Author's response to reviews

Title: Seasonality in Depression and Anxiety; Results from the NetherlandsStudy of Depression and Anxiety

Authors:

Wim H Winthorst (w.h.winthorst@umcg.nl)Wendy J Post (w.j.post@rug.nl)Ybe Meesters (y.meesters@umcg.nl)Brenda W.H.J Penninx (b.penninx@vumc.nl)Willem A. Nolen (w.a.nolen@umcg.nl)

Version: 2 Date: 25 November 2011

Author's response to reviews: see over

1 Revision MS 6137935375847103

Revision Cover Letter Groningen, November 25, 2011 Dr. Daniel Smith and Mrs. Catherine Olino Journal Editorial Office BioMed Central Dear Dr Smith and Mrs Olino, We are grateful for the reviewer’s comments on our manuscript "Seasonality in Depression and Anxiety; Results from the Netherlands Study of Depression and Anxiety” (MS: 6137935375847103). We carefully considered all comments and we believe that we were able tot adequately address all the suggestions proposed by the reviewers. We hope that this revised manuscript will be suitable for publication in your esteemed journal, BMC Psychiatry. In this letter we included a detailed overview of all changes made in response tot the suggestions by the reviewers. Comments of the reviewers are in bold and the changes tot the manuscript are in bold italics (new text) or displayed with strikeouts (deleted text). The answers to the comments are numbered I look forward to your response. Yours sincerely, W.H. Winthorst, M.D. Corresponding author Department of Psychiatry, University Medical Centre Groningen, University of Groningen. Hanzeplein 1 9713 GZ Groningen The Netherlands. Email address: w.h.winthorst@umcg.nl Tel: 0031-50- 612008

2 Revision MS 6137935375847103

Reviewer 1 - Discretionary Revisions: (1.1). My understanding is that the paper reported the results of two different sub-studies; one conducted using the K-10 on a “primary care recruitment population” and one using a larger set of scales on the “NESDA baseline population”. This is well explained in the “Methods” section, but I found it difficult to understand this from the title and from the abstract (because of the first sentence in the methods session “Data were used from the Netherlands Study of Depression and Anxiety”), where I understand that all subjects were part of the NESDA study. Perhaps the title and the abstract may be revised and the main division between the 2 sub-studies can be maintained in the results section, putting the last five sections (Inventory of depressive symptoms, total score, Inventory of depressive symptoms, atypical symptoms, Inventory of depressive symptoms, melancholic symptoms, Beck Anxiety Inventory, Fear Questionnaire) as sub-sections of “NESDA baseline population”. (Reaction 1.1): We thank the reviewer for this comment. Indeed the paper reported the results of two different sub-studies of NESDA; one conducted using the K-10 in a “primary care recruitment population” and one using a larger set of scales on the “NESDA baseline population”. The confusing point might be that recruitment phase, in which the respondents from the primary care were selected, was part of the NESDA study. Therefore we changed the term “primary care recruitment population” into “NESDA primary care recruitment population” and reframed the first paragraph of the methods section. Manuscript: page 6

This study was conducted using data from the primary care recruitment

population and from the baseline population of the Netherlands Study of

Depression and Anxiety (NESDA, www.nesda.nl) [17].

NESDA is an ongoing multi-site naturalistic 8-year longitudinal cohort study

among 2,981 adults (18-65 years), aimed at describing the long-term course

and consequences of depressive and anxiety disorders. The NESDA sample is

stratified for setting: community (n=564), primary care (n=1610) and

specialized mental health (n=807). The NESDA community sample was built

on two cohorts that were already available through prior studies described in

detail elsewhere [18]. The NESDA primary care respondents were recruited

among practices of 65 general practitioners (GPs) in the vicinity of the three

research sites. The specialized mental health patients in NESDA were

recruited from outpatient clinics of regional facilities for mental health care

around the research sites.

3 Revision MS 6137935375847103

Revised version of the manuscript: page 6 The study was conducted using data from the Netherlands Study of Depression

and Anxiety (NESDA, www.nesda.nl): (1) the NESDA primary care

recruitment population and (2&3) the NESDA baseline population [18].

NESDA is an ongoing multi-site naturalistic 8-year longitudinal cohort study

among 2,981 adults (18-65 years), aimed at describing the long-term course

and consequences of depressive and anxiety disorders. The NESDA sample

(total n=2981) is stratified for setting: community, primary care and

specialized mental health care. The community sample (n=564) was built on

two cohorts that were already available through prior studies described in

detail elsewhere [19]. The primary care participants (n=1610) were recruited

among 23,750 patients from practices of 65 general practitioners (GPs) in the

vicinity of three research sites. The specialized mental health patients (n=807)

were recruited from outpatient clinics of regional facilities for mental health

care around the research sites.

In order to make clear that the manuscript reports the results of two different sub-studies of NESDA we changed the title:

Manuscript: page 1 Seasonality in Depression and Anxiety; Results from the Netherlands Study of

Depression and Anxiety

Revised version of the manuscript: page 1 Seasonality in depressive and anxiety symptoms among primary care

patients and in patients with depressive and anxiety disorders; Results from

the Netherlands Study of Depression and Anxiety

In reaction on comment 2.8 of reviewer 2 we changed the headings of the paragraphs of the result section according to the respective research questions. The numbers of the tables and figures changed correspondingly (see also reaction 2.8). (1.2). Is there any overlap between the two populations? Are the 47 GPs of the “primary care recruitment population” part of the 65 GPs of the NESDA study? (Reaction 1.2): Yes, there is an overlap. The total of 65 GPs consists of all GPs in the field sites Amsterdam, Leiden en Groningen. The respondents from 16 GPs of the field site Leiden were excluded form the analysis because the date they filled in their questionnaire could not be recovered. Regarding this point we also found a minor mistake: We analysed the data of the remaining 5449 respondents who were the patients of 44 GPs (instead of 47 GP’s) from the field sites Amsterdam and Groningen. In Amsterdam 3 GPs were in fact substitutes for the regular GP’s. In our analyses we erroneous did

4 Revision MS 6137935375847103

count them in. In order to adhere to the overall NESDA methods paper by Penninx et al we substituted the number of 47 GPs by 44 GPs. Because the statistical analysis did not concern the number of GPs this does not affect the remainder of the manuscript.

Revised version of the manuscript: page 11 Because these dates were not recorded in Leiden the respondents from the

field site Leiden were excluded from the analysis. Off the remainder 14 K-10 questionnaires had 2 or more answers missing; and were excluded as a

consequence. The resuming 5,549 respondents from 44 GPs were included in

the analysis, consisting of 3664 (66%) women and 1885 (34%) men. The mean

age was 43.6 years (SE = 0.17).

(1.3). Methods, section “Subjects”, text “The primary care recruitment population to whom a questionnaire was sent”: I would specify what questionnaire (is it the K-10?) (Reaction 1.3): We specified the questionnaire in the following sentence: The primary care recruitment population, to whom a questionnaire was sent, consisted

of a random sample of all patients who had visited their GP during the previous four

months for any reason from January 2004 to February 2007.

Revised version of the manuscript: page 7 The NESDA primary care recruitment population to whom a the Kessler-10

screening questionnaire was sent, consisted of a random sample of all patients

who had visited their GP during the previous four months for any reason from

January 2004 to February 2007. (1.4). Methods. Although the role of latitude in the genesis of seasonal affective symptoms is controversial, it may be helpful for the reader to know the latitude of the areas. Ethnicity/country of origin have been reported as important factors for the effect of seasonality on mood (see, for example, Guzman et al, ScientificWorldJournal, 2007). Were data on ethnicity/country of origin collected? (Reaction 1.4): We agree that the role of latitude is a relevant point and therefore we added information concerning this topic in the methods section (see also reaction 2.3).

Revised version of the manuscript: page 7 The date the questionnaire was filled out was recorded for respondents from

the research sites in Amsterdam and Groningen (latitude 52,3 ○ and 53,2

○

respectively).

5 Revision MS 6137935375847103

Data on ethnicity and country of origin were collected in NESDA but not used in this analysis in order to limit the number co-variables in the analysis. 5. � (See Reaction 1.5). Methods and suggestion for further analyses/discussion. In the NESDA sample, subjects with major depressive disorder or dysthymic disorder were included. Were subjects with dysthymic disorder included also in the “Major depressive disorder” group of the current analysis? If so, it may be that seasonality has a different effect on DSM-IV major depression and DSM-IV dysthymic disorder. As stated in the “discussion”, “seasonality is usually associated with minor depression, and that once the threshold for a major depression has been passed, seasonality is less influential”. Although it is unclear whether dysthymia is fundamentally different from chronic MDD, by definition, dysthymia is characterized by depressive symptoms that are milder in severity than those found in major depressive disorder. So, it is possible that seasonality had an effect on subjects with dysthymic disorder, but this effect was masked by the subjects with a more severe episode. Moreover, the response to treatment is different in the two disorders and, as stated in the discussion “seasonal effects may be masked by treatment effects”. The relatively big sample size (n=131) should allow this further sub-analysis. (Reaction 1.5): This is an interesting point which was mentioned by reviewer 2 as well (2.9). Our study did not include participants with a minor depression or dysthymic disorder. The consideration was to enhance the contrast between healthy controls (HC) and participants in de the depression group (MDD). In a recent meta-review of depressive subtyping models (Baumeister, H., Parker, G., Meta-review of depressive subtyping models, J. Affect. Disord. 2011,doi:10.1016/j.jad.2011.07.015 ), Baumeister and Parker raised the question whether SAD is a distinct depressive subtype or rather a pronounced seasonality trait of patients with fluctuating minor and major depressions. As our study only used cross-sectional baseline data from NESDA, and did not concern participants diagnosed with a minor depression or SAD, this question could not be addressed. However, it remains an interesting hypothesis that can be tested in future analyses of the longitudinal NESDA data. 6. � (See Reaction 1.6). Methods and suggestion for further analyses/discussion. Were the subjects in the 4 NESDA groups free from any physical health problem? If not, it may be that the seasonality has a different effect on “endogenous affective disorders” and “affective disorders secondary to a physical illness”. (Reaction 1.6): The reviewer addresses an interesting point. We didn’t include these data in our analysis in order tot limit the scope of the article.

6 Revision MS 6137935375847103

7. � (See Reaction 1.7). In the Discussion, two studies on the prevalence of seasonal affective disorders (SAD) in the primary care population are reported. I would report also Eagles et al, Br J Psychiatry, 1999 (citation number 7 in this paper), who, employing the Seasonal Pattern Assessment Questionnaire, found a rate of SAD of 9.7% and a rate of 'sub-syndromal SAD’ of 11.3%. (Reaction 1.7): We added this study to the discussion about the primary care respondents.

Revised version of the manuscript: page 20 For example Blazer et al. [17] found a prevalence of 0,4 % of major

depression with a seasonal pattern and 1 % of major or minor depression with

a seasonal pattern in a community based sample, Levitt et al. [27] found a

prevalence of 2,9% of seasonal affective disorder in a community sample,

Eagles et. al. [6] found a prevalence of 5.3% of seasonal affective disorder during the winter months in a primary care population and Thompson et al.

[1] found a prevalence of 5.6% of seasonal affective disorders in a primary

care population.

(1.8). In the discussion of the limitations, I would include that the assessment of the affective symptoms was conducted using self-reported measures. For example, in the study of Eagles et al, 1999, only the 41% of SPAQ cases of SAD, also fulfilled DSM-IV and Hamilton Rating Scale for Depression - Seasonal Affective Disorder Version criteria. A similar discrepancy may have occurred also in the evaluation of the season pattern of affective symptoms in the current sample. (Reaction 1.8): In this study the CIDI (WHO version 2.1) was used to establish diagnoses according to DSM-IV criteria (American Psychiatric Association, 2001). The assessment of depressive and anxiety symptoms was conducted with self-reported measures (IDS-SR, BAI en FQ). This is now mentioned for all instruments in the method section. The self report measures were taken independently from the CIDI interview within 7 days before or after the CIDI interview. These measures had no influence on the diagnostic classification of the respondents.

Revised version of the manuscript: page 8 Within 7 days prior or after the CIDI interview, all participants completed several self-report questionnaires. Severity of depression over the past 7 days

was assessed with the Inventory of Depressive Symptoms, 30 item self-report

versions (IDS) [22]. Moreover, the IDS was used to assess the presence and

severity of atypical and melancholic features, as the IDS includes all

symptoms of these specifiers. Therefore a continuous atypical specifier was

constructed (At-IDS): a summation of the scores on the items mood reactivity,

the highest score of either weight gain or increase in appetite, hypersomnia,

7 Revision MS 6137935375847103

leaden paralysis, and interpersonal rejection sensitivity (score range 0 – 3,

total score range 0 – 15). The scores of the item mood reactivity were recoded

(reversed) resulting in an item that counts the presence of the symptom mood

reactivity in stead of its absence. Participants with one or more missing items

were excluded from the analysis. Also a continuous melancholic specifier was

constructed (Mel-IDS): a summation of the scores on the items: loss of

pleasure, lack of reactivity to usually pleasurable stimuli, depressed mood,

regularly worse in the morning, early morning awakening, psychomotor

retardation or agitation, the highest score of either anorexia or weight loss,

and excessive or inappropriate guilt (score range 0 – 3, total score range 0 –

24). Also for Mel-IDS participants with missing items were excluded from the

analysis.

The Beck Anxiety Inventory (BAI), a 21-item self-report instrument, was used

to assess overall severity of anxiety [23]. Finally the 15-item self-report

version of the Fear Questionnaire (FQ) was used to measure severity of fear

and avoidance [24].

Minor Discretionary Revisions: (1.9). As the distribution of the K-10 total score was skewed, median and range may be reported instead of mean and standard error. (Reaction 1.9): For reasons of unity over the measures taken we used the same descriptive statistics for the results of the NESDA primary care population and the NESDA baseline population. The distribution of the K-10 was skewed for the total group meaning that the majority of the respondents scored in the lower region and only few respondents scored in the higher regions. Using the median and range in this study would not enhance the comprehension of the data because the minimum score over the months was invariably 10 and the maximum score varied only from 46 to 50, with a random distribution over the months. We added the median and range in one sentence regarding the results of the K-10 scores. Manuscript: page 11

The observed total mean K-10 score was 19.2 (SE = 0.11) with the lowest

scores for summer and the highest scores for autumn.

Revised version of the manuscript: page 11 The observed total mean K-10 score was 19.2 (SE = 0.11), the median score

was 17 (range 10-50), the lowest scores were recorded in summer and the

highest scores in autumn.

2. � (See Reaction 1.10). In the discussions three possible explanations are reported for the discrepancy with the previous literature. I would put them either all in one paragraph or in three separate paragraphs (now the 1st is in a separate paragraph from the 2nd and 3rd explanations).

8 Revision MS 6137935375847103

(Reaction 1.10): We agree on this comment and have put the three explanations in one paragraph. Minor Essential Revisions 1. � (See Reaction 1.11). Background, second paragraph “the majority” [..] “HAS used” (Reaction 1.11): We changed “have” in “has”.

Revised version of the manuscript: page 4 The majority of the latter studies reported on seasonal affective disorder

(SAD), defined in DSM IV as a recurrent depressive disorder with a regular

temporal relationship between the onset of a major depressive episode and a

particular time of the year (mostly fall or winter) and have has used specific

instruments for its assessment [2, 3]. 2. � (See Reaction 1.12). Results, “NESDA baseline population” sub-section: Please review the sentence “For At-IDS and Mel-IDS 1074 resp. 1033 participants were included in the analysis”. (Reaction 1.12): We revised the following section:

Manuscript: page 12 For At-IDS and Mel-IDS 1074 resp. 1033 participants were included in the

analysis; 16 participants were excluded due to missing items on At-IDS

(1.5%) and 57 participants were excluded due to missing items on Me-IDS

(5.2%). For the BAI and the FQ there was one participant missing, resulting in

1089 included participants.

Revised version of the manuscript: page 12 The BAI and the FQ had one participant missing, resulting in 1089 included

participants. 16 Participants were excluded due to missing items on At-IDS

(1.5%) resulting in 1074 participants in the analysis of At-IDS. 57

Participants were excluded due to missing items on Me-IDS (5.2%) resulting

in 1033 participants in the analysis of Me-IDS.

3. � (See Reaction 1.13). Discussion, paragraph 5. “An explanation might be the probably low prevalence of seasonal affective disorders in this sample leading TO a very small effect on the mean scores of depression or anxiety symptoms in this sample.” (Reaction 1.13): We changed the typing error “tot” into “to”.

9 Revision MS 6137935375847103

Revised version of the manuscript: page 21 An explanation might be the probably low prevalence of seasonal affective

disorders in this sample leading tot a very small effect on the mean scores of

depression or anxiety symptoms in this sample. 4. � (See Reaction 1.14). Discussion, end paragraph 2. The [12] reference is written twice. (Reaction 1.14): We removed the double reference.

Reviewer 2 Major Compulsory Revisions: (2.1). The manuscript is very lengthy and unfocused. In the introduction, several findings of prior studies are reported. However, the following description of what is missing so far and what the aims of the present study are does not fit to this description. The authors outline that prior studies “did not measure seasonality in severity of atypical depressive symptoms, melancholic depressive symptoms and anxiety symptoms in specific patient groups with depressive and anxiety disorders”. I would expect that the present study aims to close these gaps. Thus, the research questions one and two seem to be irrelevant (as long as no other rational is provided). In summary, please specify more clearly what is already known and what your study will add to the current knowledge. Please focus on the latter. In its present form, it rather reads as a replication study of prior findings. (Reaction 2.1): We acknowledge this point. Our first point of interest (research questions 1 and 2 was whether we could demonstrate seasonal variation in depressive symptoms in a cohort of patient visiting their general practitioner for any reason and among patients with a current depressive and/or anxiety disorder. In some aspects it is indeed a replication study but the population under study is different in its compilation from most cited studies. Secondly we were interested whether we could demonstrate a seasonal variation in specific depressive symptoms (i.e. atypical and melancholic symptom profiles) because seasonal affective disorder is thought to be associated with atypical symptom profiles. This is an additional research question so we changed the first sentence of the third paragraph on page 5 of the original manuscript..

Revised version of the manuscript: page 5 In addition the studies mentioned above did not measure seasonality in

severity of atypical depressive symptoms, melancholic depressive symptoms

and anxiety symptoms in specific patient groups with depressive and anxiety

disorders.

10 Revision MS 6137935375847103

(2.2). Moreover, the manuscript would gain from deleting its focus on gender and seasonality, which has not been specified as a research question and thus should not be analysed in this study. (Reaction 2.2): We fully acknowledge this point. Gender and seasonality were not specified in our primary research questions. Before submitting the manuscript we had some internal discussions whether we should report on this (otherwise interesting) item for this reason. The reviewers comment supports our hesitation to report on this item and we decided to remove it from the discussion section in the manuscript. First version of the manuscript: page 21

Though the research questions did not primary concern gender, we did find

some differences between the sexes. Even though they may be chance findings

(i.e. false positive), we consider them interesting enough to report. Women

scored higher on depressive symptoms (IDS total score) than men in summer,

while they scored lower than men in winter, especially women with a MDD.

This is in line with findings of De Graaf who reported that among men major

depression and dysthymic disorder were more frequently observed in winter

than in summer [11]. Also, Blazer found that male gender was associated with

a higher prevalence among patients with MDD with a seasonal pattern [16].

For specific depressive symptoms, higher scores on atypical symptoms were

found among women than among men in all seasons with lowest scores in

autumn. In addition, women with a MDD scored lower than men on

melancholic symptoms in winter, while younger women scored higher on

atypical symptoms than older women. Combining these findings we

speculated that young women especially scored higher on atypical symptoms

in winter. This was confirmed by an additional three-way analysis (data not

shown). This is in line with literature regarding depressions with a seasonal

pattern, mostly defined as atypical depressions, that reports on a higher

frequency especially among young women [6, 37-39].

Regarding fear symptoms as measured with the FQ women scored higher in

summer and autumn. Comorbidity rates among depressive and anxiety

disorders have been reported to be high (30% -60%) and those disorders often

arise sequentially within the same patient [17]. Results in this study indicate

that women score higher on atypical symptoms in winter while they score

higher on fear symptoms in summer and autumn. One explanation for these

results might be that women are more vulnerable to complaints of social

phobia and agoraphobia (as measured with the FQ) in the outgoing seasons

like summer and autumn while during the winter months the symptom profile

might shift towards more atypical depressive symptoms.

Minor Essential Revisions: (2.3). - P4: “Various factors have been described to be associated with a higher prevalence of SAD: a higher latitude…”: The evidence for this statement is inconclusive (see Sohn, C.H., Lam, R.W., 2005. Update on the biology of seasonal affective disorder. CNS Spectr. 10, 635–646.;

11 Revision MS 6137935375847103

Baumeister H, Parker G.2011. Meta-review of depressive subtyping models. Journal of Affective Disorders, doi:10.1016/j.jad.2011.07.015). (Reaction 2.3): We agree on this comment and changed the text accordingly. Manuscript: page 4

Various factors have been described to be associated with a higher prevalence

of SAD: a higher latitude [5], female gender and young age[6, 7].

Revised version of the manuscript: page 4 Female gender and young age have been described to be associated with a

higher prevalence of SAD [5, 6]. The influence of a higher latitude on the

prevalence of SAD has been suggested but could not be demonstrated [7, 8].

(2.4 ). P19: “one explanation might be that patients visiting their GP have an elevation in mood and anxiety scores due to their medical condition…These effects might mask moderate seasonal differences”: As far as I know, there is no evidence for such an assumption. (Reaction 2.4): We agree on the comment that we have no literature supporting this assumption so we removed the statement. Manuscript: page 19

One explanation might be that patients visiting their General Practitioner have

an elevation in mood and anxiety scores due to their medical condition or

worry’s about it. These effects might mask moderate seasonal differences.

(2.5). - P20: “which indicates that specific diagnostic instruments are required to demonstrate seasonality”: I can´t see the findings that support this conclusion. Based on your manuscript it is also possible that there just is no marked seasonality. I would recommend altering this statement in the discussion and conclusion section. (Reaction 2.5): We agree with the reviewer that on a group level, in a cross sectional study utilising general instruments we didn’t see marked seasonality. In the text we tried to point out that in individual cases seasonality or seasonal affective disorder may be present. In that case treatment should (also) be targeted on seasonality. We decided to remove the designated part of the sentence from of the discussion and rephrased the paragraph. We also changed one sentence in de conclusion section to a more general statement on the topic.

Manuscript: page 20

The results of this study are in line with studies in the general population that

failed to demonstrate seasonal fluctuation in the prevalence of depression

using general diagnostic instruments like the CIDI and the BDI [9, 11], which

12 Revision MS 6137935375847103

indicates that specific diagnostic instruments are required to demonstrate

seasonality.

Manuscript: page 20

However the results of this study are supported by Magnusson et al. who

reported on the lack of seasonality in anxiety and depression, measured with

the HADS, in the Icelandic population and by Blacker et al., who found no

significant seasonal variation in General Health Questionnaire scores in a

primary care population [10, 12, 12]

Revised version of the manuscript: page 19 However the results of our study are in line with studies in the general

population that failed to demonstrate seasonal fluctuation in the prevalence

of depression using general diagnostic instruments like the BDI and the

CIDI [10, 12]. The results of our study are also supported by Magnusson et

al. [11] who reported on the lack of seasonality in anxiety and depression,

measured with the HADS, in the Icelandic population and by Blacker et al.

[13], who found no significant seasonal variation in General Health

Questionnaire scores in a primary care population.

Manuscript: page 23 Conclusions

Seasonal differences in severity or type of depressive and anxiety symptoms,

as measured with a general screening instrument and symptom questionnaires,

were absent or small in effectsize in a primary care population and in patient

populations with a major depressive disorder and anxiety disorders.

Seasonality may be masked by treatment. For the detection of specific

subgroups of patients with a seasonal pattern in depressive episodes more

specific questionnaires are needed. These analyses will be forthcoming in our

next study on NESDA data.

Revised version of the manuscript: page 22 Conclusions Seasonal differences in severity or type of depressive and anxiety symptoms,

as measured with a general screening instrument and symptom questionnaires,

were absent or small in effectsize in a primary care population and in patient

populations with a major depressive disorder and anxiety disorders. For the

detection of individuals with a seasonal pattern in depressive episodes more specific questionnaires and a longitudinal approach are needed. These

analyses will be forthcoming in our next study on NESDA data.

(2.6 ). - P20: “A possible explanation is that seasonal mood changes are masked by use of medication or other types of treatment”: The prominence of this explanation in the discussion, conclusion and

13 Revision MS 6137935375847103

abstract seems out of place given the only modest effectiveness of antidepressant therapies and the reported seasonal patterns of initiating antidepressant therapy in GP (Gardarsdottir et al. “5-35% more patients initiating use during winter than summer”). However, it would be very interesting, if the authors could examine this hypothesis based on their GP data (e.g. dichotomized into patients with vs. without currentantidepressant therapy). (Reaction 2.6): We agree with the reviewer that the unconfirmed role of medication is too prominent in the discussion, conclusion and abstract. We therefore made the following changes to the text:

Revised version of the manuscript: page 21 A possible explanation is that According to the results of this study it can not

be ruled out that seasonal mood changes are masked by use of medication or

other types of treatment [27-30].

In the revised version of the manuscript, abstract (page 3) and conclusion section (page 22) the following sentence was deleted: Seasonality may be masked by treatment.

We agree with the reviewer that it would be interesting tot examine this hypothesis in the primary care population. Unfortunately these data were not collected at the time of administration of the K-10. Discretionary Revisions (2.7). - Operationalization of “with melancholic” and “with atypical” features: The DSM-IV operationalization of these subtypes is controversial. It would be very interesting to examine whether your findings change if alternatively proposed conceptualisations of these subtypes are used (see Baumeister H, Parker G. 2011. Meta-review of depressive subtyping models. Journal of Affective Disorders, doi:10.1016/j.jad.2011.07.015 for an overview about alternatively discussed proposals). Otherwise, your manuscript might benefit from discussing this issue as a potential methodological reason for your findings (i.e. hypothesis: misconceptualized specifiers which lead to inconclusive results). (Reaction 2.7): We agree with the reviewer on the importance of this subject. We added the following paragraph to our discussion:

14 Revision MS 6137935375847103

Revised version of the manuscript: page 21 Another explanation might be an insufficient specificity of the atypical and

melancholic specifiers as defined by the DSM IV and used in this article.

Baumeister and Parker [34] pointed to the overlap between the different

subtyping models of depression. Like Lamers et al. [35] they drew attention

tot the ongoing debate on the best criteria to delineate melancholic

depression from atypical depression and other depressive conditions. There

is considerable discussion whether in atypical depression rejection sensitivity

should be included as a main criterion instead of mood reactivity.

(2.8). - P11: oversight: “…between de seasons” - Headlines of results: Your manuscript might benefit from changing the headlines of your results in such a way that they indicate the respective research question (rather than the assessment tools used). (Reaction 2.8): We agree on this comment and have subsequently changed the headlines of our results section. The numbers of the tables and figures changed correspondingly (see also reaction 1.1).

Revised version of the manuscript: page 5 and 6 More specific three questions were formulated:

(1) Does a seasonal pattern exist in the severity of depressive and anxiety

symptoms among patients visiting their general practitioner for any reason?

(2) Does a seasonal pattern exist in the severity of depressive or anxiety

symptoms among patients with a current depressive disorder, a current

anxiety disorder, a current depressive and anxiety disorder, and among

healthy controls; and is there a difference between these groups?

(3) Does a seasonal pattern exist in specific type of depressive symptoms (i.e.

atypical or melancholic) or anxiety among and between these groups?

1. NESDA primary care recruitment population Question 1. Seasonality in severity of depressive and anxiety symptoms among primary care patients (recruitment population)

2. NESDA baseline population Question 2 & 3. Seasonality in severity of depressive and anxiety symptoms and type of depressive symptoms in patients with a current depressive and / or anxiety disorder and in healthy controls

2.1. Inventory of depressive symptoms, total score 2.1. Severity of depressive symptoms

2.4 Beck Anxiety Inventory 2.2 Severity of anxiety symptoms (BAI)

15 Revision MS 6137935375847103

2.5 Fear Questionnaire 2.3 Severity of anxiety symptoms (FQ)

2.2. Inventory of depressive symptoms, atypical symptoms 3.1. Atypical depressive symptoms

2.3 Inventory of depressive symptoms, melancholic symptoms 3.2 Melancholic depressive symptoms

Fig 2 � Fig 2 Table 2 � Table 2 Fig 3 � Fig 5 Table 3 � Table 5 Fig 4 � Fig 6 Table 4 � Table 6 Fig 5 � Fig 3 Table 5 � Table 3 Fig 6 � Fig 4 Table 6 � Table 4

We also added the number of the research question between italics in the discussion section.

Revised version of the manuscript: page 19 The finding of no seasonal pattern in the severity of depressive and anxiety

symptoms in a primary care population (question 1) contrasts to the findings

of Mersch et al. [14] who did find seasonality in depressive symptoms as

measured with the CES-D scores, and with Oyane et al. [15] who found

modest seasonal variations in the depression scores on the Hospital Anxiety

and Depression Scale (HADS), both in samples of the general population.

Revised version of the manuscript: page 20 The finding of minimal seasonal differences in severity of depressive or

anxiety symptoms (question 2) in various clinical groups is in agreement with

the study of Posternak and Zimmerman [9], who did not find higher rates of

depressive symptoms in winter in an out-patient population.

Revised version of the manuscript: page 21 The third finding was that both atypical and melancholic symptoms were

slightly more present in winter (question 3).

(2.9 ). - What does your study add to the question of whether or not SAD can be viewed as a distinct subtype of depression (as delineated from the view of SAD as a pronounced seasonality trait of patients with fluctuating minor and major depression; compare Baumeister H, Parker G. 2011. Meta-review of depressive subtyping models. Journal of Affective Disorders, doi:10.1016/j.jad.2011.07.015 for references on this topic).

16 Revision MS 6137935375847103

(Reaction 2.9): This is an interesting point which was mentioned by reviewer 1 as well (1.5). Our study did not include participants with a minor depression or dysthymic disorder. The consideration was to enhance the contrast between healthy controls (HC) and participants in de the depression group (MDD). In their meta-review of depressive subtyping models Baumeister and Parker raised the question whether SAD is a distinct depressive subtype or rather a pronounced seasonality trait of patients with fluctuating minor and major depressions. As our study only used cross-sectional baseline data from NESDA and did not concern participants diagnosed with a minor depression or SAD , this question could not be addressed. However, it remains an interesting hypothesis that can be tested in future analyses of the longitudinal NESDA data.

Reviewer 3 (3.1). � Page 19: could it be that there is a fourth alternative explanation for not finding a seasonal component in the GP group: You select for people that do have a problem already. This may be a selection bias for trouble, independent of season. Please reply, or add to the discussion. (Reaction 3.1): We agree with the reviewer that patients who visit their General Practitioner for any reason represent a selection of the general population. As such they are not comparable to the general population. We did not conduct a literature search concerning the psychopathological differences between the general population and patients who visit their General Practitioner for any reason. One of the reviewers pointed tot this (see 2.4 and reaction 2.4.) We changed one of the sentences in the discussion regarding selection bias:

Revised version of the manuscript: page 19 These opposing findings might be explained by differences in the populations

under scrutiny: general population versus primary care population which

could reflect a source of selection bias independent of season. (3.2). � Page 20; Could it be that the former alternative explanation holds for the patient groups as well? You selected for MDD patients who suffered over the past 4 weeks. That is a bias towards depressive mood, independent of season, right? The groups may differ between winter and summer, both showing more or less severe depressed mood. These points may be a disadvantage of the data collection method. Not that the data are not sound or not important, but the conclusions that can be drawn need some extra consideration in my point of view. We agree with the reviewer on this so rephrased one of the paragraphs in the discussion.

17 Revision MS 6137935375847103

Manuscript: page 19 A second explanation may be found in the difference in data collection. While

in studies by Mersch et al., and Oyane et al. seasonality was assessed

retrospectively using questionnaires that assessed fluctuations of symptoms

over the year, in this study as in the study by Blacker et al., seasonality was

assessed using data on presence and severity of symptoms collected in a cross-

sectional way over the year. Like Nayar and Cochrane, we think that due to

recall bias the retrospective data collection leads to an overestimation of the

seasonal fluctuation in severity and prevalence of affective symptoms [24].

Revised version of the manuscript: page 19 and 20 A second explanation may be found in the difference in data collection. In the

studies by Mersch et al. [14], and Oyane et al. [15] seasonality was assessed

retrospectively using questionnaires that assessed fluctuations of symptoms

over the year. Like Nayar and Cochrane [25] we think that due to recall bias

the retrospective data collection might lead to an overestimation of the

seasonal fluctuation in severity and prevalence of affective symptoms. In this

study, as in the study by Blacker et al. [13], seasonality was assessed using

data on presence and severity of symptoms collected in a cross-sectional way

over the year. The cross sectional method of sampling in our study might

have introduced a source of selection bias with different groups of patients

in different seasons (Reaction 3.2): Minor Essential Revisions: (3.3). � Page 3, 1st paragraph; “atypical and melancholic…but did not show a differential pattern…”. The meaning of the last part of this sentence is unclear to me. (Reaction 3.3): We agree with the reviewer that the last part of the sentence is not very clear, so we removed this part.

Revised version of the manuscript: page 3 Atypical and melancholic symptoms were both elevated in winter but did not

show a differential seasonal pattern. (3.4) Page 11: Please give data (values and variance) and significance if you mention differences between season (data given in fig 1, but no statistics), men and women, age groups and research site. OR remove this information from this paragraph, since the differences between sexes and location and the statistics are given in the next paragraph (but not for age). (Reaction 3.4): We added the data to the text.

18 Revision MS 6137935375847103

Revised version of the manuscript: page 11 In table 1 the results of the multilevel regression analysis are presented for the

log transformed K-10 scores. The second model with the seasons as a

predictor (with spring as a reference), explains only little more variability

than the empty model as can be seen in the difference of the deviance (empty

model 5086.6; model with seasons 5085.6). In this second model the

difference between de seasons was not significant (summer -0.014, SE 0.019;

autumn -0.002, SE 0.021; winter -0.013, SE 0.021). Adding the covariates gender, age and site the final model showed that these

variables contribute significantly to the explanation of the model (gender

0.065, SE 0.011; age 0.002, SE 0.000; site -0.127, SE 0.019) but there was

no significant difference between the seasons (summer -0.015, SE 0.018;

autumn -0.022, SE 0.019; winter -0.002, SE 0.019). (summer -0.015, SE

0.018; autumn -0.022, SE 0.019; winter -0.002, SE 0.019). No significant

interactions were found between the seasons and these covariates, nor

between the covariates themselves.

In fig 1, which represents the data from the multilevel analysis using MLwin, the significance can be read from the ratio between the Beta and its Standard Error. (3.5).� Table 1: I presume that an asterisk is missing in the full model for the co-variable age. (Reaction 3.5): This is correct. We have added the asterix for the co-variable age in table 1. (3.6).� Table 3B: I do not understand the asterisks behind winter both in the “seasons and covariates” model and in the full model. P = 0.48 and 0.46 respectively. Is this a typing error, or my misunderstanding? In the text on page 14 you state that in the “seasons and covariates” model season is not significant, but an asterisks is shown in table 3b; in the full model you state winter is significantly different from summer. Please explain or change the text. (Reaction 3.6): This is a typing error. In both numbers a zero is missing. So 0.48 shoud be 0.048 and 0.46 should be 0.046. We have corrected this in the table and replaced 0.048 and 0.046 by < 0.05*. (3.7).� Table 4 and text p 15: in the text the authors say that women scored significantly higher than men, but p= 0.09 and no asterisk is shown in table 4. Please synchronize text and table. (Reaction 3.7): We agree with the reviewer that in the models without interaction there is no significant gender effect. We changed the text accordingly.

19 Revision MS 6137935375847103

Revised version of the manuscript: page 17 Adding the predictors gender and age revealed that there was no significant

effect for gender or age. women scored significantly higher than men but

there was no significant age effect. (3.8). � Page 17 and figure 6; Fear questionnaire. In the text the authors says that they present FQ by season for men and women and for the four groups, but in figure 6 only four groups are given. (Reaction 3.8): We agree with the reviewer. We meant that the figure comprises men and women together. We changed the text by deleting “for men and women”.

Revised version of the manuscript: page 15 In figure 6, the observed means and standard errors of the FQ are presented

by season for men and women and for the four groups.

(3.9 ).� Page 18: “A distinguishing seasonal pattern in specific (…) depressive…”; is some text missing after “depressive”? (Reaction 3.9): The word “symptoms” is missing. We added this tot the text. We also changed the word “specific” in “type of”.

Revised version of the manuscript: page 19 A distinguishing seasonal pattern in type of specific (i.e. atypical or

melancholic) depressive symptoms could not be demonstrated.

(3.10) Page 20; I do not understand the sentence: “the third finding…more present in winter and no clear seasonal pattern regarding these specific depressive symptoms could be demonstrated”. If you find “slightly more” in winter, you already have a seasonal pattern, right? Please reply, or rephrase in the text. (Reaction 3.10): We rephrased the sentence.

Revised version of the manuscript: page 21 The third finding was that both atypical and melancholic symptoms were

slightly more present in winter (question 3). and no clear seasonal pattern

regarding these specific depressive symptoms could be demonstrated.

Discretionary Revisions: (3.11).� Figures 1-6, I would advice to change the graphs to bar plots instead of singe dots, grouped per diagnostic group.

20 Revision MS 6137935375847103

(Reaction 3.11): We have tried this in an earlier version of the manuscript. Using bar plots per diagnostic group meant that we would end up with many more figures, over all resulting in a less comprehensive picture. (3.12).� Page 18 and more in the discussion: I would prefer the correct number of the references to be given immediately after the reference to the authors instead of at the end of the sentence. Otherwise it is unclear which reference belongs to which person. This also prevents double reference numbers as on page 19 (2 times 12 in line 9). Reaction (3.12): We agree with the preference of the reviewer and have changed the position of the reference.