- Email: [email protected]
Altered hippocampal formation shape in first-episode depressed patients at 5-year follow-up
Journal of Psychiatric Research 47 (2013) 50e55
Contents lists available at SciVerse ScienceDirect
Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires
Altered hippocampal formation shape in first-episode depressed patients at 5-year follow-upq Serhan Isıklı a, Onur Ugurlu a, b, Ece Durmusoglu a, Gozde Kizilates a, b, Omer Kitis a, c, Erol Ozan d, Cagdas Eker a, Kerry Coburn e, Ali Saffet Gonul a, e, * a
Ege University School of Medicine, Department of Psychiatry, SoCAT Lab, Bornova, Izmir, Turkey Ege University, Department of Mathematics, Bornova, Izmir, Turkey Ege University School of Medicine, Department of Radiology, Neuroradiology Unit, Bornova, Izmir, Turkey d Ataturk University School of Medicine, Department of Psychiatry, Erzurum, Turkey e Mercer University School of Medicine, Department of Psychiatry and Behavioral Sciences Macon, GA, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 March 2012 Received in revised form 17 August 2012 Accepted 21 August 2012
It is generally accepted that patients with major depressive disorder have smaller hippocampus size compared to healthy people. However, it is still not known if this situation exists before the onset of the disease or is a result of the toxic mechanism created by the depression itself. The findings of the longterm follow-up studies of first-episode depressed patients might contribute to solve the ongoing problem. In this study, the hippocampus of 18 first-episode patients who were followed-up for 5 years, were compared with those of healthy controls. There were no volumetric differences among groups neither at the baseline nor after 5 years of follow-up. However, shape analyses, using high dimensional mapping methods, revealed regional structural changes in the head and tail of the hippocampal formation in CA1 and subiculum regions in patients at the follow-up. Furthermore, a significant negative correlation was found with the number of days in depression without antidepressant treatment in the CA1 region in the head and tail of the hippocampal formation bilaterally. The findings of this study support the hypothesis that pathophysiological processes of depression induce structural alterations in depressed patients. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Major depressive disorder Magnetic resonance imaging Hippocampus Follow-up
1. Introduction The Hippocampal Formation (HF), including the hippocampus (Cornu Ammonis) proper, the dentate gyrus, and the subiculum, is a set of closely related limbic structures playing crucial roles in memory, spatial navigation, learning, contextual fear-conditioning and neuroendocrine regulation (Fanselow and Dong, 2010). Major depressive disorder (MDD) has several signs and symptoms that appear to be related to HF functions. Therefore, researchers have focused on structural and functional abnormalities of the HF in depressed patients. Accumulated evidence has shown that, in addition to functional abnormalities, the HF of depressed patients often has overall volumetric and regional shape abnormalities (MacQueen and Frodl, 2011; Videbech and Ravnkilde, 2004). Many q This study was conducted in Ege University School of Medicine, Department of Psychiatry, SoCAT Lab, Bornova, Izmir, Turkey. * Corresponding author. Ege University School of Medicine, Department of Psychiatry, SoCAT Lab, Bornova, Izmir, Turkey. Tel.: þ90 2323904163. E-mail address: [email protected] (A.S. Gonul). 0022-3956/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jpsychires.2012.08.022
volumetric studies, though not all, have reported smaller HF volumes compared to healthy controls (Eker and Gonul, 2010; Videbech and Ravnkilde, 2004). Some of these studies also found evidence that a longer cumulative duration of lifetime depression, a larger total number of recurrent episodes, and an earlier age of depression onset are associated with smaller HF volume (Cheng et al., 2010; MacQueen et al., 2003; McKinnon et al., 2009; Sheline et al., 2003). Thus, these findings support the view that the reduction of the HF is a cumulative process beginning around the time that symptoms first manifest and continuing over many years of depression. Physiologically, HF atrophy may be mediated by a chronic stress-induced HypothalamicePituitaryeAdrenal (HPA) axis response, which increases cortisol and reduces Brain-derived Neurotrophic Factor (BDNF). These neurohormonal changes in turn reduce neurogenesis and alter the somatodendritic, axonal, and synaptic components of neurons, as well as causing glial changes. Together these changes produce abnormal neuronal functioning (Czeh and Lucassen, 2007). However, there are also studies showing reduced HF volume only in patients carrying susceptibility genes or patients who had childhood trauma (Eker et al., 2011;
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
Frodl et al., 2010; Lupien et al., 2009; Vythilingam et al., 2002). Such findings suggest that HF atrophy is due to genetic and environmental factors during the early years of life (Lupien et al., 2009; McEwen, 2007), and that HF volume reduction might be a vulnerability factor for depression occurring before the onset of the clinical symptoms. This idea is supported by a recent meta-analysis which included only first episode depressed patients (Cole et al., 2011). One approach to investigating these two possibilities is to study HF volume change as a function of time (MacQueen and Frodl, 2011). HF atrophy at baseline in first-episode MDD patients, without further reductions over a 5 year period would constitute evidence of an atrophic process occurring prior to MDD, perhaps dating from childhood. Conversely, normal HF volume at baseline followed by HF atrophy at 5 year follow-up would be consistent with MDD as a causal mechanism. To our knowledge, there are five published studies examining HF atrophy in MDD patients over time periods of 6 months to 11 years (Frodl et al., 2008a, 2008b; Hou et al., 2012; Hviid et al., 2010; Vythilingam et al., 2004). Among those studies, the only one which used voxel-based analyses (Frodl et al., 2008b), suggested partial volume loss in the hippocampus of depressed patients in a three-year follow-up. The other four, which used tracing methods (the gold standard of volumetric analysis), did not find an on-going volume reduction in patients at follow-up. However, reduced HF volume was associated with poor clinical outcome. A limitation of these studies is that overall HF volume may not be as sensitive as regional HF volume (i.e., HF shape) to the changes associated with MDD. Therefore, we examined changes in both HF volume and shape over a 5 year period in first-episode MDD patients. 2. Methods The recruited patients were among the 68 first episode righthanded drug-free depressed patients who participated in earlier published studies (Eker et al., 2010; Gonul et al., 2011). Forty patients were still in treatment at the end of 5 years, of whom 23 agreed to a second MRI scan. However, 5 of the patients could not be scanned again. Two were very busy and failed to follow MRI schedule, one was pregnant, one had hydrocortisone treatment and one’s diagnoses was changed to bipolar disorder. All patients had been diagnosed in the earlier study with MDD (without comorbidity for any other axis I disorder) according to Diagnostic and Manual of Mental Disorders, fourth edition (DSM-IV) criteria and confirmed with the Structured Clinical Interview for DSM-IV (SCID) (First et al., 1996) by a consensus of at least 2 psychiatrists. All patients were re-evaluated with the SCID before their second scan, and those 18 whose MDD diagnosis was confirmed were included in this study. The severity of their depression was evaluated with Hamilton Depression Scale (17 items) (Hamilton, 1960). The control group consisted of 18 right-handed healthy subjects who also participated in the previous studies. During control group recruitment, an attempt was made to match age, gender, education, and time gap between the two MRI scans. However, age did not match because of recruitment problems in the control group. Control subjects were re-screened with non-patient version of SCID to exclude any Axis I disorders (First et al., 2002). A structured interview assessed exclusion criteria (previous head injury with loss of consciousness, hydrocortisone treatment, alcohol or substance abuse, neurologic disease or chronic medical disease including hypertension or diabetes mellitus, first-degree relative with bipolar or psychotic disorder) for all participants. Prior to this follow-up one psychiatrist from the research team (S. I.) screened all the hospital records of the patients and the controls and interviewed all subjects to investigate the intervening period
51
for any major social or personal trauma or conflict. The primary aim of this interview was to obtain any information that might have been missed with the structured diagnostic interview. A secondary aim was to obtain information related to any personality disorders among the subjects. However, no patients or controls were excluded for this reason. Handedness was determined using the Edinburg Inventory (Oldfield, 1971). The study design was approved by the local ethics committee and was prepared according to the ethical standards of the Declaration of Helsinki. The written informed consent was obtained from all subjects after a description of the study. 2.1. MRI image acquisition The imaging was performed on a 1.5-T MR unit (Magnetom Vision Siemens, Erlangen, Germany) with a circularly polarized head coil. The standard MRI scan included multiplanar turbo spinecho T1-weighted (T1-W) (TR/TE: 650/14 ms) and T2-weighted (T2-W) (TR/TE: 3800/90 ms) images. In addition to conventional sequences, a 3D MP-RAGE sequence (TR/TE: 9.70/4.00 ms, slice thickness: 2 mm, interslice gap: 0 mm, voxel size: 0.5 mm 0.5 mm 2 mm) in the coronal plane was obtained for volumetric analysis. All imaging data were transferred to a PC workstation for further analyses. 2.2. Hippocampal formation measurement and shape analyses We used ITK-SNAP software (www.itksnap.org) to outline the HF borders. This software provides an interactive labeling function with simultaneously coronal, sagittal and axial views and concurrently provides surface rendering of volume of interest. The definition of the HF borders was described in detail previously (Gonul et al., 2011). An experienced neuroradiologist (O.K.) and a trained psychiatrist (S.I.), who were blind to the identities of the images, manually outlined the borders. To determine the interrater reliability, the HF ROI was independently traced on 10 randomly selected brains by the two raters. There was a high intraclass correlation for both the interrater reliability (rICC ¼ 0.96) and the intrarater reliability (rICC ¼ 0.96). The 3D HF shape analysis is based on the use of spherical harmonic basis functions (SPHARM) to fit the HF boundary. This process was based on a uniform-icosahedron subdivision of the spherical parameterization: A detailed description about the algorithms can be found in previous articles (Styner et al., 2003, 2006). The binary images containing the surface points determined by boundary identification were pre-processed to fill small holes and minimally smoothed. At this point, the vertexes of the voxels described the surface. A triangulation mesh was generated from each image by dividing each exterior voxel face into two triangles. A spherical parameterization was then computed for the triangulation mesh using a distortion minimizing spherical mapping. Finally, the SPHARM coefficients were computed from the mesh and its parameterization using straightforward fitting to a spherical harmonic function. The correspondences across all surfaces were computed by aligning the first-order ellipsoids of all the parameterizations. All the triangulated meshes of the HF were spatially aligned to the template, which was obtained from SPHARM by using a rigid transform. 2.3. Statistical analysis Clinical and demographic variables were compared by t test or chi square. Repeated measures of analyses of covariance (ANCOVA) were used to compare HF volumes among the groups. Side and time were within the group factors and age, gender, and total brain
52
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
volume were covariates. We used the similar ANCOVA design without between-group factors to assess the effects on HF volume of duration of illness, duration of antidepressant treatment, and number of depressive episodes between the scans. An alpha level of p < 0.05 was used for all tests. For the shape analyses, we used 2 2 MANCOVA model with age, gender, and total brain volume as confounding factors. The analyses generated raw (unadjusted for multiple comparisons) significance maps showing raw p-values exceeding 0.05 across the surface. For post-hoc within- and between-group analyses the alpha level was adjusted to 0.01 to decrease type I errors. For visualization (Figs. 1 and 2 left, 3), surface locations are in blue if the differences are not statistically significant. In addition to significance maps, we present volume difference in millimeters among the groups in distance maps (Fig. 2 right). 3. Results 3.1. Clinical data and volumetric comparison Gender and education were similar between patient and control groups but patients were older than controls (Table 1). Among patients, 12 were in depressive episodes while 6 were in remission (Mean HAM-D scores: 17.3 and 3.2; respectively). During five years intervening between the original study and this follow-up one patient was chronically depressed, one patient had four depressive episodes, five patients had three episodes, seven patients had two episodes and four patients had one episode (the numbers include the current episode if they were depressed during second scan) Eight patients were on single antidepressant treatment, one was on antidepressant and antipsychotic combination and another patient was on a low dose amisulpride treatment. Eight patients were effectively not on antidepressant treatment due to low adherence. Comparison of total HF volume of the patient and the control groups with repeated measures ANCOVA revealed neither group nor time effects (F ¼ 0.64 df ¼ 1, 30 p > 0.05, F ¼ 0.95 df ¼ 1, 30
p > 0.05; respectively). The duration of depressed days, duration of antidepressant treatment, and number of episodes between the scans similarly had no effects (F ¼ 0.8 df ¼ 1, 12 p > 0.05, F ¼ 0.1 df ¼ 1, 12 p > 0.05, F ¼ 0.4 df ¼ 1, 12 p > 0.05; respectively). HAM-D scores were not correlated with the HF volumes. There were no differences in total brain volumes among the groups (F ¼ 2.5 df ¼ 1, 31 P > 0.05) (Table 1). 3.2. Shape analyses HF shape analyses with MANCOVA (age, gender and total brain volume as confounding factors) revealed a significant group effect in the medial aspect of the HF head and the lateral aspect of the posterior HF bilaterally (Fig. 1). We also observed a significant group effect in the subiculum from mid-body to tail in the right HF. When we investigated the time effect, we found that the left HF had multiple affected areas. One of the affected areas was at the lateral side of the head, consisting mainly of CA1. Another was a large band-like area starting in dorsal aspects of CA2e3 and extending to subiculum on the ventral side. A significant group time interaction was present in these same areas. Post-hoc analysis of the group time interaction revealed that depressed patients and healthy controls had similar HF shapes at the baseline. At the five year follow-up, controls did not show any significant gray matter HF change. On the other hand, patients showed heterogeneous gray matter changes; gray matter loss at the lateral tip of the head (subiculum) of the HF extending to the dorsal surface at CA1 region (Fig. 2). This initial thinning merged into a thickening at the end of the significant area. At the junction of the body and tail of the HF we observed a band-like area of gray matter changes, with thickening at the medial and lateral aspects of the dorsal side of the band, and thinning along the ventral side. Finally, there was apparent though non-significant gray matter loss at the head and posterior of the HF bilaterally in depressed patients. MANCOVA additionally showed that the number of days in depression without antidepressant treatment between two scans
Fig. 1. Regional structural alterations observed in the hippocampal formation of the patients with major depressive disorder relative those of the healthy comparison subjects.
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
53
Fig. 2. Structural changes observed in depressed patients at the end of five-years follow-up.
had a significant effect on CA1 and subiculum at the head and postero-lateral side on the HF bilaterally (Fig. 3). When we excluded two outliners from the analyses, the results did not change. There were no significant effects of the duration of antidepressant treatment on HF shape.
Table 1 Clinical and demographic variables of patient and control groups. Variables
Patients (N ¼ 18)
Controls (N ¼ 18)
Comparison
Age
46.11 8.5
37.56 6.15
Sex (male/female)
6/12
9/9
Education
13.1 3.03
15 3.52
T ¼ 3.45 df ¼ 34 P ¼ 0.001 X2 ¼ 1.03 P > 0.05 T ¼ 1.77 df ¼ 34 P > 0.05
Age of onset HAM-D scores MADRS Number of episodes Duration of depressed days Duration of depressed days w/o antidepressant treatment Current mood state Depressed (N%) Euthymic (N%) Duration of antidepressant treatment (days) Duration between two MRI scan (days) Right hippocampus volume (baseline) (cc) Right hippocampus volume (5 years) (cc) Left hippocampus volume (baseline) (cc) Left hippocampus volume (5 years) (cc) Total brain volume (cc)
36.89 8.84 12.61 8.2 13.89 8.6 2.33 1.08 1145.3 130 638.1 909
e e e e e e
12 (66%) 6 (33%) 658.7 763.3
e
2073.8 223.3
1903.1 356
4. Discussion This study is a 5-year follow-up assessing the changes in HF volume and shape, among first-episode MDD patients and healthy controls used in two previous studies (Eker et al., 2010; Gonul et al., 2011). Overall HF volumes of patients did not differ from those of controls, either at baseline or at follow-up. However, at follow-up, regional changes in HF shape were found among the patients but not the controls. To our knowledge, this is the first follow-up study to show HF shape alterations in MDD patients using high dimensional mapping methods. Although reduced HF volume is a common finding in MDD patients, the current literature is unclear whether the reduction is the result of a faulty neurodevelopmental process before the disease onset or due to neurotoxicity associated with depression pathophysiology, or both (Sheline, 2011). Long-term studies following MDD patients from their first episode might help elucidate the pathophysiological process. Our findings of normal HF volumes in MDD patients at the time of their first depressive episode are consistent with many (though not all) first episode studies, supporting the idea that reductions in
T ¼ 1.68 df ¼ 34 P > 0.05 3.416 0.36 3.485 0.35 F ¼ 0.2 df ¼ 1, 30 P > 0.05 3.324 0.39 3.487 0.36 F ¼ 0.3 df ¼ 1, 30 P > 0.05 3.273 0.48 3.347 0.32 F ¼ 1.5 df ¼ 1, 30 P > 0.05 3.132 0.41 3.225 0.33 F ¼ 0.5 df ¼ 1, 30 P > 0.05 1448.51 174.6 1531.5 124.7 F ¼ 2.5 df ¼ 1, 31 P > 0.05
HAM-D: Hamilton depression rating scale; MADRS: MontgomeryeÅsberg depression rating scale.
Fig. 3. The correlation between the duration without antidepressant treatment and the structural changes hippocampal formation in depressed patients.
54
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
HF volume may not precede illness onset (Kaymak et al., 2010; Kronmuller et al., 2009; MacQueen et al., 2003; van Eijndhoven et al., 2009). Although the majority of first-episode studies did not find significant volumetric difference between patient and control groups, a meta-analyses including all the studies, suggested that depressed patients had smaller hippocampi than controls (Cole et al., 2011). Therefore, individual studies using small patient samples, such as the present study, might have insufficient statistical power to show the difference (assuming one exists) between healthy control and first-episode patient groups. Our present finding of unchanged HF volume at a 5-year follow-up, although unexpected, is consistent with other studies following patients from 6 months to 11 years (Ahdidan et al., 2011; Frodl et al., 2008a; Hou et al., 2012; Soriano-Mas et al., 2011; Vythilingam et al., 2004). Null results of volumetric analyses did not exclude local structural and shape changes in the subfields of the HF. Previous shape analyses studies, despite finding no global volumetric differences between depressed and control groups, have detected structural alterations in depressed patients (Bearden et al., 2009; Posener et al., 2003). Assessing structural changes within the HF, this study found significant main effects of group, time, and a group time interaction, as shown in Fig. 1. These effects represent both local atrophic and hypertrophic changes, as shown in Fig. 2. Patients and controls did not differ at baseline, and controls showed no change at the 5 year follow-up. Regional structural changes were found among the patients at follow-up, in the head and tail of the HF in CA1 and subiculum. A significant time effect was found in the head and at the junction of the body and tail, extending into the subiculum on the ventral side of the left HF. Furthermore, a significant negative correlation was found with the number of days in depression without antidepressant treatment in the CA1 region in the head and tail of the HF bilaterally. One might speculate that over a longer period of time these mostly atrophic regional changes might eventually result in an overall HF volume reduction. Two cross-sectional studies have reported regional HF gray matter changes with distributions similar to what we observed in our depressed patients (Cole et al., 2010; Posener et al., 2003). However, in these cross-sectional studies, duration of illness varies considerably and neither of the studies specifically tested the effect of duration of illness on hippocampal shape. On the other hand, another study reported depression severity was associated with greater left HF atrophy, particularly CA1 and subiculum in depressed patients with the mean duration of illness of 11 years (Bearden et al., 2009). Our finding of 5-year regional HF shape alterations only in the depressed patients, and a negative correlation between duration of untreated depression and thinning of gray matter, may indicate an ongoing neurodegenerative process during MDD. A number of pathological process have been proposed as the basis for degenerative changes, including high cortisol levels, glutamate neurotoxicity, and reduced neurotrophin levels (Banasr et al., 2011). Preclinical studies have shown that CA1, CA3, and dentate regions are vulnerable to stress, which leads to volume reduction by diminution of the dendritiric arborization and spine density of neurons (McEwen, 2007). Post-mortem studies of depressed patients support animal studies revealing complex neuronal abnormalities in subiculum and CA1eCA3 regions (Rosoklija et al., 2000; Stockmeier et al., 2004). Although these findings help us interpret the gray matter thinning in our results, we must point out that there were regions that were associated with gray matter thickening close to thinned areas. Our study could not provide information on morphological changes in these regions, but based on previous animal and post-mortem studies, we may speculate that regional thickening of HF gray matter might be related to
adaptive changes in neurons or increased glial cells (McEwen, 2007; Stockmeier et al., 2004). Another possible explanation for thickening is neurogenesis under the influence of antidepressant treatment (Banasr and Duman, 2007; Boldrini et al., 2009). Neurogenesis is seen mainly in the dentate gyrus, and in adult mice HF neurogenesis is reduced by stress and depressive behavior. In the present study, there was no significant effect of the duration of antidepressant treatment, which may reflect poor medication compliance among some of our patient subjects. It may also reflect the insensitivity of current shape analyses approaches to changes in the dentate gyrus. Thus, we propose that during depression the HF shows structural deformations (generally thinning certain subfields but thickening of others) that over time lead to a net reduction of HF volume. The heterogeneity of study samples always handicaps interpretation of depression study results. For example, our study sample was composed of outpatients at a 5-year follow-up. Only one patient had needed to be hospitalized during the intervening time period and many had been in remission as long as they were in treatment. We believe that our sample represents the everyday outpatients who compose the most of the depressed population. Our results (HF shape change in the long-term and its relationship with the duration of untreated depression) suggest that patients with even longer durations of severe depression might have more prominent HF changes than in our sample, consistent with the findings of Frodl et al. (2008b). The main limitation of the study is the small sample size. Many patients and controls volunteering as subjects in our earlier study had moved or were otherwise lost to follow-up. Another limitation might be the medication adherence of patients during the 5 years separating baseline and follow-up evaluations, because antidepressant treatment might increase HF gray matter (Frodl et al., 2008a). Poor adherence to treatment may underlie our lack of significant findings. One additional limitation may be the clinical status of our patients, one third of who were in remission at follow-up. There are also methodological limitations of our study. During tracing, alveus and fimbria were partially included into superior border of our ROI as in many of the other hippocampus studies (discussed in Konrad et al., 2009). This unavoidable situation is intended to prevent confusion between white and gray matter when working with images obtained from 1.5T MRI scanners. Although there are studies (Frodl et al., 2002; Tupler and De Bellis, 2006) suggesting that depression may affect white and gray matter differently, it should be keep in mind that white matter constitutes just 3% of the hippocampus, and many segmentation protocols are prone to misclassifying signal intensities in hippocampal areas (Konrad et al., 2009). Another methodological limitation might be the thickness of our coronal slices, which was 2 mm. In metaanalyses, researchers could not find any significant effect of slice thickness on the results. However, during our shape analyses, we resampled the images into 0.5 mm 0.5 mm 0.5 mm voxels with SPHARM. During resampling, partial volume effects might have influenced our results. One additional limitation might be the threshold for post-hoc analyses. We adjusted the alpha level to p < 0.01 instead of using the “False discovery rate” method. That method is preferred for multiple comparison correction to prevent type I statistical error in neuroimaging studies. However, this was an exploratory study rather than a causal study, and conservative thresholds might mask weak effects that merit additional research. In conclusion, we mapped HF shape changes in a group of firstepisode MDD patients reassessed after five years, and found prominent gray matter alterations in CA1eCA3 field and subiculum. As these changes occurred in the five years following the first episode of depression, our findings constitute evidence for
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
structural alteration as a result of pathophysiological processes in depression. Financial support This study was supported by the research grants from Ege University Research Fund and Ihsan Dogramacı Foundation. Disclosure Authors declare that they have no potential conflict of interest. Acknowledgment None. References Ahdidan J, Hviid LB, Chakravarty MM, Ravnkilde B, Rosenberg R, Rodell A, et al. Longitudinal MR study of brain structure and hippocampus volume in major depressive disorder. Acta Psychiatrica Scandinavica 2011;123:211e9. Banasr M, Duman RS. Regulation of neurogenesis and gliogenesis by stress and antidepressant treatment. CNS & Neurological Disorders e Drug Targets 2007; 6:311e20. Banasr M, Dwyer JM, Duman RS. Cell atrophy and loss in depression: reversal by antidepressant treatment. Current Opinion In Cell Biology 2011;23:730e7. Bearden CE, Thompson PM, Avedissian C, Klunder AD, Nicoletti M, Dierschke N, et al. Altered hippocampal morphology in unmedicated patients with major depressive illness. American Society for Neurochemistry 2009;1:265e73. Boldrini M, Underwood MD, Hen R, Rosoklija GB, Dwork AJ, John Mann J, et al. Antidepressants increase neural progenitor cells in the human hippocampus. Neuropsychopharmacology 2009;34:2376e89. Cheng YQ, Xu J, Chai P, Li HJ, Luo CR, Yang T, et al. Brain volume alteration and the correlations with the clinical characteristics in drug-naive first-episode MDD patients: a voxel-based morphometry study. Neuroscience Letters 2010;480:30e4. Cole J, Costafreda SG, McGuffin P, Fu CH. Hippocampal atrophy in first episode depression: a meta-analysis of magnetic resonance imaging studies. Journal of Affective Disorders 2011;134:483e7. Cole J, Toga AW, Hojatkashani C, Thompson P, Costafreda SG, Cleare AJ, et al. Subregional hippocampal deformations in major depressive disorder. Journal of Affective Disorders 2010;126:272e7. Czeh B, Lucassen PJ. What causes the hippocampal volume decrease in depression? Are neurogenesis, glial changes and apoptosis implicated? European Archives of Psychiatry and Clinical Neuroscience 2007;257:250e60. Eker C, Gonul AS. Volumetric MRI studies of the hippocampus in major depressive disorder: meanings of inconsistency and directions for future research. World Journal of Biological Psychiatry 2010;11:19e35. Eker C, Kitis O, Taneli F, Eker OD, Ozan E, Yucel K, et al. Correlation of serum BDNF levels with hippocampal volumes in first episode, medication-free depressed patients. European Archives of Psychiatry and Clinical Neuroscience 2010;260: 527e33. Eker MC, Kitis O, Okur H, Eker OD, Ozan E, Isikli S, et al. Smaller hippocampus volume is associated with short variant of 5-HTTLPR polymorphism in medication-free major depressive disorder patients. Neuropsychobiology 2011;63:22e8. Fanselow MS, Dong HW. Are the dorsal and ventral hippocampus functionally distinct structures? Neuron 2010;65:7e19. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical interview for DSM-IV axis I disorders, clinician version (SCID-CV). Washington, D.C.: American Psychiatric Press, Inc.; 1996. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical interview for DSMIV-TR axis I disorders, research version, non-patient edition (SCID-I/NP). New York: Biometrics Research, New York State Psychiatric Institute; November 2002. Frodl T, Jager M, Smajstrlova I, Born C, Bottlender R, Palladino T, et al. Effect of hippocampal and amygdala volumes on clinical outcomes in major depression: a 3-year prospective magnetic resonance imaging study. Journal of Psychiatry & Neuroscience 2008a;33:423e30. Frodl T, Koutsouleris N, Bottlender R, Born C, Jager M, Scupin I, et al. Depressionrelated variation in brain morphology over 3 years: effects of stress? Archives of General Psychiatry 2008b;65:1156e65. Frodl T, Meisenzahl EM, Zetzsche T, Born C, Groll C, Jager M, et al. Hippocampal changes in patients with a first episode of major depression. American Journal of Psychiatry 2002;159:1112e8.
55
Frodl T, Reinhold E, Koutsouleris N, Donohoe G, Bondy B, Reiser M, et al. Childhood stress, serotonin transporter gene and brain structures in major depression. Neuropsychopharmacology 2010;35:1383e90. Gonul AS, Kitis O, Eker MC, Eker OD, Ozan E, Coburn K. Association of the brainderived neurotrophic factor Val66Met polymorphism with hippocampus volumes in drug-free depressed patients. World Journal of Biological Psychiatry 2011;12:110e8. Hamilton M. A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry 1960;23:56e62. Hou Z, Yuan Y, Zhang Z, Bai F, Hou G, You J. Longitudinal changes in hippocampal volumes and cognition in remitted geriatric depressive disorder. Behavioural Brain Research 2012;227:30e5. Hviid LB, Ravnkilde B, Ahdidan J, Rosenberg R, Stodkilde-Jorgensen H, Videbech P. Hippocampal visuospatial function and volume in remitted depressed patients: an 8-year follow-up study. Journal of Affective Disorders 2010;125:177e83. Kaymak SU, Demir B, Senturk S, Tatar I, Aldur MM, Ulug B. Hippocampus, glucocorticoids and neurocognitive functions in patients with first-episode major depressive disorders. European Archives of Psychiatry and Clinical Neuroscience 2010;260:217e23. Konrad C, Ukas T, Nebel C, Arolt V, Toga AW, Narr KL. Defining the human hippocampus in cerebral magnetic resonance imagesean overview of current segmentation protocols. NeuroImage 2009;47:1185e95. Kronmuller KT, Schroder J, Kohler S, Gotz B, Victor D, Unger J, et al. Hippocampal volume in first episode and recurrent depression. Psychiatry Research 2009; 174:62e6. Lupien SJ, McEwen BS, Gunnar MR, Heim C. Effects of stress throughout the lifespan on the brain, behaviour and cognition. Nature Reviews Neuroscience 2009;10:434e45. MacQueen G, Frodl T. The hippocampus in major depression: evidence for the convergence of the bench and bedside in psychiatric research? Molecular Psychiatry 2011;16:252e64. MacQueen GM, Campbell S, McEwen BS, Macdonald K, Amano S, Joffe RT, et al. Course of illness, hippocampal function, and hippocampal volume in major depression. Proceedings of the National Academy of Sciences of the United States of America 2003;100:1387e92. McEwen BS. Physiology and neurobiology of stress and adaptation: central role of the brain. Physiological Reviews 2007;87:873e904. McKinnon MC, Yucel K, Nazarov A, MacQueen GM. A meta-analysis examining clinical predictors of hippocampal volume in patients with major depressive disorder. Journal of Psychiatry & Neuroscience 2009;34:41e54. Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9:97e113. Posener JA, Wang L, Price JL, Gado MH, Province MA, Miller MI, et al. Highdimensional mapping of the hippocampus in depression. American Journal of Psychiatry 2003;160:83e9. Rosoklija G, Toomayan G, Ellis SP, Keilp J, Mann JJ, Latov N, et al. Structural abnormalities of subicular dendrites in subjects with schizophrenia and mood disorders: preliminary findings. Archives of General Psychiatry 2000;57:349e56. Sheline YI. Depression and the hippocampus: cause or effect? Biological Psychiatry 2011;70:308e9. Sheline YI, Gado MH, Kraemer HC. Untreated depression and hippocampal volume loss. American Journal of Psychiatry 2003;160:1516e8. Soriano-Mas C, Hernandez-Ribas R, Pujol J, Urretavizcaya M, Deus J, Harrison BJ, et al. Cross-sectional and longitudinal assessment of structural brain alterations in melancholic depression. Biological Psychiatry 2011;69:318e25. Stockmeier CA, Mahajan GJ, Konick LC, Overholser JC, Jurjus GJ, Meltzer HY, et al. Cellular changes in the postmortem hippocampus in major depression. Biological Psychiatry 2004;56:640e50. Styner M, Gerig G, Lieberman J, Jones D, Weinberger D. Statistical shape analysis of neuroanatomical structures based on medial models. Medical Image Analysis 2003;7:207e20. Styner M, Oguz I, Xu S, Brechbuhler C, Pantazis D, Levitt JJ, et al. Framework for the statistical shape analysis of brain structures using SPHARM-PDM. Insight Journal 2006:242e50. Tupler LA, De Bellis MD. Segmented hippocampal volume in children and adolescents with posttraumatic stress disorder. Biological Psychiatry 2006;59:523e9. van Eijndhoven P, van Wingen G, van Oijen K, Rijpkema M, Goraj B, Jan Verkes R, et al. Amygdala volume marks the acute state in the early course of depression. Biological Psychiatry 2009;65:812e8. Videbech P, Ravnkilde B. Hippocampal volume and depression: a meta-analysis of MRI studies. American Journal of Psychiatry 2004;161:1957e66. Vythilingam M, Heim C, Newport J, Miller AH, Anderson E, Bronen R, et al. Childhood trauma associated with smaller hippocampal volume in women with major depression. American Journal of Psychiatry 2002;159:2072e80. Vythilingam M, Vermetten E, Anderson GM, Luckenbaugh D, Anderson ER, Snow J, et al. Hippocampal volume, memory, and cortisol status in major depressive disorder: effects of treatment. Biological Psychiatry 2004;56:101e12.
Contents lists available at SciVerse ScienceDirect
Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires
Altered hippocampal formation shape in first-episode depressed patients at 5-year follow-upq Serhan Isıklı a, Onur Ugurlu a, b, Ece Durmusoglu a, Gozde Kizilates a, b, Omer Kitis a, c, Erol Ozan d, Cagdas Eker a, Kerry Coburn e, Ali Saffet Gonul a, e, * a
Ege University School of Medicine, Department of Psychiatry, SoCAT Lab, Bornova, Izmir, Turkey Ege University, Department of Mathematics, Bornova, Izmir, Turkey Ege University School of Medicine, Department of Radiology, Neuroradiology Unit, Bornova, Izmir, Turkey d Ataturk University School of Medicine, Department of Psychiatry, Erzurum, Turkey e Mercer University School of Medicine, Department of Psychiatry and Behavioral Sciences Macon, GA, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 March 2012 Received in revised form 17 August 2012 Accepted 21 August 2012
It is generally accepted that patients with major depressive disorder have smaller hippocampus size compared to healthy people. However, it is still not known if this situation exists before the onset of the disease or is a result of the toxic mechanism created by the depression itself. The findings of the longterm follow-up studies of first-episode depressed patients might contribute to solve the ongoing problem. In this study, the hippocampus of 18 first-episode patients who were followed-up for 5 years, were compared with those of healthy controls. There were no volumetric differences among groups neither at the baseline nor after 5 years of follow-up. However, shape analyses, using high dimensional mapping methods, revealed regional structural changes in the head and tail of the hippocampal formation in CA1 and subiculum regions in patients at the follow-up. Furthermore, a significant negative correlation was found with the number of days in depression without antidepressant treatment in the CA1 region in the head and tail of the hippocampal formation bilaterally. The findings of this study support the hypothesis that pathophysiological processes of depression induce structural alterations in depressed patients. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Major depressive disorder Magnetic resonance imaging Hippocampus Follow-up
1. Introduction The Hippocampal Formation (HF), including the hippocampus (Cornu Ammonis) proper, the dentate gyrus, and the subiculum, is a set of closely related limbic structures playing crucial roles in memory, spatial navigation, learning, contextual fear-conditioning and neuroendocrine regulation (Fanselow and Dong, 2010). Major depressive disorder (MDD) has several signs and symptoms that appear to be related to HF functions. Therefore, researchers have focused on structural and functional abnormalities of the HF in depressed patients. Accumulated evidence has shown that, in addition to functional abnormalities, the HF of depressed patients often has overall volumetric and regional shape abnormalities (MacQueen and Frodl, 2011; Videbech and Ravnkilde, 2004). Many q This study was conducted in Ege University School of Medicine, Department of Psychiatry, SoCAT Lab, Bornova, Izmir, Turkey. * Corresponding author. Ege University School of Medicine, Department of Psychiatry, SoCAT Lab, Bornova, Izmir, Turkey. Tel.: þ90 2323904163. E-mail address: [email protected] (A.S. Gonul). 0022-3956/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jpsychires.2012.08.022
volumetric studies, though not all, have reported smaller HF volumes compared to healthy controls (Eker and Gonul, 2010; Videbech and Ravnkilde, 2004). Some of these studies also found evidence that a longer cumulative duration of lifetime depression, a larger total number of recurrent episodes, and an earlier age of depression onset are associated with smaller HF volume (Cheng et al., 2010; MacQueen et al., 2003; McKinnon et al., 2009; Sheline et al., 2003). Thus, these findings support the view that the reduction of the HF is a cumulative process beginning around the time that symptoms first manifest and continuing over many years of depression. Physiologically, HF atrophy may be mediated by a chronic stress-induced HypothalamicePituitaryeAdrenal (HPA) axis response, which increases cortisol and reduces Brain-derived Neurotrophic Factor (BDNF). These neurohormonal changes in turn reduce neurogenesis and alter the somatodendritic, axonal, and synaptic components of neurons, as well as causing glial changes. Together these changes produce abnormal neuronal functioning (Czeh and Lucassen, 2007). However, there are also studies showing reduced HF volume only in patients carrying susceptibility genes or patients who had childhood trauma (Eker et al., 2011;
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
Frodl et al., 2010; Lupien et al., 2009; Vythilingam et al., 2002). Such findings suggest that HF atrophy is due to genetic and environmental factors during the early years of life (Lupien et al., 2009; McEwen, 2007), and that HF volume reduction might be a vulnerability factor for depression occurring before the onset of the clinical symptoms. This idea is supported by a recent meta-analysis which included only first episode depressed patients (Cole et al., 2011). One approach to investigating these two possibilities is to study HF volume change as a function of time (MacQueen and Frodl, 2011). HF atrophy at baseline in first-episode MDD patients, without further reductions over a 5 year period would constitute evidence of an atrophic process occurring prior to MDD, perhaps dating from childhood. Conversely, normal HF volume at baseline followed by HF atrophy at 5 year follow-up would be consistent with MDD as a causal mechanism. To our knowledge, there are five published studies examining HF atrophy in MDD patients over time periods of 6 months to 11 years (Frodl et al., 2008a, 2008b; Hou et al., 2012; Hviid et al., 2010; Vythilingam et al., 2004). Among those studies, the only one which used voxel-based analyses (Frodl et al., 2008b), suggested partial volume loss in the hippocampus of depressed patients in a three-year follow-up. The other four, which used tracing methods (the gold standard of volumetric analysis), did not find an on-going volume reduction in patients at follow-up. However, reduced HF volume was associated with poor clinical outcome. A limitation of these studies is that overall HF volume may not be as sensitive as regional HF volume (i.e., HF shape) to the changes associated with MDD. Therefore, we examined changes in both HF volume and shape over a 5 year period in first-episode MDD patients. 2. Methods The recruited patients were among the 68 first episode righthanded drug-free depressed patients who participated in earlier published studies (Eker et al., 2010; Gonul et al., 2011). Forty patients were still in treatment at the end of 5 years, of whom 23 agreed to a second MRI scan. However, 5 of the patients could not be scanned again. Two were very busy and failed to follow MRI schedule, one was pregnant, one had hydrocortisone treatment and one’s diagnoses was changed to bipolar disorder. All patients had been diagnosed in the earlier study with MDD (without comorbidity for any other axis I disorder) according to Diagnostic and Manual of Mental Disorders, fourth edition (DSM-IV) criteria and confirmed with the Structured Clinical Interview for DSM-IV (SCID) (First et al., 1996) by a consensus of at least 2 psychiatrists. All patients were re-evaluated with the SCID before their second scan, and those 18 whose MDD diagnosis was confirmed were included in this study. The severity of their depression was evaluated with Hamilton Depression Scale (17 items) (Hamilton, 1960). The control group consisted of 18 right-handed healthy subjects who also participated in the previous studies. During control group recruitment, an attempt was made to match age, gender, education, and time gap between the two MRI scans. However, age did not match because of recruitment problems in the control group. Control subjects were re-screened with non-patient version of SCID to exclude any Axis I disorders (First et al., 2002). A structured interview assessed exclusion criteria (previous head injury with loss of consciousness, hydrocortisone treatment, alcohol or substance abuse, neurologic disease or chronic medical disease including hypertension or diabetes mellitus, first-degree relative with bipolar or psychotic disorder) for all participants. Prior to this follow-up one psychiatrist from the research team (S. I.) screened all the hospital records of the patients and the controls and interviewed all subjects to investigate the intervening period
51
for any major social or personal trauma or conflict. The primary aim of this interview was to obtain any information that might have been missed with the structured diagnostic interview. A secondary aim was to obtain information related to any personality disorders among the subjects. However, no patients or controls were excluded for this reason. Handedness was determined using the Edinburg Inventory (Oldfield, 1971). The study design was approved by the local ethics committee and was prepared according to the ethical standards of the Declaration of Helsinki. The written informed consent was obtained from all subjects after a description of the study. 2.1. MRI image acquisition The imaging was performed on a 1.5-T MR unit (Magnetom Vision Siemens, Erlangen, Germany) with a circularly polarized head coil. The standard MRI scan included multiplanar turbo spinecho T1-weighted (T1-W) (TR/TE: 650/14 ms) and T2-weighted (T2-W) (TR/TE: 3800/90 ms) images. In addition to conventional sequences, a 3D MP-RAGE sequence (TR/TE: 9.70/4.00 ms, slice thickness: 2 mm, interslice gap: 0 mm, voxel size: 0.5 mm 0.5 mm 2 mm) in the coronal plane was obtained for volumetric analysis. All imaging data were transferred to a PC workstation for further analyses. 2.2. Hippocampal formation measurement and shape analyses We used ITK-SNAP software (www.itksnap.org) to outline the HF borders. This software provides an interactive labeling function with simultaneously coronal, sagittal and axial views and concurrently provides surface rendering of volume of interest. The definition of the HF borders was described in detail previously (Gonul et al., 2011). An experienced neuroradiologist (O.K.) and a trained psychiatrist (S.I.), who were blind to the identities of the images, manually outlined the borders. To determine the interrater reliability, the HF ROI was independently traced on 10 randomly selected brains by the two raters. There was a high intraclass correlation for both the interrater reliability (rICC ¼ 0.96) and the intrarater reliability (rICC ¼ 0.96). The 3D HF shape analysis is based on the use of spherical harmonic basis functions (SPHARM) to fit the HF boundary. This process was based on a uniform-icosahedron subdivision of the spherical parameterization: A detailed description about the algorithms can be found in previous articles (Styner et al., 2003, 2006). The binary images containing the surface points determined by boundary identification were pre-processed to fill small holes and minimally smoothed. At this point, the vertexes of the voxels described the surface. A triangulation mesh was generated from each image by dividing each exterior voxel face into two triangles. A spherical parameterization was then computed for the triangulation mesh using a distortion minimizing spherical mapping. Finally, the SPHARM coefficients were computed from the mesh and its parameterization using straightforward fitting to a spherical harmonic function. The correspondences across all surfaces were computed by aligning the first-order ellipsoids of all the parameterizations. All the triangulated meshes of the HF were spatially aligned to the template, which was obtained from SPHARM by using a rigid transform. 2.3. Statistical analysis Clinical and demographic variables were compared by t test or chi square. Repeated measures of analyses of covariance (ANCOVA) were used to compare HF volumes among the groups. Side and time were within the group factors and age, gender, and total brain
52
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
volume were covariates. We used the similar ANCOVA design without between-group factors to assess the effects on HF volume of duration of illness, duration of antidepressant treatment, and number of depressive episodes between the scans. An alpha level of p < 0.05 was used for all tests. For the shape analyses, we used 2 2 MANCOVA model with age, gender, and total brain volume as confounding factors. The analyses generated raw (unadjusted for multiple comparisons) significance maps showing raw p-values exceeding 0.05 across the surface. For post-hoc within- and between-group analyses the alpha level was adjusted to 0.01 to decrease type I errors. For visualization (Figs. 1 and 2 left, 3), surface locations are in blue if the differences are not statistically significant. In addition to significance maps, we present volume difference in millimeters among the groups in distance maps (Fig. 2 right). 3. Results 3.1. Clinical data and volumetric comparison Gender and education were similar between patient and control groups but patients were older than controls (Table 1). Among patients, 12 were in depressive episodes while 6 were in remission (Mean HAM-D scores: 17.3 and 3.2; respectively). During five years intervening between the original study and this follow-up one patient was chronically depressed, one patient had four depressive episodes, five patients had three episodes, seven patients had two episodes and four patients had one episode (the numbers include the current episode if they were depressed during second scan) Eight patients were on single antidepressant treatment, one was on antidepressant and antipsychotic combination and another patient was on a low dose amisulpride treatment. Eight patients were effectively not on antidepressant treatment due to low adherence. Comparison of total HF volume of the patient and the control groups with repeated measures ANCOVA revealed neither group nor time effects (F ¼ 0.64 df ¼ 1, 30 p > 0.05, F ¼ 0.95 df ¼ 1, 30
p > 0.05; respectively). The duration of depressed days, duration of antidepressant treatment, and number of episodes between the scans similarly had no effects (F ¼ 0.8 df ¼ 1, 12 p > 0.05, F ¼ 0.1 df ¼ 1, 12 p > 0.05, F ¼ 0.4 df ¼ 1, 12 p > 0.05; respectively). HAM-D scores were not correlated with the HF volumes. There were no differences in total brain volumes among the groups (F ¼ 2.5 df ¼ 1, 31 P > 0.05) (Table 1). 3.2. Shape analyses HF shape analyses with MANCOVA (age, gender and total brain volume as confounding factors) revealed a significant group effect in the medial aspect of the HF head and the lateral aspect of the posterior HF bilaterally (Fig. 1). We also observed a significant group effect in the subiculum from mid-body to tail in the right HF. When we investigated the time effect, we found that the left HF had multiple affected areas. One of the affected areas was at the lateral side of the head, consisting mainly of CA1. Another was a large band-like area starting in dorsal aspects of CA2e3 and extending to subiculum on the ventral side. A significant group time interaction was present in these same areas. Post-hoc analysis of the group time interaction revealed that depressed patients and healthy controls had similar HF shapes at the baseline. At the five year follow-up, controls did not show any significant gray matter HF change. On the other hand, patients showed heterogeneous gray matter changes; gray matter loss at the lateral tip of the head (subiculum) of the HF extending to the dorsal surface at CA1 region (Fig. 2). This initial thinning merged into a thickening at the end of the significant area. At the junction of the body and tail of the HF we observed a band-like area of gray matter changes, with thickening at the medial and lateral aspects of the dorsal side of the band, and thinning along the ventral side. Finally, there was apparent though non-significant gray matter loss at the head and posterior of the HF bilaterally in depressed patients. MANCOVA additionally showed that the number of days in depression without antidepressant treatment between two scans
Fig. 1. Regional structural alterations observed in the hippocampal formation of the patients with major depressive disorder relative those of the healthy comparison subjects.
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
53
Fig. 2. Structural changes observed in depressed patients at the end of five-years follow-up.
had a significant effect on CA1 and subiculum at the head and postero-lateral side on the HF bilaterally (Fig. 3). When we excluded two outliners from the analyses, the results did not change. There were no significant effects of the duration of antidepressant treatment on HF shape.
Table 1 Clinical and demographic variables of patient and control groups. Variables
Patients (N ¼ 18)
Controls (N ¼ 18)
Comparison
Age
46.11 8.5
37.56 6.15
Sex (male/female)
6/12
9/9
Education
13.1 3.03
15 3.52
T ¼ 3.45 df ¼ 34 P ¼ 0.001 X2 ¼ 1.03 P > 0.05 T ¼ 1.77 df ¼ 34 P > 0.05
Age of onset HAM-D scores MADRS Number of episodes Duration of depressed days Duration of depressed days w/o antidepressant treatment Current mood state Depressed (N%) Euthymic (N%) Duration of antidepressant treatment (days) Duration between two MRI scan (days) Right hippocampus volume (baseline) (cc) Right hippocampus volume (5 years) (cc) Left hippocampus volume (baseline) (cc) Left hippocampus volume (5 years) (cc) Total brain volume (cc)
36.89 8.84 12.61 8.2 13.89 8.6 2.33 1.08 1145.3 130 638.1 909
e e e e e e
12 (66%) 6 (33%) 658.7 763.3
e
2073.8 223.3
1903.1 356
4. Discussion This study is a 5-year follow-up assessing the changes in HF volume and shape, among first-episode MDD patients and healthy controls used in two previous studies (Eker et al., 2010; Gonul et al., 2011). Overall HF volumes of patients did not differ from those of controls, either at baseline or at follow-up. However, at follow-up, regional changes in HF shape were found among the patients but not the controls. To our knowledge, this is the first follow-up study to show HF shape alterations in MDD patients using high dimensional mapping methods. Although reduced HF volume is a common finding in MDD patients, the current literature is unclear whether the reduction is the result of a faulty neurodevelopmental process before the disease onset or due to neurotoxicity associated with depression pathophysiology, or both (Sheline, 2011). Long-term studies following MDD patients from their first episode might help elucidate the pathophysiological process. Our findings of normal HF volumes in MDD patients at the time of their first depressive episode are consistent with many (though not all) first episode studies, supporting the idea that reductions in
T ¼ 1.68 df ¼ 34 P > 0.05 3.416 0.36 3.485 0.35 F ¼ 0.2 df ¼ 1, 30 P > 0.05 3.324 0.39 3.487 0.36 F ¼ 0.3 df ¼ 1, 30 P > 0.05 3.273 0.48 3.347 0.32 F ¼ 1.5 df ¼ 1, 30 P > 0.05 3.132 0.41 3.225 0.33 F ¼ 0.5 df ¼ 1, 30 P > 0.05 1448.51 174.6 1531.5 124.7 F ¼ 2.5 df ¼ 1, 31 P > 0.05
HAM-D: Hamilton depression rating scale; MADRS: MontgomeryeÅsberg depression rating scale.
Fig. 3. The correlation between the duration without antidepressant treatment and the structural changes hippocampal formation in depressed patients.
54
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
HF volume may not precede illness onset (Kaymak et al., 2010; Kronmuller et al., 2009; MacQueen et al., 2003; van Eijndhoven et al., 2009). Although the majority of first-episode studies did not find significant volumetric difference between patient and control groups, a meta-analyses including all the studies, suggested that depressed patients had smaller hippocampi than controls (Cole et al., 2011). Therefore, individual studies using small patient samples, such as the present study, might have insufficient statistical power to show the difference (assuming one exists) between healthy control and first-episode patient groups. Our present finding of unchanged HF volume at a 5-year follow-up, although unexpected, is consistent with other studies following patients from 6 months to 11 years (Ahdidan et al., 2011; Frodl et al., 2008a; Hou et al., 2012; Soriano-Mas et al., 2011; Vythilingam et al., 2004). Null results of volumetric analyses did not exclude local structural and shape changes in the subfields of the HF. Previous shape analyses studies, despite finding no global volumetric differences between depressed and control groups, have detected structural alterations in depressed patients (Bearden et al., 2009; Posener et al., 2003). Assessing structural changes within the HF, this study found significant main effects of group, time, and a group time interaction, as shown in Fig. 1. These effects represent both local atrophic and hypertrophic changes, as shown in Fig. 2. Patients and controls did not differ at baseline, and controls showed no change at the 5 year follow-up. Regional structural changes were found among the patients at follow-up, in the head and tail of the HF in CA1 and subiculum. A significant time effect was found in the head and at the junction of the body and tail, extending into the subiculum on the ventral side of the left HF. Furthermore, a significant negative correlation was found with the number of days in depression without antidepressant treatment in the CA1 region in the head and tail of the HF bilaterally. One might speculate that over a longer period of time these mostly atrophic regional changes might eventually result in an overall HF volume reduction. Two cross-sectional studies have reported regional HF gray matter changes with distributions similar to what we observed in our depressed patients (Cole et al., 2010; Posener et al., 2003). However, in these cross-sectional studies, duration of illness varies considerably and neither of the studies specifically tested the effect of duration of illness on hippocampal shape. On the other hand, another study reported depression severity was associated with greater left HF atrophy, particularly CA1 and subiculum in depressed patients with the mean duration of illness of 11 years (Bearden et al., 2009). Our finding of 5-year regional HF shape alterations only in the depressed patients, and a negative correlation between duration of untreated depression and thinning of gray matter, may indicate an ongoing neurodegenerative process during MDD. A number of pathological process have been proposed as the basis for degenerative changes, including high cortisol levels, glutamate neurotoxicity, and reduced neurotrophin levels (Banasr et al., 2011). Preclinical studies have shown that CA1, CA3, and dentate regions are vulnerable to stress, which leads to volume reduction by diminution of the dendritiric arborization and spine density of neurons (McEwen, 2007). Post-mortem studies of depressed patients support animal studies revealing complex neuronal abnormalities in subiculum and CA1eCA3 regions (Rosoklija et al., 2000; Stockmeier et al., 2004). Although these findings help us interpret the gray matter thinning in our results, we must point out that there were regions that were associated with gray matter thickening close to thinned areas. Our study could not provide information on morphological changes in these regions, but based on previous animal and post-mortem studies, we may speculate that regional thickening of HF gray matter might be related to
adaptive changes in neurons or increased glial cells (McEwen, 2007; Stockmeier et al., 2004). Another possible explanation for thickening is neurogenesis under the influence of antidepressant treatment (Banasr and Duman, 2007; Boldrini et al., 2009). Neurogenesis is seen mainly in the dentate gyrus, and in adult mice HF neurogenesis is reduced by stress and depressive behavior. In the present study, there was no significant effect of the duration of antidepressant treatment, which may reflect poor medication compliance among some of our patient subjects. It may also reflect the insensitivity of current shape analyses approaches to changes in the dentate gyrus. Thus, we propose that during depression the HF shows structural deformations (generally thinning certain subfields but thickening of others) that over time lead to a net reduction of HF volume. The heterogeneity of study samples always handicaps interpretation of depression study results. For example, our study sample was composed of outpatients at a 5-year follow-up. Only one patient had needed to be hospitalized during the intervening time period and many had been in remission as long as they were in treatment. We believe that our sample represents the everyday outpatients who compose the most of the depressed population. Our results (HF shape change in the long-term and its relationship with the duration of untreated depression) suggest that patients with even longer durations of severe depression might have more prominent HF changes than in our sample, consistent with the findings of Frodl et al. (2008b). The main limitation of the study is the small sample size. Many patients and controls volunteering as subjects in our earlier study had moved or were otherwise lost to follow-up. Another limitation might be the medication adherence of patients during the 5 years separating baseline and follow-up evaluations, because antidepressant treatment might increase HF gray matter (Frodl et al., 2008a). Poor adherence to treatment may underlie our lack of significant findings. One additional limitation may be the clinical status of our patients, one third of who were in remission at follow-up. There are also methodological limitations of our study. During tracing, alveus and fimbria were partially included into superior border of our ROI as in many of the other hippocampus studies (discussed in Konrad et al., 2009). This unavoidable situation is intended to prevent confusion between white and gray matter when working with images obtained from 1.5T MRI scanners. Although there are studies (Frodl et al., 2002; Tupler and De Bellis, 2006) suggesting that depression may affect white and gray matter differently, it should be keep in mind that white matter constitutes just 3% of the hippocampus, and many segmentation protocols are prone to misclassifying signal intensities in hippocampal areas (Konrad et al., 2009). Another methodological limitation might be the thickness of our coronal slices, which was 2 mm. In metaanalyses, researchers could not find any significant effect of slice thickness on the results. However, during our shape analyses, we resampled the images into 0.5 mm 0.5 mm 0.5 mm voxels with SPHARM. During resampling, partial volume effects might have influenced our results. One additional limitation might be the threshold for post-hoc analyses. We adjusted the alpha level to p < 0.01 instead of using the “False discovery rate” method. That method is preferred for multiple comparison correction to prevent type I statistical error in neuroimaging studies. However, this was an exploratory study rather than a causal study, and conservative thresholds might mask weak effects that merit additional research. In conclusion, we mapped HF shape changes in a group of firstepisode MDD patients reassessed after five years, and found prominent gray matter alterations in CA1eCA3 field and subiculum. As these changes occurred in the five years following the first episode of depression, our findings constitute evidence for
S. Isıklı et al. / Journal of Psychiatric Research 47 (2013) 50e55
structural alteration as a result of pathophysiological processes in depression. Financial support This study was supported by the research grants from Ege University Research Fund and Ihsan Dogramacı Foundation. Disclosure Authors declare that they have no potential conflict of interest. Acknowledgment None. References Ahdidan J, Hviid LB, Chakravarty MM, Ravnkilde B, Rosenberg R, Rodell A, et al. Longitudinal MR study of brain structure and hippocampus volume in major depressive disorder. Acta Psychiatrica Scandinavica 2011;123:211e9. Banasr M, Duman RS. Regulation of neurogenesis and gliogenesis by stress and antidepressant treatment. CNS & Neurological Disorders e Drug Targets 2007; 6:311e20. Banasr M, Dwyer JM, Duman RS. Cell atrophy and loss in depression: reversal by antidepressant treatment. Current Opinion In Cell Biology 2011;23:730e7. Bearden CE, Thompson PM, Avedissian C, Klunder AD, Nicoletti M, Dierschke N, et al. Altered hippocampal morphology in unmedicated patients with major depressive illness. American Society for Neurochemistry 2009;1:265e73. Boldrini M, Underwood MD, Hen R, Rosoklija GB, Dwork AJ, John Mann J, et al. Antidepressants increase neural progenitor cells in the human hippocampus. Neuropsychopharmacology 2009;34:2376e89. Cheng YQ, Xu J, Chai P, Li HJ, Luo CR, Yang T, et al. Brain volume alteration and the correlations with the clinical characteristics in drug-naive first-episode MDD patients: a voxel-based morphometry study. Neuroscience Letters 2010;480:30e4. Cole J, Costafreda SG, McGuffin P, Fu CH. Hippocampal atrophy in first episode depression: a meta-analysis of magnetic resonance imaging studies. Journal of Affective Disorders 2011;134:483e7. Cole J, Toga AW, Hojatkashani C, Thompson P, Costafreda SG, Cleare AJ, et al. Subregional hippocampal deformations in major depressive disorder. Journal of Affective Disorders 2010;126:272e7. Czeh B, Lucassen PJ. What causes the hippocampal volume decrease in depression? Are neurogenesis, glial changes and apoptosis implicated? European Archives of Psychiatry and Clinical Neuroscience 2007;257:250e60. Eker C, Gonul AS. Volumetric MRI studies of the hippocampus in major depressive disorder: meanings of inconsistency and directions for future research. World Journal of Biological Psychiatry 2010;11:19e35. Eker C, Kitis O, Taneli F, Eker OD, Ozan E, Yucel K, et al. Correlation of serum BDNF levels with hippocampal volumes in first episode, medication-free depressed patients. European Archives of Psychiatry and Clinical Neuroscience 2010;260: 527e33. Eker MC, Kitis O, Okur H, Eker OD, Ozan E, Isikli S, et al. Smaller hippocampus volume is associated with short variant of 5-HTTLPR polymorphism in medication-free major depressive disorder patients. Neuropsychobiology 2011;63:22e8. Fanselow MS, Dong HW. Are the dorsal and ventral hippocampus functionally distinct structures? Neuron 2010;65:7e19. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical interview for DSM-IV axis I disorders, clinician version (SCID-CV). Washington, D.C.: American Psychiatric Press, Inc.; 1996. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical interview for DSMIV-TR axis I disorders, research version, non-patient edition (SCID-I/NP). New York: Biometrics Research, New York State Psychiatric Institute; November 2002. Frodl T, Jager M, Smajstrlova I, Born C, Bottlender R, Palladino T, et al. Effect of hippocampal and amygdala volumes on clinical outcomes in major depression: a 3-year prospective magnetic resonance imaging study. Journal of Psychiatry & Neuroscience 2008a;33:423e30. Frodl T, Koutsouleris N, Bottlender R, Born C, Jager M, Scupin I, et al. Depressionrelated variation in brain morphology over 3 years: effects of stress? Archives of General Psychiatry 2008b;65:1156e65. Frodl T, Meisenzahl EM, Zetzsche T, Born C, Groll C, Jager M, et al. Hippocampal changes in patients with a first episode of major depression. American Journal of Psychiatry 2002;159:1112e8.
55
Frodl T, Reinhold E, Koutsouleris N, Donohoe G, Bondy B, Reiser M, et al. Childhood stress, serotonin transporter gene and brain structures in major depression. Neuropsychopharmacology 2010;35:1383e90. Gonul AS, Kitis O, Eker MC, Eker OD, Ozan E, Coburn K. Association of the brainderived neurotrophic factor Val66Met polymorphism with hippocampus volumes in drug-free depressed patients. World Journal of Biological Psychiatry 2011;12:110e8. Hamilton M. A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry 1960;23:56e62. Hou Z, Yuan Y, Zhang Z, Bai F, Hou G, You J. Longitudinal changes in hippocampal volumes and cognition in remitted geriatric depressive disorder. Behavioural Brain Research 2012;227:30e5. Hviid LB, Ravnkilde B, Ahdidan J, Rosenberg R, Stodkilde-Jorgensen H, Videbech P. Hippocampal visuospatial function and volume in remitted depressed patients: an 8-year follow-up study. Journal of Affective Disorders 2010;125:177e83. Kaymak SU, Demir B, Senturk S, Tatar I, Aldur MM, Ulug B. Hippocampus, glucocorticoids and neurocognitive functions in patients with first-episode major depressive disorders. European Archives of Psychiatry and Clinical Neuroscience 2010;260:217e23. Konrad C, Ukas T, Nebel C, Arolt V, Toga AW, Narr KL. Defining the human hippocampus in cerebral magnetic resonance imagesean overview of current segmentation protocols. NeuroImage 2009;47:1185e95. Kronmuller KT, Schroder J, Kohler S, Gotz B, Victor D, Unger J, et al. Hippocampal volume in first episode and recurrent depression. Psychiatry Research 2009; 174:62e6. Lupien SJ, McEwen BS, Gunnar MR, Heim C. Effects of stress throughout the lifespan on the brain, behaviour and cognition. Nature Reviews Neuroscience 2009;10:434e45. MacQueen G, Frodl T. The hippocampus in major depression: evidence for the convergence of the bench and bedside in psychiatric research? Molecular Psychiatry 2011;16:252e64. MacQueen GM, Campbell S, McEwen BS, Macdonald K, Amano S, Joffe RT, et al. Course of illness, hippocampal function, and hippocampal volume in major depression. Proceedings of the National Academy of Sciences of the United States of America 2003;100:1387e92. McEwen BS. Physiology and neurobiology of stress and adaptation: central role of the brain. Physiological Reviews 2007;87:873e904. McKinnon MC, Yucel K, Nazarov A, MacQueen GM. A meta-analysis examining clinical predictors of hippocampal volume in patients with major depressive disorder. Journal of Psychiatry & Neuroscience 2009;34:41e54. Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9:97e113. Posener JA, Wang L, Price JL, Gado MH, Province MA, Miller MI, et al. Highdimensional mapping of the hippocampus in depression. American Journal of Psychiatry 2003;160:83e9. Rosoklija G, Toomayan G, Ellis SP, Keilp J, Mann JJ, Latov N, et al. Structural abnormalities of subicular dendrites in subjects with schizophrenia and mood disorders: preliminary findings. Archives of General Psychiatry 2000;57:349e56. Sheline YI. Depression and the hippocampus: cause or effect? Biological Psychiatry 2011;70:308e9. Sheline YI, Gado MH, Kraemer HC. Untreated depression and hippocampal volume loss. American Journal of Psychiatry 2003;160:1516e8. Soriano-Mas C, Hernandez-Ribas R, Pujol J, Urretavizcaya M, Deus J, Harrison BJ, et al. Cross-sectional and longitudinal assessment of structural brain alterations in melancholic depression. Biological Psychiatry 2011;69:318e25. Stockmeier CA, Mahajan GJ, Konick LC, Overholser JC, Jurjus GJ, Meltzer HY, et al. Cellular changes in the postmortem hippocampus in major depression. Biological Psychiatry 2004;56:640e50. Styner M, Gerig G, Lieberman J, Jones D, Weinberger D. Statistical shape analysis of neuroanatomical structures based on medial models. Medical Image Analysis 2003;7:207e20. Styner M, Oguz I, Xu S, Brechbuhler C, Pantazis D, Levitt JJ, et al. Framework for the statistical shape analysis of brain structures using SPHARM-PDM. Insight Journal 2006:242e50. Tupler LA, De Bellis MD. Segmented hippocampal volume in children and adolescents with posttraumatic stress disorder. Biological Psychiatry 2006;59:523e9. van Eijndhoven P, van Wingen G, van Oijen K, Rijpkema M, Goraj B, Jan Verkes R, et al. Amygdala volume marks the acute state in the early course of depression. Biological Psychiatry 2009;65:812e8. Videbech P, Ravnkilde B. Hippocampal volume and depression: a meta-analysis of MRI studies. American Journal of Psychiatry 2004;161:1957e66. Vythilingam M, Heim C, Newport J, Miller AH, Anderson E, Bronen R, et al. Childhood trauma associated with smaller hippocampal volume in women with major depression. American Journal of Psychiatry 2002;159:2072e80. Vythilingam M, Vermetten E, Anderson GM, Luckenbaugh D, Anderson ER, Snow J, et al. Hippocampal volume, memory, and cortisol status in major depressive disorder: effects of treatment. Biological Psychiatry 2004;56:101e12.