Delayed sleep phase syndrome is related to seasonal affective disorder

Abstract

Background

Both delayed sleep phase syndrome (DSPS) and seasonal affective disorder (SAD) may manifest similar delayed circadian phase problems. However, the relationships and co-morbidity between the two conditions have not been fully studied. The authors examined the comorbidity between DSPS and SAD.

Methods

We recruited a case series of 327 DSPS and 331 controls with normal sleep, roughly matched for age, gender, and ancestry. Both DSPS and controls completed extensive questionnaires about sleep, the morningness-eveningness trait, depression, mania, and seasonality of symptoms, etc.

Results

The prevalences of SAD and subsyndromal SAD (S-SAD) were higher in DSPS compared to controls (χ²=12.65, p=0.002). DSPS were 3.3 times more likely to report SAD (odds ratio, 3.34; 95% CI, 1.41–7.93) compared to controls as defined by the Seasonal Pattern Assessment Questionnaire (SPAQ). Correspondingly, DSPS showed significantly higher seasonality scores compared to controls in mood, appetite, and energy level subscores and the global seasonality score (t=3.12, t=0.002; t=2.04, p=0.041; t=2.64, p=0.008; and t=2.15, p=0.032, respectively). Weight fluctuation during seasons and winter-summer sleep length differences were also significantly higher in DSPS than controls (t=5.16, p<0.001 and t=2.64, p=0.009, respectively). SAD and S-SAD reported significantly higher eveningness, higher depression self-ratings, and more previous mania symptoms compared to non-seasonal subjects regardless of whether they were DSPS or controls.

Conclusions

These cases suggested that DSPS is partially comorbid with SAD. These data support the hypothesis that DSPS and SAD may share a pathophysiological mechanism causing delayed circadian phase.

1. Introduction

Delayed sleep-phase syndrome (DSPS) is a circadian rhythm sleep disorder. DSPS patients tend to fall asleep some hours after midnight and have difficulty waking up in the morning. Not only their timing of sleep but also their peak times of performance, body temperature and hormonal rhythms are much delayed from societal norms. When allowed to choose their preferred schedules, DSPS patients are said to exhibit normal sleep quality and duration and maintain a delayed, but stable, phase of entrainment to the 24-h sleep–wake pattern (American Academy of Sleep Medicine, 2005). However, some studies suggest that sleep quality is poorer even when sleep and wake times have been self-selected (Campbell and Murphy, 2007). A prevalence in the adult population, equally distributed among women and men, has been reported at approximately 0.15% (Schrader et al., 1993; Yazaki et al., 1999), but Ando et al. (2002) found that the prevalence of mild DSPS symptoms might be much higher. Light therapy and melatonin administration have been reported to be effective treatments (Barion and Zee, 2007; Mundey et al., 2005).

Seasonal affective disorder (SAD), which is usually synonymous with winter depression, is a mood disorder in which people experience recurrent depressive symptoms in the winter while they are asymptomatic throughout much of the year. It has been estimated that 1.5–9% of adults in the US experience SAD (Modell et al., 2005). An excellent Canadian study estimated that the seasonal subtype of depression represents 11% of all subjects with major depression yielding a point prevalence of 2.9% (Levitt et al, 2000). However, a milder form, so-called subsyndromal SAD (S-SAD), is much more common than SAD. Rosenthal (1989) estimated that for the United States about 6.1% have SAD and about 14.3% have S-SAD, which represents a higher estimate than Levitt’s for a Canadian sample.

SAD patients may report sleeping too much, display little energy, and may also feel depressed during winter, though there may be more lethargy about getting out of bed than objective hypersomnia. Many different treatments for SAD have been recommended, including light therapy with sunlight or bright lights, antidepressant medications, cognitive-behavioral therapy, and timed administration of the hormone melatonin. (Lavoie et al., 2009; Lam et al., 2006; Rohan et al., 2007; Rohan et al., 2009; Terman and Terman, 2010; Lewy et al., 2009) DSPS and SAD might share similar pathophysiological mechanisms, since both manifest problems of delayed circadian rhythm phase and both are treated by morning light therapy and melatonin before bedtime. However, as far as we know, there are no large scale studies of the relationships between DSPS and SAD yet except one case report. (Uruha et al., 1991)

Although the dim light melatonin onset (DLMO) is delayed in relation to the timing of their sleep in most SAD patients (Lewy et al., 2006), there also appears to be a subgroup of SAD patients in which symptom severity correlates with a DLMO advance in reference to sleep, suggesting internal circadian misalignment. Interestingly, DSPS has also been reported associated with internal circadian misalignment which can be either a phase delay or a phase advance (Hughes and Lewy, 1998). Therefore, the relationships and comorbidity of these two conditions are interesting research topics in chronobiology and psychiatry. The purpose of this report was to investigate the relationship between DSPS and SAD in a DSPS case and control sample.

2. Methods

2.1. DSPS sample recruitment

The study subjects were recruited for a case-control genetic study of DSPS. In the initial part of the study, recruitment was limited to the Southern California region, particularly San Diego County, to enable the investigators to make home visits when necessary. The study was later expanded to other states of the US. Recruitment of the sample took place between June, 2004 and September, 2008. Some data from the early portion of the sample was published previously (Kripke et al., 2008). Recruitment of DSPS participants utilized contacts with sleep physicians, word of mouth, media contacts, newspaper advertising, late night radio advertising, UCSD minority outreach programs, paid context-based internet ads, and free internet advertising. A web site was constructed to facilitate contacts, provide potential volunteers with information, and allow potential volunteers to self-screen themselves versus the Horne-Östberg questionnaire: those with scores < 30 were encouraged to volunteer for the full study.

Only participants 25 years of age or older were accepted (with a few exceptions), to reduce the likelihood that participants might acquire late hours entirely through the social influences of youth culture or a college environment. Elderly participants were accepted if there was no suggestion of dementia or other illness which might distort circadian rhythms, and if there was a life-long history of delayed sleep.

Potential participants were provided written and verbal information about the study and signed written informed consent under the supervision of the UCSD Human Research Protection Program. Subjects referred or self-referred as potential DSPS cases completed a series of questionnaires, contributed a sample of blood or saliva for DNA, and the initial half of DSPS participants underwent a 2-week wrist activity and illumination recording, using a wrist-activity recorder (Kripke et al., 2008). DSPS participants were reimbursed $100 for their time and effort, which included completing the questionnaires, providing a blood or saliva sample, wearing the wrist actigraph for 2 weeks, completing sleep logs, and associated travel.

2.2. Control recruitment

While the DSPS case series was being assembled, the investigators recruited normal controls who claimed to have healthy and normal sleep. Control volunteers were recruited by word-of-mouth, outreach at health fairs and community meetings, the internet, and by a campus poster seeking volunteers who “sleep like a baby.” Control recruitment was targeted, so far as possible, to match the ancestry of the case series. Each control received an explanation of the study and signed written informed consent. They completed the questionnaires and contributed a sample of blood or saliva, but they were not asked to wear the actigraphs or to provide sleep logs. Since control participation usually required less than 2 hours, they were reimbursed $25. The principal investigator (DFK) reviewed the record of each participant volunteering as a control. Based on their questionnaires, control volunteers were retrospectively rated as 1) certain DSPS, 2) possible DSPS, 3) neither, 4) possible ASPS, or 5) certain ASPS. Only those with ratings 3 or 4 were included among the normal controls in this analysis.

2.3. Measures

2.3.1. Horne-Östberg questionnaire

An U.S.-idiom rephrasing was adopted for the Horne-Östberg morningness-eveningness questionnaire (Horne and Östberg, 1976; Terman et al., 2001a). This is a 19-item form yielding Horne-Östberg scores (HOS) from 16–30 (extreme eveningness) to above 70 (extreme morningness), as diagnostic ranges were originally defined.

2.3.2. BALM

The Basic Language Morningness scale (BALM) (Brown, 1993) was administered as a simplified-language revision of the 13-item Composite Scale (CS) (Smith et al., 1989). The CS was reported to have superior psychometric properties to the HOS as well as greater brevity.

2.3.3. SPAQ

All subjects completed a modified version of the Seasonal Pattern Assessment Questionnaire (SPAQ) (Rosenthal et al., 1987). The SPAQ is a retrospective, self-rating questionnaire that assesses seasonality of mood and behaviors. The first part of the SPAQ inquires “To what degree do the following change with the seasons?” for six items (sleep length, social activity, mood, weight, appetite, and energy level). For each item, responses were 0 (none), 1 (mild), 2 (moderate), 3 (marked), or 4 (extreme). The sum of these six items is the Global Seasonality Score (GSS), which therefore ranges from 0 (no seasonality) to 24 (extreme seasonality). Another section of the SPAQ requests that respondents fill in the month or months of the year (if any) when they feel best, feel worst, eat least, eat most, etc. The severity of problems with seasonality was also queried by “How much of a problem are these seasonal changes?” with choices being 0 (none), 1 (mild), 2 (moderate), 3 (marked), 4 (severe), or 5 (disabling). Another SPAQ section evaluated the weight fluctuation during the course of the year in 6 levels (1, 0–3 lbs; 2, 4–7 lbs; 3, 8–11 lbs; 4, 12–15 lbs; 5, 16–20 lbs; 6, Over 20 lbs). Sleep duration including napping in each season, and changes of food preference during the seasons were also reported by subjects. We calculated the sleep duration difference between winter and summer and included this difference in our analyses.

2.3.4. Other measures

Participants completed the Mood Disorders Questionnaire (MDQ), a lifetime screening scale for history of mania with very high specificity (Hirschfeld et al, 2000). They also completed the Quick Inventory of Depression Symptomatology Self Report (QIDS-SR), a self-rating scale for current major depression with high correlation to the psychiatrist-administered Hamilton Depression rating scale (Rush et al., 2003; Rush et al., 2006). Residence of subjects and test date were recorded. Subjects completed other questionnaires which will be reported in separate papers.

2.4. Clinical Evaluation

2.4.1. DSPS

Once all data were assembled, the principal investigator (DFK) reviewed the record of each participant who had volunteered as a DSPS case and recorded the participant’s DSPS classification as 1) absolutely certain, 2) fairly certain, 3) questionable, 4) unlikely, or 5) very doubtful. The main criterion for classification was the HOS, recognizing that the criterion for definite evening type of < 30 was too strict for the San Diego population (Kripke et al., 2008). Confirmatory classification criteria included the score on the BALM, reported prior-week and adult-life bedtimes and awakening times, the actigraphic recordings, and whether the participant reported going to sleep “somewhat later” or “much later” than most people their age, both as a child and as an adult. Whether the participant reported distress about falling asleep, reported related social or vocational problems, or had sought medical attention for a sleep problem was also considered. The consistency of the data supporting a classification of DSPS was evaluated, together with the presence of depression, other mental illnesses, or other sleep disorders which might confuse the classification. However, if depression or other disorders had their first incidence after the onset of delayed sleep and did not appear to be causing the delay, these disorders were not considered exclusionary. As many DSPS patients cannot consistently report to work by 8–9 AM, evening or night shift work was not considered exclusionary if the history indicated that the delay in sleep occurred before shift work was adopted, and the delay tended to persist when the participant was off work.

2.4.2. SAD

SAD and subsyndromal SAD were diagnosed by Kasper’s criteria based on SPAQ scores (Kasper et al., 1989). To be diagnosed with SAD, subjects must have met the following criteria: (1) GSS of 11 or more; (2) a ‘problem’ rating with seasonal mood changes of at least a moderate degree (2 or above); and (3) feeling worst during one of the winter months. To be categorized as having subsyndromal SAD (S-SAD), the subjects would either have a GSS of 11 or more with a ‘problem’ rating of not more than mild (0 or 1), or have a GSS of 9 or 10 with a ‘problem’ rating of at least mild (1 or above). Furthermore, the S-SAD subject had to feel worst in the winter months. We classified both SAD and S-SAD as a winter “seasonals” category and all other participants were classed as “non-seasonals”.

2.5. Statistical Analysis

We compared the differences of basic, clinical and seasonal characteristics between DSPS and controls by t-tests for continuous data or by χ² tests for categorical data. We also performed comparisons among SAD, S-SAD, and non-seasonals by analyses of variance (ANOVAs) for continuous data or by χ² tests for categorical data. All of the analyses were performed using SPSS for Windows version 11 (SPSS, Chicago, Ill), with the cutoff for statistical significance set at P < 0.05.

3. Results

3.1. Subjects

By the recruitment process described above, we recruited a case series of 359 DSPS along with 389 controls. Out of these subjects, 658 participants (327 DSPS and 331 controls) who completed the SPAQ were included in our analysis. Their mean age was 37.84 (SD=11.94). The gender distribution was 69.4% female and 30.6% male. Among 658 participants, 29 (4.4%) were categorized as a SAD, 44 (6.7%) were categorized as a S-SAD, and 585 (88.9%) were categorized as non-seasonals according to Kasper’s criteria (Kasper et al., 1989). Residents of California were 75.4% and the others were residents of 42 different states. The seasonal distribution of questionnaire completions was 20.8% winter, 21.0% spring, 28.6% summer, 26.9% fall, and 2.7% unknown.

3.2. DSPS and SAD

Figure 1 shows the difference in prevalence of SAD and S-SAD between DSPS and controls. The prevalence of SAD and S-SAD in DSPS were 6.7% and 8.6% respectively, which were significantly higher than the prevalence in controls (SAD, 2.1%; S-SAD, 4.8%) (χ²=12.65, p=0.002). DSPS were 3.3 times more likely to report SAD (odds ratio, 3.34; 95% CI, 1.41–7.93) and 2.4 times more likely to be seasonals (SAD + S-SAD) (odds ratio, 2.42; 95% CI, 1.44–4.07) compared with controls.

Figure 1.

Differences in prevalence of SAD and S-SAD between DSPS and controls.

Abbreviation; DSPS, delayed sleep phase syndrome; SAD, seasonal affective disorder; S-SAD, subsyndromal SAD

Table 1 shows comparisons of basic, clinical and circadian characteristics between DSPS and controls. DSPS showed significantly higher seasonality scores compared to controls in mood, appetite, and energy level subscores and global seasonality scores (t=3.12, t=0.002; t=2.04, p=0.041; t=2.64, p=0.008; and t=2.15, p=0.032, respectively). Although 3 seasonality subscores (weight, sleep length, and social activity) were not significantly different between DSPS and controls, other SPAQ items concerning weight fluctuation during seasons and sleep duration variation between winter and summer were significantly higher in DSPS (t=5.16, p<0.001; t=2.64, p=0.009, respectively). Thus among 6 seasonality characteristics, all except social activity were significantly different between DSPS and controls.

Table 1.

Contrast between DSPS and controls.

	DSPS (n=327)	Controls (n=331)
Sex (male/female)	90/237	110/221	X2=2.54	p=0.11
Age (yrs)	37.37±11.13	38.31±12.68	t=−1.00	p=0.32
Horne Östberg score	25.40±5.73	55.56±10.77	t=−44.91	p<0.001
BALM	17.86±4.11	38.28±7.96	t=−41.44	p<0.001
QIDS-SR	5.72±3.90	3.75±3.37	t=6.94	p<0.001
MDQ	3.66±3.81	3.36±3.60	t=1.02	p=0.31
SPAQ
Global Seasonality Score	5.48±4.90	4.71±4.23	t=2.15	p=0.032
Sleep length seasonality	0.79±1.01	0.67±0.97	t=1.53	p=0.13
Social activity seasonality	1.07±1.15	1.00±1.03	t=0.83	p=0.41
Mood seasonality	1.16±1.16	0.90±1.01	t=3.12	p=0.002
Weight seasonality	0.67±0.89	0.65±0.82	t=0.35	p=0.73
Appetite seasonality	0.66±0.90	0.50±1.09	t=2.04	p=0.041
Energy level seasonality	1.12±1.11	0.91±0.95	t=2.64	p=0.008
Problem Severity of Seasonality	0.61±1.06	0.30±0.72	t=4.50	p<0.001
Seasonal weight fluctuation*	2.28±1.21	1.85±0.94	t=5.16	p<0.001
Sleep duration-winter (hrs)	8.18±1.95	7.94±1.35	t=1.88	p=0.061
Sleep duration-spring (hrs)	7.57±1.70	7.54±1.25	t=0.26	p=0.80
Sleep duration-summer (hrs)	7.43±1.71	7.44±1.30	t=−0.072	p=0.94
Sleep duration-fall (hrs)	7.73±1.66	7.68±1.29	t=0.46	p=0.65
Sleep duration difference (winter – summer) (hrs)	0.75±1.37	0.50±1.05	t=2.64	p=0.009
Food preference change (yes/no)	121/206	127/204	X2=0.13	p=0.72

SPAQ, Seasonal Pattern Assessment Questionnaire; BALM, Basic Language Morningness Scale; QIDS-SR, Quick Inventory of Depression Symptomatology-Self Report; MDQ, Mood Disorders Questionnaire

^*1, 0–3 lbs; 2, 4–7 lbs; 3, 8–11 lbs; 4, 12–15 lbs; 5, 16–20 lbs; 6, Over 20 lbs

Depression scores evaluated by QIDS-SR were significantly higher in DSPS than controls (t=6.94, p<0.001), however, the life-time mania history evaluated by MDQ was not different between them (t=1.02, p=0.31).

3.3. Comparison among SAD, S-SAD and non-seasonals

Table 2 shows contrasts of basic, clinical and circadian characteristics among three groups (SAD, S-SAD and non-seasonals) divided by SPAQ seasonality criteria irrespective of DSPS or control status. The BALM scores and Horne-Östberg scores (HOS), which evaluate morningness-eveningness, were significantly lower in SAD and S-SAD groups than non-seasonals (HOS, F(2,655)=5.20, p=0.006; BALM, F(2,655)=6.26, p=0.002). This indicated that SAD and S-SAD groups tended to report less morningness (more eveningness) than non-seasonals. Depression scores evaluated by QIDS-SR were significantly higher in SAD and S-SAD than non-seasonals (F(2,655)=47.50, p<0.001), Furthermore, scores in MDQ, a lifetime screening scale for mania, were significantly higher in SAD and S-SAD than non-seasonals (F(2,655)=15.63, p<0.001). Sleep duration during the seasons highlighted interesting contrasts, in which SAD reported sleeping less in summer but more in winter compared to non-seasonals (summer, F(2,652)=4.90, p=0.008; winter, F(2,652)=6.32, p=0.002). The mean sleep length difference between winter and summer of SAD subjects was 2.24 hours, and this was significantly longer than S-SAD and non-seasonals (F(2,652)=40.42, p<0.001). SAD and S-SAD showed more food preference changes during the seasons (χ²=16.45, p<0.001). Most frequently they reported more carbohydrate intake during winter. Due to the previous reports that SPAQ scores are influenced by the latitude and season, we analyzed the relationships between seasonality, questionnaire completion dates, and latitude. The latitudes were not significantly different among SAD, S-SAD and non-seasonals (SAD, 33.89±3.29; S-SAD, 34.12±3.46; non-seasonals, 33.55±3.58; F(2,637)=0.60, p=0.55). The seasonal distribution of questionnaire completions were also not different among three groups (χ²=8.1, p=0.23).

Table 2.

Contrasts between SAD, S-SAD and non-seasonals irrespective of DSPS or control status

	SAD (n=29)	S-SAD (n=44)	Non-seasonals (n=585)
Sex (male/female)	5/24	13/31	182/403	X2=2.53	p=0.28
Age (yrs)	41.59±12.86	35.75±10.38	37.81±11.98	F(2,655)=2.11	p=0.12
Horne-Östberg score	33.86±13.31a	34.82±14.81	41.34±17.61a	F(2,655)=5.20	p=0.006
BALM	22.55±8.68a	24.18±10.43b	28.70±12.17a,b	F(2,655)=6.26	p=0.002
QIDS-SR	9.55±5.00 a	7.68±4.72 b	4.26±3.34 a,b	F(2,655)=47.50	p<0.001
MDQ	6.52±4.08a	5.07±3.85b	3.24±3.59a,b	F(2,655)=15.63	p<0.001
SPAQ
Global Seasonality Score	15.79±3.31a,c	11.41±0.47b,c	4.09±3.62a,b	F(2,655)=227.79	p<0.001
Sleep length seaonality	2.52±1.02 a,c	1.61±0.99 b,c	0.58±0.86 a,b	F(2,655)=90.89	p<0.001
Social activity seasonality	2.76±0.99 a	2.30±0.77 b	0.86±0.98 a,b	F(2,655)=93.97	p<0.001
Mood seasonality	3.10±0.67 a,c	2.48±0.66 b,c	0.82±0.94 a,b	F(2,655)=146.75	p<0.001
Weight seasonality	2.21±0.90 a,c	1.43±1.04 b,c	0.52±0.72 a,b	F(2,655)=93.45	p<0.001
Appetite seasonality	2.21±0.94 a,c	1.39±0.89 b,c	0.44±0.91 a,b	F(2,655)=70.76	p<0.001
Energy level seasonality	3.00±0.76 a,c	2.20±0.67 b,c	0.82±0.90 a,b	F(2,655)=127.06	p<0.001
Problem Severity of Seasonality	2.93±1.16 a,c	0.84±0.96 b,c	0.30±0.69 a,b	F(2,655)=181.48	p<0.001
Seasonal weight fluctuation*	3.21±1.24a,c	2.41±1.30 b,c	1.98±1.04 a,b	F(2,655)=20.62	p<0.001
Sleep duration-winter (hrs)	9.07±2.24 a	8.30±1.86	7.99±1.61 a	F(2,652)=6.32	p=0.002
Sleep duration-spring (hrs)	7.76±1.38	7.50±1.64	7.55±1.48	F(2,652)=0.31	p=0.74
Sleep duration-summer (hrs)	6.83±1.37	6.98±1.66	7.50±1.51	F(2,652)=4.90	p=0.008
Sleep duration-fall (hrs)	8.00±1.46	7.59±1.62	7.70±1.47	F(2,652)=0.71	p=0.49
Sleep duration difference (winter – summer) (hrs)	2.24±1.86 a,c	1.32±1.39 b,c	0.49±1.09 a,b	F(2,652)=40.42	p<0.001
Food preference change (yes/no)	20/9	22/22	206/379	X2=16.45	p<0.001

SAD, seasonal affective disorder; S-SAD, subsyndromal SAD; SPAQ, Seasonal Pattern Assessment Questionnaire; BALM, Basic Language Morningness Scale; QIDS-SR, Quick Inventory of Depression Symptomatology-Self Report; MDQ, Mood Disorders Questionnaire

^*1, 0–3 lbs; 2, 4–7 lbs; 3, 8–11 lbs; 4, 12–15 lbs; 5, 16–20 lbs; 6, Over 20 lbs

^a,b,cmeans significantly different 2 × 2 according to Bonferroni post hoc tests

4. Discussion

The pathophysiologic mechanisms of DSPS remain to be clarified. Genetic factors, especially in circadian oscillator genes, may be associated with this syndrome (Archer et al., 2010; Takano et al., 2004). Our group is performing circadian genetic association studies on DSPS using this sample and will report results separately. In addition to genetic factors, environmental factors such as exposure to light are important in the development of DSPS. Decreased light exposure in the morning can exacerbate DSPS.

SAD has been regarded as an ideal model for understanding the role of circadian rhythms in human mood and behavior, because of predictable depression in winter and predictable recovery as days lengthen in the spring. SAD patients become depressed in the winter at least in part because of a phase delay with respect to the sleep/wake cycle (Lewy et al., 1987) and bright light should usually be scheduled in the morning in order to provide a corrective phase advance. However, studies of melatonin administration suggest a more complicated etiology (Lewy et al., 2006; Rahman et al., 2010).

Both DSPS and SAD have been considered to arise from delayed circadian phases; however, this study was the first to demonstrate comorbidity in a large number of cases. This comorbidity supports the hypothesis that DSPS and SAD share similar pathophysiological mechanisms of delayed circadian phase and low mood.

However, there must be further factors explaining why DSPS experience fluctuating SAD symptoms in winter, if their circadian phases are delayed throughout the year. Two possibilities are that 1) the timing of bright light exposures may change in reference to internal rhythms and 2) changes may occur in the internal phase-angles between melatonin-related circadian rhythms and the sleep-wake cycle. These alternatives resemble the internal coincidence and external coincidence models of photoperiodic responses and the phenomena of photoperiodic sensitivity to changing photoperiod independent of absolute photoperiod duration (Lewy et al., 2006; Terman et al., 2001b; Lewy, 2010).

Patients with major depression which does not meet SAD criteria apparently tend to display delayed melatonin in reference to sleep timing throughout the year (Sekula et al., 1997; Tuunainen et al., 2002; Emens et al., 2009). The internal phase malsynchronization is hypothesized to have a causal role in the depression. In contrast, patients with DSPS, in the original description, have no malsynchronization between sleep and internal rhythms (Weitzman et al., 1981). However, careful study showed that patients with DSPS may display melatonin rhythms which are either early or late in reference to their sleep times, and vice versa (Uchiyama et al. 2000; Cole et al., 2002). In DSPS, both melatonin and sleep are usually late, but sometimes melatonin might display timing within the normal range or even somewhat advanced (Hughes and Lewy, 1998). The possibility of subgroups of different pathophysiology in DSPS and SAD may be able to explain why most DSPS did not have seasonality and why some seasonals do not have DSPS in our sample. Patients with overlap of SAD and DSPS have a delay in circadian rhythms with respect to the sleep/wake cycle. These patients should preferentially respond to the morning bright light treatment so as to provide a corrective phase advance and restore optimum alignment between the endogenous circadian rhythms and those rhythms that are related to the sleep/wake cycle. Therefore, for the overlap patients, it is recommended to undergo bright light treatment as soon after the usual awakening time as possible.

Another interesting finding is that SAD is related to bipolar disorder. The MDQ is a brief, self-report screening instrument that can be used to identify patients most likely to have bipolar disorder (Hirschfeld et al. 2000). The mean MDQ score of SAD was 6.52 (SD=4.08), which was significantly higher than non-seasonals (3.24±3.59) and close to the score of ‘7’ that was chosen as the optimal cutoff of MDQ for bipolar disorder in the previous study (Hirschfeld et al. 2000). This is another indication that SAD is often related to bipolar disorder, as was reported in the original description (Rosenthal et al., 1984).

There are some limitations in this study. First, as a diagnostic tool, the SPAQ discriminates poorly between SAD and S-SAD, and hence it has a poor case-finding ability for SAD (Magnusson, 1996). Furthermore, it is a self-report scale which does not always agree well with longitudinal observations (Lund and Hansen, 2001; Christensen et al., 2003). Second, the Southern California concentration of the subjects is a limitation. This restricted our opportunity to observe an association between latitude and SAD prevalence and may have produced a relatively low prevalence of SAD and S-SAD compared to the U.S. average. Nevertheless, the large scale of the DSPS and well-matched control sample are appropriate to the aim of study. Further studies for the molecular mechanisms and similar pathophysiology of DSPS and SAD are needed.