The sample was composed of 40 patients, 26 of whom were males (65%) and 14 of whom were females (35%). 9 of these patients received neoadjuvant chemotherapy alone and 3 patients received neoadjuvant chemotherapy and radiotherapy 1 month prior to surgery. None of these patients received any neoadjuvant treatment within 1 month of surgery in order to eliminate the influence of treatment on the preoperative CBC and so the lymphocyte percentage, PLR and NLR, as well as other preoperative labs [11]. The average number of comorbidities was 2.2 with comorbid diabetes present in 25% of patients, hypertension in 37.5% and cardiovascular disease in 42.5%. 12.5% had a second primary malignancy at the time of presentation. 55% of patients had a history of smoking.
The median operative time was 255 minutes (interquartile range (IQR) = 1.1) with a minimum operative time of 180 minutes and a maximum of 420 minutes. The median day of oral feeding resumption was 8 days with a minimum of 7 days and a maximum of 19 days. The median LOICUS was 3 days (IQR = 4.5) with a maximum stay of 22 days observed for 1 patient and a minimum stay of 0 days observed for 4 patients. The median LOHS was 14 days (IQR = 6.25) with a maximum stay of 45 days for 1 patient and a minimum of 9 days for 1 patient.
18 patients (47.4%) developed postoperative complications. Mild complications (Clavien-Dindo Grade 1) were observed in 12 patients (30%). Major complications (Clavien-Dindo Grade 2, 3 or 4) were seen in 6 patients (15%). Of these patients, 2 developed a cervical anastomotic leak that required no surgical intervention. It is worth noting that 1 patient developed pneumonia and ARDS post-operatively and, due to a prolonged ICU stay of 15 days, required a feeding jejunostomy for nutritional support.
Lymphocyte Percentage
The mean pre-operative lymphocyte percentage was 24.98%. Pearson correlation revealed a negative correlation between lymphocyte percentage (Pearson coefficient = -0.365, p-value = 0.043) and LOHS. Similarly, a negative association existed for LOICUS (Pearson coefficient = -0.385, p-value = 0.033) and day of oral feeding resumption (Pearson coefficient = 0.423, p-value = 0.022).
Neutrophil to Lymphocyte Ratio
The average preoperative NLR was 2.94. Pearson correlation revealed that a higher preoperative NPR was positively correlated with an increase in LOHS (Pearson coefficient = 0.454, p-value = 0.01), LOICUS (Pearson coefficient = 0.422, p-value = 0.018) and delayed resumption of oral feeding (Pearson coefficient = 0.517, p-value = 0.004). Figure 1 demonstrates the association between LOICUS and preoperative lymphocyte percentage as well as LOICUS and preoperative NLR. Figure 2 demonstrates the association between postoperative day of feeding resumption and the preoperative lymphocyte percentage. Figure 3 demonstrates the association between postoperative day of feeding resumption and the NLR.
Platelet to Lymphocyte Ratio
The average preoperative PLR was 165.6. Similar to the pre-op NLR, Pearson correlation revealed that a higher preoperative PLR was positively correlated with an increase in LOICUS (Pearson coefficient = 0.364, p-value = 0.044) and delayed resumption of oral feeding (Pearson coefficient = 0.578, p-value = 0.001). A correlation between pre-op PLR and LOHS was not statistically significant (Pearson coefficient = 0.284, p-value = 0.12). Figure 4 demonstrates the association between postoperative day of feeding resumption and the PLR.
Peri-operative Packed Red Blood Cell Transfusions
Patients received a median of 1 transfusion (IQR = 2.5) and an average of 1.625 transfusions of PRBCs. The number of PRBCs transfusions was found to be positively correlated with an increase in LOHS (Pearson coefficient = 0.383, p-value = 0.03), LOICUS (Pearson coefficient = 0.597, p-value <0.001) and a delayed resumption of oral feeding (Pearson coefficient = 0.566, p-value <0.001).
Patient Comorbidities
The presence of DM, HTN and CVD was correlated with and increased LOICUS (p-value = 0.022, 0.011 and 0.028 respectively) but only DM was correlated with a delay in feeding resumption (p-value = 0.033). Table 1 demonstrates the correlation of patients’ comorbidities with LOICUS in days.
Univariate Analysis
Univariate analysis showed that lymphocytes percentage, NLR, PLR, PRBCs and patients’ comorbidities were all correlated with an increased LOHS, LOICUS and a delay in feeding resumption. All were statistically significant with p <0.05. The correlation between operative time and LOHS, LOICUS and delay in oral feeding resumption was, however, not statistically significant.
Multivariate Analysis
On multivariate analysis, a delay of feeding resumption is shown to be the most powerful predictor of increased LOHS (p-value < 0.01) and LOICUS (p-value < 0.01). In turn, LOICUS (p-value <0.01) and pre-op PLR (p-value <0.01) are found to be the most powerful predictors of a delay in resumption of oral feeding. Moreover, multivariate analysis did not demonstrate a statistically significant correlation between operative time, LOHS, LOICUS and delay in feeding resumption.