Houyem Mansouri1 , Ines Zemni2,3*, Mohamed Ali Ayadi2,3, Ines Ben Safta2,3, Tarek Ben Dhiab2 , Najet Mahjoub4 , Leyla Achouri1 , Khaled Rahal2
Mansouri H, Zemni I, Ali Ayadi M, Ben
Safta I, Ben Dhiab T, Mahjoub N, Achouri L,
Rahal K. Determinants of surgical morbidity
in gastric cancer: experience of a single
center and literature review. J Gastric Surg
2020; 2(4).
Received: December 5, 2020
Accepted: December 12, 2020
Article in press: December 13, 2020
Published online: December 15, 2020
Dr. Ines Zemni
Department of surgical oncology
Salah Azaiez Institute
University Tunis El Manar
Faculty of Medicine of Tunis
Avenue 9 April Bab Saadoun
1006 Tunis, Tunisia
ines.zemni@yahoo.fr
Telephone: +216 25 560736
gastric cancer, surgery, Morbidity, postoperative complications, Prognosis.
Management of gastric cancer has undergone in the last decade remarkable progress both in the neo-adjuvant and adjuvant treatments and in the surgical procedure as regards the extent of resection and the lymph node dissection. These advances are certainly associated with an improvement in the overall prognosis of the neoplastic disease. But this gain has been associated with over- morbidity and mortality that we need to know to prevent it. The rate of postoperative morbidity reported by the largest randomized controlled series varies from 11 to 46%.[1] This significant variability in the literature is explained not only by the differences in the characteristics of the studied populations, the therapeutic procedures but also by the definition of the morbidity used in each study. Thus, to limit the subjectivity in the analysis of the complications, many authors used the Clavien and Dindo classification of postoperative complications published in 2004. This classification seems more straightforward, objective, but above all reproducible, since it evaluates the severity of the complications according to their therapeutic management.[2] This study aimed to assess the severity of intraoperative and postoperative complications of gastric surgery and to investigate the predictive factors correlated to surgical morbidity.
Approval to conduct this study was obtained from the Institutional Review Board of the Salah Azaiz Institute. We conducted a retrospective study including 145 patients with histologically proven gastric adenocarcinoma who underwent curative or palliative gastrectomy from January 2005 to December 2015. Noninclusion criteria were as follows:(1) metastatic disease, (2) other synchronous tumoral location, (3) unknown surgery status, (4) unknown vital status; (5) incomplete pathological data.
The surgical variables included the type of procedure (total versus partial gastrectomy), the extent of lymph node dissection (D0, D1.5, and D2) based on the Japanese Gastric Cancer Treatment Guidelines 2010 (version 3)[3] and the number of retrieved lymph nodes (NRLN), multivisceral organ resection (MVR), operative time measured from the first skin incision to the closure of all skin incisions and confirmed by the attending anesthesiologist in the operating room, intraoperative allogeneic blood transfusion and the length of hospital stay.
The extent of the lymphadenectomy was based on the individual surgeon’s judgment. Splenectomy was performed in cases of metastatic lymph nodes at the hilum of the spleen or because of iatrogenic injury. The extent of stomach resection was related to the primary tumor site: total gastrectomy was performed in all proximal tumor locations and total gastric tumors, and subtotal gastrectomy was performed for distal tumor locations, provided that a 5 to 6-cm safety margin was present. A multi-visceral resection was performed in all cases of T4 or suspected T4 tumors. After total gastrectomy, reconstruction was made by Roux-en-Y esophagojejunostomy. Esophago-jejunal anastomoses were performed by hand-suture or mechanically with a circular stapler according to the surgeon’s choice. The stapled esophageal and jejunal doughnuts were examined for completeness of anastomosis. A hand-sutured or stapler Billroth II followed subtotal gastrectomy. A prophylactic antibiotic of a second- or third-generation cephalosporin was administered to all patients and usually lasted for 5 days following the operation.
The integrity of the anastomosis was routinely evaluated on the seventh postoperative day by a water-soluble contrast swallow and/or by ingestion of methylene blue before reintroducing an oral intake. The anastomotic fistula was diagnosed with extravasation of contrast product during radiological examination and/or with discharge or gastrointestinal content through a drain. Clinical leakage was defined as the presence of either clinical symptoms suggesting potential leakages such as abdominal pain abnormal with drain discharge, fever, and leukocytosis.
Postoperative morbidity and mortality included all adverse events reported within the first 30 days after surgery or during the same hospitalization, and late complications were not included in this study. Surgical complications included anastomotic fistula and leakage, early anastomotic stenosis, septic collection, and wound infection, pleural effusion, ileus, and bleeding. Non-surgical complications were heart failure, pneumonia or respiratory failure, urinary tract infection, decompensation of diabetes, and thromboembolic events. All complication data were graded retrospectively according to the Clavien-Dindo classification from Grade 1 to Grade 5 according to the treatment for each complication. Major complications included all complications classified as grade III A or more according to the Clavien Dindo system.
The categorical variables were presented as numbers and percentages. Continuous variables were expressed as the mean ± standard Deviation (SD) with ranges, and groups were compared using the Test t of Student. Moreover, some continuous variables were converted to dichotomous variables for convenience, including the age (≤65 years versus > 65 years), the body mass index BMI (<20 kg/m2 versus ≥ 20kg/m2), hemoglobin rate (≤10 g/dl versus >10 g/dl), the tumor size (<50mm versus≥50 mm) and the NRLN (≤15, 16-25 and≥25). Clinical and pathological data were compared between patients to analyze the major risk factors associated with complications, length of hospital stay, operative time and intraoperative blood transfusion. The groups were compared using the chi-square test or Fisher’s exact test. All variables with a p-value <0.2 in the univariate analysis were entered into a multivariate analysis using a logistic regression model to identify independent factors of perioperative blood transfusion and operative time. We used the Statistical Package for Social Science (SPSS), version 20.0 for Windows, and ap-value less than 0.05 was considered significant.
The clinicopathological characteristics of the patients are summarized in
The operative data are summarized in
The Mean operative time was 194 minutes ± 53.22 with extremes ranging from 110 to 350mn. Univariate analysis of the factors influencing the operating time is shown in
Intraoperative blood transfusion was performed in 77 patients (53.1%), with a mean amount of 2.14 units and extremes ranging from 1 to 4 units. Of the transfused patients, 19 patients (24.7%) required more than three units. Univariate analysis of factors associated with intraoperative blood transfusion is shown in table4. Transfused patients had a longer mean operative time compared to non-transfused patients (23.04mn vs 183.83mn, p=0.054). Significant factors on the univariate analysis were included in the multivariate analysis (
A total of 113 out of 145 (77.9%) patients had no complications, and 32 (22.1%) had at least one medical or surgical complications. Details of postoperative complications are listed in
Surgical complications occurred in 18 patients within a delay of 10 days (range, 4 to 20 days). The anastomotic fistula was the most common surgical complication occurring in 10 cases (6.9%), and postoperative peritonitis secondary to complete anastomotic leakage was diagnosed in only one patient. From the 145 patients, four (2.8%) required re-laparotomies for serious complications: one case of stenosis of the esophago jejunal anastomosis, one case of anastomotic leakage, one case of the esophago-jejunal fistula with major ionic disorders and severe dehydration, and one case of a deep abdominal abscess that was not accessible to percutaneous radiologic drainage.
Using the Clavien-Dindo classification, the 32 complicated patients had the following grades: 21 patients (14.5%) had grade II, 3 patients (2.1%) had grade IIIa, 2 patients (1.4%) had grade IIIB and one (1.6%) had grade IVa. Postoperative death (grade V of Clavien-Dindo classification) occurred in five patients (3.4%) within a mean delay of 30 days after surgery (range 18 to 74 days). Three patients died because of anastomotic fistula, one patient died after reoperation for anastomotic stenosis and the last patient died from respiratory failure secondary to pneumonia.
Univariate analysis of clinical and therapeutic predictive factors of postoperative complications is shown in
p = 0.066). Patients with BMI greater than or equal to 20 had a higher rate of complications than those with a BMI of less than 20 (23.3% vs. 20.8%, p = 0.842). However, we found that age, gender, locally advanced tumors, extended lymphadenectomy, splenectomy or SPC, total gastrectomy, and combined organ resection were not significantly associated with higher postoperative morbidity.
The mean length of postoperative hospital stay was 15.08 days ± 7.48 (range 7-74 days). Univariate analysis of factors associated with the length of hospital stay is represented in
The results of the univariate analysis of patient-related risk factors of fistula and leakage showed that these complications were more frequent in diabetic patients (23.5% vs 4.7%; p=0.004). According to laboratory variables, preoperative hypoproteinemia (11.5% in case of hypoproteinemia vs 4.3% with normal proteinemia, p=0.168) and anemia with hemoglobin rate less than 10 g/dl (12.5 % vs 4.8%, p=0.139) were likely to be associated with a higher rate of fistula even if the difference was not
statistically significant. In the analysis of tumor-related factors, anastomotic fistula was significantly associated with tumor size exceeding 50mm (10.3% vs 1.7%, p=0.045), gastric walls rigidity (33.8% vs 5.1%; p=0.016), and the length of the proximal margin of resection (the mean length of the proximal margin in patients with fistula was 41± 24.358 mm vs 57.55± 26.008, p=0.042). When analyzing surgical related factors, we found that the type of gastrectomy, the extent of lymphadenectomy, and combined organ resection were not significantly associated with a more frequent fistula. Nevertheless, the number of retrieved lymph nodes (23.24 ± 11.036 vs 27.5 ± 12.059 in case of anastomotic fistula, p=0.073), as well as the duration of the operation (190.77 ± 12.59 min vs 238±16.96 min in case of fistula, p=0.006), were identified as predictive factors of fistula.
In our study, the rate of postoperative complications was 22.1% which seems consistent with the data in the literature.[1] However, we found a lower rate of major complications according to Clavien and Dindo classification (7.6%), and this can be explained not only by the small number of the patients but also by a retrospective classification of complications based on data collected from medical records. The most common complications reported in the literature are respiratory complications occurring in 1.1% to 12.32% of cases[4], decompensation of chronic disease, and gastroparesis.
Anastomotic leakage occurring in 1 to 11.5%, intraabdominal abscesses, pancreatic fistulas, intraperitoneal hemorrhages, postoperative occlusions, postoperative pancreatitis, and eviscerations were the most common surgical complications.[5, 6] However, Marrelli reported a higher rate of intra-abdominal infection of 14.2% with more extensive lymph node surgery.[7] In this study, we reported a rate of 9% of pneumonia, 6.9% of anastomotic fistula, and 4.2% of suppurative complications. According to Baiocchi, the postoperative mortality rate was also very variable. Western centers reported a mortality rate of 5%, while Eastern centers reported a lower rate of 2%.[1] In our series, postoperative deaths occurred in 3.4% of cases within an average of 30 days after the intervention, which seemed consistent with the literature data.
The extent of surgical resection, particularly the type of gastrectomy and lymphadenectomy, represent the most crucial factor of postoperative morbidity in gastric cancer.
Many studies evaluated the impact of extended lymphadenectomy on perioperative morbidity and mortality and the results were variable between the Asian series, particularly Japanese, and the Western series (
Several studies investigated the impact of the type of gastrectomy on postoperative morbidity and mortality with variable results. In fact, since the extent of the gastrectomy depended on the site of the tumor and its size, most of the studies had essentially compared total gastrectomy to partial gastrectomy in distal tumors. These studies were included in a recent metaanalysis published in 2016 by Qi et al.[23], combining data from 5447 patients included in 10 retrospective studies and one prospective randomized study. In this meta-analysis, TG was associated with a higher risk of postoperative complications (RR = 1.76, 95% CI = 1.31- 2 .36, p = 0.0002) and especially more frequent intraabdominal abscesses (RR = 3.41, 95 % CI = 1.21-9.63, p <0.05) compared to partial gastrectomy with a similar rate of postoperative mortality (RR = 1.48, 95% CI = 0.90-2.44, p = 0.12). Moreover, the impact of the type gastrectomy on morbidity and mortality, regardless of the tumor location, has been reported in several other studies suggesting that TG is associated with higher rates of postoperative complications and morbidity, such as the Dutch trial (OR = 2.04, 95% CI = 1.01-3.79, p = 0.02). In the analysis of surgical morbidity and mortality of the Randomized “Critics” trial where 636 patients were included, total gastrectomy and oesophago-gastrectomy were independent risk factors for both surgical and medical postoperative complications (OR = 1.88, 95% CI = 1.30-2.72, p = 0.001) with higher rates of surgical revision and longer hospital stay.[24] Nakagawa et al. performed a retrospective analysis of the risk factors of postoperative complications in 539 patients who were previously prospectively collected.[25]
In this study, total gastrectomy was the only independent factor of high-grade complications according to the Clavien and Dindo classification (OR = 2.075, 95% CI = 0.26-0.896, p = 0.021). However, several other studies had reported comparable morbidity rates between the
two surgical procedures. Indeed, the retrospective study of Park et al.[26] involving 719 patients did not report a significant difference between GT and GP regarding postoperative morbidity (19.3% versus 13.6% respectively, p = 0.103). However, Persiani found that the length of hospital stay exceeding ten days was particularly observed with total gastrectomy.[27] These results support the data from our series, where postoperative morbidity was not significantly correlated with the type of gastrectomy with a comparable rate of fistula and length of hospital stay even though the incidence of complications appeared to be higher in case of GT (27.3%) compared to GP (16.2%). However, TG was associated with higher perioperative morbidity with and increased rate of blood transfusion (63.6% vs 41.2%, p= 0.007) which is consistent with the results of the meta-analysis of Sun et al.[28] and also a longer operative time (210mn vs 189.5mn, p= 0.015) which is consistent with the results of Gockel et al. and Papenfus et al.[29, 30]
The benefit of multivisceral resection (MVR) for locally advanced gastric ADK is controversial because of the increased mortality and morbidity. In the systematic review carried out by Brar[31] including 17 studies with 1343 patients, the morbidity rate varied between 11.8% and 90% and perioperative mortality ranged from 0 to 15%. In contrast, in our study, MVR did not significantly increase the rate of postoperative complications (23.5% vs 21.6%, p = 0.814) and fistula (11.8% vs. 5.4%, p = 0.245) but was correlated to higher intraoperative morbidity with a significant lengthening of the operative time and an increase of blood transfusion requirement. Our findings were similar to that of the large multicenter and observational study published by Pacelli in 2013, who found that MVR was not associated with an increase in mortality (p = 0.55) or morbidity (33.9% vs 31.6%, p = 0.38).[32]
The impact of intraoperative blood transfusions on short-term and long-term outcomes has been widely debated. In our study, blood transfusion had led to a significant increase in the rate of complication and fistula. Indeed, many authors had identified blood transfusions as an independent risk factor of major complications[30, 33] and supported the hypothesis that allogeneic blood transfusions induced immunosuppression that might increase the risk of infectious complications and leading to an extension of hospital stay.[34] Otherwise, in this study, transfused patients had a longer mean operation time than non-transfused patients (203.4 minutes vs183.82mn respectively, p=0.054), and the same results were also reported by Xiao et al.[34] and Ojima et al.[35] We also found that postoperative complication was correlated to the duration of surgical procedure and we stated that patients who developed an anastomotic fistula had a significantly longer operative time (238 min vs 190.77 min, p = 0.006) which is consistent with the results of the study of Migita et al.[5] who reported a significant correlation between anastomotic leakage and the duration of the surgical procedure (330 min vs 290 min, p = 0.0416) and the finding of Nakagawa et al. who reported that operating time exceeding 240mn was a risk factor for high-grade complications.[25]
Several intrinsic factors of postoperative morbidity of gastric cancer have been reported in the literature such as age, sex, nutritional status, and co-morbidities as well as tumor characteristics. The impact of gender and hormonal status on postoperative morbidity and mortality remains controversial. Although our results as well as those of Persiani et al.[27], Nakagawa et al.[25] and Lee et al.[36] did not support the correlation between gender and postoperative morbidity and mortality, the results of Critics trial[24] as well the 15-year results of the Dutch trial[37] have shown that men were exposed to a higher risk of post-operative complications and death. On the other hand, Sah et al. stated that women aged between 46 to 56 years were significantly predisposed to a higher risk of postoperative complications suggesting that hormonal instability related to menopausal status could result in a change of the host’s response to stress and surgical trauma.[38]
Malnutrition can lead to the abnormal function of macrophages, neutrophils, and lymphocytes, which can inhibit the immune response and increase the surgical morbidity[39] and the incidence of major postoperative complications according to Clavien Dindo classification.[36] In the recent study by Zheng et al. including 1976 patients, the group of malnourished patients with hypoproteinemia (412 patients) had a significantly higher complication rate (21.4% vs 15.5%, p = 0.005).[40] Although we only recorded the level of proteinemia, we found that patients with hypoproteinemia had a higher rate of complications (28.8% vs 18.3%, p=0.066) and fistula (11.5% vs 4.3%, p = 0.168). Obesity (BMI> 25) is frequently associated with other co-morbidities such as diabetes, cardiovascular and respiratory diseases and according to The Japanese study by Ojima T, overweight patients had a longer duration of intervention and a higher rate of postoperative complications (anastomotic release, pancreatic fistula).
[41] However, in our series, postoperative complications were not significantly increased in patients with a BMI greater than 20 kg/m2. These findings are comparable to those of several other studies which had not objectified a correlation between the weight of the patients and the specific surgical complications, and which had identified hypoalbuminemia as the only independent factor of morbidity.[25, 42]
Although in the studies by Persiani et al., Gil-Rendo et al., and Nakagawa et al., morbidity and mortality were correlated neither to age nor to co-morbidities and ASA score, but to the extensive surgery[25, 27, 43], other studies including that of Papenfuss reported that the implication of age as a factor of morbidity and mortality would most probably be related to the increased incidence of comorbidities with higher ASA score after 60 years and the alteration of the immune mechanisms.[30] In our series, the age of patients was not identified as a risk factor for morbidity even though we found that the average age of patients who died postoperatively was 75 years (range, 65-84 years). However, we found that postoperative complications were more frequent in the case of comorbidities with a higher rate of anastomotic fistula in diabetic patients and a longer duration of hospitalization. The predictive morbidity value of the ASA score in our study was similar to that of Lee et al. who reported a linear increase of surgical morbidity with ASA score without significant difference (19.5% for ASA1, 24% for the ASA2, and 31.4% if the ASA score was greater than or equal to 3; p= 0.088).[36]
Some authors had identified clinical and histological features of gastric tumors as intrinsic risk factors for postoperative morbidity. It has been reported that morbidity and mortality were higher in the upper and middle third gastric tumors.[43] On the contrary, in our study, there was no difference regarding postoperative morbidity according to the tumor location (p=0.424). However, intraoperative morbidity was found to be higher in the proximal, middle, and total gastric tumors compared to distal tumors and the tumor site represents an independent risk factor of intraoperative blood transfusion. Also, the operative time was significantly shorter in distal tumors compared to other locations. These findings were like those of Yu et al. and Liu et al. who suggested that the incrimination of the proximal tumor site in postoperative morbidity would be more related to larger resections and more difficult anastomosis with more blood transfusion and longer operative time and not to the tumor location itself.[44, 45]
The correlation between the risk of postoperative complications and the tumor size can be explained by the fact that large tumors are associated with increased surgical technical difficulties, an extension of the type of gastrectomy, and the need for multi-visceral resections. [33, 42] Contrariwise, in our study, we found no significant difference in terms of post complications according to the tumor size. However, when analyzing the specific surgical complications, we found that tumors larger than 50 mm
were significantly associated to a higher rate of fistula and an increase of blood transfusion requirement which is supported by the finding of Wang et al. who divided 513 patients into four groups according to the tumor size (≤2, ≤3, ≤5,> 5 cm) and stated that the rate of postoperative complications was comparable between the four groups (p = 0.682) with a significant increase in blood transfusion with larger tumor (30% for sizes ≤2, 33.7% for sizes ≤3, 46.3% for sizes ≤5 and 63.9% for sizes> 5 cm, p <0.001).[46] The majority of authors as well as our finding had not demonstrated a strong correlation between postoperative morbidity and tumor stage and the depth of parietal invasion even if locally advanced tumors would be particularly associated with larger resections and more frequent surgical difficulties.[23, 24, 26, 27] However, we stated that surgery of T3-T4 tumors was associated with an increase of the intraoperative morbidity which was assessed by a longer operative time and more frequent blood transfusion supporting the results of Ojima et al. and Zhou et al.[35, 47]
Our study has some limitations not only its retrospective nature and the small number of included patients but also a lack of evaluation of the predictive factors of surgical mortality because of the low number of postoperative deaths.
Our study was able to analyze the intraoperative morbidity by determining the factors influencing the operative time and the intraoperative transfusions and suggested that it depends not only on the extent of the surgery and particularly the multi-visceral resections and the extension of the lymphadenectomy but also on the tumoral characteristics. In addition, the use of the Clavien and Dindo classification allowed an objective assessment of postoperative morbidity that depends on the patient’s terrain and the association of comorbidities and tumor characteristics. We also identified the predictive factors of an anastomotic fistula which represented the most severe complication. Moreover, the occurrence of complications and their grade was the determinant of the length of postoperative hospital stay. Further studies focusing on late complications, functional results, and quality of life are needed to improve the surgical outcomes.
All authors thank Professor Ibtissem Abderrazek for the revision of the English language.
HM, IZ, MAA, IBS, TBD, NM, LA, KR conceptualized and designed the study, acquired, and analyzed data, interpreted the study results, drafted the manuscript, and critically revised the final version of the manuscript.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-forprofit sectors.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
Further information is available from the corresponding author on reasonable request.
The ethics review board approved this study and did not require informed consent from study participants since this was a strictly registry-based study.
Not commissioned; externally peer reviewed.
This is an Open Access article distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work noncommercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/ licenses/by-nc/4.0/