Tolerability of Antiseizure Medications in Individuals With Newly Diagnosed Epilepsy

Key Points

Question  Has the overall tolerability of antiseizure medications (ASMs) improved after the introduction of more than a dozen second-generation ASMs in the past 3 decades?

Findings  In this 30-year longitudinal cohort study of 1795 individuals with newly diagnosed and treated epilepsy, the intolerable adverse effect rate observed was not improved despite the increased using of second-generations ASMs.

Meaning  Efficacy and tolerability are 2 sides of the same coin in achieving successful epilepsy treatment; to improve the overall effectiveness of ASM treatment, future drug development should not only focus on seizure control, but also tolerability and safety profile.

Importance  Tolerability is a key determinant of the effectiveness of epilepsy treatment. It is important to evaluate whether the overall tolerability has improved.

Objective  To identify factors associated with poor tolerability of antiseizure medications (ASMs) and examine temporal changes in tolerability.

Design, Setting, and Participants  This was a longitudinal cohort study at a specialist clinic in Glasgow, Scotland. Patients with newly diagnosed and treated epilepsy between July 1982 and October 2012 were included from 2282 eligible individuals. They were followed up until April 2016 or death. Data analysis was completed in August 2019.

Exposures  Antiseizure medications.

Main Outcomes and Measures  Univariable and multivariable survival analyses were performed to examine associations between potential risk factors and development of intolerable adverse effects (AEs). Intolerable AE rates of the ASMs as the initial monotherapy were compared between 3 epochs (July 1982-June 1992, July 1992-June 2002, and July 2002-April 2016).

Results  Of 1795 patients, 969 (54.0%) were male, and the median (interquartile range) age was 33 (21-50) years. A total of 3241 ASMs were prescribed during the period, of which 504 (15.6%) were discontinued within 6 months owing to intolerable AEs. Children younger than 18 years had lower intolerable AE rates than adults (vs aged 18-64 years: adjusted hazard ratio [aHR], 1.58; 95% CI, 1.07-2.32; vs aged ≥65 years: aHR, 1.90; 95% CI, 1.19-3.02) while female individuals (aHR, 1.60; 95% CI, 1.30-1.96) and those who had more than 5 pretreatment seizures (aHR, 1.24; 95% CI, 1.03-1.49) were associated with having higher risk. For each ASM trial, the risk of intolerable AEs increased with the number of previous drug withdrawals due to AEs (aHR, 1.18; 95% CI, 1.09-1.28) and the number of concomitant ASMs (aHR, 1.31; 95% CI, 1.04-1.64). The proportion of second-generation ASMs prescribed as the initial monotherapy increased from 22.3% (33 of 148) in the first epoch to 68.7% (645 of 939) in the last (P < .001). Although differences in intolerable AE rates and types of AEs were found between the ASMs, there was no difference in the overall intolerable AEs rates to the initial monotherapy across the 3 epochs (first: 10.1% [15 of 148]; second: 13.8% [98 of 708]; third: 14.0% [131 of 939]; P = .41).

Conclusions and Relevance  In this cohort study, the increased use of the second-generation ASMs had not improved overall treatment tolerability. Greater effort to improve tolerability in ASM development is needed.

Antiseizure medications (ASMs) are the mainstay of treatment for people with newly diagnosed epilepsy. The introduction of more than a dozen second-generation ASMs with different mechanisms of action throughout the past 3 decades has expanded the range of treatment options for this patient population.¹ Despite the increased use of these medications, long-term seizure control has not fundamentally improved.²

However, the overall effectiveness of ASM treatment is determined not only by its efficacy, but also safety and tolerability,³ which have rarely been specifically studied. Antiseizure medications are associated with a range of adverse effects (AEs), including central nervous system problems, idiosyncratic reactions, neurocognitive and psychiatric symptoms, and long-term complications. It has been reported that up to 88% of people taking ASMs experience at least 1 AE,⁴ imposing substantial negative effect on their quality of life.⁵ Adverse effects are a common cause of early treatment discontinuation and a barrier to optimizing dosage for seizure control.⁶

Some of the second-generation ASMs have demonstrated similar efficacy but better tolerability than the first-generation agents in individual comparative studies,^7-10 raising the hope that the former might bring improved overall effectiveness to epilepsy treatment. In the present study, we identified factors associated with ASM tolerability and examined whether the tolerability of ASM treatment has changed in everyday practice throughout the past 30 years. The findings have implication for ASM development strategy.

As previously described,² the study population included 1795 people in whom epilepsy was newly diagnosed and the first ASM prescribed at the Epilepsy Unit of the Western Infirmary in Glasgow, Scotland, between July 1, 1982, and October 31, 2012. For this analysis, all individuals were followed up for a minimum of 1 year until April 30, 2016, or death. This study protocol was ruled exempt by the institutional review board of Western Infirmary in Glasgow, Scotland. Patient consent was waived because all data were deidentified prior to analysis.

Details of initial clinical assessment, ASM treatment, and follow-up protocols have been described previously.^2,11,12 Antiseizure medications were selected for an individual based on seizure type and other drug and personal factors.¹³ Titration schedule¹⁴ and doses were adjusted based on seizure control and any emergence of AEs. Individuals were reviewed every 2 to 6 weeks for the first 6 months after commencing treatment, then every 4 months thereafter. At each clinic visit, any AEs were recorded and their association with the ASM was determined by the clinic physician, who documented whether any drug withdrawal was mainly due to the AEs or other reasons.

Adverse effects were categorized into 27 classes according to the Medical Dictionary for Regulatory Activities.^15,16 These categories included nervous system disorders (eg, dizziness, headache, and memory impairment), psychiatric disorders (eg, drowsiness and mood and behavioral changes), general disorders (eg, tiredness), skin and subcutaneous disorders (eg, rash and hair loss), gastrointestinal tract disorders, and weight gain.^17-19 Adverse effects were regarded as intolerable if they were stated as the main reason of discontinuation within 180 days of commencement of the ASM. This was chosen because it is the typical duration for assessment of primary outcome in monotherapy trials of ASM.²⁰

The ASMs prescribed that were developed before 1980 were considered to be first-generation ASMs, and those introduced after were classified as second-generation ASMs (eTable 1 in the Supplement).²¹ An ASM regimen was defined as either a single drug (monotherapy) or a combination of 2 or more drugs.² The first ASM regimen was always monotherapy. Drug load for each ASM was calculated as a ratio of the World Health Organization–defined daily dose (eTable 1 in the Supplement)²² at the time of withdrawal due to AE or at the last follow-up. For individuals receiving more than 1 ASM, the defined daily dose ratios for all the ASMs were summated.

To assess the change in tolerability of the initial ASM treatment over time, the study period was divided into 3 epochs according to the ASM start date, ie, July 1, 1982, to June 30, 1992; July 1, 1992, to June 30, 2002; and July 1, 2002, to April 30, 2016.

Analysis began November 2018. Pearson χ² test was performed to assess the associations between proportions of second-generation ASMs prescribed as the initial monotherapy and the 3 epochs. Fisher-Freeman-Halton exact test was used to assess the associations between rates of individual types of AEs to the initial monotherapy and the 3 epochs. Mann-Whitney test was used for comparing summated defined daily dose ratios for tolerated and intolerable ASM therapies. Univariable Andersen-Gill model for ordered multiple failure events was used to identify clinically relevant potential risk factors associated with intolerable AEs. Variables with univariable P value less than .20 were selected for the multivariable Andersen-Gill models.²³ Kaplan-Meier survivor function was used to estimate crude intolerable AE rates to the initial ASM monotherapy in the 3 epochs and adjusted for the relevant risk factors identified in the previous model.

Subgroup analysis was performed in individual ASMs used as the initial monotherapy in at least 50 patients. This provided 80% power for detecting a medium effect size (hazard ratio [HR], 1.78) in log-rank test using the Freedman method.²⁴ Pairwise comparisons of intolerable AE rates between individual ASMs were performed for AE types reported in at least 50 initial monotherapies using univariable log-rank test. Cox regression with adjustments of risk factors identified in the previous models was used. Further compete risk analysis was performed using the Fine-Gray model²⁵ to compare individual intolerable AE types between these initial ASM monotherapies included in the previous analysis.

Statistical significance level was set at P value less than .05. Holm-Bonferroni method²⁶ was applied to correct for multiple comparisons. All statistical tests were performed by using Stata version 15 (StataCorp).

Among the 1795 individuals included, 969 (54.0%) were male, 1409 (78.5%) were classified as having focal epilepsy, and 386 (21.5%) were classified as having generalized epilepsy. Their median age at treatment initiation was 33 (range, 9-93; interquartile range [IQR], 21-50) years, 259 (14.4%) were younger than 18 years, 1335 (74.4%) were adults aged 18 to 64 years, and 201 (11.2%) were 65 years or older. The median (IQR) follow-up duration was 12.7 (8.1-17.4) years. A total of 3241 ASM regimens were prescribed, including 2346 (72.4%) as monotherapies and 895 (27.6%) as combinations. The proportion of second-generation ASMs prescribed increased from the first 10-year epoch to the last epoch (eTable 2 in the Supplement).

Of 3241 ASMs, 646 (19.9%) were discontinued within the first 180 days of commencement in 436 of 1795 individuals (24.3%). Adverse effects were attributed as the reason of discontinuation of 504 ASMs (15.6%) in 364 individuals (20.3%). A total of 133 ASMs were discontinued primarily owing to lack of efficacy and 9 for reasons unrelated to efficacy or tolerability. The summated defined daily dose ratio for intolerable ASM regimens (median [IQR], 0.67 [0.40-1.33]) was significantly lower than the tolerated regimens (median [IQR], 1.00 [0.67-1.67]; P < .001). Nervous system disorders were the most common type of intolerable AEs, accounting for 35.3% (178 of 504) of all withdrawals. Psychiatric, general (eg, tiredness and ataxia), skin and subcutaneous tissue, and gastrointestinal tract disorders also frequently led to ASM withdrawal (eTable 3 in the Supplement).

Univariable screening demonstrated the following clinically relevant variables had P values less than .20 for association with intolerable AEs (eTable 4 in the Supplement): age at starting the ASM, sex, epilepsy type, pretreatment seizure number, history of alcohol abuse, number of concomitant ASMs, number of previous intolerable AEs, number of previous ASMs failed owing to inadequate seizure control, and whether the ASM commenced was a first- or second-generation ASM.

Including these factors in the multivariable analysis (Table 1), female individuals had higher intolerable AE rates than male individuals (adjusted HR [aHR], 1.60; 95% CI, 1.30-1.96; P < .001). Adults and elderly individuals had significantly higher intolerable AE rates compared with children (aHR, 1.58; 95% CI, 1.07-2.32; P = .02 and aHR, 1.90; 95% CI, 1.19-3.02; P = .007, respectively), but the rate was not different between elderly individuals and adults (aHR, 1.20; 95% CI, 0.91-1.60; P = .20).

Table 2 shows the AE rates with successive ASM regimens. Individuals who had previously stopped using more ASMs owing to intolerable AEs had a higher rate of developing intolerable AEs to the current ASM (aHR, 1.18; 95% CI, 1.09-1.28; P < .001). The rate of withdrawal due to AEs increased for each additional concomitant ASM (aHR, 1.31; 95% CI, 1.04-1.64; P = .02). Individuals who had more than 5 seizures before commencing ASM therapy demonstrated significantly higher intolerable AE rates than those who had 5 or fewer seizures (aHR, 1.24; 95% CI, 1.03-1.49; P = .02). Epilepsy type, history of alcohol abuse, and whether the current ASM was a first- or second-generation ASM were not associated with the probability of withdrawal owing to AEs (Table 1).

The proportion of second-generation ASMs prescribed as the initial monotherapy increased from 22.3% (33 of 148) in the first epoch to 41.4% (293 of 708) in the second and 68.7% (648 of 939) in the last epoch (P < .001; Cramer v = 0.32; 95% CI, 0.28-0.36) (eTable 5 in the Supplement). The crude probabilities (Figure 1) and crude rates of patients developing intolerable AEs to the initial ASM in the 3 epochs were similar (first epoch: 15 of 148 [10.1%]; second epoch: 98 of 708 [13.8%]; third epoch: 131 of 939 [14.0%]; P = .41; Cramer v = 0.03; 95% CI, 0.002-0.058). Analysis after adjustment of the relevant risk factors for intolerable AEs (sex, age at starting the ASM, and pretreatment seizure number) confirmed no significant difference in the probability of withdrawing the initial ASM owing to AEs between the patients who commenced treatment in the 3 epochs (second vs first: aHR, 1.39; 95% CI, 0.80-2.41; P = .24; third vs first: aHR, 1.42; 95% CI, 0.82-2.44; P = .21; third vs second: aHR, 0.98; 95% CI, 0.75-1.28; P = .88). Separate subanalyses for first- and second-generation ASM monotherapies showed no difference in intolerable AEs rates between the 3 epochs (eTable 6 in the Supplement).

Analysis of the specific types of intolerable AEs (Table 3) showed that skin and subcutaneous tissue AE rate reduced from 8.11% (12 of 148) in the first epoch to 4.24% (30 of 708) in the second and 4.05% (38 of 939) in the third epochs, although the difference was not statistically significant (corrected P = .12 and P = .11, respectively). The rate of intolerable nervous system AEs in the first epoch (1 of 148 [0.68%]) was significantly lower than the second (37 of 708 [5.23%]; corrected P = .04) and third (43 of 939 [4.58%]; corrected P = .04) epochs (eTable 7 in the Supplement). The rate of withdrawal due to psychiatric AEs in the last epoch (39 [4.15%]) was significantly higher than the first (0; corrected P = .01) and second (11 [1.55%]; corrected P = .006) epochs. The intolerable gastrointestinal AE rate in the last epoch (13 [1.38%]) was significantly lower than the second (23 [3.25%]; corrected P = .048). The rates of other AEs leading to drug withdrawal did not differ between the 3 epochs.

During the study period, lamotrigine (551 [30.7%]), valproate (489 [27.2%]), carbamazepine (323 [18.0%]), levetiracetam (215 [12.0%]), oxcarbazepine (74 [4.12%]), and topiramate (63 [3.51%]) were each prescribed to at least 50 patients as the initial monotherapy. eTable 8 in the Supplement shows the pairwise comparison of crude intolerable AE rates between individual ASMs. As shown in Figure 2, the intolerable AE rate of oxcarbazepine was significantly higher than lamotrigine (aHR, 2.48; 95% CI, 1.41-4.35; P = .002), valproate (aHR, 2.02; 95% CI, 1.16-3.50; P = .01), and levetiracetam (aHR, 1.93; 95% CI, 1.04-3.58; P = .04) but was similar to carbamazepine (aHR, 1.70; 95% CI, 0.96-3.02; P = .07) and topiramate (aHR, 1.26; 95% CI, 0.61-2.58; P = .53). Topiramate also demonstrated significantly higher intolerable AE rates than lamotrigine (aHR, 1.97; 95% CI, 1.10-3.52; P = .02), but the rate was similar to than other ASMs. Carbamazepine, levetiracetam, lamotrigine, and valproate had similar intolerable AE rates.

Nervous system disorders (81 [4.72%]), skin and subcutaneous tissue disorders (80 [4.66%]), general disorders (54 [3.15%]), and psychiatric disorders (50 [2.92%]) were each reported to lead to withdrawal of these 6 ASMs as initial monotherapy in at least 50 patients. As shown in eTable 9a in the Supplement, patients were least likely to withdraw lamotrigine owing to nervous system disorder and were more likely to withdraw topiramate owing to this AE type than carbamazepine and valproate. Carbamazepine, lamotrigine, and oxcarbazepine were associated with higher risk of intolerable skin and subcutaneous disorders than levetiracetam and valproate (eTable 9b in the Supplement). No patient withdrew topiramate owing to this AE. Lamotrigine was associated with lower risk of intolerable general disorders (eg, tiredness) than levetiracetam and topiramate (eTable 9c in the Supplement). Patients were more likely to withdraw levetiracetam, oxcarbazepine, or topiramate than valproate, carbamazepine, or lamotrigine owing to psychiatric disorders (eTable 9d in the Supplement).

In this cohort of people with newly diagnosed epilepsy followed up longitudinally since commencing ASM therapy, nearly 1 in 6 ASMs prescribed was withdrawn owing to intolerable AEs, often at low doses. Female individuals, those who had higher number of pretreatment seizures, higher number of previous ASMs withdrawn owing to AEs, and higher number of concomitant ASMs were more likely to discontinue treatment owing to AEs. The association between female individuals and higher AE rate to ASMs has been reported in previous studies of people with epilepsy²⁷ and is also observed in other medication classes.^28,29 Potential explanations include sex differences in health-seeking behavior and effect of ASMs on endocrine functions and sex steroids,³⁰ which may in turn lead to various metabolic, cognitive, and mood disorders. Unfavorable pharmacokinetic and pharmacodynamic interactions likely explain the higher intolerable AE rate associated with higher number of concomitant ASMs.³¹ Previous withdrawal due to AEs as a risk factor for future discontinuation may point to individual-specific susceptibility, including genetic, that predisposes to the development of AEs to ASMs, many of which share similar molecular targets.³² Higher pretreatment seizure number may be a marker of epilepsy severity, requiring more aggressive or combination therapy for seizure control, thereby increasing the risk of AEs. Children were less likely to withdraw treatment owing to AEs than adults. Differences in AE profile between children and adults may be attributed to differing underlying diseases, comedications, pharmacodynamics, pharmacokinetics, and developmental issues,^33,34 although the possibility that children were less likely to self-report AEs could not be excluded.

Despite increased use of the second-generation ASMs as the initial monotherapy, overall treatment tolerability has not changed in the past 30 years. This is likely because some AEs are shared by both generations of ASMs. For instance, the reduced rates of skin and subcutaneous AEs in the second and last epochs were likely due to the decreased use of carbamazepine, but the reduction was somewhat neutralized by the increased use of other second-generation ASMs, which can also induce these AEs, such as lamotrigine and oxcarbazepine.^7,8,35 This is supported by the individual drug analysis, which showed that oxcarbazepine was associated with similar rate and lamotrigine with even higher rates of withdrawal owing to skin and subcutaneous AEs compared with carbamazepine. However, lamotrigine was less likely to be withdrawn due to nervous system symptoms than the older agents. Further, while the second-generation ASMs may be less likely to cause certain types of AEs compared with the first-generation drugs, they are associated with other AEs. For example, the higher psychiatric AE rate in the last epoch might be attributed to the increased use of levetiracetam,³⁵ which was associated with higher risk of this AE type, yet it had lower rate of skin AEs.

These differences in AE profiles were reflected in the analysis of overall tolerability, which showed that lamotrigine was better tolerated than oxcarbazepine and topiramate and might be better tolerated than carbamazepine although the difference was not statistically significant after adjusting for other clinical factors. Levetiracetam had similar tolerability to other ASMs. On the other hand, valproate was overall better tolerated than oxcarbazepine and demonstrated similar tolerability to other ASMs. These observations were in line with those reported in previous randomized clinical trials^7,8,10,36 and with the recent treatment guidelines.^37,38

In our previous 4 sets of consecutive analyses^2,11,39,40 in this expanding population of adolescents and adults with newly diagnosed epilepsy, despite substantial increased use of second-generation ASMs, we observed no fundamental improvement in treatment efficacy, with more than one-third of patients remaining uncontrolled.² The finding of lack of improvement in overall tolerability in the present study complements our previous observations in efficacy and may explain, at least in part, why the second-generation ASMs have not been shown to improve the overall seizure outcome in people with new-onset epilepsy.

This observation also has important implications for ASM development strategy. Regulatory monotherapy trials of ASMs have typically compared the proportion of patients remaining seizure-free for 6 consecutive months as the primary outcome.²⁰ Individuals who withdrew treatment owing to AEs would not be able to reach this end point even if they did not have seizures. Our findings of early occurrence of intolerable AEs at low doses in a substantial proportion of patients suggest that effort to improve tolerability of ASMs might also improve this primary outcome. For instance, a recent noninferiority randomized clinical trial²⁰ comparing lacosamide and carbamazepine (controlled release) in people with newly diagnosed epilepsy demonstrated an absolute 5.0% (11% vs 16%) difference in intolerable AE rate and an absolute 4.1% (73.6% vs 69.7%) difference in 6-month seizure-free rate, both in favor of lacosamide. Theoretically, if a future ASM can be perfectly tolerated while having the same efficacy, it is possible for it to show a higher absolute seizure-free rate sufficient to demonstrate superiority in efficacy to its comparator. Hence, greater effort should be made to reduce AEs in ASM development, targeting those that commonly lead to drug withdrawal including nervous system, psychiatric, and skin and subcutaneous tissue AEs.

This study has limitations. First, the uptake of the second-generation ASMs varies widely between different countries and centers. As a single-center study, our findings may not be generalizable to all settings. Routine recording of AEs in our unit relied on reports from patient and clinician assessment rather than use of screening instruments; thus, milder AEs might have been overlooked. Instead of all AEs, we focused the analysis on those that led to drug discontinuation, which is arguably the most clinically relevant. The main reason of continuation was determined based on the treating clinician’s judgment. For robustness, we analyzed AEs that led to withdrawal within 180 days of commencement of the ASMs, although this could have resulted in underestimation of long-term AEs. Owing to the lack of reliable obstetric record, teratogenicity was not analyzed. As an observational study, treatment plans were determined based on the patient’s characteristics, which might lead to differences in response to individual ASMs. Lastly, we could not exclude the possibility that as more ASMs are made available, it may have lowered the threshold for clinicians and patients to switch drugs owing to AEs compared with 30 years ago. Therefore, when analyzing any change in the rates of intolerable AEs over time, we included only the first monotherapy.

Efficacy and tolerability are 2 sides of the same coin in achieving successful epilepsy treatment. Despite increased used of the second-generation agents, overall treatment tolerability has not improved over the past 3 decades. However, certain advantages of the second-generation ASMs over their older counterparts should not be overlooked. These include a wider therapeutic index, generally lower potential for drug interactions,⁴¹ and lower risk for long-term complications (including osteoporosis, sexual dysfunction, and vascular disease) through enzyme induction.^42-44 Importantly, some of the newer ASMs such as lamotrigine and levetiracetam have been shown to have lower teratogenic risk than valproate,⁴⁵ offering safer choices for women of childbearing potential. To improve the overall effectiveness of ASM treatment, future drug development should not only focus on seizure control, but also tolerability and safety profile. Current strategies to improve tolerability of ASMs under investigation include the use of intracerebroventricular drug delivery to avoid systemic toxicity (NCT02899611), on-demand drug delivery to epileptogenic focus to reduce AEs from continuous administration,⁴⁶ and design of safer drugs through better understanding of the molecular mechanisms of AEs.

Corresponding Authors: Patrick Kwan, FRACP, PhD, Department of Neuroscience, Central Clinical School, Monash University, Level 6, The Alfred Centre, 99 Commercial Rd, Melbourne, VIC 3004, Australia (patrick.kwan@monash.edu); Martin J. Brodie, MD, Epilepsy Unit, Scottish Epilepsy Initiative, 11 Somerset Place, Glasgow G3 7JT, Scotland (martin.brodie@glasgow.ac.uk).

Accepted for Publication: December 30, 2019.

Published Online: February 24, 2020. doi:10.1001/jamaneurol.2020.0032

Author Contributions: Drs Kwan and Brodie had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Walters, Kwan.

Acquisition, analysis, or interpretation of data: Alsfouk, Brodie, Kwan, Chen.

Drafting of the manuscript: Kwan, Chen.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Chen.

Obtained funding: Kwan.

Administrative, technical, or material support: Alsfouk.

Supervision: Brodie, Walters, Kwan.

Conflict of Interest Disclosures: Dr Brodie serves on the scientific advisory boards of Eisai, UCB Pharma, GlaxoSmithKline, Lundbeck, Bial, GW Pharmaceuticals, and Takeda; is on the speakers’ bureau for Eisai, UCB Pharma, GlaxoSmithKline, Lundbeck, Sanofi Aventis, and Abbott Laboratories; and has accepted travel grants for scientific meetings from Eisai, UCB Pharma, and Lundbeck. Dr Kwan reports grants from the National Health and Medical Research Council of Australia, the Australian Research Council, the US National Institutes of Health, Hong Kong Research Grants Council, Innovation and Technology Fund, Health and Medical Research Fund, Biscayne Pharmaceuticals, Eisai, GW Pharmaceuticals, LivaNova, Novartis, UCB Pharma, and Zynerba outside the submitted work; speaker fees from Eisai, LivaNova, and UCB Pharma outside the submitted work; and is supported by the Medical Research Future Fund Practitioner Fellowship (MRF1136427). Dr Chen reports grants from UCB Pharma outside the submitted work; is supported by the National Health and Medical Research Council of Australia Early Career Fellowship (GNT1156444); and has received research grants from University of Melbourne Early Career Researcher Grant Scheme. No other disclosures were reported.