Introduction

The idea behind the current Enhanced Recovery After Surgery (ERAS) protocols were first introduced by Kehlet [1]. The aim is to optimize recovery from major surgery through patient guidance and multidisciplinary collaboration, taking into account evidence-based perioperative care [2]. This treatment model was originally created for patients undergoing colorectal surgery [3]. The ERAS Society was officially registered in 2010, and since then, ERAS guidelines have been applied to many surgical fields. For major head and neck surgery with free flap reconstruction, the recommendations for optimal perioperative care were published by the ERAS Society in 2017, to optimize clinical outcome by influencing perioperative care in an evidence-based and structured manner [4].

Reconstruction of tissue defects with microvascular  free flaps is the standard of care in the surgical treatment of advanced head and neck cancer (HNC). However, complications tend to be common, reportedly even in over 70% of patients [5]. Many of these patients are older, with significant comorbidities, and are often heavy users of alcohol and tobacco predisposing them to both surgical and medical complications.

Several studies on HNC patients undergoing free tissue reconstruction have compared outcome before and after implementation of ERAS guidelines (Supplementary Table 1). The most common finding seems to be significantly reduced hospital length of stay (LOS). In a systematic review and meta-analysis recently published by Chorath et al., the number of wound complications was lower, LOS shorter, and the readmission rate reduced in patients treated using ERAS guidelines [6].

The aim of this study was to investigate the effects of the implementation of ERAS protocol at our tertiary care center. Special emphasis was placed on the effects on LOS, length of intensive care stay (ICU–LOS), complications, and mortality within 6 months after operation. We hypothesized that ERAS implementation would reduce these parameters.

Materials and methods

Retrospective data were collected from the hospital registry at the Helsinki University Hospital (HUS) (Helsinki, Finland) and comprised 283 patients undergoing HNC surgery with free flap reconstruction between 2013 and 2020. The ERAS protocol was implemented in October 2017. Patients operated before that timepoint comprised the pre-ERAS group (N = 169). Patients operated after October 2017 comprised the ERAS group (N = 114). The collected parameters to be compared between the groups were patient demographics (age, sex), ACE-27 comorbidity index [7], tumor location, TNM classification and stage, the type of free flap used, medical and surgical complications, mortality within 30 days or 6 months after surgery, LOS, ICU–LOS, and the need for re-operation or re-admission to intensive care. The primary endpoints were LOS, ICU–LOS, and complications (both surgical and medical).

The ERAS protocol used at our institution was designed by a multidisciplinary group, including otolaryngologist—head and neck surgeons, plastic surgeons, oral and maxillofacial surgeons and anesthesiologists. The cornerstone of our protocol was the recommendation for optimal perioperative care in major HNC surgery with free flap reconstruction published by the ERAS Society, which was modified to match our local operational environment [4]. During the study period, our ERAS protocol was updated once: in October 2018 a recommendation of rapid awakening after operation was added. The principles of our ERAS guidelines are presented in Supplementary Table 2.

The aim of the statistical modelling was to study difference between pre-ERAS and ERAS periods. All data were manually extracted from the electronic hospital charts and transferred to SPSS statistical software. Statistical analyses were performed using IBM SPSS Statistics 25 for Windows (IBM Corp., Armonk, NY) and R language [8].

We modeled continuous variables with linear regression. Three models were applied: Model 1 only with ERAS period as predictor; Model 2 with ERAS period, sex, and age; Model 3 with ERAS period, sex, age, stage, and ACE-27 score. Linear models estimate difference between pre-ERAS and ERAS with 95% confidence intervals.

Dichotomic (no/yes) variables were modelled with modified Poisson regression that estimates relative risk between pre-ERAS and ERAS [9]. This modelling approach was used instead of standard logistic regression, because some of the depended variables have high prevalence, and thus odds ratios produced by logistic regression do not estimate relative risk  so well [10]. Results of modified Poisson regression are presented as relative risks with 95% confidence intervals.

Institutional permission to conduct this study was granted by the Research Administration of the HUS District (HUS/419/2018, HUS/307/2019). This research involved only patient charts, and therefore, no formal Research Ethics Board approval or informed consent was needed according to the Finnish legislation.

Results

The patient characteristics are shown in Table 1.

Table 1 Descriptive statistics of the study cohort
Full size table

In the pre-ERAS and ERAS groups the mean LOS and ICU–LOS were 20 days (range 7–79) and 13 days (range 3–70), and 6 days (range 1–32) and 5 days (range 1–24), respectively. Both LOS and ICU–LOS were statistically reduced by a linear logistic regression model in the ERAS group compared to the pre-ERAS group (Table 2).

Table 2 Linear regression model for LOS and ICU–LOS compared to ERAS and pre-ERAS groups
Full size table

No difference was found in the overall complication rate: in the pre-ERAS and ERAS groups, 115 patients (68%) and 76 patients (67%) had complications (p = 0.808), respectively (Table 3). Neither were there any significant differences in the rate of surgical complications between the two groups. The rate of medical complications was significantly lower in the ERAS group using modified Poisson regression model (Table 4). In both the pre-ERAS and ERAS groups, the majority of complications belonged to Clavien–Dindo grades IIIb (pre-ERAS 26.6%, ERAS 21.1%) and IVa (pre-ERAS 26%, ERAS 22.8%) (data not shown) [11].

Table 3 Comparison of postoperative data
Full size table
Table 4 Modified Poisson regression model for independent variables comparing ERAS group to pre-ERAS group
Full size table

No statistically significant differences in mortality were found: in the pre-ERAS group the 30-day and 6-month mortality were 1.8% and 12.4%, and in the ERAS group 4.4% and 12.3%, respectively (Tables 3, 4).

The number of re-operations within 30 days in the pre-ERAS and ERAS groups were 64 (37.9%) and 35 (30.7%), respectively (p = 0.266) (Table 3).

The rate of re-admissions to the ICU or prolonged stay in the ICU was significantly higher in the ERAS group compared to the pre-ERAS group (38% vs. 19%, p < 0.001) (Table 3).

Discussion

In 2017 at our institution, an evidence-based ERAS recommendation was implemented aiming at improving the quality of perioperative care of patients undergoing major HNC surgery. In this study, we focused on the effects of ERAS guidelines especially on LOS, ICU–LOS, and complication rate. In patients undergoing surgery after implementation of ERAS guidelines, the mean LOS was reduced by 7 days and ICU–LOS by 1 day, but no difference was found in overall or surgical complication rates. However, the rate of medical complications decreased. Our results are in line with several former studies. In most published studies, the LOS has been reduced after implementation of ERAS guidelines, but only in  one study was the complication rate affected (Supplementary Table 1) [5, 12,13,14,15,16,17,18,19,20].

Postoperative complications are common in HNC surgery with free flap reconstruction and have been associated with increased LOS, morbidity, and decreased overall survival [21]. In our study, the overall complication rate in the pre-ERAS group was 68.0% and in the ERAS group 66.7%. These figures are relatively high compared to some previous studies, in which complication rates of 20.3–41% have been reported [22, 23]. On the other hand, several studies are in line with our study showing higher complication rates of 54–72% [5, 24,25,26]. However, there may be methodological differences in reporting complications between different studies. In many studies the complication rate has not been significantly altered after implementation of the ERAS protocol [5, 13, 14, 16, 18,19,20]. In the study by Kiong et al., as in our study, the use of ERAS guidelines resulted in significantly fewer medical complications [17]. In our study the use of the ERAS protocol reduced the rate of pneumonia, pulmonary edema, prolonged mechanical ventilation, delirium, and surgical site infection. We evaluated all complications extremely carefully and might have reported minor complications more precisely than some other studies. However, the high number of complications did not negatively affect the LOS or ICU–LOS, which were comparable to that reported by others. Twomey et al. found that patient mobilization later than 24 h after surgery is associated with all types of complications, including major complications, encouraging early mobilization [27].

The flap loss rate in the ERAS group was surprisingly high (Tables 3, 4). It seems unlikely that the changes in our treatment protocol related to ERAS would affect flap survival. As the recent review by Ronen et al. suggests, around half of surgical complications are preventable, so we will evaluate our flap loss rate carefully in the future [28]. In addition, the number of re-operations was high in both groups compared to other studies [16, 17, 19]. Methodological differences in reporting complications may at least partly explain this finding as we recorded all minor wound revisions as re-operations.

The rate of re-admissions to the ICU or prolonged stay in the ICU was significantly higher in the ERAS group compared to the pre-ERAS group (Table 4). In both groups, the reasons for re-admissions included both re-operations and medical complications. This is of cause contradictory to the aims of using the ERAS guidelines. One explanation could be the increased need for new reconstruction procedures due to increased number of flap losses in the ERAS group. This issue warrants careful preoperative patient assessment and selection.

A major change in our practice along with the implementation of the ERAS protocol has been in the preoperative nutritional care. All patients will be provided with preoperative supplementary nutrition preparations (immunonutrition, Oral Impact®) for 1 week before surgery. Mueller et al. showed benefit of preoperative immunonutrition in HNC surgery, i.e., a decrease in the overall complication rate and reduced LOS [29]. Dort et al. recommended that preoperative fasting should be minimized [4]. According to our ERAS protocol, patients are offered a carbohydrate-rich supplement (200 ml) 2 h before anesthesia, which has been shown to reduce postoperative insulin resistance and LOS [30, 31].

In the ERAS guidelines published by the ERAS Society in 2017, and also in our ERAS protocol, antibiotic prophylaxis is recommended to be continued for only 24 h postoperatively [4]. Before we launched the ERAS guidelines, prophylactic antibiotics were used for significantly longer periods at our institution. According to the present results, the number of infections did not increase after the significant shortening of the prophylactic antibiotic treatment. Instead, the rate of surgical site infections and pneumonias decreased, and the changes were statistically significant in the modified Poisson regression model (Table 4). Numerous studies have shown that long-term postoperative antibiotics do not protect against surgical site infections [32,33,34]. On the contrary, they may cause antibiotic-related complications, such as Clostridium difficile infections [35].

At our institution, prolonged mechanical ventilation after HNC surgery with free flap reconstruction has been routine practice. The patients have been sedated and ventilated overnight in the ICU and extubated or weaned from the ventilator on the day after surgery. However, several ERAS guidelines recommend avoiding prolonged sedation and mechanical ventilation [36,37,38]. In a study by Clemens et al. rapid awakening protocol in patients undergoing HNC surgery significantly decreased complication rate [39]. After being in use for 1 year, we updated our ERAS protocol and added a recommendation of rapid awakening, whenever assessed safe and possible; the patient should have a guaranteed airway and no need for ventilatory support. The surgeon and the anesthesiologist together must decide whether rapid awakening is advisable. The adherence to this recommendation has not been optimal at our institution but is gradually improving. It might be that the patients are easier to manage in the ICU during the night following operation when the sedation is continued until the next morning. In addition, old unproved beliefs that prolonged immobility somehow protects the microvascular anastomosis may affect these practices. Many studies have shown that these patients benefit from rapid awakening [17, 18, 27]. According to the study by Clemens et al. the patients that benefit most from this are older HNC patients and those with significant comorbidities with limited cardiopulmonary reserve [39].

In the current study, the use of the ERAS protocol had no effect on mortality within 30 days or 6 months after surgery. In a recently published systematic review and meta-analysis by Chorath et al., 4 of the 16 studies included data on early mortality and there was no difference between patients treated according to the ERAS guidelines and those that were not [6].

The strength of our study is the reasonably large number of patients compared to many other similar studies. There are some limitations that should be considered when interpreting our data. This study has limitations related to possible biases caused by the retrospective design. The patient groups are heterogeneous with possible confounding factors. For reasons unclear to us, the UICC cancer stages of the patients were not evenly distributed: in the pre-ERAS group there were more patients with stage IV disease than in the ERAS group. We hypothesize, however, that this does not cause a remarkable bias, because the same kind of surgery with microvascular flap reconstruction was performed in all patients. In addition, we did not monitor the adherence to the ERAS protocol during the study period. Only in a few studies, has the adherence to the ERAS protocol been monitored in HNC patients [12, 15, 20]. The ERAS protocol includes multiple interventions which makes it difficult to demonstrate the benefits of a single intervention.

Conclusion

In conclusion, the implementation of the ERAS protocol for major HNC surgery at our institution was feasible and safe and resulted in shortening of LOS, ICU–LOS, and reduced rate of medical complications, while no effect was found on the rate of surgical complications or short-term postoperative mortality. The multimodal nature of the ERAS protocol warrants updating and collaboration between different disciplines to improve its application.