The present study was designed as a cross-sectional, single-centre study conducted in the General Surgery Clinics of Ankara Numune Training and Education Hospital in Ankara, Turkey, between May 2010 and May 2014. The Institutional Ethics Committee approved the study protocol prior to its commencement (Scientific Research Evaluation Committee of Ankara Numune Training and Research Hospital, Number: 20796219-E.Kurul-869/2014). The study was conducted in accordance with the ethical considerations outlined in the Declaration of Helsinki. Written informed consent could not be obtained from the patients due to the study’s retrospective design and the anonymity of the data. Data regarding patients who underwent PD were collected throughout 2014, with the last patient included in the study undergoing surgery in the last quarter of 2014. To account for the postoperative 30-day mortality period, data collection began in the first quarter of 2015. Following the completion of data collection in early 2015, data analysis and interpretation were conducted, ensuring that all relevant outcomes were thoroughly evaluated. This timeline allowed the authors to capture comprehensive postoperative outcomes and perform a detailed analysis to accurately interpret the data.

Inclusion and Exclusion criteria: All consecutive adult patients aged 18 years and above who underwent elective PD for resectable malignant periampullary tumours were included. Patients with incomplete medical records, benign final postoperative histopathology, those who underwent PD along with other procedures and those who received PD after neoadjuvant chemotherapy were excluded from the study. In the end, the study sample consisted of 64 patients.

Patients’ preoperative demographic (age, gender), clinical (co-morbidities, ASA class) and haematological and biochemical characteristics were collected and recorded in a prospectively maintained database for hepatobiliary surgical patients. Patients’ imaging findings and interventions were noted, including Percutaneous Transhepatic Cholangiography (PTC), Endoscopic Retrograde Cholangiography (ERCP) and biliary stenting. Patients’ periampullary tumours were classified according to their location: pancreatic head, distal common bile duct, ampulla of Vater and the second part of the duodenum.

The same surgical team performed PD on all patients using a standard technique [2]. Intraoperative data, including the PD subtype (i.e., pylorus-preserving or classical), duration of surgery, anastomotic details, blood loss and need for transfusion, as well as postoperative morbidity and mortality data, were recorded in a prospectively maintained database. Complications occurring within 30 days after PD were categorised as major or minor [15]. Major complications included anastomotic leakage, intra-abdominal abscess, pulmonary embolism and significant cardiovascular and respiratory adverse events, such as myocardial infarction. Minor complications encompassed morbidities such as surgical site infections, evisceration, atelectasis and delayed gastric emptying.

The PD-specific complications, including pancreatic fistula, the grade of pancreatic fistula, delayed gastric emptying, haemorrhage and bile leakage, were diagnosed based on the International Study Group of Pancreatic Surgery Guidelines [16-18]. Postoperative mortality, defined as death within 30 days of PD due to a medical or surgical cause, was the primary outcome of the study. The impact of sociodemographic, clinical and intraoperative characteristics on mortality was considered the secondary outcome.

The descriptive statistics from the collected data were tabulated using either mean±standard deviation or median with minimum and maximum values for continuous (numerical) variables, depending on whether they conformed to a normal distribution. For categorical variables, numbers and percentage values were used. The normal distribution characteristics of the numerical variables were analysed using the Shapiro-Wilk, Kolmogorov-Smirnov and Anderson-Darling tests. Pearson’s Chi-square, Fisher’s-exact and Fisher-Freeman-Halton tests were employed to compare the differences in categorical variables between the groups in 2×2 tables with five or more expected cells, 2×2 tables with expected cells fewer than five and RxC tables with expected cells fewer than five, respectively. The independent samples t-test and the Mann-Whitney U test were used to compare differences in numerical variables that conformed or did not conform to a normal distribution between the groups. Statistical analyses were performed using Jamovi Project 2.3.28 (Jamovi, version 2.3.28, 2023, retrieved from https://www.jamovi.org) and JASP 0.17.3 (Jeffreys’ Amasing Statistics Program, version 0.17.3, 2023, retrieved from https://jasp-stats.org) software packages. A probability (p) value of ≤0.05 was deemed to indicate statistical significance.

The study sample consisted of 64 patients with a mean age of 60.1±12.4 years. Approximately three-fifths 39 (60.9%) of the patients were male. In terms of anatomic locations, the most common site was the pancreatic head 31 (48.44%), followed by the ampulla of Vater 19 (29.69%), the distal common bile duct 10 (15.63%) and the second part of the duodenum 4 (6.25%). The other clinical characteristics of the patients are provided in [Table/Fig-1]. A total of 12 (18.75%) patients died within 30 days of PD. Non survivors were significantly older (p=0.006) and had significantly higher ASA classes (3 and 4) (p=0.008). There was no significant difference in other demographic and clinical characteristics between the groups (p>0.05) [Table/Fig-1].

The preoperative laboratory data of the survivors and non survivors are presented in [Table/Fig-2]. Accordingly, non survivors had significantly lower Neutrophil-to-lymphocyte Ratios (NLR) (p<0.001) and serum albumin levels (p=0.038) than survivors. There was no significant difference in Carbohydrate Antigen (CA) 19-9 and carcinoembryonic antigen levels between the groups (p>0.05) [Table/Fig-2].

Of the 64 patients, 34 (53.1%) underwent Endoscopic Retrograde Cholangiopancreatography (ERCP), three (4.7%) underwent Percutaneous Transhepatic Cholangiography (PTC) and 27 (42.2%) underwent biliary stenting preoperatively. Classical PD 51 (79.7%) and duct-to-mucosa anastomosis 47 (73.4%) were performed more frequently than pylorus-preserving PD 13 (20.3%) and dunking anastomosis 17 (26.6%), respectively. The rate of patients experiencing blood loss exceeding 1000 mL intraoperatively was significantly higher in non survivors compared to survivors (41.67% vs. 7.69%, p<0.009) and therefore the median number of blood units transfused was significantly greater in non survivors than in survivors (p<0.001) [Table/Fig-3].

All non survivors experienced at least one complication. Both general complications and major complications were significantly more common in non survivors than in survivors (p=0.002 and p=0.012, respectively). However, there was no significant difference in the incidence of anastomotic leakage and pancreatic fistula between the groups (p=0.351 and p=0.999, respectively) [Table/Fig-4].

The study findings revealed that PD performed for malignant periampullary tumours was associated with a high 30-day postoperative mortality rate, which varies according to the patients’ clinical characteristics and intraoperative findings. Significant risk factors for 30-day mortality after PD included older age, higher ASA classes, intraoperative blood loss, perioperative blood transfusion and postoperative complications.

Postoperative complications, overall morbidity and mortality rates reported in the literature vary significantly depending on various factors, including patient-related, tumour-related and surgeon-related parameters [2-4,8,11,19]. In a study conducted with 1,348 patients who underwent PD for malignancy, Russell TB et al., reported overall morbidity and mortality rates of 53% and 4%, respectively [1]. Other studies reported mortality rates of less than 5% [4,7-9,20,21]. Nevertheless, several studies have reported relatively high mortality rates following PD [3,11,12,22]. In the study by Agrawal S et al., from Nepal, the in-hospital mortality rate was 12.9% [22]. The present study included 62 patients who underwent PD for benign and malignant conditions between 2015 and 2019. The authors found that morbidity and mortality rates decreased as the number of PDs performed increased over the study period. Another study from Turkey reported a mortality rate of 9.5% among geriatric patients who underwent PD for malignancy [12]. The authors suggested that this high mortality rate could be attributed to the advanced age and underlying malignancy of the patients. In the current study, the mortality rate was 18.75%. The concurrent presence of multiple independent risk factors within a patient cohort may result in a relatively higher mortality rate [11]. Similarly, poor postoperative outcomes can be attributed to increased patient load, inadequate resources, a lack of a multidisciplinary approach and insufficient critical care management tools, which could lead to poor postoperative outcomes [3]. However, a causality analysis due to the study’s retrospective design could not be conducted. Additionally, a year-wise trend analysis of postoperative morbidity and mortality was not performed due to the small number of mortalities each year.

Therefore, large-scale prospective studies are needed to clarify the factors contributing to the relatively higher mortality rate after PD.

Several studies have investigated the potential relationship between various aspects of PD, such as postoperative complications and mortality and the demographic and clinical characteristics of patients who underwent PD for malignancy [1,6]. In one of these studies, Russell TB et al., reported that obesity, ASA class ≥2 and performing PD using the classic Whipple approach were risk factors for overall morbidity after PD [1]. However, they found no significant relationship between 90-day mortality and various demographic and clinical parameters. Zhao Z et al., found that age and albumin levels were significantly associated with major morbidities after PD [6]. On the other hand, several studies did not find a significant relationship between age and mortality after PD [5,8,20]. Parasyris S et al., reported that morbidity and mortality rates after PD were similar in patients under 80 and those over 80 years old [5]. Similar findings have been observed in other studies [13,14]. In comparison, older age, higher ASA classes and lower albumin levels were significantly associated with 30-day mortality after PD, suggesting that selecting appropriate candidates for PD based on perioperative risk factors is critical.

Some studies have suggested that major complications after PD are essential risk factors for 90-day mortality [4,10,19]. The causes of mortality after PD also vary depending on the frequency of PD performed at the centre [19,22,23]. Accordingly, sepsis, multiple system organ failure secondary to aspiration and bleeding, followed by cardiac arrest and pulmonary embolism, were the leading causes of death in centres where PD was frequently performed. In contrast, septic shock, massive bleeding and thromboembolism were the leading causes of death in centres where PD was performed less frequently. In the current study, postoperative morbidity or complications were significantly associated with the development of mortality following PD. Kapoor D et al.’s study also identified clinically relevant pancreatic fistula and post-pancreatectomy haemorrhage as significant predictors of mortality [8]. Although the type of PD-associated complications, such as anastomotic leakage or pancreatic fistula, was not directly related to mortality in this patient group, anastomotic leakage was more frequently observed among non survivors. However, other complications were not analysed separately, which limited the ability to establish causality. Therefore, large-scale studies are needed to clarify the impact of perioperative clinical factors on the development of mortality following PD.

The negative effect of higher ASA classes on the outcomes of PD has been well documented [1,8,24]. Similarly, it has been reported that patients with morbidity are more likely to experience mortality after PD [1,3]. In contrast, Russell TB et al., found that higher ASA classes were significantly correlated with overall morbidity and major morbidity but not with 90-day mortality [20]. In comparison, a significant relationship was observed between higher ASA classes and 30-day mortality after PD. Therefore, patients with co-morbidities that lead to higher ASA classes should be informed about the potentially adverse outcomes of PD in advance.

The surgical approach used in PD and technical modifications have been investigated in the literature for their impact on PD-specific morbidity and mortality [1]. Some studies have reported that conventional PD is a risk factor for more complications [1,25]. However, there was no significant relationship found between the type of PD and pancreatic anastomosis and mortality.

Although the mechanism is not fully understood, high neutrophil-to-lymphocyte Ratio (NLR) values have been associated with poor short- and long-term outcomes after PD [26]. Russell TB et al., reported a higher risk of major morbidity in patients with higher NLR values [20]. However, they did not find a significant relationship between NLR and the 90-day mortality rate. On the other hand, a significant relationship between lower NLR values and increased 30-day mortality rate after PD was detected, which contradicts the findings of other studies [11,27,28]. Prospective large-scale studies are needed to elucidate the relationship between NLR and mortality in PD patients.

The study’s primary limitations were its retrospective design, small sample size and the absence of intraoperative data regarding the main pancreatic duct diameter and parenchymal texture of the pancreas. Additionally, the lack of a causal analysis of the factors impacting the mortality rate was another limitation of the study.

In conclusion, the present study highlights that older age, a higher ASA class, a lower NLR, hypoalbuminemia, increased intraoperative blood loss, the necessity for blood transfusion and postoperative complications were significant risk factors for 30-day mortality following PD for malignant periampullary tumours. The findings emphasise the importance of preoperative risk assessment and optimisation in improving postoperative outcomes. Tailoring surgical approaches to individual patient risk profiles and ensuring meticulous perioperative care can help reduce mortality rates. Future prospective studies with larger sample sizes must validate these risk factors and develop comprehensive risk stratification models. Ultimately, the goal is to enhance patient selection and perioperative management, thereby improving survival rates and the overall quality of care for patients undergoing PD.