An extensive description of the MOTTO study design has been published previously [13]. In brief, from day 5 after LT patients received twice-daily, immediate-release (IR) tacrolimus. After achieving stable tacrolimus trough levels between 8–10 μg/L, patients were randomized in a 1:1 ratio to either a LCP-tacrolimus regimen or an ER-tacrolimus regimen. To prevent for to toxicity (renal insufficiency or tremors), rejection or to prevent recurrence of hepatocellular carcinoma: in the LCP-tacrolimus group 19 patients used a combination regimen of LCP-tacrolimus with mycophenolic acid and 2 LT recipients to LCP-tacrolimus and sirolimus. In the ER-tacrolimus group 22 LT recipients used to a regimen of ER-tacrolimus and mycophenolic acid. The study was performed at two centers in the Netherlands: The Erasmus MC, University Medical Center Rotterdam, Rotterdam, Netherlands and Leiden University Medical Center, Leiden, Netherlands. The study was approved by the institutional Ethical Committees of these institutions, registered in the EudraCT database (EudraCT: 2018-002856-34) and conducted in accordance with the latest version of the declaration of Helsinki. The inclusion period ran from April 2019 until October 2021.

The evaluation of the severity of tremors and the HRQoL comprised a pre-defined secondary objective of the MOTTO study. The MOTTO study was initially designed to investigate whether LCP-tacrolimus compared to ER-tacrolimus results in a difference in the prevalence of post-transplant diabetes mellitus, new onset hypertension and chronic kidney disease at 12 months after transplantation.

HRQoL was assessed with the validated Dutch version of the EQ-5D-5L questionnaire and the SF-36 questionnaire (two generic HRQoL instruments) and the quality of life in essential tremor (QUEST) questionnaire (a domain specific HRQoL instrument). The questionnaires were distributed at the day of randomization, month 3, 6, and 12.

The EQ-5D-5L questionnaire is based on a descriptive system that defines health in terms of 5 states: Mobility, Self-Care, Usual Activities, Pain/Discomfort, and Anxiety/Depression [14]. Each dimension has 5 response categories corresponding to no problems, slight problems, moderate problems, severe problems, and extreme problems. EQ-5D-5L scores were transformed to societal values based on the Dutch tariff for the EQ-5D-5L established by Versteegh et al. [15].

In the EQ-5D-5L questionnaire, the respondents’ overall health on the day of the interview (patient’s self-rated HRQoL scores) was rated on a 0–100 hash-marked, vertical visual analogue scale (EQ-VAS). The threshold for the minimally important difference (MID), indicating a clinical meaningful improvement, in the EQ-VAS score was defined as ≥7 points [16].

The SF-36 questionnaire contains 36 items grouped in eight domains: physical functioning, role limitation-physical, pain, general health, energy/fatigue, social functioning, emotional wellbeing, role limitation-emotional. Each domain is scored between 0 and 100 points, with higher scores indicating better HRQoL.

The QUEST questionnaire is a self-administered questionnaire with 30 items on a five-point scale (0–4), corresponding to the frequency (never, rarely, sometimes, frequently, always) with which tremor is perceived to currently impact five domains: physical, psychosocial, communication, hobbies/leisure and work/finance [17, 18]. The score on each domain is expressed as a percentage of the total score possible on that domain, with a higher score indicating greater dissatisfaction with that domain of QoL. A total score was computed by calculating the mean of the five domain scores.

Given that the QUEST is “domain specific” for “patients with essential tremors,” this questionnaire is most likely more sensitive than the generic EQ-5D-5L and SF-36. The value of those two questionnaires is the ability to formulate “values” of quality of life for cost effectiveness analysis, and these generic questionnaires can measure side effects outside the measuring domain of the QUEST.

In this study treating physicians were allowed to apply the current common practice in order to manage the severity of tremors. This includes either reduce the dose of tacrolimus while maintaining the LT recipient on monotherapy tacrolimus or start combination therapy of tacrolimus and another immunosuppressive agent. No comedication to actively treat tremors such as beta-blockers or anticonvulsants were allowed to start for the treatment of tremors.

Variables collected included recipient socio-demographic, clinical and transplantation parameters, the HRQoL and tremor severity and trough levels tacrolimus.

The HRQoL analysis included all patients within the MOTTO study who responded to at least one questionnaire. The EQ-5D-5L, SF-36 and QUEST questionnaire included in the analysis missed <5% based on the total number of measurements across all patients and questions. The missing data were considered as missing completely at random.

Two generalized linear mixed effect models were fitted to examine the HRQoL (EQ-VAS and the societal values of the EQ-5D-5L) over the course of the study. The models included covariates shown or suggested to be relevant: time since transplantation, study group, tacrolimus trough concentrations, kidney function, hemoglobin, recipient age and sex, primary disease, diabetes mellitus and hypertension pre-transplantation as well as the interaction between visit and the study group. Participant specific random intercepts were included to account for correlation among repeated measurement nested within each participant. Natural cubic splines were used to model the potentially nonlinear trajectories of the EQ-VAS and societal values of the EQ-5D-5L over time. The need for these splines was evaluated using likelihood-ratio tests. Splines provide a convenient non-parametric way to flexibly model (potentially) non-linear associations in regression models. Instead of using one polynomial (e.g., a quadratic or cubic function) that spreads over the whole range of the covariate, splines use a set of several polynomial functions that are defined over smaller intervals. This allows the resulting fit to be more flexible and less influenced by outliers than when using a single polynomial. To visualize the estimated associations, the expected HRQoL across the course of the study was calculated while fixing the values of all other covariates to the median or reference category.

Secondary endpoints were analyzed using the Pearson’s Chi-square test or Mann-Whitney U Test. Confidence intervals for binomial proportions were calculated using the binconf package for R software. For all statistical tests, a (two-sided) p-value of <0.05 was considered to indicate statistical significance.

All data were collected in CastorEDC and analysis was conducted with R software (version 4.2.1) [19, 20].

A total of 108 LT recipients was included and randomized. No LT recipient included was diagnosed with a neurological movement disorder pre-transplantation. At baseline, 100% of the LT recipients responded to the EQ-5D-5L, SF-36 and QUEST questionnaires. The response rate decreased during follow up to a minimum of 75.5% at the end of the study (Figure 1).

Table 1 shows the baseline characteristics. No relevant differences in the baseline characteristics for the EQ-5D-5L questionnaire existed between the two regimens. However, more LT recipients randomized to the LCP-tacrolimus regimen experienced tremors compared to the LT recipients randomized to the ER-tacrolimus regimen [30.2% (16/53), 95%-confidence interval (CI) 19.9%–44.3% versus 19.2% (10/52), 95%-CI 10.8%–31.9%]. The mean tacrolimus trough level at the day of randomization to the LCP-tacrolimus regimen was 7.5 ± 3.3 μg/L and in the ER-tacrolimus regimen 6.9 ± 3.1 μg/L, p = 0.38. LT recipients in the LCP-tacrolimus regimen were converted to that formulation after 11 days (IQR: 9.25–15.25 days) and in the ER-tacrolimus regimen, LT recipients were converted to that formulation after 13.5 days (IQR: 9–15.75 days).

Figure 2 shows the proportion of LT recipients experiencing tremors during study follow up. Supplementary Table S1 shows the QUEST questionnaire results and the tacrolimus levels of the LT recipients during the study. No statistically significant differences between the two regimens were found at 3, 6 and 12 months in the frequency and severity of tremors. At 12 months 25% [10/40], 95%-CI 14.2%–40.2% of the LT recipients in the LCP-tacrolimus regimen versus 30.4% [14/46], 95%-CI 19.1%–44.8% of the LT recipients in the ER-tacrolimus regimen experienced tremors; risk difference: 0.054; 95%CI -0.151–0.249; p = 0.63.

The mean tacrolimus trough level at 12 months in the LCP-tacrolimus regimen was statistically significantly higher compared to the ER-tacrolimus regimen: 7.6 ± 3.1 μg/L versus 6.3 ± 2.2 μg/L, p = 0.026. No statistically significant differences were observed in any of the five domains and the total score of the QUEST (Supplementary Figure S1). No relevant differences were found for tacrolimus levels versus regimen and the presence of tremors (Supplementary Figure S5).

Interestingly we did see effects of switching and dose reduction of tacrolimus. During the study period, some patients changed the immunosuppressive therapy because of tremors: one patient switched from ER-tacrolimus to LCP-tacrolimus, two patients switched from monotherapy LCP-tacrolimus to combination therapy of low-exposure LCP-tacrolimus with mycophenolic acid and one patient switched from monotherapy ER-tacrolimus to combination therapy of low-exposure ER-tacrolimus with mycophenolic acid. In all four LT recipients a reduction in the severity of tremors and an improved QUEST score after the switch was observed. The other patients experiencing tremors were managed by reducing the dose of the tacrolimus formulation while maintaining these patients on monotherapy with tacrolimus.

Supplementary Figure S2 shows the proportion of responses by level of severity for the EQ-5D-5L dimensions during the study period. Overall, patients reported the least issues in the states of Self-Care and Anxiety/Depression and the most problems in the states of Usual Activities and Pain/Discomfort. No evidence for differences between the study groups in any of the five domains was found.

The likelihood-ratio tests indicated non-linear patient specific trajectories of HRQoL scores and the societal values of the EQ-5D-5L. No evidence was found for between-group differences over the course of the study based on the mixed effect models. The hemoglobin level was statistically significantly associated with a higher EQ-VAS and EQ-5D-5L score, whereas tacrolimus trough levels were statistically significantly associated with a lower EQ-VAS and EQ-5D-5L score (Supplementary Table S2). Figure 3 visualize the expected HRQoL scores and societal values of the EQ-5D-5L together with the corresponding observed values per time point and study group. At the end of the study, the patient’s self-rated HRQoL scores as expressed with the EQ-VAS approximate the mean self-reported EQ-VAS score by the general Dutch population. For both arms, the societal values of the EQ-5D-5L were below those of the general Dutch population.

LT recipients in both study groups achieved a clinically meaningful improvement (>7 points) in the EQ-VAS score at 12 months (LCP-tacrolimus: 20.8 points and ER-tacrolimus: 14.3 points difference with the moment of randomization).

Supplementary Figure S3 shows the results from the SF-36 questionnaire. Every domain of the SF-36 questionnaire improved during the follow-up. Most improvement was shown in the domains: physical functioning, social functioning and pain. No statistically significant differences were found between both study groups on any of the eight domains.

An analysis of the EQ-VAS score in relation to tremors did not show statistically significant differences between LT recipients with and without tremor as indicated by the QUEST questionnaire (Supplementary Figure S4).