Personality traits and emotional domains have been hypothesised to influence surgery outcomes. This cross-sectional study explored associations between alexithymia and negative affectivity and early postoperative outcomes following colorectal resection.
MethodsFrom a cohort of patients who underwent laparoscopic colorectal surgery at the Division of General and Hepatobiliary Surgery, University of Verona Hospital Trust between March 2015 and November 2021, 146 patients completed the “Toronto Alexithymia Scale 20 items” (TAS-20) and the “Type-D Scale” (DS14) during a postoperative interview. Clinicopathological data and postoperative outcomes were prospectively collected. Statistical analysis included regression models using the TAS-20 and DS14 negative affectivity subscale as explanatory variables, adjusting for sociodemographic factors and correcting for multiple testing according to Benjamini-Hochberg.
ResultsOf the 146 patients, 43.1% met criteria for alexithymia, while 43.2% met criteria for negative affectivity. An association was found between negative affectivity and increased odds of postoperative complications (OR 4.42; p = 0.002) and need for analgesic intervention after the third postoperative day (OR 3.79; p = 0.015). Higher levels of alexithymia were associated with longer time to achieve a numeric pain rating scale score < 4 (β 0.062; p = 0.038) on a scale going from 0 to 10.
ConclusionsIn this cross-sectional study alexithymia and negative affectivity were associated with selected early unfavourable postoperative outcomes. While these results may have implications for improving quality of life of individuals undergoing major surgery, they require confirmation in larger prospective studies that include preoperative psychological assessment.
Psychological factors and personality traits have been shown to significantly influence perioperative outcomes.1–3 Among these factors, alexithymia and negative affectivity deserve particular attention, as they represent two distinct constructs that may both shape the perioperative trajectory through partially independent pathways. Alexithymia is conceptualised as a cognitive-affective processing deficit characterised by difficulties identifying and describing one's own feelings and distinguishing emotions from bodily sensations.4,5 Because emotional information is poorly processed at a symbolic level, somatic correlates of distress tend to be amplified and misinterpreted as physical symptoms, with poorer outcomes in conditions such as neurological disorders, chronic pain, and cardiovascular disease due to heightened pain sensitivity and limited stress regulation.4–6 Negative affectivity, on the other hand, is a stable emotional temperament reflecting the dispositional tendency to experience aversive emotional states such as anxiety, irritability, and sadness across situations and over time.7 Through heightened stress reactivity and a pessimistic appraisal of bodily sensations, it has been associated with poorer recovery, greater postoperative pain, and an adverse long-term prognosis in several chronic conditions.7–9
Thanks to the introduction of laparoscopy in the 1990s, colorectal surgery procedures have reduced surgical trauma, facilitated earlier mobilisation, and improved functional recovery without compromising safety outcomes.10,11 At the same time, Enhanced Recovery After Surgery (ERAS) protocols, developed by Kehlet and colleagues, optimised perioperative care, shortened hospital stays, and improved outcomes, even in open surgery.12–14 Integrating ERAS with minimally invasive techniques was a logical progression, yielding fewer complications and a quicker return to daily life.15,16 ERAS is now considered the standard of care in colorectal surgery thanks to its multidisciplinary, cost-effective approach emphasising prehabilitation through physical preparation, medical optimisation, nutritional support, and patient engagement.17,18
Although prior studies have linked alexithymia and negative affectivity to surgical outcomes in other populations,19–21 and Type D personality has been associated with poorer quality of life and slower recovery in colorectal cancer survivors,9,22 no study has investigated the role of these traits in patients undergoing minimally invasive colorectal resection within an ERAS protocol. Because ERAS standardises perioperative care and minimises iatrogenic variability,23 residual differences in early postoperative outcomes may be more closely related to patient-level factors, including stable psychological traits.24 We therefore hypothesised that higher levels of alexithymia and negative affectivity would be independently associated with poorer early postoperative outcomes.
Materials and methodsCohort under study and inclusion criteriaData from patients who underwent elective minimally invasive colorectal resection for colorectal neoplasia or diverticular disease between March 2015 and November 2021 were retrieved from the prospective database of the Division of General and Hepatobiliary Surgery at the University of Verona Hospital Trust, Italy, and considered for study accrual. All patients aged 18 to 80 years who underwent elective minimally invasive surgery for colorectal neoplasia or diverticular disease within our institutional ERAS program were originally eligible for inclusion (n = 710). Exclusion criteria included: age below 18 or above 80 years (n = 110), open surgery (n = 200), urgent surgery (n = 44), patients suffering from inflammatory bowel disease or familial adenomatous polyposis (n = 21), or formally diagnosed mental health disorders or with a history of drug or alcohol addiction (n = 15). Patients were also excluded if procedures required a permanent stoma (n = 34), in case of palliative surgery (n = 8), or recurrent cancer (n = 3) (participants may present more than one exclusion criterion). This left a cohort of 410 clinically eligible patients. Among these, 264 were unwilling or unable to participate in the psychometric assessment, leaving 146 patients who completed the questionnaires and were included in the analysis. Fig. 1 illustrates this two-stage selection process.
The study was approved by the Institutional Review Board and Ethics Committee of the University of Verona Hospital (ID: 42763 - CRINF-1034 CESC). Informed consent was obtained from all participants at the time of the interview. The manuscript was prepared in accordance with the STROBE guidelines for observational studies.25
Management of surgical conditionsPreoperative work-up and histopathology staging have been previously reported in detail.26 All patients were staged with colonoscopy, chest-abdomen and pelvis computed tomography (CT), and carcinoembryonic antigen (CEA). Other imaging modalities were performed on a case-by-case basis. Pathology specimens were analysed and reported according to the 8th Edition of the AJCC and UICC guidelines as pTNM.27 Adjuvant chemotherapy was reserved for cancers staged as pTNM Stage III or pTNM Stage II with high-risk features in the absence of general contraindications. Post-operative follow-up was regularly performed according to pTNM stage through imaging modalities, colonoscopy, and CEA measurement for at least 5 years. The indication for sigmoidectomy or left hemicolectomy was based on clinical symptoms, colonoscopy, and/or CT colonography findings. After surgical resection, patients underwent follow-up visits at 3 and 12 months, including a colonoscopy at 12 months. Thereafter, no routine follow-up was scheduled, and patients were re-evaluated upon request in case of symptom recurrence.
Perioperative management was conducted in accordance with the ERAS® guideline,17 which has been the standard of care for patients undergoing colorectal surgery since 2014. Intravenous antibiotics, thromboprophylaxis, and prophylactic antiemetics were administered in all cases. Pain control was achieved through locoregional anaesthesia, both with transversus abdominis plane block or surgical wounds infiltration, while epidural or spinal analgesia was never adopted.28 Nasogastric tube was placed intraoperatively and removed at the end of the procedure. Additionally, measures were taken to maintain normothermia and avoid intraoperative overhydration.
Patients were allowed to drink fluids on the day of surgery and advanced to a solid diet as tolerated starting from postoperative day (POD) 1. Intravenous fluids were discontinued on POD 1, along with the removal of the urinary catheter. Patients were encouraged to sit and walk as soon as possible.
Postoperative pain was managed with opioid-sparing multimodal analgesia. Patients requiring an overnight stay in the intensive care unit (ICU) adhered to the ERAS® protocol from POD 1.
During the study period, multiport laparoscopic, single-incision laparoscopic, and natural orifice specimen extraction procedures were performed. The site for minilaparotomy was determined by the surgeon, as was the choice between intracorporeal and extracorporeal anastomosis. Conversion was defined as the inability to complete all intended steps using a minimally invasive approach.
PredictorsEach participant who underwent surgery during the study period and fulfilled the inclusion criteria was contacted by a dedicated researcher with appropriate clinical expertise. Following acquisition of consent, the researcher performed a structured interview on several demographic and clinical variables and administered the “Toronto Alexithymia Scale 20 items” (TAS-20) and the “Type-D Scale 14 items” (DS14). The TAS-20 is a 20-item self-reported instrument developed to assess alexithymia. Respondents rate each item on a five-point Likert scale ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). The cumulative score ranges from 20 to 100, where scores ≤ 51 are indicative of non-alexithymic profiles, scores between 52 and 60 suggest possible alexithymia, and scores ≥ 61 are considered reflective of alexithymia.29 The original TAS-20 has shown good internal consistency (Cronbach's α = 0.81), satisfactory three-week test-retest reliability (r = 0.77), and a stable three-factor structure across clinical and non-clinical samples.29 The Italian version used in the present study showed adequate internal consistency (Cronbach's α between 0.75 and 0.82) and test-retest reliability.30
The DS14 is a 14-item, self-reported instrument designed to assess the Type D (distressed) personality construct, which is conceptualised as the interaction of two personality traits: negative affectivity and social inhibition. The DS14 comprises 14 items, with 7 per subscale. The scale is scored using a five-point Likert scale ranging from 0 (“false”) to 4 (“true”), yielding a total score ranging from 0 to 56.31 The original DS14 has demonstrated good internal consistency (Cronbach's α = 0.88 for negative affectivity and 0.86 for social inhibition) and three-month test-retest reliability (r = 0.72 and 0.82, respectively).31 The Italian version used in the present study was validated in cardiac and general samples and replicated the original two-factor structure with adequate reliability indices.32
For the purposes of this study, we did not include Type D personality as a predictor because the construct, defined as a combination of negative affectivity and social inhibition, is controversial, and simulation studies have shown that apparent Type D associations may be accounted for by the influence of one underlying trait rather than a true synergistic effect.33,34 Conversely, social inhibition primarily reflects interpersonal behavioural style and was not considered a plausible correlate of postsurgical complications, including pain. We therefore focused on negative affectivity as the clinically relevant DS14 dimension, given its stronger links to stress reactivity, emotional processing, and pain perception.35–37 The negative affectivity subscale is categorically classified: individuals scoring ≥10 are designated as having negative affectivity.
Baseline demographic and clinical characteristics, surgical data, and post-operative course were retrieved from a prospectively maintained retrospective database. The TAS-20 and DS14 were administered after surgery, beyond the immediate postoperative phase. Comorbidity status was assessed using the American Society of Anesthesiologist (ASA) Physical Status Classification System38 and the Charlson-Age Comorbidity Index (CCI).39
OutcomesWe categorised outcomes into three groups: “post-operative complications”, “recovery process”, and “post-operative pain”. “Post-operative complications” include “length of stay” (continuous), indicating the duration of hospitalisation after surgery in days and “presence and severity of complications” (dichotomised) based on the Clavien-Dindo system.40 Presence and severity of complications were differentiated as no complications (Clavien-Dindo 0), and moderate (grades 1–2) or severe (grade 3 or greater) complications. The “post-operative pain” group of outcomes includes the number of days required for the patient to achieve a numeric rating scale (NRS) score of less than 4 out of a pain scale going from 0 to 10 (continuous), and the need for analgesic intervention (binary), which we expressed as the patient requiring at least three doses of painkillers (PKs) after POD 3. The “recovery process” includes post-operative nausea and vomiting (binary), prolonged postoperative ileus (binary), the day of the first bowel movement (continuous), the day solid oral intake was resumed (continuous), the POD the patient first ambulated (continuous), and the POD the urinary catheter was removed (continuous).
Statistical analysisFirst, we identified three subpopulations based on alexithymia level, according to the commonly used clinical cutoffs of the TAS-20: individuals without alexithymia (score ≤51); individuals with possible/borderline alexithymia (score between 52 and 60); individuals with definite alexithymia (score ≥61).29 Differences in sociodemographic and clinical variables between these groups were examined using the χ2 test for categorical and the Kruskal-Wallis test for continuous variables.
Second, we used the TAS-20 total score and the DS-14 negative affectivity subscale as predictors for the set of postoperative outcomes described above (in the case of the day of the first bowel movement and the post-operative day the urinary catheter was removed, the logarithmic transformation was adopted to reduce right-skewness in data). Initially, we used three global tests to evaluate the joint statistical significance of the predictors across the three groups of outcomes: postoperative complications, recovery process, and postoperative pain (step 1). These global tests were conducted using seemingly unrelated estimation,41 allowing simultaneous examination of the predictors' effects on multiple dependent variables. Ordinary least squares regressions were performed for continuous outcomes, and probit regressions for binary outcomes. The Benjamini-Hochberg correction42 was used to control for Type I error inflation due to repeated testing. Subsequently, each group of outcomes showing global significance was further explored through hypothesis testing on the set of parameters related to each of the two predictors separately (step 2). In case of further statistical significance, results from every single regression were considered (step 3). For dichotomous outcomes, we performed logistic models by penalised maximum likelihood regression43 using the “firthlogit” command,44 to address the issue of quasi-separation. For continuous outcomes, linear regression models with robust standard errors have been used. All analyses were adjusted for age, gender, body mass index (BMI), education level, occupation, alcohol intake, smoke, number of psychosomatic symptoms, presence/absence of social inhibition according to the DS14 subscale, ASA score, type of condition (diverticulis vs neoplasia), CCI, former neoplastic disease, number of prescribed medications (any type), drug regimen inclusive of psychotropics and associated surgical procedures (stoma needed, surgery duration). Analyses were performed through Stata 19.45
ResultsSociodemographic and clinical characteristics for the overall study population are reported in Table 1. A comparison between included (n = 146) and excluded (n = 264) participants showed no substantial differences across demographic, clinical, oncological, or surgical characteristics, apart from a slightly lower proportion of TNM stage III tumours among included patients (16%vs 29% for included and excluded participants, respectively; p = 0.002). The analysis included 146 participants. The mean age of the participants was 62.2 years (SD 10.6), and the mean BMI of the participants was 26.3 (SD 4.9). Male and female participants were represented in nearly equal proportions. Regarding education level, most participants had middle (37%) or high school education (38%), whereas 15% had elementary and 10% university education. The employment status of the participants varied, with the largest categories being unemployed (28.1%), employees (19.9%), and retired (19.9%). In terms of lifestyle factors, 56.8% reported no lifetime alcohol consumption, 28.1% consumed alcohol occasionally, 9.6% regularly, 3.4% were ex-drinkers, and 2.1% were active drinkers. Forty-eight per cent were non-smokers, 38.4% were ex-smokers, 8.9% were active smokers consuming less than 20 cigarettes/day, and 4.8% were active smokers consuming more than 20 cigarettes/day. As for clinical characteristics, the mean number of psychosomatic symptoms was 0.5 (SD 1.0). Most participants (82.2%) had an ASA score of 2. The diagnoses that led to the surgical procedure were split between diverticulitis (43.8%) and neoplasia (56.2%). Aside from the participants who underwent surgery because of cancer, most participants (86.3%) had never had a cancer diagnosis before, and 76.7% had undergone previous surgery. The mean CCI (age-adjusted) was 4.06 (SD 1.86). The mean number of drugs prescribed was 1.65 (SD 1.94), and 14.4% of participants were prescribed psychotropics. Regarding surgical details, 21.9% had an associated surgical procedure, and 1.4% of participants required conversion from laparoscopy to open surgery. A temporary stoma was needed in 11.6% of patients. The mean surgery time was 246.7 min (SD 87.4). For those with neoplasia, the TNM stages were distributed as 46.3% with stages < 2, 30.5% with stage 2, 19.5% with stage 3, and 3.7% with stage 4. Regarding psychological characteristics, the overall population showed a mean score of 51.8 (SD 8.0) on the TAS-20, with 83 individuals (56.8%) without alexithymia, 41 individuals (28.1%) with possible/borderline alexithymia, and 22 individuals (15.1%) with definite alexithymia. “Negative Affectivity” was found in 43.2% of participants, according to the DS14. According to the same scale, 30.8% could be considered “Socially Inhibited”.
Sociodemographic and clinical characteristics.
Legend.
ASA: American Society of Anesthesiologists; BMI: Body Mass Index; CCI: Charlson Comorbidity Index; SD: standard deviation; TAS: Toronto Alexithymia Scale; TNM: tumor, node, metastasis.
The bivariate approach assessing differences in the distribution of sociodemographic factors showed no differences between the comparison groups. Clinical variables across such categories showed that people with definite alexithymia were significantly more likely to suffer from psychosomatic symptoms and be prescribed psychotropic medications. Furthermore, individuals with definite alexithymia were more likely to have negative affectivity, and participants with no alexithymia were less likely to be socially inhibited.
Table 2 reports the findings of the three-step multivariate approach that we first implemented to investigate the joint effects of the two predictors (TAS-20 total score and DS-14 negative affectivity subscale) on the three postoperative outcome groups: “post-operative complications”, “recovery process”, and “post-operative pain” (step 1). The global tests indicated statistically significant multivariate associations between the psychological predictors and both “postoperative complications” (adjusted p = 0.013) and “postoperative pain” (adjusted p = 0.015). No significant global association was observed for the “recovery process” domain (p = 0.572). When disaggregated (step 2), the global conditional association of negative affectivity remained significant for both complications (p = 0.004) and pain outcomes (p = 0.021), whereas the TAS-20 total score was only slightly significant for pain outcomes (p = 0.046). We provided further insight into the above-mentioned significant associations with subsequent analyses of individual outcomes, within each domain (step 3). Within the post-operative complications group, the presence of negative affectivity was associated with a significantly increased likelihood of experiencing moderate to severe postoperative complications according to the Clavien-Dindo system (OR 4.427; 95%CI 1.745 to 11.232; p = 0.002), although it was not significantly associated with length of hospital stay (β 0.94; 95%CI –0.63 to 2.51; p = 0.240). Higher TAS-20 scores were significantly associated with an increased number of days needed to achieve an NRS score below 4, in terms of postoperative pain (β 0.062; 95%CI 0.003 to 0.121; p = 0.038), suggesting that higher levels of alexithymia are associated with a longer time to satisfactory pain control (Fig. 2).
Results of the three-step regression model.
Significant results are in bold.
*Includes: post-operative nausea and vomiting; prolonged postoperative ileus; POD of the first bowel movement; POD of resuming solid oral intake; POD of first ambulation; POD of removal of the urinary catheter.
NA: negative affectivity; NRS: numeric rating scale; POD: Postoperative day; TAS: Toronto Alexithymia Scale.
Conversely, negative affectivity was significantly associated with the need for three or more PKs per day after postoperative day 3 (OR 3.793; 95%CI 1.300 to 11.064; p = 0.015). This implies that having negative affectivity (rather than not having it) is associated with a higher probability of Clavien-Dindo complications from 21.57% to 50.87%, and with a higher probability of needing at least three PKs after POD 3 from 13.95% to 33.57%, with all other variables held equal.
DiscussionIn this cross-sectional study of patients undergoing elective minimally invasive colorectal resection within an ERAS program, we identified exploratory associations between specific psychological traits and early postoperative outcomes. Negative affectivity, as identified by the dedicated subscale of the DS14, was associated, after adjustment for relevant covariates, with an increased risk of moderate-to-severe complications and with a greater likelihood of requiring analgesic intervention beyond the third postoperative day. At the same time, people scoring high levels on the TAS showed an association with delayed pain resolution after surgery, as indicated by a longer time to achieve satisfactory pain control. While these associations are adjusted for a broad range of demographic and clinical covariates, the wide confidence intervals around point estimates indicate statistical imprecision.
The prevalence of alexithymia (15%) and negative affectivity (43%) in our cohort is consistent with that observed in the general population. Alexithymia has been observed in roughly 10–13% of the general population, with some studies reporting rates as high as 17%.46,47 By comparison, traits of negative affectivity are present in about 13–32% of individuals overall,9 and their prevalence increases to 18–53% among patients with cardiac conditions.48 These psychological traits are common in the general population, yet they are often overlooked in routine clinical practice.49 Our findings suggest that they may be associated with selected surgical outcomes.
To facilitate the understanding of the patient's journey after the surgical procedure, we organised the outcomes in three groups addressing both adverse events and the progression towards functional recovery. The “post-operative complications” group of outcomes delineates various untoward events that may arise in the immediate and early post-surgical period. The “recovery process” group is dedicated to tracking the patient’s physiological and functional recuperation. Lastly, the “post-operative pain” group is representative of the need for analgesic administration and, consequently, of the patient’s experience of pain after surgery. We then adopted a stringent three-step regression framework to minimise the likelihood of type I errors. By first examining overall associations between predictors and outcome domains, then testing predictors conditionally, and considering individual outcomes only when justified, we were able to substantially limit the number of statistical tests conducted. This hierarchical strategy, combined with the Benjamini-Hochberg correction, increases confidence in the robustness of the observed associations.50–52 The finding that alexithymia was associated with slower pain resolution, a finding generally known for chronic pain populations,53–55 reinforces the case also for the acute surgical setting.19 Alexithymic individuals may have diminished capacity to recognise and articulate internal states, leading to suboptimal communication of symptoms, reduced engagement in active coping strategies, and increased reliance on pharmacological pain control. Furthermore, alexithymia has been associated with altered central pain modulation, which may prolong nociceptive processing after surgical injury.56,57 Our analysis did not find significant associations between either psychological predictor and the “recovery process” group of outcomes, which included, among others, time to first bowel movement, resumption of solid oral intake, and early ambulation. One possible explanation is that recovery milestones in our cohort were strongly standardised under ERAS protocols, thereby reducing variability and limiting the detectable association of psychological traits with these outcomes. Alternatively, these parameters may be less sensitive to psychological modulation in the immediate postoperative setting.
Several limitations warrant discussion. First, as we set out to examine emotional traits pertinent to affective regulation as correlates of postoperative outcomes, only the negative affectivity subscale of the DS-14 was considered, as it reflects a stable disposition toward emotional distress. Although the social inhibition subscale mainly identifies interpersonal behavioural restraint and does not align with the theoretical framework of the study, it was included as a confounding factor in the statistical model. Second, both psychological measures relied on self-report, which may be affected by limited insight or social desirability bias. Third, the study was conducted at a single centre and only included patients undergoing minimally invasive colorectal resection. For this reason, findings may not be generalisable to other settings, surgical populations, or outside of ERAS programs. Fourth, while we adjusted for a comprehensive set of potential confounders, unmeasured variables, such as depression symptomatology, anxiety personality traits, or preoperative pain, may have contributed to the observed associations. Fifth, the participants were self-selected: only 146 of the 410 eligible patients agreed to be interviewed. This may have introduced selection bias, limiting the representativeness of the sample and the generalisability of the findings. Sixth, the categorical TAS-20 cut-offs used to describe the sample were derived from psychiatric and mixed clinical populations, and their predictive validity in surgical patients has not been formally established. This limitation does not affect the main analyses, which were based on the TAS-20 total score as a continuous variable. Seventh, no physiological measures of stress reactivity or systemic inflammation (e.g., cortisol, heart rate variability, inflammatory markers) were collected, which limits our ability to test the biological pathways through which alexithymia and negative affectivity may influence postoperative outcomes. Eight, included participants had a lower proportion of TNM stage III tumours compared to excluded participants. Patients with advanced-stage disease may have been less available for participation due to greater postoperative burden, potentially biasing the sample toward less complicated cases. This could attenuate observed associations with adverse outcomes and limit generalisation to populations with more advanced disease. Finally, the TAS-20 and DS14 were administered after surgery. Although alexithymia and negative affectivity are conceptualised as stable traits and their assessment at a distance from surgery is expected to reflect preoperative dispositions, retrospective administration may be affected by state-dependent reporting biases, including recall of the surgical experience. The cross-sectional nature of this assessment also reinforces the need to interpret all observed associations as non-causal.
The significance of these results emerges in two ways. First, they may be regarded as hypothesis-generating, raising the possibility that brief preoperative psychological assessment could help characterise patients potentially at greater risk of less favourable early outcomes.58 Second, they may motivate future research aimed at testing more focused prehabilitation measures alongside ERAS protocols, for example by paying attention to personality traits that could slow down recovery.59 If such issues are recognised before surgery, they could inform the design of prehabilitation pathways, pending confirmation in adequately powered prospective studies.
Future research should confirm these associations in larger, prospective cohorts with preoperative assessments and with long-term outcomes, including chronic postsurgical pain and quality of life. Studies integrating psychometric assessment with physiological markers of stress and inflammation may also clarify the biological pathways through which personality traits are associated with recovery.
Ethical considerationsThe study was approved by the Institutional Review Board and Ethics Committee of the University of Verona Hospital (approval number: 42763 - CRINF-1034 CESC). All procedures performed in this study were conducted in accordance with the ethical standards of the institutional research committee and with the Declaration of Helsinki. Written informed consent was obtained from all participants.
FundingThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflict of interestThe authors declare that they have no relevant financial or non- financial interests related to the preparation and submission of this manuscript.
CRediT authorship contribution statementDavide Papola: Conceptualization, Methodology, Investigation, Data curation, Writing – original draft, Writing – review & editing. Giulia Turri: Conceptualization, Methodology, Writing – review & editing. Federico Tedeschi: Methodology, Formal analysis, Data curation, Writing – review & editing. Corrado Barbui: Methodology, Writing – review & editing, Supervision. Giovanni Ostuzzi: Conceptualization, Methodology, Writing – review & editing, Supervision. Corrado Pedrazzani: Conceptualization, Methodology, Writing – review & editing, Supervision.
