Lamotrigine for people with borderline personality disorder: a RCT

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 10/103/01. The contractual start date was in February 2013. The draft report began editorial review in July 2017 and was accepted for publication in November 2017. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Peter Tyrer is a member of the Health Technology Assessment (HTA) Commissioning Board. Joseph G Reilly has received project funding from the Drug Safety Research Unit as part of an unrestricted grant provided by Merck Pharmaceuticals. Alan Montgomery is part of the HTA Clinical Evaluation and Trials Board. Institutions for all authors have received funding from the National Institute for Health Research for other studies.

Copyright statement

© Queen’s Printer and Controller of HMSO 2018. This work was produced by Crawford et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2018 Queen’s Printer and Controller of HMSO

The importance of borderline personality disorder

Borderline personality disorder (BPD) is a severe mental health condition that is characterised by affective instability, recurrent suicidal behaviour and impaired interpersonal functioning. 1 It is estimated that between 0.5% and 2% of people have BPD. 2 The levels of BPD among people in contact with mental health services are far higher; as many as one-fifth of people who are admitted to inpatient mental health units in the UK have this diagnosis. 3

People with BPD are more likely to experience other mental health problems such as anxiety, depression and substance misuse. Of those who attend emergency medical services following deliberate self-harm, 1 in 10 have BPD,4 and the rate of completed suicide among people with this condition is 50 times higher than in the general population. 5 People with this condition have poor social functioning; many are socially isolated and most are unemployed or on long-term sick leave. 6 People with BPD are also more likely to experience poor physical health7 and mortality due to cardiovascular disease and other physical health problems is higher. 8 The reasons for this are unclear. Although it is possible that the high levels of emotional distress that people with BPD experience are associated with more somatic symptoms,9 it seems likely that higher levels of smoking and substance misuse are important. 10 People with BPD may neglect themselves, and problems in maintaining interpersonal relationships may make it more difficult for them to obtain the physical health care they need when unwell. 10

Treatment of borderline personality disorder

Concerns have been expressed about the quality of services for people with BPD. 11 Many people who have this diagnosis report that they are dissatisfied with the treatment they receive,12,13 and mental health practitioners often find it difficult to work with people with this condition. 14

Although psychological treatments, such as dialectical behaviour therapy and mentalisation-based therapy, have been shown to improve the mental health of people with BPD,15 most people with this disorder do not have access to specialist psychological treatment services. Among those who do, many do not engage with psychological treatment, and as many as half of those who do engage drop out before the treatment has been completed. 16 People with the most severe problems are less likely to engage successfully in psychological treatments than those with milder forms of the disorder. 16,17

No drug treatments are licensed for the treatment of BPD. Despite this, people with this condition are often prescribed large amounts of psychotropic medication. 18 Antidepressants are widely used, despite evidence that they do not improve the mental health or social functioning of people with BPD. 19 The results of randomised trials of antipsychotic medications are equivocal. Although some studies have shown short-term reductions in symptoms of anger and hostility, the longer-term effects of these drugs are not known. 19

The role of mood stabilisers

Affective instability and higher than expected levels of comorbidity with bipolar disorder among people with BPD have led to considerable interest in the role that mood stabilisers may play in improving the mental health of people with this disorder. 20 Investigation of the role of mood stabilisers was highlighted as a priority for future research in the National Institute for Health and Care Excellence (NICE) guidelines on the recognition and management of BPD. 11 Research into the effects of established mood stabilisers, such as lithium and carbamazepine, has been limited because of their toxicity in overdose, which is a not infrequent occurrence among people with this condition. 21 Another concern about the use of mood stabilisers in people with BPD is the increased incidence of birth defects among children born to women taking these drugs. 22 Most people with BPD who are in contact with mental health services are women of child-bearing age. Many women with BPD report impulsive behaviour, including unplanned and unprotected sex. Data from women who take these drugs for epilepsy have shown that levels of major congenital malformations are higher among those taking valproate than among those taking lamotrigine. 23 Concerns have also been raised about long-term cognitive impairment among children born to women taking valproate. 24

Lamotrigine is a mood stabiliser with antiepileptic and analgesic properties. 25,26 The mechanism of action of lamotrigine in patients with bipolar affective disorder is poorly understood but may relate to enhancing the action of the inhibitory neurotransmitters, including gamma-aminobutyric acid. 27

Evidence in support of the use of lamotrigine for people with BPD comes from three open-label studies and two placebo-controlled trials. 19 The two randomised controlled trials of lamotrigine for people with BPD have reported positive findings. The first trial, by Tritt et al. ,28 involved 24 women who were recruited mainly from advertisements placed in primary care practices. In comparison with those women taking the placebo, women taking up to 200 mg of lamotrigine were found to have lower levels of anger 8 weeks later. The second trial, by Reich et al. ,29 recruited 28 men and women through websites and television and radio advertisements. Those participants who were randomised to receive up to 225 mg of lamotrigine were subsequently found to have lower levels of affective instability and impulsiveness [assessed using the Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD)] 12 weeks later. A key limitation of both studies is that the degree of severity may have been lower than that seen among people with BPD who are treated in secondary care mental health services. Data on levels of global functioning at the time of the baseline assessment in the trial by Reich et al. 29 show that mean levels of impaired social function were in the ‘moderate’ range. Both studies focused on short-term effects of lamotrigine and neither examined the costs or cost-effectiveness of this intervention.

Lamotrigine is associated with a range of side effects, which include a skin rash that is reported in up to 10% of people taking the drug and a more severe cutaneous reaction, Stevens–Johnson syndrome, which is estimated to occur in < 0.1% of people taking this drug. 30 The incidence of this problem is reduced by gradual dose escalation and care with interacting agents, such as valproate and oral contraceptives. However, serious events are rare, and the drug is widely used in the UK for the treatment of people with bipolar disorder; therefore, psychiatrists are familiar with its dose titration requirements and the need for vigilance regarding severe cutaneous adverse reactions.

The LABILE study

The Lamotrigine And Borderline personality disorder: Investigating Long-term Effectiveness (LABILE) trial was designed to compare the clinical effectiveness and cost-effectiveness of lamotrigine plus usual care with an inactive placebo plus usual care over a 1-year period. The main aim of the study was to test whether or not prescribing lamotrigine in addition to usual treatment reduces symptoms of this condition, improves social functioning and quality of life, reduces the incidence of suicidal behaviour, reduces the level of alcohol and substance misuse and lowers the amount of antipsychotic and other psychotropic medication that people are prescribed. The trial also examined the cost, cost-effectiveness and cost–utility of adding lamotrigine to usual care for adults with BPD.

Design

The LABILE trial was a multicentre, two-arm, parallel-group, double-blind, placebo-controlled, individually randomised trial of lamotrigine versus placebo with 12-, 24- and 52-week follow-up assessments. The trial included an integrated clinical and economic evaluation.

Study setting

Study participants were recruited from secondary care mental health services in England including inpatient units, outpatient clinics and community mental health teams. There were six recruitment centres altogether in London (Central and North West London NHS Foundation Trust, Oxleas NHS Foundation Trust, West London Mental Health NHS Trust), the East Midlands (Nottinghamshire Healthcare NHS Foundation Trust, Derbyshire Healthcare NHS Foundation Trust) and the north-east of England (Tees, Esk and Wear Valleys NHS Foundation Trust).

Participants

To be eligible to take part in the study, potential participants had to be aged ≥ 18 years, meet the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV) diagnostic criteria for BPD and be willing and able to provide written informed consent to take part in the study. Potential participants were excluded if they:

1. had a coexisting diagnosis of bipolar affective disorder (type I and II) or psychotic disorder (schizophrenia, schizoaffective disorder or mood disorder with psychotic features)

2. were already being prescribed a mood stabiliser (lithium, carbamazepine or valproate) or had had one within the past 4 weeks

3. had a known medical history of liver or kidney impairment

4. had cognitive or language difficulties that prevented them from providing informed consent.

In addition to this, women were excluded from the study if they were pregnant, planning a pregnancy or of child-bearing age and not using adequate contraception.

Interventions

Those who were allocated to the active arm of the trial were prescribed encapsulated generic lamotrigine, titrated according to the established British National Formulary protocol30 but with the titration occurring at standardised 14-day intervals. The dose was altered for participants who were taking the combined oral contraceptive pill, which affects the metabolism of lamotrigine. For all participants, the starting dose was 25 mg per day and this was increased to 50 mg after 2 weeks, 100 mg after 4 weeks and 200 mg per day after 6 weeks. The dose was maintained at 200 mg unless the participant was taking the combined oral contraceptive pill, in which case it was further increased to 300 mg after 8 weeks and 400 mg per day after 10 weeks. However, the same dose could be prescribed again for an additional 2 weeks during titration and a lower maintenance dose utilised throughout participation when this was clinically indicated, such as when tolerability or emergent side effects were a concern.

Those who were allocated to the placebo arm of the trial were given capsules identical in appearance to those containing active lamotrigine, but backfilled with lactose monohydrate. This was prescribed in the same regime as that used in the active arm of the trial.

Trial medication was issued to patients fortnightly to cover the dose titration period, with a 17-day supply provided in case the next supply was delayed for any reason, such as a participant not attending a scheduled meeting. Once the maintenance dose was reached, further trial medication was provided either fortnightly or 4-weekly, as decided by the prescriber based on an assessment of risk of intentional overdose. In any instance when a participant had, intentionally or unintentionally, stopped taking trial medication for a period of ≥ 5 consecutive days, they were returned to a prescribed dosage of 25 mg daily and re-titrated gradually to their maintenance dose.

Usual care

Usual care in the trial comprised contact with primary care and secondary care health services, including access to psychological treatment services and inpatient admission if required. No restrictions were imposed on the use of other treatments, except that those who remained in the trial were not to be prescribed lamotrigine (aside from trial medication) or any other mood stabiliser (lithium, carbamazepine or sodium valproate).

Assessments

Blinding

All patients, carers and referring psychiatrists were blinded to treatment assignment until the participant had left the trial or until 52 weeks post randomisation (whichever was the longer). Blinding of investigators, researchers, the trial manager and the trial statistician was maintained until all data were entered, the database was locked and initial analyses of trial data were complete. The exception to this was for participants whose referring psychiatrist was also the principal investigator, in which case the allocation for that particular participant was revealed following the final assessment.

Site pharmacies were unblinded to trial arm allocation and were provided with a list of the randomisation codes and corresponding trial arm allocation for that site. The trial medication was produced with tear-off labels that identified it as being lamotrigine or placebo in a coded format, so that pharmacy staff could dispense the appropriate medication for a participant. Pharmacy procedures required that the tear-off label was removed during dispensing and added to trial documents for accountability. The need to maintain the blinding of researchers and other individuals at the site was made clear to those delegated to work on the trial within the pharmacy.

Study logistics

Data management

Data were entered onto a secure web-based database. Access was restricted by user identifiers and passwords (encrypted using a one-way encryption method). Study data will be archived securely and then safely destroyed after 15 years.

Sample size

We based the sample size calculation for the study on our primary hypothesis: for people with BPD who are in contact with mental health services, the addition of lamotrigine to usual treatment will reduce symptoms of their disorder at 52 weeks’ follow-up, according to the total score on the ZAN-BPD.

The ZAN-BPD has been used to examine the clinical effectiveness of a range of psychological and pharmacological treatments for people with BPD. In a randomised trial of a modified form of group-based cognitive behavioural therapy, Blum et al. 40 found that there were improvements in mental health and reduced use of emergency medical services among those who were randomised to problem-solving therapy. These improvements were associated with a difference of 3.6 [standard deviation (SD) 6.9] in total ZAN-BPD score. The ZAN-BPD rating scale was also used to examine the clinical effectiveness of lamotrigine for people with BPD in a randomised trial conducted by Reich et al. 29 In this small trial (n = 28), a non-statistically significant difference of 5.6 (SD 6.75) in total score on the ZAN-BPD was found at 12 weeks. Seventeen (61%) people in the trial completed all 12 weeks of the study and the levels of adherence to trial medications in those that completed the study were judged to be high.

Anticipating that levels of adherence to trial medications would be lower in the LABILE trial than in the study by Reich et al. ,29 we powered the study on the basis of a smaller difference in ZAN-BPD score of 3.0 (SD 6.75). The sample size was calculated using Stata^® version 13.1 (StataCorp LP, College Station, TX, USA).

A total of 214 participants (107 receiving lamotrigine and 107 receiving placebo) would need to be randomised to have 90% power to detect a minimal clinically relevant difference of 3.0 (SD 6.75) in total score on the ZAN-BPD at 52 weeks, using a 0.05 level of statistical significance. To take account of 15% loss to follow-up at 52 weeks, the sample size was increased to 252. However, this was further revised to 266 during the course of the trial to account for a greater loss to follow-up of 25%.

Statistical analyses

The analysis and reporting of the trial was conducted in accordance with Consolidated Standards of Reporting Trials (CONSORT) guidelines. 55 A detailed statistical analysis plan was developed and agreed with the Independent Data Monitoring and Ethics Committee, and this was finalised prior to the completion of data collection, the database lock and the unblinding of the study. Continuous variables were summarised in terms of the mean, SD, median, lower and upper quartiles, minimum, maximum and number of observations. Categorical variables were summarised in terms of frequency counts and percentages. All data were analysed using Stata, version 13.1.

Health economics analysis

The primary economic evaluation took a NHS/personal social services perspective, including only costs incurred to health and social care services, following guidance from NICE. 56 Although previous studies in people with BPD have found that health and social care are the key cost drivers in this patient group,44 it is also clear that BPD can have an impact on not only an individual’s ability to work but also their absence from work as a result of sickness. 57 Therefore, productivity losses were included in a sensitivity analysis.

Service user involvement

Plans for the study were presented at a research seminar at the British and Irish Group for the Study of Personality Disorder, which included service user representatives. These discussions, together with feedback from Fenella Lemonsky (an expert by experience and a co-applicant on the study), helped us to decide which mood stabiliser we should examine. We also used feedback from service users and results of a Delphi study of users of personality disorder services to help us decide. 58

Fenella Lemonsky remained an active member of the project management group throughout the study. Additional input from people with lived experience of using services was provided by Sally Strange and Jennie Parker. Service users reviewed materials for publicising the study and commented on a draft version of the patient information sheet.

Service users also contributed to the communication of study findings; Sally Strange and Jennie Parker helped us interpret study findings and develop our recommendations for services and for future research. Jennie Parker also commented on a draft of the lay summary of the study findings, which we have distributed to study participants, and Fenella Lemonsky presented the results of the study at the annual conference of the British and Irish Group for the Study Personality Disorders in 2017.

An independent service user, Jenny Trite, was an active member of the Trial Steering Group throughout the course of the study.

Ethics approval and governance

The trial was approved by the London-Central Research Ethics Committee (reference number 2/LO/1514). In accordance with the current revision of the Declaration of Helsinki68 (amended October 2000, with additional footnotes added in 2002 and 2004), a participant had the right to stop trial treatment and to withdraw from the trial at any time and for any reason, without prejudice to his or her future medical care by the physician or at the institution, and was not obliged to give his or her reasons for doing so. The investigator could also withdraw a participant from trial treatment at any time in the interest of the participant’s health and well-being or for administrative reasons. Trial follow-up continued after treatment was withdrawn, unless the participant withdrew consent.

All potential participants were provided with written and verbal information about the study before being asked to provide written informed consent to participate in the study.

Progress of the study was overseen by a Trial Steering Committee and a Data Monitoring and Ethics Committee.

Changes to trial design

Baseline characteristics of randomised participants

The demographic and clinical data of each trial arm are summarised in Table 2. The results of the outcome assessments at baseline are presented for each arm in Table 3. In terms of baseline comparability, the lamotrigine and placebo groups were well matched.

The ZAN-BPD is a measure of symptoms and behavioural problems and the total score ranges from 0 to 36; a lower score indicates a better outcome. The BDI measures the severity of depression symptoms and the score ranges from 0 to 63; a lower score indicates a better outcome. The SFQ score measures social functioning problems and the score ranges from 0 to 24; a lower score indicates a better outcome.

Adherence to protocol and trial medication

Details of the number of participants who completed each of the follow-up assessments are presented in Table 4, broken down by study arm. The median time between randomisation and the completion of these assessments is also presented. In total, 195 (71%) participants completed the 52-week follow-up.

Table 5 summarises the key parameters of trial medication adherence and prescribing over the course of participation for participants in each arm. In total, 93 (34%) participants completed the trial medication per protocol, and similar proportions were shown in both arms. Although 191 (69%) study participants were taking the trial medication 12 weeks after randomisation, only 107 (39%) were taking it at the end of the 1-year follow-up period. Self-rated adherence, assessed using the total score on the four-item Morisky Medication Adherence Scale at 12, 24 and 52 weeks’ follow-up (Table 6), also showed similar levels of compliance with trial medication.

Primary outcome

The total score on the ZAN-BPD decreased for study participants as a whole between baseline and 12 weeks and then remained fairly stable throughout the remainder of the follow-up. This pattern was seen in both arms. No difference was seen in adjusted mean total ZAN-BPD score at 52 weeks (Table 7).

The lack of treatment effect was supported by the results of the sensitivity analyses (Table 8), which included the following:

• using repeated measures analysis to include ZAN-BPD scores at all visits

• adjusting baseline variable with imbalance (no formal statistical testing for differences)

• multiple imputation of missing data using chained equations

• CACE analysis to investigate the impact of compliance level on treatment effect.

Secondary outcomes

Comparison of secondary outcomes at 52 weeks between those in the two treatment arms of the trial revealed no statistically significant differences (Table 9). The proportion of people who were prescribed psychotropic medication in the year following randomisation was 93% of those in the lamotrigine arm and 95% of those in the placebo arm of the trial, of whom 16% in the lamotrigine arm and 17% in the placebo arm of the trial were prescribed antipsychotic medication. We also found no evidence of clinically important differences in secondary outcomes at 12 or 26 weeks (see Appendix 4). No differences were seen on any of the subscores of the ZAN-BPD at 52 weeks. The adjusted mean difference for those prescribed lamotrigine was –0.1 (95% CI –0.9 to 0.7) on the affective disturbance subscore, 0.1 (95% CI –0.4 to 0.7) for cognitive disturbance, –0.1 (95% CI –0.6 to 0.5) for impulsivity and 0.0 (95% CI –0.5 to 0.5) for disturbed relationships compared with those in the placebo arm of the trial (see Appendix 4, Table 22).

Safety

Tables 10 and 11 show that there was an excess of AEs in the placebo group but the incidence of those classified as serious did not differ across treatment arms. There were three deaths in the trial, all of which occurred in the placebo arm. No suspected unexpected serious adverse reactions were recorded.

Using a pro forma to enquire about the presence of specific known side effects of lamotrigine revealed no difference between the groups at any assessment time points (Table 12).

Service use

All resources used by study participants over the 52-week follow-up are summarised in Table 14. There are some noticeable differences in the use of hospital services over the follow-up period; participants in the lamotrigine group had an average of 12 nights of inpatient care compared with six nights of inpatient care in the placebo group. The SDs and ranges reported alongside the means in the table suggest that this difference in mean costs was because of a small number of participants who had long stays in hospital. Outpatient appointments and accident and emergency attendances were similar between groups.

In general, the use of community health services was similar in both groups. Between 98% and 100% of participants saw their GP at least once over the period of follow-up, and, on average, the number of times a participant saw their GP was between 17 and 20. Use of other health-care services (practice nurse, walk-in clinic, dietitian, physiotherapist, podiatrist) and community mental health services (psychiatrist, psychologist, care co-ordinator/key worker, psychiatric nurse, counsellor/therapist, group therapy, day centre, drug and alcohol services, eating disorder services) was equally high.

Cost

At baseline, on average, costs were £5618 in the lamotrigine group and £3555 in the placebo group. The adjusted difference of £2376 was not statistically significant (95% CI –£108.13 to £4860.37; p = 0.061). The mean average cost differences between the groups are detailed in Table 15. The average drug and prescribing cost of lamotrigine was £242.69. The difference in hospital costs between the lamotrigine and placebo groups (£7294.09 vs. £4711.22) reflects the different average duration of inpatient stays. Other medication costs were also higher in the lamotrigine group (£674.10 vs. £302.64), resulting in higher average total costs in the lamotrigine group (£12,244.32) than in the placebo group (£8495.41), although this difference in cost was not statistically significant (95% CI –£1886.61 to £3169.83; p = 0.617).

Outcomes

The EQ-5D-3L scores at baseline and all follow-up points are detailed in Table 16. There were very few between-group differences in EQ-5D-3L scores and the resulting QALYs. The QALYs in the lamotrigine group were 0.287 and in the placebo group were 0.299, not significantly different (95% CI –0.057 to 0.034; p = 0.612).

Cost-effectiveness analysis

The ICER for the QALY outcome using adjusted mean differences is not reported. As it was negative, the lamotrigine treatment was dominated by the placebo. The ICER for ZAN-BPD using adjusted mean differences was £641.61/0.22 = £2916 per unit change in ZAN-BPD score.

The uncertainty around the ICER for QALYs is shown in Figure 2. The bootstrapped replications are present in all four quadrants of the plane: 44% appear in the less effective, more costly, quadrant; 23% in the more costly, more effective, quadrant; 21% in the less effective, less costly, quadrant; and 12% in the less costly, more effective, quadrant. The green line denotes the willingness-to-pay value of £20,000 per QALY. As there are only relatively few replications below this line, it suggests that lamotrigine treatment is not cost-effective in terms of QALYs. The cost-effectiveness acceptability curve in Figure 3 gives a clear representation of the cost-effectiveness plane. There are no willingness-to-pay values at which the probability of lamotrigine treatment being cost-effective is > 35%.

FIGURE 2.

Scatterplot on a cost-effectiveness plane of differences in costs vs. differences in QALYs.

FIGURE 3.

Cost-effectiveness acceptability curve showing the probability that lamotrigine treatment is cost-effective compared with placebo at different values that a decision-maker might be willing to pay for increases in QALYs.

The bootstrapped replications for the ZAN-BPD outcomes are shown in Figure 4. Here, the replications are mainly to the right of the y-axis, suggesting that outcomes were slightly better in the lamotrigine treatment group; however, the lack of a difference in cost between the groups is reflected in 54% of replications being in the more costly, more effective, quadrant. The resulting cost-effectiveness acceptability curve in Figure 5 can only be indicative, as we do not know the willingness-to-pay values for a unit change in ZAN-BPD score. The curve suggests that there is a probability of > 60% that lamotrigine treatment is cost-effective, but only when willingness-to-pay values for a unit change in ZAN-BPD score are greater than £1000.

FIGURE 4.

Scatterplot on a cost-effectiveness plane of differences in costs vs. differences in ZAN-BPD.

FIGURE 5.

Cost-effectiveness acceptability curve showing the probability that lamotrigine treatment is cost-effective compared with placebo at different values that a decision-maker might be willing to pay for decreases in ZAN-BPD scores.

Sensitivity analysis

The sensitivity analyses are presented in Table 17 and show that varying the economic perspective to include productivity losses and examining the impact of missing data through multiple imputation of missing cases have no impact on the finding that lamotrigine treatment is not cost-saving and thus it is unlikely to be cost-effective compared with placebo.

Strengths and weaknesses of the study

The LABILE trial is the first ever UK-based study of a medical treatment for people with BPD. Data were collected from participants receiving secondary care from NHS mental health services in six large mental health trusts in the north, south and centre of England. The study was designed to maximise internal validity. This included using independent remote randomisation to avoid unmasking researchers and adhering to an analysis plan that was finalised and shared with the Independent Data Monitoring and Ethics Committee and Trial Steering Group prior to the start of data analysis. Members of the research team drafted their conclusions and recommendations while still masked to treatments received by the two trial arms.

We used a validated measure of severity of BPD that is acceptable to patients and sensitive to change. 37,40 All researchers were trained to use the measure, and inter-rater reliability between researchers was high.

One of the main strengths of the LABILE trial is that participants were followed up over a 12-month period. BPD is a long-term condition, but previous drug trials have not conducted double-blind assessments beyond 12 weeks.

We recruited 11% more participants than we originally planned and the study was sufficiently powered to detect a minimum clinically significant difference in the severity of symptoms of BPD. We over-recruited, as it became clear that we were not going to achieve our ambitious target of following up 85% of participants to 12 months after randomisation. The 71% rate of follow-up that we achieved was very similar to that achieved in a UK study of problem-solving therapy for people with personality disorder71 and may represent a more realistic target for studies that aim to follow up people with personality disorder in community-based studies. A planned secondary analysis using multiple imputation to account for missing data found no difference between the study arms.

In this pragmatic trial, we attempted to replicate clinical practice in the NHS. However, one area in which we were unable to do this was in the means by which participants obtained their medication. Most participants typically had medication delivered to them in person or by post, rather than collecting medication from a local pharmacy as required by this study. This meant that participants had more regular contact with staff than they would have done in normal clinical practice (once every 2 weeks during titration and once a month for the majority of participants once the recommended dose was achieved). Although levels of adherence to medication were low, we believe that the additional contact that participants had with study researchers meant that the level of adherence may have been higher in the trial than would be seen in routine clinical practice.

Comparison with results of previous trials

In contrast to the results of the LABILE study, the two previous randomised trials of lamotrigine treatment for people with BPD both reported positive effects. 28,29 Both trials were smaller, had a larger number of exclusion criteria and followed up participants for a shorter period of time. There are a number of other differences between the LABILE study and these other trials, which are summarised in Table 18.

In their trial of 24 women with BPD recruited from advertisements in primary care clinics, Tritt et al. 28 reported statistically significant reductions in four out of five scales of the State–Trait Anger Expression Inventory at 8 weeks. Reich et al. 29 randomised 28 men and women with BPD who were recruited from websites and advertisements on local television and radio stations. Although a statistically significant difference was not seen in total score on the ZAN-BPD at 12 weeks, the team reported differences in two out of the four subscales of this measure (affective lability and impulsivity). 29

The primary outcome of the LABILE study was assessed at 1 year, but in exploring the reasons for differences in outcomes between these trials, we have focused on data from the 12-week follow-up assessment for comparison. We believe that a number of factors may have resulted in the differences in the results between the LABILE study and the two previous randomised trials. First, we cannot rule out the possibility that apparent differences between groups in the two earlier trials result from chance. Randomisation does not guarantee that treatment arms are balanced in small-scale trials and it is possible that differences in study outcomes resulted from confounding.

The LABILE study was designed to generate evidence that can be used by secondary care mental health services. We therefore limited our exclusion criteria and recruited people who had high levels of contact with mental health services and major impairments in social functioning. This approach meant that we were able to recruit people with the type of complex and severe problems that people with BPD who use NHS secondary care mental health services generally have. A commonly used marker of severity is whether or not people have other coexisting personality problems in addition to BPD. 33 It is of note that all but one of the people who took part in the LABILE trial met criteria for other personality disorders and had the types of complex personality-related problems that people in contact with NHS secondary services generally have. In contrast, the two previous trials of lamotrigine treatment for people with BPD excluded people who were misusing alcohol or other drugs, were actively suicidal or were receiving inpatient treatment. Therefore, there are important differences in the sample that we recruited compared with those in the two previous trials. For instance, 73% of participants in the trial conducted by Tritt et al. 28 were employed, compared with only 22% in the LABILE trial, and none of those recruited by Tritt et al. had a recent history of deliberate self-harm, compared with two-thirds of participants in the LABILE trial. It is possible that lamotrigine treatment reduces the symptoms of BPD among people who have less complex mental health problems, higher level of social functioning and lower levels of substance misuse than those we recruited to the LABILE trial.

Previous research has shown that people with more complex and severe personality problems are less likely to adhere to treatment. 17 It is possible that participants in the LABILE study were less adherent to trial medication than those in the two previous trials. Tritt et al. 28 did not report data on adherence. Reich et al. 29 did not provide detailed information about adherence but did note that 3 (20%) out of the 15 participants in the active arm of the trial were no longer taking medication at 12 weeks, compared with 86 (31%) of 276 participants in the LABILE trial who were no longer taking lamotrigine at 12 weeks. Although we cannot rule out the possibility that differences in levels of adherence are responsible for differences in outcomes between these trials, the results of the CACE analysis do not suggest that higher levels of adherence to lamotrigine result in greater improvements in mental health.

It is possible that other aspects of the design of previous trials contributed to their positive effects. For instance, in the LABILE trial we had a rigorous process for maintaining blinding through the use of a computerised system, managed by an independent team that allocated study participants. In contrast, Tritt et al. 28 used an internal group to oversee treatment allocation. Insufficient information is provided in the paper by Reich et al. 29 to establish the process they used to randomise participants.

The LABILE trial is the only one of the three studies to be publicly funded. The trial by Reich et al. 29 was funded by a manufacturer of Lamictal^® (a proprietary form of lamotrigine; GlaxoSmithKline UK Ltd, Brentford, UK) and the trial by Tritt et al. 28 was self-funded.

Finally, we are not aware of any unpublished trials of lamotrigine for people with BPD, but we cannot rule out the possibility that publication bias helps to explain the absence of previous negative trials of this treatment approach.

Implications for clinical practice

People with BPD experience high levels of emotional distress at times of crisis and may behave in an erratic or impulsive way that puts their health at risk. Emotional distress, suicidal behaviour or thoughts of suicide may lead people into contact with mental health services at such times. Interpersonal problems and negative experiences of previous contact with services often make it difficult for people with BPD to trust health-care professionals or feel reassured by their attempts to provide support and advice. In such circumstances, the prescription of medication can provide a clear signal to the patient that they are being taken seriously and health-care staff may also feel reassured that they are ‘doing something’ under difficult circumstances. Data from a national audit of prescribing practice suggest that one-quarter of people who are in contact with mental health services are being prescribed a mood stabiliser. 72 Some patients report that prescribing medication during a crisis can help bring about short-term reductions in emotional distress. The results of the LABILE study support the clinical impression that prescribing lamotrigine can lead to improvements in mental health but suggest that this benefit is no greater than would be found if a placebo were to be prescribed.

Current NICE guidelines73 for the treatment of BPD recommend that future research should be conducted to examine the impact of mood stabilisers on people with the condition, but they conclude that there was insufficient evidence to recommend any drug treatment at the time of publication. The results of the LABILE trial provide evidence to support current NICE recommendations and emphasise the need to use alternative approaches to help people with BPD cope with crises and take steps to improve their mental health. The two best-established treatments for BPD are dialectical behaviour therapy and mentalisation-based treatment. 74 More recently, evidence has begun to emerge that Systems Training for Emotional Predictability and Problem Solving (STEPPS),40 schema-focused therapy75 and day therapeutic community treatment76 can also improve the health of people with BPD. Clinicians who are working with people in secondary care mental health services should be aware of how they can help support people through mental health crises and how to access evidence-based psychological treatments when these are available.

In the LABILE trial, we took great care not to recruit women who were pregnant, wanting to become pregnant or having regular unprotected sex. Despite the assurances that potential recruits gave us, six participants subsequently became pregnant during the course of the trial. Although lamotrigine treatment has been shown to be relatively safe in pregnancy, this is not true of all mood stabilisers – notably sodium valproate. 23 In January 2015, the Medicines and Healthcare products Regulatory Agency issued a warning advising clinicians to avoid prescribing sodium valproate to women of child-bearing age when possible. 77 Despite this, a recent national audit showed that 11% of women with BPD who are in contact with secondary care mental health services are currently being prescribed this drug. 72 The data from the LABILE study showing the high level of unplanned pregnancies among women with BPD emphasise the importance of avoiding the use of unlicensed medications, such as sodium valproate, that are potentially teratogenic.

Future research

Existing evidence-based psychological treatments for BPD are lengthy, intensive and expensive. Previous research has demonstrated that, when compared with treatment as usual, these interventions are associated with reduced levels of deliberate self-harm and contact with health-care services. 74 However, studies that compare the effects of these interventions with high-quality control treatments delivered in a consistent manner show less, if any, additional benefit. 78,79 Reductions in self-harm and contact with health-care services raise the possibility that specialist psychological treatment for BPD provides an effective use of available resources, but very few data on the costs and cost-effectiveness of these treatments exist. Further research is needed to examine the costs and cost-effectiveness of specialist psychological treatment for people with BPD compared with high-quality structured general psychiatric management. Such research is needed to help policy-makers make better decisions about investing in specialist treatment services, which many patients are currently unable to access in the NHS.

Even were such services to become more widely available, a substantial minority of people with BPD would be unwilling or unable to use these group-based treatments. There is a pressing need for new research to examine alternative psychological approaches to helping people with BPD cope better with emotional crises and improve their mental health. Computer-assisted self-help interventions and training and support for friends and family of people with BPD are all areas worthy of further exploration. 80

An area of particular concern is how best to help people with severe BPD who are treated in inpatient mental health units. Following a number of open-label studies of clozapine treatment for people with BPD,81,82 the use of this atypical antipsychotic drug is becoming more widespread. Evidence from cohort studies suggests that this type of medication may help reduce impulsive self-harming behaviour and reduce the need for inpatient treatment. 82 However, this drug is also associated with potential harms, including blood dyscrasia,82 which have the potential to be fatal if not properly monitored and treated. The large number of potential benefits and harms of clozapine for the treatment of inpatients with severe BPD provides a strong case for a placebo-controlled randomised trial of this drug. Given the high rate of drop-out in this and other Phase III trials of community-based interventions for people with BPD, consideration should be given using primary outcomes that do not require direct patient contact.

Conclusions

Based on the results of this trial, we have not shown any benefits of prescribing lamotrigine for people with BPD. These results provide further support for current NICE guidelines, which state that medication should not be used specifically for BPD and consideration should be given to referring people for specialist psychological treatment when these services are available.

Contributions of authors

Mike J Crawford (Professor of Mental Health, Centre for Psychiatry, Imperial College London) was the chief investigator of the study, and contributed to the design of the study and the study protocol and the preparation of this manuscript.

Rahil Sanatinia (Research Associate, Centre for Psychiatry, Imperial College London) provided clinical leadership, supported recruitment of study participants and contributed to the preparation of this manuscript.

Barbara Barrett (Reader, King’s College London) developed plans for the economic evaluation, contributed to the design of the study and the study protocol and preparation of this manuscript.

Gillian Cunningham (Research Associate, Tees, Esk and Wear Valleys NHS Foundation Trust) recruited and followed up study participants and contributed to the preparation of this manuscript.

Oliver Dale (Consultant Psychiatrist, West London Mental Health NHS Trust) was a principal investigator and contributed to the conduct of the study and the preparation of this manuscript.

Poushali Ganguli (Research Associate, King’s College London) contributed to the management, analysis and interpretation of economic data and to the preparation of this manuscript.

Geoff Lawrence-Smith (Consultant Psychiatrist, Oxleas NHS Foundation Trust) was a principal investigator, and contributed to the conduct of the study and the preparation of this manuscript.

Verity C Leeson (Clinical Trials Manager, Centre for Psychiatry, Imperial College London) co-ordinated the trial, contributed to the development of the study protocol and preparation of this manuscript.

Fenella Lemonsky (Expert by Experience, Centre for Psychiatry, Imperial College London) oversaw service user involvement in the study and contributed to the design of the study protocol and preparation of this manuscript.

Georgia Lykomitrou-Matthews (Research Associate, University of Nottingham) recruited and followed up study participants and contributed to the writing of this manuscript.

Alan Montgomery (Professor, University of Nottingham) contributed to the development of the analysis plan for the study and the study protocol and the preparation of this manuscript.

Richard Morriss (Professor, University of Nottingham) was a principal investigator, and contributed to the design and conduct of the study and the preparation of this manuscript.

Jasna Munjiza (Consultant Psychiatrist, Central and North West London NHS Foundation Trust) contributed to the design of the study and the study protocol and preparation of this manuscript.

Carol Paton (Chief Pharmacist, Oxleas NHS Foundation Trust) was the lead pharmacist on the project, and contributed to the design of the study and the study protocol and preparation of this manuscript.

Iwona Skorodzien (Research Associate, University of Nottingham) recruited and followed up study participants and contributed to the preparation of this manuscript.

Vineet Singh (Consultant Psychiatrist, Derbyshire Healthcare NHS Foundation Trust) was a principal investigator, and contributed to the conduct of the study and the preparation of this manuscript.

Wei Tan (Statistician, University of Nottingham) contributed to the development of analysis plan for the study and the study protocol and the preparation of this manuscript.

Peter Tyrer (Professor, Centre for Psychiatry, Imperial College London) contributed to the design of the study and the study protocol and preparation of this manuscript.

Joseph G Reilly (Professor, University of Durham and Tees, Esk and Wear Valleys NHS Foundation Trust) was a principal investigator, and contributed to the design and conduct of the study and the preparation of this manuscript.

All authors read and approved the final manuscript.

Publications

Crawford MJ, McLaren T, Reilly J. Role of mood stabilisers in treatment of borderline personality disorder. BMJ 2014;349:g5378.

Crawford MJ, Sanatinia R, Barrett B, Byford S, Cunningham G, Gakhal K, et al. Lamotrigine versus inert placebo in the treatment of borderline personality disorder: study protocol for a randomized controlled trial. Trials 2015;16:308.

Crawford MJ, Sanatinia R, Barrett B, Cunningham G, Dale O, Ganguli P, Lawrence-Smith G, et al. The clinical effectiveness and cost-effectiveness of lamotrigine in borderline personality disorder: a randomized placebo-controlled trial [published online ahead of print April 6 2018]. Am J Psychiatry 2018. https://doi.org/10.1176/appi.ajp.2018.17091006

Data sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to anonymised data may be granted following review.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it is important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care.