University of Groningen
Evidence for a differential role of HPA-axis function, inflammation and metabolic syndrome in
melancholic versus atypical depression
Lamers, F.; Vogelzangs, N.; Merikangas, K. R.; de Jonge, P.; Beekman, A. T. F.; Penninx, B.
W. J. H.
Published in:
Molecular Psychiatry
DOI:
10.1038/mp.2012.144
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2013
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Lamers, F., Vogelzangs, N., Merikangas, K. R., de Jonge, P., Beekman, A. T. F., & Penninx, B. W. J. H.
(2013). Evidence for a differential role of HPA-axis function, inflammation and metabolic syndrome in
melancholic versus atypical depression.
Molecular Psychiatry
,
18
(6), 692-699.
https://doi.org/10.1038/mp.2012.144
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.
More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-
amendment.
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
ORIGINAL ARTICLE
Evidence for a differential role of HPA-axis function,
inflammation and metabolic syndrome in melancholic versus
atypical depression
F Lamers
1
, N Vogelzangs
2
, KR Merikangas
1
, P de Jonge
3
, ATF Beekman
2
and BWJH Penninx
2,3,4
The hypothalamic–pituitary–adrenal (HPA) axis and the inflammatory response system have been suggested as pathophysiological
mechanisms implicated in the etiology of major depressive disorder (MDD). Although meta-analyses do confirm associations
between depression and these biological systems, effect sizes vary greatly among individual studies. A potentially important factor
explaining variability is heterogeneity of MDD. Aim of this study was to evaluate the association between depressive subtypes
(based on latent class analysis) and biological measures. Data from 776 persons from the Netherlands Study of Depression and
Anxiety, including 111 chronic depressed persons with melancholic depression, 122 with atypical depression and 543 controls were
analyzed. Inflammatory markers (C-reactive protein, interleukin-6, tumor necrosis factor-a), metabolic syndrome components, body
mass index (BMI), saliva cortisol awakening curves (area under the curve with respect to the ground (AUCg) and with respect to the
increase (AUCi)), and diurnal cortisol slope were compared among groups. Persons with melancholic depression had a higher AUCg
and higher diurnal slope compared with persons with atypical depression and with controls. Persons with atypical depression had
significantly higher levels of inflammatory markers, BMI, waist circumference and triglycerides, and lower high-density lipid
cholesterol than persons with melancholic depression and controls. This study confirms that chronic forms of the two major
subtypes of depression are associated with different biological correlates with inflammatory and metabolic dysregulation in atypical
depression and HPA-axis hyperactivity in melancholic depression. The data provide further evidence that chronic forms of
depressive subtypes differ not only in their symptom presentation, but also in their biological correlates. These findings have
important implications for future research on pathophysiological pathways of depression and treatment.
Molecular Psychiatry (2013) 18, 692–699; doi:10.1038/mp.2012.144; published online 23 October 2012
Keywords: depression; inflammation; metabolic syndrome; salivary cortisol; subtypes
INTRODUCTION
Three often-studied pathophysiological systems that have a role in
the etiology of major depressive disorder (MDD) are the
hypothalamic–pituitary–adrenal (HPA) axis, the inflammatory
response system and metabolic abnormalities. HPA-axis hyper-
activity has been demonstrated in depressed persons compared
with controls, and has been further implicated as a potential
mechanism through which depression increases the risk of
cardiovascular disease and other somatic diseases.
1
Alterations
in the immune response system have been reported as well, with
depressed persons having higher serum levels of pro-
inflammatory cytokines such as C-reactive protein (CRP),
interleukin (IL)-6 and tumor necrosis factor (TNF)-a compared
with healthy individuals.
2–4
Related to this, more metabolic
abnormalities such as obesity and adverse lipoprotein patterns
are also associated with MDD,
5–7
and several studies have shown
an association between depression and metabolic syndrome.
8,9
Although meta-analyses have confirmed significant associations
between depression and HPA-axis measures (cortisol, adrenocor-
ticotropic hormone) and inflammation (CRP, IL-6, TNF-a), there is
substantial variability in the effect sizes across studies.
1,2,10
Such
variability could be attributable to sampling (for example, clinical
sample versus community), composition of the sample (for
example, age and ethnic composition) or to methodological
differences in measures of depression and biological correlates.
For example, previous studies have indicated that chronic
depression, specifically dysthymia, may differ from non-chronic
depression in levels of inflammatory markers.
11
However,
variability could also be associated with the heterogeneity of
the MDD diagnosis, particularly to differences in biological
systems among those with different depressive subtypes. We
hypothesize that the heterogeneity of MDD significantly
contributes to this variability.
Some evidence in support of this hypothesis suggests that
depressive subtypes contribute to variability in associations with
biological measures. Subtypes represent more homogeneous
groups of cases, and may potentially have different underlying
pathophysiological processes. For instance, the association
between melancholic depression and the HPA-axis hyperactivity
has been replicated in some studies, whereas persons with
1
Genetic Epidemiology Research Branch, Intramural Research Program, National Institute of Mental Health, National Institutes of Health, Bethesda, MD, USA;
2
Department of
Psychiatry/EMGO þ Institute for Health and Care Research, VU University Medical Center, Amsterdam, The Netherlands;
3
Interdisciplinary Center for Psychiatric Epidemiology
(ICPE), Department of Psychiatry and Department of Internal Medicine, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands and
4
Department of Psychiatry, Leiden University Medical Center, Leiden, The Netherlands. Correspondence: Dr F Lamers, Genetic Epidemiology Research Branch, Intramural
Research Program, National Institute of Mental Health, National Institutes of Health, 35 Convent Drive, Room 1A108, Bethesda, MD 20892-3720, USA.
E-mail: lamersf@mail.nih.gov
Received 10 February 2012; revised 11 July 2012; accepted 20 August 2012; published online 23 October 2012
Molecular Psychiatry (2013) 18, 692–699
&
2013 Macmillan Publishers Limited All rights reserved 1359-4184/13
www.nature.com/mp
atypical depression have been found to have a lower cortisol
values than persons without atypical depression.
1,12,13
Atypical
features were recently linked to decreased IL-4 and increased IL-2
compared with persons without atypical features in one study,
14
while another study reported decreased IL-2 in atypical depression
compared with controls.
11
Another study found no differences in
CRP, IL-6 and TNF-a between melancholic and atypical depression,
and a higher CRP in atypical depression compared with controls in
multivariable analyses.
15
Findings on inflammatory markers
among those with melancholic depression have been
contradictory; whereas one study reported higher IL-1b among
melancholics,
16
others found lower IL-1b compared with non-
melancholics.
17
Based on meta-analytic work, Howren et al.
3
concluded that body mass index (BMI) may interact with CRP and
IL-6 to yield a potential tridirectional relationship between
adiposity, inflammation and depression. The high BMI levels of
those with atypical depression
18,19
may indicate a differential
association between atypical depression with inflammation
compared with melancholic depression, as was also postulated
by Gold and Chrousos.
12
In the past, substantial research has been devoted to the
identification of subtypes of depression
20
based on clinical
correlates, treatment response or observed symptom profiles.
Opposed to following these paths, a more novel way of identifying
subtypes has become available in the form of data-driven
techniques, which result in more empirically based subtypes.
Using such techniques, we previously observed subtypes closely,
but not exactly, matching the DSM-IV (Diagnostic and Statistical
Manual of Mental Disorders, 4th Edition) atypical and melancholic
depression.
18,21
In line with a previous finding that metabolic
syndrome was associated with neurovegetative symptoms of
depression,
22
we observed differences between atypical and
melancholic depression, including higher rates of metabolic
syndrome.
18
If more homogeneous subtypes have different
biological profiles, this may help bring us closer to underlying
etiologies, and can help direct future etiological research and
treatment development. Although several researchers have
addressed differences in biological measures across depressive
subtypes, to our knowledge, no previous studies have rigorously
evaluated different biological systems among depressive subtypes
and controls simultaneously.
Our previous empirical work on the Netherlands Study of
Depression and Anxiety (NESDA) sample distinguished two severe
depressive subtypes that differed in depressive symptom profile—
one resembling melancholic and one resembling atypical depres-
sion.
18,21
The aim of the current study was to compare different
pathophysiological indicators of HPA-axis function, the
inflammatory response system and metabolic syndrome across
these two subtypes of MDD and healthy controls. We expect that
melancholic depression will show more pronounced HPA-axis
hyperactivity, whereas atypical depression will, in this respect,
not differ from controls. Further, we expect that atypical
depression will show a profile of immune activation and
metabolic abnormalities compared with melancholic depression,
whereas melancholic depression and controls will not differ from
each other in this aspect.
MATERIALS AND METHODS
Sample
Data from NESDA were used.
23
NESDA is a longitudinal naturalistic cohort
study, consisting of 2981 persons (18–65 yrs), including those with lifetime
or current anxiety and/or depressive disorders (n ¼ 2329; 78%) and healthy
controls (n ¼ 652; 22%).
23
Participants were recruited from the community
(n ¼ 564; 19%), primary care (n ¼ 1610; 54%) and specialized mental health
care (n ¼ 807; 27%) from September 2004 through February 2007 at three
study sites (Amsterdam, Groningen and Leiden). Approval of the study
protocol was granted by the Ethical Review Boards of all the participating
centers and all the participants gave written informed consent. Exclusion
criteria were (1) a primary clinical diagnosis of psychotic disorder,
obsessive compulsive disorder, bipolar disorder or severe addiction
disorder and (2) not being fluent in Dutch. At baseline, participants were
invited for a 4-h interview at one of the three study sites. Information was
collected on psychopathology, demographic characteristics, physical and
psychosocial functioning. It further included a blood draw, medical
assessment, computer tasks, two self-administered questionnaires and
salivary cortisol assessment. For the current study, we included 776
persons, including 233 persons with a current diagnosis of MDD whose
subtype has been established (see below for details), and 543 controls
without lifetime depressive or anxiety disorders.
Depressive subtypes
MDD was diagnosed with the Composite International Diagnostic Inter-
view (CIDI) lifetime version 2.1,
24
which was conducted by trained clinical
research staff. For the current study, we used two depressive subtypes, a
Severe Melancholic (n ¼ 111) and a Severe Atypical (n ¼ 122) subtype,
which were previously identified based on a latent transition analysis (LTA)
of patients with persistent chronic depression. Chronic depression was
defined as having a current (12-month) diagnosis of MDD at both the
baseline and at the 2-year follow-up measurement.
A detailed description of the LTA analyses and subtypes can be found
elsewhere.
21
In short, 10 depressive symptoms from the CIDI were used as
indicator variables to identify depression subtypes at each measurement.
In the first step of these data-driven analyses, persons with similar
symptom-endorsement patterns were clustered in classes. In a second
step, after identifying the best-fitting models for each time point,
transitions over time were modeled in a longitudinal analysis. The
analyses revealed three subtypes of depression that were labeled by the
researchers based on their observed symptom patterns: a Severe
Melancholic/Typical subtype (average prevalence across time points:
27.2%) characterized by decreased appetite and weight loss, a Severe
Atypical subtype (32.2%) characterized by overeating and weight gain, and
a Moderate subtype (40.5%) that was characterized by lower symptom
probabilities and overall lower severity. Transition analyses showed that
76% of the sample endorsed the same subtype at both measurements,
indicating that this group had a relatively stable depressive subtype. It
should be noted that our labels for subtypes do not refer to the DSM-
classifiers. However, inclusion of the symptoms of mood reactivity and
interpersonal sensitivity in the current definition atypical depression has
been debated,
25–28
and this was underlined in our latent class analysis
(LCA) analysis showing that these two symptoms did not discriminate
subtypes.
18
Also, other LCA studies have found similar symptom patterns
with appetite and weight being the most discriminating symptoms,
showing robustness of the identified subtypes.
19,29
Because our labels
were used to describe the classes in previous work, we use the same labels
here as well for consistency, but it is important to point out that our LCA-
based subtypes of melancholic and atypical depression are, thus, not
literally resembling DSM-classifications but are based on LCA-driven
analyses. More information about the LTA method, identified subtypes
and their correlates are provided in the Supplementary Materials.
In the current analyses, we included only persons with chronic
depression and a stable depressive subtype over time for the following
reasons. Firstly, stable subtypes are likely more homogeneous. Within
chronic forms of depression some heterogeneity may still exist with
different episodes having a different presentation or etiology,
30
but this
can be avoided by including only persons with a stable clinical
presentation. Secondly, by including persons with a chronic depression
(that is, diagnoses present at two different measurements) we capture the
more disabling, severe forms of depression. There is increasing awareness
that single episodes may be distinct from recurrent and chronic forms,
with the latter type being increasingly viewed upon as a chronic recurring
condition characterized by high disability.
30,31
These chronic forms
furthermore may have higher heritability than single episodes of
depression.
32
Finally, having more homogeneous subtypes may bring us
closer to the underlying etiologies of subtypes.
Severity may also be an important correlate for many biological markers
that makes direct comparison of subtypes of different severity somewhat
difficult. Therefore, we only included the Severe Melancholic and the
Severe Atypical subtypes as they were of similar severity. Furthermore,
melancholic and atypical depression are well recognized in the literature
and are also currently included in DSM-IV. The controls used in the study
were persons without lifetime depressive or anxiety disorders at both the
baseline and 2-year assessment.
Differential pathophysiology in depressive subtypes
F Lamers et al
693
& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 692 – 699
Biological measures
Metabolic syndrome and BMI. Serum levels of triglycerides, high-density
lipid (HDL) cholesterol and fasting plasma glucose were determined in
plasma samples, for which blood samples were collected after an overnight
fast. Blood samples were drawn into vacuum tubes between 0730–0930 hrs
and kept frozen at 80 1C. Blood pressure, height, weight and waist
circumference were measured during the medical assessment. Blood
pressure was measured twice during supine rest on the right arm with
the Omron M4-I, HEM 752A (Omron Healthcare Europe B.V., Hoofddorp, The
Netherlands) and was averaged over the two measurements. In the analyses,
positive outliers (mean þ / 3 s.d.) were trimmed to the mean þ / 3s.d.
(waist circumference n ¼ 3, triglycerides n ¼ 5, HDL n ¼ 5, systolic blood
pressure n ¼ 5, diastolic blood pressure n ¼ 2, glucose n ¼ 5). Metabolic
syndrome was defined according to the Adult Treatment Panel III.
33,34
The
criteria are as follows: (1) waist circumference 4102 cm in men or 488 cm
in women, (2) triglycerides X1.7 mmol l
1
(150 mg dl
1
), (3) HDL
cholesterol o1.03 mmol l
1
(40 mg dl
1
) in men or o1.30 mmol l
1
(50 mg dl
1
) in women, (4) blood pressure X130/85 mm Hg or use of
antihypertensives and (5) fasting glucoseX 6.1 mmol l
1
(110 mg dl
1
)or
drug treatment for elevated glucose. Further, a count of the total number
of positive criteria was constructed. BMI was calculated (kg m
2
).
Inflammatory markers. CRP and IL-6 were assayed at the Clinical
Chemistry department of the VU University Medical Center. High-sensitivity
plasma levels of CRP were measured in duplicate by an in-house ELISA
based on purified protein and polyclonal anti-CRP antibodies (Dako,
Glostrup, Denmark). The lower detection limit of CRP is 0.1 mg l
1
and the
sensitivity is 0.05 mg l
1
. Intra- and inter-assay coefficients of variation
were 5% and 10%, respectively. Plasma IL-6 levels were measured in
duplicate by a high-sensitivity enzyme-linked immunosorbent assay
(PeliKine CompactTM ELISA, Sanquin, Amsterdam, the Netherlands).
Comparison of IL-6 levels according to this assay with that of the IL-
6 R&D array (R&D systems Minneapolis, MN, USA) in 77 random NESDA
participants showed a correlation of 0.88, confirming comparability of
these methods. The lower detection limit of IL-6 is 0.35 pg ml
1
and the
sensitivity is 0.10 pg ml
1
. Intra- and inter-assay coefficients of variation
were 8% and 12%, respectively. Plasma TNF-a levels were assayed in
duplicate at Good Biomarker Science, Leiden, the Netherlands, using a
high-sensitivity solid phase ELISA (Quantikine HS Human TNF-a Immu-
noassay, R&D systems, Minneapolis, MN, USA). The lower detection limit of
TNF-a is 0.10 pg ml
1
and the sensitivity is 0.11 pg ml
1
. Intra- and inter-
assay coefficients of variation were 10% and 15%, respectively.
Cortisol. At the baseline interview, participants were instructed to collect
saliva samples at home on a regular day (preferably, a working day). This
method has been shown to be a reliable and minimally intrusive method
to assess the active, unbound form of cortisol.
35
Samples for the cortisol
awakening response were obtained using Salivettes (Sarstedt AG and Co,
Nu¨ mbrecht, Germany) at awakening (T1), and 30 (T2), 45 (T3) and 60 (T4)
minutes later. Additionally, participants collected a sample at 1100 hrs.
Samples were stored in refrigerators and returned by mail. After receipt,
Salivettes were centrifuged at 2000 g for 10 min, aliquoted and stored at
80 1C. Cortisol analysis was performed by competitive electro-
chemiluminescence immunoassay (E170; Roche, Basel, Switzerland) as
described by van Aken et al.
36
The functional detection limit was
0.07 mgdl
1
and the intra-assay and inter-assay variability coefficients in
the measuring range were o10%. Values collected outside a margin of
5 min before or after the time protocol were recoded as missing. Persons
using corticosteroids or who were pregnant or breastfeeding were
excluded from analyses (n ¼ 47). Values that were larger than the
mean þ 2 s.d. were coded as missing (n ¼ 24).
For the cortisol awakening response, the area under the curve with
respect to the increase (AUCi) and with respect to the ground (AUCg) were
calculated using trapezoid formulas.
37
To calculate AUCi and AUCg,
samples of at least three time points had to be available. For those with
one missing sample (n ¼ 26), the missing value was imputed using a linear
regression model, including information on the available three cortisol
levels, age, sex, awakening time and smoking status.
38
To assess the
diurnal cortisol slope, we calculated the slope mean decline per hour as:
diurnal slope ¼ (T
awakening
–T
evening
)/(time T
evening
—time T
awakening
).
39
Covariates and descriptive variables
Potential confounding variables that were considered included age, sex,
educational level and smoking (yes/no). Models for cortisol AUCg and AUCi
were additionally corrected for awakening time on the day of saliva
collection.
To describe groups, we used several clinical characteristics, including
severity of depressive symptoms as measured with the Inventory of
Depressive Symptoms 30-item self-report,
40
age of onset assessed in the
CIDI interview, duration of symptoms in the 4 years before baseline as
measured with the Lifechart method
41
and family history of depression in
first-degree relatives (not including offspring) based on self-report.
42
Antidepressant use was based on drug-container inspection of all the
drugs used in the past month (on at least 50% of days), classified according
to the World Health Organization Anatomical Therapeutic Chemical
classification, and included selective serotonin reuptake inhibitors (SSRI;
ATC code N06AB), serotonin-norepinephrine reuptake inhibitors (SNRI;
N06AX16, N06AX21), tricyclic antidepressants (TCAs; N06AA) and
tetracyclic antidepressants (TeCAs; N06AX03, N06AX05, N06AX11).
Duration of antidepressant treatment was also assessed in number of
months. Also, use of anti-inflammatory medication (M01A, M01B, A07EB,
A07EC), statins (C10AA, C10B) and corticosteroids (H02, R03BA, R03AK,
D07) were assessed. Diagnoses of cardiovascular disease and diabetes
were based on self-report.
Statistical analyses
All analyses were performed in SPSS, version 19. In case of non-normal
distributions, biological measures were log-transformed when necessary.
First, means of metabolic syndrome variables, inflammatory markers and
cortisol variables were compared across groups (controls, melancholic
depression, atypical depression) using analyses of variance. In multivariable
analyses we corrected for potential confounders. To test our hypotheses,
we tested differences between atypical and melancholic depression. We
also tested differences between atypical and melancholic depression
versus controls, to evaluate whether they were different from controls (that
is, inflammation and metabolic syndrome in atypical depression, and
cortisol in melancholic depression) or similar to controls (that is,
inflammation and metabolic syndrome in melancholic depression, and
cortisol in atypical depression). In addition, we calculated effect sizes
(Cohen’s d). For cortisol, we also analyzed the four cortisol awakening
measurements in a linear mixed model, adjusting for age, sex, educational
level, smoking and awakening time. As mixed models can handle missing
data through maximum likelihood estimation, we included all persons who
had at least two cortisol measurements available.
RESULTS
In Table 1, we describe the socio-demographic and clinical
characteristics of the two LCA-based depressive subtypes and
controls. Groups did not differ in age, but the depressed groups
were more often female, had lower educational levels and were
more often smokers. In terms of clinical characteristics, persons
with melancholic depression had a slightly higher severity and
chronicity than persons with atypical depression, but otherwise
the depressed subtypes were comparable. With the exception of
TCAs and TeCAs, used by only nine and seven persons,
respectively, duration of treatment was relatively short. Duration
of TeCAs was longer in atypical depressed persons. There were no
differences in prevalence of diabetes and cardiovascular disease,
statin use and corticosteroid use across groups, but the depressed
groups used anti-inflammatory medication more often.
Unadjusted means and s.ds. for metabolic syndrome variables
and BMI, inflammatory markers and cortisol are presented in
Table 2. Of the metabolic measures, waist circumference, BMI and
the number of metabolic syndrome criteria were significantly
more common in atypical depression compared with controls and
melancholic depression, and the levels of triglycerides were
elevated in the atypical group compared with controls. Systolic
blood pressure was significantly elevated in controls compared
with both the depressive groups. CRP, IL-6 and TNF-a levels were
most elevated in atypical depression. Of the cortisol measures,
AUCg was significantly higher in melancholic depression com-
pared with controls and atypical depression, while diurnal slope
was lower in atypical depression compared with controls and
melancholic depression.
Differential pathophysiology in depressive subtypes
F Lamers et al
694
Molecular Psychiatry (2013), 692 – 699 & 2013 Macmillan Publishers Limited
Table 1. Group characteristics of depressive subtype groups and controls
N Controls Melancholic
a
Atypical
a
P-value
N ¼ 543 N ¼ 111 N ¼ 122
Demographics and health indicators
Female (%) 776 60.6 65.8 79.5 o0.0001
Age, mean (s.d.) 776 41.3 (14.6) 40.2 (12.1) 39.6 (12.1) 0.40
Educational level (years), mean (s.d.) 776 12.9 (3.2) 10.8 (3.1) 11.3 (3.3) o0.0001
Smoking (% yes) 776 24.7 59.5 32.8 o0.0001
Clinical characteristics
Severity of depression, mean (s.d.) 233 NA 39.1 (10.0) 35.2 (11.5) 0.01
Age of onset, median (IQR) 233 NA 22.0 (17.0–34.0) 22.0 (16.0–32.0) 0.32
Duration sx(% time), median (IQR) 232 NA 0.40 (0.19–0.66) 0.29 (0.15–0.58) 0.09
Positive family history depression (%) 231 NA 81.7 82.8 0.82
Medication use baseline, N (%)
a
SSRI 233 NA 39 (35.1) 39 (32.0) 0.61
SNRI 233 NA 10 (9.0) 10 (8.2) 0.83
TCA 233 NA 4 (3.6) 5 (4.1) 0.85
Tetracyclic antidepressants 233 NA 4 (3.6) 3 (2.5) 0.61
Duration use among users (months), median (IQR)
SSRI 78 NA 6 (3–35) 5 (2–35) 0.66
SNRI 20 NA 5.5 (2.5–24) 4 (2–30) 1.00
TCA 9 NA 20 (3.25–97.5) 36 (3.5–168) 0.73
Tetracyclic antidepressants 7 NA 3 (1.25–4.75) 72 (-
b
) 0.06
Physical health indicators
Medication use, N (%)
Anti-inflammatory medication 776 8 (1.5) 6 (5.4) 4 (3.3) 0.03
Statins 776 33 (6.1) 5 (4.5) 10 (8.2) 0.50
Corticosteroids 776 24 (4.4) 5 (4.5) 7 (5.7) 0.82
Chronic diseases
DM 776 22 (4.1) 5 (4.5) 8 (6.6) 0.49
CVD 766 20 (3.7) 4 (3.6) 8 (6.6) 0.63
Abbreviations: CVD, cardiovascular disease; DM, diabetes mellitus; IQR, interquartile range; NA, not applicable; SNRI, serotonin-norepinephrine reuptake
inhibitors; SSRI, selective serotonin reuptake inhibitors; sx symptoms; TCA, tricyclic antidepressants.
a
Subtypes based on latent class and latent transition analyses.
b
IQR not available due to low numbers.
Table 2. Means (s.d.) of biological measures across stable depressive subtypes and controls (N ¼ 776)
N Controls Melancholic
a
Atypical
a
P-value
N ¼ 543 N ¼ 111 N ¼ 122
Metabolic syndrome and BMI
Waist circumference (cm) 774 88.1 (13.5) 86.2 (13.1) 94.0 (15.7) o0.0001 B, C
Triglycerides (mmol l
1
)
b
764 1.08 (1.67) 1.11 (1.68) 1.22 (1.62) 0.055 B
HDL cholesterol (mmol l
1
) 760 1.63 (0.43) 1.59 (0.42) 1.55 (0.42) 0.16
Systolic BP (mm Hg) 774 137.5 (19.9) 133.4 (17.3) 132.7 (15.9) 0.012 A, B
Diastolic BP (mm Hg) 774 80.9 (11.5) 80.8 (10.3) 81.4 (9.5) 0.91
Fasting plasma glucose (mmol l
1
) 766 5.13 (0.78) 5.13 (0.79) 5.21 (0.94) 0.62
No. of MetSyn components, mean (s.d.) 765 1.4 (1.3) 1.4 (1.2) 1.8 (1.4) 0.01 B, C
BMI, mean (s.d.) 776 25.1 (4.6) 24.2 (4.8) 28.7 (6.0) o 0.0001 B, C
Inflammatory markers
CRP (mg l
1
), mean (s.d.)
b
768 1.12 (3.23) 1.18 (3.57) 1.86 (3.48) o0.0001 B, C
IL-6 (pg ml
1
), mean (s.d.)
b
769 0.73 (2.58) 0.75 (2.64) 1.04 (2.42) 0.001 B, C
TNF-a, (pg ml
1
), mean (s.d.)
b
762 0.84 (1.90) 0.78 (1.89) 1.03 (1.97) 0.002 B, C
Cortisol
AUCg 504 18.47 (6.85) 21.82 (8.34) 17.16 (6.13) o0.0001 A, C
AUCi 504 1.62 (6.61) 3.27 (8.74) 2.90 (5.61) 0.10
Diurnal cortisol slope 507 0.75 (0.46) 0.86 (0.51) 0.59 (0.34) 0.001 B, C
Abbreviations: AUCg, area under the curve with respect to the ground; AUCi, area under the curve with respect to the increase; BMI, body mass index; BP,
blood pressure, CRP, C-reactive protein; HDL, high-density lipid; IL-6, interleukin-6; MetSyn, metabolic syndrome; TNF-a, tumor necrosis factor-alpha.
A ¼ Controls different from Melanchol ic. B ¼ Controls different from Atypical. C ¼ Atypical different from Melancholic.
a
Subtypes based on latent class and latent transition analyses.
b
Test on log-transformed variable, means are back-transformed.
Differential pathophysiology in depressive subtypes
F Lamers et al
695
& 2013 Macmillan Publishers Limited Molecular Psychiatry (2013), 692 – 699