JAK2V617F variant allele frequency, non-driver mutations, single-nucleotide variants and polycythemia vera outcome

Introduction Despite comparatively favourable prognosis in polycythemia vera (PV) patients (pts), the overall survival is shorter compared to the age-matched general population. The aim of the study was to evaluate the impact of chosen laboratory and genetic factors on the individual disease outcome, i.e. risk of thrombosis, myelofibrosis/blastic transformation and death. Materials and methods The study group consisted of 151 pts and 57 healthy donors (HD). Results JAK2V617F mutation was found in 96.7% (146/151) of the studied pts. JAK2 exon 12 mutations were identified in 2 individuals. The coexistence of JAK2V617F and JAK2 exon 12 mutation was confirmed in 2 other pts. In one case, neither JAK2V617F nor JAK2 exon 12 mutation was found. The presence of ten different non-driver mutations (ASXL1, SRSF2, U2AF1, IDH2) in eight of the analyzed pts (5.3%) was confirmed. The overall frequency of thrombotic events (TE) in the studied PV group was 23.8% (36/151). In patients with TE, median platelet count was lower than in pts without TE. Thrombotic risk did not depend on JAK2 rs12343867, TERT rs2736100, OBFC1 rs9420907 SNV, however, we found a novel strong tendency towards statistical significance between the CC genotype miR-146a rs2431697 and thrombosis. The disease progression to fibrotic phase was confirmed in 9% of the pts. Fibrotic transformation in PV pts was affected mainly by JAK2V617F variant allele frequency (VAF) and the presence of coexisting non-driver variants. The high JAK2V617F VAF and elevated white blood cell (WBC) count at the time of diagnosis were associated with an increased risk of death. Conclusion Therefore, in our opinion, complex, laboratory and genetic PV pts evaluation at the time of diagnosis should be incorporated into a new prognostic scoring system to more precisely define the PV prognosis and to optimize the therapeutic decision-making process. Supplementary Information The online version contains supplementary material available at 10.1007/s00432-022-04327-0.


Introduction
Polycythemia vera (PV) together with essential thrombocythemia (ET) and primary myelofibrosis (PMF) is classified as a classical Philadelphia negative myeloproliferative neoplasm (MPN Ph-) (Arber et al. 2016). Among various registries, the annual incidence of PV is estimated at 0.4-2.8 per 100.000 and the prevalence at 1/3.300, with a male/female ratio of 1.2:1. PV occurs at all ages but is most common in subjects aged 50-70 years (Moulard et al. 2014;Shallis et al. 2020a). The disease is mainly characterized by an increase of red blood cells mass and content; an increase of haemoglobin and haematocrit value, and thrombotic tendency, including microvascular complications. Thrombotic episodes at the time of diagnosis are relatively frequent and are observed in 12 to 39% of PV patients (pts) (Tefferi and Elliott 2007;Falanga and Marchetti 2012). More frequently, the microvascular complications occurs, including numbness (66.2%), concentration difficulties (61.2%), sad mood (65%) and sexuality difficulties (56.8%), headache (52.2%), and dizziness (52.1%) (Cuthbert and Stein 2019).
Initial reports documented that 95% of PV pts carry the V617F mutation (exon 14) of the JAK2 gene, resulting in a diminished JH2 inhibitory effect on the adjacent JH1 kinase domain, thus keeping JAK2 in a constitutively active conformation (Chen and Mullally 2014). Most of the JAK2V617F negative cases involves JAK2 exon12 mutations (different in nature-synonymous substitutions, deletion variants, duplications) (Scott 2011). In small number of PV pts, the driving mutation leading to the clinical symptoms occurrence has been not identified yet. The JAK2V617F VAF was shown to have laboratory and clinical implications (Borowczyk et al. 2015). Its quantification, however, is not obligatory for the diagnosis nor the thrombotic risk assessment of patients with PV according to the WHO criteria. It has been shown that JAK2V617F homozygosity positively correlates with erythropoiesis and granulopoiesis, splenomegaly, pruritus, requiring cytoreductive therapy and negatively correlates with the platelet (PLT) count (Vannucchi et al. 2007a). In comparison to the JAK2V617F positive pts, the individuals carrying the JAK2 exon 12 mutation are characterised at the time of diagnosis by younger age, higher Hb level, lower WBC and PLT count. Despite the differences in the laboratory patients' characteristics, the rates of survival and constitutive symptoms are similar to patients carrying the JAK2V617F mutation (Passamonti et al. 2011;Tondeur et al. 2021). However, the molecular landscape of nondriver mutations in not as homogenous as driver mutations. The question of the impact of JAK2V617F VAF, single-nucleotide variants (SNVs) and the coexisting nondriver variants on the individual PV disease course (risk of: thrombosis, fibrotic/blastic progression, death) is still not fully answered.
So far ASXL1, SRSF2, IDH2 have been identified as unfavourable risk factors affecting overall (OS), leukemiafree (LFS), or myelofibrosis-free survival (MFFS) in PV (Tefferi et al. 2016). In 2020, the SRSF2 mutation positivity has been incorporated in the prognostic scoring systems-MIPSS-PV, as an independent genetic risk factor in PV, together with age > 67 years, leukocyte count ≥ 15G/L, and previous thrombosis (Tefferi et al. 2020). A lot of effort has been made to better understand the PV biology and to identify risk factors associated with the unfavourable PV outcome at the time of diagnosis. The risk factors associated with thrombosis in PV have been not fully recognized. The data suggest a relevant role of regulatory miRNAs involved in the platelet activation process and immune system control (Krammer et al. 2022). Among others, the miR-146a, a brake in NF-κB signaling, has been shown as a key mediator of inflammation-induced carcinogenesis and in thromboinflammation (Boldin et al. 2011;Arroyo et al. 2021).
Herein, we present the results of our study evaluating the impact of the factors postulated to be associated with PV occurrence, manifestation and outcome, including the SNVs' genotypes (JAK2 rs12343867, TERT rs2736100, OBFC1 rs9420907, miR-146a rs2431697) and non-driver genetic variants (SRSF2, ASXL1, U2AF1, IDH1, IDH2). Special attention was paid to their impact on the risk of thrombosis, fibrotic progression, blastic transformation, and death of the evaluated patients.

Study group characteristics
The study group consisted of 151 pts recruited from the 2 Polish university centres -the Department of Hematology and Bone Marrow Transplantation of Poznan University of Medical Sciences in Poznan and the Department of Hematology of Pomeranian Medical University in Szczecin, and  (Arber et al. 2016). The grade of the bone marrow fibrosis was assessed according to the European Consensus on grading bone marrow fibrosis and the assessment of cellularity (Thiele et al. 2005). A characteristic of the patients studied is presented in (Table 1)

Materials and methods
DNA was extracted from whole-blood leukocytes at the time of diagnosis or first evaluation at our Department using QIAmp DNA Mini Kit (Qiagen). The assessment for the presence of the JAK2V617F mutation was conducted by quantitative allele-specific RQ-PCR according to Larsen et al. (2007) standardized by cooperation with MPN&MPNr EuroNet (Jovanovic et al. 2013

Statistical analysis
Nominal data were described using counts and percentages for each category. A comparison of such data between the study and control group was performed using the chi-square test or its correction (Fisher's exact test) when the numbers in individual categories were too low. In the case of variables with several categories, when the effect obtained was statistically significant, multiple comparisons were also applied using the chi-square test and Fisher's exact test, but adding the Benjamini-Hochberg correction. In addition, to describe the magnitude of the obtained effect, an odds ratio was determined together with a 95% confidence interval, giving the chance of occurrence of the event in the exposed group in relation to the reference category. A Cochran-Armitage chi-square test for trend was calculated to look for a trend in the obtained proportions. JAK2V617F VAF was presented in four equally sized categories defined by the boundaries: 0%; 25%; 50%; 75%; 100%. Continuous data were described using mean ± standard deviation and median with quartiles. Comparisons of continuous variables according to normality of distribution and equality of variance (tested by Shapiro-Wilk test and Fisher-Snedecor test, respectively) were made using unpaired t test or exact version of Mann-Whitney test dedicated to small sample sizes. The strength and direction of the association between JAK2V617F % and continuous variables were assessed using Spearman's monotonic correlation coefficient with 95% confidence interval and the test of its statistical significance. The survival analysis was performed using Kaplan-Meier curves which were compared with the log-rank test. The predictive ability for individual parameters was examined using receiver operating characteristic (ROC) curves. The size of the area under the curve (AUC) was determined along with the 95% confidence interval, the cut-off point (via Youden's index) for the studied parameters was identified, and the sensitivity and specificity values for it were given. All analyses were performed in PQStat v1.8.4 software. A significance level of 0.05 was assumed.

SNVs' frequency
The distribution of the different SNVs' genotypes (JAK2 rs12343867 T>C, TERT rs2736100 A >C, OBFC1 9420907 A> C and miR-146a rs2431697 C >T) in PV patients and healthy donor groups is presented in Table 3. Statistically significant differences in SNVs' genotypes frequency between PV and control group were found in the case of JAK2 rs12343867 (p = 0.0001) and TERT rs2736100 (p < 0.0001). The differences were evident for each of the genotype subgroups, i.e. for JAK2 rs12343867 TT/TC/CC and TERT rs2736100 AA/AC/CC, which was confirmed by the Benjamini-Hochberg corrected multiple comparison. The risk of PV is the greatest and lowest in the case of JAK2 rs12343867 in CC and TT individuals, respectively, and in the case of TERT rs2736100 in CC and AA individuals, respectively.

JAK2V617F VAF and complete blood count results
JAK2V617F VAF correlated positively with total WBC count, neutrophil (NEU) count, red cell distribution width (RDW), and negatively with PLT count, mean cell haemoglobin (MCH), mean cell haemoglobin concentration (MCHC) (Supplementary Table 2). JAK2V617F VAF was higher in PV JAK2V617F positive patients with coexisting variants than in the JAK2V617F mutation positive only (61 vs 26%. p = 0.0121). No significant differences were found in terms of complete blood count parameters between the groups with and without coexisting variants (Supplementary Table 3).
There was no death due to thrombotic complications. Thrombotic events risk did not depend on the genotypes of JAK2 rs12343867, TERT rs2736100, OBFC1 rs9420907 (Table 4), JAK2V617F VAF nor the presence of additional variants (Table 5). However, it is worth mentioning that in the case of miR-146a rs2431697, the obtained p-value was 0.0511 (Cochran-Armitage test for trend) and the obtained odds ratios were more than twice lower in the CT and TT groups, as compared to CC.

PV progression to the fibrotic phase
The bone marrow trephine biopsy was performed in 73 of PV pts, with the clinical suspicion of disease progression to the fibrotic phase (the presence of at least one symptom of the disease progression, like anemia or sustained loss of requirement of either phlebotomy in the absence of cytoreductive therapy or cytoreductive treatment for erythrocytosis; a leukoerythroblastic peripheral blood picture;  It should be stressed, however, that in 11 other pts qualified for trephine biopsy, the fibrosis grade ≥ 2 was found, but the post-PV-MF was not diagnosed due to the non-fulfillment of the other obligatory IWG Criteria (Barosi et al. 2008).
Median JAK2V617F VAF at the time of PV diagnosis was 74 and 24% in patients who progressed and who did not progress to post-PV-MF, respectively (p = 0.0002) (Fig. 1a). The fibrosis grade depending on the baseline JAK2V617F VAF and Kaplan-Meier curves for MFFS time are presented in Fig. 1b and Supplementary Fig. 1, respectively. The JAK2 rs12343867 CC genotype (Table 6) and the presence of coexisting non-driver variant (Table 5) were identified as strongly associated with an increased risk of post-PV-MF development.  Genotype analysis did not confirm the impact of miR-146a rs2431697, a brake in NF-κB signaling and a key mediator of inflammation-induced carcinogenesis, on the risk of PV progression to the fibrotic phase.

PV progression to the blastic phase
Three out of 151 of PV patients (2%) progressed to the blastic phase during the observation follow-up. The risk of blastic transformation was ten times higher in patients with coexisting non-driver genetic variants (Table 5). A molecular characteristic of the PV patients transformed to the blastic phase is shown in Table 7.

Overall survival of the PV patients
The risk of death in the studied patient group was closely associated with JAK2V617F VAF at the time of diagnosis. A detailed analysis confirmed a simple association: the higher VAF, the lower chance of survival (p = 0.0028) ( Table 8). The association between the different SNVs' genotypes and OS time are presented in Table 9 and in Supplementary  Fig. 2, respectively. JAK2 rs12343867 CC genotype is associated with an increased risk of death than CT.

Factors associated with unfavourable PV outcome by ROC analysis
The  (Fig. 2). For JAK2V617F, the specificity is higher (the ability to exclude the preterm death), while for WBC, the sensitivity is higher (the ability to detect patients who will die). Moreover, the occurrence of coexisting variants is associated with higher JAK2V617F VAF (AUC[95% CI] = 0.77[0.59;0.95], p = 0.0110, cut-off point = 47%).

Discussion
The prognosis in PV pts is comparatively favourable, but individual disease course is very heterogeneous. It is mainly due to the different risk of thrombosis and disease evolution to the fibrotic or blastic phase. The overall survival of PV patients is shorter compared to the age-matched general population (Srour et al. 2016;Abu-Zeinah et al. 2022), superior to that in PMF and inferior to that in ET patients, with estimated median OS of 11.9-14.0, 4.0-6.0 and 12.1-20.0 years, respectively (Tefferi et al. 2014;Shallis et al. 2020b;Smith et al. 2021). According to Kiladjian et al., the risk of progression to the fibrotic phase is 4.9 and 9.4% after 10 and 15 years, and according to Bonicelli    (Kiladjian et al. 2003;Bonicelli et al. 2013). The rates of leukemic transformation are 1.0-3.9% in PV, 1.5-2.6% in ET and 3.3-9.2% in PMF (Szuber et al. 2019;Shallis et al. 2020b;Smith et al. 2021). In the case of PV blastic transformation (secondary acute myeloid leukemia), the reported OS is about 4 months (Abdulkarim et al. 2009). According to the Weill Cornell Medicine (PV-WCM) Research Database Repository data, an excess late mortality was observed after 17 years, potentially due to an increased incidence of post-PV-MF progression, affecting almost 50% of PV pts after 25 years (Abu-Zeinah et al. 2022 (Smith et al. 2021).
The above-mentioned data support the hypothesis that early identification of the factors influencing the individual risk of thrombosis, fibrotic/blastic transformation and death in PV patients may be very helpful in the decisionmaking process of treatment individualization. It has been postulated that individual genetic profile may be responsible for the differences in individual disease outcome. It was shown that 46/1 haplotype (Jones et al. 2009;Olcaydu et al. 2009;Anelli et al. 2018) and TERT rs2736100 A > C genotype predisposed to MPN (Oddsson et al. 2014). Lastly, Giaccherini et al., using a combined "teloscore" from 11 telomere-length-associated SNPs, indicated that genetically determined longer telomeres were a risk marker for the MPN development and also reported a novel association of the OBFC1 rs9420907 C allele and increased MPN risk (Giaccherini et al. 2020). Our result confirm that JAK2 rs12343867 CC and TERT rs2736100 CC are strongly correlated with PV predisposition, as was previously established, but contrary to other reports, we did not find the association between OBFC1 rs9420907 CC genotype nor C allele and PV predisposition (Jones et al. 2009;Olcaydu et al. 2009;Anelli et al. 2018).   (Ferrer-Marín et al. 2020) and overrepresentation of post-PV-MF cases (40 PV vs 21 post-PV-MF) in the second one (Aref et al. 2021). In the available literature also, another polymorphism was associated with an unfavorable MPN Phoutcome. The negative impact of the JAK2 rs12343867 TT genotype on OS (poorer outcome) was also confirmed by Tefferi et al. in patients with PMF (Tefferi et al. 2019). Our results revealed the JAK2 rs12343867 CC genotype was a risk factor of post-PV-MF development and shorter OS in comparison to CT but not TT genotype in PV patients. Cardiovascular complications and cerebrovascular disease are a prominent cause of death, especially in younger patients with PV (Smith et al. 2021). The factors influencing the thrombotic risk in PV pts include age over 60 years and previous thrombosis. However, arterial hypertension, leukocytosis, and hypoxia should be considered as additional risk factors in certain circumstances (Stavik et al. 2016;Narita et al. 2016;Tefferi et al. 2018;Pilli et al. 2018).
We have been the first to find that PLT count lower than 506 G/L at the time of diagnosis is as a predictor of thrombosis. Additionally, we report a novel strong trend towards statistical significance between miR-146a rs2431697 CC genotype and thrombotic risk, as well. The above-mentioned association between the miR-146a rs2431697 CC genotype, thrombosis and platelet count seems to be important, especially evident in the light of 1) the discovery that rs2431697 C allele binds NF-κB more effectively than rs2431697 T allele, increasing the expression of mi-R146a (Hou et al. 2021), 2) the confirmation that mi-R146a activates JAK2/ STAT3 (Wang et al. 2019) and 3) the documentation of the role of both NF-κB and mi-R146a in platelet signaling process, as reviewed by Kojok (Kojok et al. 2019) and Czajka (Czajka et al. 2021). It cannot be excluded that elevated miR-146a expression increases the thrombotic risk in PV pts by the impairment of fibrinolysis. The above-mentioned hypothesis is supported by the observation of Alfano et al., showing that overexpression of miR-146a reduces urokinase-type plasminogen activator receptor (uPAR) expression in AML cell lines (Alfano et al. 2015). Limited data are available on the role of fibrinolysis in the pathogenesis of thrombotic complications in MPN Ph-pts. Most of them postulated hypofibrinolysis as a prothrombotic mechanism in MPN Ph-. Specifically, hypofibrinolysis is driven by elevated levels of plasminogen activator inhibitor-1 (PAI-1) and platelet factor 4 (PF4). These proteins are associated with excessive platelet reactivity, not with the platelet count, which is in line with our results confirming the relationship between lower platelet count and thrombosis risk (Pósán et al. 1998;Birdane et al. 2005;Malecki et al. 2016). It should be stressed that activated platelets and leukocytes in ET patients were found to be the main source of tissue factor (TF) and that reduced activity of the TF pathway inhibitor additionally increases the prothrombotic risk in JAK2V617F positive ET patients (Gadomska et al. 2016).
Several studies have shown an association between leukocytosis and an increased risk of thrombosis (Landolfi et al. 2007;Gangat et al. 2007;De Stefano et al. 2010;Lim et al. 2015;Barbui et al. 2015;Cerquozzi et al. 2017;Carobbio et al. 2019). Data from the REVEAL study (prospective Observational Study of Patients with Polycythemia Vera in US Clinical Practices) signal that an elevated WBC count of more than 11 G/L increases the risk of thrombotic events and suggest the need to incorporate blood count values into the risk stratification (Gerds et al. 2021). Similar to Zhao et al., we did not find an association between the WBC count nor Hb level (Zhao et al. 2016). According to Ronner et al., the risk of thrombosis is not significantly associated with any hematologic laboratory value trajectory (Ronner et al. 2020). The results of our study suggest that rs2431697 CC genotype of the miR-146a, a brake in NF-κB signaling and a key mediator of inflammation-induced carcinogenesis and thromboinflammation, may be associated with thrombotic risk. The latter requires confirmation in larger groups of PV pts.
On the basis of the results of our study, we concluded that the WBC (≥ 12.15 G/L) at the time of diagnosis could serve as a predictor of preterm death in PV patients. This is in good compliance with the results obtained by Ronner et al. who found that persistently elevated leukocyte trajectories were significantly associated with an increased risk of unfavorable outcome of the disease due to the evolution to myelofibrosis, myelodysplastic syndrome, and acute myeloid leukemia (Ronner et al. 2020). For this reason, leukocytosis (≥ 15 G/L) has been determined to be an independent survival factor according to MIPSS-PV (Tefferi et al. 2020).
Another issue which is still debated is the impact of JAK2V617F VAF at the time of diagnosis on the individual disease outcome. JAK2V617F VAF at time of diagnosis in our patients was 31 and 20% in the case of venous and arterial thrombosis, respectively. Similar to Lee et al. (Lee et al. 2021), we did not confirm results which identified JAK2V617F VAF > 50% as an independent strong predictor of venous but not arterial thrombosis (Cerquozzi et al. 2017;Guglielmelli et al. 2021). Our results also documented that high JAK2V617F VAF at the time of diagnosis was associated with the risk of PV progression to post-PV-MF, which had been established previously by others (Vannucchi et al. 2007b;Passamonti et al. 2010;Alvarez-Larrán et al. 2014;Senín et al. 2018). The presence of such an association was neglected by Lee et al. (2021). Our results confirmed also that higher JAK2V617F VAF at the time of diagnosis was associated with impaired OS, as was also reported by Lee et al. (2021).
Complex molecular fingerprint seems to be an indispensable element of PV prognosis. It was documented that the presence of non-driver mutations in PV patients significantly changed the PV outcome. Until now, a large spectrum of coexisting somatic gene mutations affecting epigenetic regulation, messenger RNA splicing, signalling, transcriptional regulation and DNA repair in PV pts have been identified. In a study by Tefferi et al., the most frequently identified mutations in JAK2V617F positive pts were TET2, ASXL1 and SH2B3, with the frequency of 22%, 12%, 9%, respectively. In a study by Song et al., the mutations of ASXL1, KMT2A and TP53 coexisted with the driver mutation with the frequency of 6.25, 13.64 and 6.25%, respectively. Rarely, the presence of other gene mutations including splicing machinery genes such as SRSF2, U2AF1, SF3B1 or ZRSR2 has been confirmed (Tefferi et al. 2016;Song et al. 2017).
We detected 10 non-driver gene variants in 8 of 151 pts. Half of them experienced disease progression during the follow-up. Two pts -the first carrying the U2AF1 and the second carrying the ASXL1 variant, progressed to post-PV-MF after 184 and 200 months, respectively. The patient carrying the IDH2 variant progressed to the fibrotic and blastic phase after 288 and 312 months of follow-up, respectively. The most rapid fibrotic progression we observed in a patient carrying the ASXL1 splice variant and SRSF2 variant (Kanduła et al. 2022). Four of our pts, carrying the ASXL1 mutation, who have not progressed so far, were observed for 36, 48 and 180 months, respectively. When it comes to the analysis of ASXL1 solely, we detected variants of this gene in 3% of pts who did not progress to post-PV-MF and in 23% who had, which it is in good agreement with results of Guo et al. (4 vs 26%) who demonstrated that PV patients with coexisting mutations of ASXL1 had a poor MFFS (Guo et al. 2019). A study by Senín et al. indicates that the coexistence of additional genetic variants, especially in SF3B1 and IDH1/2, is a strong predictor of MF transformation (Senín et al. 2018). According to the results of Lee et al., the SF3B1, IDH1/2 and ASXL1 variant occurrence is associated with fibrotic disease progression, and RUNX1, TP53 and IDH1/2 are associated with high leukemic transformation risk. ASXL1, Fig. 2 The ROC curves of the factors influencing the PV outcome a the JAK2V617F VAF and WBC as death predictors b the PLT count as thrombosis predictor c JAK2V617F VAF as predictor of PV evolu-tion to fibrotic phase (post-PV-myelofibrosis) d JAK2V617F as predictor of coexisting variants 1 3 SRSF2, IDH2 were classified by Tefferi et al. as adverse variants/mutations, in terms of overall-(median survival 7.7 versus 16.9 years), leukemia-free-or myelofibrosis-free survival, with a combined prevalence of 15% (Tefferi et al. 2016). Bartels revealed that patients with SRSF2, U2AF1, and IDH1/2 mutations at the time of diagnosis showed rapid transformation to the blastic phase and proposed to consider these mutations as negative predictors of rapid blastic progression in newly diagnosed MPN Ph-at the chronic stage (Bartels et al. 2021).
In conclusion, our study confirms the necessity of a complex clinical, laboratory and genetic PV patients work-up. It especially concerns the JAK2V617F VAF and ASXL1, SRSF2, IDH2, U2AF1 mutation screening at the time of diagnosis to more accurately define the risk of progression to the fibrotic and/or blastic phase. The presence of the above-mentioned aberrations was confirmed by us in only 5.3% of PV patients. However, even in this case, the confirmation of their presence may improve the individual risk stratification and allow to optimize the therapeutic decision-making process in patients at risk of unfavourable disease outcome.
Acknoweldgments The authors would like to thank Katarzyna Lewandowska for proofreading the final version of the manuscript.
Author contributions ZK and KL: designed the study. ZK: prepared material and carried out molecular genetic testing. ZK, MJ, EP: collected data. BW: performed the statistical analysis. ZK and KL: interpreted data. ZK: wrote the manuscript. KL: reviewed and edited the manuscript. All authors read and approved the final manuscript.
Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Conflict of interest
The authors have no relevant financial or non-financial interests to disclose.
Ethical approval This study was conducted in accordance with the Declaration of Helsinki. The study protocol was approved by the Ethics Committee of Poznan University of Medical Sciences (number: 1056/16, 181/18 and 846/21).

Consent to participate Informed consent was obtained from all individual participants included in the study.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.