Introduction

In addition to the health complications associated with morbid obesity, levels of psychological distress, such as depression and anxiety, are significantly higher than those reported in the general population [1,2,3]. The National Co-morbidity Survey Replication [4] reports lifetime prevalence rates of 46.4% for axis I disorders, 20.8% for any mood disorder, 14.6% for any substance use disorder and 13.2% for alcohol use disorder (AUD). Between 21 and 61% of weight loss surgery (WLS), candidates suffer from a psychiatric illness, with one third having a history of substance abuse [5]. While Roux-en-Y gastric bypass (RYGB) is the most effective type of WLS [6], with concomitant improvements in mental health, a significant minority reported the reversal of mental health gains experienced in the first few years or an absence of any significant psychological benefit [7].

Among the various psychological models of overeating, there is growing interest in viewing overeating as resulting from addictive behaviour [8]; the homeostatic regulation of calorie intake is overridden in the pursuit of immediate reward when highly palatable food is available and the neurobiological processes underlying both reward sensitivity and inhibitory control are disrupted. Changes in brain function, particularly decreased basal dopamine secretion, similar to that seen in drug abuse can also be seen in chronic consumption of highly palatable foods [9]. Further support for the addictions model comes from the apparent contradiction that although there is an inverse relationship betweem BMI and alcohol consumption in the general population [10], lifetime prevalence rates of AUD are higher among individuals with morbidobesity than the general population [11]. Thus the very lowrates of alcohol consumption noted among bariatric surgery candidates [12] cannot be exclusively attributed to under-reporting. AUD following RYGB has been consistently reported since 2008 [11, 13, 14] with the largest published study [15] (reporting data from 11,115 post-bariatric surgery patients) finding a twofold increased risk of inpatient care for alcohol abuse among patients who had RYGB compared with those who had gastric banding. Identified risk factors included pre-operative AUD [13, 14], male gender [13, 16], younger age [13] and any pre-operative alcohol use [16].

The link between RYGB, specifically, and an increased risk of post-surgery AUD appears to be due to faster absorption of alcohol as well as a marked reduction in available alcohol dehydrogenase, which is mainly secreted in the now circumvented stomach [17]. Fifteen participants used as their own controls demonstrated significantly increased peak blood alcohol at 3 months and significantly higher again at 6 months post-surgery after drinking 5 oz of wine compared to pre-surgery [18]. Reduced body mass was ruled out as a causal factor suggesting that the anatomical changes alone were responsible. Only one published study [19] using a qualitative methodology addressing post-RYGB AUD has been carried out, yielding themes associated with its aetiology: unresolved psychological problems, addiction transfer, faster onset or stronger effects from substances and increased pain medication availability. However, these themes were descriptive rather than conceptual and did not offer a coherent explanatory model. The present study sought to build on this work by developing a grounded theory of post-RYGB AUD.

Qualitative research methods are particularly well suited to exploring poorly understood and complex areas of human experience as well as giving an important voice to those affected. Among various qualitative approaches that seek to understand psychological phenomena in exploring the totality of a situation, investigating the why and how of a phenomenon [20], grounded theory is particularly well suited to questions about influencing factors by generating a contextually situated theory that comes from a close inspection of the data collected in a concrete local setting [21]. Grounded theory methods broadly consist of a set of guidelines for collecting and analysing data that are both systematic and flexible.

Method

The eight participants attended the same publicly funded weight management service in the Republic of Ireland and had voiced concerns to a health professional regarding problematic alcohol use. The inclusion criterion for AUD was a score of eight or more on the Alcohol Use Disorders Identification Test (AUDIT) [22], the cut-off score used to determine problematic alcohol use. Scores ranged from 12 to 36 with a mean of 24. The four males and four females ranged in age from 30 to 67 with a mean of 48 and their pre-surgery BMI ranged from 41 to 74 kg/m2 with a mean of 55 kg/m2. Recent BMI ranged from 26 to 41 kg/m2 with a mean of 31 kg/m2. Time since bariatric surgery ranged from 3 to 12 years (mean = 8.5 years). Four participants were abstinent from alcohol at the time of the interview and were asked about their previous problematic alcohol intake when responding to items on the AUDIT; three of these participants had been treated in residential substance abuse facilities and the fourth in an outpatient addictions service, a further two completed residential substance abuse treatment but relapsed. Other patients identified with post-RYGB AUD were either unwilling to participate in the study, were uncontactable or deceased. Informed consent was obtained and ethical approval granted by the research ethics committee of the National University of Ireland, Galway.

A semi-structured individual interview was carried out. The grounded theory method used in the study required that questions should be open-ended, exploratory and designed to elicit information about the relevant actions and processes and included topics relating to what changes in participants’ lives were noticed after surgery, including personality changes. They were also asked about coping skills, drinking habits and personal relationships before and after surgery as well as what triggers or influences to drink that they could identify. In keeping with the grounded theory method, the possibility of revising the interview schedule was considered following an initial analysis of each interview. This practice is based on the assumption that the theoretical understanding of what is being studied is constantly growing and evolving, and may move in unexpected directions which may necessitate a significant shift in the direction an interview schedule may follow. Since each participant displayed such openness and frankness, very little prompting was required to elicit dense and rich data and it was not necessary to make more than very minor changes to the interview schedule.

The data was analysed using a constructivist grounded theory method, that is, a relativist and pragmatic version of grounded theory, which does not assume that theories are discovered, as in classical grounded theory, but are constructed by the researcher within a particular cultural and personal context [21]. After verbatim transcription of the interviews, each segment of data was coded. Initial codes were reviewed and assimilated into a more abstract conceptualization in a process known as theoretical coding. Core categories were then constructed as the data analysis moved from a linear or chronological process view of the development of AUD after RYGB to a more vertical explanation linking disordered eating and excess drinking to proposed underlying generative mechanisms. These core categories underpin the final grounded theory because they contain the most explanatory power and provide the most plausible account for the participants’ main concern (Table 1).

Table 1 Overview of theoretical categories and illustrative quotes
Full size table

Results

Following the grounded theory analysis, an overarching theory was developed to provide an explanatory account of the emergent themes. The theory of ‘filling the void’, which can be seen in Fig. 1, explains the pattern of excess drinking employed by the participants to manage the problem of their unmet emotional and psychological needs when they were physically prevented from comfort eating after bariatric surgery. This ‘void’ refers to both an inner void of personhood described by most of the participants as well as the void left in their coping repertoire which was previously occupied by comfort eating. Seven of the eight participants revealed significant trauma histories, as well as attachment difficulties in childhood, major losses and consequent challenges in regulating their emotions. In their distress, they found comfort in palatable foods and this pattern persisted until they had WLS and were no longer physically able to eat the same foods or in the same quantities. Self-soothing became difficult to achieve and they sought the same calming effect from alcohol. The desire to replace hedonic eating (eating as a source of sensory pleasure in the absence of an energy deficit) was another contributing factor in their initial drinking as a ‘new buzz’. Unrestrained drinking then began to mirror their previous unrestrained eating patterns.

Fig. 1
figure 1

Overview of the theory of ‘filling the void’ of unmet psychological and emotional needs and its relation to the problem of ‘unresolved psychological problems’

Full size image

In grounded theory research, the participants’ main concern is often used as a starting point in theory construction, as it is, this common issue that is at the root of the phenomenon under investigation. The main concern of ‘unresolved psychological issues’ (Fig. 2) was developed from four key categories: ‘psychological problems’, ‘trauma’, ‘loss’ and ‘internally unchanged’.

Fig. 2
figure 2

Conceptual development of participants’ main concern: ‘unresolved psychological problems’

Full size image

Managing unresolved psychological issues before WLS was achieved with the help of external coping mechanisms, that is, using things or activities outside themselves to cope with the burden of what they experience as intense or intolerable emotions. While there was some reporting of adaptive coping in the form of appropriate support seeking as well as using hobbies, interests and physical activity to reduce stress, removal of supports or additional stressors led to use of eating to cope before WLS. Eating behaviours in the form of comfort eating, binge eating and eating for pleasure were so altered after WLS that there was a gap in their coping repertoire and the ability to manage stress.

The period immediately following WLS, accounting for the hiatus before the onset of AUD and lasting up to 2 years, is referred to in our theory as the ‘honeymoon’ because of the typically elevated mood experienced by patients as they rapidly lose weight, resulting in dramatic improvements in physical health and mobility, praise and attention and increasing social acceptance after occupying a socially excluded position previously. The honeymoon ended for this group of participants due to the persistence of psychological problems or new psychosocial stressors and the realisation that they were still the same person (‘internally unchanged’). As well as needing a replacement coping strategy, there was a need for a new source of pleasure. Alcohol provided this ‘new buzz’, as it was described with enthusiasm by several participants. The rapid effect of alcohol is regarded as an enabling factor within the current theoretical model rather than the driving force behind the development of AUD. The concept of a substitute behaviour that serves the same purpose as the one it replaced is core to the theory of ‘filling the void’. The drive to fill this vacuum previously occupied by eating was briefly suspended during the honeymoon period. After the honeymoon period, alcohol served the same purpose as eating did, as a ‘new buzz’, and it helped to regulate emotion through the counteracting effects of pleasant sensations. It was also associated with a calming effect on unpleasant emotions, such as the depression, anxiety, post-traumatic stress disorder (PTSD) and low self-esteem that persisted for several participants.

The theory, ‘filling the void’, represents the conclusion of the participants’ attempts to use alcohol to meet emotional and psychological needs that they had previously met through eating. The concept of a ‘void’ has two aspects: that of an inner emptiness and that of a vacuum previously occupied by eating. Moreover, it has passive and active facets: passive in the sense that it represents a vacuum waiting to be filled and active in the sense that there are deeply felt unmet needs that motivated participants to act. These unmet needs relate to issues of attachment, security and safety as well as stimulation and pleasure. The void can also be understood as a sense of not having a firm place in the world, needing something to feel anchored to in the face of turbulence and using first eating and then alcohol to meet a range of emotional and psychological needs.

AUD treatment and recovery formed part of the participants’ narratives and included inpatient treatment (n = 5), outpatient addictions counselling (n = 1), self-help groups such as Alcoholics Anonymous (n = 3), use of oral disulfiram (Antabuse) (n = 1), disulfiram implant (n = 1), anti-depressant medication (n = 6) and anti-anxiety medication (n = 1). At the time of interview, four participants continued to drink at problematic levels.

Discussion

The theory of ‘filling the void’ describes the dynamic of managing symptoms relating to unresolved psychological issues that continued in a different form after RYGB. Participants’ main concern was how to cope with experiences that they either found overwhelming or needed the help of external coping mechanisms to manage. This theory has two main aspects. The first is that the participants used eating to control their affective state. In most cases, this was to neutralise or minimise negative affective states, and in some cases, it was also to induce positive affect. The second aspect is the behavioural replacement of eating with drinking as a coping mechanism. This theory incorporates elements of the two existing theories in the literature about post-surgery AUD and also elaborates upon them. The current theory includes the concept of ‘addiction transfer’ or ‘behavioural substitution’ previously proposed [23]. However, a second proposed theory [19], attributing post-surgery AUD to faster absorption of alcohol due to anatomical changes, is included only as an enabling factor and is not represented as the driving force behind their development of AUD. Rather, the underlying mechanism facilitating AUD in this context is psychological rather than biological. It refers to the participants’ use of external coping mechanisms to manage unresolved psychological issues.

The key finding in this study is the persistence of the use of external coping mechanisms, arising due to the non-remission of pre-existing psychological disorders. This finding is congruent with recent research findings, such as the significant association that obesity has with mood disorders and anxiety disorders (especially specific phobias, social phobias and PTSD [24]). AUD is itself strongly associated in the general population with both PTSD [25] and social anxiety [26], suggesting that the risk of AUD in the presence of a combination of severe obesity and psychopathology is greater than has been heretofore appreciated. A close relationship between a history of sexual abuse and the occurrence of obesity [27] has been reported as well, along with an association between a history of sexual abuse and AUD [28]. The level of significant trauma history among participants in this study was remarkable; this, combined with their mood and anxiety disorders, further highlights their elevated risk of AUD. Findings in the current study are similar to those reported in previous qualitative research regarding unmet expectations after weight loss surgery [29, 30], both reporting that participants expressed disappointment that weight loss by itself did not automatically lead to all of the anticipated changes, particularly among participants who used food to regulate emotions and who also referred to unaddressed psychological problems [31].

The strengths of the current study are that it is the first known qualitative European study published in English addressing this important problem and that it moves beyond mere description of participants’ experience, to the conceptualising of those experiences into a theoretical account of post-RYGB AUD. The main limitation of this study is the small number of participants. The sample size was limited by the availability of participants who met the inclusion criteria. The constructivist epistemology of the grounded theory methodology used implies that the resulting theoretical model (‘filling the void’) represents only one of many possible perspectives on the data. Nonetheless, we believe the model provides an enhanced theoretical framework to guide future research.

It is reasonable to conclude that treatment for alcohol addiction increased the level of insight noted in these participants, which may not be generalizable to all post-bariatric surgery patients with AUD who have not participated in AUD treatment. However, the generative mechanisms may be the same, regardless of the patients’ level of insight.

Future research is needed to develop interventions, both before and after WLS, designed to reduce the incidence of AUD among those identified as being at greater risk, that is, those patients with psychological disorders. All candidates should have a psychological evaluation, and the role of food for emotional regulation should form part of this assessment as well as current and past alcohol use. Further psychological work should be undertaken to develop alternative coping strategies if indicated by the psychologist. All post-bariatric surgery patients should be offered psychological input routinely, and if any member of the post-operative treatment team identifies a significant increase in alcohol intake, even when not in excess of recommended limits, a referral to a psychologist experienced with this population is indicated. All patients should be informed about the increased risk of alcohol abuse even when pre-operative alcohol use is minimal. In particular, the second post-surgical year and beyond, when the honeymoon period has ended, may be when patients are more receptive to addressing unresolved psychological issues and may forestall the onset of problematic compensatory behaviours. Post-bariatric surgery support groups using structured programmes to address eating pathology as well as promoting the use of adaptive coping skills are recommended [32].

Conclusion

This research provides an enhanced conceptual account of the development of AUD following RYGB. Participants’ main concern was identified as ‘unresolved psychological issues’. The theory of ‘filling the void’ accounts for how the participants in this study attempted to cope with their unmet psychological and emotional needs through external means of coping. Through ‘behavioural substitution’, participants managed the symptoms of their unresolved psychological issues by drinking alcohol instead of excess eating. A grounded theory methodology allowed exploration of very personal and sensitive details of participants’ lives and adds to our collective understanding of the risks bariatric surgery candidates may encounter that contribute to the development of AUD.