International Journal of Colorectal Disease https://doi.org/10.1007/s00384-018-3100-0

ORIGINAL ARTICLE

A new dimensional-reducing variable obtained from original inflammatory scores is highly associated to morbidity after curative surgery for colorectal cancer Martin Bailon-Cuadrado 1 & Baltasar Perez-Saborido 1 & Javier Sanchez-Gonzalez 1 & Mario Rodriguez-Lopez 1 & Agustin Mayo-Iscar 2 & David Pacheco-Sanchez 1 Accepted: 13 June 2018 # Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract Purpose Several scores have been developed to define the inflammatory status of oncological patients. We suspect they share iterative information. Our hypothesis is that we may summarise their information into one or two new variables which will be independent. This will help us to predict, more accurately, which patients are at an increased risk of suffering postoperative complications after curative surgery for CRC. Methods Observational prospective study with those patients undergoing curative surgery for CRC between September 2015 and February 2017. We analysed the influence of inflammatory scores (PNI, GPS, NLR, PLR) on postoperative morbidity (overall and severe complications, anastomotic leakage and reoperation). Results Finally, 168 patients were analysed. We checked these four original scores are interrelated among them. Using a complex and innovative statistical method, we created two new independent variables (resultant A and resultant B) which resume the information coming from them. One of these two new variables (resultant A) was statistically associated to overall complications (OR, 2.239; 95% CI, 1.541–3.253; p = 0.0001), severe complications (OR, 1.773; 95% CI, 1.129–2.785; p = 0.013), anastomotic leakage (OR, 3.208; 95% CI, 1.416–7.268; p = 0.005) and reoperation (OR, 2.349; 95% CI, 1.281–4.305; p = 0.006). Conclusions We evinced the four original scores we used share redundant information. We created two new independent new variables which resume their information. In our sample of patients, one of these variables turned out to be a great predictive factor for the four complications we analysed. Keywords Colorectal cancer . Postoperative complications . Morbidity . Inflammatory status . Inflammatory scores

Introduction Colorectal cancer (CRC) represents about 10% of all tumours worldwide, being the second most frequent in Originality - We declare this manuscript has not been previously presented as a communication in any congress or meeting. - We declare the manuscript has not been previously published and is not under consideration for publication in any other journal. * Martin Bailon-Cuadrado [email protected] 1

General and Digestive Surgery Department, Rio Hortega University Hospital, C/ Dulzaina, n° 2, 47012 Valladolid, Spain

2

Statistics Department, University of Valladolid, Valladolid, Spain

males and the third in females [1]. It predominates in western countries, where the incidence is about 30 new cases per 100,000 people every year. This constitutes three or four times what we can find in developing countries [2]. It appears mainly from the sixth decade of life onwards and it is more common in men. It has a multifactorial aetiology since it is influenced by genetic and environmental factors [3]. CRC surgery presents a relevant rate of postoperative complications. These patients are usually elderly, and they may suffer several previous diseases (arterial hypertension, diabetes mellitus, dyslipidaemia, smoking, heart failure, ischemic heart disease, chronic obstructive pulmonary disease [COPD]). Postoperative morbidity not only affects short-term quality of life, but it also affects long-term oncological outcomes. Some publications have shown

Int J Colorectal Dis

that infectious complications (and anastomotic leakage in particular) decrease both overall survival and relapse-free survival in patients who have undergone surgery for CRC [4, 5]. This is an essential reason why we should improve our capability of detecting patients who are at an increased risk of suffering postoperative complications after this type of surgery. Interaction between the tumour and the host triggers a systemic inflammatory response, in which several immunologic and neuroendocrine changes appear [6]. There is an increase of positive acute phase reactants (C-reactive protein and fibrinogen) and a decrease of negative acute phase reactants (albumin) [7]. A tendency towards anorexia is developed due to the increase in the levels of leptin, which regulates the balance among orexigenic and anorexigenic signals [8]. High cortisol levels, as well as resistance to insulin and growing hormone, provoke an increase of catabolism [9]. In addition, tumour itself releases factors which stimulate catabolism (proteolysis inducing factor and lipid mobilising factor) and slow down anabolism [10]. Subsequently to these findings, several scores have been developed in the last decades in order to define the inflammatory status of oncological patients. Prognostic Nutritional Index (PNI) was created by Onodera and his team from the Tokyo University; it takes into account albumin and lymphocytes [11]. Glasgow Prognostic Score (GPS) was elaborated by McMillan and his team from the Royal Infirmary of Glasgow; it considers albumin and Creactive protein [12]. Neutrophil/lymphocyte ratio (NLR) represents the quotient between the level of neutrophils and lymphocytes [13]. In the same way, platelet/lymphocyte ratio (PLR) represents the quotient between the level of platelets and lymphocytes [14]. These scores present, in our opinion, some disadvantages. First, they share reiterative information since some variables are included in several scores (albumin is part of PNI and GPS, lymphocytes are part of PNI, NLR, and PLR). Secondly, they have been mainly studied in relation to longterm oncological outcomes and much less in relation to the appearance of postoperative complications [15–18]. Finally, there are very few articles which have compared simultaneously the predictive capability of several scores on postoperative morbidity after CRC surgery. Our hypothesis is that we may summarise the information coming from these scores into one or two new variables which will be independent. This will help us to predict, more accurately, which patients are at an increased risk of suffering postoperative complications after curative surgery for CRC.

Method We performed an observational prospective study with those patients who underwent curative surgery for CRC

at our institution between September 2015 and February 2017. Our inclusion criteria were 18 years or older, having signed a specific informed consent for the extraction of a peripheral blood sample in the 24 h prior to surgery, the existence of a pathologic diagnosis of colorectal adenocarcinoma and the indication of curative surgery. Preoperative analytics included the variables needed to calculate PNI, GPS, NLR and PLR. We excluded urgent surgeries due to the impossibility of assessing adequately the inflammatory status of these patients, palliative surgeries and those procedures in which there were intraoperative findings resulting in a diagnosis other than colorectal adenocarcinoma. Our institution’s ethics committee approved this study. We analysed the influence of several variables on the appearance of postoperative complications: sex, age, body mass index (BMI), active diseases (arterial hypertension, diabetes mellitus, dyslipidaemia, smoking, heart failure, ischemic heart disease, COPD), anaesthetic risk (we considered high risk to American Society of Anaesthesiology [ASA] grades III and IV and low risk to grades I and II), preoperative placement of intraluminal stent, neoadjuvant therapy (chemotherapy [CT] and radiotherapy [RT]), tumour location (colon or rectum), surgical procedure (colectomy, proctectomy with anastomosis, Hartmann’s procedure or abdominoperineal excision), intraoperative transfusion and operative time. Inflammatory status was also analysed and it was defined by using PNI, GPS, NLR and PLR. Considering these four variables, we performed the Kolmogorov-Smirnov test to check if these scores’ distribution conformed to normal pattern. Those which did not adapt to normal curve underwent a decimal logarithmic transformation. Later, we carried out the Pearson correlation test in order to find out if these scores were interrelated among them, as we initially suspected. Those which were effectively interrelated were included in factor analysis. With this statistical method, we pretend to resume the information coming from the four original scores into one or two new variables, which will be independent. Starting from the principal components and after applying a Varimax rotation, we performed a dimensional reduction to assess the inflammatory status more accurately, by using only one or two independent new variables. Complications were carefully registered during the first 30 days of postoperative period or until home discharge in the case of longer hospital stays. They were defined with four dichotomic variables: overall complications, severe complications (grades III-V from Clavien-Dindo classification), anastomotic leakage and reoperation. Anastomotic leakage was considered when visualising enteric content through the drain or wound, or when detecting contrast outflow through the

Int J Colorectal Dis

anastomosis during a computed tomography. This complication was analysed exclusively in those patients who underwent an anastomosis, so we discarded abdominoperineal excisions and Hartmann’s procedures. In univariate analysis (UVA), we used chi-square or Fisher tests (when appropriate) for discrete variables and Student t test for continuous variables. A p value < 0.10 was considered for the entry of variables in multivariate analysis (MVA). Logistic regression was used in MVA, in which a p < 0.05 was used to establish statistical significance. Statistical analysis was performed with SPSS (version 18.0, IBM Corporation).

Results Finally, we analysed 168 patients. Mean age was 69 years old and 63.7% of them were males. Considering our dichotomic anaesthetic risk variable, 34.5% were at high risk. Seventytwo percent of the tumours were located at colon, and most common techniques were colectomy (58.3%), proctectomy

with anastomosis (29.8%), abdominoperineal excision (7.1%) and Hartmann’s procedure (4.8%). PNI adapted to normal pattern, with a mean value of 44.88 ± 6.08. GPS distribution was as follows: 61.9% grade 0, 21.4% grade 1 and 16.7% grade 2. NLR and PLR distributions did not conform to normal curve, with median values of 2.19 (interquartile range 1.68– 3.28) and 135.28 (interquartile range 94.11–196.83), respectively. These last two variables underwent a decimal logarithmic transformation, since they had a non-normal distribution. Pearson correlation showed that the four original scores were statistically interrelated among them (p < 0.05). Thereby, they all were included in factor analysis. This dimension-reducing method generated two new independent variables, which we named resultant A and resultant B. The first was influenced by PNI and GPS, while the second was by NLR and PLR [Table 1]. In addition, we obtained the mathematical formula needed to calculate the numeric value of these two new variables starting from the original four scores:

Resultant A ¼ 4; 069–ð0; 074 x PNIÞ þ ð0; 966 x GPSÞ–ð0; 607 x Log NLRÞ–ð0; 491 x Log PLRÞ Resultant B ¼ ‐4; 645–ð0; 019 x PNIÞ–ð0; 304 x GPSÞ þ ð2; 322 x Log NLRÞ þ ð2; 236 x Log PLRÞ

Our rates of postoperative complications were 32.7% for overall complications, 10.7% for severe complications, 7.4% for anastomotic leakage (11 cases in 148 patients with anastomosis) and 6% of reoperation. Up to nine variables were relevant in UVA for overall complications: age, anaesthetic risk, tumour location, surgical technique, PNI, GPS, NLR, PLR and resultant A. In MVA, three of them were statistically significant: age (odds ratio [OR], 1.047; 95% confidence interval [CI], 1.012–1.084; p = 0.008), tumour location (OR, 3.520; 95% CI, 1.550–7.994; p = 0.003) and resultant A (OR, 2.239; 95% CI, 1.541–3.253; p = 0.0001) [Table 2]. For severe complications, three variables were significant in UVA: age, PNI and resultant A. But only resultant A maintained this condition in MVA (OR, 1.773; 95% CI, 1.129– 2.785; p = 0.013) [Table 3]. A total of six variables were relevant in UVA for anastomotic leakage: chemotherapy, tumour Table 1

Factor analysis: scores influencing the new variables

Factor analysis

Resultant A

Resultant B

PNI GPS Log NLR Log PLR

− 0.737 0.949 0.118 0.148

− 0.507 − 0.002 0.924 0.931

location, surgical technique, operative time, PNI and resultant A. In MVA, two of them achieved statistical significance for this complication: tumour location (OR, 27.733; 95% CI, 4.253–180.850; p = 0.001) and resultant A (OR, 3.208; 95% CI, 1.416–7.268; p = 0.005) [Table 4]. For reoperation, three scores were significant in UVA: PNI, GPS and resultant A. But only this last variable sustained statistical significance in MVA (OR, 2.349; 95% CI, 1.281–4.305; p = 0.006) [Table 5].

Discussion and conclusions In oncological patients, a systemic inflammatory response is generated because of the interaction with the tumour. This provokes several immunological and neuroendocrine changes which lead to anorexia, an increase in catabolism and a decrease in anabolism. This way, systemic inflammatory response also impairs nutritional status, which has a negative influence on healing and immune response. This situation may be evinced by albumin, whose value decreases due to a direct absorption by the tumour and due to the extravasation mainly induced by tumour necrosis factor alpha. There exist many indexes which may estimate inflammatory status, but we wanted to focus on the most objective ones (those which we may obtain from a peripheral blood analytics) and on the scores which have been mainly used, in literature,

Int J Colorectal Dis Table 2

Univariate and multivariate analysis for overall complications

Overall complications

Sex

Univariate analysis

Multivariate

Yes

No

p value

Male Female

36 (33.6) 19 (31.1)

71 (66.4) 42 (68.9)

N.S.

Age

Mean value (years)

BMI Arterial hypertension

Mean value (kg/m2) Yes No

72.95 26.95 34 (34.7) 21 (30.0)

67.09 26.69 64 (65.3) 49 (70.0)

0.001 N. S. N.S.

Diabetes mellitus

Yes No

11 (45.8) 44 (30.6)

13 (54.2) 100 (69.4)

N.S.

Dyslipidaemia

Yes No Yes No Yes No

19 (31.7) 36 (33.3) 13 (28.3) 42 (34.4) 4 (57.1) 51 (31.7)

41 (68.3) 72 (66.7) 33 (71.7) 80 (65.6) 3 (42.9) 110 (68.3)

N.S.

Yes No Yes No High Low

5 (35.7) 50 (32.5) 3 (33.3) 52 (32.7) 26 (44.8) 29 (26.4)

9 (64.3) 104 (67.5) 6 (66.7) 107 (67.3) 32 (55.2) 81 (73.6)

Endoscopic stent

Yes No

6 (33.3) 49 (32.7)

12 (66.7) 101 (67.3)

N.S.

Chemotherapy

Yes No Yes No

8 (29.6) 47 (33.3) 12 (37.5) 42 (31.1)

19 (70.4) 94 (66.7) 20 (62.5) 93 (68.9)

N.S.

Colon Rectum Colectomy Proctectomy

34 (28.1) 21 (44.7) 29 (29.6) 14 (28.0)

87 (71.9) 26 (55.3) 69 (70.4) 36 (72.0)

AAP HARTMANN Mean value (min) Yes No Mean value 0 1 2 Mean value Mean value Mean value Mean value

6 (50.0) 6 (75.0) 220.36 2 (66.7) 53 (32.1) 42.28 22 (21.2) 15 (41.7) 18 (64.3) 3.16 195.25 0.501 0.177

6 (50.0) 2 (25.0) 211.77 1 (33.3) 112 (67.9) 46.15 82 (78.8) 21 (58.3) 10 (35.7) 2.51 148.83 − 0.244 − 0.086

Smoking Heart failure Ischaemic heart disease COPD Anaesthetic risk

Radiotherapy Tumour location Surgical technique

Operative time Intraoperative transfusion PNI GPS

NLR PLR Resultant A Resultant B

to define the inflammatory status in patients with colorectal cancer. Therefore, we decided to use PNI, GPS, NLR and PLR. These present some problems; they share redundant information since some variables are part of several of them, they have not been compared among them and they have been principally studied in their relation to oncological prognosis. With the intention of solving these problems, we have

p value

0.008

N.S. N.S. N.S. N.S. 0.015

N.S.

N.S. 0.04

0.003

0.037

N. S.

N.S. N.S. 0.001 0.0001

N. S. N. S.

0.088 0.063 0.0001 N. S.

N.S. N.S. 0.0001

compared original scores, we have checked their correlation and we have tried to resume the information coming from all of them. Our distribution for demographic, preoperative and intraoperative variables is equal to what we may find in other publications with patients undergoing curative surgery for CRC and with a similar methodology [19–21]. Inflammatory

Int J Colorectal Dis Table 3

Univariate and multivariate analysis for severe complications

Severe complications

Sex

Univariate analysis

Multivariate

Yes

No

p value

Male Female

11 (10.3) 7 (11.5)

96 (89.7) 54 (88.5)

N.S.

Age

Mean value (years)

BMI Arterial hypertension

Mean value (kg/m2) Yes No

73.83 26.87 11 (11.2) 7 (10.0)

68.43 26.77 87 (88.8) 63 (90.0)

0.056 N.S. N.S.

Diabetes mellitus

Yes No

1 (4.2) 17 (11.8)

23 (95.8) 127 (88.2)

N.S.

Dyslipidaemia

Yes No Yes No Yes No

6 (10.0) 12 (11.1) 5 (10.9) 13 (10.7) 2 (28.6) 16 (9.9)

54 (90.0) 96 (88.9) 41 (89.1) 109 (89.3) 5 (71.4) 145 (90.1)

N.S.

Yes No Yes No High Low

0 (0.0) 18 (11.7) 0 (0.0) 18 (11.3) 9 (15.5) 9 (8.2)

14 (100.0) 136 (88.3) 9 (100.0) 141 (88.7) 49 (84.5) 101 (91.8)

Endoscopic stent

Yes No

2 (11.1) 16 (10.7)

16 (88.9) 134 (89.3)

N.S.

Chemotherapy

Yes No Yes No

3 (11.1) 15 (10.6) 3 (9.1) 15 (11.1)

24 (88.9) 126 (89.4) 30 (90.9) 120 (88.9)

N. S.

Colon Rectum Colectomy Proctectomy

11 (9.1) 7 (14.9) 10 (10.2) 7 (14.0)

110 (90.9) 40 (85.1) 88 (89.8) 43 (86.0)

AAP HARTMANN Mean value (min) Yes No Mean value

0 (0.0) 1 (12.5) 216.67 0 (0.0) 18 (10.9) 41.56

12 (100.0) 7 (87.5) 214.33 3 (100.0) 147 (89.1) 45.28

7 (6.7) 6 (16.7) 5 (17.9) 2.51 156.41 0.574

97 (93.3) 30 (83.3) 23 (82.1) 2.75 164.94 − 0.069

N.S.

NLR PLR Resultant A

0 1 2 Mean value Mean value Mean value

Resultant B

Mean value

− 0.326

0.039

N.S.

Smoking Heart failure Ischaemic heart disease COPD Anaesthetic risk

Radiotherapy Tumour location Surgical technique

Operative time Intraoperative transfusion PNI GPS

status was established by using originally described scores, which were obtained from values coming from peripheral blood analytics extracted the day before surgery. PNI and GPS values were similar to other articles, while NLR and PLR values were slightly lower than reported in other papers [22–25]. These two scores represent the humoral fraction of

p value

N. S.

N.S. N.S. N.S. N.S. N.S.

N.S. N.S. N.S.

N.S. N.S. 0.014

N.S. N.S. 0.009

N.S.

0.013

the inflammatory response. Possibly, the explanation for these lower values lies in our inclusion and exclusion criteria. We have discarded urgent procedures, which usually present as stenosing and perforated lesions. These situations may suppose peritonitis or intestinal ischemia, triggering the inflammatory response and overestimating NLR and PLR values.

Int J Colorectal Dis Table 4

Univariate and multivariate analysis for anastomotic leakage

Anastomotic leakage

Sex

Univariate analysis

Multivariate

Yes

No

p value

Male Female

6 (6.7) 5 (8.6)

84 (93.3) 53 (91.4)

N.S.

Age

Mean value (years)

BMI Arterial hypertension

Mean value (kg/m2) Yes No

67.91 26.18

68.49 26.84

N.S. N.S.

8 (9.5) 3 (4.7)

76 (90.5) 61 (95.3)

N.S.

Diabetes mellitus

Yes No Yes No

1 (5.6) 10 (7.7) 3 (6.0) 8 (8.2)

17 (94.4) 120 (92.3) 47 (94.0) 90 (91.8)

N.S.

Smoking

Yes No

5 (12.2) 6 (5.6)

36 (87.8) 101 (94.4)

N.S.

Heart failure

Yes No Yes No Yes No

1 (16.7) 10 (7.0) 0 (0.0) 11 (7.9) 0 (0.0) 11 (7.7)

5 (83.3) 132 (93.0) 9 (100.0) 128 (92.1) 6 (100.0) 131 (92.3)

N.S.

High Low Yes No

4 (8.7) 7 (6.9) 0 (0.0) 11 (8.3)

42 (91.3) 95 (93.1) 16 (100.0) 121 (91.7)

Chemotherapy

Yes No

4 (22.2) 7 (5.4)

14 (77.8) 123 (94.6)

0.03

Radiotherapy

Yes No Colon Rectum Colectomy Proctectomy

3 (17.6) 8 (6.1) 4 (3.4) 7 (23.3) 4 (4.1) 7 (14.0)

14 (82.4) 123 (93.9) 114 (96.6) 23 (76.7) 94 (95.9) 43 (86.0)

N.S.

Mean value (min) Yes No Mean value 0 1

240.91 0 (0.0) 11 (7.5) 40.77 4 (3.8) 4 (11.1)

204.45 2 (100.0) 135 (92.5) 45.17 100 (96.2) 32 (88.9)

2 Mean value Mean value Mean value Mean value

3 (10.7) 3.35 247.56 0.581 0.274

25 (89.3) 2.68 158.17 − 0.062 − 0.128

Dyslipidaemia

Ischaemic heart disease COPD Anaesthetic risk Endoscopic stent

Tumour location Surgical technique Operative time Intraoperative transfusion PNI GPS

NLR PLR Resultant A Resultant B

Complications were carefully collected during the first 30 days of postoperative period or until home discharge in the case of longer hospital stays. Our rate of complications is also consistent with literature for this kind of surgery [20, 26]. We proved that the four original scores were interrelated with each other, and we created two new independent variables to resume the information coming from them. Both original and new scores were analysed in relation to the

p value

N.S.

N.S. N.S. N.S. N.S. N.S.

0.001

0.001

0.044

N.S.

0.077 N.S.

N.S.

0.02

N.S.

N.S.

N.S. N.S. 0.042 N.S.

0.005

appearance of postoperative morbidity. One of our two new variables (resultant A) turned out to be an independent predictive factor for overall and severe complications, anastomotic leakage and reoperation. This new score is mainly influenced by PNI and GPS, which represent the cellular fraction of the inflammatory response, and it is more related to nutritional status. No other score reached significance for any complication in MVA.

Int J Colorectal Dis Table 5

Univariate and multivariate analysis for reoperation

Reoperation

SEX

Univariate analysis

Multivariate

Yes

No

p value

MALE Female

5 (4.7) 5 (8.2)

102 (95.3) 56 (91.8)

N.S.

p value

Age

Mean value (years)

BMI Arterial Hypertension

Mean value (kg/m2) Yes No

72.3 25.67

68.8 26.85

N.S. N.S.

6 (6.1) 4 (5.7)

92 (93.9) 66 (94.3)

N.S.

Diabetes mellitus

Yes No Yes No

1 (4.2) 9 (6.3) 3 (5.0) 7 (6.5)

23 (95.8) 135 (93.8) 57 (95.0) 101 (93.5)

N.S.

Smoking

Yes No

2 (4.3) 8 (6.6)

44 (95.7) 114 (93.4)

N.S.

Heart failure

Yes No Yes No Yes No

1 (14.3) 9 (5.6) 0 (0.0) 10 (6.5) 0 (0.0) 10 (6.3)

6 (85.7) 152 (94.4) 14 (100.0) 144 (93.5) 9 (100.0) 149 (93.7)

N.S.

High Low Yes No

5 (8.6) 5 (4.5) 1 (5.6) 9 (6.0)

53 (91.4) 105 (95.5) 17 (94.4) 141 (94.0)

Chemotherapy

Yes No

1 (3.7) 9 (6.4)

26 (96.3) 132 (93.6)

N.S.

Radiotherapy

Yes No Colon Rectum Colectomy Proctectomy

0 (0.0) 10 (7.4) 7 (5.8) 3 (6.4) 7 (7.1) 3 (6.0)

33 (100.0) 125 (92.6) 114 (94.2) 44 (93.6) 91 (92.9) 47 (94.0)

N.S.

0 (0.0) 0 (0.0) 200 0 (0.0) 10 (6.1) 40.35 2 (1.9) 5 (13.9) 3 (10.7) 2.25 137.59 0.929

12 (100.0) 8 (100.0) 215.51 3 (100.0) 155 (93.9) 45.17 102 (98.1) 31 (86.1) 25 (89.3) 2.75 165.7 − 0.059

0.015 0.019

N.S. N.S.

NLR PLR Resultant A

AAP HARTMANN Mean value (min) YES NO Mean value 0 1 2 Mean value Mean value Mean value

N.S. N.S. 0.002

0.006

Resultant B

Mean value

− 0.329

0.021

N.S.

Dyslipidaemia

Ischaemic heart disease COPD Anaesthetic risk Endoscopic stent

Tumour location Surgical technique

Operative time Intraoperative transfusion PNI GPS

Not many publications have analysed the influence of inflammatory scores on postoperative morbidity. Mohri et al., Tokunaga et al. and Cao et al. have studied the influence of PNI on overall and severe complications after colorectal surgery. Mohri et al. published a retrospective article with 365 patients. They used receiver operating characteristic (ROC) curves to establish a cut-off point for PNI, thus creating two

N.S.

N.S. N.S. N.S. N.S.

N.S. N.S.

N.S. N.S.

risk groups. PNI seemed to be an independent predictive factor for both overall and severe complications in MVA [27]. Tokunaga et al. presented a retrospective paper with 556 patients. They also used ROC curves to establish a cut-off value for PNI. In MVA, this score was statistically significant for severe but not for overall complications [28]. This same result was obtained by Cao et al. in their retrospective article with 583

Int J Colorectal Dis

patients. ROC curves were also used by these authors to turn PNI into a dichotomous variable [29]. Josse et al. analysed the influence of NLR on severe complications and anastomotic leakage. They also used ROC curves to create two risk groups. NLR turned out to be an independent predictive factor for severe complications in MVA, but not for anastomotic leakage [30]. Unlike these authors, we have respected the continuous nature of PNI, NLR and PLR. Cut-off points are different in every single publication, so conclusions are difficultly applicable to other samples of patients. Another difference is that our work has been prospective. In our opinion, retrospective collection of complications might not be as strict as with a prospective approach. Finally, we may perceive that every article analyses the influence of only a score. Considering it is suspected that inflammatory scores share iterative information among them, we defend that several scores should be studied simultaneously. This is a lack we have tried to solve. Our work presents obvious limitations. It has been performed in a single centre and our sample is not very large. However, other publications with a similar approach and methodology have an equal or even smaller sample size [31–33]. Our strict inclusion and exclusion criteria (limited to programmed and curative surgery for colorectal adenocarcinoma) provoke that our results might not be easily extrapolated to other samples of patients. Finally, results involving our two new variables should be validated in other sample coming from the same population. Inflammatory scores have been generally analysed in relation to long-term outcomes [34–40]. When publications have assessed their influence on postoperative morbidity, they have used one or two scores and they have studied few complications, usually with a retrospective approach [27–30]. The novelty of our work is that we have compared four scores, in a prospective way, and we have analysed four different complications. In addition, we have obtained two new independent variables which resume the information from the four original scores. One of them (resultant A) has been significantly related to the four postoperative complications we considered. Interesting future applications may be developed, as long as our results concerning this new score will be validated. Its numeric formula might be included in a mobile phone application or a web page in order to ease the calculation of resultant A directly from the variables obtained in the peripheral blood analytics, thus avoiding the calculation of the four original scores. This value will be translated into a risk (a percentage) of suffering a certain complication. For high-risk patients, it might be proposed the possibility of delaying a surgery, whenever it is possible, with the intention of improving the nutritional status, as European Society for Clinical Nutrition and Metabolism suggests [41]. This could be done with supplements rich in proteins. Nevertheless, this possible clinical management should only be taken

into account when this new score will be validated in a much larger population. We may conclude that the four inflammatory scores we used (PNI, GPS, NLR and PLR) present redundant information among them. We have created two new independent variables which resume the information coming from them. In our sample of patients, one of these two new scores (Resultant A) has turned out to be a great predictive factor for all the complications we studied (overall and severe complications, anastomotic leakage and reoperation) after curative surgery for CRC. Utility of this new score must be validated in other samples of patients. Author contribution We declare every author has actively participated in the drafting and design of this manuscript. We declare every author is pleased with the final version of this manuscript. Funding We have received funding from a governmental public health agency (Gerencia Regional de Salud de la Junta de Castilla y León) through a research project with title BInfluencia del estado nutricional e inflamatorio preoperatorio en la morbilidad y mortalidad postoperatoria de la cirugía curativa del cáncer de colon y recto^ and file number BGRS 1316/A/16.^

Compliance with ethical standards Conflict of interest The authors declare that they have no conflict of interest. Research involving human participants and/or animals All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent Informed consent was obtained from all individual participants included in the study.

References 1.

Aran V, Victorino AP, Thuler LC, Ferreira CG, Levine JS, Ahnen DJ et al (2016) Colorectal cancer: epidemiology, disease mechanisms and interventions to reduce onset and mortality. Clin Colorectal Cancer 15(3):195–203 2. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D (2011) Global cancer statistics. CA Cancer J Clin 61(2):69–90 3. Doubeni CA, Major JM, Laiyemo AO, Schootman M, Zauber AG, Hollenbeck AR, Sinha R, Allison J (2012) Contribution of behavioral risk factors and obesity to socioeconomic differences in colorectal cancer incidence. J Natl Cancer Inst 104(18):1353–1362 4. Law WL, Choi HK, Lee YM, Ho JWC, Seto CL (2007) Anastomotic leakage is associated with poor long-term outcome in patients after curative colorectal resection for malignancy. J Gastrointest Surg 11(1):8–15 5. Pucher PH, Aggarwal R, Qurashi M, Darzi A (2014) Meta-analysis of the effect of postoperative in-hospital morbidity on long-term patient survival. Br J Surg 101(12):1499–1508 6. Fearon K, Strasser F, Anker SD, Bosaeus I, Bruera E, Fainsinger RL, Jatoi A, Loprinzi C, MacDonald N, Mantovani G, Davis M,

Int J Colorectal Dis

7. 8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Muscaritoli M, Ottery F, Radbruch L, Ravasco P, Walsh D, Wilcock A, Kaasa S, Baracos VE (2011) Definition and classification of cancer cachexia: an international consensus. Lancet Oncol 12(5): 489–495 Lu H, Ouyang W, Huang C (2006) Inflammation, a key event in cancer development. Mol Cancer Res 4(4):221–233 Alemán MR, Santolaria F, Batista N, de La Vega MJ, GonzálezReimers E, Milena A et al (2002) Leptin role in advanced lung cancer. A mediator of the acute phase response or a marker of the status of nutrition? Cytokine 19(1):21–26 Strasser F, Palmer JL, Schover LR, Yusuf SW, Pisters K, Vassilopoulou-Sellin R, DeGracia B, Willey JS, Bruera E (2006) The impact of hypogonadism and autonomic dysfunction on fatigue, emotional function, and sexual desire in male patients with advanced cancer: a pilot study. Cancer 107(12):2949–2957 Blum D, Stene GB, Solheim TS, Fayers P, Hjermstad MJ, Baracos VE, Fearon K, Strasser F, Kaasa S, Block Lieve V, Koen M, Sarah B, Augusto C, Joachim C, Massimo C, Anneke F, Richard H, Higginson Irene J, Stein K, Karen L, Guido M, Bregje OP, Koen P, Roeline P, Sophie P, Sheila P, Luc D (2014) on behalf of EuroImpact Validation of the consensus-definition for cancer cachexia and evaluation of a classification model—a study based on data from an international multicentre project (EPCRC-CSA). Ann Oncol 25(8):1635–1642 Onodera T, Goseki N, Kosaki G (1984) Prognostic nutritional index in gastrointestinal surgery of malnourished cancer patients. Nihon Geka Gakkai Zasshi 85(9):1001–1005 Forrest L, Mcmillan D, Mcardle C, Angerson W, Dunlop D (2004) Comparison of an inflammation-based prognostic score (GPS) with performance status (ECOG) in patients receiving platinum-based chemotherapy for inoperable non-small-cell lung cancer. Br J Cancer 90:1704–1706 Walsh SR, Cook EJ, Goulder F, Justin TA, Keeling NJ (2005) Neutrophil-lymphocyte ratio as a prognostic factor in colorectal cancer. J Surg Oncol 91(3):181–184 Zou ZY, Liu HL, Ning N, Li S-Y, Du XH, Li R (2016) Clinical significance of pre-operative neutrophil lymphocyte ratio and platelet lymphocyte ratio as prognostic factors for patients with colorectal cancer. Oncol Lett 11(3):2241–2248 Nozoe T, Kohno M, Iguchi T, Mori E, Maeda T, Matsukuma A, Ezaki T (2012) The prognostic nutritional index can be a prognostic indicator in colorectal carcinoma. Surg Today 42(6):532–535 Choi KW, Hong SW, Chang YG, Lee WY, Lee B, Paik IW, Lee H (2014) Inflammation-based score (Glasgow prognostic score) as an independent prognostic factor in colorectal cancer patients. Ann Surg Treat Res 86(6):309–313 Pine JK, Morris E, Hutchins GG, West NP, Jayne DG, Quirke P, Prasad KR (2015) Systemic neutrophil-to-lymphocyte ratio in colorectal cancer: the relationship to patient survival, tumour biology and local lymphocytic response to tumour. Br J Cancer 113(2):204–211 Tan D, Fu Y, Su Q, Wang H (2016) Prognostic role of plateletlymphocyte ratio in colorectal cancer: a systematic review and meta-analysis. Medicine (Baltimore) 95(24):e3837 Siegel RL, Miller KD, Fedewa SA, Ahnen DJ, Meester RGS, Barzi A, Jemal A (2017) Colorectal cancer statistics, 2017. CA Cancer J Clin 67(3):177–193 Artinyan A, Orcutt ST, Anaya DA, Richardson P, Chen GJ, Berger DH (2015) Infectious postoperative complications decrease longterm survival in patients undergoing curative surgery for colorectal cancer: a study of 12,075 patients. Ann Surg 261(3):497–505 Hede P, Sörensson MÅ, Polleryd P, Persson K, Hallgren T (2015) Influence of BMI on short-term surgical outcome after colorectal cancer surgery: a study based on the Swedish national quality registry. Int J Color Dis 30(9):1201–1207

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Shibutani M, Maeda K, Nagahara H, Ohtani H, Iseki Y, Ikeya T, Sugano K, Hirakawa K (2015) The prognostic significance of the postoperative prognostic nutritional index in patients with colorectal cancer. BMC Cancer 15:521 Ishizuka M, Nagata H, Takagi K, Iwasaki Y, Kubota K (2012) Inflammation-based prognostic system predicts postoperative survival of colorectal cancer patients with a normal preoperative serum level of carcinoembryonic antigen. Ann Surg Oncol 19(11):3422– 3431 Emir S, Aydin M, Can G, Bali I, Yildirim O, Öznur M et al (2015) Comparison of colorectal neoplastic polyps and adenocarcinoma with regard to NLR and PLR. Eur Rev Med Pharmacol Sci 19(19):3613–3618 You J, Zhu GQ, Xie L, Liu WY, Shi L, Wang OC, Huang ZH, Braddock M, Guo GL, Zheng MH (2016) Preoperative platelet to lymphocyte ratio is a valuable prognostic biomarker in patients with colorectal cancer. Oncotarget 7(18):25516–25527 Slankamenac K, Slankamenac M, Schlegel A, Nocito A, Rickenbacher A, Clavien P-A, Turina M (2017) Impact of postoperative complications on readmission and long-term survival in patients following surgery for colorectal cancer. Int J Color Dis 32(6):805–811 Mohri Y, Inoue Y, Tanaka K, Hiro J, Uchida K, Kusunoki M (2013) Prognostic nutritional index predicts postoperative outcome in colorectal cancer. World J Surg 37(11):2688–2692 Tokunaga R, Sakamoto Y, Nakagawa S, Miyamoto Y, Yoshida N, Oki E, Watanabe M, Baba H (2015) Prognostic nutritional index predicts severe complications, recurrence, and poor prognosis in patients with colorectal cancer undergoing primary tumor resection. Dis Colon Rectum 58(11):1048–1057 Cao X, Zhao G, Yu T, An Q, Yang H, Xiao G (2017) Preoperative prognostic nutritional index correlates with severe complications and poor survival in patients with colorectal cancer undergoing curative laparoscopic surgery: a retrospective study in a single Chinese institution. Nutr Cancer 69(3):454–463 Josse JM, Cleghorn MC, Ramji KM, Jiang H, Elnahas A, Jackson TD, Okrainec A, Quereshy FA (2016) The neutrophil-tolymphocyte ratio predicts major perioperative complications in patients undergoing colorectal surgery. Color Dis 18(7):O236–O242 Gohil R, Rishi M, Tan BHL (2014) Pre-operative serum albumin and neutrophil-lymphocyte ratio are associated with prolonged hospital stay following colorectal cancer surgery. Br J Med Med Res 4(1):481–487 Shimazaki J, Tabuchi T, Nakachi T, Motohashi G, Nishida K, Ubukata H et al (2015) Pre-operative granulocyte/lymphocyte ratio as a predictive marker of post-operative complications in patients with colorectal cancer. Oncol Lett 9(1):425–429 MacKay GJ, Molloy RG, O’Dwyer PJ (2011) C-reactive protein as a predictor of postoperative infective complications following elective colorectal resection. Color Dis 13(5):583–587 Park BK, Park JW, Han EC, Ryoo SB, Han SW, Kim TY, Chie EK, Jeong SY, Park KJ (2016) Systemic inflammatory markers as prognostic factors in stage IIA colorectal cancer. J Surg Oncol 114(2): 216–221 Lin MS, Huang JX, Yu H (2015) Prognostic significance of Glasgow prognostic score in patients with stage II colorectal cancer. Int J Clin Exp Med 8(10):19138–19143 Kubo H, Murayama Y, Arita T, Kuriu Y, Nakanishi M, Otsuji E (2016) The prognostic value of preoperative neutrophil-to-lymphocyte ratio in colorectal cancer. World J Surg 40(11):2796–2802 Ozawa T, Ishihara S, Nishikawa T, Tanaka T, Tanaka J, Kiyomatsu T, Hata K, Kawai K, Nozawa H, Kazama S, Yamaguchi H, Sunami E, Kitayama J, Watanabe T (2015) The preoperative platelet to

Int J Colorectal Dis lymphocyte ratio is a prognostic marker in patients with stage II colorectal cancer. Int J Color Dis 30(9):1165–1171 38. Sugimoto K, Komiyama H, Kojima Y, Goto M, Tomiki Y, Sakamoto K (2012) Glasgow prognostic score as a prognostic factor in patients undergoing curative surgery for colorectal cancer. Dig Surg 29:503–509 39. Li MX, Liu XM, Zhang XF, Zhang JF, Wang WL, Zhu Y, Dong J, Cheng JW, Liu ZW, Ma L, Lv Y (2014) Prognostic role of neutrophil-to-lymphocyte ratio in colorectal cancer: a systematic review and meta-analysis. Int J Cancer 134(10):2403–2413

40.

41.

Kwon HC, Kim SH, Oh SY, Lee S, Lee JH, Choi HJ, Park KJ, Roh MS, Kim SG, Kim HJ, Lee JH (2012) Clinical significance of preoperative neutrophil-lymphocyte versus platelet-lymphocyte ratio in patients with operable colorectal cancer. Biomarkers 17(3): 216–222 Weimann A, Braga M, Carli F, Higashiguchi T, Hübner M, Klek S, Laviano A, Ljungqvist O, Lobo DN, Martindale R, Waitzberg DL, Bischoff SC, Singer P (2017) ESPEN guideline: clinical nutrition in surgery. Clin Nutr 36(3):623–650