Association between non-suicidal self-injury and suicidal behavior in Borderline Personality Disorder: a retrospective study

Abstract

Suicidal behavior (SB) spectrum is nuclear in the clinic and management of Borderline Personality Dis- order (BPD). Although in recent research papers non-suicidal self-injury behavior (NSSI) and suicidal behavior (SB) differ in intentionality, frequency and lethality; these two behaviors have been described concurrently with a controversial distinction. Few works talk about the reason for the co-occurrence between both entities in the psychiatric population in general and in BPD in particular. The aim of the report is to analyze the link between SB and NSSI in BPD.

ABSTRACT

Introduction

Suicidal behavior (SB) spectrum is nuclear in the clinic and management of Borderline Personality Disorder (BPD). Although in recent research papers non-suicidal self-injury behavior (NSSI) and suicidal behavior (SB) differ in intentionality, frequency and lethality; these two behaviors have been described concurrently with a controversial distinction. Few works talk about the reason for the co-occurrence between both entities in the psychiatric population in general and in BPD in particular. The aim of the report is to analyze the link between SB and NSSI in BPD.

Methods

A cross-sectional, observational and retrospective study was carried out on a sample of 134 patients between 18 and 56 years old, diagnosed with BPD according to DSM-5 criteria. The association between variables was analyzed through a negative binomial and multivariate logistic regression model.

Results

77.6% report a history of at least one suicide attempt (SA), while 30.4% none. The average number of SA is 2.69. For NSSI, 64.2% presented them, while 35.8% did not. A statistically significant association is found between both of them. NSSI are also significantly related to performing a greater number of SA according to the multivariate analysis.

Conclusions

The results suggest that these behaviors are nuclear and frequent in BPD. Both appear significantly related to each other. Looking ahead, longitudinal studies are needed to confirm the relationship between these variables.

INTRODUCTION

Borderline Personality Disorder (BPD) is a multifactorial etiology disorder characterized by emotional instability, feeling of emptiness and impulsive behaviors(1). Among the impulsive behaviors associated with BPD, suicidal behavior (SB) is considered a core aspect(2)-(4). Between 40 and 85% of BPD patients commit suicide attempts (SA), with an average of three per patient(5)-(6). Completed suicide rate amid BPD patients is between 5 and 10%, which is about 400-times higher than the estimated for general population(5)-(7). It is considered a symptom with great clinical relevance and important prognostic value, and it is one of the manifestations that most affects the functionality of these patients(7).

The spectrum of SB and related behaviors takes into account behaviors that cause direct and deliberate harm to oneself(8). Nonsuicidal self-injury (NSSI), SB, and completed suicide itself are included(9). Although in recent research, NSSI and SB differ in intent, frequency and lethality(10), it has been described that these two behaviors often occur concurrently(11)-(12). For some authors, NSSI are a risk factor and in some way, precursor behaviors, for the future appearance of SA(13)-(16). Likewise, the iceberg model has been used to illustrate the great prevalence of undetected self-harm behaviors as part of a spectrum that encloses SB and finishes with completed suicide(18)(18).

The distinction between SB and NSSI is a controversial issue, from which different theories have been postulated, such as the “Gateway Theory”, that of the “Third variable” and that of “Acquisition of capacity for suicide” by Joiner(19). Furthermore, few studies deal with the reason for the co-occurrence between both entities in the psychiatric population in general and in Borderline Personality Disorder (BPD) in particular.

From the hypothesis that SA in patients with BPD are associated with the concomitant presence of NSSI and reciprocally, NSSI are associated with the presence of SA; the objectives of this work are:

Analyze the relationship between SB and NSSI in BPD

Determine if SA are related to NSSI presence and NSSI to SA

Know if the presence of NSSI are related to a greater number of SA

METHODS

We performed a cross-sectional, observational and retrospective study which aimed to analyze the relationship between the NSSI and the SB using a sample of 134 patients between 18 and 56 years old, diagnosed with BPD according to DSM-5 criteria (listed in Table 1). These patients were recruited consecutively in the admission process to the Personality Disorders Day Unit of the Hospital Clínico San Carlos in Madrid, which is a specific and nationwide reference unit for the treatment of patients with this diagnosis.

1. Table 1.

Diagnostic criteria for Borderline Personality Disorder, according to DSM-5 (APA. American Psychiatric Association. Diagnostic Manual and Statistics of Mental Disorders (DSM-5). Arlington V, editor. Madrid: Editorial Médica Panamericana; 2014.)

1	Desperate or exaggerated efforts to avoid an experience of helplessness
2	Pattern of interpersonal relationships characterized by instability, which oscillates between idealization and rejection of other people
3	Alteration of the experience of one's own identity
4	Impulsiveness
5	Suicidal behaviors
6	Affective instability
7	Chronic feeling of emptiness
8	Inappropriate bouts of anger or rage
9	Transitory paranoid ideas

The main descriptive characteristics of the patients are shown in (Table 2).

2. Table 2.

Descriptive variables of the studied sample /n/Percentage (%)

			n	Percentage (%)
Sex (N = 134)
Male			37	27.6
Woman			97	72.3
Marital status (N = 114)
Single			83	72.8
Married or with a partner			26	22.8
Divorced or separated			5	4.4
Children (N = 134)
No			109	81.4
Yes			25	18.6
Current activity (N = 134)
Unemployed			78	58.2
Working			16	11.9
Student			24	17.9
Leave from work			16	11.9
Educational level (N = 134)
Primary studies			18	13.5
Secondary studies			53	39.8
Vocational training			27	20.3
University studies			35	26.6
Socioeconomic level (N = 134)
Low			20	22.7
Intermediate			37	42
High			31	35.2
	Years of evolution of the disease (N = 134)
Up to five years			23	17.2
Between five and ten years			26	19.4
More than ten years			85	63.4
	Type of previous treatment (N = 134)
None			3	2.2
Psychopharmacological			34	25.4
Psychotherapeutic			12	8.9
Psychopharmacological + Psychotherapeutic			85	63.4

Patients who met criteria for other diagnoses, had an IQ of less than 85 or severe neurological disease, a history of traumatic brain injury, severe medical illness, current abuse of psychoactive substances –except for tobaccoor declined to participate in the study were excluded. The Hospital Ethics Committee approved the evaluation protocol, and all the participants signed the informed consent.

For the clinical evaluation we used the validation into Spanish of the Columbia-Suicide Severity Rating Scale(20), assessing the presence or not of SA and NSSI, number of SA and modality of the attempts of autolysis. Patients were individually evaluated by a psychiatrist and a clinical psychologist for approximately 120 minutes in the Personality Disorder Day Hospital of the Hospital Clínico San Carlos in Madrid (Spain). In order to reduce variability, all tests were performed at similar times (between 10 and 12 a.m.).

Statistical analysis

The mean and the standard deviation were used for the description of continuous data and the percentages for categorical data. Regarding the quantitative variables, their concordance to a normal distribution was determined using the Kolmogorov-Smirnov test. The sample was divided into two groups according to whether or not there was a history of NSSI. Variable comparisons were made using Chi-squared test and Student’s t-test. The association between variables was analyzed through a multivariate negative binomial logistic regression model. Data analysis was performed using the SPSS statistical package, version 19.0. The significance level established for all of the hypothesis testing was 0,05.

RESULTS

For suicide / suicide attempts (SA), 104 patients out of 134 (77.6%) reported a history of at least one SA, while 30 patients (22.4%) did not have NSSI, as shown in (Figure 1). The mean number of SA is 2.69 for each patient, with a standard deviation of 1.774.

Fig. 1.

Suicidal and related behaviors (n = 134)

For non-suicidal self-injurious behaviors (NSSI), 86 patients (64.2%) reported a history of NSSI, while 48 patients (35.8%) did not present this history, as also detailed in (Figure 1).

Within the method followed to carry out the SA, the distribution shown in (Table 3) is observed. The most frequent method is the combination of methods (53.5%), followed by drug overdose (46.5%).

3. Table 3.

Autolytic attempt method in the studied sample

		n	Percentage (%)
Drug overdose
Yes		40	46.5
No		46	53.5
Venoclysis
Sí		6	7
No		80	93
Poisoning
Sí		1	1.2
No		85	98.8
Hanging
Sí		1	1.2
No		85	98.8
Precipitation
Sí		6	7
No		80	93
White weapon
Sí		1	1.2
No		85	98.8
Throwing oneself into vehicles
Sí		2	2,3
No		84	97.7
Other methods / combination of methods
Sí		46	53.5
No		40	46.5

With the multivariate analysis using logistic regression, a statistically significant association (p = 0.038) was found with the NSSI, with an odds ratio of 3.218, for a 95% confidence interval (1.069-9.690). With the multivariate analysis using negative binomial regression, the NSSI (64.2%) were also significantly related to performing a greater number of SA according to the multivariate analysis.

In turn, NSSI are statistically significantly associated with SA (p = 0.006) with an odds ratio of 4.037, and a 95% confidence interval (1.491-10.932).

DISCUSSION

The data presented show slightly higher SA history figures (77.6%) than those collected in the literature (4070%)(5)-(21)-(22). This may be due to the fact that the patients in the studied sample come from a clinical population with distinctive characteristics compared to other populations of patients with BPD. They are patients with BPD with a clinical course classified as severe, and who have been referred to the Personality Disorders Unit of the Hospital Clínico San Carlos de Madrid, either from the health area of the Hospital in which the Unit is located, or from other health centers, as a specific unit for BPD treatment, to benefit from more individualized programs for this type of population(23).

Another peculiar characteristic of the sample studied is that the presence of a history of suicide attempts (SA) (77.6%) is more frequent than the presence of self-injurious behaviors without suicidal intent (NSSI) (64.2%). The literature collects precisely the opposite data, that is, that NSSI are more frequent than SA(24). This characteristic could also be associated with the greater clinical severity of the patient sample mentioned previously. Given that, in the spectrum of suicidal behaviors and related behaviors, the way of reacting to circumstances experienced as stressful in moments of emotional overflow will be more likely to carry out a disruptive act or behavior with the purpose of death than to an act without that purpose. Precisely the main distinguishing feature between an SA and NSSI is the intention to die(8)-(9)-(25).

The distinction between SB and NSSI is a subject in debate in the scientific literature. The model of Hamza et al 2012(19) has been accepted, in which a model is proposed that integrates the “Gateway Theory”, that of the “Third variable” and that of “acquisition of capacity for suicide” of Joiner to explain the relationship between NSSI and SB. The BPD can function as this third variable, although the study model presented as will be pointed out later limits being able to conclude that it leads a progressive training from the NSSI to the SB itself through an increase in the perceived load or the feeling of frustrated belonging.

The present study finds a strong statistical association in the multivariate analysis between NSSI and SB in patients diagnosed with BPD, a finding also present in other studies(14). Suicide risk is associated with NSSI, particularly repeated self-harm. A statistically significant association was found in the multivariate analysis performed by negative binomial regression between the NSSI and the number of SA. That is, having presented NSSI is related not only to the presence of SA, but also to a greater number of attempts. Thus, the present study reproduces results found by other authors, who affirm that the history of NSSI is also associated with the risk of repetition of self-injurious behavior, especially the first month after the hospital evaluation(26).

There are few longitudinal studies that allow establishing the direction of the link between NSSI and SB. But, in any case, the statistical relationship between history of NSSI and the risk of future SB(19) can be affirmed. It is found that NSSI behaviors are a greater predictor of SA than vice versa, that is to say that SA with respect to NSSI, as recent works have pointed out(15)-(27)-(28). It is observed that both are closely related and can form a continuum in the expression of a basic discomfort common to them(29). Therefore, the identification of NSSI could be of interest in the prevention of more disruptive and serious behaviors such as SA, based on these results.

The main limitation of the study is that it is an observational, descriptive and cross-sectional study, of concurrent and retrospective temporality. This design model does not allow the establishment of causal links between the statistical associations described and does not have the statistical power comparable to that of a prospective study. Likewise, the presence of a control group could make the study more consistent, an aspect that was rejected at first because the statistics service and the Ethics Committee considered it unnecessary.

CONCLUSIONS

NSSI and SB are core behaviors in BPD. BPD is associated with high SA and NSSI rates. Both appear to be related to each other in BPD in the present study in a significant way according to the multivariate analysis. Likewise, in its quantitative assessment, a greater number of SA are significantly associated with NSSI.

ACKNOWLEDGEMENTS

We want to thank all the members of the Institute of Psychiatry and Mental Health of the Hospital Clínico San Carlos for their support and help in collecting the data for this study.

CONFLICT OF INTEREST

None