Accepted Manuscript

NUTRISCORE: A new nutritional screening tool for oncological outpatients

Lorena Arribas, Laura Hurtós, Maria José Sendrós, Inmaculada Peiró, Neus Salleras,
Eduard Fort, Jose Manuel Sánchez-Migallón

PII: S0899-9007(16)30158-7
DOI: 10.1016/[Link].2016.07.015
Reference: NUT 9817

To appear in: Nutrition

Received Date: 14 March 2016

Revised Date: 20 July 2016
Accepted Date: 22 July 2016

Please cite this article as: Arribas L, Hurtós L, Sendrós MJ, Peiró I, Salleras N, Fort E, Sánchez-Migallón
JM, NUTRISCORE: A new nutritional screening tool for oncological outpatients, Nutrition (2016), doi:
10.1016/[Link].2016.07.015.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Title: NUTRISCORE: A new nutritional screening tool for oncological outpatients

Author names and affiliations:
Lorena Arribas a *, Laura Hurtós a, Maria José Sendrós b, Inmaculada Peiró a, Neus Salleras c,
Eduard Fort a, Jose Manuel Sánchez-Migallón b
a Clinical Nutrition Unit, Institut Català d’Oncologia (ICO), L’Hospitalet de Llobregat, Barcelona,
Spain
b Clinical Nutrition and Dietetics Department, Institut Català d’Oncologia (ICO), Badalona,

PT
Barcelona, Spain
c Nutrition and Dietetics Department, Institut Català d’Oncologia (ICO), Girona
Corresponding author:
* Corresponding author Tel. +34 93 260 7751

RI
Email address: larribas@[Link] (L. Arribas)
Present/permanent address:
Clinical Nutrition Unit

SC
Institut Català d’ Oncologia (ICO)
Granvia de l’Hospitalet, 199-203
08908 L’Hospitalet de Llobregat (Barcelona)
Spain

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

1 NUTRITION MANUSCRIPT

3 Introduction

4 Malnutrition is a problem that greatly affects cancer patients throughout the course of their illness, and it may

PT
5 be present from the moment of diagnosis until the end of treatment. It is estimated that up to 84% of patients

6 have lost weight by the time they are diagnosed, with just over half having lost over 5% of their body weight1.

The prevalence of malnutrition varies between oncological inpatients (44.1%) and outpatients (27.7%)1.

RI
7

8 Among patients with haematological malignancies, the prevalence of malnutrition has not been studied

SC
9 extensively, but some studies report a 27% risk in patients who undergo haematopoietic stem cell

10 transplantation2. Numerous studies have demonstrated the negative impact that malnutrition has on

11 oncological patients, reducing the tolerance and the efficacy of their treatment3, increasing the risk of clinical

12
U
and surgical complications4, and lengthening hospital stay with the concomitant increase in healthcare costs
AN
5,6
13 . Despite its importance, malnutrition remains an unsolved and poor visibility problem for the professionals

14 in charge of caring for these patients.

M

15 Many cancer patients lose weight as a direct result of their treatment, whose side effects may include

16 anorexia, mucositis, nausea and/or vomiting, xerostomy, disgeusia and diarrhoea, all of which contribute to
D

17 the deterioration of their nutritional status7. Weight loss associated with decreased caloric intake, together
TE

18 with metabolic changes, characterize cancer cachexia. This concept has been defined by expert consensus as a

19 multifactorial syndrome characterised by steady loss of skeletal muscle mass (with or without loss of adipose
EP

20 tissue), which cannot be completely reversed through conventional nutritional support and which leads to a

21 gradual functional deterioration8. Early detection of malnutrition in the pre-cachectic phase could avert the
C

22 progression of this syndrome to more advanced stages.

AC

23 The first step in the early detection of malnutrition is the implementation of a screening tool to identify

24 patients at nutritional risk. A variety of instruments exist, such as the Nutritional Risk Screening (NRS-

25 2002)9, Malnutrition Universal Screening Tool (MUST)10, Malnutrition Screening Tool (MST)11,12, Short

26 Nutritional Assessment Questionnaire (SNAQ)13 and the Subjective Global Assessment (SGA)14. Although

27 most of them have been designed for the general population, some have been validated for oncological

28 patients in both inpatient and outpatient settings15–17.

1
 ACCEPTED MANUSCRIPT

29 The Oncology Nutrition Dietetic Practice Group of the American Dietetic Association has adopted the Scored

30 Patient-Generated Subjective Global Assessment (PG-SGA) as a specific nutritional assessment and screening

31 tool for oncological patients18, but its use as a screening tool is limited by the need for specially trained staff

32 and the length of time needed to carry out the assessment estimated in approximately 15 minutes19.

33 To achieve efficiency gains among nutrition professionals without unduly increasing workload, screening

PT
34 tools must be accurate, fast and simple20. The objective of nutritional screening is to identify patients at high

35 nutritional risk so that a more thorough evaluation can be carried out specifically on them. Although there are

RI
36 some tools in the literature, none of them is specifically designed to detect nutritional risk among cancer

37 outpatients. The rising number of oncological outpatient treatments has created the need for a new tool that is

SC
38 especially designed to the ambulatory setting. This new tool needs to be fast, simple and only detect those

39 patients at risk of malnutrition aimed to minimise the false positives in order to avoid overwork and maximise

40 optimize the professional resources available.

U
AN
41

42 The aim of our study was to design a new nutritional screening tool for oncological outpatients to detect
M

43 nutritional risk.

44
D

45 Methods
TE

46 2.1. Study design

47 This was a multi-centre, cross-sectional study designed to validate the NUTRISCORE tool (Figure 1).
EP

48 Assessment of nutritional risk among outpatients was performed at the Catalan Institute of Oncology (ICO). It

49 is a public centre working exclusively in the field of cancer and, to date, has three centres in Catalonia

50 (L’Hospitalet de Llobregat –Barcelona-, Girona and Badalona) that work together with three public general
C

51 university hospitals.
AC

52 2.2 Subjects

53 Adult outpatients (≥18 years of age) diagnosed with malignant neoplasm, including solid tumours and

54 haematological malignancies, who visited our centre for oncological evaluation, treatment, or palliative and

55 symptomatic care were eligible for the study. Patients were excluded if they were unable to understand and

56 speak Spanish, had a definitive diagnosis of dementia or lacked capacity to understand the purpose of the

57 study.

2
 ACCEPTED MANUSCRIPT

58 Due to the characteristics of our center, all patients included were in oncospecific treatment, evaluation,

59 follow up after treatment or palliative care. Study subjects presented different stages of cancer, from early

60 diagnosis to more advanced stages. Surgery may be present throughout the treatment.

61 Recruitment to the study was designed to obtain a representative sample in terms of tumour location and

62 treatment groups within each hospital. Random patients lists based on all diagnoses and treatments were

PT
63 generated by the Research Data Management and Statistics Unit the day prior to study recruitment. Patients

64 from each list were approached randomly and asked if they would like to participate.

RI
65

SC
66 2.3. Ethics

67 The study protocol was approved by the Hospital Universitari de Bellvitge Ethics Committee for Clinical

68 Research (PR285/12). All patients provided written informed consent.

69

U
AN
70 2.4. Development of the NUTRISCORE tool

71 The MST is a good nutritional screening test for the oncological patient, consisting of two questions related to
M

72 weight loss and decreased caloric intake. This is the screening tool recommended by the Spanish oncology

73 societies (the Spanish Society for Medical Oncology, SEOM, and the Spanish Society for Radiation
D

74 Oncology, SEOR), as well as by the Spanish Society for Enteral and Parenteral Nutrition (SENPE). However,
TE

75 in our experience, its implementation in a comprehensive cancer centre is not cost-effective. The high number

76 of false positives makes difficult to provide nutritional assessment to all patients due to the limited resources

available21. Modifying the criterion to increase specificity would result in a reduced sensitivity. Moreover a
EP

77

78 greater complexity would be introduced with the triage, by adding a test to compensate this lower sensitivity.

79 Therefore we decided to improve the test by designing a new screening tool increasing sensitivity and
C

80 specificity. We have used the MST as a basis and introduced specific timeframes for both questions into our
AC

81 tool. Our purpose was to improve the tool by incorporating two differential items: tumour location and

82 treatment

83 Given that the tumour site22 and the oncological treatment23 have an enormous impact on the nutritional

84 status11,12,24–28, we aimed to design a method that considered these parameters. We organised an expert

85 consensus meeting with professionals from different dietetic and nutrition units from each centre belonging to

86 the Catalan Institute of Oncology (L’Hospitalet de Llobregat –Barcelona-, Girona and Badalona), with the

3
 ACCEPTED MANUSCRIPT

87 participation of specialists in oncology, haematology and palliative care, to define the key characteristics of

88 high nutritional risk patients for all of the pathologies attended, considering tumour site and treatment course

89 in addition to other aspects. We also used data available from the literature classifying the nutritional risk

90 according to the tumour location and the treatment22,23.

91

PT
92 Given that oncological treatment can affect metabolic and nutritional status29, some related considerations

93 were taken into account:

RI
94 - Patients were considered to be undergoing chemotherapy from the beginning of their first cycle

95 until three weeks after the last, including during the periods between cycles, even if treatment

SC
96 was delayed due to toxicity concerns.

97 - Patients were considered to be undergoing radiotherapy from the first session until two weeks

98 after the last.

U
AN
99 - With regard to haematopoietic stem cell transplantation, this treatment was considered until one

100 week after hospital discharge post transplantation.

M

101 - Other treatments: this section of the tool considered monoclonal antibodies against membrane

102 receptors, tyrosine kinase inhibitors, mTor inhibitors, anti-angiogenesis drugs, hormone therapy,
D

103 and other treatments, including for symptomatic care (corticoids, analgesia, etc.).
TE

104

105 The PG-SGA tool30 was used as the reference method of nutritional screening to classify patients without
EP

106 (PG-SGA Stage A) or with (PG-SGA Stages B+C) nutritional risk. For the purpose of comparison we have

107 decided to unify the classification of PG-SGA B (moderately malnourished or at risk of malnutrition) and C

108 (severely malnourished). We are validating a screening tool that only gives you the risk but it does not tell
C

109 you if the patient is malnourished, for that we will need to perform a validated nutritional assessment. To be
AC

110 able to used PG-SGA as a reference method to validate nutriscore and for comparative purpose we can only

111 classify patients between at risk or not at risk of malnutrition.

112 Since NUTRISCORE was designed to categorise oncological outpatients according to the presence of

113 nutritional risk, using a scoring system, patients who obtained ≥ 5 points were considered at risk, while those

114 who scored <5 were not at risk. These results were obtained from a pilot study (unpublished) using the cutoff

115 point that maximized the AUC of 0.94 comparing our screening with the reference method (PG- SGA).

4
 ACCEPTED MANUSCRIPT

116

117 2.5. Assessment of the NUTRISCORE

118 Face-to-face interviews were carried out by a trained dietician when patients were attending routine visits at

119 any time during the course of the disease during a study period of eight months. Although NUTRISCORE is a

120 tool thought to be used by any health professional, we used a trained dietician in this study to assure an

PT
121 accurate validation with the reference method. The interviewer evaluated nutritional risk using three different

122 tools: first with the new tool NUTRISCORE, then with the MST and finally with the PG-SGA (reference

RI
123 method).

124 Weight and height was registered. Patients were weighed with lightweight clothing and no shoes in an upright

SC
125 position if possible; if patients were unable to stand on the scale, a sitting scale was used. All scales were

126 regularly calibrated. For the measurement of the height, patients were barefoot in an upright position if

127

U
possible, and if not, height was estimated by doubling the arm span measurement (from the patient´s sternal
AN
128 notch to the end of the longest finger)31. These values were recorded along with weight loss, changes in

129 dietary intake and current symptoms, along with data from the electronic patient record: age, sex, diagnosis,
M

130 treatment and albumin levels.

131
D

132 2.6. Statistical analysis

TE

133 Sample size was based on the correlation between screening tools in patients with different types of cancer. A

134 sample size of 394 patients was calculated to detect as significant a 0.7 correlation between screening tools
EP

135 and a proportion of malnourished patient around 0.3, with a significance level of 0.05 and a statistical power

136 of 0.8.

137 For qualitative variables, frequency and proportions were used and compared by chi-square test, with Yate's
C

138 correction. For quantitative variables, means and standard deviation (SD) values are presented and compared
AC

139 with t-students test or ANOVA, as appropriate.

140 Sensitivity, specificity, and positive and negative predictive values were calculated for NUTRISCORE and

141 MST using the PG-SGA as a reference method. The performance of NUTRISCORE and MST in relation to

142 the PG-SGA was checked by the ROC curve analysis based upon the area under the curve (AUC).

143 To compare the time (in minutes) required for the application of each nutritional screening tool we conducted

144 a comparison of means (t-student).

5
 ACCEPTED MANUSCRIPT

145 The overall agreement between pairs of nutritional screening tools the Kappa coefficient (k) was also

146 assessed. The value of k was classified as: 0.2 < k ≥ 0.0, denoting poor agreement between both tools; 0.4 < k

147 ≥ 0.2, fair agreement; 0.6 < k ≥ 0.4, moderate agreement; 0.8 < k ≥ 0.6, good agreement; and ≥ 0.8, almost

148 excellent agreement32.

149 Data analysis was performed using the SPSS software v.20 (SPSS Inc, Chicago, IL, USA) and R Software

PT
150 version 3.2.3 use to evaluate diagnostic tests. Statistical significance was reported at the conventional P < 0.05

151 level.

RI
152

SC
153 Results

154 General characteristics of study participants

155 Out of 400 eligible patients, a total of 394 were recruited while six patients declined to participate in the

156

U
study. Detailed characteristics of the study participants are shown in Table 1. The mean age was 61.5 ± 12.1
AN
157 years in the study group, which consisted of 217 (55.1%) men and 177 women (44.9%). Our study population

158 included patients with different types of solid tumours (87.5%) and haematological malignancies (12.5%).
M

159 At the time of screening, 83% of the patients reported no weight loss, or a loss of < 5% of their basal body

160 weight in the last three months, while 67 patients (17%) had lost >5% in that time period.
D

161
TE

162 Comparison for classifying risk of malnutrition

163 The classification of the risk of malnutrition according to NUTRISCORE and the other screening methods
EP

164 used is presented in Table 2.

165 The PG-SGA classified 19% of the patients as malnourished or at risk of malnutrition (Stages B+C); 14 of

166 these patients were severely malnourished. This proportion was higher for the MST, with 28.2% of the
C

167 patients classified as at risk of malnutrition, while our new screening tool (NUTRISCORE), found that 22.6%
AC

168 of patients were at risk of malnutrition.

169 When looking at the tumour sites, grouped according to the nutritional risk (low, moderate, high) (table 2), the

170 highest proportion of patients at risk of malnutrition in the three categories was observed for the MST. The

171 NUTRISCORE and the PG-SGA classified as at risk of malnutrition about the same proportion of patients

172 with tumour sites with low and moderate risk (12% and 17% respectively), while a higher proportion was

173 observed for the NUTRISCORE (49.5%) as compared with the PG-SGA (34.1%) among the patients

6
 ACCEPTED MANUSCRIPT

174 According to the oncological treatment, patients undergoing concomitant chemotherapy or haematopoietic

175 stem cell transplantation presented a greater risk of malnutrition, as evaluated by all three methods (table 2).

176 Time taken to carry out the NUTRISCORE compared to the PG-SGA, confirmed to be quicker for our test.

177 0.33 (SD± 0.21) minutes were used for NUTRISCORE, while for the PG-SGA 3.27 (SD± 0.69) minutes were

178 required.

PT
179

180

RI
181 Validity of the NUTRISCORE

182 Using PG-SGA as a reference method, the MST had a sensitivity of 84% and a specificity of 85.6%, whereas

SC
183 NUTRISCORE exceeded these values, at 97.3% sensitivity and 95.9% specificity (table 3). This better

184 performance of NUTRISCORE against MST was confirmed by the ROC curve analysis, with area under the

185

U
curve (AUC) values of 0.95 (CI 0.92-0.98) for NUTRISCORE and 0.84 (CI 0.79-0.89) for the MST.
AN
186 The agreement between NUTRISCORE and the PG-SGA was high, with a kappa index of κ= 0.88 (p<

187 0.0001; CI 95(0.82-0.94), compared to kappa κ= 0.59 (p< 0.0001; CI 95(0.50-0.68) between the MST and the
M

188 PG-SGA that was lower.

189
D

190 Discussion
TE

191 In this study, we developed a new nutritional screening tool for oncological outpatients: NUTRISCORE.

192 When compared with PG-SGA as a reference, the new tool has demonstrated better sensitivity and specificity.
EP

193 Moreover application of NUTRISCORE requires very short time.

194 None of the nutritional screening tools published and validated up to now have been designed specifically for
C

195 oncological outpatients. Some authors have designed screening tools specifically for oncological inpatients.

196 Kim et al.30 validated a tool through an equation that considered changes in intake, weight loss, functional
AC

197 state and body mass index (BMI). Compared to the PG-SGA, it showed sensitivity of 94% and specificity of

198 84.2%. Recently, Zekri et al.33 proposed a new screening tool that classified risk of malnutrition into three

199 stages, taking into account gastrointestinal symptoms as well as changes in oral intake, BMI and weight loss.

200 However, these authors provide no details on the sensitivity or specificity of the screening method. Shaw et

201 al.34 have also published a new nutritional screening method for hospitalised cancer patients with four items:

7
 ACCEPTED MANUSCRIPT

202 weight loss, subjective assessment of low weight, changes in intake and the presence of symptoms that

203 influence intake. Their tool had a sensitivity of 93% and a specificity of 53% compared to the PG-SGA, the

204 method of reference.

205 The multidisciplinary clinical guidelines on nutrition management for cancer patients35, published in Spain in

206 2008 as a result of a consensus process between Spanish societies of nutrition and oncology (SENPE, SEOM

PT
207 and SEOR), recommend using the MST as the screening tool of choice to detect nutritional risk in oncological

208 patients. As reflected in Van Bokhorst et al.36, the MST is considered a good method of nutritional screening

RI
209 for oncological outpatients, and it is one of the most commonly used in oncology. Although studies show it to

210 be highly specific, in our experience, implementing this nutritional screening tool in our comprehensive

SC
211 cancer centres made it difficult to provide nutritional services for all the patients identified as at nutritional

212 risk, because of the high number of false positives and the limited resources available to our centres’ nutrition

213

U
teams. The false positives are probably attributable to the early detection and diagnoses of cancer as well as
AN
214 the combination of conventional antineoplastic treatments with new, targeted therapies that make a smaller

215 nutritional impact. Using the MST as a basis, we designed a specific nutritional screening tool for oncological
M

216 outpatients.

217 Our screening test considers information on weight loss and changes in food intake in a specific period of
D

218 time, in addition to tumour site and type of treatment. These aspects have never been included in existing
TE

219 screenings, despite being important factors that can condition nutritional status37. With regard to weight loss,

220 in our sample, 83% of patients had not lost more than 5% of their body weight in the previous three months, if
EP

221 they’d lost any at all. All the nutritional screenings published until now have the weight as an easy parameter

222 to measure without consideration of other circumstances such as imbalance in the hydration status. We have

223 used this same parameter to be able to compare it with the rest of the nutritional screenings available.
C

224 However, adding the tumour site and treatment may take into account signs of imbalance in the hydration
AC

225 status to some extent.

226 Beyond the overall proportion of patients with risk of malnutrition according to the screening method (MST,

227 28.2%; NUTRISCORE, 22.6%; and PG-SGA stage B+C, 19%), the type of treatment and the tumour site

228 significantly condition nutritional status and future risk of malnutrition. Sites in the head and neck, upper GI

229 tract and some lymphomas are associated with a high nutritional risk1,28, while locations in the lung and most

230 locations in the abdominal and pelvic area present a moderate risk. Some studies that focused on the same

8
 ACCEPTED MANUSCRIPT

231 tumour site and were treated with radiation38, chemotherapy39 or combinations of both40–42, present different

232 figures on malnutrition, demonstrating the importance of this factor when evaluating nutritional risk.

233 Although none of the existing screening methods consider both of these aspects, upon analysis of our tool, we

234 found significant differences with the MST and the PG-SGA according to the tumour site and the treatment

235 (p<0.05). Moreover, and unlike other methods, ours did not include BMI because this measure does not take

PT
236 into account the influence of other non-nutritional factors such as edema or ascites.

237 Although it would have been interesting to include also the tumour stage, we needed to use easy parameters to

RI
238 be able to fill the nutritional screening as simple as possible. Furthermore it is often difficult to have reliable

239 information on the stage and it may change during the course of the treatment and diagnosis. That is why we

SC
240 have chosen to add only tumour location and treatment as easier parameters for the purpose of the screening

241 and we have obtained better sensitivity and specificity.

242

U
As compared with the reference (PG-SGA), NURISCORE has better screening features than MST, reaching
AN
243 sensitivity and specificity (97.3% and 95.9%, respectively and AUC of 0.95. In our population these

244 parameter from NUTRISCORE result in reasonable good predictive values for the risk of malnutrition
M

245 (PPV=84.8%; PPV=73/86). Apart from it good performance, NUTRISCORE has also proven to be a quick

246 and simple tool of screening, which is useful at a clinical level. It is likely to be easy to implement for any
D

247 health professional involved in patient treatment, just like other existing methods including the MST and
TE

248 NRS-2002, although the later has not been assessed in our study. In our case, the dietitian was highly trained

249 for the purpose of the study and we had some data available prior the nutritional assessment so we are aware
EP

250 that the time consumed to perform the nutritional assessment is probably quicker than the time cited in the

251 literature19.

252
C

253 Our study has a few limitations. One of the main limitations is the potential impact on the results, the clinical
AC

254 situation and the patients’ level of cooperation that may have been generated from the need for patients to

255 respond to three different and consecutive nutritional screening tools. Another limitation is interviewer bias,

256 as the interviewer was a single, trained dietician with broad experience in cancer patients. As noted above, the

257 reproducibility of the screening by health professionals with no specific training in oncology nutrition has not

258 been assessed in our study.

9
 ACCEPTED MANUSCRIPT

259 In conclusion, NUTRISCORE proved to be a novel, fast and valid nutritional screening tool for cancer

260 patients, performing excellently in the ambulatory setting. Its simplicity and high level of accuracy in

261 detecting nutritional risk facilitates its applicability. Further studies are currently evaluating the

262 implementation of this new screening method by other types of health professionals to better characterise its

263 reproducibility.

PT
264

265 References

RI
266 1. Hébuterne X, Lemarié E, Michallet M, de Montreuil CB, Schneider SM, Goldwasser F. Prevalence of
267 malnutrition and current use of nutrition support in patients with cancer. JPEN J Parenter Enteral Nutr

SC
268 [Internet]. 2014 Feb [cited 2014 Oct 9];38(2):196–204. Available from:
269 [Link]
270 2. Horsley P, Bauer J, Gallagher B. Poor nutritional status prior to peripheral blood stem cell
271 transplantation is associated with increased length of hospital stay. Bone Marrow Transplant

U
272 [Internet]. 2005 Jun [cited 2014 Oct 13];35(11):1113–6. Available from:
273 [Link]
AN
274 3. Andreyev HJN, Norman a. R, Oates J, Cunningham D. Why do patients with weight loss have a
275 worse outcome when undergoing chemotherapy for gastrointestinal malignancies? Eur J Cancer.
276 1998;34(4):503–9.
277 4. Bozzetti F, Gianotti L, Braga M, Di Carlo V, Mariani L. Postoperative complications in
M

278 gastrointestinal cancer patients: the joint role of the nutritional status and the nutritional support. Clin
279 Nutr [Internet]. 2007 Dec [cited 2014 Aug 25];26(6):698–709. Available from:
280 [Link]
D

281 5. Van Cutsem E, Arends J. The causes and consequences of cancer-associated malnutrition. Eur J
282 Oncol Nurs [Internet]. 2005 Jan [cited 2014 Jul 29];9 Suppl 2:S51–63. Available from:
283 [Link]
TE

284 6. Tong H, Isenring E, Yates P. The prevalence of nutrition impact symptoms and their relationship to
285 quality of life and clinical outcomes in medical oncology patients. Support Care Cancer [Internet].
286 2009 Jan [cited 2014 Sep 2];17(1):83–90. Available from:
287 [Link]
EP

288 7. Bosaeus I. Nutritional support in multimodal therapy for cancer cachexia. Support Care Cancer
289 [Internet]. 2008 May [cited 2014 Aug 29];16(5):447–51. Available from:
290 [Link]
C

291 8. Fearon K, Strasser F, Anker SD, Bosaeus I, Bruera E, Fainsinger RL, et al. Definition and
292 classification of cancer cachexia: an international consensus. Lancet Oncol [Internet]. Elsevier Ltd;
AC

293 2011 May [cited 2014 Jul 15];12(5):489–95. Available from:

294 [Link]
295 9. Kondrup J. Nutritional risk screening (NRS 2002): a new method based on an analysis of controlled
296 clinical trials. Clin Nutr [Internet]. 2003 Jun [cited 2014 Jul 21];22(3):321–36. Available from:
297 [Link]
298 10. Elia M. Screening for malnutrition: a multidisciplinary responibility. Development and use of the
299 “Malnutrition Universal Screening Tool” (MUST) for adults. Malnutrition Advis Group, a Standing
300 Com BAPEN. 2003;Redditch: .
301 11. Ferguson M, Capra S, Bauer J, Banks M. Development of a valid and reliable malnutrition screening
302 tool for adult acute hospital patients. Nutrition [Internet]. 1999 Jun;15(6):458–64. Available from:
303 [Link]

10
 ACCEPTED MANUSCRIPT

304 12. Isenring E, Cross G, Daniels L, Kellett E, Koczwara B. Validity of the malnutrition screening tool as
305 an effective predictor of nutritional risk in oncology outpatients receiving chemotherapy. Support
306 Care Cancer [Internet]. 2006 Nov [cited 2014 Sep 2];14(11):1152–6. Available from:
307 [Link]
308 13. Kruizenga HMM, Seidell JCC, de Vet HCWCW, Wierdsma NJJ, van Bokhorst–de van der Schueren
309 MAE, van Bokhorst-de van der Schueren M a E. Development and validation of a hospital screening
310 tool for malnutrition: the short nutritional assessment questionnaire (SNAQ). Clin Nutr [Internet].
311 2005 Feb [cited 2014 Sep 26];24(1):75–82. Available from:

PT
312 [Link]
313 14. Baker JP, Detsky AS, Wesson DE, Wolman SL, Stewart S, Whitewell J, et al. Nutritional assessment:
314 a comparison of clinical judgement and objective measurements. N Engl J Med [Internet]. 1982 Apr
315 22 [cited 2014 Nov 3];306(16):969–72. Available from:

RI
316 [Link]
317 15. Kyle UG, Kossovsky MP, Karsegard VL, Pichard C. Comparison of tools for nutritional assessment
318 and screening at hospital admission: a population study. Clin Nutr [Internet]. 2006 Jun [cited 2014

SC
319 Aug 20];25(3):409–17. Available from: [Link]
320 16. Leuenberger M, Kurmann S, Stanga Z. Nutritional screening tools in daily clinical practice: the focus
321 on cancer. Support Care Cancer [Internet]. 2010 May [cited 2014 Sep 2];18 Suppl 2:S17–27.
322 Available from: [Link]

U
323 17. Boléo-Tomé C, Monteiro-Grillo I, Camilo M, Ravasco P. Validation of the Malnutrition Universal
324 Screening Tool (MUST) in cancer. Br J Nutr [Internet]. 2012 Jul [cited 2014 Sep 2];108(2):343–8.
AN
325 Available from: [Link]
326 18. Bauer J, Capra S, Ferguson M. Use of the scored Patient-Generated Subjective Global Assessment
327 (PG-SGA) as a nutrition assessment tool in patients with cancer. Eur J Clin Nutr [Internet]. 2002 Aug
328 [cited 2014 Jul 10];56(8):779–85. Available from: [Link]
M

329 19. Kwang AY, Kandiah M. Objective and subjective nutritional assessment of patients with cancer in
330 palliative care. Am J Hosp Palliat Care [Internet]. 2010 Mar [cited 2014 Nov 3];27(2):117–26.
331 Available from: [Link]
D

332 20. Kondrup J. ESPEN Guidelines for Nutrition Screening 2002. Clin Nutr [Internet]. 2003 Aug [cited
333 2014 Jul 11];22(4):415–21. Available from:
334 [Link]
TE

335 21. Sanchez-Migallon J, Joaquim C, Sendros M. Validación de un nuevo método de cribado nutricional
336 para pacientes onco-hematológicos: resultados preliminares. Barcelona; 2010.
337 22. Wie G-A, Cho Y-A, Kim S-Y, Kim S-M, Bae J-M, Joung H. Prevalence and risk factors of
EP

338 malnutrition among cancer patients according to tumor location and stage in the National Cancer
339 Center in Korea. Nutrition [Internet]. Elsevier Ltd; 2010 Mar [cited 2014 Oct 9];26(3):263–8.
340 Available from: [Link]
341 23. Shike M, MF B. Supportive Care of the Cancer Patient. Cancer Princ Pract Oncol Philadelphia. 1989;
C

342 24. Stratton RJ, King CL, Stroud MA, Jackson AA, Elia M. “Malnutrition Universal Screening Tool”
343 predicts mortality and length of hospital stay in acutely ill elderly. Br J Nutr [Internet]. 2006 Feb
AC

344 [cited 2014 Sep 17];95(2):325–30. Available from: [Link]

345 25. Stratton RJ, Green CJ, Elia M. Disease-related malnutrition: an evidence-based approach to treatment.
346 CABI Publishing, CAB International. OXON UK; 2003.
347 26. Isenring E a, Capra S, Bauer JD. Nutrition intervention is beneficial in oncology outpatients receiving
348 radiotherapy to the gastrointestinal or head and neck area. Br J Cancer [Internet]. 2004 Aug 2 [cited
349 2014 Aug 26];91(3):447–52. Available from:
350 [Link]
351 bstract
352 27. Ferguson ML, Bauer J, Gallagher B, Capra S, Christie DR, Mason BR. Validation of a malnutrition
353 screening tool for patients receiving radiotherapy. Australas Radiol [Internet]. 1999 Aug [cited 2014

11
 ACCEPTED MANUSCRIPT

354 Sep 17];43(3):325–7. Available from: [Link]

355 28. Dewys WD, Begg C, Lavin PT, Band PR, Bennett JM, Bertino JR, et al. Prognostic effect of weight
356 loss prior to chemotherapy in cancer patients. Eastern Cooperative Oncology Group. Am J Med
357 [Internet]. 1980 Oct [cited 2014 Sep 17];69(4):491–7. Available from:
358 [Link]
359 29. Tallman MS, Gray R, Robert NJ, LeMaistre CF, Osborne CK, Vaughan WP, et al. Conventional
360 adjuvant chemotherapy with or without high-dose chemotherapy and autologous stem-cell
361 transplantation in high-risk breast cancer. N Engl J Med [Internet]. 2003 Jul 3;349(1):17–26.

PT
362 Available from: [Link]
363 30. Kim J-Y, Wie G-A, Cho Y-A, Kim S-Y, Kim S-M, Son K-H, et al. Development and validation of a
364 nutrition screening tool for hospitalized cancer patients. Clin Nutr [Internet]. Elsevier Ltd; 2011 Dec
365 [cited 2014 Sep 2];30(6):724–9. Available from: [Link]

RI
366 31. Jensen GL, Hsiao PY, Wheeler D. Adult nutrition assessment tutorial. JPEN J Parenter Enteral Nutr
367 [Internet]. 2012 May [cited 2014 Aug 15];36(3):267–74. Available from:
368 [Link]

SC
369 32. Altman DG. Practical statistics for medical research. New York: Chapman and Hall; 1991.
370 33. Zekri J, Morganti J, Rizvi A, Sadiq B Bin, Kerr I, Aslam M. Novel simple and practical nutritional
371 screening tool for cancer inpatients: a pilot study. Support Care Cancer [Internet]. 2014 May [cited

U
372 2014 Sep 2];22(5):1401–8. Available from: [Link]
373 34. Shaw C, Fleuret C, Pickard JM, Mohammed K, Black G, Wedlake L. Comparison of a novel, simple
AN
374 nutrition screening tool for adult oncology inpatients and the Malnutrition Screening Tool (MST)
375 against the Patient-Generated Subjective Global Assessment (PG-SGA). Support Care Cancer
376 [Internet]. 2014 Jun 20 [cited 2014 Sep 2]; Available from: [Link]
377 014-2319-8
M

378 35. Álvarez Hérnandez J, Muñoz Carmona D, Planas Vila M, Rodriguez Rodriguez I, Sánchez Rovira P,
379 Seguí Palme M. Guía clínica multidisciplinar sobre el manejo de la nutrición en el paciente con
380 cáncer. Madrid: Prodrug Multimedia; 2008.
381 36. van Bokhorst-de van der Schueren M a E, Guaitoli PR, Jansma EP, de Vet HCW. Nutrition screening
D

382 tools: does one size fit all? A systematic review of screening tools for the hospital setting. Clin Nutr
383 [Internet]. Elsevier Ltd; 2014 Feb [cited 2014 Oct 7];33(1):39–58. Available from:
TE

384 [Link]
385 37. Gomez-Candela C, Rodriguez L, Luengo L, Zamora P, Celaya S, Zarazaga A, et al. Intervención
386 Nutricional en el Paciente Oncológico Adulto. Editorial Glosa, editor. Barcelona; 2003.
387 38. Lescut N, Personeni E, Desmarets M, Puyraveau M, Hamlaoui R, Servagi-Vernat S, et al. [Evaluation
EP

388 of a predictive score for malnutrition in patients treated by irradiation for head and neck cancer: a
389 retrospective study in 127 patients]. Cancer Radiother [Internet]. Elsevier Masson SAS; 2013 Nov
390 [cited 2014 Oct 7];17(7):649–55. Available from: [Link]
C

391 39. Silver HJ, Dietrich MS, Murphy BA. Changes in body mass, energy balance, physical function, and
392 inflammatory state in patients with locally advanced head and neck cancer treated with concurrent
393 chemoradiation after low-dose induction chemotherapy. Head Neck [Internet]. 2007 Oct [cited 2014
AC

394 Oct 9];29(10):893–900. Available from: [Link]

395 40. Langius J a E, Zandbergen MC, Eerenstein SEJ, van Tulder MW, Leemans CR, Kramer MHH, et al.
396 Effect of nutritional interventions on nutritional status, quality of life and mortality in patients with
397 head and neck cancer receiving (chemo)radiotherapy: a systematic review. Clin Nutr [Internet].
398 Elsevier Ltd; 2013 Oct [cited 2014 Sep 19];32(5):671–8. Available from:
399 [Link]
400 41. Berg MGA Van Den, Ru H, Rasmussen-conrad EL, Knuijt S, Takes RP. Nutritional status , food
401 intake , and dysphagia in long-term survivors with head and neck cancer treated with
402 chemoradiotherapy : A cross-sectional study. 2014;36(1):1–6.
403 42. Valentini V, Marazzi F, Bossola M, Miccichè F, Nardone L, Balducci M, et al. Nutritional

12
 ACCEPTED MANUSCRIPT

404 counselling and oral nutritional supplements in head and neck cancer patients undergoing
405 chemoradiotherapy. J Hum Nutr Diet [Internet]. 2012 Jun [cited 2014 Oct 7];25(3):201–8. Available
406 from: [Link]
407
408
409

PT
RI
U SC
AN
M
D
TE
C EP
AC

13
 ACCEPTED MANUSCRIPT

410 Figure 1-. NUTRISCORE (novel nutritional screening tool).

411 NUTRISCORE
412
413 A. Have you lost weight involuntarily in the last 3 months?
414
No 0
415
I am not sure 2
416

PT
417 If yes, how much weight (in kilograms) have you lost?
418
1-5 1
419
6-10 2
420
11-15 3

RI
421
>15 4
422
Unsure 2
423 B. Have you been eating poorly in the last week because of a decreased

SC
424 appetite?
425
No 0
426
Yes 1
427

U
Location / Neoplasm Nutritional risk Score
Head and neck High* +2
AN
Upper GI tract: oesophagus, gastric, pancreas,
intestines
Lymphoma that compromised GI tract
Lung Medium +1
Abdominal and pelvis: liver, biliary tract, renal,
M

ovaries, endometrial
Breast Low +0
Central Nervous System
Bladder, prostate
D

Colorectal
Leukaemia, other lymphomas
TE

Others
Treatment YES (+2) NO (+0)
The patient is receiving concomitant chemo
radiotherapy
The patient is receiving hyper fractionated radiation
EP

therapy
Haematopoietic stem cell transplantation

YES (+1) NO (+0)

C

The patient is receiving chemotherapy

The patient is only receiving radiotherapy
AC

YES (+0) NO (+0)

Other treatments or only symptomatic treatment

428 *Please repeat the screening every week for those patients at high risk
429
430 Total Score
431
432 Score ≥ 5: the patient is at nutritional risk. Please refer to a dietician.
433

434 TABLES

14
 ACCEPTED MANUSCRIPT

435 Table 1. Patient characteristics.

Variables N (% or ±SD)

Male 217 (55.1)

Female 177 (44.9)

Age (years) 61.55 ± 12.09

PT
Height (cm) 164.3 ± 9.6

Weight (kg) 71.4 ± 13.9

RI
BMI (kg/m2) 26.3 ± 4.87

Site (diagnosis) N (%)

SC
Abdominal and pelvic: liver, biliary tract, renal, 74 (18.8)

gynaecological

U
Head & neck 49 (12.4)

Colorectal 38 (9.6)
AN
Lymphomas that compromised GI tract 3 (0.8)

Leukaemia and other lymphomas 46 (11.7)

M

Breast 57 (14.5)

Prostate 31 (7.9)
D

Lung 41 (10.4)
TE

CNS (Central nervous system) 15 (3.8)

Upper GI Tract (oesophagus, gastric, pancreas, 39 (9.9)

intestinal)
EP

Others 1 (0.3)

Treatment N (%)
C

Only chemotherapy 142 (36.0)

AC

Only radiation therapy 74 (18.8)

Concomitant chemo radiotherapy 37 (9.4)

Haematopoietic stem cell transplantation 28 (7.1)

113 (28.7)
Other treatments, or exclusively symptomatic
treatment
436

437 Table 2. Risk of malnutrition according to tool, to tumour site and to treatment.

15
 ACCEPTED MANUSCRIPT

438
439
NUTRISCOREa,b MSTa,b PG-SGA* a,b

n % (CI 95%) n % (CI 95%) n % (CI 95%)

All patients at risk of
89 22.6 (18.5-26.7) 111 28.2 (23.7-32.6) 75 19.0 (15.2-22.9)
malnutrition
According to tumour site

PT
Site associated with low 18 12.0 (6.8-17.2) 33 22.0 (15.3-28.7) 18 12.0 (6.8-17.2)
nutritional risk
Site associated with 26 17.0 (11-23) 32 20.9 (14.4-27.4) 26 17.0 (11-23)
moderate nutritional risk

RI
Site associated with high 45 49.5 (39.1-59.8) 46 50.5 (40.2-60.9) 31 34.1 (24.3-43.9)
nutritional risk
According to treatment

SC
Chemotherapy 26 18.3 (11.9-24.7) 39 27.5 (20.1-34.8) 27 19.0 (12.5-25.5)

Radiation therapy 12 16.4 (7.9-25) 9 12.3 (4.7-19.9) 5 6.8 (1.0-12.7)

U
Concomitant chemo
24 64.9 (49.3-80.5) 23 62.2 (46.3-78) 16 43.2 (27.1-59.4)
radiotherapy
AN
Haematopoietic stem cell
15 53.6 (34.8-72.4) 24 85.7 (72.5-98.9) 15 53.6 (34.8-72.4)
transplantation
Other treatments or
12 10.5 (4.9-16.2) 16 14.0 (7.6-20.4) 12 10.6 (4.9-16.2)
symptomatic treatment
M

440 * PG-SGA stages at risk: B+C.

441 a: significant differences according to site (MST: χ2=31.407, p<0.001; PG-SGA: χ2=21.549, p<0.001;
442 NUTRISCORE:χ2=49.908, p<0.001).
443 b: significant differences according to treatment (MST: χ2=91.219, p<0.001; PG-SGA: χ2=48.170, p<0.001;
D

444 NUTRISCORE:χ2=59.220, p<0.001).

TE
C EP
AC

16
 ACCEPTED MANUSCRIPT

445 Table 3. Validity of the NUTRISCORE.

446
447

PT
PG-SGA

Malnourished Well-nourished Total Sensitivity Specificity PPV NPV AUC*

n n n % (CI 95%) % (CI 95%) %(CI 95%) %(CI 95%)

RI
At risk 73 13 86
97.3 (91- 84.8 (76- 0.95 (CI
NUTRISCORE 95.9 (93-98) 99 (98-100)
100) 92) 0.92-0.98)

SC
Without risk 2 306 308

At risk 63 46 109
57.7 (48- 0.84 (CI
MST 84 (74-91) 85.6 (81-89) 95.7 (93-98)

U
12 273 285 67) 0.79-0.89)
Without risk

AN
448 PPV=Positive Predictive Value
449 NPV=Negative Predictive Value
450 AUC=Area under the curve

M
451
452 *The two AUC are significantly different with p value =0.001754
453

D
454
455

TE
456 EP
C
AC

17