DOI: 10.31038/AWHC.2020333

Abstract

Objective: Video electroencephalography monitoring (VEM) is an important tool for the diagnosis and classification of seizures and for the presurgical evaluation of patients with drug-resistant epilepsy. The aim of this study was to assess the utility of VEM in patients referred for differential diagnoses (epileptic events versus nonepileptic episodes) and/or improving diagnosis accuracy in epilepsy.

Methods: Three hundred and eight VEM studies were analyzed retrospectively over a period of 3 years. Only studies obtaining seizure classification and diagnostic clarification to determine the nature of paroxysmal events were included (n = 125). The clinical diagnoses before and after VEM were compared. VEM was useful if it changed the diagnosis and/or therapy or if it answered the clinical question raised by the referring physician.

Results: One hundred twenty-five patients were included (64% women) with a mean age of 43.0 ± 1.6. During VEM, 61 patients had typical clinical events; there were 21 seizures, 25 physiological events, and 18 psychogenic nonepileptic seizures (PNES). In the PNES group, we found that women and younger patients were more frequent. Epileptic patients had a shorter evolution time, and the physiological events group had an older age at event onset compared to the epilepsy group. The provisional diagnosis changed in 35% of the cases after VEM. The diagnostic usefulness of VEM was 89%. After VEM, treatment changed in 50% of patients.

Significance: VEM is an essential tool to differentiate seizure from nonepileptic paroxysmal events. It is imperative to achieve an accurate diagnosis to determine the most suitable therapeutic approach.

Key Point Box

• Misdiagnosis of nonepileptic paroxysmal events and epilepsy represents a problem with important therapeutic and social repercussions.

• Inpatient VEM has been demonstrated to be a useful tool for the diagnosis and classification of seizure events.

• VEM achieved an accurate diagnosis and accomplished the most suitable therapeutic approach in epileptic and nonepileptic patients.

Keywords

epileptic seizures, nonepileptic paroxysmal events, physiological events, psychogenic nonepileptic seizure, syncope

Introduction

An Epileptic Seizure (ES) is defined as a transient occurrence of behavioral alterations produced by abnormal, excessive, and hypersynchronous neuronal activity in the brain [1, 2]. However, since the symptoms are diverse, diagnosis of ES may be challenging given the differential diagnoses [3, 4]. Nonepileptic paroxysmal events of physiological and psychological origin, such as syncope, sleep disorders, migraines or psychogenic nonepileptic seizures (PNES), can also manifest as behavioral disturbance events or transient alterations of consciousness [3]. It is extremely important to achieve an accurate diagnosis in epilepsy, given its morbidity associated with undiagnosed and untreated seizures [2]. In the same way, misdiagnosed epilepsy can result in side effects from antiepileptic drugs, economic costs, and impact on quality of life.

Video electroencephalography monitoring (VEM) is the most useful diagnostic tool for the classification of ES, as well as being the current gold standard for distinguishing epileptic versus nonepileptic paroxysmal events [5]. The Internal League Against Epilepsy (ILAE) recommends VEM for i) differential diagnosis for epileptic seizures, ii) characterization and classification of seizures types and epilepsy syndrome, iii) quantifying seizures, iv) intensive care unit monitoring, and v) presurgical evaluation of drug-resistant epilepsy [6].

Long-term VEM (1–6 days) has been shown to improve diagnostic accuracy compared with standard EEG (20–30 minutes) [7]. VEM not only obtains more complete information regarding the EEG background and characterization of the interictal activity but also analyzes the clinical semiology with the electrophysiological phenomenology during clinical events [8]. Moreover, during preoperative evaluation of drug-resistant temporal lobe epilepsy patients, it has been shown that no other routine tests, including imaging studies, were as reliable as VEM in identifying and characterizing epilepsy seizures and defining the epileptogenic zone in patients evaluated for epilepsy surgery [9]. Nonetheless, VEM is an expensive tool that needs sophisticated equipment, highly qualified staff, and admission of patients to the hospital during variable periods.

A highly variable range of diagnostic usefulness for VEM has been described (19%–75%), which depends first on how utility is defined and on the selection of the patients evaluated [10]. However, another factor is also involved in this issue. There is no standard protocol for the duration of VEM. In fact, some units carry out 12-hour studies, while in other units, monitoring lasts several days.

In our unit, we systematically used two protocols of VEM: 24-hour VEM for the differential diagnosis and follow-up of epilepsy patients and a longer-lasting VEM (2–10 days) for the presurgical evaluation of drug-resistant epilepsy.

The purpose of this study was to assess the diagnostic utility of VEM for the classification and differential diagnosis of epilepsy and nonepileptic paroxysmal events in a national reference unit for refractory epilepsy. We define a VEM study as useful when either a tentative diagnosis was changed or confirmed or when patient management was modified as a result of the information obtained from VEM.

Methods

Patients

We retrospectively analyzed the clinical chart and VEM records of consecutive patients who underwent inpatient VEM at the video electroencephalography (VEEG) unit of the National Reference Unit for Refractory Epilepsy at Hospital Universitario de la Princesa, over a period of three years (n = 308). Only those patients who had been referred to i) differentiate between epileptic and nonepileptic events and/or ii) to classify the kind of seizure and epilepsy syndrome type were selected. Finally, the number of patients fulfilling these conditions was 125. Those patients with known medically refractory epilepsy undergoing presurgical evaluation were excluded.

Clinical charts, including age, sex, age at symptomatology onset, duration, provisional clinical diagnosis, antiepileptic drugs (AED), brain magnetic resonance imaging (MRI), and ambulatory standard EEG, were revised. In cases where patients had undergone an ambulatory EEG and neuroimaging at institutions other than our hospital, only the reports were available, and the studies themselves were not reviewed.

The provisional diagnosis of the physician who referred the patient to the VEEG unit (pre-VEM diagnosis) was compared with the final clinical diagnostic (post-VEM diagnosis), and both diagnoses were classified into the following categories: i) epilepsy, ii) PNES, and iii) nonepileptic paroxysmal events of physiological origin, including a cardiogenic or metabolic cause or event-related to other neurological diseases (e.g., sleep disorders, movement disorders, migraine, or cognitive disturbance).

Video electroencephalography monitoring

VEM was performed using a 64-channel digital VEEG system (EMU64, NeuroWorks. XLTEK^®, Oakville, Canada) with 19 scalp stainless steel electrodes fixed with collodion according to the 10–20 international system; electrocardiography (ECG) and simultaneous video images were recorded continuously for 24 h. If needed, one or two electromyography channels (EMG) were added, too. Recordings were performed at a 512 Hz sampling rate with a 0.5–70 Hz bandwidth, 50 Hz notch on. EMG bandwidth was 1.5–200 Hz, notch on, and ECG bandwidth was 1.5–30 Hz, notch on. Impedances for EEG were under 25 kW.

Patients had partial sleep deprivation, but medication withdrawal was not undertaken. If considered, induction techniques involving suggestion and administration of placebo were used in some patients [4].

To avoid biases in the assessment of VEM utility, we considered three periods chronologically: i) t₁defining the putative diagnosis and treatment considered as the basal line usually performed 1–2 months before, ii) VEM (performed at t_VEM), and finally, iii) the period t₂that includes the first clinical interview after VEM (usually 1–2 months later). No other complementary studies were undertaken between t₁ and t₂; thus, we were sure that changes either in diagnosis or treatment would be due to the VEM result.

The patient’s function state (p) is defined at time t, as a two-variables function, i.e., treatment (T) and diagnosis (d), stated as p_t(T,d). In this definition, t is not a variable but a parameter. Therefore, we have different possibilities of changes at consecutive periods (Figure 1), depending on either T or d or both changing between t₁ and t₂. VEM was considered useful when variables T, d or both changed. However, in some cases where both T and d remained the same, the study can still be considered useful if it confirms a previously suspected but not well-established diagnosis, e.g., suspected PNES with no pharmacological treatment where, after VEM, a positive result confirms PNES. In such cases, the utility derived from confirmation is written as p₁(T₁,d₁) → p_2,conf (T₁,d₁). In other words, only when the VEM result did not change any variable or confirm a suspected previous diagnosis (p_2,not (T₁,d₁)) was it considered to be not useful.

Figure 1. Scheme used to evaluate the utility of VEM.

Additional, clinical demographics and VEM bioelectrical features for each of the three clinical diagnostic categories were also analyzed.

Statistical analysis

Statistical comparisons between groups were performed using Student’s t-test or ANOVA for data with normal distribution. Normality was evaluated using the Kolmogorov–Smirnov test. The Mann–Whitney rank sum test or ANOVA on ranks was used when normality failed. In the last case, Dunn’s method was used for all pairwise post hoc comparisons of mean ranks of treatment groups. Chi-square test (χ²) was used to assess the differences between groups of patients.

The SigmaStat 3.5 software (SigmaStat, Point Richmond, CA, USA) was used for statistical analysis. The significance level was set at p < 0.05. The results are presented as the mean ± SEM, except where otherwise indicated.

Results

Patients

A total of 125 patients were included (64% women) with a mean age of 43.0 ± 1.6 years. The mean disease duration was 8.8 ± 1.0 years. A total of 82 (66%) patients received treatment with AED at the time of the study, and the mean number of AED was 1.3 ± 0.1. Brain magnetic resonance imaging (MRI) was available in 113 patients, with abnormal findings in 57 (50%) patients; of these, only 8 patients (14%) had epileptogenic lesions. Standard EEG was performed in 92 patients, 37% of which displayed epileptiform activity, 24% was reported as abnormal nonepileptic activity, and 39% was normal.

Regarding the provisional clinical diagnosis, ES was suspected in 67 patients (54%). Nonepileptic events of physiological origin were suspected in 45 patients, accounting for syncope (30 patients), movement disorders (7 patients), cognitive impairment (3 patients), sleep disorders (1 patient), and others (4 patient). A diagnosis of PNES was suspected in 10 patients, and in three of them, it was considered that it could also coincide with the diagnosis of epilepsy.

Video electroencephalography monitoring

During VEM, typical events, described as similar to accustomed, occurred in 61 patients (49%). Of these, 21 patients (36%) had ES, 25 patients (41%) hadnonepileptic paroxysmal events of physiological origin and in 13 cases (21%), the diagnosis was PNES. One of these latter patients had ES along with PNES.

Regarding patients who did not have typical events (n = 65), we found epileptiform activity in 36 patients (37.5%) and nonepileptiform abnormalities in 20 patients (21.5%). Table 1 shows the electroencephalographic findings in all patients.

The three diagnostic categories, before and after VEM, are shown in Table 2. Patients with the double diagnosis of epilepsy and PNES have been included. As we can observe from this table, the diagnosis of epilepsy was confirmed in 41% of patients, 24% less than the initial presumptive diagnosis. Additionally, the diagnosis of PNES increased by 80% after VEM. In five patients, diagnosis of both epilepsy and PNES was made.

We compared the demographic characteristics between patients with final clinical diagnosis of epilepsy, events of physiological origin, and PNES (Table 3). We found that in the PNES group, women were more frequent in comparison with the other two groups. Additionally, this group had the lowest average age compared to nonepileptic events of physiological origin group. This last group had the highest mean age of onset of symptoms compared to the other two groups. It also had the shorter evolution time compared with the epilepsy group.

On the other hand, epileptiform activity on standard EEG was associated with the occurrence of epilepsy diagnosis compared with physiological events and PNES. Moreover, patients with epilepsy (59.6%) were more likely to have an abnormal brain MRI scan.

We assessed the overall structure for all the three groups according to sex proportion, age, age at onset, and abnormal findings in EEG. Paired comparison by groups yielded highly significant differences for ES vs. physiological events (χ²= 40.65, p < 0.001), ES vs. PNES (χ²= 37.30, p < 0.001), and physiological events vs. PNES (χ²= 38.17, p < 0.001), with 5 degrees of freedom.

Utility

We considered that VEM was useful for the clinician when the results of the study led to a change in the previous diagnosis, in treatment or both or when a suspected but not well-defined diagnosis was reinforced by the study.

According to our classification, VEM was defined as useful in 112 patients (89.6%). Specifically, treatment was changed (p₂(T₂,d₁)) in 32 cases, diagnosis was modified in 10 patients (p₂(T₁,d₂)), both of them were modified in 35 patients (p₂(T₂,d₂)), and the suspected diagnosis was confirmed in 35 patients (p_2,con(T₁,d₁)). Therefore, after VEM, the treatment was modified in 67 patients (p₂(T₂,d₁) + p₂(T₂,d₂)) and diagnosis changed in 45 patients (p₂(T₁,d₂) + p₂(T₂,d₂)).

In 10.4% of cases, the VEM results were inconclusive and could not be termed as useful.

Regarding the group of patients with a final diagnosis of epilepsy (n = 51), 44 of the patients (86.3%) were on at least one AED at the time of admission. After VEM, change in therapy was reported in 33 patients (64.7%) of the total group, among whom a new AED was introduced or the dose was increased in 26 patients (78%) and the dose was reduced in 7 patients (21%).

With respect to the group of patients with physiological events (n = 50), 21 patients (42%) had been on at least one AED at the time of the study. After VEM, the treatment was modified in 18 patients (36%) with a discontinued or decreased treatment in 15 patients (71%). There were 12 patients diagnosed with PNES who had previous AED treatment. Discontinuation of therapy was seen in 3 patients.

Discussion

We have shown that VEM is an extremely helpful tool in the diagnosis and therapeutic management of patients with paroxysmal behavioral events. Misdiagnosis and misclassification of nonepileptic paroxysmal events and ES can lead to inappropriate treatment. High costs have been incurred annually on diagnostic evaluations, inappropriate antiepileptic drugs, and emergency unit utilization [11].

In many cases, clinical information alone can be incomplete or misleading due to descriptions made by untrained witnesses. Furthermore, the correct diagnosis may not be apparent during the short period of outpatient EEG. VEM helps to correlate electroencephalographic changes with clinical events and detect epileptiform activity in long-term records, which also include a sleep period [2, 12].

VEM provided a useful yield of recorded clinical events. We have found that the 24-h VEM is able to detect typical clinical events in a 49% of cases. ES represents 36% of the cases, which is comparable to other series [13]. Most of the recorded events corresponded to nonepileptic paroxysmal events (62%), although one-third (21%) were of psychogenic origin. Thus, an accurate correlation between electroclinical findings is essential to properly characterize different paroxysmal events [10, 14]. This finding was revealed in a meta-analysis of 135 published studies on VEEG, which describe that 59% of referrals were for diagnostic reasons [15].

In this sense, PNES has a special importance. VEM is an indispensable tool because it allows simultaneous analysis of both clinical and ictal EEG findings to make the most accurate differential diagnosis between seizure and PNES [4, 16]. The results of VEM between patients with epilepsy and PNES revealed a sustained decline in AED use from discharge to follow-up [17], suggesting that VEM may contribute to a beneficial elimination of unnecessary medications in the PNES group once a definitive diagnosis is made.

It has been previously shown that in those patients with a change in diagnosis, the most common change involves distinguishing epilepsy from physiological events (68.2%) [18]. In our work, we found that VEM reduced the previous diagnosis of epilepsy in 24% of the cases. The reference physicians are more likely to misdiagnose nonepileptic events as seizures than the opposite. This finding may be due to a diagnosis that was made based on clinical history and a routine EEG that can be deceptive. Physiological paroxysmal events were more frequently misdiagnosed as PNES.

A total of 35% of patients saw their previous diagnosis change. Other authors [19] described a change in the diagnosis in 58% of patients. The higher figure could possibly be observed because these researchers’ studies were longer (1–13 days) and included patients with refractory epilepsy and who were gradually tapering AED. However, recent studies indicate that the VEEG clarifies diagnosis in 56.3% of patients and changes the diagnosis in 35.6% of patients [15, 20]. We have shown that VEM was useful in establishing or shaping the diagnosis in 112 patients (89%). This technique helped to confirm the reference diagnosis with certainty, classify patients with ES, and select the best treatment according to each diagnosis of either epilepsy or nonepileptic events. The high diagnostic yield of VEM in adult patients with recurrent and unprovoked events has been confirmed previously [21-23]; however, the diagnostic usefulness is widely variable (19%–75%) due to a variation in the definition of utility [10].

We found in the PNES group that women were more frequent in comparison to the other two groups. They were also younger at the time of the study and had a younger age of onset of events compared to those experiencing physiological events. We found a longer disease duration in epilepsy patients, as has been previously described [24], showing a delay in diagnostic confirmation. On the other hand, physiological event patients were older at the onset of symptoms compared to the PNES group, which is probably related to the etiology [25]. It is quite interesting to observe that all three groups are really different. This fact can help to establish a predictive model based on electroclinical findings to help the clinician to classify a patient in daily practice.

We had a final diagnosis of event of physiological origin in 40% of our patients and of PNES in 14% of the cases (see table 3). It is important to highlight that in both these groups, we found epileptiform activity in 20% and 28% of cases, respectively, and in both, ≥30% of abnormal nonepileptic activity. There are some studies describing between 17% and 26% of patients eventually being diagnosed with nonepileptic events who had interictalepileptiform discharges (IEDs) recorded during VEEG [26, 27]. Similar findings were caused by overreading a standard EEG as abnormal [27-30]. In a study that analyzed the significance of epileptiform abnormalities in patients without epilepsy, the researchers found that of 521 patients with a follow-up of 230 person-years with no history of unprovoked seizures, 64 (12.3%) had IEDs on their EEG. These patients had associated structural neurological conditions (e.g., tumors and vascular disorders), which would explain the 20% of epileptiform abnormalities found in our patients with physiological events.

VEM also helped to determine the best treatment for the individual patient based on the type of witnessed events and the electrographic characteristics in VEM. This finding caused a treatment change in 50% of patients. Moreover, the largest change was seen in the group of patients diagnosed with epilepsy. This finding shows the value of VEM when it comes to influencing the overall care pathway of patients. Optimization of AED may result in avoiding drug adverse events, a better quality of life, and reduction in health costs [31]. In our results, the discontinuation of AED treatment in PNES was low for what we might have expected. The reasons can be different, but it is essential to mention that some clinical physicians often analyze VEM results in an imprecise way for the diagnosis of PNES [32], even though VEM is the gold standard for diagnosis of PNES [4].

Despite being considered an expensive technique with limited availability (a neurophysiologist is needed with healthcare staff and specialized technical equipment) [33], VEM has been demonstrated to be a useful test with robust therapeutic benefits. There are recent recommendations and algorithms based on high-level evidence for the use of VEM for the diagnosis and monitoring of patients with epilepsy [34]. Conversely, the financial and social cost of unclassifiable behavioral disturbances to the patient and the family is considerable, and poorly controlled ES has been associated with impaired psychosocial skills and an increased risk of death [35]. Therefore, in the absence of a study that compares the cost–benefit of VEM and the economic and social cost of chronic uncontrolled seizures (epileptic or nonepileptic), it seems logical to consider that VEM should be a mandatory tool in the differential diagnosis of ES.

Acknowledgement

This work was financed by a grant from the Ministerio de Sanidad FIS PI17/02193 and was partially supported by FEDER (FondsEuropeen de DeveloppementEconomiqueet Regional).

References

1. Tatum WO, Rubboli G, Kaplan PW,Mirsatari SM, Radhakrishnan K et al. (2018) Clinical utility of EEG in diagnosing an monitoring epilepsy in adults. ClinNeurophysiol 129:1056-1082. [crossref]
2. Fisher RS, Acevedo C, Arzimanoglou A, Bogacz A, Cross JH et al. (2014) ILAE official report: a practical clinical definition of epilepsy. Epilepsia 55:475-82. [crossref]
3. Krumholz A (1999) Nonepileptic seizures: diagnosis and management. Neurology 53: 76-83.[crossref]
4. Vega-Zelaya L, Alvarez M, Ezquiaga E, et al. (2014) Psychogenic Non-Epileptic Seizures in a Surgical Epilepsy Unit: Experience and a Comprehensive Review. Epilepsy Topics. Prof. Mark Holmes (Ed.)InTech.
5. Cuthill FM, Espie CA (2005) Sensitivity and specificity of procedures for the differential diagnosis of epileptic and non-epileptic seizures: a systematic review. Seizure14:293-303.[crossref]
6. Velis D, Plouin P, Gotman J, et al. (2007) Recommendations regarding the requirements and applications for long-term recordings in epilepsy. Epilepsia48:379-384.[crossref]
7. Park KI, Lee SK, Chu K, Jung Ju Lee, Dong Wook Kimet al. (2009) The value of video-EEG monitoring to diagnose juvenile myoclonic epilepsy. Seizure18:94-99.[crossref]
8. Delgado Nunes V, Sawyer L, Neilson J, GrammatiSarri, J Helen Crosset al. (2012) Diagnosis and management of the epilepsies in adults and children: summary of updated NICE guidelines. Br Med J 344: 281.[crossref]
9. Pastor J, Ortega GJ, Herrera-Peco I,Marta Navas-García, Eduardo G Navarrete et al. (2010) Differential contribution of preoperatory studies to diagnosis in temporal lobe epilepsy surgery. Rev Neurol51:393-402.[crossref]
10. Alving J, Beniczky S (2009) Diagnostic usefulness and duration of the inpatient long-term video-EEG monitoring: findings in patients extensively investigated before the monitoring. Seizure18:470-473.[crossref]
11. Koblar, SA, Black, AB and Schapel, GJ (1992) Video-audio/EEG monitoring in epilepsy—The Queen Elizabeth Hospital Experience. Clinical and Experimental Neurology29:70-73. [crossref]
12. Pastor J (2011) Vídeo-electroencefalografía de scalp. En: Neurocirugía Funcional y Estereotáxica. Eds: E.G. Navarrete y R. G. Sola. Viguera Editores Barcelona; pp: 109-120.
13. McGonigal A, Russell AJC, Mallik AK, M Oto, R Duncanet al. (2004) Use of short term video EEG in the diagnosis of attack disorders.  J NeurolNeurosurg Psychiatry75:771-772.[crossref]
14. Lawley A, Manfredonia F, Cavanna AE (2016) Video-ambulatory EEG in a secondary care center: A retrospective evaluation of utility in the diagnosis of epileptic and nonepileptic seizures. Epilepsy & Behavior57: 137-140.[crossref]
15. Sauro KM, Wiebe N, Macrodimitris S, Wiebe S, Lukmanji S et al. (2016) Quality and safety in adult epilepsy monitoring units: A systematic review and meta-analysis. Epilepsia57:1754-1770.[crossref]
16. Gallego I, Ezquiaga E, Betancor D, et al. (2011) Psychogenic non-epileptic seizures in an epilepsy surgery unit. Rev Neurol52:449-446.[crossref]
17. Zhang YC, Bromfield EB, Hurwitz S, Aaron Nelson, Kristen Sylviaet al. (2009) Comparison of outcomes of video/EEG monitoring between patients with epileptic seizures and those with psychogenic nonepileptic seizures. Epilepsy Behav15:303-307.[crossref]
18. Yogarajah M, Powell HWR, Heaney D, et al. (2009) Medium term monitoring in refractory epilepsy: the Gowers Unit experience. J NeurolNeurosurg Psychiatry80:305-311.
19. Ghougassian DF, d’Souza W, Cook MJ, Terence J O’Brien(2004) Evaluating the utility of inpatient video-EEG monitoring. Epilepsia45:928-932.[crossref]
20. Sauro KM, Macrodimitris S, Krassman C, et al. (2014) Epilepsy Monitoring Unit Quality Improvement Team. Quality indicators in an epilepsy monitoring unit. Epilepsy Behav33:7-11.
21. Alsaadi TM, Thieman C, Shatzel A, Sarah Farias(2004) Video-EEG telemetry can be a crucial tool for neurologists experienced in epilepsy when diagnosing seizure disorders. Seizure13:32-34.[crossref]
22. Lee Y, Lee M, Chen I, Yu-TaiTsai, Chung-Yang Sunget al. (2009) Long-term Video-EEG Monitoring for Paroxysmal Events. Chang Gung Med J32:305-312.[crossref]
23. Mohan KK, Markand ON, Salanova V (1996) Diagnostic utility of video EEG monitoring in paroxysmal events. ActaNeurolScand94:320-325.[crossref]
24. Foong M, Seneviratne U (2016) Optimal duration of video-electroencephalographic monitoring to capture seizures. J ClinNeurosci28:55-60. [crossref]
25. Mari F, Di Bonaventura C, Vanacore N, JinaneFattouch, Anna ElisabettaVaudanoet al. (2006) Video-EEG study of psychogenic nonepileptic seizures: Differential characteristics in patients with and without epilepsy. Epilepsia5:64-67.[crossref]
26. McBride AE, Shih TT, Hirsch LJ (2002) Video-EEG monitoring in the elderly: a review of 94 patients. Epilepsia43:165-169.[crossref]
27. Friedman DE, Hirsch LJ (2009) How long does it take to make an accurate diagnosis in an epilepsy monitoring unit? J ClinNeurophysiol26:213-217.[crossref]
28. Benbadis SR, Tatum WO (2003) Overintepretation of EEGs and misdiagnosis of epilepsy. J ClinNeurophysiol20:42-44. [crossref]
29. Benbadis SR (2013) “Just like EKGs!” Should EEGs undergo a confirmatory interpretation by a clinical neurophysiologist? Neurology 80: 47-51.[crossref]
30. Benbadis SR (2007) Errors in EEGs and the misdiagnosis of epilepsy: importance, causes, consequences, and proposed remedies. Epilepsy Behav11:257-2562.[crossref]
31. Manfredonia F, Lawley A, Cavanna AE (2016) Impact of video-ambulatory electroencephalography on the medical management of epilepsy. J NeurolSci365:139-1342.[crossref]
32. Harden CL, Burgut FT, Kanner AM (2003) The diagnostic significance of video-EEG monitoring findings on pseudoseizure patients differs between neurologists and psychiatrists. Epilepsia44:453-456.[crossref]
33. Parnell KJ, Cascino GD, So EL, et al. (1999) Long-term EEG monitoring in patients with spells: clinical characteristics and predictive factors. Neurology 52: 371-372.
34. Cho YW, Motamedi GK, Kim KT (2019) The clinical utility of non-invasive video-electroencephalographic monitoring has been diversifying. NeurolSci 40:2625-2631. [crossref]
35. Morrel MJ, Pedley TA (2000) “The Scarlet E”: epilepsy is still a burden. Neurology54:1882-1883.[crossref]

DOI: 10.31038/AWHC.2020332

Abstract

Background: Sadomasochistic fantasies are often stigmatized and not easily disclosed to friends and family members. Although the nature of these fantasies is still incompletely understood, more frequent unconventional sexual fantasies seem to be associated with male gender with non-heteronormative sexual orientation and with higher educational level.

Methods: This was a cross-sectional study in which subjects provided information through a self-reported questionnaire in a face-to-face interview. This tool included questions assessing sociodemographic characteristics, the Beck Depression Inventory, and the subscale “sadomasochistic fantasies” of the Wilson Sexual Fantasy Questionnaire. A total of 412 medical students aged 18 and over attending first through sixth year at a medical school were randomly selected and recruited to participate.

Results: Non-heteronormativity and illicit drug use were directly and positively correlated with higher scores on sadomasochistic sexual fantasies. In addition, non-heteronormativity was a mediator variable in the model between being male and having higher scores on sadomasochistic sexual fantasies.

Conclusion: It is possible that unconventional sexual fantasies are more frequent in certain social groups, such as non-heteronormative males with high educational level. Although the use of psychoactive substances was correlated with sadomasochistic sexual fantasies, there are scarce scientific data to support this finding.

Keywords: Sadomasochistic fantasies, University Students,Heteronormativity

Introduction

The deliberate act of mentally envisioning a sexual scenario involving a target and/or behavior is a normal part of human sexuality. The content of the mental imagery or sexual fantasy frequently reflects one’s sexual interest and is experienced as sexually arousing [1]. In fact, some studies have shown a positive correlation between the experience of arousing sexual fantasies and sexual satisfaction [2].

There are important reasons for studying the diversity of sexual fantasies. First, although sexual fantasies are universally experienced, they can affect sexual behavior; second, sexual fantasies can be influenced by what people have previously seen, read, or done; third, repetitive sexual fantasies can help to shape our sexual schema or script; fourth, as sexual fantasies are private, they may be more revealing than actual behavior [1]. In sum, sexual fantasies can reflect our sexual behavior, which in turn can reflect them.

Of the enormous variety of sexual fantasies, sadomasochistic fantasies are often stigmatized and not easily disclosed to friends and family members by the fantasizers [3,4]. In addition, individuals with ingrained sadomasochistic sexual fantasies might need to reshape their identity to cope with shame, guilt, self-labeling, and self-hatred, and might need to overcome phases of dissatisfaction and depression before assuming or even expressing these sexual fantasies [5].

In fact, sadism and masochism have also been stigmatized medically. It was only in the 1970s and 1980s that a growing body of studies from the social sciences took a non-pathological view of sadomasochistic fantasies, practices, and behaviors [6]. Studies of consensual sadomasochistic practices have shown the healthy aspects of this behavior, such as improvement of intimacy between practitioners and greater creative stimulation [7]. In addition, associations of sadomasochistic practices with mental instability, depression, anxiety, and antisocial or psychotic traits have not been supported [8, 9]. Thus, since the latest edition of the Diagnostic and Statistical Manual of Mental Disorders [10] the terms sexual sadism and sexual masochism were changed to sexual sadism disorder and sexual masochism disorder in order to draw a line between non-pathological and pathological sexual behavior. In truth, one of the most important distinctions between deviant and normal sadomasochistic behavior is the presence or lack of consent. This perspective is also assumed within sadomasochistic communities, where consensual play and sex are unbreakable principles [11]. Despite this, the nature of sadomasochism is still incompletely understood, even though many studies have applied a broad variety of qualitative and quantitative methods [12].

Despite the demedicalization of consensual sadomasochistic behaviors, we should nonetheless consider that certain psychosocial factors seem to be linked to a higher diversity of sexual fantasies and practices in general. Higher diversity of sexual practices is consistently found to be associated with male gender, non-heteronormative sexual orientation, and higher educational level [13-16]. Considering these findings, nonclinical clusters of people with more intense unconventional sexual fantasies could be grouped into certain socio-demographic profiles [17]. In a somewhat different way, some studies have suggested that, although men show more frequent sadomasochistic fantasies, there are a large number of women in sadomasochistic communities [18-20]. In addition, [21] contends that “those who are most attracted to masochism may be the women (…). The women turn to masochism to live out the humiliation, the submission that they no longer have to endure anywhere else, or to remind themselves of other realities.” Based on these contrasting assumptions, it is possible that there is a mediator variable between biological sex and sadomasochistic sexual fantasies.

Still considering differences between sadomasochistic fantasizers and non-fantasizers, drug use has been scarcely investigated. It is important to note that it is possible that some sexual fantasies or feelings may be related to urges or cravings for drugs. Many drug users become trapped in a “reciprocal relapse” pattern in which a sexual behavior precipitates relapse to drugs and vice-versa [22]. According to some sadomasochistic communities, “just as in the greater population, there are many people in these communities who identify as clean and sober, and there are many who do not” [23]. This statement demonstrates the heterogeneity of those that have sadomasochistic sexual fantasies in terms of drug use. Examining such heterogeneity, a study with 164 sadomasochistically oriented males showed that the use of psychoactive substances before or during sadomasochistic sessions is not negligible. About 26%, 17%, and 5% of the participants of this sample admitted to using alcohol, poppers, and marijuana, respectively [24]. It is important to note here that poppers (volatile nitrites) are forbidden for recreational use in Brazil by the Brazilian Health Regulatory Agency (volatile nitrites are approved for use only for industrial purposes).

Also in line with other studies linking psychological problems with unconventional sexual fantasies, an association of traumatic childhood with sadomasochism and intermittent depression has been suggested [25]. In truth, depression symptoms can consist of a phase that individuals with sadomasochistic fantasies have to overcome before accepting and expressing them. Although these mental problems, drug use and depressive symptoms, may be directly or indirectly associated with unconventional sexual fantasies and behaviors, there does not seem to be any causal nexus. That said, we understand that sexual science would benefit from more systematic assessments of sadomasochistic fantasies and behaviors in clinical and nonclinical samples.

This study aimed to investigate whether psychosocial aspects, such as male gender and non-heteronormative sexual orientation, are associated with more frequent sadomasochistic sexual fantasies. In addition, we evaluated whether drug use and depressive symptoms are correlated with these sexual fantasies in a nonclinical sample of university students.

Method

Procedure

Permission to use the Wilson Sexual Fantasy Questionnaire (WSFQ) was obtained from the instrument’s marketers (CymeonTM Research, Sydney, Australia). Prof. Glenn Wilson was also contacted by our staff, and he referred us to the CymeonTM Research team. The original version was translated using the standard processes of back translation [26]. The English version of the instrument was translated into Portuguese by a team including one professor, four psychiatrists, and two psychologists with experience in sexual disorders and competency in both English and Portuguese languages, and two independent bilingual native speakers. The staff worked collaboratively to ensure that the instrument had semantic equivalence across the languages and conceptual equivalence across cultures. The translation coordinator compared both versions and reconciled any differences. Finally, the team compiled the Portuguese version and chose the most appropriate wording for clarity and similarity to the original. The final Portuguese version was formalized after the team discussed culturally problematic issues.

The Portuguese version was then independently translated back to English by two separate translators, neither of whom had previously seen the original scale. The back-translated versions were also evaluated and discussed by the team. A pilot study was then performed on a small sample (N = 10) of healthy individuals from diverse educational levels to examine whether any items on the WSFQ were perceived as difficult. No problematic items requiring revision were found.

Subsequently, a cross-sectional study was conducted to investigate associations or correlations between some psychosocial variables, mainly among those potentially related to unconventional sexual fantasies, such as biological sex, sexual orientation, illicit drug use, and depression. The investigators were specially trained medical graduate and postgraduate students. This study was approved by the Ethics Committee of ABC Medical School, Santo André, São Paulo, Brazil.

Participants

Between November 2016 and August 2019, a total of 412 medical students aged 18 and over attending the first through sixth year at one medical school were randomly selected and recruited to participate in this study. They were assured that their participation was voluntary, that only the researchers would see the data, and that all data would be kept confidential. A financial reward was not provided because this is not allowed under Brazilian law.

Important participant outcomes were compared based on 11 variables: biological sex, age, race or ethnicity, marital status, lifetime alcohol use, lifetime illicit drug use (marijuana, poppers, and cocaine were grouped into just one group), family members with alcohol use problems, family members with illicit drug use problems, sexual orientation, scores on depression symptoms, and scores on sadomasochistic sexual fantasies. Sex was coded as male, female, and intersex. Monthly income was not coded because our participants follow a full-time course of study.

Measures

This was a cross-sectional study in which subjects provided information through a self-reported questionnaire. This tool included questions assessing sociodemographic characteristics and the following inventories: The Beck Depression Inventory and the subscale “sadomasochistic fantasies” of the Wilson Sexual Fantasy Questionnaire.

The Beck Depression Inventory (BDI)

This inventory measures behavioral responses related to depression among adults and adolescents. In this 21-item instrument, scores above 10 (score range: 0–63) indicate the presence of a depressive syndrome [27, 28]. A sensitivity of 100% and specificity of 0.83 are obtained with a cut-off score of 9/10.

The Wilson Sex Fantasy Questionnaire (WSFQ)

This questionnaire is a 40-item self-report measure of sexual fantasies comprising a range of sexual themes “from the normal to the deviant and potentially harmful” [29]. Each item is scored on a four-point scale ranging from Never (0) to Often (3) across five different contexts (e.g., Daytime fantasies, Fantasies during intercourse or masturbation, Dream while asleep, Have done in reality, and Would do in reality). When assessing the frequency of sexual fantasy use, it is considered advisable only to use responses for Daytime fantasies, since scores for the other four contexts are highly correlated with Daytime dreams [30, 31]. Four subscales are derived from this instrument, including intimate (e.g., kissing passionately, having intercourse with a loved partner, being masturbated to orgasm by a partner), impersonal (e.g., sex with strangers, watching others having sex, fetishism), exploratory (e.g., group sex, promiscuity, mate-swapping), and sadomasochistic sexual fantasies (e.g., whipping or spanking, being forced to have sex). Each subscale has 10 items and this 4-factors structure has demonstrated consistency across multiple assessments [32, 33]. In line with the aims of this study, we only used the sadomasochistic sexual fantasy subscale. The following question was constructed for the participants: How often do you fantasize about the theme below at various times?

Analysis

Univariate analyses were used to compare the sociodemographic and psychometric features between men and women and between heteronormative and non-heteronormative participants. Categorical variables were compared using the χ2 or Fisher’s exact tests, following the Monte Carlo method. Continuous variables were compared using Student’s t-test.

In order to develop a correlational model, we performed a Structural Equation Modelling (SEM). Maximum likelihood estimation was used to estimate the fit of the model. TheComparative Fit Index (CFI), Tucker-Lewis Index (TLI), Goodness of Fit Index (GFI), Adjusted GFI (AGFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR) were used to evaluate model fit. Some standard recommendations regarding values for global model fit were followed. Specifically, CFI, TLI, GFI, and AGFI values greater than 0.90 and RMSEA and SRMR values lower than 0.08 were deemed indicative of acceptable model fit [34, 35]. As the chi-square value is dependent on the sample size, we calculated the ratio of chi-square to the degrees of freedom (c2/df), where a value of 2 or lower is an acceptable c2/df ratio [36].

Results

Of the questionnaires applied, 8 (1.94%) were discarded due to incomplete answers, leaving 402 participants. Of the participants, 200 (49.75%) were male and the mean age of the total sample was 21.45 (SD = 2.35) years old. Our sample had neither intersex nor transgender participants.

Descriptive analysis

As it is shown in Table 1, when biological male and female respondents were compared, male students showed more frequent non-heteronormative sexual orientation, and female students demonstrated higher mean scores on the BDI. There were no significant differences between the sexes in age, marital status, alcohol and illicit drug use, family members with alcohol and drug use problems, or mean scores on sadomasochistic sexual fantasies. As shown in Table 2, when heteronormative and non-heteronormative students were compared, those that admitted to non-heteronormativity demonstrated higher scores on the BDI and on sadomasochistic sexual fantasies, and more frequent family members with alcohol use problems. There were no statistically significant differences regarding the other psychosocial variables.

Table 1. Psychosocial and Psychometric features between male and female Medical University students.

Note: * p < 0.04; ** p < 0.01; BDI = Beck Depression Inventory

Table 2. Psychosocial and Psychometric features between Medical University students in accordance with the sexual orientation.

Note: * p < 0.04; ** p < 0.01; BDI = Beck Depression Inventory

SEM analysis

For the purposes of our SEM analysis, items were loaded uniquely on their respective factors and the factor loadings were fixed at 1.0. The sample was then evaluated using bootstrapping (800 bootstrap samples) with the Bollen-Stine Bootstrap statistic being calculated to verify absolute fit. As shown in Figure 1, the model fitted the data well, with c2/df= 0.69; CFI = 0.99; TLI = 0.99; GFI = 0.99; AGFI = 0.98; RMSEA = 0.012 [95% CI = 0.010-0.191]; SRMR = 0.02; and a Bollen-Stine statistic of p = 0.51.

Figure 1. Psychosocial and psychometric features correlated with sadomasochistic fantasies.

This model showed that non-heteronormativity and illicit drug use were directly and positively correlated with higher scores on sadomasochistic sexual fantasies. In addition, this model showed that non-heteronormativity was a mediator variable between having family members with alcohol problems and higher scores on sadomasochistic sexual fantasies. Furthermore, the variables of female sex and non-heteronormativity were directly and positively correlated with higher scores on depression. Non-heteronormativity also mediated the correlation between female sex and higher scores on depression.

Discussion

This study supports previous ones that have shown no correlation between depression symptoms and sadomasochistic fantasies, and that non-heteronormative persons have a higher frequency of these sexual fantasies. In addition, it found a correlation between illicit drug use and sadomasochistic fantasies.

Being male was not correlated with more frequent sadomasochistic sexual fantasies; however, being male was correlated with non-heteronormativity, which in turn was correlated with these unconventional sexual fantasies. We should consider that non-heteronormativity was a mediator variable between being male and having more frequent sadomasochistic fantasies in the correlational model. That said, it is possible that unconventional sexual fantasies are more frequent in certain social groups, such as non-heteronormative males with high educational level.

With regard to drug use, there is scarce evidence of an association with unconventional sexual fantasies. A few studies have investigated this use among sadomasochistically oriented people [24] and therefore it is necessary to devote greater attention to this theme, since our study shows a positive and direct correlation. It is possible that users of a variety of psychoactive substances experience strong aphrodisiac effects and disinhibition. This combination may result in obsessive pornography viewing, diverse sexual fantasies (including unconventional ones), and unsafe sexual practices [37]. Nonetheless, there is no definitive link between “kinky” sex and drug use to date.

Although comparisons between male and female and between heteronormative and non-heteronormative students in terms of depression symptoms were not the main goal of this study, the higher mean scores on depression in females than in males, mainly among young people [38-40] and in non-heteronormative persons [41-43] are widely supported in the scientific literature. Furthermore, studies have demonstrated that sexual minorities show a higher risk of having a family history of alcohol use problems than heterosexuals [44, 45]. Although the interpretation of this last association is politically sensitive [46] the fact is that this study does not show a causal relationship between sexual orientation and parental problems. It is possible to say here that non-heteronormative students may show higher self-reflection during the coming-out process or have a desire to endorse their own sexual orientation, leading them to reveal this fact more emphatically [47].

Although some authors have found significant associations between non-heteronormativity and alcohol and/or illicit drug use [48, 49] our study was not able to show this type of association. It is possible that given the notable and welcome reduction of the stigma, shame, and secrecy surrounding non-heteronormativity, the differences in the incidence of alcohol and drug use may be diminishing between heterosexuals and homosexuals/bisexuals. Also, the social context where people live can have a notable influence on the differences between heteronormative and non-heteronormative persons regarding alcohol and drug use [50] and this sample consisted of university students at a private institution.

In addition, the correlation between being male and non-heteronormative is not surprising. In Brazil, like in other diverse countries, the prevalence of homosexuals and bisexuals is higher in men, that is, almost 6% of males identify as gay and 2% identify as bisexual, while almost 3% of women identify as lesbian and 1% as bisexual [51].

This study showed that sadomasochistic fantasies in this sample of university students are more frequent in a group characterized by a non-heteronormative sexual orientation and reporting drug use. Although the first assertion has empirical support, the drug use among fantasizers needs further investigation.

There are several limitations in this study that need to be pointed out:

a) Response accuracy of research using self-response questionnaires may be less than fully satisfactory;

b) The university population is unrepresentative of the general population; and

c) The study’s cross-sectional design precludes drawing causal inferences and only provides information about population frequency and characteristics in a “snapshot” at a specific time.

References

1. Leitenberg H, Henning K(1995) Sexual fantasy. Psychological Bulletim117: 469-496.
2. Colon Vilar G, Concepcion E, Galynker I, Tanis T, Ardalan F, et al.(2016) Assessment of sexual fantasies in psychiatric inpatients with mood and psychotic disorders and comorbid personality disorder traits. The Journal of Sex Medicine 13:262-269.[Crossref]
3. Bezreh T, Weinberg TS, Edgar T (2012) BDSM disclosure and stigma management: Identifying opportunities for sex education. American Journal of Sexuality Education7: 37-61.[Crossref]
4. Wright, S. (2006). Discrimination of SM-identified individuals. Journal of Homosexuality, 50, 217-231.
5. Kamel GWL (1983) The leather career: On becoming a sadomasochist. In T. K. Weinberg (Ed.), S and M: Studies in sadomasochism. Buffalo: Prometheus Books.
6. Weinberg TS (2006) Sadomasochism and the social sciences: A review of the sociological and social psychological literature. Journal of Homosexuality 50: 17-40.[Crossref]
7. Williams DJ (2009) Deviant leisure: Rethinking “the good, the bad, and the ugly”. Leisure Science31: 207-213.
8. Connoly PH (2006) Psychological functioning of Bondage Domination Sado-Masochism (BDSM) practitioners. Journal of Psychology & Human Sexualitys 18: 79-120.
9. Cross PA, Matheson K (2006) Understanding sadomasochism: An empirical examination of four perspectives. Journal of Homosexuality, 50: 133-166.[Crossref]
10. American Psychiatric Association. (2013)Diagnostic and statistical manual of mental disorders (5th edn.). Washington: Author.
11. Sagarin BJ, Cutler B, Cutler  N, Lawler-Sagarin KA, Matuszewich L (2009) Hormonal changes and couple bonding in consensual sadomasochistic activity. Archives of Sexual Behavior38: 186-200.[Crossref]
12. Weierstall R, Giebel G (2017) The sadomasochism checklist: A tool for the assessment of sadomasochistic behavior. Archives of Sexual Behavior, 46: 735-745.[Crossref]
13. Bhugra DRB,  Rahul B (2006) Sexual fantasy in gay men in India: A comparison with heterosexual men. Sexual and Relationship Therapy21: 197-207.
14. Gagnon JH, Simon W (1987) The sexual scripting of oral genital contacts. Archives of Sexual Behavior16: 1-25.[Crossref]
15. Moser C, LevittEE (1987) An exploratory-descriptive study of a sadomasochistically oriented sample. The Journal of Sex Research23: 322-337.
16. Nordling N, Sandnabba NK, Santtila P, Alison L (2006) Differences and similarities between gay and straight individuals involved in the sadomasochistic subculture. Journal of Homosexuality50: 41-57.[Crossref]
17. Joyal CC (2015) Defining “normophilic” and “paraphilic” sexual fantasies in a population-based sample: On the importance of considering subgroups. Sexual Medicine3: 321-330.[Crossref]
18. Breslow N, Evans L, Langley J (1985) On the prevalence and roles of females in the sadomasochistic subculture: Report of an empirical study. Archives of Sexual Behavior 14: 303-317. [Crossref]
19. Holvoet L, Huys W, Coppens V, Seeuws J, Goethals K, e (2017) Fifty shades of Belgian gray: The prevalence of BDSM-related fantasies and activities in the general population. The Journal of Sex Medicine14: 1152-1159.[Crossref]
20. Rehor JE (2015) Sensual, erotic, and sexual behaviors of women from the “kink” community. Archives of Sexual Behavior44: 825-836.[Crossref]
21. Phillips A (1998)A Defence of Masochism. London: Faber and Faber Ltda.
22. Washton AM, Stone-Washton N (1993) Outpatient treatment of cocaine and crack addiction: A clinical perspective. Washington: National Institute on Drug Abuse (NIDA).
23. Ortmann DM, Sprott RA (2013)Sexual outsiders: Understanding BDSM sexualities and communities.Toronto: Rowman & Littlefield Publishers.
24. Sandnabba NK, Santilla P, Nordling N (1999) Sexual behavior and social adaptation among sadomasochistically-oriented males. Journal of Sex Research36: 273-282.
25. Gabriel JBS (2007) Early trauma in the development of masochism and depression. International Forum of Psychoanalysis 6: 231-236.
26. Guillemin F, BombardierC, Beaton D (1993) Cross-cultural adaptation of health-related quality of life measures: Literature review and proposed guidelines. Journal of Clinical Epidemiology46: 1417-1432.[Crossref]
27. Beck AT, Rial WY, Rickels K (1974) Short form of depression inventory: Cross-validation. Psychological Reports 34: 1184-1186. [Crossref]
28. Furlanetto LM, Mendlowicz MV, Romildo Bueno J (2005) The validity of the Beck Depression Inventory-Short Form as a screening and diagnostic instrument for moderate and severe depression in medical inpatients. Journal of Affective Disorder86: 87-91. [Crossre]
29. Wilson G (1988) Measurement of sex fantasy. Sex and Marital Therapy3: 45-55.
30. Bartels RM, Lehmann RJB, Thornton D (2019) Validating the utility of the Wilson Sex Fantasy Questionnaire with men who have sexually offended against children. Frontiers in Psychiatry10: 206. [Crossref]
31. Wilson G (1978)The Secrets of Sexual Fantasy.Toronto: JM Dent & Sons Ltda.
32. Plaud JJ, Bigwood SJ (1997) A multivariate analysis of the sexual fantasy themes of college men. Journal of Sex & Marital Therapy23: 221-230.[Crossref]
33. Sierra JC, Ortega V, Zubeidat I (2006) Confirmatory factor analysis of a Spanish version of the sex fantasy questionnaire: Assessing gender differences. Journal of Sex & Marital Therapy32: 137-159.
34. Gilson KM, Bryant C, Bei B, Komiti A, Jackson H, et al., (2013) Validation of the Drinking Motives Questionnaire (DMQ) in older adults. Addictive Behaviors38: 2196-2202.[Crossref]
35. Hu L, Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling6: 1-55.
36. Tabachnick BG, Fidell LS (2007)Using Multivariate Statistics.Boston: Pearson Education.
37. Slavin MN,  Kraus SW, Ecker A, Sartor C, Blycker GR, et al., (2017) Marijuana use, marijuana expectancies, and hypersexuality among college students. Sexual Addiction & Compulsivity24: 248-256.[Crossref]
38. lbert PR (2015) Why is depression more prevalent in women? [Editorial]. Journal of Psychiatry Neurosciences40:219-221.[Crossref]
39. Cyranowski JM, Frank E, Young E,  Shear MK (2000) Adolescent onset of the gender difference in lifetime rates of major depression: A theoretical model. Archives of General Psychiatry57: 21-27. [Crossref]
40. Ford DE, Erlinger TP (2004)Depression and C-reactive protein in US adults: Data from the Third National Health and Nutrition Examination Survey. Archives of Internal Medicine164: 1010-1014.[Crossref]
41. King M, Semlyen J, Tai SS, Killaspy H, Osborn D, et al., (2008) A systematic review of mental disorder, suicide, and deliberate self harm in lesbian, gay and bisexual people. BMC Psychiatry8: 70.[Crossref]
42. Lee C, Oliffe JL, Kelly MT, Ferlatte O (2017) Depression and suicidality in gay men: Implications for health care providers. American Journal of Men`s Health11: 910-919.[Crossref]
43. Pompili M, Lester D, Forte A, Seretti ME, Erbuto D, et al., (2014) Bisexuality and suicide: A systematic review of the current literature. The Journal of Sexual Medicine11: 1903-1913.[Crossref]
44. McCabe SE, West BT, Hughes TL, Boyd CJ (2013) Sexual orientation and substance abuse treatment utilization in the United States: Results from a national survey. Journal of Substance Abuse Treatment44: 4-12.[Crossref]
45. McCabe SE, West BT, Hughes TL, Boyd CJ (2013) Sexual orientation and substance abuse treatment utilization in the United States: Results from a national survey. Journal of Substance Abuse Treatment44: 4-12.[Crossref]
46. Roberts AL, Glymour MM, Koenen KC (2013) Does maltreatment in childhood affect sexual orientation in adulthood? Archives of Sexual Behavior42: 161-171.[Crossref]
47. Corliss HL, Cochran SD,  Mays VM (2002) Reports of parental maltreatment during childhood in a United States population-based survey of homosexual, bisexual, and heterosexual adults. Child Abuse & Neglect 26: 1165-1178.[Crossref]
48. Green KE,  Feinstein BA (2012) Substance use in lesbian, gay, and bisexual populations: An update on empirical research and implications for treatment. Psychology of Addictive Behaviors26:265-278.[Crossref]
49. Hughes T, McCabe SE, Wilsnack SC, West BT, Boyd CJ (2010) Victimization and substance use disorders in a national sample of heterosexual and sexual minority women and men. Addiction105: 2130-2140.[Crossref]
50. Duncan DT, Hatzenbuehler ML, Johnson RM (2014) Neighborhood-level LGBT hate crimes and current illicit drug use among sexual minority youth. Drug and Alcohol Dependence135: 65-70.[Crossref]
51. Abdo C (2004)Estudo da vida sexual do brasileiro (Study on the Sexual Life of Brazilians) (1st edn.). São Paulo: Bregantini.