The Impact of Panic-Agoraphobic Comorbidity on Suicidality in Hospitalized Patients with Major Depression

Abstract

Background

Previous research in outpatient samples suggests that panic and agoraphobic comorbidity is related to suicidality in outpatients with major depression. The purpose of the study was to further investigate this relationship specifically in a hospitalized sample.

Method

The current study examined the severity of current suicidal ideation and behaviors in a psychiatric hospital sample diagnosed with major depressive disorder alone (MDD; n = 28) versus MDD plus panic-agoraphobic spectrum disorders (MDD+PAS; n = 69).

Results

Members of the MDD+PAS group were significantly more likely to have had a suicide attempt history, higher current depression severity, and higher current suicidal severity compared with individuals in the MDD alone group. The relationship between current suicidality and comorbid PAS remained significant after controlling for overall depression severity and other clinical factors.

Conclusions

These findings suggest that panic-agoraphobic comorbidity is associated with a greater risk for suicidality in hospitalized patients which cannot be adequately explained by the level of current depression alone. The clinical and research implications for these findings are discussed.

Individuals who have major depression and comorbid panic disorder or agoraphobia tend to have more severe symptomatology than individuals with depression alone. Previous research has reported that comorbid depression and panic/agoraphobia are associated with increased agitation, hypochondriasis, guilt, indecisiveness, hopelessness, feelings of inadequacy, social isolation, as well as a poorer treatment response and psychosocial outcome, and a greater number of hospitalizations.^[1–3] Overall depression severity has also been found to be more severe in comorbid depression and panic/agoraphobia.^{[4, 5]}

Several studies have also reported an important relationship among anxiety, depression, and suicidal ideation. Fawcett et al.^[6] and Fawcett^[7] found a strong correlation between suicide and the presence of panic attacks in mixed samples with affective and nonaffective disorders. Early suicides (within one year of assessment) were found to occur significantly more often in patients with panic disorder compared to those who did not have this comorbidity. However these studies included patients with bipolar disorder, schizoaffective disorder, and minor depression and therefore it is not possible to draw clear conclusions solely about individuals with unipolar depression and panic disorder. Norton, Temple, and Pettit^[8] found that individuals with panic disorder had more severe suicidal ideation than those without panic disorder even when depression severity was accounted for; however, this study was conducted by surveying college students in their classrooms which restricts the potential generalizability to a treatment seeking population, or an inpatient population more specifically. In an earlier study, Grunhaus et al.^[2] also found increased reports of suicidality in a comorbid panic and depression group compared to those with depression alone. The sample included both inpatients and outpatients; however, conclusions were limited by the fact that the authors did not control for depression severity. Sareen et al.^[9] found that panic disorder and agoraphobia were significantly related to a history of suicide attempts. This is an important finding as a history of past attempts has been found to be a predictor of future suicide attempts.^[10] They also found that people with comorbid anxiety and mood disorders were at a higher risk for suicide attempts than people with mood disorders alone. However, this study used an outpatient sample and combined bipolar patients within the mood disorder group.

Other research suggests that the effect of depression and panic/agoraphobic comorbidity may not be related to suicidality. Placidi et al.^[11] found that a lifetime history of suicide attempts in inpatients with major depression was unrelated to a history of panic disorder. In this study, anxiety symptoms such as agitation, psychic anxiety, and hypochondriasis were proposed to act as protective factors against suicidal behavior because individuals with these symptoms were thought to be more fearful of death. However, other research has suggested a link between psychomotor agitation and suicidal ideation, which contradicts the hypothesis that anxiety symptoms may protect against suicide.^{[12, 13]} Additionally, the Placidi et al.^[11] study also combined patients with major depression alone and bipolar disorder, and unfortunately cannot clearly determine the relationship between unipolar depression, panic, and suicidality.

The increased risk of suicidal ideation found in some previous studies is of particular concern for clinicians working with comorbid depression and panic/agoraphobia patients, especially when the disturbance has become severe enough to require hospitalization. Sanderson, Beck, and Beck^[14] found that about 42% of inpatients in psychiatric hospitals suffer with comorbid anxiety and depression, 12% suffer with comorbid panic and depression, and 6% suffer with comorbid panic and dysthymia. Because the number of individuals suffering with symptoms of depression and panic is high, the potential concern of an increased suicidality risk is particularly relevant. In order to ensure the safety of these patients both while they are hospitalized and following their discharge, it would be useful to further clarify their rates of suicidality. However, much previous research in this area (particularly recent studies) has focused on outpatients or college samples, and these subjects may not have the same symptom presentation as a hospitalized sample, which is likely to evidence greater suicidality and depression severity.

There may be factors associated with panic-agoraphobic comorbidity that increase the risk for suicidal ideation and behavior beyond the impact of depression alone. Norton et al.^[8] postulate that the increased risk for suicide in a comorbid anxious and depressed population may be due to distress about the anxiety, particularly as it affects functional, social, and occupational facets of life. Individuals with comorbid anxiety and depression may turn to suicide more often than depressed but nonanxious patients because it functions as an escape from the distress associated with their symptoms.^[9] Comorbid panic attacks during depression have been found to be associated with greater impairment in psychosocial adjustment over a 5 year follow-up period,^[15] which may shed light on the link with increased suicidal ideation.

Therefore, the current study examined suicidal ideation and behaviors in a sample of hospitalized patients to examine potential differences between patients with major depressive disorder only (MDD only) versus those with MDD and comorbid panic and agoraphobic disorders (MDD+PAS). First, we hypothesized that overall depression severity would be greater in the MDD+PAS group than the MDD only group. Second, we hypothesized that the MDD+PAS group would report greater suicidal ideation and have a greater likelihood of past suicide attempts than the MDD only group. Third, we hypothesized that this relationship would remain even after controlling for depression severity and other clinical differences.

Materials and Methods

Sample

This study was approved by the local Institutional Review Board (IRB). Participants (n=121) were recruited from a psychiatric inpatient or partial hospital to participate in a larger clinical trial investigating psychotherapy and pharmacotherapy for depression.^[16] All participants were diagnosed with Major Depressive Disorder according to the Structured Clinical Interview for DSM-III for Axis I (SCID-I),^[17] spoke English, were at least 18 years of age, and were living with at least one other family member who would be willing to participate in the study. Exclusion criteria included cognitive impairment or diagnoses of bipolar disorder, substance dependence, severe borderline personality disorder (characterized by chronic suicidality or a history of therapy-interfering behavior), somatization disorder, organic mental disorder, or schizophrenia. Participants also could not undergo electroconvulsive therapy (ECT) at the time of recruitment or while in the study or have a medical condition that would prevent the use of antidepressant medication. As part of the larger study, participants were randomized to receive pharmacotherapy alone, pharmacotherapy plus cognitive behavior therapy, or pharmacotherapy plus family therapy. For further information on study methods, see Miller et al..^[16]

Data analysis for the current study was only conducted on a subgroup of the larger sample. The first group included participants who had major depressive disorder and no comorbid current or past anxiety disorder (n=69). The second group included participants who had major depressive disorder and who were diagnosed with panic-agoraphobic spectrum disorders (n=28). This group included individuals who had current threshold or subthreshold (i.e., meeting all but one of the symptom criteria and showing evidence of functional impairment) panic disorder with agoraphobia (threshold n=10, subthreshold n=2), panic disorder without agoraphobia (threshold n=10, subthreshold n=1), or agoraphobia without panic disorder (threshold n=2, subthreshold n=3). We decided to include participants with subthreshold but clinically significant symptoms as previous research has indicated that these patients tend to have more severe depression and poorer outcomes.^{[18, 19]}

Measures

Psychiatric Diagnosis

The presence of depressive and anxiety disorders were assessed with the SCID-I.^[17] Presence of borderline personality disorder was assessed with the Structured Clinical Interview for DSM-II-R Personality Disorders (SCID-II).^[20]

Modified Scale for Suicidal Ideation (MSSI)

The MSSI^[21] is a modified version of the Scale for Suicidal Ideation Scale.^{[21, 22]} It is a 26-item structured interview scale that measures suicidal ideation for the most severe 48-hour period in the week prior to the interview. Each item includes a severity rating ranging from 0–3 and the total score is a sum of the core 18 items. Total scores ranging from 0–8 indicate no suicidal ideation or low severity, scores ranging from 9–20 indicate mild/moderate ideation, and a score of greater than 21 indicates severe suicidal ideation. The MSSI has demonstrated good interrater reliability and high convergent validity.^[21] Two subscales of the MSSI that were based on a previous factor analysis of the scale were also examined.^[23] The Suicidal Desire and Ideation subscale assesses desire for death, frequency of ideation, deterrents to suicide, expectation for suicide, and talk about suicide. The Resolved Plans and Preparations subscale assesses development of a suicidal plan, intensity of suicidal ideation, and the patient’s report of courage and competence to commit suicide, as well as if the patients had the means to complete suicide. These subscales have demonstrated good internal consistency.^[23]

Modified Hamilton Rating Scale for Depression (MHRSD)

The MHRSD^[24] is an adapted version of the 17-item Hamilton Rating Scale for Depression.^[25] It includes additional questions to improve the standardized administration of the interview. The MHRSD was used to assess overall depression severity for the week prior to hospitalization. Items on the MHRSD are rated on either a three-point or a five-point severity scale. The MHRSD has been found to have good interrater reliability and high convergent validity.^[24] For the current sample, the MHRSD total scores ranged from 13 to 36 with a score above 20 indicating severe depression.

Procedure

Assessors were trained to an interrater reliability >.80 and were continually monitored throughout the study to ensure continued reliability. As part of the larger study, participants were randomly assigned to six months of outpatient treatment following discharge from the hospital. In the current study, data were only analyzed from the baseline time-point. Informed consent was obtained for each participant while he/she was hospitalized according to the procedures approved by the IRB.

Results

Demographics

Preliminary analyses of demographic variables are presented in Table I. These variables were examined using two-tailed independent-samples t-tests or chi square analyses and alpha was set at .05. The only demographic variable that revealed a significant between-group difference was education (t(95)=2.430, p<.05), with the MDD only group achieving a higher level of educational attainment compared with the MDD-PAS group. No significant group differences were found for age (t(95)= −.93, p=.355), sex (χ²= 3.597, p=.058) race/ethnicity (white vs. nonwhite; Fisher’s Exact test, p=.432), or marital status (married/living together vs. single/living alone; χ²= 3.597, p=.970).

Table I.

Demographic Variables in Depressed Patients with versus without Panic-Agoraphobic Comorbidity

	MDD (n=69)	PAS (n=28)
	M(SD)	M(SD)
Age (y)	37.6(12.7)	40.1(10.6)
Education (years)	13.7(2.6)	12.2(2.9)
	%(n)	%(n)
Sex
Male	33.3(23.0)	14.3(4.0)
Female	66.7(46.0)	85.7(24.0)
Race/Ethnicity
American Indian/Alaskan	0.0(0.0)	0.0(0.0)
Asian	0.0(0.0)	0.0(0.0)
African American	7.2(5.0)	0.0(0.0)
Hispanic	2.9(2.0)	3.6(1.0)
White	89.9(62.0)	96.4(27.0)
Martial status
Single/living alone	24.6(17.0)	25.0(7.0)
Married/Living together	75.4(52.0)	75.0(21.0)

History of Illness

Table II displays severity of illness variables, including the duration of the current episode, total number of psychiatric hospitalizations (including current), age of onset of the first depressive episode, total number of depressive episodes (including current), and history of suicide attempts. The MDD+PAS group was more likely to have had multiple hospitalizations compared to the MDD only group (χ²= 3.970, p<.05). Participants in the MDD+PAS group were also more likely to report recurrent major depression compared to the MDD only group (χ²= 8.319, p<.01). Regarding current depression severity, MHRSD-17 scores were significantly higher in the MDD+PAS group compared to the MDD only group (t(95)= −2.018, p<.05). A chi square analysis was conducted to examine suicide attempt history. It revealed a significant difference between the MDD only and the MDD+PAS group with the comorbid group being more likely to have made a past attempt (χ²= 4.110, p<.05). Finally, a chi square analysis was performed to investigate differences for borderline personality disorder diagnosis, which could also explain suicide risk. However, the analysis revealed no significant difference (Fisher’s Exact Test, p=.408).

Table II.

Clinical Differences between Depressed Patients with versus without Panic-Agoraphobic Comorbidity

	MDD (n=69)	PAS (n=28)
	M(SD)	M(SD)
Age of onset of first MDE	29.2(13.4)	26.3(13.3)
MHRSD-17 Item Total	23.6(4.7)	25.7(4.7)
MSSI Total Score	17.1(13.6)	25.04(13.3)
MSSI-Suicidal Desire and Ideation Subscale	9.6(6.9)	14.5(6.7)
MSSI-Resolved Plans and Preparation	5.8(5.4)	7.0(4.7)
	%(n)	%(n)
Duration of current episode
2–4 weeks	44.9(31.0)	46.4(13.0)
5 or more weeks	55.1(38.0)	53.6(15.0)
Total number of hospitalizations
1 hospitalization	68.1(47.0)	46.4(13.0)
2 or more hospitalizations	31.9(22.0)	53.6(15.0)
Total number MDEs (including current)
1 episode	60.9(42.0)	28.6(8.0)
2 or more episodes	39.1(27.0)	71.4(20.0)
Borderline Personality Disorder Diagnosis
Borderline Diagnosis	5.8(4.0)	10.7(3.0)
No borderline diagnosis	94.2(65.0)	89.3(25.0)
History of Suicide Attempts
No suicide attempts	65.2(45.0)	42.9(12.0)
1 or more suicide attempts	34.8(24.0)	57.1(16.0)

Note. MHRSD = Modified Hamilton Rating Scale for Depression; MDE = Major Depressive Episode

Current Suicidality

T-tests also were conducted on the total MSSI score, and indicated that the MDD+PAS group had significantly higher levels of current suicidal ideation than the MDD only group (t(95)= −2.626, p=.01). Therefore, MSSI subscales scores were examined. A significant difference was found for the MSSI-Suicidal Desire and Ideation subscale, in which the MDD+PAS group had significantly higher scores than the MDD only group (t(94)= −3.029, p<.01). However, the MSSI-Resolved Plans and Preparation subscale was not significantly different (t(95)= −1.033, p=.30).

Follow-up analyses of covariance (ANCOVAs) were conducted on MSSI scores to control for the potential effects of other potentially confounding variables. We included the following variables as covariates: education level, presence of borderline personality disorder, and current depression severity according to the MHRSD (16-item total excluding the suicidal ideation item). When these three variables were covaried, the total MSSI score remained significantly higher in the MDD+PAS group (F=5.01, dfs = 1,92, p<.05). Additionally, the MSSI-Suicidal Desire and Ideation subscale also remained significantly higher in the MDD+PAS group compared to the MDD group (F=7.24, dfs = 1,91, p<.01).

Discussion

The results of this study provide additional evidence that hospitalized patients with comorbid depression and panic-agoraphobic comorbidity tend to exhibit more severe suicidal ideation compared to individuals with depression alone. Our results were not consistent with the Placidi et. al^[11] study, which found no relationship between suicide attempts and comorbid panic and affective disorders using a more heterogeneous sample. In the current study, not only did the MDD+PAS group have a greater likelihood of a past suicide attempt compared to the MDD only group, but they also demonstrated increased suicidal ideation at the time of hospitalization compared to the MDD only group. Furthermore, the relationship between MDD+PAS and current suicidality remained even after controlling for current depression severity, presence of a borderline diagnosis, and education level. This suggests that the relationship between MDD+PAS and suicide is not solely due to increased depression severity. These results indicate that comorbid panic and agoraphobia more specifically relate to increased risk for suicidal ideation and behavior in this population.

Researchers such as Norton et al.^[8] and Sareen et al.^[9] found comparable results in studies with non-hospitalized samples. Norton et al.^[8] reported increased suicidal ideation in a non-treatment seeking group of college students with depression and panic even when depression severity was covaried. Sareen et al.^[9] also found that panic disorder comorbid with depression was related to an increased history of past suicide attempts in a community sample. Additionally, Grunhaus et al.^[2] found similar results in a sample of both inpatients and outpatients. Placidi et al.^[11] found that there was no relationship between comorbid panic and depression; however, this study included bipolar patients in the sample as well which limits the conclusions that can be drawn for patients with unipolar depression. In contrast, the current study is the only one to our knowledge to examine unipolar depression, panic/agoraphobia, and suicidality in a hospitalized sample that controls for depression severity.

The results of this and other studies indicate that there may be features of panic-agoraphobic comorbidity that increase the risk for suicidal ideation beyond the effects of depression alone. The higher chronicity of depression in our comorbid sample may have been related to these patients having more severe suicidal ideation at hospital admission. Another explanation previously offered by Norton et. al^[8] attributes the increased suicide risk to the distress about anxiety that comorbid patients experience. Additionally, the higher rate of suicidality could have been due to the desire to escape from this distress, as hypothesized by Sareen et al.^[9] There are a number of explanations that could account for the increased rate of suicidality that are beyond the scope of the current investigation but which could be explored in future research.

There are limitations to our study that require consideration when interpreting the results. The sample was only modest in size and consisted of treatment-seeking individuals who were mainly Caucasian females. Additionally, inclusion criteria required that participants have a family member who would be willing to participate in the program. Thus, our results may not generalize to all hospitalized patients. Furthermore, our sample size was not large enough to investigate whether there were differences among subgroups of patients, such as those with panic disorder with versus without agoraphobia. Despite these limitations, the results of the study were consistent with results from other studies using a number of different samples, lending support to our findings.

This study has important implications for clinicians who are working with comorbid depression and panic-agoraphobia in a hospital setting. Considering the large number of individuals who are hospitalized with suicidal ideation, the fact that this comorbidity pattern appears to signify an even greater risk for suicidal ideation than depression alone should alert clinicians to the need to pay even greater clinical attention to these patients. It appears to be particularly important to comprehensively screen for suicidality and panic-agoraphobic symptoms at hospital intake and discharge when treating depressed patients. Additionally, a greater portion of the treatment while these individuals are in the hospital may need to focus on suicide prevention and management of panic and agoraphobic symptoms following discharge from the hospital.

Future research should investigate treatment methods that not only teach coping skills for the panic-agoraphobic and depression symptoms, but additionally provide methods for decreasing suicidal ideation and preventing future attempts. Since researchers have determined an important link between panic-agoraphobia and depression, it would appear useful to design tailored treatments to directly address the factors that impact suicide risk in this population. Our study found an increased likelihood of past suicide attempts in this population, which is particularly concerning due to previous research showing that a history of a suicide attempt increases the risk of future attempts.^[10] Thus, more work must be done to investigate and improve psychosocial suicide interventions in this population.