Epilepsy & Behavior 50 (2015) 61–66

Contents lists available at ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

Validation of diagnostic tests for depressive disorder in drug-resistant mesial temporal lobe epilepsy Bianca de Lemos Zingano a,b,⁎, Ricardo Guarnieri a,b,c, Alexandre Paim Diaz a,b, Marcelo Liborio Schwarzbold a,b, Maria Alice Horta Bicalho a, Lucia Sukys Claudino b,d, Hans J. Markowitsch g, Peter Wolf d,h, Katia Lin b,d,f, Roger Walz b,e,f a

Centro de Epilepsia de Santa Catarina (CEPESC), Hospital Governador Celso Ramos (HGCR), Florianópolis, SC, Brazil Centro de Neurociências Aplicadas (CeNAp), Hospital Universitário (HU), Universidade Federal de Santa Catarina (UFSC), Florianópolis, SC, Brazil c Serviço de Psiquiatria, HU-UFSC, Florianópolis, SC, Brazil d Serviço de Neurologia, HU-UFSC, Florianópolis, SC, Brazil e Departamento de Cirurgia, HU-UFSC, Florianópolis, SC, Brazil f Departamento de Clínica Médica, HU-UFSC, Florianópolis, SC, Brazil g Physiological Psychology, University of Bielefeld, Bielefeld, Germany h Danish Epilepsy Center Filadelfia, Dianalund, Denmark b

a r t i c l e

i n f o

Article history: Received 9 February 2015 Revised 3 June 2015 Accepted 4 June 2015 Available online xxxx Keywords: Drug-resistant mesial temporal lobe epilepsy Hippocampal sclerosis Depressive disorder Diagnostic tests Diagnostic accuracy

a b s t r a c t Purpose: This study aimed to evaluate the diagnostic accuracy of the Hamilton Rating Scale for Depression (HRSD), the Beck Depression Inventory (BDI), the Hospital Anxiety and Depression Scale (HADS), and the Hospital Anxiety and Depression Scale-Depression subscale (HADS-D) as diagnostic tests for depressive disorder in drug-resistant mesial temporal lobe epilepsy with hippocampal sclerosis (MTLE-HS). Methods: One hundred three patients with drug-resistant MTLE-HS were enrolled. All patients underwent a neurological examination, interictal and ictal video-electroencephalogram (V-EEG) analyses, and magnetic resonance imaging (MRI). Psychiatric interviews were based on DSM-IV-TR criteria and ILAE Commission of Psychobiology classification as a gold standard; HRSD, BDI, HADS, and HADS-D were used as psychometric diagnostic tests, and receiver operating characteristic (ROC) curves were used to determine the optimal threshold scores. Results: For all the scales, the areas under the curve (AUCs) were approximately 0.8, and they were able to identify depression in this sample. A threshold of ≥9 on the HRSD and a threshold of ≥8 on the HADS-D showed a sensitivity of 70% and specificity of 80%. A threshold of ≥19 on the BDI and HADS-D total showed a sensitivity of 55% and a specificity of approximately 90%. The instruments showed a negative predictive value of approximately 87% and a positive predictive value of approximately 65% for the BDI and HADS total and approximately 60% for the HRSD and HADS-D. Conclusions: HRSD ≥ 9 and HADS-D ≥ 8 had the best balance between sensitivity (approximately 70%) and specificity (approximately 80%). However, with these thresholds, these diagnostic tests do not appear useful in identifying depressive disorder in this population with epilepsy, and their specificity (approximately 80%) and PPV (approximately 55%) were lower than those of the other scales. We believe that the BDI and HADS total are valid diagnostic tests for depressive disorder in patients with MTLE-HS, as both scales showed acceptable (though not high) specificity and PPV for this type of study. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Epilepsy is a chronic disease associated with high rates of disability and functional impairment [1]. Mesial temporal lobe epilepsy (MTLE)

⁎ Corresponding author at: Centro de Neurociências Aplicadas (CeNAp), Hospital Universitário (HU), Universidade Federal de Santa Catarina (UFSC), Florianópolis, SC, Brazil. Tel./fax: +55 48 99102242. E-mail address: [email protected] (B. de Lemos Zingano).

http://dx.doi.org/10.1016/j.yebeh.2015.06.004 1525-5050/© 2015 Elsevier Inc. All rights reserved.

is cited in a recent review [2] as an example of an aggressive epileptic condition that seems to have a relationship with increased seizure frequency and worsening cognitive function. In addition to disability and functional impairment, it is well known that there is a higher lifetime prevalence of mental disorders in people with epilepsy compared with the general population (23.5%) [3]; this prevalence reaches approximately 50% in patients with MTLE and drug-resistant epilepsy [4]. Depression is the most common psychiatric comorbidity found in patients with epilepsy [5]. It has been identified as a predictor of drug resistance [6], associated with a poorer result for epilepsy surgery [7,

62

B. de Lemos Zingano et al. / Epilepsy & Behavior 50 (2015) 61–66

8] and related to unsatisfactory quality of life [9,10]. Despite the growing recognition of the importance of mood and anxiety disorders to the morbidity and mortality of patients with epilepsy, such disorders remain underrecognized and undertreated [11,12]. The diagnosis of depressive disorder is primarily based on the Diagnostic and Statistical Manual of Mental Disorders-Revised 4th edition (DSM-IV-TR) [13] for major depressive disorder (MDD) and the International League Against Epilepsy (ILAE) Commission of Psychobiology classification criteria [14] for interictal dysphoric disorder (IDD), a psychiatric comorbidity specifically associated with epilepsy. The application of these diagnostic standards usually requires psychiatric expert consultation, but this care is not always available, especially in countries with limited access to medical care. Diagnostic questionnaires can provide a viable alternative approach, depending on their accuracy and reliability. Psychometric scales such as the Hospital Anxiety and Depression Scale (HADS) [15,16], the Hamilton Rating Scale for Depression (HRSD) [17–19], and the Beck Depression Inventory (BDI) [18,20,21] are widely used to screen for depression in clinical practice. Several studies investigating the available diagnostic tests for assessing depression comorbidity in epilepsy have shown that these tests have adequate performance and usefulness [22–24]. One study, however, found that sensitivity was low for most screening instruments, and the selection of instruments should consider the questions being addressed and the associated costs [25]. The purpose of the application of these scales in this patient population was to establish the presence of depressive disorder, which is known to be more prevalent than in the general population. Priority is given to highly specific tools that aimed to establish the presence of depressive disorder in a population of symptomatic but undiagnosed individuals. Because either sensitivity or specificity may be more relevant in different settings (i.e., clinical practice vs. research), it is possible to prioritize one over the other for a particular purpose. For example, the sensitivity would be more relevant in clinical practice, and the specificity would be more relevant in a research environment. In the case of epilepsy, these scales have been used for diagnosing depression in heterogeneous samples composed of various types of epilepsy and epileptic syndromes. However, it is possible that the sensitivity and specificity of these diagnostic tests for depressive disorder when used in patients with epilepsy vary according to particular genetic and environmental characteristics and the severity of particular types of epilepsy and epileptic syndromes. Mesial temporal lobe epilepsy is the most frequent form of drugresistant epilepsy that is treated surgically [26]. It seems to be more closely associated with depressive disorders and with poorly controlled seizures compared to other types of epilepsy [27]. Moreover, seizure focus locations that involve the limbic structures (as in the case of temporal lobe epilepsy) are more strongly associated with depressive disorders compared with types of epilepsy that do not involve the limbic structures [28]. The neocortical and limbic structures involved in MTLE are also part of the neuronal network involved in depression; consequently, the sensitivity and specificity of diagnostic tests for depressive disorder could vary in MTLE compared to neocortical temporal lobe epilepsy, extratemporal epilepsies, or epilepsies related to thalamocortical synchronization mechanisms. Consistent with this idea, the results from one study involving healthy individuals suggest that hippocampal volume reduction may be associated with a predisposition toward developing depression [29]. The results from another study indicate that neuroimaging changes in patients with MTLE seem to be magnified when they are associated with comorbid, untreated depression, suggesting a neuroanatomical overlap between the two diseases [30]. For these reasons, this study aimed to validate diagnostic tests for detecting depressive disorder in a sample of Brazilian patients who are candidates for surgical treatment of drug-resistant mesial temporal lobe epilepsy with hippocampal sclerosis (MTLE-HS).

2. Methods 2.1. Subjects One hundred three consecutive patients with drug-resistant MTLEHS were enrolled. All patients underwent a presurgical evaluation at the Centro de Epilepsia de Santa Catarina (CEPESC) between October 2008 and March 2013, as well as a complete medical history, seizure semiology, neurological and neuropsychological examination, psychiatric interview, interictal and ictal video-EEG analyses, and magnetic resonance imaging (MRI) of their brain, as previously described [31,32]. Inclusion criteria were age older than 18 years, diagnosis of epilepsy according to the ILAE criteria [33], focal slowing, and interictal spikes and sharp waves over the anterior, inferior, and mesial temporal regions on an interictal scalp electroencephalogram (EEG). Furthermore, the patients were also required to have hippocampal atrophy on T1 and an increased hippocampal signal on T2 MRI sequences consistent with MTLE-HS [31,32]. Refractoriness was defined as a failure to respond to adequate trials of at least 2 antiepileptic drugs with a period of at least 12 months without seizures [34]. We excluded patients with extrahippocampal lesions, focal motor or sensory abnormalities on physical examination, generalized or extratemporal interictal spikes, and marked cognitive impairment that could confound the psychiatric and MTLE diagnosis. Patients presenting with acute psychotic symptoms during evaluation were also excluded from the study. The sociodemographic features that were analyzed in the patients that met the inclusion criteria were sex, age, marital status, years of education, and occupation. The clinical data included epilepsy duration until the presurgical evaluation, monthly complex partial seizure frequency in the year before the psychiatric evaluation, the side of MTLE-HS, antiepileptic drug use (monotherapy or polytherapy) and current intake of antidepressants, defined as the use of antidepressants 30 days prior to the psychiatric interview. The local Research Ethics Committee approved the study, and informed consent was obtained from all patients. 2.2. Psychiatric evaluation The psychiatric interviews lasted approximately 120 min. During the interview, all patients were assessed alone first, and then with the caregiver, by a psychiatrist with experience in psychiatric comorbidities associated with epilepsy [7]. Axis I disorder diagnoses were classified according to a semistructured interview for the diagnostic algorithm of the DSM-IV-TR to determine the current presence of MDD. In this interview, the experienced psychiatrist actively sought to obtain data to complete the medical and psychiatric history of the patients and to perform the examination of their mental functions to investigate the diagnostic criteria established by the DSM-IV-TR. All patients were allocated to one of two groups: with or without depressive disorders. Moreover, we included in the former group the diagnosis of interictal dysphoric disorder (IDD), a psychiatric comorbidity specifically associated with epilepsy by the ILAE Commission of Psychobiology [14], in the depressive disorders group. Four psychiatric questionnaires were completed by each patient at the end of the interview. The patients completed the self-report scales (HADS, HADS-D, BDI), and finally, the same expert psychiatric interviewer administered the HRSD [15,19,21]. 2.3. Diagnostic tests The HADS was designed to measure psychological distress in nonpsychiatric inpatient populations [15]. It is widely used and consists of 14 multiple-choice items divided into anxiety and depression subscales. The items are rated on a 4-point Likert scale scored from 0 to 3, resulting in a final score ranging from 0 to 21. Bjelland et al. [35] found that a cutoff score of 8 + on the HADS-D had a sensitivity and specificity of approximately 0.80 in a general nonpsychiatric population, with a better

B. de Lemos Zingano et al. / Epilepsy & Behavior 50 (2015) 61–66

balance between sensitivity and specificity compared with the HADS as a screening tool. In this study, these results support the two-factor structure of the HADS. Accordingly, a recent meta-analysis for nonpsychiatric patient populations found that a cutoff score of 8 or more results in a sensitivity of 82% and a specificity of 74% for identifying major depressive disorder [36]. In patients with epilepsy, the same cutoff for the depression subscale was suggested, with similar sensitivity and better specificity (80.2%) [23]; for the HADS total, a threshold of ≥7 was recommended (sensitivity of 80.0% and specificity of 85.7%) [24]. The BDI is a self-rating questionnaire designed to assess the severity of current (past 2 weeks) depression symptoms [21]. It consists of 21 items. Each item is assigned a score ranging from 0 to 3 to indicate the degree of severity of the symptom, with the total score ranging from 0 to 63. A previous study using the Brazilian version of the BDI in a community sample suggested a score of 10/11 as the best cutoff point for identifying depression (sensitivity of 70%, specificity of 87%) [20]. In patients with epilepsy, the recommended cutoff was ≥17 (sensitivity of 88.6% and specificity of 91.2%) in a Brazilian study [23]. The HRSD is widely used to assess depression. It has the advantage of being applicable to semiliterate and seriously ill patients who are unable to complete self-rating scales [19]. The scale contains 17 questions measured either on a five-point (ranging from 0 to 4) or a three-point (ranging from 0 to 2) scale. The total score ranges from 0 to 50. In a validation study of the HRSD-17 in patients with epilepsy [22], a cutoff score of 6 was found to yield the best sensitivity (94%) and specificity (80%) threshold for detecting depression without subdividing it into degrees of severity. A cutoff score of 6 showed a positive predictive value of 46% and a negative predictive value of 99%. 2.4. Statistical analysis For the comparison of categorical variables, the chi-square or Fisher's exact test was used, if indicated. The Kolmogorov–Smirnov test was used to determine whether continuous variables could be assumed to be normally distributed. To compare continuous variables between groups, we used Student's t-distribution for normally distributed variables and the Mann–Whitney U test for nonparametric variables. Empirical receiver operating characteristic (ROC) curves were generated using the DSM-IV-TR diagnostic criteria for depression and the ILAE Commission of Psychobiology criteria as the gold standards and the HRSD, HADS, HADS-D, and BDI scores as predictors. Statistical analyses were performed using SPSS version 17.0 for Windows (SPSS, Inc., Chicago, IL, USA). Sensitivity, specificity, and positive and negative predictive values at possible cutoff points (coordinates of the curve) were calculated using MedCalc 12.7 for Windows (MedCalc Software, Ostend, Belgium) with 95% exact binomial confidence intervals. The predictive value is the probability that those who test positive have the disorders in question. The values are defined by sensitivity and specificity and by the condition's prevalence in the sample in which the test is applied [37]. The area under the ROC curve provides a measure of the overall performance of a diagnostic test. It allows the selection of optimal cutoff values for sensitivity and specificity and to assess the diagnostic validity of a test and compare the usefulness of two or more diagnostic tests. The coordinate nearest to the upper left corner and the Youden index, which is the maximum vertical distance between the ROC curve and the diagonal line, are the optimal cutoff points at which the sensitivity and specificity are maximized [38]. The optimal cutoff points are illustrated as gray and white circles in the curves. 3. Results One hundred three patients with drug-resistant MTLE-HS underwent a psychiatric evaluation. Table 1 describes the demographic, clinical, and psychiatric characteristics of the sample. The patients consisted of 57.3%

63

females; 53.4% were unmarried and 38.8% were unemployed. The mean age was 36.42 years, the median years of education was 8, the median frequency of seizures was 5 per month, and the mean epilepsy duration was 22.27 years. Among the participating patients, 48.6% had left lateralization on the MRI, and 52.5% of the sample was undergoing polytherapy with 2 antiepileptic drugs. One-fourth of the patients (25.2%) were diagnosed with depressive disorder, and approximately 46% were taking an antidepressant at the time of evaluation. Between the groups with and without depressive disorders, there were no significant differences with respect to duration of disease, seizure frequency, MRI side, and use of polytherapy (Table 1; p N 0.05); however, significant differences existed between the groups that were and were not currently taking antidepressants (p b 0.005). Table 2 lists the area under the curve (AUC) values (HRSD = 0.0774, BDI = 0.784, HADS total = 0.770, HADS-D = 0.792) referring to the curves in Fig. 1 and selected coordinates (HRSD ≥ 9, BDI ≥ 19, HADS total ≥ 19, HADS-D ≥ 8) and their respective sensitivity (HRSD = 73.7, BDI = 57.9, HADS total = 56.0, HADS-D = 72.0), specificity (HRSD = 77.2, BDI = 90.0, HADS total = 88.4, HADS-D = 79.7), positive predictive values (HRSD: PPV = 51.9, BDI: PPV = 64.7, HADS total: PPV = 63.6, HADS-D: PPV = 56.3), and negative predictive values (HRSD: NPV = 89.8, BDI: NPV = 87.1, HADS total: NPV = 84.7, HADS-D: NPV = 88.7) with the confidence intervals for the HRSD (0.664–0.862), BDI (0.677– 0.869), HADS total (0.672–0.851), and HADS–D (0.696–0.869). The diagnosis of depressive disorder was used as a gold standard, and the HRSD, BDI, HADS total, and HADS-D scores were predictors. A cutoff point of ≥ 9 on the HRSD and a cutoff point of ≥ 8 on the HADS-D showed a sensitivity of approximately 70% and a specificity of approximately 80%. A cutoff point of ≥ 19 on the BDI and HADS total showed a sensitivity of approximately 55% and a specificity of 90%. The BDI and HADS total showed a negative predictive value of approximately 87% and a positive predictive value of approximately 65%; the positive predictive value of the HRSD and the HADS-D was approximately 60%. All of the curves had an AUC significantly above 0.5 (p = 0.0001 for all comparisons), but they did not differ between each other (p = 0.05 for all comparisons; Fig. 1). 4. Discussion As expected, we found a high point prevalence (25.2%) of depressive disorders in our sample of patients with drug-resistant MTLE-HS undergoing treatment at a tertiary center. The clinical profile was similar to that of patients at other tertiary epilepsy centers [11,23,24]. The AUC is a good test to distinguish between patients with depression and those without depression according to all screening tests. Except for the HADS [24], our diagnostic tests indicated optimal cutoffs that are higher than those reported in other recent epilepsy studies [11,23]. A possible reason may be the homogeneity of the sample; all of our patients had seizures that were drug-resistant and had the same type of epilepsy (MTLE-HS). These are severe cases in which seizures are uncontrolled. According to Barry et al., because of the substantial functional impairment caused by epilepsy, patients with epilepsy believe that depressive symptoms are part of their epilepsy and a normal reaction to disease, and they “become accustomed to living in a chronic depressed state to the point where they end up forgetting to live in a euthymic mood” [12]. It is noteworthy that almost half of the patients diagnosed with depressive disorder were taking an antidepressant and were still symptomatic enough to be diagnosed with depression. This factor may be a reason for the higher scores needed to detect depression, as patients with epilepsy tend not to perceive depression as a different disorder. This issue would require that the tools used in the samples with epilepsy not only measure patients' perceptions but also interviewers' perceptions because the cutoff score was higher than expected for both scales, whether they were answered by the patients themselves or by others. However, the threshold scores of the scales were similar in terms of their capability to predict depression, even in

64

B. de Lemos Zingano et al. / Epilepsy & Behavior 50 (2015) 61–66

Table 1 Sociodemographic, clinical, and psychiatric characteristics of the sample. Variables

All patients n = 103 (%)

Depressive disordera Not n = 77 (74.8%)

Sociodemographic Sexb Male Female Age (years) Mean (±SD) Years of educationc Median (IR) Marital statusb Married Unmarried Labor statusb Employed Unemployed Retired because of the disease Discharged because of the disease Clinical Duration of disease, years Mean (± SD) Frequency of seizures/monthc Median (IR) RMI sideb Left Right Bilateral AEDsb Monotherapy Polytherapy (2 AEDs) Polytherapy (3 or more AEDs) Psychiatric Antidepressants Not current intake Current intake HDRS-17 Mean (±SD) HADS Mean (±SD) HADS-D Mean (±SD) HADS-A Mean (±SD) DI Mean (±SD)

44 (42.7) 59 (57.3)

39 (50.6) 38 (49.4)

36.42 (±10.72)

34.9 (±10.54)

8 (4/11)

8 (4/11)

48 (46.6) 55 (53.4)

30 (39) 47 (61)

16 (15.5) 40 (38.8) 11 (10.7) 36 (35.0)

p value Yes n = 26 (25.2%)

5 (19.2) 21 (80.8)

0.005

40.92 (±10.14)

0.013

6 (3.75/10)

0.142

18 (69.2) 8 (30.8)

0.070

13 (16.9) 32 (41.5) 4 (5.2) 28 (36.4)

3 (11.5) 8 (30.8) 7 (26.9) 8 (30.8)

0.021

22.27 (±10.60)

22.27 (±10.44)

25.54 (±10.86)

0.175

5 (3/8)

5 (3/8)

50 (48.6) 47 (45.6) 6 (5.8)

35 (45.5) 37 (48) 5 (6.5)

15 (57.7) 10 (38.5) 1 (3.8)

0.542

40 (38.8) 54 (52.5) 9 (8.7)

35 (45.5) 36 (46.7) 6 (7.8)

5 (19.2) 18 (69.3) 3 (15.5)

0.600

89 (86.4) 14 (13.6)

75 (97.4) 2 (2.6)

14 (53.85) 12 (46.15)

0.000

7.03 (±4.46)

5.75 (±4.00)

10.84 (±5.72)

0.000

14.09 (±6.94)

12.14 (±5.43)

19.44 (±7.91)

0.000

6.23 (±3.67)

5.16 (±2.95)

9.20 (±3.89)

0.000

7.74 (±3.92)

6.84 (±3.10)

10.24 (±4.85)

0.000

11.56 (±9.32)

9.08 (±7.4)

19.37 (±10.63)

0.000

6 (3/10.5)

0.700

Abbreviations: SD = standard deviation; IR = interquartile range; MRI = magnetic resonance image; AED = antiepileptic drugs. a According to DSM-IV-TR for MDD or to ILAE Commission of Psychobiology for IDD. b Chi-square test. c Significance level analyzed by Mann Whitney U test because the distributions violated assumptions for parametric testing.

a very specific sample of patients such as ours (namely, patients with drug-resistant MTLE-HS, a disease with a high prevalence of depression). Patients with epilepsy may manifest atypical features of depression, with a predominance of anxiety, fear, panic, and euphoric mood; this combination of features is known as interictal dysphoric disorder (IDD) and is related to hippocampal sclerosis, and these results may not be representative of the general population with epilepsy. It is necessary to perform test–retest validation of the HRSD, BDI, HADS, and HADS–D as diagnostic tests for depression in drug-resistant MTLE-HS to obtain a better reliability of these results. The optimal choice of tests will depend on the purpose for which they are applied. Because we aimed to validate diagnostic tests for detecting depressive disorder in patients with MTLE-HS, specificity was more relevant than sensitivity in terms of the accuracy of the test. The PPV value is important because it shows the applicability of the tests to patients with MTLE-HS in clinical settings, in which the suspected diagnosis and positive diagnostic tests for depressive disorder will guide treatment decisions. Table 2 shows that the BDI and HADS total have a specificity of approximately 90% and a PPV of approximately 65% at the threshold of

≥ 19. Therefore, they appear more useful than other diagnostic tests (the HRSD and HADS-D) for diagnosing depressive disorder in this sample. Hamilton Rating Scale for Depression ≥ 9 and HADS-D ≥ 8 had the best balance between sensitivity (approximately 70%) and specificity (approximately 80%); however, with these cutoffs, these diagnostic tests are not useful for identifying depressive disorder in this population with epilepsy because specificity (approximately 80%) and PPV (approximately 55%) are lower than those of the other scales. We believe that the BDI and HADS total are valid diagnostic tests for depressive disorder in patients with MTLE-HS because both have acceptable (though not high) specificity and PPV for this type of study. For example, a recent study showed ≥ 6 as the best cutoff for the HADS total, with a specificity of approximately 80% and a PPV of approximately 40% [22], whereas the present study found ≥19 as the optimal cutoff with a specificity of approximately 90% and a PPV of approximately 65%. Regarding the BDI and HADS-D, another study [21] showed that ≥17 was the best cutoff for BDI and ≥8 was the best for HADS-D, with BDI showing a specificity of approximately 90% and PPV showing a specificity of approximately 80%. In our study, we found different optimal thresholds. For the BDI, the optimal threshold was ≥ 19, with a

B. de Lemos Zingano et al. / Epilepsy & Behavior 50 (2015) 61–66

65

Table 2 Statistics and selected coordinates of the ROC curves generated using scores from HRSD, BDI and HADS as predictors and clinical diagnosis of depressive disorder as gold standard in patients with mesial temporal epilepsy. Scale and cutoff points

HRSD ≥6 ≥9a,b ≥12 BDI ≥8 ≥12 ≥19a,b HADS total ≥13 ≥15b ≥19a HADS-Depression subscale ≥5 ≥8a,b ≥9

Sensitivity (%)

Specificity (%)

PPV (%)

NPV (%)

AUC

Value (95% CI)

Value (95% CI)

Value (95% CI)

Value (95% CI)

Value (95% CI)

84.2 (60.4–96.6) 73.7 (48.8–90.9) 42.1 (20.3–66.5)

54.4 (40.7–67.6) 77.2 (64.2–87.3) 91.2 (80.7–97.1)

38.1 (23.6–54.4) 51.9 (31.9–71.3) 61.5 (31.6–86.1)

91.2 (76.3–98.1) 89.8 (77.8–96.6) 82.5 (70.8–91.0)

0.774 (0.664–0.862)

84.2 (60.4–96.6) 68.4 (43.4–87.4) 57.9 (33.5–79.7)

51.7 (38.4–64.8) 63.3 (49.9–75.4) 90.0 (79.5–96.2)

35.6 (21.9–51.2) 37.1 (21.5–55.1) 64.7 (38.3–85.8)

91.2 (76.3–98.1) 86.4 (72.6–94.8) 87.1 (76.1–94.3)

0.784 (0.677–0.869)

84.0 (63.9–95.5) 68.0 (46.5–85.1) 56.0 (34.9–75.6)

55.1 (42.6–67.1) 68.1 (55.8–78.8) 88.4 (78.4–94.9)

40.4 (27.0–54.9) 43.6 (27.8–60.4) 63.6 (40.7–82.8)

90.5 (77.4–97.3) 85.5 (73.3–93.5) 84.7 (74.3–92.1)

0.770 (0.672–0.851)

88.0 (68.8–97.5) 72.0 (50.6–87.9) 60.0 (38.7–78.9)

40.6 (28.9–53.1) 79.7 (68.3–88.4) 85.5 (75.0–92.8)

34.9 (23.3–48.0) 56.3 (37.7–73.6) 60.0 (38.2–79.2)

90.3 (74.2–98.0) 88.7 (78.1–95.3) 85.5 (75.0–92.8)

0.792 (0.696–0.869)

ROC = receiver operating characteristic; HRSD = 17-item Hamilton Rating Scale for Depression; BDI = Beck Depression Inventory; HADS = Hospital Anxiety and Depression Scale; CI = confidence interval; PPV = positive predictive value; NPV = negative predictive value; AUC = area under the curve. a Cutoff point indicated by the Youden index. b Cutoff point corresponding to the coordinates closest to the upper left corner.

specificity of approximately 90% similar to the previous study, but a lower PPV of approximately 65%. In relation to the cutoff for the HADS-D (≥8), the results of the study showed a specificity of

approximately 80% and a PPV of approximately 60%, similar to the results that we found with the same cutoff. For the HRSD, the study by Mula et al. [22] showed that N6 was the best cutoff, with a specificity

Fig. 1. Receiver operating characteristic (ROC) curves generated using depressive disorder diagnosed according to DSM-IV-TR for MDD or to ILAE Commission of Psychobiology for IDD as gold standard and scores on the 17-item Hamilton Rating Scale for Depression (HRSD), Beck Depression Inventory (BDI), Hospital Anxiety and Depression Scale (HADS), and Hospital Anxiety and Depression Scale-Depression subscale (HADS-D) as predictors in patients with drug-resistant mesial temporal lobe epilepsy related to hippocampal sclerosis. Curves and coordinates are represented by solid lines with blue dots. The gray circle corresponds to the optimal cutoff point indicated by the Youden index; the white circle corresponds to the optimal cutoff point indicated by the coordinates closest to the upper left corner; the overlapping gray and white circles corresponds to the optimal cutoff point indicated by both overlapping indices in each curve. Dashed lines represent the confidence interval of the coordinates. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

66

B. de Lemos Zingano et al. / Epilepsy & Behavior 50 (2015) 61–66

of approximately 80% and a PPV of approximately 45% [20]. We found a different optimal cutoff for the HRDS (≥9), but our specificity and PPV were similar to those of the previous study. The differences between these recent studies and ours may be explained by the fact that we exclusively assessed patients with MTLE-HS, whereas the other studies included patients with heterogeneous types of epilepsy. The HADS total was one of the two best diagnostic tests for depressive disorder in our patients. The other was the BDI, which demonstrated specificity and PPV similar to those found in previous studies but with different cutoff points. Our study addresses the fact that depression is one of the comorbidities that must be considered in neurological tertiary epilepsy centers, where the gold standard of diagnostic reliability (i.e., investigation by a psychiatrist with expertise in epilepsy or a neurologist trained in semistructured interviews based on the DSM-IV-TR) is not easily available. We evaluated the validity of available diagnostic tests for depressive disorders in patients with MTLE-HS so that neurologists could be trained to administer these tests, which require approximately 15 min to conduct. A limitation of the present study is that there is still no consensus on which test(s) is or are the most suitable for use in tertiary centers of epilepsy and whether they have the same validity for diagnosing depressive disorder in patients with other types of epilepsy. Another limitation is that the use of the HADS questionnaire is licensed by GL Assessment. A license agreement must be completed, and a user fee is required of all users (commercial & academic). Additionally, the HRSD must be administered by a health professional; in 1988, a standardized administration of this scale for health professionals was developed [39]. Therefore, we believe that more research on diagnostic tests for depressive disorders is needed in populations with specific types of epilepsy, such as patients with extratemporal epilepsy surgery or outpatients with good seizure control, because there is some evidence that the available tests may not be equally valid for various types of epilepsy. This evidence could be the beginning of a discussion based on research domain criteria (RDoC) [40,41] on various types of epilepsy and levels of therapeutic response. Acknowledgments This work was supported by the Program of Excellence Centers of Santa Catarina State of FAPESC (NENASC Project — TO 1262/2012-9). Peter Wolf (grant 88881.030478/2013-0) and Hans J. Markowitsch (grant 406929/2013-0) are Special Visiting Researchers supported by the program Science Without Borders of the Foundation for the Coordination of Improvement of Higher Education Personnel (CAPES) and the National Council of Scientific and Technological Development (CNPQ), Ministry of Science and Technology, Brazil. Disclosures The authors report no conflicts of interest. References [1] Dulac O. Epileptic encephalopathy. Epilepsia 2001;42(Suppl. 3):23–6. [2] Avanzini G, Depaulis A, Tassinari A, de Curtis M. Do seizures and epileptic activity worsen epilepsy and deteriorate cognitive function? Epilepsia 2013;54(Suppl. 8): 14–21. [3] Tellez-Zenteno JF, Patten SB, Jette N, Williams J, Wiebe S. Psychiatric comorbidity in epilepsy: a population-based analysis. Epilepsia 2007;48:2336–44. [4] Kanner AM. Depression in epilepsy: prevalence, clinical semiology, pathogenic mechanisms, and treatment. Biol Psychiatry 2003;54:388–98. [5] Gaitatzis A, Trimble MR, Sander JW. The psychiatric comorbidity of epilepsy. Acta Neurol Scand 2004;110:207–20. [6] Hitiris N, Mohanraj R, Norrie J, Sills GJ, Brodie MJ. Predictors of pharmacoresistant epilepsy. Epilepsy Res 2007;75:192–6. [7] Guarnieri R, Walz R, Hallak JE, Coimbra E, de Almeida E, Cescato MP, et al. Do psychiatric comorbidities predict postoperative seizure outcome in temporal lobe epilepsy surgery? Epilepsy Behav 2009;14:529–34. [8] Kanner AM, Byrne R, Chicharro A, Wuu J, Frey M. A lifetime psychiatric history predicts a worse seizure outcome following temporal lobectomy. Neurology 2009;72:793–9.

[9] Boylan LS, Flint LA, Labovitz DL, Jackson SC, Starner K, Devinsky O. Depression but not seizure frequency predicts quality of life in treatment-resistant epilepsy. Neurology 2004;62:258–61. [10] Pauli C, Thais ME, Claudino LS, Bicalho MA, Bastos AC, Guarnieri R, et al. Predictors of quality of life in patients with refractory mesial temporal lobe epilepsy. Epilepsy Behav 2012;25:208–13. [11] Jones JE, Hermann BP, Woodard JL, Barry JJ, Gilliam F, Kanner AM, et al. Screening for major depression in epilepsy with common self-report depression inventories. Epilepsia 2005;46:731–5. [12] Barry JJ, Ettinger AB, Friel P, Gilliam FG, Harden CL, Hermann B, et al. Consensus statement: the evaluation and treatment of people with epilepsy and affective disorders. Epilepsy Behav 2008;13(Suppl. 1):S1–S29. [13] American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Press; 2000. [14] Krishnamoorthy ES, Trimble MR, Blumer D. The classification of neuropsychiatric disorders in epilepsy: a proposal by the ILAE Commission on Psychobiology of Epilepsy. Epilepsy Behav 2007;10:349–53. [15] Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70. [16] Cosco TD, Doyle F, Ward M, McGee H. Latent structure of the Hospital Anxiety and Depression Scale: a 10-year systematic review. J Psychosom Res 2012;72:180–4. [17] Zimmerman M, Martinez JH, Young D, Chelminski I, Dalrymple K. Severity classification on the Hamilton Depression Rating Scale. J Affect Disord 2013;150:384–8. [18] Oliveira GNM, Araujo Filho GM, Kummer A, Salgado JV, Portela EJ, Sousa-Pereira SR, et al. Inventário de Depressão de Beck (BDI) e Escala de Avaliação de Depressão de Hamilton (HAM-D) em pacientes com epilepsia. J Bras Psiquiatr 2011;60:131–4. [19] Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23: 56–62. [20] Gomes-Oliveira MH, Gorenstein C, Neto FL, Andrade LH, Wang YP. Validation of the Brazilian Portuguese version of the Beck Depression Inventory-II in a community sample. Rev Bras Psiquiatr 2012;34:389–94. [21] Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4:561–71. [22] Mula M, Iudice A, La Neve A, Mazza M, Mazza S, Cantello R, et al. Validation of the Hamilton Rating Scale for Depression in adults with epilepsy. Epilepsy Behav 2014;41C:122–5. [23] de Oliveira GN, Lessa JM, Goncalves AP, Portela EJ, Sander JW, Teixeira AL. Screening for depression in people with epilepsy: comparative study among neurological disorders depression inventory for epilepsy (NDDI-E), hospital anxiety and depression scale depression subscale (HADS-D), and Beck depression inventory (BDI). Epilepsy Behav 2014;34:50–4. [24] Fiest KM, Patten SB, Wiebe S, Bulloch AG, Maxwell CJ, Jette N. Validating screening tools for depression in epilepsy. Epilepsia 2014;55:1642–50. [25] Hesdorffer DC, Baldin E, Caplan R, Berg AT. How do we measure psychiatric diagnoses? Implications of the choice of instruments in epilepsy. Epilepsy Behav 2014;31: 351–5. [26] Engel Jr J. Surgery for seizures. N Engl J Med 1996;334:647–52. [27] Kanner AM, Balabanov A. Depression and epilepsy: how closely related are they? Neurology 2002;58:S27–39. [28] Kanner AM. Mood disorder and epilepsy: a neurobiologic perspective of their relationship. Dialogues Clin Neurosci 2008;10:39–45. [29] Baare WF, Vinberg M, Knudsen GM, Paulson OB, Langkilde AR, Jernigan TL, et al. Hippocampal volume changes in healthy subjects at risk of unipolar depression. J Psychiatr Res 2010;44:655–62. [30] Salgado PC, Yasuda CL, Cendes F. Neuroimaging changes in mesial temporal lobe epilepsy are magnified in the presence of depression. Epilepsy Behav 2010;19:422–7. [31] Araujo D, Santos AC, Velasco TR, Wichert-Ana L, Terra-Bustamante VC, Alexandre Jr V, et al. Volumetric evidence of bilateral damage in unilateral mesial temporal lobe epilepsy. Epilepsia 2006;47:1354–9. [32] Nunes JC, Zakon DB, Claudino LS, Guarnieri R, Bastos A, Queiroz LP, et al. Hippocampal sclerosis and ipsilateral headache among mesial temporal lobe epilepsy patients. Seizure 2011;20:480–4. [33] Fisher RS, van Emde Boas W, Blume W, Elger C, Genton P, Lee P, et al. Epileptic seizures and epilepsy: definitions proposed by the International League Against Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE). Epilepsia 2005;46:470–2. [34] Kwan P, Arzimanoglou A, Berg AT, Brodie MJ, Allen Hauser W, Mathern G, et al. Definition of drug resistant epilepsy: consensus proposal by the ad hoc Task Force of the ILAE Commission on Therapeutic Strategies. Epilepsia 2010;51:1069–77. [35] Bjelland I, Dahl AA, Haug TT, Neckelmann D. The validity of the Hospital Anxiety and Depression Scale. An updated literature review. J Psychosom Res 2002;52:69–77. [36] Brennan C, Worrall-Davies A, McMillan D, Gilbody S, House A. The Hospital Anxiety and Depression Scale: a diagnostic meta-analysis of case-finding ability. J Psychosom Res 2010;69:371–8. [37] Akobeng AK. Understanding diagnostic tests 1: sensitivity, specificity and predictive values. Acta Paediatr 2007;96:338–41. [38] Akobeng AK. Understanding diagnostic tests 3: receiver operating characteristic curves. Acta Paediatr 2007;96:644–7. [39] Williams JB. A structured interview guide for the Hamilton Depression Rating Scale. Arch Gen Psychiatry 1988;45:742–7. [40] Insel T, Cuthbert B, Garvey M, Heinssen R, Pine DS, Quinn K, et al. Research domain criteria (RDoC): toward a new classification framework for research on mental disorders. Am J Psychiatry 2010;167:748–51. [41] Morris SE, Cuthbert BN. Research Domain Criteria: cognitive systems, neural circuits, and dimensions of behavior. Dialogues Clin Neurosci 2012;14:29–37.

Validation of diagnostic tests for depressive disorder in drug-resistant mesial temporal lobe epilepsy.

This study aimed to evaluate the diagnostic accuracy of the Hamilton Rating Scale for Depression (HRSD), the Beck Depression Inventory (BDI), the Hosp...
496KB Sizes 0 Downloads 13 Views