Morbidity analysis in minimally invasive esophagectomy for oesophageal cancer versus conventional over the last 10 years, a single institution experience.

BACKGROUND: There has been an increasing inclination towards minimally invasive esophagectomies (MIEs) at our institute recently for resectable oesophageal cancer.
OBJECTIVES: The purpose of the present study is to report peri-operative and long-term procedure specific outcomes of the two groups and analyse their changing pattern at our institute.
METHODS: All adult patients with a diagnosis of oesophageal cancer managed at our institute from 2005 to 2015 were included in this retrospective study. Patients' demographic and clinical characteristics were recorded through our hospital information system. The cohort of esophagectomies was allocated into two groups, conventional open esophagectomy (OE) or total laparoscopic MIE; hybrid esophagectomies were taken as a separate group. The short-term outcome measures are an operative time in minutes, length of hospital and Intensive Care Unit (ICU) stay in days, post-operative complications and 30 days in-hospital mortality. Complications are graded according to the Clavien-Dindo classification system. Long-term outcomes are long-term procedure related complications over a minimum follow-up of 1 year. Trends were analysed by visually inspecting the graphic plots for mean number of events in each group each year.
RESULTS: Our results showed no difference in mortality, length of hospital and ICU stays and incidence of major complications between three groups on uni- and multi-variate analysis (P > 0.05). The operative time was significantly longer in MIE group (odds ratio [OR]: 1.66, confidence interval [CI]: 2.4-11.5). The incidence of long-term complication was low for MIE (OR: 1.0, CI: 133-1.017). However, all post-operative surgical outcomes trended to improve in both groups over the course of this study and stayed better for MIE group except for the operative time.
CONCLUSION: MIE has overall comparable surgical outcomes to its conventional counterpart. Furthermore, the peri-operative outcomes tend to improve in our centre with the maturation of program and experience.

INTRODUCTION

Esophagectomy the main treatment option for resectable oesophageal cancer is a complex operation with significant morbidity and mortality.[123] Minimally invasive esophagectomy (MIE) was described in 1990s in an endeavour to reduce operative morbidity.[4] Luketich et al. in 1998 demonstrated the potential feasibility of the procedure by publishing their results on 8 MIEs using either laparoscopic and/or thoracoscopic techniques with no perioperative mortalities and one anastomotic leak.[15] MIE since then is becoming the routine procedure for resectable oesophageal cancer with apparently similar peri-operative short and long-term outcomes.[26] In the absence of strong evidence confirming to the efficacy of the technique with a single published randomised controlled trial and another in process,[78] most of the evidence from literature comes from various large retrospective case series and their meta-analysis.[9101112] Further problems such as those inherent to the learning phase of MIE have been less frequently addressed.[3131415]

Our institute being the largest cancer centre of the country has also seen a shift from conventional to minimally invasive techniques for resectable oesophageal cancer over the past decade.[16] Moreover, during the same period our surgical oncology program has evolved into a high volume centre for the management of this disease. With the start of our first minimally invasive hybrid esophagectomy in 2011, it has grown into a standard treatment modality over recent few years, limiting the conventional open approach mainly for emergency esophagectomies.

The purpose of the present study is to analyse the peri-operative and long-term procedure related outcomes of the two groups at our institute and to report the results. In addition, the idea is to assess the changes in the occurrence of these outcomes in a time-dependant fashion.

METHODS

All adult patients with a histo-pathological diagnosis of oesophageal cancer managed at our institute after a multidisciplinary meeting decision for surgical resection from September 2005 to September 2015 were included in the study.

An exemption status was granted by the hospital Ethical Review Committee for conduction of this study. All patients demographic and baseline clinical and pathological characteristics are recorded Table 1. Furthermore, radiologic, endoscopic, operative and post-operative details were documented. The co-morbid conditions for each patient were graded according to Charlson et al. co-morbidity scoring system.[1718] The cases with more complex surgery including en bloc resection of adjacent organs (splenectomy/gastrectomy) and tumours of upper one-third of the oesophagus with concomitant laryngectomy or pharngectomy were excluded from the study. Similarly, pathological subtypes other than adeno and squamous cell cancer of the oesophagus were excluded.

Table 1

Basic demographic, clinical and histo-pathological variables stratified by the type of surgery

The cohort of esophagectomies was allocated into two groups depending on the type of surgery conventional open esophagectomy (OE) or MIEs. To maintain a clear segregation of MIE and conventional groups, hybrid procedures with only one of the abdominal or thoracic approaches being done through laparoscope or thoracoscope are taken as a separate group.

The short-term outcome measures are operative time in minutes, length of hospital stay in days, length of post-operative Intensive Care Unit (ICU) stay in days, post-operative complications and 30 days in-hospital mortality. Complications are graded according to the Clavien-Dindo classification system.[19] Long-term outcomes are long-term procedure related complications recorded over a minimum follow up of 1 year.

All variables were obtained through our hospital information system by three main investigators and were reviewed independently by the main investigator (Khan M). The frequency and nature of post-operative complications were determined based on daily physician progress notes supplemented by relevant investigation reports. The hospital length of stay was determined by the surgery and discharge date and operative times were determined based on anaesthesia records. The long-term complication rate was calculated for patients with a minimum follow-up of 1 year. Furthermore, the changes in above mentioned morbidity parameters from 2005 to 2015, in minimally invasive group as compared to conventional were analysed. Results in the two arms MIE versus OE are further evaluated for a subset analysis according to the type of surgery performed (three-stage, two-stage and trans hiatal) for assessment.

For the purpose of division, Conventional techniques of esophagectomy included procedures without any laparoscopic or thoracoscopic component including transhiatal and three stage McKeown esophagectomy via right thoracotomy approach. While MIE techniques included total laparoscopic transhiatal or video-assisted thoracoscopic three-stage procedures with a laparoscopic abdominal part. Stomach was used for reconstruction in all of our cases with a hand sewn gastro-oesophageal anastomosis except one. All cases were jointly performed by one of the two surgical oncologists and one thoracic surgeon over the period of this study. The video-assisted thoracoscopic or thoracotomy portion of the operation is performed in the left lateral decubitus position for optimal thoracic lymphadenectomy, and the abdominal part of both open and laparoscopic portion is done in a modified Lloyd-Davis position for better assistant and instrument positioning. All patients are followed up in surgical oncology clinic at 1 week post-surgery followed by 1 month, 3 months, and then every 6 months.

Statistical analysis

This is a retrospective comparative cohort study. All analysis was performed with IBM SPSS Statistical Software version 19.0. Armonk, NY, USA. We looked at frequencies and proportions. Chi-square test for categorical and ANOVA test for continuous variables were utilized. Continuous variables were dichotomized according to the clinical importance or median value of each variable. Uni- and multi-variate logistic regression analysis was performed for the main outcome measures controlling all other variables in the study. Differences that achieved a two-tailed P < 0.05 were considered statistically significant for the present study. Trends were analysed by visually inspecting the graphic plots for mean number of events in each group each year.

RESULTS

Out of 247 consecutive esophagectomies performed at our institute from September 2005 to September 2015, 216 patients with diagnosed squamous or adeno-carcinoma of oesophagus and Type I, II gastro-oesophageal junction managed with a standard esophagectomy without a laryngectomy, pharyngectomy were included in the study. Among these, 90 were conventional open, 95 were minimally invasive and 31 hybrid esophagectomies. 7 minimally invasive surgeries converted to open transhiatal and 4 MIE into open three-stage procedures were treated as conventional open esophagectomies due to retrospective nature of the study.

The patient and tumour characteristics were similar among groups on the basis of gender distribution, body mass index, co-morbidity index, initial tumour stage according to the seventh edition of the American Joint Committee on Cancer guidelines and tumour grade. However, due to the inclusion of emergency oesophageal resections done mainly via open approach (n = 7; 6 endoscopy related and 1radiation necrosis), proportion of patients without neoadjuvant treatment and less pathological treatment response was high in OE group. The OE group also had a higher number of transhiatal procedures done for GE junction adeno-carcinomas reaching statistical significance.

The outcome variables distribution among various types of esophageal resections is shown in Table 2. Median length of follow-up for all patients was 12 months (range 0–90). Overall 30-day mortality rate, rate of major complications requiring re-intervention or re-exploration and length of hospital and post-operative ICU stay were statistically insignificant between the groups. The results for these outcomes stayed insignificant on multivariate analysis performed by controlling all other variables including age, tumour location, tumour grade and morphology, radiological and pathological stage, neo-adjuvant or adjuvant treatment and type of esophagectomy, that is, transhiatal, two-stage or three-stage [Table 3]. In our series, rate of long term complications remained low for MIE and the mean operative time longer for MIE and hybrid groups with statistically significant difference on multivariate analysis.

Table 2

Outcome distribution among various groups

Table 3

Results of uni- and multi-variate analysis for risk of occurrence of outcome measures in minimally invasive and hybrid group as compared to conventional

The Class I and II minor complication group included minor respiratory and wound complications, fluid electrolyte disturbances or any other complications including atrial fibrillations, diarrhoea, urinary tract infection, surgical emphysema, recurrent laryngeal nerve injury etc., managed either conservatively or with medications. The major complications group included major Class III and IV respiratory complications requiring additional radiological or surgical intervention, in addition to re-explorations for thoracic duct or tracheo-bronchial tree injury, conduit failure, haemorrhage or anastomotic leak.

Long-term complications were 22.2%, 11.6% of them were anastomotic strictures requiring multiple dilations and stent placement in 25 cases. Rest were 8.7% long term reflux and aspiration related symptoms and 1.9% incisional hernias including one diaphragmatic hernia.

However, when the results were analysed in a time dependent fashion we found out that the overall rate of all peri-operative morbidity parameters utilised in our study stayed better for minimally invasive group and showed a uniform improvement for both the groups. It was only the mean operative time which stayed consistently longer for the MIE group [Figure 1].

Figure 1

Changes observed in outcome parameters over time among the comparison groups. OE: Open esophagectomy, MIE: Minimally invasive esophagectomy

DISCUSSION

Minimally invasive technique is being utilized for resectable oesophageal cancers since 2011 at our centre and has resulted in a constant decrease in the number of conventional esophagectomies with only two conventional transhiatal esophagectomy procedures done during year 2015 for emergency oesophageal perforation.

Recent systematic reviews and meta-analysis comparing MIE to conventional esophagectomies including Watanabe et al., Dantoc et al., and Kim et al.[1620212223] have shown no statistically significant difference between groups in terms of 30-day mortality, similar to our data with an overall mortality lower in MIE group (2.1% vs. 8.9% conventional and 3.2% hybrid) without any statistically significant difference on multi-variate analysis.

The major complication rate (22%) and length of ICU stay (mean 1 day) was lower for minimally invasive group in our series, but values did not reach a statistical significance. The median length of hospital stay (median 9 days) is comparable or less than most of the series but without significant difference between groups.[9121622] Larger series by Schoppmann et al. (MIE 12 vs. OE 24) and Kauppi et al. (MIE 13 vs. OE 14,) have reported their results with a length of hospital stay significantly shorter for MIE group with no significant difference for operative time, whereas the median ICU stay was better for MIE in series by Schoppmann et al. and stayed similar for groups in Kauppi et al.[912]

The operative time in our series was significantly longer for minimally invasive group (MIE median 330 min vs. 240 OE) similar to what was shown in the meta-analysis by Watanabe et al. and a recently published Society of Thoracic Surgeons National Database analysis of USA (443 vs. 312 min; P < 0.001).[1202224] Their results also showed a shorter median hospital stay (9 vs. 10 days) and higher re-operation rate (9.9% vs. 4.4%) in MIE group than conventional open. The complication rates were described quite heterogeneously by all of the above series with peri-operative morbidity advantage of the two groups staying still arguable.[22]

The frequency of individual procedure specific complications in our series in different groups is shown in Table 4. The overall rate of minor pulmonary complications managed with chest physiotherapy and incentive spirometry and Class III complications requiring bronchoscopy, radiological drainage or chest tube insertion for a pleural effusion, pneumothorax or lung collapse was higher for MIE and hybrid techniques in our series accounting for a higher number of thoracic procedures in these groups as compared to conventional. However, rate of Class IV pulmonary complications involving prolonged ventilation and re-explorations were similar between the groups. Overall rate of pulmonary complications by many series have reported a low incidence of pulmonary complications with MIE group,[912] also the randomised control trial by Biere et al. published a very low over all pulmonary infection rate,[7] these series however did not include minor class I complications such as basal atelectasis, pleural effusion or minor pneumothorax in their analysis. Furthermore, the true rate of respiratory infections in our series was MIE 28.4% versus OE 21.1% and hybrid 29%. While it was 21.8 for transhiatal and 28% for 3-stage esophagectomies.

Table 4

Frequency of occurrence of various major complications in each group

The long term complication rate for anastomotic stricture (MIE 9.5% vs. 15.6% in conventional group) is superior to recently published data by Maas et al. showing high rates of 44% MIE and 39% OE on the 1 year follow-up of their multi-centric randomised controlled trial patients.[725] In their previously published review of 12 studies, describing various intra-thoracic anastomotic techniques for minimally invasive Ivor-Lewis esophagectomy anastomotic stricture rates ranged from 0% to 27.5%.[26]

Out of thoracic duct injuries only two required thoracotomy and duct ligation rest were managed conservatively with total parenteral nutrition (TPN) and fat free diet. Also in case of anastomotic leak on contrast study, 50% of the leaks were successfully managed conservatively with TPN followed by a prolonged naso-jejunal (NJ) tube feed, with remaining 7 cases requiring a re-exploration (3 conduit failures, 3 tracheo-esophageal fistulas and 1 anastomotic site leak). Other major complications requiring re-exploration were 2 azygous vein rebleeds, 2 gastric outlet obstructions at hiatus, 2 tracheo-bronchial tree injuries, 1 neck haemorrhage and 1 large hiatus diaphragmatic hernia in the long term.

Other complications in Table 4 included post-operative medical problems like fluid and electrolyte disturbances, cardiac issues such as arrhythmias, uncontrolled hypertension, thrombo-embolic phenomenon and urinary tract infections. Aspiration was taken as a separate complication because of its multifactorial relationship to anastomotic narrowing or recurrent laryngeal nerve injury (RLNI), and its wide spectrum consequences of minor cough and aspiration pneumonia to cases requiring prolonged NJ or a feeding jejunostomy for repeated aspirations.

In terms of oncologic efficacy of the specimen as published recently, the mean number of lymphnode harvest in our series was 13.7 ± 5.6 and the rate of R0 resection in patients with post neo-adjuvant residual disease (50% of study population) was 53.4% with statistically insignificant difference among the groups (P > 0.05).[27]

In-general, all post-operative surgical outcomes tended to improve over the course of this study. With the maturation of our program, we recognised a gradually decreasing frequency of complications and lengths of ICU and hospital stay. Very few studies have attempted to show the evolution of minimally invasive techniques in comparison to conventional and hybrid groups.[3141528] Tapias and Morse in their study described the learning curve for MIE of a single surgeon and observed improved operative time, blood loss, median hospital stay and morbidity in their last forty patients compared with the first 40.[14] Similar results with improved operative and peri-operative outcomes were shown by Arlow et al. in their series of 200 open transhiatal esophagectomies performed over 13 years by same set of surgeons.[15] In another recent series on 180 consecutive minimally invasive McKeown esophagectomy by Mu et al. no significant differences in post-operative morbidity were found between their first sixty patients group as compared to second and third sixty; however, there was significantly longer duration of surgery in first group as compared to other two.[3]

Limitations

Although we separated hybrid esophagectomies in our study design for data comparison of outcome variables, yet there were more number of transhiatal esophagectomies in conventional group with a higher percentage of gastro-oesophageal junction tumours in the same group and more three-stage in MI group with an equally higher percentage of oesophageal tumours, that can lead to a potential source of bias for overall results. This was statistically minimised by controlling these variables in multivariate analysis.

Hence, a retrospective study design and heterogeneity of the cases with regards to the type of MIE (more three stage) and conventional open (more transhiatal) resulting in a lack of true match to match comparison of MIE and conventional groups are main limitations for our study.

CONCLUSION

The data suggest that MIE has overall comparable surgical outcomes to its conventional counterpart with a longer operative time and low long-term complication rate. Also, the peri-operative outcome tended to improve in our centre with the maturation of program and experience.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.