for the behavioral sciences. New York, NY: Psychology

Press, Taylor & Francis Group; 1988.

33. Olivier J, May W, Bell M. Relative effect sizes for measures of risk. Commun Stat Theory

Methods 2017;46:6774-81.

34. Zimmerman M. Diagnosing Personality Disorders. Arch Gen Psychiatry 1994;51:225.

 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 20

35. Gratz K, Roemer L. Multidimensional Assessment of Emotion Regulation and Dysregulation:

Development, Factor Structure, and Initial Validation of the Difficulties in Emotion Regulation

Scale. J Psychopathol Behav Assess 2004;26:41-54.

36. Pincus A, Lukowitsky M. Pathological Narcissism and Narcissistic Personality Disorder. Annul

Rev Clin Psychol 2010;6:421-46.

37. Brand S, Kirov R, Kalak N, Gerber M, Pühse U, Lemola S et al. Perfectionism related to self-

PT
reported insomnia severity, but not when controlled for stress and emotion regulation.

RI
Neuropsychiatr Dis Treat 2015; 263-71.

SC
38. Kamphuis J, Karsten J, de Weerd A, Lancel M. Sleep disturbances in a clinical forensic

psychiatric population. Sleep Med 2013;14:1164–9.

NU
39. Van Veen MM, Karsten J, Lancel M. Poor Sleep and Its Relation to Impulsivity in Patients with

Antisocial or Borderline Personality Disorders. Behav Med 2017;43:218-26.

MA

40. Sabouri S, Gerber M, Lemola S, Becker SP, Shamsi M, Shakouri Z, et al. Examining Dark

Triad traits in relation to sleep disturbances, anxiety sensitivity and intolerance of uncertainty in
ED

young adults. Compr Psychiatry 2016;68:103-10.

41. Cohen J, Cohen P, West S, Aiken L. Applied multiple regression/correlation analysis for the
PT

behavioral sciences. Mahawah, New Jersey: Lawrence Erlbaum Associates; 2003.

CE
AC
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 21

Table 1.
Demographic Characteristics of the Insomnia Participant Sample (n = 171) and Age- and Sex-
Matched Psychotherapy Patient Participants (n = 171).

Insomnia patients (n = 171) Psychotherapy patients (n = 171)

Male (%) 75 (43.9%) 75 (43.9%)
Female (%) 96 (56.1%) 96 (56.1%)
Participants’ mean age 40.88 years 40.88 years
(SD) (SD = 10.48 years) (SD = 10.48 years)
Age range 18-67 years 18-67 years

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 22

Table 2.
DSM-IV Axis II/DSM-5 Section II Personality Disorders in Insomnia Patients (n = 171): Comparisons with
Epidemiological Mean and Median Prevalence Estimates [31] in General Population.

Insomnia Comparisons with General Population Prevalence

Patients Estimates from Torgersen’s (2014) Study

DSM-IV Axis II/DSM-5 Section II PDs P Mdn P z 𝑃̅ z

Paranoid PD 0.0% 1.8% -1.77 1.7% -1.72
Schizoid PD 1.2% 0.8% 0.59 1.3% -0.12
Schizotypal PD 0.0% 0.7% -1.10 1.3% -1.50

PT
Antisocial PD 0.6% 1.0% -0.53 1.8% -1.18
Borderline PD 6.4% 1.7% 4.75*** 1.6% 5.00***
Histrionic PD 8.2% 0.7% 11.76*** 1.2% 8.41***
***
Narcissistic PD 19.9% 0.7% 30.11 0.8% 28.04***

RI
Avoidant PD 1.2% 2.3% -0.96 2.7% -1.21
Dependent PD 0.6% 1.0% -0.53 1.0% -0.53
Obsessive-Compulsive PD 3.5% 2.0% 1.40 2.5% 0.84

SC
Any PD diagnosis 52.0% 11.9% 16.20*** 11.0% 17.14***
Note. PD: Personality Disorder; P: Base rate expressed as percentage; Mdn P: Median prevalence estimate in
Torgersen’s [31] study expressed as percentage; 𝑃̅ : Mean prevalence estimate in Torgersen’s [31] study expressed as
percentage; z: Binomial test z.
NU
*** p (two-tailed) <.0023
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 23

Table 3.
DSM-IV Axis II/DSM-5 Section II Personality Disorders in Insomnia Patients (n = 171) and Psychotherapy Patients Matched on Age and Gender (n = 171):
Descriptive Statistics, Nonparametric Within-group and Between-group Comparisons.

Continuously-Scored (n. of Symptoms) Categorically-Scored

DSM-IV Axis II/DSM-5 Section II PDs DSM-IV Axis II/DSM-5 Section II PDs

Insomnia Psychotherapy Insomnia

P TPsychotherapy

DSM-IV Axis II/

Patients Patients Patients

R I Patients

SC
M SD M SD U z rES n % n % RR z
DSM-5 Section II PDs
a, b b *** a a
Paranoid PD 0.11 0.40 0.50 0.96 11650.00 -4.84 -.25 0 0.0 4 2.3 0.11 1.47
0.08 a 0.08 a 2 a, b 1a

U
Schizoid PD 0.57 0.31 14036.50 -1.80 -.09 1.2 0.6 2.00 0.57
Schizotypal PD 0.02 a 0.13 0.12 a 0.60 13846.50 -2.39 -.12 0 a, b 0.0 3a 1.8 0.14 1.29
Antisocial PD
Borderline PD
Histrionic PD
0.02 a
0.70 b, c
1.04 c
0.24
1.66
1.52
0.16 a
1.37 c
1.08 b, c
0.76
2.15
1.70
13765.00
11357.50
14204.00
-2.73

A
-4.23***
-0.51
N
-.14
-.22
-.03
1 a, b
11 c, d, e
14 d, e
0.6
6.4
8.2
4 a, b
20 c, d
14 b, c
2.3
11.7
8.2
0.25
0.55
1.00
1.25
1.66
0.00
Narcissistic PD
Avoidant PD
2.12 d
0.49 c
1.97
0.90
2.23 d
0.56 b, c
2.14
1.10
14466.50
14578.50 M -0.17
-0.06
-.01
.00
34 f
2 a, b, c
19.9
1.2
39 d
6 a, b, c
22.8
3.5
0.87
0.33
0.66
1.36
Dependent PD
Obsessive-Compulsive PD
0.47 c
0.68 c
0.88
1.19
0.79 b, c
0.72 b, c
1.18
1.11
E D
12514.50
13901.50
-2.70
-0.91
-.14
-.05
1 a, b
6 b, c, d
0.6
3.5
5 a, b
4 a, b
2.9
2.3
0.20
1.50
1.48
0.64

T
Not otherwise specified/with
24 e, f 14.0 38 d 22.0 0.60 2.17
other specification PD

N. of PD criteria
Any PD diagnosis
5.64 3.34
E P
7.46 4.82 11568.50 -3.35*** -.17
89 52.0 124 72.5% 0.72 3.80***

C C
Note. PD: Personality Disorder; rES: Rosenthal’s [29] effect size measure; U: Mann-Whitney U statistic; RR: Relative risk. Means and frequencies with different superscripts were
significantly different (i.e., p <.05) in repeated measure Dunn-Bonferroni post-hoc contrasts (e.g., Narcissistic PD showed a significantly higher mean number of PD traits than all

A
other PDs, whereas the mean number of Paranoid PD and Schizoid PD was not significantly different).
*** p <.0045
 ACCEPTED MANUSCRIPT
DSM-IV AXIS II/DSM-5 SECTION II PDS IN INSOMNIA 24
Highlights

 We evaluated 171 insomnia patients and a matched sample of 171 psychotherapy patients.

 52.0% of insomnia patients received at least one personality disorder diagnosis.

 Personality dysfunction may play a role in insomnia.

PT
RI
SC
NU
MA
ED
PT
CE
AC