RESEARCH ARTICLE
Subjective sleep characteristics associated with anxiety and depression in older adults: a population-based study Olivier Potvin1, Dominique Lorrain2, Geneviève Belleville3, Sébastien Grenier4 and Michel Préville2 1
Centre de recherche de l’Institut universitaire en santé mentale de Québec, Québec, Canada Université de Sherbrooke, Sherbrooke, Canada 3 Université Laval, Québec, Canada 4 Centre de recherche de l’Institut universitaire de gériatrie de Montréal, Montréal, Canada Correspondence to: O. Potvin, PhD, E-mail: [email protected] 2
Objective: Sleep complaints are often associated with anxiety and depression, but the specific complaints related to each syndrome are poorly characterized, especially in older adults. The objective was to identify subjective sleep characteristics specific to anxiety and depression in this population. Methods: A random sample of 2393 individuals aged 65 years or older was used. Anxiety and depression were categorized using DSM-V criteria for phobias, panic disorder, generalized anxiety disorder, unspecified anxiety disorder, major depressive episode, and depressive episode with insufficient symptoms. Subjective sleep characteristics were measured using the Pittsburgh Sleep Quality Index. Logistic regression models predicting anxiety or depression were used to determine the independent sleep characteristics associated with each syndrome adjusting for age, sex, education level, cognitive functioning, anxiolytic/sedative/hypnotic use, antidepressants use, subjective health, chronic diseases, cardiovascular conditions, and anxiety or depression (as appropriate). Results: Nearly all Pittsburgh Sleep Quality Index subscales were significantly associated with anxiety, but these subscales shared variance and only sleep duration in hours, sleep disturbance score, and daytime functioning score were independently related to anxiety. Within these significant subscales, the main specific sleep complaints associated with anxiety were daytime sleepiness and sleep disturbances related to coughing/snoring, feeling cold, and bad dreams. The use of sleeping medication was the only specific sleep characteristic associated with depression. Conclusions: These results suggest that in older adults, symptoms of short sleep duration, daytime sleepiness and sleep disturbances are independently related to anxiety while the use of sleep medication is independently associated to depression. Copyright # 2014 John Wiley & Sons, Ltd. Key words: anxiety disorder; mood disorder; sleep; insomnia; elderly; community survey History: Received 10 October 2013; Accepted 26 February 2014; Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/gps.4106
Introduction It is well known that poor sleep quality is associated with anxiety and depression. According to the Diagnostic and Statistical Manual of Mental Disorders (DSM) (American Psychiatric Association, 2000, 2013), sleep difficulties in generalized anxiety disorder (GAD) are described as difficulty falling or staying asleep and restless unsatisfying sleep. Previous studies also indicated that other anxiety disorders including panic Copyright # 2014 John Wiley & Sons, Ltd.
disorder/agoraphobia (Stein et al., 1993a; Lepola et al., 1994; Ramsawh et al., 2009), social phobia (Stein et al., 1993b; Ramsawh et al., 2009), specific phobia (Ramsawh et al., 2009), as well as subclinical anxiety symptoms (Taylor et al., 2005) are associated with poor sleep quality. For major depressive episode (MDE), it is proposed that sleep difficulties generally involve waking up during the night or too early in the morning and difficulty falling asleep may also occur. Int J Geriatr Psychiatry 2014
O. Potvin et al.
Although difficulties falling or staying asleep are considered as part of anxiety disorders and depression, these symptoms can occur with other sleep-related features such as shorter sleep duration, poorer sleep efficiency, daytime dysfunctions, etc. Whether anxiety disorders and depression display particular patterns of subjective sleep characteristics remains to be established. The identification of specific subjective sleep characteristics associated with either syndrome could eventually help in targeting sleep symptoms comprised within a syndrome rather than sleep difficulties that are unrelated to a syndrome. This aspect is even more challenging in older adults because they are more likely to have objective and subjective sleep difficulties compared with younger adults (Prinz et al., 1990; Newman et al., 1997; Schubert et al., 2002), and the prevalence rate of poor sleep in this population can reach more than 40% (Spira et al., 2012). Older adults also display more physical diseases, which is a condition associated with anxiety, depression, and poor sleep quality (Egede, 2007; Quan et al., 2005; Scott et al., 2007). Moreover, although some results suggested that sleep disturbance is a key feature to distinguish older adults with and without GAD symptoms (Wetherell et al., 2003), other results do not agree with the view proposed by the DSM concerning the nature of the sleep difficulties related to anxiety and depression (Brenes et al., 2009; Yokoyama et al., 2010; Soehner and Harvey, 2012). First, results in adults indicated that difficulty in initiating and maintaining sleep and early morning awakening have similar associations with anxiety and mood disorders (Soehner and Harvey, 2012). Second, in older adults, waking up too early in the morning, but not difficulty initiating sleep, was associated with GAD (Brenes et al., 2009), whereas difficulty initiating sleep, but not maintaining sleep, was related to depression (Yokoyama et al., 2010), suggesting that subjective sleep characteristics associated with anxiety and depression may differ in older adults compared with younger adults. At the moment, population-based data on the subjective sleep characteristics related to anxiety disorders in older adults are scarce (Spira et al., 2005; Magee and Carmin, 2010). Previous research in this population has mainly studied the sleep characteristics associated with depression. However, because anxiety disorders and depression are highly comorbid (Zimmerman et al., 2000; Kessler et al., 2003), both need to be studied simultaneously in order to determine the specific sleep characteristics related to each syndrome. Moreover, sleep complaints are likely to correlate with each other (e.g., difficulty staying asleep, unsatisfying Copyright # 2014 John Wiley & Sons, Ltd.
sleep, and shorter sleep duration, etc.), and to our knowledge, no previous study has assessed the independent subjective sleep characteristics associated with anxiety disorders and depression. The objective of this study was to identify, in a population-based sample of older adults, independent subjective sleep characteristics specific to anxiety and depression. Methods Participants
Participants were part of a population-based cohort, the ESA Study (Enquête sur la santé des aînés; Survey on elders’ health)(Préville et al., 2008). A random sample of 2811 community-dwelling French-speaking adults aged 65 years or older was recruited in the province of Québec, Canada. The sampling method used a random dialing procedure. Data were collected by research nurses through in-home structured interviews using a computer-assisted questionnaire. They received a training of 2 days by the principal investigator (Michel Préville) for the administration of the computerassisted questionnaire, which included familiarization periods and administration practices. Written informed consent was obtained at the beginning of the interview from all participants. The research procedures were previously reviewed and authorized by the Ethics Committee of the Institut universitaire de gériatrie de Sherbrooke. The response rate was 76.5%. In order to exclude participants with moderate and severe cognitive dysfunction, participants with Mini-mental State Examination (MMSE)(Folstein et al., 1975) scores below 22 were not evaluated (n = 26). Data from the interview were linked with medical records from the Régie de l’assurance maladie du Québec (Québec’s public health insurance plan). Medical records were not available for 291 individuals due to refusal of consent to provide medical records, moving outside Québec, or having additional drug insurance. Participants with schizophrenia or other forms of psychosis were excluded from the sample (n = 17). Education level was missing for nine participants, and mental health assessment was incomplete for 75 others. The final sample used for the present study included 2393 participants. Mental health
Anxiety and depression were assessed for the period of 1 year prior to the interview using a computer-assisted questionnaire based on DSM-IV criteria (American Int J Geriatr Psychiatry 2014
Anxiety, depression, and sleep
Psychiatric Association, 2000), the Diagnostic Interview Schedule and the Composite International Diagnostic Interview, which demonstrated satisfactory reliability and validity (Robins et al., 1981; Erdman et al., 1992). For the present study, algorithms were modified in order to reflect DSM-V categories for anxiety and mood disorders (American Psychiatric Association, 2013). The main changes for anxiety and depression criteria between DSM-IV and DSM-V include the exclusion of obsessive-compulsive disorder, the removal of the bereavement exclusion for depression, and new terms for unspecified disorders. Two DSM-V categories for anxiety were used: anxiety disorders and unspecified anxiety disorder (USAD). Anxiety disorders included specific phobia, social phobia, agoraphobia, panic disorder, and GAD. USAD included participants having anxiety symptoms causing significant distress or functional impairment without meeting the full criteria for a particular anxiety disorder. Two DSM-V categories for depression were used: MDE and depressive episode with insufficient symptoms (DEIS). DEIS, which is similar to the DSM-IV minor depression category, is defined as depressed affect in addition of at least one other MDE symptom causing significant distress or functional impairment. Subjective sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI) (Buysse et al., 1989; Blais et al., 1997), and poor sleepers were defined as a score higher than five. Covariates
Sociodemographic characteristics (age, sex, and education level), cognitive functioning level, physical health characteristics, and psychotropic drug use were included as covariates. Cognitive functioning was assessed by the MMSE (Folstein et al., 1975), and the number of chronic diseases was measured using a list of 14 health problems. Subjective health was measured by asking participants to evaluate their health compared with other individuals of their age. A cardiovascular conditions score was computed using self-report of high blood pressure, diabetes, and heart disease at baseline and medical records for the period of 1 year before baseline. The score was the total of the presence of these three health problems. Anxiolytics/sedatives/ hypnotics and antidepressants use was assessed by the Régie de l’assurance maladie du Québec medical records, and drugs were coded according to the American Hospital Formulary Service (American Society of Health System Pharmacist, 2001). Drug use was defined as at least one prescription during the year prior the interview. Copyright # 2014 John Wiley & Sons, Ltd.
Statistical analyses
First, to compare with results from previous study, crude and adjusted odds ratio (OR) with 95% confidence intervals (CI) were computed using a multinomial logistic regression with anxiety (disorder, USAD, and no anxiety) or depression (MDE, DEIS, and no depression) as the dependent variable and global PSQI score as predictor. For adjusted ORs, covariates included age, education level, anxiety, or depression according to the dependent variable, MMSE score, subjective health, anxiolytic/sedative/hypnotic use, antidepressants use, cardiovascular conditions score, and the number of chronic diseases. Second, to determine the PSQI subscales independently associated with anxiety and depression, an adjusted logistic model including all PSQI subscales was conducted. In order to reduce the number of predictors, nonsignificant PSQI subscales were manually removed one by one, and the final model included only PSQI subscales with a p-value below 5%. ORs were also computed for each PSQI subscale in order to show which subscales had overlapping of variance for the prediction of anxiety/depression. Third, to identify which sleep complaints appear to be more important within the PSQI subscales independently associated with anxiety and depression, sleep complaints in these subscales were examined individually. These sleep complaints were dichotomized (at least once a week/not during the last month), and ORs were adjusted for the aforementioned covariates and for the other independently associated PSQI subscales. For these post hoc analyses, a corrected p-value was used (0.05/number of tests). Finally, because previous studies observed nonlinear relationships (U-shaped) between anxiety/depression and sleep duration (van den Berg et al., 2009; van Mill et al., 2010), nonlinear relationships between sleep duration, anxiety, and depression were tested using a quadratic term. All adjusted analyses on anxiety included depression as covariate, whereas all adjusted analyses on depression included anxiety as covariate. All PSQI subscales (subjective sleep quality, sleep latency, sleep duration habitual sleep efficiency, sleep disturbances, use of sleeping medications, and daytime dysfunction) were used as continuous variables with their original score, except sleep duration. The number of hours of sleep duration was preferred to the sleep duration PSQI subscale because it was used in previous studies observing nonlinear relationships (van den Berg et al., 2009; van Mill et al., 2010). Statistical assumptions were verified, and analyses were performed using IBM SPSS Statistics 19 software (New York, USA). Int J Geriatr Psychiatry 2014
O. Potvin et al.
Results The mean age was 73.8 years old (SD: 5.7). Sixty-nine percent of the sample were women, and 35% of the participants had postsecondary education. The mean cardiovascular condition score was of 1.2 (SD: 0.9), and the mean number of chronic diseases was 3.1 (SD: 1.9). Approximately one-third of the sample (31.2%) was using anxiolytics/sedatives/hypnotics while 13.9% were using antidepressants. Eighty participants had an anxiety disorder (3.3%), and 166 had a USAD (6.9%). One hundred and seventeen participants had an MDE (4.9%) while 192 had a DEIS (8.0%). Eighty-four participants had both anxiety disorder/USAD and MDE/DEIS (21 with anxiety disorder and MDE/DEIS, 37 with MDE and anxiety disorder/USAD, and 26 with USAD and DEIS). For clarity purposes, the terms anxiety and depression are referring to the variables (Anxiety: anxiety disorder/USAD; Depression: MDE/DEIS). Table 1 displays the prevalence rates of poor sleep according to anxiety and depression. The prevalence of poor sleepers was of 64% in anxiety disorders and 55% in USAD compared with 37% in participants without anxiety. Poor sleep was significantly associated with all anxiety categories. The prevalence of poor sleepers was of 54% in MDE and 58% in DEIS compared with 36% in those without depression. After adjustments for covariates, DEIS, but not MDE, was significantly related to poor sleep. Table 2 presents the ORs related to the associations between PSQI subscales and anxiety. Model 1 shows that after adjustment for covariates, all PSQI subscales, except use of sleeping medication, were significantly related to anxiety. Model 2, which initially included all PSQI subscales, revealed that these subscales shared
variance and sleep duration, sleep disturbances, and daytime dysfunction were the only sleep variables independently and significantly associated with anxiety. Table 3 shows the specific sleep complaints among the sleep disturbance and daytime dysfunction subscales associated with anxiety. Sleep complaints significantly associated with anxiety were related to coughing or snoring, feeling too cold bad dreams, and trouble staying awake (daytime) while using the bathroom, breathing problems and pain were close to significance. Significant predictors for anxiety disorders were coughing or snoring and bad dreams. Significant predictors for USAD were feeling too cold and having trouble staying awake. Table 4 presents the ORs related to the association between PSQI subscales and depression. When tested separately with Model 1, sleep disturbances, use of sleeping medication, and daytime dysfunction subscales significantly predicted depression. Model 2, which initially included all PSQI subscales, revealed that use of sleeping medication and daytime dysfunction subscales were the only independent and significant subscales associated with depression. Among the daytime dysfunction subscale, keeping up enough enthusiasm (MDE: OR 1.99, CI 1.33–2.96, p < 0.001; DEIS: 2.24, 1.63–3.07, p < 0.001), but not trouble staying awake (MDE: 1.41, 0.75–2.64, p = 0.282; DEIS: 1.58, 0.96–2.58, p = 0.070), was significantly associated with depression. The subscale use of sleeping medication was not further investigated because it comprises a single question. No significant nonlinear relationship between total sleep time and anxiety or depression was observed. Finally, because depression can occur in shorter time windows (2-week episodes) than anxiety disorders and that psychiatric disorders and sleep quality were
Table 1 Prevalence of poor sleep in older adults with anxiety and depression (N = 2393) Adjusteda Symptomatology Anxiety Disorder USAD No anxiety Depression MDE DEIS No depression
Good sleepers (PSQI ≤ 5) n (%)
Poor sleepers (PSQI > 5) n (%)
Crude OR (95% CI)
OR (95% CI)
29 (36.3) 74 (44.6) 1358 (63.3)
51 (63.8) 92 (55.4) 789 (36.7)
3.03 (1.90–4.82) 2.14 (1.56–2.94) 1
2.16 (1.30–3.60) 1.51 (1.06–2.15) 1
54 (46.2) 81 (42.2) 1326 (63.6)
63 (53.8) 111 (57.8) 758 (36.4)
2.04 (1.40–2.97) 2.40 (1.78–3.24) 1
1.07 (0.70–1.64) 1.70 (1.22–2.36) 1
p 0.001 0.003 0.021 0.007 0.740 0.002
CI, confidence interval; DEIS, depressive episode with insufficient symptoms; MDE, major depressive episode; OR, odds ratio; PSQI, Pittsburgh Sleep Quality Index; USAD, unspecified anxiety disorder. a Logistic regression models adjusted for age, sex, education level, anxiety or depressive disorders, Mini-mental State Examination, anxiolytic/ sedative/hypnotic use, antidepressants use, subjective health, chronic diseases, and cardiovascular conditions.
Copyright # 2014 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014
Anxiety, depression, and sleep Table 2 Association between anxiety and sleep complaints in older adults Model 1a (Separated models) PSQI subscale score Subjective sleep quality Anxiety disorder USAD No anxiety Sleep latency Anxiety disorder USAD No anxiety Sleep duration (hours) Anxiety disorder USAD No anxiety Habitual sleep efficiency Anxiety disorder USAD No anxiety Sleep disturbances Anxiety disorder USAD No anxiety Use of sleeping medication Anxiety disorder USAD No anxiety Daytime dysfunction Anxiety disorder USAD No anxiety
Mean (SD)
OR (95% CI)
1.06 (0.86) 0.99 (0.79) 0.76 (0.64)
1.50 (1.07–2.09) 1.26 (0.99–1.61) 1
1.35 (1.15) 1.24 (1.10) 0.94 (0.96)
1.33 (1.06–1.66) 1.19 (1.01–1.40) 1
6.56 (1.56) 6.91 (1.39) 7.20 (1.39)
0.77 (0.66–0.90) 0.90 (0.81–1.01) 1
1.15 (1.23) 0.98 (1.15) 0.71 (1.01)
1.32 (1.08–1.61) 1.17 (1.01–1.35) 1
1.05 (0.53) 1.02 (0.57) 0.80 (0.51)
1.88 (1.20–2.95) 1.81 (1.31–2.49) 1
1.11 (1.37) 1.01 (1.33) 0.67 (1.17)
1.12 (0.89–1.40) 1.01 (0.85–1.19) 1
0.56 (0.73) 0.70 (0.85) 0.37 (0.62)
1.16 (0.84–1.60) 1.53 (1.24–1.89) 1
p
Model 2b (Single model) OR (95% CI)
p
0.80 (0.68–0.94) 0.95 (0.85–1.07)
0.018 0.005 0.412
0.018 0.018 0.064 0.009 0.013 0.036 0.002 0.001 0.071 0.006 0.007 0.040
Subjective sleep characteristics associated with anxiety and depression in older adults: a population-based study Olivier Potvin1, Dominique Lorrain2, Geneviève Belleville3, Sébastien Grenier4 and Michel Préville2 1
Centre de recherche de l’Institut universitaire en santé mentale de Québec, Québec, Canada Université de Sherbrooke, Sherbrooke, Canada 3 Université Laval, Québec, Canada 4 Centre de recherche de l’Institut universitaire de gériatrie de Montréal, Montréal, Canada Correspondence to: O. Potvin, PhD, E-mail: [email protected] 2
Objective: Sleep complaints are often associated with anxiety and depression, but the specific complaints related to each syndrome are poorly characterized, especially in older adults. The objective was to identify subjective sleep characteristics specific to anxiety and depression in this population. Methods: A random sample of 2393 individuals aged 65 years or older was used. Anxiety and depression were categorized using DSM-V criteria for phobias, panic disorder, generalized anxiety disorder, unspecified anxiety disorder, major depressive episode, and depressive episode with insufficient symptoms. Subjective sleep characteristics were measured using the Pittsburgh Sleep Quality Index. Logistic regression models predicting anxiety or depression were used to determine the independent sleep characteristics associated with each syndrome adjusting for age, sex, education level, cognitive functioning, anxiolytic/sedative/hypnotic use, antidepressants use, subjective health, chronic diseases, cardiovascular conditions, and anxiety or depression (as appropriate). Results: Nearly all Pittsburgh Sleep Quality Index subscales were significantly associated with anxiety, but these subscales shared variance and only sleep duration in hours, sleep disturbance score, and daytime functioning score were independently related to anxiety. Within these significant subscales, the main specific sleep complaints associated with anxiety were daytime sleepiness and sleep disturbances related to coughing/snoring, feeling cold, and bad dreams. The use of sleeping medication was the only specific sleep characteristic associated with depression. Conclusions: These results suggest that in older adults, symptoms of short sleep duration, daytime sleepiness and sleep disturbances are independently related to anxiety while the use of sleep medication is independently associated to depression. Copyright # 2014 John Wiley & Sons, Ltd. Key words: anxiety disorder; mood disorder; sleep; insomnia; elderly; community survey History: Received 10 October 2013; Accepted 26 February 2014; Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/gps.4106
Introduction It is well known that poor sleep quality is associated with anxiety and depression. According to the Diagnostic and Statistical Manual of Mental Disorders (DSM) (American Psychiatric Association, 2000, 2013), sleep difficulties in generalized anxiety disorder (GAD) are described as difficulty falling or staying asleep and restless unsatisfying sleep. Previous studies also indicated that other anxiety disorders including panic Copyright # 2014 John Wiley & Sons, Ltd.
disorder/agoraphobia (Stein et al., 1993a; Lepola et al., 1994; Ramsawh et al., 2009), social phobia (Stein et al., 1993b; Ramsawh et al., 2009), specific phobia (Ramsawh et al., 2009), as well as subclinical anxiety symptoms (Taylor et al., 2005) are associated with poor sleep quality. For major depressive episode (MDE), it is proposed that sleep difficulties generally involve waking up during the night or too early in the morning and difficulty falling asleep may also occur. Int J Geriatr Psychiatry 2014
O. Potvin et al.
Although difficulties falling or staying asleep are considered as part of anxiety disorders and depression, these symptoms can occur with other sleep-related features such as shorter sleep duration, poorer sleep efficiency, daytime dysfunctions, etc. Whether anxiety disorders and depression display particular patterns of subjective sleep characteristics remains to be established. The identification of specific subjective sleep characteristics associated with either syndrome could eventually help in targeting sleep symptoms comprised within a syndrome rather than sleep difficulties that are unrelated to a syndrome. This aspect is even more challenging in older adults because they are more likely to have objective and subjective sleep difficulties compared with younger adults (Prinz et al., 1990; Newman et al., 1997; Schubert et al., 2002), and the prevalence rate of poor sleep in this population can reach more than 40% (Spira et al., 2012). Older adults also display more physical diseases, which is a condition associated with anxiety, depression, and poor sleep quality (Egede, 2007; Quan et al., 2005; Scott et al., 2007). Moreover, although some results suggested that sleep disturbance is a key feature to distinguish older adults with and without GAD symptoms (Wetherell et al., 2003), other results do not agree with the view proposed by the DSM concerning the nature of the sleep difficulties related to anxiety and depression (Brenes et al., 2009; Yokoyama et al., 2010; Soehner and Harvey, 2012). First, results in adults indicated that difficulty in initiating and maintaining sleep and early morning awakening have similar associations with anxiety and mood disorders (Soehner and Harvey, 2012). Second, in older adults, waking up too early in the morning, but not difficulty initiating sleep, was associated with GAD (Brenes et al., 2009), whereas difficulty initiating sleep, but not maintaining sleep, was related to depression (Yokoyama et al., 2010), suggesting that subjective sleep characteristics associated with anxiety and depression may differ in older adults compared with younger adults. At the moment, population-based data on the subjective sleep characteristics related to anxiety disorders in older adults are scarce (Spira et al., 2005; Magee and Carmin, 2010). Previous research in this population has mainly studied the sleep characteristics associated with depression. However, because anxiety disorders and depression are highly comorbid (Zimmerman et al., 2000; Kessler et al., 2003), both need to be studied simultaneously in order to determine the specific sleep characteristics related to each syndrome. Moreover, sleep complaints are likely to correlate with each other (e.g., difficulty staying asleep, unsatisfying Copyright # 2014 John Wiley & Sons, Ltd.
sleep, and shorter sleep duration, etc.), and to our knowledge, no previous study has assessed the independent subjective sleep characteristics associated with anxiety disorders and depression. The objective of this study was to identify, in a population-based sample of older adults, independent subjective sleep characteristics specific to anxiety and depression. Methods Participants
Participants were part of a population-based cohort, the ESA Study (Enquête sur la santé des aînés; Survey on elders’ health)(Préville et al., 2008). A random sample of 2811 community-dwelling French-speaking adults aged 65 years or older was recruited in the province of Québec, Canada. The sampling method used a random dialing procedure. Data were collected by research nurses through in-home structured interviews using a computer-assisted questionnaire. They received a training of 2 days by the principal investigator (Michel Préville) for the administration of the computerassisted questionnaire, which included familiarization periods and administration practices. Written informed consent was obtained at the beginning of the interview from all participants. The research procedures were previously reviewed and authorized by the Ethics Committee of the Institut universitaire de gériatrie de Sherbrooke. The response rate was 76.5%. In order to exclude participants with moderate and severe cognitive dysfunction, participants with Mini-mental State Examination (MMSE)(Folstein et al., 1975) scores below 22 were not evaluated (n = 26). Data from the interview were linked with medical records from the Régie de l’assurance maladie du Québec (Québec’s public health insurance plan). Medical records were not available for 291 individuals due to refusal of consent to provide medical records, moving outside Québec, or having additional drug insurance. Participants with schizophrenia or other forms of psychosis were excluded from the sample (n = 17). Education level was missing for nine participants, and mental health assessment was incomplete for 75 others. The final sample used for the present study included 2393 participants. Mental health
Anxiety and depression were assessed for the period of 1 year prior to the interview using a computer-assisted questionnaire based on DSM-IV criteria (American Int J Geriatr Psychiatry 2014
Anxiety, depression, and sleep
Psychiatric Association, 2000), the Diagnostic Interview Schedule and the Composite International Diagnostic Interview, which demonstrated satisfactory reliability and validity (Robins et al., 1981; Erdman et al., 1992). For the present study, algorithms were modified in order to reflect DSM-V categories for anxiety and mood disorders (American Psychiatric Association, 2013). The main changes for anxiety and depression criteria between DSM-IV and DSM-V include the exclusion of obsessive-compulsive disorder, the removal of the bereavement exclusion for depression, and new terms for unspecified disorders. Two DSM-V categories for anxiety were used: anxiety disorders and unspecified anxiety disorder (USAD). Anxiety disorders included specific phobia, social phobia, agoraphobia, panic disorder, and GAD. USAD included participants having anxiety symptoms causing significant distress or functional impairment without meeting the full criteria for a particular anxiety disorder. Two DSM-V categories for depression were used: MDE and depressive episode with insufficient symptoms (DEIS). DEIS, which is similar to the DSM-IV minor depression category, is defined as depressed affect in addition of at least one other MDE symptom causing significant distress or functional impairment. Subjective sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI) (Buysse et al., 1989; Blais et al., 1997), and poor sleepers were defined as a score higher than five. Covariates
Sociodemographic characteristics (age, sex, and education level), cognitive functioning level, physical health characteristics, and psychotropic drug use were included as covariates. Cognitive functioning was assessed by the MMSE (Folstein et al., 1975), and the number of chronic diseases was measured using a list of 14 health problems. Subjective health was measured by asking participants to evaluate their health compared with other individuals of their age. A cardiovascular conditions score was computed using self-report of high blood pressure, diabetes, and heart disease at baseline and medical records for the period of 1 year before baseline. The score was the total of the presence of these three health problems. Anxiolytics/sedatives/ hypnotics and antidepressants use was assessed by the Régie de l’assurance maladie du Québec medical records, and drugs were coded according to the American Hospital Formulary Service (American Society of Health System Pharmacist, 2001). Drug use was defined as at least one prescription during the year prior the interview. Copyright # 2014 John Wiley & Sons, Ltd.
Statistical analyses
First, to compare with results from previous study, crude and adjusted odds ratio (OR) with 95% confidence intervals (CI) were computed using a multinomial logistic regression with anxiety (disorder, USAD, and no anxiety) or depression (MDE, DEIS, and no depression) as the dependent variable and global PSQI score as predictor. For adjusted ORs, covariates included age, education level, anxiety, or depression according to the dependent variable, MMSE score, subjective health, anxiolytic/sedative/hypnotic use, antidepressants use, cardiovascular conditions score, and the number of chronic diseases. Second, to determine the PSQI subscales independently associated with anxiety and depression, an adjusted logistic model including all PSQI subscales was conducted. In order to reduce the number of predictors, nonsignificant PSQI subscales were manually removed one by one, and the final model included only PSQI subscales with a p-value below 5%. ORs were also computed for each PSQI subscale in order to show which subscales had overlapping of variance for the prediction of anxiety/depression. Third, to identify which sleep complaints appear to be more important within the PSQI subscales independently associated with anxiety and depression, sleep complaints in these subscales were examined individually. These sleep complaints were dichotomized (at least once a week/not during the last month), and ORs were adjusted for the aforementioned covariates and for the other independently associated PSQI subscales. For these post hoc analyses, a corrected p-value was used (0.05/number of tests). Finally, because previous studies observed nonlinear relationships (U-shaped) between anxiety/depression and sleep duration (van den Berg et al., 2009; van Mill et al., 2010), nonlinear relationships between sleep duration, anxiety, and depression were tested using a quadratic term. All adjusted analyses on anxiety included depression as covariate, whereas all adjusted analyses on depression included anxiety as covariate. All PSQI subscales (subjective sleep quality, sleep latency, sleep duration habitual sleep efficiency, sleep disturbances, use of sleeping medications, and daytime dysfunction) were used as continuous variables with their original score, except sleep duration. The number of hours of sleep duration was preferred to the sleep duration PSQI subscale because it was used in previous studies observing nonlinear relationships (van den Berg et al., 2009; van Mill et al., 2010). Statistical assumptions were verified, and analyses were performed using IBM SPSS Statistics 19 software (New York, USA). Int J Geriatr Psychiatry 2014
O. Potvin et al.
Results The mean age was 73.8 years old (SD: 5.7). Sixty-nine percent of the sample were women, and 35% of the participants had postsecondary education. The mean cardiovascular condition score was of 1.2 (SD: 0.9), and the mean number of chronic diseases was 3.1 (SD: 1.9). Approximately one-third of the sample (31.2%) was using anxiolytics/sedatives/hypnotics while 13.9% were using antidepressants. Eighty participants had an anxiety disorder (3.3%), and 166 had a USAD (6.9%). One hundred and seventeen participants had an MDE (4.9%) while 192 had a DEIS (8.0%). Eighty-four participants had both anxiety disorder/USAD and MDE/DEIS (21 with anxiety disorder and MDE/DEIS, 37 with MDE and anxiety disorder/USAD, and 26 with USAD and DEIS). For clarity purposes, the terms anxiety and depression are referring to the variables (Anxiety: anxiety disorder/USAD; Depression: MDE/DEIS). Table 1 displays the prevalence rates of poor sleep according to anxiety and depression. The prevalence of poor sleepers was of 64% in anxiety disorders and 55% in USAD compared with 37% in participants without anxiety. Poor sleep was significantly associated with all anxiety categories. The prevalence of poor sleepers was of 54% in MDE and 58% in DEIS compared with 36% in those without depression. After adjustments for covariates, DEIS, but not MDE, was significantly related to poor sleep. Table 2 presents the ORs related to the associations between PSQI subscales and anxiety. Model 1 shows that after adjustment for covariates, all PSQI subscales, except use of sleeping medication, were significantly related to anxiety. Model 2, which initially included all PSQI subscales, revealed that these subscales shared
variance and sleep duration, sleep disturbances, and daytime dysfunction were the only sleep variables independently and significantly associated with anxiety. Table 3 shows the specific sleep complaints among the sleep disturbance and daytime dysfunction subscales associated with anxiety. Sleep complaints significantly associated with anxiety were related to coughing or snoring, feeling too cold bad dreams, and trouble staying awake (daytime) while using the bathroom, breathing problems and pain were close to significance. Significant predictors for anxiety disorders were coughing or snoring and bad dreams. Significant predictors for USAD were feeling too cold and having trouble staying awake. Table 4 presents the ORs related to the association between PSQI subscales and depression. When tested separately with Model 1, sleep disturbances, use of sleeping medication, and daytime dysfunction subscales significantly predicted depression. Model 2, which initially included all PSQI subscales, revealed that use of sleeping medication and daytime dysfunction subscales were the only independent and significant subscales associated with depression. Among the daytime dysfunction subscale, keeping up enough enthusiasm (MDE: OR 1.99, CI 1.33–2.96, p < 0.001; DEIS: 2.24, 1.63–3.07, p < 0.001), but not trouble staying awake (MDE: 1.41, 0.75–2.64, p = 0.282; DEIS: 1.58, 0.96–2.58, p = 0.070), was significantly associated with depression. The subscale use of sleeping medication was not further investigated because it comprises a single question. No significant nonlinear relationship between total sleep time and anxiety or depression was observed. Finally, because depression can occur in shorter time windows (2-week episodes) than anxiety disorders and that psychiatric disorders and sleep quality were
Table 1 Prevalence of poor sleep in older adults with anxiety and depression (N = 2393) Adjusteda Symptomatology Anxiety Disorder USAD No anxiety Depression MDE DEIS No depression
Good sleepers (PSQI ≤ 5) n (%)
Poor sleepers (PSQI > 5) n (%)
Crude OR (95% CI)
OR (95% CI)
29 (36.3) 74 (44.6) 1358 (63.3)
51 (63.8) 92 (55.4) 789 (36.7)
3.03 (1.90–4.82) 2.14 (1.56–2.94) 1
2.16 (1.30–3.60) 1.51 (1.06–2.15) 1
54 (46.2) 81 (42.2) 1326 (63.6)
63 (53.8) 111 (57.8) 758 (36.4)
2.04 (1.40–2.97) 2.40 (1.78–3.24) 1
1.07 (0.70–1.64) 1.70 (1.22–2.36) 1
p 0.001 0.003 0.021 0.007 0.740 0.002
CI, confidence interval; DEIS, depressive episode with insufficient symptoms; MDE, major depressive episode; OR, odds ratio; PSQI, Pittsburgh Sleep Quality Index; USAD, unspecified anxiety disorder. a Logistic regression models adjusted for age, sex, education level, anxiety or depressive disorders, Mini-mental State Examination, anxiolytic/ sedative/hypnotic use, antidepressants use, subjective health, chronic diseases, and cardiovascular conditions.
Copyright # 2014 John Wiley & Sons, Ltd.
Int J Geriatr Psychiatry 2014
Anxiety, depression, and sleep Table 2 Association between anxiety and sleep complaints in older adults Model 1a (Separated models) PSQI subscale score Subjective sleep quality Anxiety disorder USAD No anxiety Sleep latency Anxiety disorder USAD No anxiety Sleep duration (hours) Anxiety disorder USAD No anxiety Habitual sleep efficiency Anxiety disorder USAD No anxiety Sleep disturbances Anxiety disorder USAD No anxiety Use of sleeping medication Anxiety disorder USAD No anxiety Daytime dysfunction Anxiety disorder USAD No anxiety
Mean (SD)
OR (95% CI)
1.06 (0.86) 0.99 (0.79) 0.76 (0.64)
1.50 (1.07–2.09) 1.26 (0.99–1.61) 1
1.35 (1.15) 1.24 (1.10) 0.94 (0.96)
1.33 (1.06–1.66) 1.19 (1.01–1.40) 1
6.56 (1.56) 6.91 (1.39) 7.20 (1.39)
0.77 (0.66–0.90) 0.90 (0.81–1.01) 1
1.15 (1.23) 0.98 (1.15) 0.71 (1.01)
1.32 (1.08–1.61) 1.17 (1.01–1.35) 1
1.05 (0.53) 1.02 (0.57) 0.80 (0.51)
1.88 (1.20–2.95) 1.81 (1.31–2.49) 1
1.11 (1.37) 1.01 (1.33) 0.67 (1.17)
1.12 (0.89–1.40) 1.01 (0.85–1.19) 1
0.56 (0.73) 0.70 (0.85) 0.37 (0.62)
1.16 (0.84–1.60) 1.53 (1.24–1.89) 1
p
Model 2b (Single model) OR (95% CI)
p
0.80 (0.68–0.94) 0.95 (0.85–1.07)
0.018 0.005 0.412
0.018 0.018 0.064 0.009 0.013 0.036 0.002 0.001 0.071 0.006 0.007 0.040
