Blood pressure changes during smoking cessation in a randomized, double-blind, placebo-controlled trial of dulaglutide treatment

Abstract

Aims

Cigarette smoking cessation reduces cardiovascular risk via various mechanisms. Thereby, the role of blood pressure remains unclear, with studies reporting both decreased and increase blood pressure values after cessation, potentially influenced by weight change. We previously showed that the glucagon like peptide-1 analogue dulaglutide mitigates weight gain after smoking cessation. This secondary analysis investigates the effect of smoking cessation on blood pressure changes in dulaglutide- vs. placebo-treated individuals. We hypothesized a beneficial effect of smoking cessation on blood pressure, particularly in dulaglutide-treated participants.

Methods and results

This is a predefined secondary analysis of a randomized, double-blind, placebo-controlled trial. Participants (n = 255) underwent a 12-week smoking cessation programme including standard of care (behavioural counselling and varenicline) with weekly injections of dulaglutide 1.5 mg or placebo, followed by a follow-up of 52 weeks. The primary outcome was change in systolic blood pressure after 52 weeks in abstinent vs. smoking individuals. Further outcomes included blood pressure and body weight changes at Week 12 and 52 according to smoking status and treatment arms. A path analysis was performed to estimate direct and indirect effect of different variables on systolic blood pressure changes. Two hundred and eighteen out of 255 participants with complete blood pressure readings were included in the analyses. Across the entire study population, systolic blood pressure was stable over the period of 52 weeks after smoking cessation despite a weight gain of +3 kg (0, 5.4) at Week 52. Blood pressure reductions were seen in the subgroups of participants with minimal weight gain ≤3 kg [−4.6 mmHg (−9, 3)] and in individuals with hypertensive blood pressure values at baseline [−16 mmHg (−22, 2)]. Dulaglutide treatment reduced body weight and blood pressure initially, followed by a weight rebound and a blood pressure increase of +7.5 mmHg (−1, 15) at Week 52. The path analysis identified weight as an important factor influencing blood pressure during smoking cessation.

Conclusion

Our analysis suggests that smoking cessation may have a beneficial effect on blood pressure- especially in hypertensive individuals-, counteracting the expected blood pressure increase caused by post-cessation weight gain. However, it also underlines the importance of weight control after smoking cessation as a crucial factor in smoking cessation.

Registration

ClinicalTrials.gov: NCT 03204396.

Graphical Abstract

Open in new tabDownload slide

Introduction

Cigarette smoking is an established major risk for atherosclerosis, cardiovascular morbidity, and mortality. The underlying mechanisms are complex and mediated by both combustion products and nicotine, leading to inflammation, endothelial dysfunction, lipid alterations, platelet activation, and sympathetic nervous system activation.^1–3

Whether blood pressure fluctuations also contribute to this increased cardiovascular risk is a matter of debate. In fact, repetitive blood pressure peaks after cigarette smoking leading to profound blood pressure fluctuations during the day have been linked to masked hypertension in smoking individuals.^4–8 On the other hand, cross-sectional studies show that smoking individuals tend to have lower blood pressures values than non-smoking individuals and suggest a negative dose-response relationship between amount of smoking and blood pressure.^6–11

In line with that, a prospective cohort study of 1550 smoking females followed for ∼9 years showed that participants who successfully quit smoking had increased odds of hypertension (OR 1.3) compared with those who continued smoking.¹² However, it remained unclear whether or to what extent the increased incidence of hypertension could be explained by post-cessation weight gain (in this cohort, participants who quit smoking had a 3–4 kg higher weight gain compared with those who continued smoking).

In contrast, other longitudinal studies, such as the Framingham study did not show significant increases in blood pressure in subjects after smoking cessation.^7,13,14 Of note, a recent randomized controlled trial even showed decreased blood pressure values in a subgroup of hypertensive participants following a smoking cessation intervention of 12 weeks despite weight gain.¹⁵ Therefore, the role of smoking and smoking cessation in blood pressure regulation remains unclear and is complicated by weight changes during smoking cessation. We have recently demonstrated in a randomized controlled trial that the glucagon like peptide-1 (GLP-1) analogue dulaglutide effectively reduces weight gain during a 12-week smoking cessation programme.^16,17 In addition to their well-known weight- and glucose-lowering effects, studies have demonstrated reductions in blood pressure in obese individuals, with or without Type 2 diabetes, treated with GLP-1 analogues even before substantial weight loss has occured.^18–21

In this predefined secondary analysis, we aimed to explore the effect of smoking cessation on systolic blood pressure changes after 12 weeks of dulaglutide vs. placebo treatment and after 52 weeks of follow-up. We hypothesized a beneficial effect of smoking cessation, particularly in dulaglutide-treated participants, which may be related to, or partly independent of, reduced weight gain after smoking cessation.

Methods

Study design

This is a predefined secondary analysis of a randomized, placebo-controlled, double-blind trial conducted at University Hospital Basel in Switzerland from June 2017 to March 2022. The study protocol and primary study results have been published previously.^17,22,23 In brief, the trial included 255 participants who underwent a smoking cessation programme of 12 weeks’ duration, receiving Swiss standard of care with behavioural counselling and varenicline, a nicotine receptor partial agonist. Participants were randomly assigned 1:1 to either weekly subcutaneous injections of the study drug dulaglutide 1.5 mg or placebo for 12 weeks. Total follow-up time was 52 weeks (see Supplementary material online, Figure S1). Written informed consent was obtained from all study participants, and the trial was conducted according to the established ethical guidelines of the Declaration of Helsinki and the International Conference on Harmonization guidelines on Good Clinical Practice. The study was registered on ClinicalTrials.gov, identifiers NCT03204396.

Participants

Inclusion criteria were active daily smoking, with at least moderate nicotine dependence (Fagerstrom score ≥5 points), ≥10 cigarettes per day or a tobacco associated disease, 18–75 years of age and the desire to quit smoking.²⁴ Main exclusion criteria were pregnancy, severe renal insufficiency, previous treatment with GLP-1 analogues, or unstable psychiatric conditions. The whole list of eligibility criteria can be found in the study protocol.¹⁷ For this secondary analysis, all participants for whom systolic blood pressure measurements at baseline and Week 12 were available and who were adherent to treatment (≤4 missed injections in the dulaglutide group) were included in the analysis (as treated set), see Figure 1. Participants with the following blood pressure values at baseline were included in the subgroup of hypertensive participants: systolic blood pressure ≥135 mmHg and/or diastolic blood pressure ≥85 mmHg.²⁵

Study procedure

Study visits took place at baseline and Weeks 12, 24 and 52. At each visit, blood pressure was measured by an automated office blood pressure measurement (AOBPM), automatically starting without premonishing the participant after 5 min of rest, executing three measurements spaced ∼2 min apart. The mean value of three measurements was used. Automated office blood pressure measurement has been shown to correlate more strongly with a 24 h blood pressure measurement than a common office blood pressure measurement.²⁵ Because this secondary analysis focused on physiological and mechanistic trends, smoking abstinence was defined by a stricter abstinence definition than the one used in the main analysis of the trial.²² Smoking abstinence was defined as self-reported abstinence during the last 7 days and an end-expiratory carbon monoxide (CO) ≤ 10 ppm, measured by the Micro+™ Smokerlyzer®,²⁶ and testing negative for urinary cotinine, measured in the spot urine by the ‘Urine Cotinine All Test COT 3 in 1’.²⁷ Prolonged abstinence at Weeks 24 and 52 required that these criteria were met at the respective visit as well as at all previous study visits.

Study outcomes

The primary outcome was the change in systolic blood pressure at Week 52 compared with baseline between abstinent participants and those who continued smoking. Secondary outcomes were change in systolic blood pressure at Weeks 12 and 24 as well as change in weight, diastolic blood pressure, and heart rate at Weeks 12, 24, and 52, respectively, in abstinent and smoking individuals and according to treatment dulaglutide vs. placebo. Predefined subgroup analyses consisted of abstinent participants only, individuals with substantial weight gain (>3 kg) vs. moderate weight gain (≤3 kg) at Week 52, individuals with hypertensive blood pressure values at baseline (systolic >135 mmHg and/or diastolic >85 mmHg) vs. with normotensive blood pressure values at baseline, male vs. female participants, extent of cigarette dependence (assessed by the Fagerstrom score and numbers of cigarettes per day), and age ≤50 vs. >50 years. In a path analysis (see below), we aimed to analyse the direct effect of smoking cessation on systolic blood pressure while taking into consideration other variables influencing blood pressure such as weight change, gender, age, body mass index (BMI), and blood pressure at baseline.

Statistical analysis

Data were summarized using descriptive statistics to avoid potential errors associated with multiple testing. Unless otherwise stated, all data are expressed as mean and standard deviation or median with interquartile range, as deemed appropriate. Change in blood pressure, heart rate, and weight for respective subgroups (smoking vs. abstinent and dulaglutide vs. placebo-treated participants) was visualized and analysed descriptively at Weeks 12, 24, and 52. The analyses performed in this predefined secondary analysis are exploratory in nature. The P-values that are reported in the path analyses provide a signal for potentially important pathways that should be further assessed in future confirmatory studies. No adjustment for multiple testing was performed. Participants without blood pressure measurements at Week 12 were excluded, missing data were not imputed in this analysis. All statistical analyses were conducted using R version 4.4.1.²⁸

Path analysis using R package lavaan²⁹ was performed to estimate direct and indirect effects of different variables on change in systolic blood pressure considering the influence of other variables included in the model. The rationale for choice of variables of interest was clinically based. Hypothesized relationships and interactions of variables were predefined and are graphically displayed in Supplementary material online, Figure S2. To handle missing data, we used casewise full-information maximum likelihood estimation as implemented in the lavaan package.²⁹ Sex and age were considered as important covariates influencing change in blood pressure and weight. Body mass index at baseline was included as a predictor of weight change. Blood pressure at baseline was included as a predictor of change in blood pressure. In a first step, we tested whether there was an interaction between treatment with dulaglutide and smoking cessation on weight change. Because the interaction was not significant, it was excluded from the final model.

Results

Patient characteristics

Between June 2017 and March 2021, 255 participants were included and randomized. Two hundred and eighteen of them were eligible for this analysis with 103 (47%) participants in the dulaglutide treatment group and 115 (53%) in the placebo group, see Figure 1.

Baseline characteristics were similarly distributed in the dulaglutide and placebo groups except for a higher percentage of females in the dulaglutide group (64 vs. 56%). Median age was 41 years in the dulaglutide and 44 years in placebo group, with a median BMI of 26 kg/m² in both groups. Median pack years in the dulaglutide group were 18 and 20 in the placebo group, see Table 1.

Eighteen (18%) participants in the dulaglutide group and 23 (20%) in the placebo group had a medical history of hypertension. A large majority of participants had blood pressure values within normal range with a median of 118/77 mmHg in the dulaglutide and 121/78 mmHg in the placebo group. Based on these baseline values, the hypertensive subgroup consisted of 15 (15%) in the dulaglutide and 18 (16%) in the placebo-treated group.

At the end of the intervention period (Week 12), smoking abstinence was achieved in 56 (54%) participants in the dulaglutide group, and 66 (57%) participants in the placebo group. At the end of follow-up (Week 52), prolonged smoking abstinence was achieved in 30 (29%) participants in the dulaglutide group, and 30 (26%) participants in the placebo group.

Effect of smoking cessation

Systolic blood pressure

Overall systolic blood pressure was fairly stable over the study period of 52 weeks with no difference between continuing smoking and abstinent participants [Figure 2A, Week 12: smoking −1.5 mmHg (−6, 5), abstinent −2.0 mmHg (−11, 5); Week 24: smoking −1.0 mmHg (−8, 5), abstinent −1.0 mmHg (−7, 7); Week 52: smoking +1.0 mmHg (−4, 9), abstinent +2.5 mmHg (−7, 12)]. Diastolic blood pressure and heart rate were similarly stable (see Supplementary material online, Figure S3A and B).

Weight

Across the entire study population, weight change was +3 kg (0, 5.4) at Week 52. Abstinent participants showed a profound weight gain over the study period with a median weight gain of +4.2 kg (1, 7) after 52 weeks follow-up, while participants continuing smoking showed a median weight gain of +2.1 kg (−1, 4; Figure 2B).

Effect of dulaglutide treatment

Systolic blood pressure

Systolic blood pressure was stable over the study period of 52 weeks in the placebo treatment group. Dulaglutide-treated participants showed a median reduction in systolic blood pressure of −3 mmHg (−8, 3) at Week 12, followed by a systolic blood pressure increase of a median of +4 mmHg (−2, 11) at Week 52, compared with 0 mmHg (−7, 5) in placebo group at Week 52. A similar dynamic was seen for diastolic blood pressure. Heart rate increased slightly in both groups at Week 52 with dulaglutide-treated participants also showing an increase at Week 12 (see Supplementary material online, Figure S4A and B).

Weight

There was a difference in course of weight gain between dulaglutide- and placebo-treated participants (Figure 2D). While the placebo-treated group showed a continuous weight gain of +1.8 kg (−1, 2) at Week 12 with further weight increase of absolute +3 kg (1, 5) at Week 52, dulaglutide-treated participants showed a slight weight reduction of −0.9 kg (−3, 1) at Week 12, reaching similar weight gain levels as the placebo-treated group of absolute +2.9 kg (0, 5) at Week 52.

Subgroup of abstinent participants

Systolic blood pressure

There was a difference in systolic blood pressure in the subgroup of abstinent participants between dulaglutide and placebo treatment (Figure 2E). While the placebo-treated group showed a slight blood pressure reduction of—3.5 mmHg (−14, 6) at Week 52 [−2.5 mmHg (−13, 5) at Week 12], dulaglutide-treated participants showed a profound blood pressure increase of +7.5 mmHg (−1, 15) at Week 52 [−2 mmHg (−6, 8) at Week 12]. A similar dynamic was seen for diastolic blood pressure, while heart rate decreased in the dulaglutide treated group at Week 52 (see Supplementary material online, Figure S5A and B).

Weight

There was a difference in weight change in the subgroup of abstinent participants between dulaglutide and placebo treatment (Figure 2F).

While the placebo-treated group showed a continuous weight gain, with +2.3 kg (1, 3) at Week 12 and a further increase of absolute +4.2 kg (1, 7) at Week 52, dulaglutide treated participants showed a slight weight reduction of −0.8 kg (−3, 1) at Week 12, reaching similar weight gain levels like the placebo-treated group of absolute +4.2 kg (2, 8) at Week 52.

Substantial weight gain

Analyses of participants with substantial weight gain >3 kg (n = 40) and participants with no or moderate weight gain ≤3 kg (n = 20) after 52 weeks showed divergent systolic blood pressure changes with +5.4 mmHg (−4, 13.5) in the substantial weight gain group, compared with a weight reduction of—4.6 mmHg (−9, 3) in moderate weight gain group (Figure 3A). Relationship between weight gain and systolic blood pressure change at Week 52 is shown in Supplementary material online, Figure S6A.

Hypertensive blood pressure at baseline

While the subgroup of abstinent participants with hypertensive blood pressure values at baseline (n = 11) showed a considerable reduction in systolic blood pressure with a median of −16 mmHg (−22, 2) at Week 52, abstinent participants with normotensive blood pressure values at baseline (n = 49) showed a median blood pressure increase of +3 mmHg (−2, 11) (Figure 3B). Median weight gain at Week 52 in the subgroup of abstinent participants with hypertensive blood pressure values at baseline was +4.8 kg (3.3, 7.9) compared with +3.8 kg (1.1, 7.4) in normotensive abstinent participants.

Extent of cigarette dependence

Analyses of the extent of cigarette dependence, assessed by the Fagerstrom score, showed somewhat higher systolic blood pressure reduction (see Supplementary material online, Figure S6B). Similarly, a greater blood pressure reduction was observed in participants with higher reduction in the number of cigarettes per day after 52 weeks (see Supplementary material online, Figure S6C). A positive correlation of the Fagerstrom score at baseline and weight gain was also noted (see Supplementary material online, Figure S6D).

Gender and age

There was a small difference in change in systolic blood pressure between male [median increase of +3 mmHg (−2, 10)] and female participants of the entire cohort [+0 mmHg (−6.5, 8)] at Week 52 (see Supplementary material online, Figure S7A) with similar weight gain between male and female participants (see Supplementary material online, Figure S7B). There was no difference in change in systolic blood pressure between participants ≤50 and >50 years of age of the entire cohort (see Supplementary material online, Figure S8).

Path analysis

In the path analysis for Week 12, dulaglutide treatment showed a reducing effect on weight change of—3.0 kg on average (−3.7, −2.4), P < 0.001, Figure 4A, Supplementary material online, Figure S9A. Weight change had a direct effect on systolic blood pressure: each kg weight gain resulted in a systolic blood pressure increased by +0.67 mmHg [(0.0, 0.1), P = 0.01]. Dulaglutide treatment had an indirect effect on systolic blood pressure, mediated by weight loss [−2.0 mmHg (−3.7, −0.4), P = 0.02]; there was no direct effect of dulaglutide on systolic blood pressure. Path analysis showed no effect of smoking cessation on blood pressure at Week 12, neither directly, nor via the indirect path mediated by weight gain.

In the path analysis for Week 52, dulaglutide treatment had a direct impact on systolic blood pressure change of +5.2 mmHg (1.5, 9.0), P < 0.01, Figure 4B, Supplementary material online, Figure S9B. Smoking cessation had a direct impact on weight gain of +3.0 kg [(1.3, 4.6), P < 0.001]. Mediation analysis showed no effect of smoking cessation or weight change on blood pressure.

Discussion

Across the entire study population, blood pressure was reassuringly stable over the period of 52 weeks, despite a clinically relevant average post-cessation weight gain of more than 4 kg. This may imply that smoking cessation could lead to blood pressure benefits, which is also suggested by the favourable blood pressure results in participants with only moderate post-cessation weight gain and in hypertensive individuals.

Regarding dulaglutide and placebo treatment, weight and blood pressure trajectories were considerably different in the two groups. While dulaglutide treatment reduced body weight and blood pressure initially, it was followed by a weight rebound and a profound blood pressure increase after 1 year.

The observed stability in systolic blood pressure despite post-cessation weight gain is noteworthy, especially in the connection with the results of the subgroup analyses: in abstinent participants with only moderate or no post-cessation weight gain, a reduction in systolic blood pressure after smoking cessation was noted. A positive correlation between the reduction in blood pressure and the extent of cigarette dependence also suggested that the beneficial effects on blood pressure might be more pronounced in individuals with a high extent of cigarette dependence (e.g. Fagerstrom score). The most pronounced blood pressure benefits were observed in the subgroup of participants with hypertensive blood pressure values at baseline who showed a blood pressure reduction of median −16 mmHg after 52 weeks of smoking cessation, despite a median weight gain of +4.8 kg. Of note, there was no difference in newly started antihypertensive medication between the groups, excluding treatment changes as possible confounding factors (data not shown). Interestingly, in a recent subgroup analysis of a randomized controlled trial of 113 hypertensive smoking individuals, Gaya et al.¹⁵ indicate a blood pressure reducing effect of −5.4 mmHg after 12 weeks of smoking cessation. In line with our data, this effect was observed despite an accompanied post-cessation weight gain of around 1.5 kg. Therefore, we suggest that smoking cessation may have important beneficial effects on blood pressure that are particularly relevant or clinically evident in people with pre-existing hypertension. These beneficial effects are thought to be mediated by changes in sympathetic tone, restoration of endothelial function, decreased oxidative stress, inflammation, and metabolic changes.^3,30 Our study was not designed for a hypertensive cohort and the number of participants with hypertensive blood pressure values was small with most participants having normal blood pressure ranges, but our findings warrant further investigation in a future study of hypertensive individuals.

Beside pre-existing hypertension, post-cessation weight gain was another important factor modulating blood pressure in our study. This was also evident in the path analysis at Week 12, which showed that each kg of weight gain was associated with a +0.7 mmHg increase in systolic blood pressure. While the highly beneficial effect of smoking cessation on cardiovascular risk reduction is unquestioned,^2,31,32 substantial post-cessation weight gain may mitigate these beneficial effects, as illustrated by a clinically relevant systolic blood pressure increase of ∼5 mmHg at Week 52 in these participants. This is consistent with a recent study by Ninomiya et al.³³ in which post-cessation weight gain (>1.8 kg) was associated with an increase in systolic blood pressure of around 10 mmHg 6 years after smoking cessation. These findings emphasize the need for comprehensive interventions aimed at effective weight management in smoking cessation.

Lifestyle modifications with increased physical activity has been shown to be a major factor for weight control and blood pressure improvements, with lean mass being one important factor influencing blood pressure.^34–38 Therefore, physical activity should be encouraged as a major factor in smoking cessation to prevent post-cessation weight gain. Whether the combination of physical activity and long-term GLP-1-analogue treatment, which has been shown to be an effective strategy in obesity,³⁹ could also be a beneficial option in the context of post-cessation weight control, remains to be investigated.

In our cohort, dulaglutide treatment was not given as a long-term treatment, but only for 12 weeks. While body weight was successfully reduced during the treatment, a rapid weight rebound occurred after stop of dulaglutide treatment, accompanied by a profound systolic blood pressure increase (+7.5 mmHg) after 12 months that was not observed in participants treated with placebo. This increase is most presumably attributed to rapid weight rebound following the treatment phase. Results of the STEP 4 and STEP 1 trial extension underline the challenging concern of weight regain after stopping GLP-1 analogue treatment.^36,37 After a 68-week treatment phase of once-weekly semaglutide 2.4 mg, participants regained two-thirds of their prior weight loss by Week 120. Similar findings were seen for cardiometabolic improvements, as systolic and diastolic blood pressure reductions after 68 weeks treatment (mean systolic blood pressure change of −8 mmHg compared with baseline) reverted to baseline levels 1 year after stopping of GLP-1 analogue treatment (mean systolic blood change +2 mmHg compared with baseline).³⁶ These findings emphasize the critical role of concomitant lifestyle changes during GLP-1 analogue treatment.

Limitations

Some limitations should be considered when interpreting our findings. First, participants who continued smoking also underwent a cessation attempt, potentially masking the true effect of smoking cessation on blood pressure due to reduced nicotine use, especially at Week 12. Second, demonstrating differences in blood pressure benefits was challenging in mostly young and healthy participants with low hypertension prevalence. The number of hypertensive blood pressure values at baseline was too small to apply statistical analyses to this subgroup in a robust manner. A larger sample size, especially a bigger population of hypertensive patients might clarify the suggested benefits more distinctly. Third, our data are not able to provide information on blood pressure changes independently of varenicline treatment, as both groups received this treatment as standard of care. Lastly, we have not assessed changes in body composition or waist circumference-to-height ratio, a only recently proposed new framework for weight assessment.⁴⁰

Strengths

The randomized controlled setting of this predefined analysis allowed for comprehensive investigations into various aspects of blood pressure, including the effects of smoking cessation over a period of up to 52 weeks and the impact of a 12-week dulaglutide treatment during smoking cessation. Additionally, the utilization of standardized AOBPM at each visit, which has been shown to correlate well with 24 h blood-pressure measurements,²⁵ ensured highly reliable blood pressure data. Measurement of end-expiratory carbon monoxide and urinary cotinine provided a robust and objective definition of smoking abstinence.

Conclusion and outlook

This analysis suggests that smoking cessation may have a beneficial effect on blood pressure- especially in hypertensive individuals-, counteracting the expected blood pressure increase caused by post-cessation weight gain. At the same time, it underlines the importance of weight control after smoking cessation as a crucial factor influencing blood pressure outcome. A 12-week treatment with GLP-1 analogues did not provide long-term benefits on weight and blood pressure in our cohort but highlights the risk of weight rebound after stopping treatment, which was associated with a blood pressure increase. Therefore, studies with long-term GLP-1 analogue treatment are needed to determine if these drugs may offer additional benefits for weight and blood pressure control and cardiovascular risk in this setting.

Supplementary material

Supplementary material is available at European Journal of Preventive Cardiology.

Acknowledgements

We are grateful to our participants for taking part in the trial. We thank support staff, study and laboratory personnel at the University Hospital Basel, especially Nina Hutter, Joyce Santos de Jesus, Silke Purschke, and Nicole Salvisberg and all members of the clinical neuroendocrinology research team and the group of the doctoral students for their most helpful support during the analyses. We thank the tobacco treatment specialists Uta Engler and Caroline Beer for screening and supporting recruitment of possible participants and for taking over standard of care during the study.

Author contributions

J.B. analysed and interpreted the data, wrote, and finalized the manuscript. F.H. analysed the data and did the literature search. T.B. and A.M. were involved in the study design and contributed to the manuscript. L.W. planned, performed, and interpreted the statistical analyses and contributed to the manuscript. S.L., A.M., C.B., contributed to data collection. D.V. was involved in the statistical analysis. S.L. was involved in the study design. M.C.-C. gave input to the study design and contributed to the manuscript. B.W. designed the study, wrote the protocol, collected, analysed and interpreted data and supervised all steps of the conduct of the study. All authors edited and approved the final manuscript. The funders had no role in study design, data collection, data analysis, data interpretation, or data reporting. The first and last authors had access to all data and have final responsibility for the decision to submit for publication.

Funding

Swiss National Science Foundation (PZ00P3_193206), Gottfried and Julia Bangerter-Rhyner Foundation, Goldschmidt-Jacobson Foundation, HEMMI Foundation, University and University Hospital of Basel, Swiss Academy of Medical Science.

Data availability

We may share de-identified, individual participant-level data that underlie the results reported in this article and related documents, including the study protocol and the statistical analysis plan. Data will be available with the publication of our main manuscript on receipt of a request detailing the study hypothesis and statistical analysis plan. The data underlying this article will be shared on reasonable request to the corresponding author. The steering committee of this study will discuss all requests and decide based on the scientific rigour of the proposal whether data sharing is appropriate. All applicants are asked to sign a data access agreement.

References

Author notes