Preoperative Risk Factors for Suicide in Candidates for Weight Loss Surgery

Introduction

Although bariatric surgery continues to be one of the most efficacious treatments for morbidly obese individuals, ¹ among postsurgery deaths, suicide is one of the leading causes.^2,3 This elevated rate of suicide has been documented by numerous researchers. For example, after examining bariatric surgery records in Pennsylvania, Omalu et al. ² reported that among the 16,683 patients who underwent bariatric operations in a 9-year period, 16 (4%) died by suicide, while 14 (3%) died due to drug overdoses that potentially could have been suicide, but were not classified as such. In a follow-up study of these same data, Tindle et al. reported that suicide rates of bariatric surgery patients were significantly higher than those of the general population. ³ The researchers note that 30% of deaths by suicide occurred in the first two postoperative years, with 70% occurring by the third postoperative year.^2,3 As such, the risk for suicide appears to increase relatively shortly after weight loss surgery.

The above findings are intriguing when contrasted with research suggesting an inverse relationship between body–mass index (BMI) and suicide, further confounded by gender. Mukamal et al. found that among men, risk of death by suicide was inversely related to BMI. ⁴ In contrast, the Carpenter et al. study revealed that female patients with a 10-U increase in BMI reported a 22% increase in suicide attempts and ideation. Among men, a 10-U increase in BMI reduced the risk of ideation and attempts by 26% and 55%, respectively. ⁵ Dutton et al. determined that the relationship between obesity and suicidality has not been well defined or understood because conventional risk factors have not adequately explained the relationship. In an attempt to further define the mechanisms of the relationship between BMI and suicidal ideation, the authors demonstrated that increased suicidal ideation and perceptions of burdensomeness were associated with higher BMI in a group of university students and persons recruited from a community-based behavioral weight management (WM) clinic. ⁶

Although several studies have focused on the link between obesity and suicidal ideation, fewer have focused on suicidality in bariatric populations. A segment of the bariatric literature has focused on risk factors such as depression and lack of social support in the immediate postoperative phase. The studies to date have provided evidence for some suicide risk factors in this population; however, the majority of the studies of suicidality in surgical weight loss (SWL) groups have been descriptive in nature and have focused on variables that are not unique to suicidality (e.g., depression). To our knowledge, there have been no studies that have investigated suicide risk in bariatric patients utilizing a theoretical empirically supported model of suicide, although the need has been cited. ⁷ The purpose of this study was to describe the prevalence of specific risk factors for suicide in a preoperative SWL sample compared with individuals with similar levels of obesity who have not opted to undergo surgery by applying the interpersonal–psychological theory of suicidal behavior (IPTS). ⁸

The IPTS proposes that suicide results from the interaction of three necessary and sufficient components—thwarted belongingness, perceived burdensomeness, and the acquired capability for suicidal behavior. It is important to note that all three variables must be present for one to die by suicide; the desire for death alone is insufficient. Thwarted belongingness and perceived burdensomeness are related constructs that together represent suicidal ideation. Thwarted belongingness is a feeling of being disconnected from others. Perceived burdensomeness involves feeling as though one is a liability to others and that people would fare better if one were dead. ⁸ Prior research has demonstrated a robust relationship between thwarted belongingness, perceived burdensomeness, and the desire for death. ⁹

In contrast to suicidal ideation, acquired capability is the physical capacity to inflict lethal self-injury in the event that one desires to do so. Acquired capability is distinct from the desire to die by suicide. Death by suicide is contradictory to humans' instincts to protect their lives. Joiner asserts that one must override self-preservation to die by suicide and that this is a fearsome task. This capability is hypothesized to result from a process of habituation after one is faced with repeated exposure to painful and provocative events. Continued experience with such painful or provocative situations leads to fearlessness and an elevated tolerance for physical pain. This is true especially if the painful events include prior suicide attempts. ¹⁰ Thus, it follows that those who have high pain tolerance and do not fear death are more likely to die by suicide should the desire for suicide arise. Acquired capability has been demonstrated in various populations that are frequently exposed to painful and/or provocative situations, including those with eating disorders.^11,12

Although acquired capability has yet to be investigated specifically in those with obesity, or those undergoing bariatric surgery, it is plausible that acquired capability may be relevant to both situations. Given that increased BMI is associated with a host of health complications—many of which are painful—it stands to reason that having an increased BMI could influence one's acquired capability. For example, body weight may be positively correlated with acquired capability due to increased pain and mobility issues associated with a larger body habitus. While it may be the case that increased BMI displays a specific relationship with acquired capability, this risk factor—as well as thwarted belongingness and perceived burdensomeness—may also change from pre- to postsurgery in those undergoing bariatric surgery.

An increase in suicidal behavior from the preoperative phase to the postoperative phase in those undergoing bariatric procedures may indicate that certain experiences that occur during or after surgery could potentially increase thwarted belongingness, perceived burdensomeness, or one's acquired capability for suicide. In fact, research suggests that significant psychosocial changes—including those of romantic relationships—may result after surgery, potentially influencing feelings of belongingness. ¹³ Additionally, relying on others for caretaking postsurgery may lead to feelings of burdensomeness. Furthermore, a number of physical outcomes after bariatric surgery constitute situations that are either painful, provocative, or both (e.g., dumping syndrome, strictures). Given these potentially painful and provocative experiences and changes in psychosocial situations, the IPTS may be a promising means of exploring suicidality in those undergoing bariatric surgery. However, the IPTS has yet to be applied to individuals presenting for bariatric surgery.

To provide a baseline of sorts, from which to examine suicidality in SWL patients, we sought to report correlates of the IPTS in a preoperative bariatric population. This type of descriptive study will provide cross-sectional information to practitioners regarding suicide risk during the preoperative phase. Additionally, providing information on preoperative risk allows for researchers and clinicians to assess how risk may change during the critical period between preoperative intake and postoperative follow-up. Additionally, we sought to utilize a comparison group that did not pursue surgery as an option for managing weight, but that demonstrated a similar weight status nonetheless. To do so, we collected study measures from a behavioral WM group. Utilizing a comparison group with a similar BMI allows for a more specific evaluation of differences between those who are obese and are seeking weight loss surgery and those who are obese and are not seeking surgery. As such, relevant group differences on IPTS constructs may suggest characteristics unique to an SWL population that have the potential to influence suicidality in this group.

Methods

Results

Table 1 provides group comparisons on study variables. The SWL group was primarily female (75.9%) and the average patient age was 44.52 (range 22–70). Average BMI of SWL participants at intake was 48.21 (range 34–87). Sixty-one percent of participants in the WM group were female, and the average patient age was 44.79 (range 18–70). Average BMI of WM patients was 43.84 (range 26–73). One patient (0.7%) in the SWL group and nine patients (8.7%) in the WM group reported prior suicide attempts. The groups did not differ significantly on age. BMI differed significantly between the groups (F = 8.07, p < 0.05) such that the SWL group had significantly higher BMI scores. Additionally, gender differed significantly between groups (χ² = 6.17, p < 0.05), in that there were significantly more males in the WM group than the SWL group. As such, gender and BMI were entered as covariates in all analyses.

To compare group means on components of the IPTS, we utilized a one-way analysis of covariance (ANCOVA) for all statistical tests. Group (SWL vs. WM) was entered as the independent variable, BMI and gender were entered as covariates, and the relevant IPTS component was entered as the dependent variable in each analysis. Results indicated that both perceived burdensomeness (F = 21.97, p < 0.001) and thwarted belongingness (F = 34.62, p < 0.001) differed significantly by group even while controlling for gender and BMI. Specifically, those in the WM group scored higher on both constructs. In contrast, acquired capability for suicide did not differ between groups (F = 0.107, p = ns).

We also sought to examine group differences on several other variables known to be related to risk for suicide, including suicidal ideation and past history of physical and sexual abuse. A one-way ANCOVA revealed significant differences on suicidal ideation (F = 7.35, p < 0.01), with those in the WM group reporting higher scores. Finally, chi-square analyses revealed that those in the WM group were more likely to have experienced physical abuse compared with those in the SWL group (χ² = 6.21, p < 0.05). Sexual abuse was not reported with differing frequencies between the groups (χ² = 1.41, p = ns)

Discussion

As previously outlined, research suggests that suicide rates among SWL patients may be higher than that of the general population, specifically in the postsurgery period.^1,3 Previous researchers have described this trend; however, there have been no studies, to our knowledge, investigating established correlates of suicidality in preoperative samples. This is the first such study to examine the IPTS within an SWL group while utilizing a comparison group with similar levels of obesity. Although the groups did differ significantly on BMI, the average BMI in both groups exceeded 40, which would result in a classification of morbidly obese. BMI was utilized as a covariate in all analyses to help ensure that observed differences were not accounted for by differences in BMI.

Analyses revealed greater perceived burdensomeness and thwarted belongingness in WM patients compared with SWL patients. Given that these two constructs are highly correlated with suicidal ideation, ⁹ it is not surprising then that suicidal ideation was also greater in WM patients. However, because suicidal ideation has also been linked to excess weight, ⁶ it was unexpected that the two groups would not score similarly on these constructs. There are a few potential explanations for these observed differences. First, given that SWL candidates are providing the information utilized in the current study as a part of their standard preoperative intake procedures, they may be motivated to answer in socially desirable ways, fearing that if they report psychological distress they may be denied surgery. In a recent study, White et al. suggested that patients who participate in presurgical psychological evaluations produce elevated social desirability scores and typically deny mood symptoms. ²⁰ Impression management in preoperative assessment is a consideration that has been investigated in previous studies. ²¹ Impression management may be less likely to occur when assessing WM patients who are not anticipating a major surgery and, by contrast, are often soliciting constructive advice for behavior change.

Recent data gathered by the authors of this study found that in a sample of preoperative candidates, mean score on the Marlowe-Crowne Social Desirability Scale Short Form C (MC-C)—a questionnaire assessing socially desirable responding—was 9.25. ²² This score exceeds that of other populations noted to display elevated scores on the MC-C, including forensic populations (mean score of 7.61) and student and military populations (mean score of 5.37). ²³ This elevation would suggest that some socially desirable responding may be present. However, these data were taken from a different sample than the current study. Furthermore, the items related to perceived burdensomeness and thwarted belongingness are not particularly face valid (e.g., “these days, I feel disconnected from other people”). While prior research has shown a clear relationship between these constructs and suicidal ideation, the items assessing burdensomeness and belongingness do not directly ask individuals to report on suicidal thoughts or psychological symptoms. As such, it is less likely that patients would deliberately engage in impression management when completing these items.

A further explanation of lower scores displayed by SWL patients may be related to optimism stemming from the potential of drastic weight reduction through surgery. It is possible that preoperative SWL patients are in a hopeful state, envisioning significant, and at times unrealistic, changes to their bodies as a result of surgical intervention. Therefore, they may be likely to report fewer feelings of isolation and burdensomeness because they believe that surgery will soon improve their lives.

WM patients also reported more frequent histories of lifetime physical abuse than their SWL counterparts, which may partially explain the increased suicidal ideation in the WM group given that a history of physical abuse has been found to predict increased suicidal ideation. ²⁴ Additionally, 8.7% of the WM group (nine patients) reported prior suicide attempts, while only 0.7% (one patient) in the SWL group did. According to the IPTS, suicide attempts will occur when thwarted belongingness, perceived burdensomeness, and acquired capability are all sufficiently present. The increased rate of attempts in WM patients would suggest that these three factors are more often simultaneously present in this group. Indeed, two of the three IPTS constructs were significantly higher in the WM group.

No group differences were noted on acquired capability, which is somewhat surprising given the group differences in previous attempts. However, physical abuse (more frequently reported in the WM group) represents a particular type of painful and provocative experience—the kind of experience that has been found to predict increased acquired capability for suicide. As such, the physically painful and emotionally provocative experience of prior abuse may partially explain the elevated rate of suicide attempt in WM patients. Additionally, if both groups have experienced ongoing pain related to carrying increased weight, it is less surprising that acquired capability would be similar between the two groups. It could also be the case that acquired capability changes from the pre- to postsurgical period and that the painful and provocative nature of undergoing surgery contributes to these increases. However, this conclusion is precluded by the lack of follow-up data.

The question remains as to how and why risk for suicide appears to increase from the preoperative to postoperative phase. It is possible that SWL patients demonstrate a level of acquired capability sufficient for a lethal suicide attempt, but that their lower levels of perceived burdensomeness, thwarted belongingness, and suicidal ideation serve as protective factors. This consideration leads to a theoretical situation in which preoperative candidates do not demonstrate increased suicidal ideation, but do demonstrate acquired capability similar to those with higher rates of suicide attempt (WM patients). Given that the rates of postoperative suicide are significantly elevated compared with that of the general population, it would follow that suicidal ideation must increase at some point from pre- to postsurgery.

As previously delineated, suicide risk in this population seems to increase ∼3 years postsurgery.^2,3 This may suggest that while patients are not at particularly high risk before surgery, the unique aspects of undergoing a painful and life-altering major surgery may contribute to increased risk of suicide during the postoperative period. Clearly, longitudinal data spanning the time frame from pre- to postsurgery are necessary to explore this possibility. The current study provides a baseline descriptive report of such variables. Perhaps the change in risk occurs gradually as patients fail to make expected weight loss goals, are disappointed by their surgical outcome, or feel like a burden to those who assist them in their recovery period. This could explain the similar levels of acquired capability between the preoperative SWL group and the WM group while also taking into account the differing levels of suicidal ideation, thwarted belongingness, and perceived burdensomeness.

The data gathered in the current study may be useful to clinicians both preoperatively and postoperatively. First, clinicians should be encouraged to thoroughly explain preoperative procedures—particularly the psychological evaluation—to encourage honest responding in surgical candidates. If candidates view psychological evaluations as a gatekeeping tool that may have the potential to prevent them from receiving surgery, they may be less likely to respond honestly regarding their risk for suicide. Practitioners may wish to inform patients that the psychological evaluation is just one component of a comprehensive strategy to gather information that will help patients to obtain the best possible outcome after surgery. Furthermore, providing psychoeducation to patients regarding increased risk for suicide as well as significant physical and emotional changes that can happen during and after surgery is necessary.

Additionally, given that suicide risk appears to be most prominent in the first 3 years postsurgery, postoperative attrition is concerning. The data gathered in the current study also suggest that risk factors may continue to change during the postoperative period, particularly those related to thwarted belongingness and perceived burdensomeness. As such, patients may benefit from follow-up treatment planning that includes important members of their support system. Including these individuals in follow-up visits and recovery may not only help to encourage follow-up itself but may also help to educate both the patient and significant others of the increased risks that may occur postoperatively.

A final suggestion would be to emphasize a team approach in evaluating risk for suicide and assessing patients during postoperative care. Given that many patients cease attending follow-up appointments, communication between practitioners becomes even more important. For example, if a patient misses a follow-up appointment with their surgeon, the patient's primary care physician or dietitian may be the next person to see the patient and should make an effort to assess the patient not only physically but also emotionally and make referrals as necessary.

While the above findings provide useful clinical information for practitioners regarding the risk for suicide in the preoperative phase, several limitations are worth noting. First, the data are cross-sectional in nature and were assessed preoperatively—before rates of suicidality are noted to increase in this population. Additionally, preoperative SWL patients may respond in an overly favorable way because they are highly motivated to seek surgery. Given that response bias has previously been reported in the SWL literature (e.g., Glinski et al. ²⁵ ), it is necessary to consider this possibility when interpreting the results. However, the use of several measures that do not directly ask questions related to suicide or psychopathology (e.g., the INQ) may somewhat assuage this concern. Additionally, the ACSS demonstrated lower reliability in comparison with other scales. However, internal consistency of the measure is similar to that of other studies using the ACSS.^16–19 Additionally, internal reliability is influenced by the number of items on a scale. ²⁶ Given that there are only five items on the ACSS, it is less surprising that the internal consistency was 0.65. Finally, suicidal ideation was assessed by utilizing only one item (item 9 from the BDI-II). Given that suicidal ideation is a complex experience, using only one question to assess this construct precludes any understanding of different facets of suicidal ideation (e.g., intensity or frequency of suicidal thoughts).

A strength of the study includes the comparison of two groups with excess BMI. The data provided help to inform clinicians regarding the suicide risk of preoperative candidates compared with a group that has not elected to undergo weight loss surgery, but demonstrates similar levels of obesity. This comparison allows for potential prognostic data to be assessed longitudinally in both groups. A further strength is the use of a theoretically driven and empirically supported framework of suicidal behavior. The IPTS has garnered support in assessing suicidality in a wide array of populations^9,18 and may be a promising means by which to assess suicide risk factors in both SWL and WM groups.

Conclusions

SWL and behavioral WM groups appear to display differential risk for suicide. Although suicide risk is increased in the years following bariatric surgery, risk appears to be lower in preoperative SWL populations compared with a similarly obese WM population. Future longitudinal studies aimed at investigating changes in suicide risk from pre- to postsurgery will help to clarify the unique aspects of weight loss surgery that may contribute to increased risk of suicide in SWL populations as well as help to elucidate the controversial literature regarding suicidality and BMI.