Predictive Value of Body Composition Parameters for Postoperative Complications in Patients Who Received Pancreaticoduodenectomy

Pancreaticoduodenectomy (PD) maintains its important curative therapeutic approach for periampullary tumors with an operative mortality that has decreased to approximately 5%. However, even when performed in high-quality, large volume centers, the operative morbidity still averages 20–40% [1]. This clearly prolongs and increases the difficulty of postoperative recovery but also inevitably delays or even prevents receiving postoperative adjuvant therapy. Postoperative pancreatic fistula (POPF) has long been considered the most hazardous postoperative complication, potentially causing serious infection and hemorrhage from a vascular pseudoaneurysm, with markedly increased costs [2]. Significant efforts have been made over the last decade to identify preventable or remediable factors associated with POPF, although the efficacy has been questioned [3]. Recently, the assessment of body composition, specifically the loss of skeletal muscle mass and accumulation of visceral adipose tissue (VAT), has been linked to both poor short- and long-term outcomes following gastrointestinal surgery [4]. However, additional similar evidence would be valuable regarding complications following PD. Hence, we used the comprehensive complication index (CCI), a novel scoring method that combines all Clavien-Dindo classification (CDC) I through V complications, to quantify the complications more accurately. Therefore, the aim of the present study was to evaluate the impact of body composition parameters measured from preoperative CT scans on postoperative CCI scores and clinically relevant postoperative pancreatic fistula (CR-POPF) following PD.

This study was approved by the Ethics Committee of Drum Tower Hospital affiliated to Nanjing University Medical School (No. 2021-437-01), and individual consent was waived due to its retrospective nature. Between January 2018 and January 2021, 129 consecutive patients with radiology-identified or pathology-diagnosed periampullary diseases were enrolled. The inclusion criteria were patients who (a) underwent PD for periampullary neoplasms; (b) had no other active cancer; (c) were >18 years of age; and (d) underwent enhanced CT within 1 month prior to operation. The exclusion criteria were patients whose: (a) CT images were not obtained within 1 month prior to operation; (b) clinical data were incomplete.

CT images were acquired from the picture archiving and communication system. Different tissues were distinguished by specific Hounsfield units (HU), ranging from −190 HU to −30 HU for subcutaneous adipose tissue, −150 HU to −50 HU for VAT, and −29 HU to +150 HU for skeleton muscle [5]. Two physicians (Yifei Yang, Xu Fu) used 3D Slicer software (v.4.10.2, www.slicer.org) to analyze the images and calculated the areas of subcutaneous adipose tissue, VAT, and total abdominal muscle area at the 3rd lumbar vertebra level (Fig. 1); the average radio-attenuation of skeletal muscle was also assessed. One experienced physician (Liang Mao) reviewed and validated the results. These physicians were blinded to patient information. TAMA was normalized to stature by dividing the muscle area by the patient's height squared which was termed the skeletal muscle index (SMI = TAMA [cm²]/height [m²]). Sex-specific SMI cut-off values of 42.2 cm²/m² for men and 33.9 cm²/m² for women were used to define sarcopenia [6]. Visceral obesity was defined as areas of visceral fat > 100 cm² in both sexes [4]. Myosteatosis was defined according to body mass index (BMI); the muscle attenuation threshold was < 41HU for patients with BMI <25 kg/m² and <33 HU for obese patients with BMI ≥25 kg/m² [7]. Typical images of abnormal body compositions were listed in Figure 2.

Patients who met the following criteria were managed with preoperative biliary drainage: preoperative serum total bilirubin level >256 mmol/L or preoperative cholangitis and poor nutritional status. Endoscopic nasobiliary drainage was utilized in most patients, whereas few patients were managed with percutaneous transhepatic cholangial drainage. PD and pylorus-preserving pancreaticoduodenectomy were performed by five experienced surgeons. The scope of lymphadenectomy was determined according to tumor characteristics. Reconstructive anastomoses included a modified Blumgart pancreaticojejunostomy, continuously sutured end to side cholangiojejunostomy, and gastrojejunostomy; two drainage tubes were then positioned anterior and posterior to the pancreaticojejunal anastomosis. For patients with severe malnutrition, a nasojejunal feeding tube was inserted intraoperatively. Prophylactic antibiotics were administered 30 min before incision and continued for 48 h postoperatively. Development of POPF and/or abdominal infection was documented by amylase measurements and bacterial cultures of the drainage, performed 1, 3, and 5 days following the procedure. Drainage tubes were removed when a postoperative CT scan indicated no evidence of an encapsulated effusion in the abdominal cavity.

Clinical patient data that were collected included (1) age, sex, hypertension, diabetes mellitus, nutrition risk score, BMI, blood test results, preoperative jaundice, preoperative biliary drainage, and tumor characteristics; (2) duration of the operation, blood loss, diameter of the main pancreatic duct, and texture of the pancreas; (3) postoperative length of stay (PLOS), occurrence of POPF, delayed gastric emptying, bile leakage, chyme leakage, postpancreatectomy hemorrhage, surgical site infection (SSI), and other standard complications.

CR-POPF was defined as grade B/C POPF which required clinical interventions according to the International Study Group of Pancreatic Fistula guidelines. All complications were ranked according to their magnitude and a major complication was defined as CDC ≥ IIIa. A free online program (https://www.assessurgery.com) was used to calculate the cumulative CCI for each patient.

Statistical analysis was performed by SPSS 23.0 software (SPSS Inc.). Continuous variables with normal distribution were presented as mean ± SD and compared with independent t test; Mann-Whitney U test was used when they showed skewed distribution and expressed as median (interquartile range). Categorical variables were compared using χ² test or Fisher’s exact test. On the basis of previous studies, 26.2 was set as the cut-off value to distinguish high and low CCI, for one CDC-IIIa complication scored 26.2 [8]. Then, the univariate and multivariate logistic regression models were applied to identify factors associated with CR-POPF and high CCI. For all analyses, p < 0.05 was considered statistically significant.

Demographic and clinical characteristics of the 129 patients are shown in Table 1. The cohort included 78 men (60.5%) and 51 women (39.5%), with an average age of 62.4 years. Sarcopenia, visceral obesity, and myosteatosis were detected in 47 (36.4%), 38 (29.4%), and 50 (38.7%) patients of the study population, respectively. The incidence of major complications was 16.3% among the 129 patients, with 51 (39.5%) of them developing CR-POPF, 44 (34%) delayed gastric emptying, 11 bile leakage, 15 chyme leakage, 5 PPH, and 40 SSI. Overall, 123 patients had CDC grade I complications, 101 grade II, 16 grade IIIa, 5 grade IIIb. Median CCI score of the group was 30.8 (22.6–36.2), and high CCI score accounted for 54.3%. Details of minor and major complications are depicted in Table 2.

The sarcopenic group had a lower albumin concentration (37.7 [34.8–39.4] g/L vs. 38.4 [36.4–40.5] g/L; p = 0.013) and higher CRP level (6.5 [4.2–34.6] g/L vs. 4.7 [2.9–9.8] g/L; p = 0.007) (Table 3). The incidences of CR-POPF (53.2% vs. 31.7%; p = 0.016), major complications (36.2% vs. 4.8%; p = 0.001), PLOS (22 [17–33] d vs. 17 [14–23] d; p = 0.002), and CCI score (34.8 [24.2–39.2] vs. 25.7 [19.4–32]; p = 0.001) were all increased in sarcopenic patients. Significant sex differences were found in the visceral obesity group (33 men vs. 5 women; p = 0.001) and the BMI was higher (26.1 ± 3.5 kg/m² vs. 22.1 ± 2.4 kg/m²; p = 0.001). The operative time of patients with visceral obesity was longer (434.2 ± 109.5 min vs. 375.4 ± 94.4 min, p = 0.003). CR-POPF developed in 57.8% of patients in the visceral obesity group, greater than 31.8% in the non-visceral obesity group. Moreover, visceral obesity was associated with higher rates of bile leakage (18.4% vs. 4.4%; p = 0.015). Although there was no difference in morbidity between patients with or without myosteatosis, it appeared that patients with myosteatosis were older than patients with normal muscle density (68.8 ± 7.9 y vs. 58.3 ± 12.5 y; p = 0.001).

Univariate analysis of factors that predicted CR-POPF and high CCI is shown in Table 4. Patients who developed CR-POPF had statistically significantly higher frequencies of the following characteristics than patients without CR-POPF: ampullary carcinoma (45.1% vs. 26.9%, p = 0.03), main pancreatic duct diameter (p = 0.029), positive postoperative drainage fluid culture (84.3% vs. 61.5%, p = 0.006), sarcopenia (49% vs. 28.2%, p = 0.016), and visceral obesity (43.1% vs. 20.5%, p = 0.006). Age, preoperative biliary drainage (44.3% vs. 25.4%, p = 0.026), positive postoperative drainage fluid culture (84.3% vs. 54.2%, p = 0.001), sarcopenia (47.1% vs. 23.7%, p = 0.006), and visceral obesity (38.6% vs. 18.6%, p = 0.013) had significant difference between patients with and without high CCI score.

A multivariate analysis was performed and significant indicators are listed in Table 5. For CR-POPF, sarcopenia (OR: 2.999; 95% CI: 1.365–6.596) and visceral obesity (OR: 3.569; 95% CI: 1.560–8.163) were found as independent factors. Age (OR: 1.054; 95% CI: 1.02–1.089), preoperative biliary drainage (OR: 2.369; 95% CI: 1.067–5.261), positive postoperative drainage fluid culture (OR: 4.571; 95% CI: 1.973–10.591), sarcopenia (OR: 3.425; 95% CI: 1.541–7.613), and visceral obesity (OR: 3.357; 95% CI: 1.428–7.894) were independent variables that predicted high CCI.

This study demonstrated that sarcopenia and visceral obesity were significant predictors of CR-POPF and high CCI score following PD, testimony that body composition alteration impaired the normal recovery process. Age, preoperative biliary drainage, and a positive postoperative drainage fluid culture contributed to the development of postoperative complications as well. Therefore, we would conclude that radiologic assessment of body composition was helpful in predicting postoperative morbidity following PD.

Sarcopenia has been attributed to progressive loss of skeletal muscle mass and impaired function resulting from poor physical and nutritional status. Sarcopenia has previously been demonstrated to have negative effects on postoperative morbidity and prognosis [9]. The study by Joglekar et al. [10] found that sarcopenia was an independent predictor of major grade III complications and prolonged length of stay. Interestingly, however, they did not find any significant relationship between sarcopenia and POPF. Nevertheless, due to the importance of CR-POPF, and according to international guidelines, POPF was separated into two groups in the current study: biochemical leakage and CR-POPF. Consistent with the study by Nishida, sarcopenia was significantly correlated with CR-POPF in our series [11]. Importantly, the incidence of CR-POPF in our study was somewhat higher than previously reported. A plausible explanation for such a difference could be attributed to VAC in 34.1% of our patients, which promoted the development of CR-POPF (OR: 1.824, 95% CI: 1.087–3.060), consistent with the study by Yang et al. [12]. Notably, most of the CR-POPF patients (47/51) were grade B, whereas only 4 patients experienced grade C POPF. We credited this favorable percentage to performance of routine postoperative CT scans with abdominal puncture when a fluid collection was detected. Additionally, we measured the mean radio-attenuation from preoperative staging CT scans to assess the function of skeletal muscle. Murnane et al. [7] noted that myosteatosis reflected poor physical fitness and capacity to recover from rigorous trauma due to metabolic risk factors including insulin resistance and excessive inflammatory cytokines. Additionally, Margadant et al. [13] found that myosteatotic individuals were more likely to experience serious complications after colectomy. In line with the cut-off values defined in the Murnane study, 50 patients in the present study were classified as myosteatotic, but it had no effect on CR-POPF or high CCI scores, consistent with the findings of the study by Srpcic et al. [14]. The outcomes varied because of different cut-off values of ≤22.0 HU for males and ≤23.5 HU for females in the Margadant study [13].

In addition to sarcopenia, visceral obesity also elevated the incidence of CR-POPF by 3.57 times in the present study. These results were consistent with those of the study by Jang et al. [15] and could be explained by increased cytokines generated from VAT including leptin, TNF-α, IL-1, and IL-6. Excessive release of pro-inflammatory factors might impair the immune function and delay wound healing, particularly the critically important pancreaticojejunal anastomosis. Moreover, excess accumulation of visceral fat in the abdominal cavity added difficulty to exposure in the operative field. Adhesions between adipose tissues tend to obscure the anatomy that requires extra time for dissection of the fat from around the vessels in the course of performing the lymphadenectomy. Furthermore, visceral obesity constituted the principal component of the metabolic syndrome, associated with insulin resistance and chronic inflammation, which could contribute to poor surgical outcomes after major abdominal operations [16]. These same conclusions were reached by David and Pecorelli et al. [5] who concluded that visceral obesity increased the risk of POPF and SSI [17]. Combining sarcopenia and visceral obesity, termed sarcopenia obesity, Sandini et al. [18] found the combination was a strong predictor of major complications following PD for cancer (OR: 3.20; 95% CI: 1.35–7.60; p = 0.008). Because of the limited sample size, we did not pursue further analysis of sarcopenia obesity.

The CCI, a composite score depicting all complications following an operation, was closely related to PLOS, medical expense, and duration in the ICU. Perhaps surprisingly, Hass et al. [19] found that minor complications, as opposed to the most serious ones, were more responsible for the overall medical burden and stress. In the present series, 21 (16.3%) patients developed major complications (CDC ≥ III), whereas the proportion having high CCI accounted for 54.3%. Multivariate analysis (Table 5) identified both sarcopenia and visceral obesity significantly correlated with a high CCI score following PD. The mechanisms are likely attributable to malnutrition and inflammation. Lack of appetite with consequent reduced food intake, widely recognized as a consequence of periampullary cancers, together with obstructed pancreaticobiliary ducts jointly contributed to poor nutrition status, objectively documented in these sarcopenic patients by a low albumin concentration (Table 3). The malignancies and VAT were sources of pro-inflammatory factors which might accelerate the loss of skeletal muscle and lead to poor wound healing [15].

Some measures have been proposed to ameliorate preoperative sarcopenia. Oral nutrition supplementation for patients with pancreatic cancer was found to improve both patient fatigue scores and patient-generated subjective global assessment (PG-SGA) scores [20]. Additionally, physical activity was prescribed to patients at the time of admission because myosteatosis and sarcopenia were reversible forms of skeletal muscle degeneration and dysfunction. Based on these types of data, patients scheduled for PD in our center underwent a 6-min walk test (6MWT) with a distance of 400 m or more deemed acceptable. Throughout the entire perioperative period, additional daily activities and balloon blowing were advised for the weak and elderly.

Increased age, the need for preoperative biliary drainage, and positive postoperative drainage fluid cultures were all independent risk factors for high CCI score by multivariate analysis. As anticipated, the elderly were more susceptible to fatigue, and the mean age was older in the sarcopenia and myosteatosis groups. Furthermore, Fu et al. [21] observed that preoperative biliary drainage raised the incidence of SSI because the duodenum and biliary tree were directly connected and thereby contaminated with an indwelling tube. In our study, 54.3% of the patients were jaundiced and 65% of them received preoperative biliary drainage, undoubtedly contributing to the high postoperative SSI rate. Significantly, K. pneumoniae and E. faecalis were frequently found in postoperative ascites and bile in the Fu study, and we further confirmed these findings in our earlier research that positive bacteria cultures, particularly K. pneumoniae, were closely related to CR-POPF and serious complications [22].

This study presented some limitations. First, it was conducted in a single center and the sample size was limited; additional research will be necessary to validate our findings. Second, because the study was retrospective, the term “sarcopenia” described anatomic muscle mass. Results of a grip strength test or a 6-min walk test would add a functional component for enhanced accuracy. Finally, a postoperative time limit of 30 days admittedly excludes a certain percentage of patients that develop only late complications. Ultimately, the CCI was calculated based on all complications that occurred during the hospital stay.

In conclusion, the current investigation found that visceral obesity and preoperative sarcopenia were independent predictors of CR-POPF and high CCI scores. Additionally, age, preoperative biliary drainage, and a positive postoperative drainage fluid culture were strongly associated with postoperative mild and severe complications. Therefore, preoperative sarcopenia and visceral obesity in patients scheduled for PD appear to be reliable indicators of high risk for postoperative complications. If verified with subsequent validation studies, safe and expeditious means of ameliorating these conditions would seem potentially valuable.

The authors thank members of the multidisciplinary pancreatic cancer team of the Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School for their guidance. The authors also would like to express their gratitude to EditSprings (https://www.editsprings.cn) for the expert linguistic services provided.

This study protocol was reviewed and approved by the Health Research Ethics Board of Drum Tower Hospital of Nanjing University Medical School (No. 2021-271-01). Informed consents were waived due to its retrospective nature.

All authors declare that they have no competing interests.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Yudong Qiu designed this research. Liang Mao gave concrete guidance and checked the results. Yifei Yang and Xu Fu collected the clinical data and analyzed the CT images. Zhenghua Cai conducted statistical analysis and drafted the manuscript. All authors contributed to the article and approved the submitted version.

All data generated or analyzed during this study are included in the article. Further inquiries can be directed to the corresponding author.

Zhenghua Cai and Yifei Yang contributed equally to this work.