Article Text
PDF
Abstract
Background Preprocedural predictors of outcome in patients with acute basilar artery occlusion (ABAO) who have undergone endovascular treatment (EVT) remain controversial. Our aim was to determine if pre-EVT diffusion-weighted imaging cerebellar infarct volume (CIV) is a predictor of 90-day outcomes.
Methods We analyzed consecutive MRI-selected endovascularly treated patients with ABAO within the first 24 hours after symptom onset. Successful reperfusion was defined as a modified Thrombolysis in Cerebral Infarction score of 2b–3. Using the initial MRI, baseline CIV was calculated in mL on an apparent diffusion coefficient map reconstruction (Olea Sphere software). CIV was analyzed in univariate and multivariable models as a predictor of 90-day functional independence (modified Rankin Scale (mRS) 0–2) and mortality. According to receiver operating characteristic (ROC) analysis, the optimal cut-off was determined by maximizing the Youden index to evaluate the prognostic value of CIV.
Results Of the 110 MRI-selected patients with ABAO, 64 (58.18%) had a cerebellar infarct. The median CIV was 9.6 mL (IQR 2.7–31.4). Successful reperfusion was achieved in 81.8% of the cases. At 90 days the proportion of patients with mRS ≤2 was 31.8% and the overall mortality rate was 40.9%. Baseline CIV was significantly associated with 90-day mRS 0–2 (p=0.008) in the univariate analysis and was an independent predictor of 90-day mortality (adjusted OR 1.79, 95% CI 1.25 to 2.54, p=0.001). The ROC analysis showed that a CIV ≥4.7 mL at the initial MRI was the optimal cut-off to discriminate patients with a higher risk of death at 90 days (area under the ROC curve (AUC)=0.74, 95% CI 0.61 to 0.87, sensitivity and specificity of 87.9% and 58.1%, respectively).
Conclusions In our series of MRI-selected patients with ABAO, pre-EVT CIV was an independent predictor of 90-day mortality. The risk of death was increased for baseline CIV ≥4.7 mL.
- MRI
- posterior fossa
- stroke
- thrombectomy
Data availability statement
Data are available upon reasonable request. Data are available from the corresponding author on reasonable request.
Statistics from Altmetric.com
Introduction
Acute basilar artery occlusion (ABAO) is a rare and devastating type of stroke.1 Endovascular treatment (EVT) is routinely performed in real-world practice, encouraged by guidelines from learned societies, despite conflicting results of clinical studies.2–5 Recently, the Basilar Artery International Cooperation Study (BASICS), a randomized controlled study, showed that EVT was effective in patients with ABAO and moderate or severe deficit.6
ABAO may result in infarcts in the brainstem, cerebellar lobes, thalamus and subthalamic area, or occipitotemporal lobes. The relationship between posterior circulation diffusion-weighted imaging (DWI) infarction and outcome has been previously examined. In posterior circulation occlusive disease, the DWI lesion volume did not correlate significantly with the National Institutes of Health Stroke Scale (NIHSS) score, and studies using the DWI posterior circulation Alberta Stroke Program Early CT (ASPECT) score or brainstem score for predicting outcome in patients with ABAO showed conflicting results.7–11 The rationale for the brainstem scoring system was based on the anatomic architecture and the regional clinical relevance of the brainstem, which may negatively influence the functional outcome.
A predictive model of a good outcome combining an initial DWI infarct volume of <10 mL, onset-to-puncture time of <8 hours, and embolic origin has recntly been reported in 71 Korean patients undergoing EVT for vertebrobasilar occlusion.12
One of the most significant and immediate complications in patients with ABAO and cerebellar infarcts is the development of a space-occupying edema. Due to the small volume of the posterior fossa, cerebellar infarcts with mass effect (CIMASS) contribute to a risk of catastrophic herniation with direct brainstem compression, hydrocephalus due to blockage of the fourth ventricle, or both. Up to 25% of patients with CIMASS deteriorate clinically resulting in morbidity and death in 85% of patients without intervention.2 This potential risk of CIMASS has led clinicians to strengthen clinical and radiological monitoring and to modify therapeutic management with recourse to life-saving suboccipital decompressive craniectomy (SDC), as recommended by learned societies.2
Little is known about the predicting factors of outcomes in patients with ABAO treated by EVT and associated cerebellar infarct. No research has focused on assessing whether the baseline cerebellar infarct volume (CIV) correlates with 90-day clinical outcome and mortality or if such lesions would affect a decision on reperfusion.
Using data in our prospective registry, our aim was to evaluate the frequency and the clinical and radiological characteristics of cerebellar infarcts in a large cohort of patients with ABAO selected by MRI prior to EVT, and to investigate whether the baseline CIV is a predictor of 90-day functional outcome and mortality.
Methods
From our prospectively collected databases, we reviewed all patients referred to our hospital from January 2011 to October 2018 who presented with ABAO confirmed by MRI and subsequently treated by EVT, with or without intravenous thrombolysis (IVT).
Inclusion criteria were the following: (1) clinical diagnosis of acute stroke in the posterior circulation within 24 hours from symptom onset; (2) pertinent clinical deficit following physician evaluation (no NIHSS limit); (3) acute basilar ischemia assessed on MRI matching clinical symptoms and including DWI; and (4) ABAO confirmed by intra-arterial DSA and treated using mechanical thrombectomy.
Patients were excluded if they were ineligible for an MRI or ineligible for mechanical thrombectomy for the following reasons: (1) prestroke modified Rankin Scale (mRS) score >2; (2) life expectancy <3 months; (3) brainstem ischemia on DWI involving more than 80% of the area in axial view.
The study was approved by the local ethics committee and was in accordance with the 1964 Helsinki Declaration and its later amendments. Informed consent was obtained from all patients or their relatives before treatment. NIHSS and the Glasgow Coma Scale (GCS) were assessed by a stroke neurologist before the thrombectomy. A NIHSS score of 40 was assigned to an unconscious/comatose patient. Demographic characteristics, stroke presentation, and severity, time of symptom onset, and imaging data were prospectively collected.
Imaging protocol
All patients underwent 1.5T MRI (Magneton Aera or Avanto Siemens Healthcare, Erlangen, Germany or Intera, Ingénia, Philips Healthcare, Best, The Netherlands) or 3T MRI (Magneton Skyra Siemens Healthcare, Erlangen, Germany or Signa HDx, General Electric, Milwaukee, USA) at admission before EVT.
The MRI protocols included DWI with apparent diffusion coefficient (ADC) map reconstruction (to identify the necrotic core), fluid-attenuated inversion recovery (FLAIR) (to corroborate information about the time from symptom onset), T2 gradient echo (to screen for intracranial hemorrhage), and three-dimensional intracranial time of-flight MR angiography (MRA) centered in the region of the circle of Willis or MRA (to screen for large vessel occlusion).
DWI parameters included a repetition time of 3600–8000 ms, echo time of 78–98.3 ms, voxel size of 0.833×0.833×5 mm (3.47 µL) to 1875×1875×5 mm (17.57 µL), slice thickness of 5 mm, number of slices 22–30, and interslice gap of 0–1 mm. All DWI images used the parallel acquisition technique and two different b values (0 and 1000 s/mm2). Moreover, we used one semiquantitative grading system, the posterior circulation ASPECTS.8
The level of ABAO and the presence of pretreatment posterior communicating artery (PCOM) collateral flow was also collected and defined as previously reported.13 The location and vascular territory of the cerebellar infarct was determined according to the anatomical map of the cerebellum published by Tatu et al—namely, infarct in the territory of the posterior inferior cerebellar artery (PICA), superior cerebellar artery (SCA), or anterior inferior cerebellar artery (AICA), and unilateral or bilateral cerebellar infarction.14
Diagnosis of associated infarcts in the brainstem, thalamus, or posterior cerebral artery (PCA) and CIMASS were also recorded on the basis of the initial MRI. The criteria for mass effect included: deformity or displacement of the fourth ventricle, obstructive hydrocephalus, effacement of the basal cisterns, and deformity or anterior bowing of the brainstem.15
Pretreatment imaging findings were retrospectively assessed by a neurointerventionalist (MM) blinded to the clinical outcome and not involved in the treatment.
Cerebellar infarct volumes
Infarct volume was estimated in mL on an ADC map reconstruction using Olea Sphere software Version 3.0-SP16 (Olea Medical SAS, La Ciotat, France). It was semi-automatically segmented after applying a threshold of 0.620 10−3 mm2/s with manual correction when necessary.
Two segmentation approaches were used: (1) evaluation of ischemic volume at the level of the cerebellar structures (the mean value of the ADC was also recorded); and (2) assessment of the total posterior circulation ischemic volume which was defined as an infarct involving the brainstem, cerebellar structures, thalamosubthalamic area, and the occipitotemporal lobes.
Therapeutic regimens
Treatment of ABAO involved either (1) a EVT after IVT (as a bridging therapy) or (2) a EVT alone in patients with a contraindication for IVT. Moreover, IVT was not administered in cases of extensive cerebellar ischemia at DWI sequencing according to the results of a previous report.16 EVT was performed using a stent retriever technique associated with distal aspiration. Antiplatelet drugs were started 24 hours after IVT and based on the CT or MRI control.
Successful recanalization was defined as a modified Thrombolysis in Cerebral Infarction (mTICI) grade of 2b or 3. Periprocedural complications included vessel perforation, arterial dissection, distal embolization of a previous uninvolved territory, vasospasm, and subarachnoid hemorrhage.
Post-EVT neurosurgical procedures including SDC and/or external ventricular drain (EVD) were recorded.
Outcome measures
All patients underwent cerebral imaging (CT or MRI) within 18–24 hours after the procedure.
Symptomatic intracranial hemorrhage (sICH) was defined as a decline of ≥4 points in the NIHSS score associated with any intracranial hemorrhage which was the dominant brain pathology causing neurological deterioration, and the absence of another explanation for neurological deterioration. The location of the hemorrhage was also noted (cerebellar or extracerebellar). A favorable outcome was defined as a 90-day mRS score of ≤2. Mortality and the cause of death at 90 days were also recorded.
Statistical analysis
Data are presented as mean (SD) or median (IQR) for continuous variables and number (percentage) for categorical variables. Normality of distribution was assessed using histograms and the Shapiro–Wilk test. The distribution of CIV (considering patients without a cerebellar infarct as 0 mL) was compared according to favorable outcome (mRS 0–2) and mortality status at 90 days using the Mann–Whitney U test. We further assessed the association of MRI parameters (CIV, number of cerebellar arterial territories involved, brainstem involvement) with favorable outcome and all-cause mortality in univariable and multivariable logistic regression models by considering the following prespecified confounders: age, intravenous thrombolysis, onset MRI time, GCS score (<8 vs ≥8), and collateral PCOM flow pattern. CIVs were analyzed as categorical variables according to tertiles in order to examine the shape of the relationships and as continuous variables after log transformation to reduce the skewness. ORs were derived from logistic regression models as effect sizes. In patients with a cerebellar infarct we further evaluated the prognostic performance of CIV by calculating the area under the receiver operating characteristic curve and determined the optimal cut-off value by maximizing the Youden index. Finally, we assessed the heterogeneity associated with the CIV with outcomes according to the collateral PCOM flow pattern and with successful reperfusion by plotting the distribution of volumes; no formal heterogeneity test was done regarding the small subgroup size and only a qualitative description was done. Statistical testing was performed at the two-tailed α level of 0.05. Data were analyzed using the SAS software package, release 9.4 (SAS Institute, Cary, North Carolina, USA).
Results
Overall, 226 consecutive patients with ABAO were admitted to our hospital between January 2011 and October 2018. Of these, 110 MRI-selected patients who underwent thrombectomy were included in the present study. Patient and treatment characteristics are shown in tables 1 and 2.
Baseline characteristic of patients with acute basilar artery occlusions treated with intention to receive thrombectomy (n=110)
Treatment and outcomes of patients with acute basilar artery occlusions treated with intention to receive thrombectomy (n=110)
Neuroradiological findings
The median delay from symptom onset to MRI was 295 min (IQR 177–425).
Overall, the median total posterior circulation CIV was 4.4 mL (IQR 1.3–21.2) and the median CIV was 0.7 mL (IQR 0–13.5). In the 64 patients with a cerebellar infarct (representing 58.2% of the study sample) the median CIV was 9.6 mL (IQR 2.7–31.4). In these patients the median ADC value was 0.48 10-3 mm2/s (IQR 0.45–0.53 10-3 mm2/s) and cerebellar infarcts were bilateral in 32 patients (50%). Cerebellar infarcts represented 45% (IQR 0–80%) of the total infarct lesions.
The arterial distribution of cerebellar infarcts included the SCA in 41 patients (64.1%), AICA in 27 patients (42.2%), and PICA in 25 patients (39.1%) (online supplemental table 1). Additional brainstem infarction was found in 91 patients (82.7%, bilateral in 47 patients) and PCA infarction in 52 patients (47.2%, bilateral in 21 patients).
Supplemental material
CIMASS was detected in two patients on baseline MRI and in 34 patients CIMASS was radiologically detected during hospitalization. Nine of these 36 patients survived, 23 died due to CIMASS, and four died from other causes.
Treatment
Successful reperfusion (mTICI 2b/3) was achieved in 81.8% (n=90, 95% CI 73.3% to 88.6%) within a median time from groin puncture of 45 min (IQR 22–70). Neurosurgical procedures were performed in six patients (5.4%), including three SDCs with evacuation of necrotic tissue combined with EVD in one patient. Three patients received EVD alone. All patients treated by SDC survived whereas only one patient survived after treatment with EVD alone.
Outcome
At 90 days, 35 patients (31.8%, 95% CI 23.2% to 41.4%) achieved a good outcome (mRS 0–2) and 45 patients (40.9%, 95% CI 31.6% to 50.7%) died within a median time of 6 days (range 1–79).
Mortality was related to the extension of infarction in 38 patients (38/110, 34.5%). Of these, 23 (23/45, 51.5%) showed progressive CIMASS with brainstem compression and hydrocephalus. The remaining 15 (15/45, 33.3%) patients died from extensive posterior circulation infarction. Seven patients (15.5%) died from other causes not related to the infarction. According to these three groups of causes of death, the median time of death was 5 days (IQR 4–7), 8 days (IQR 6–19), and 3 days (IQR 2–22), respectively.
In the 23 patients with CIMASS-related death, the baseline median CIV was significantly larger than in the remaining 22 patients who died from other causes (32.7 mL (IQR 10.6–51.0) vs 0.24 mL (IQR 0–8.1), p<0.001). The median ADC value in these two groups was not significantly different (0.45 10−3 mm2/s (IQR 0.40–0.46) vs 0.43 10−3 mm2/s (IQR 0–0.52, p=0.75).
The overall rates of ICH and sICH within 24 hours were 11.8% (n=13, 95% CI 6.4% to 19.4%) and 2.7% (n=3, 95% CI 0.05% to 7.8%), respectively. There were seven cerebellar hemorrhagic complications, of which two were symptomatic.
DWI baseline cerebellar infarct volume and 90-day outcomes
Compared with patients who did not experience clinical recovery (mRS 3–6) at 90 days, a favorable outcome was associated with a smaller CIV (median 0 vs 3.4 mL; Mann–Whitney U test, p=0.008). Mortality was associated with a larger CIV (median 9.2 vs 0 mL; Mann–Whitney U test, p<0.001). After categorization of CIV into tertiles, the probability of a favorable outcome decreased gradually with increasing tertiles of volumes with an OR per 1 log increase of 0.67 (95% CI 0.48 to 0.92) (table 3). Conversely, the probability of mortality increased with increasing tertiles of volumes, with an adjusted OR per 1 log increase of 1.79 (95% CI 1.25 to 2.54) (table 4).
Association of baseline cerebellar infarct volume in DWI with favorable outcome at 90 days in patients with acute basilar artery occlusions treated by thrombectomy
Association of baseline cerebellar infarct volume in DWI with all-cause mortality at 90 days in patients with acute basilar artery occlusions treated by thrombectomy
In multivariable analyses, baseline CIV was an independent factor associated with mortality (p=0.001). In comparison with the model including only clinical factors (AUC=0.72, 95% CI 0.62 to 0.82), the prognostic value of the multivariable model including baseline CIV was increased (AUC=0.77, 95% CI 0.68 to 0.86) but did not reach a statistically significant level (p=0.17, DeLong test).
After excluding patients without cerebellar infarct, the optimal cut-off of CIV for predicting mortality was ≥4.7 mL (AUC=0.74, 95% CI 0.61 to 0.87), with a sensitivity and specificity of 87.9% and 58.1%, respectively (figure 1).
Receiver operating characteristic (ROC) curves for predicting (A) 90-day favorable outcome and (B) 90-day all-cause mortality by cerebellar diffusion-weighted imaging volume after exclusion of patients without cerebellar infarct. AUC, area under the ROC curve; SE, sensitivity; Sp,specificity.
CIV was not independently associated with the likelihood of 90-day mRS 0–2 in the multivariable model (table 3). The number of cerebellar arterial territories involved and brainstem infarct were not associated with a favorable outcome or with mortality.
MRI infarct volume and outcomes according to collateral PCOM flow and successful reperfusion status
There is no evidence of heterogeneity in the association between CIV and outcomes (either favorable or death) according to collateral PCOM flow or successful reperfusion status (online supplemental figure 1). Whatever the collateral flow pattern and the reperfusion status, patients with a favorable outcome or alive at 90 days had a smaller infarct volume than their counterparts.
Discussion
Our study underlines important findings about the relationship between the CIV and the 90-day clinical outcomes after EVT for MRI-selected patients with ABAO. The major finding of our study was that baseline DWI CIV was an independent predictor of mortality at 90 days. Interestingly, relatively small initial CIVs (≥4.7 mL) were predictive of death, which occurred within the first week in the majority of patients. To the best of our knowledge, this is the first study to investigate the impact of CIV in MRI-selected patients with ABAO who underwent EVT.
Baseline cerebellar infarct characteristics
The baseline characteristics of the cerebellar infarcts in our cohort showed some discrepancies with our current knowledge based on clinicopathologic and neuroimaging features. Cerebellar infarcts accounted for 1.9% of 1000 consecutive initial strokes, with a male preponderance (2–3 times).17 Similar to our cohort, patients had the usual vascular risk factors, a mean age of 65±13 years, and a cardioembolic cause that occurs more frequently than an atherothrombotic cause (except for AICA infarcts).17 Arterial occlusion is located in the basilar artery in one-quarter of patients (50% in the vertebral artery and 20% in a cerebellar artery).17 While, in the literature, cerebellar infarcts are more frequently attributed to PICA, in our cohort the main arterial territory involved was the SCA. Bilateral cerebellar infarctions are considered rare lesions; however, they are now more frequently detected by the use of DWI-MRI.18 Accordingly, in our cohort, half of the baseline cerebellar infarctions were bilateral.
Two perspectives may explain the discrepancies between our study and earlier findings. First, previous reports focused on cerebellar infarcts, whatever the occlusion site, whereas we included only patients with ABAO. Second, all patients were selected with MRI before EVT. DWI is known to be more effective in providing optimal visualization of acute posterior circulation stroke,3 especially bilateral cerebellar infarct or brainstem involvement.
Cerebellar infarct mass effect and clinical outcome
In our cohort of MRI-selected patients undergoing EVT for ABAO, more than 58% of subjects had a cerebellar ischemic lesion on baseline MRI, with a median ischemic volume of 9.6 mL. The main cause of mortality was related to CIMASS in 23 patients (51.5% of deaths), representing 21% of our overall population, emphasizing the prognostic value of cerebellar infarction in this population.
Data on the incidence of CIMASS are sparse and estimated to be between 4% and 39% of all cases of cerebellar infarction.15 19 20 Despite advances in medical treatment for strokes, 20–48% of patients with CIMASS deteriorate clinically.15 21 22
The timing of edema formation can vary between days 2 and 9, suggesting the importance of closely monitoring these patients using MRI rather than CT.15 17 19 21 23
In our study, CIMASS and clinical deterioration were more frequent than previously reported. More than 54% of our 64 patients with baseline cerebellar infarctions had signs of CIMASS (ie, 32.7% of our overall population), and 65.7% of the patients with CIMASS died due to hydrocephalus, brainstem compression, or both. More important, in our cohort baseline CIV was an independent predicting factor of 90-day mortality. The baseline CIV was associated with mortality even for small volumes. Accordingly, we found that an initial volume ≥4.7 mL was significantly related to the risk of mortality at 3 months. It is likely that relatively small initial CIVs are associated with the risk of death due to the potential evolution of the cerebellar infarct and cerebellar edema during the following days of ABAO. Indeed, in the group of 23 patients who died from CIMASS, baseline median CIV was 32.7 mL compared with 0.24 mL in patients who died from other causes.
A search of the literature suggests that deterioration seems to be more dependent on initial infarct volume than on any specific vascular territory.15 However, other factors such as infarcts affecting multiple arterial territories, hemorrhagic transformation, and inadequate collateral blood flow have been reported.15 23
In addition, despite the univariate analysis showing an association between CIV and the rate of 90-day independence with volumes >4.4 mL associated with an increased likelihood of a poor outcome, our multivariable model failed to confirm the CIV as an independent predictor of neurological outcome at 3 months. It is probable that other factors such as the involvement of the brainstem may influence the probability of neurological recovery.
Our results suggest that ABAO is a severe pathology with high rates of mortality, and that involvement of the cerebellum may be an important prognosticator of mortality due to the development of posterior fossa mass effect. It is therefore possible that cerebellar infarcts should be checked during pretreatment imaging and patient selection. Moreover, these patients should be closely monitored with MRI-FLAIR or DWI for the early detection of CIMASS and for consideration for pre-emptive neurosurgery.
Cerebellar infarct and craniectomy
In contrast with craniectomy for malignant MCA stroke, which has been examined with well-designed randomized controlled trials, the optimal management of cerebellar stroke remains controversial.24 However, SDC is a life-saving procedure recommended by current guidelines associated with better outcomes compared with decompressive surgery for hemispheric infarctions.2 22 24 The findings of a recent meta-analysis suggest that a shorter time to surgery (<48 hours after onset), younger age (<60 years), concomitant EVD insertion, debridement of infarcted tissue, and higher preoperative GCS score may contribute to lower rates of death and unfavorable outcomes.22
Insertion of EVD as the sole treatment appears not to be recommended.18 However, in bilateral cerebellar infarctions early surgical intervention could be beneficial, particularly when there is no established brainstem infarction, and MRI may help in recognizing the extent of brainstem destruction in such cases.18 19 However, in these surgical studies, patients with ABAO with coma or very poor neurological conditions were usually excluded or under-represented.20 22
In our study, among the 36 patients with CIMASS, six (16.6%) neurosurgical procedures were performed (5.4% of our overall population) including three lifesaving SDCs and three EVDs only. Non-eligibility criteria for SDC for patients with CIMASS were concomitant extensive bilateral brainstem infarct and extensive posterior circulation infarct with persistent basilar artery occlusion (unsuccessful thrombectomy or basilar reocclusion).
The paucity of high-quality data for SDC after cerebellar stroke is challenging and requires further studies to define the optimal timing of surgery and the optimal surgical procedure.
Finally, it is important to point out that cerebellar lobes seem to be more vulnerable than other posterior circulation structures to ICH. In our study, 11% of patients (7/64) with baseline acute cerebellar infarct developed cerebellar ICH within 24 hours of stroke onset. Moreover, 66.6% of sICH were in the cerebellar lobes and, in one case, sICH contributed to CIMASS and death.
These findings are in accordance with previous reports, showing that hemorrhagic transformation is frequent in patients with acute cerebellar infarction (43%) and that the initial infarct diameter and volumes >4.5 mL were independent factors associated with hemorrhagic transformation.16 These results emphasize the importance of analyzing the involvement of the cerebellum in the ABAO, and its ischemic volume before reperfusion therapy and during the early follow-up period after mechanical thrombectomy.
Study limitations
Our study has several limitations. First, it is a single-center, retrospective, non-randomized observational registry study. Second, because of the small sample size we cannot exclude a lack of adequate statistical power, especially in the multivariable models. Third, the specificity of the optimal cut-off of CIV for the prediction of mortality was quite low (58.1%), suggesting the effect of other prognostic factors. Further studies are therefore required to confirm whether the baseline CIV is reliable for predicting the clinical outcome in patients with ABAO treated with EVT.
Conclusion
In our series of MRI-selected patients with ABAO, the initial CIV was an independent predictor of 90-day mortality. The risk of death increased with CIVs ≥4.7 mL. The prognostic value of baseline CIV in this population has been previously underestimated and should be checked during pretreatment imaging to help clinicians to identify patients exposed to a higher risk of CIV-related mortality. Further studies are warranted to confirm these results.
Data availability statement
Data are available upon reasonable request. Data are available from the corresponding author on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
The study was approved by the local ethics committee, the Institutional Review Board (IRB) of the University Hospital of Montpellier, (IRB-MTP_2020_10_202000630). It was carried out in accordance with the 1964 Helsinki Declaration and its later amendments. Clinical Trials.gov ID: NC T04611178.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors IM, MM, CD, JL, ELB, and FC participated in the conception and design of the study, analyzed and interpreted the data and were responsible for drafting of the article. JL wrote the statistical analysis plan and was responsible for statistical expertise. IM, MM, CD, FP, CA, ID, NG, GBF, PHL, YB, GG, LC, VC, ELB, and FC were responsible for the provision of study materials or patients and provided feedback on the paper. IM, MM, CD, FP, ELB, and FC were responsible for collection, assembly, and possession of the raw data.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.