Original article

te

rne b a

aQueensland Health, Logan Hospital, Nutrition & DietetbQueensland Health, Beaudesert Hospital, Nutrition & D

a r t i c l e i n f o

Article history:Received 4 December 2013Accepted 6 February 2014

Keywords:

. All rights reserved.

ia and hypo-ulmonary, cardiac,ations.1 Effectiveutrition by dieteticreatment of these

length of stay.RFS studies currently focus on adolescents with eating disor-

ders, and patients receiving enteral or parenteral refeeding, with55% of patients (n 133/243) on enteral or parenteral feedingregimens having one or more risk factors for RFS but only threedeveloping severe conditions associated with RFS.3 No study hasexamined identication rates of RFS risk of patients on acutemedical and surgical wards, particularly focussing on RFS

Abbreviations: RFS, refeeding syndrome.q Conference presentation: Select results from this study are planned to be

presented at the Dietitians Association of Australia conference in May 2014 inBrisbane, Queensland, Australia.* Corresponding author. Nutrition and Dietetics Department, Logan Hospital,

Queensland Health, Meadowbrook, QLD 4131, Australia. Tel.: 61 7 3299 8253;fax: 61 7 3299 8280.

Contents lists availab

Clinical N

journal homepage: ht tp: / /ww

Clinical Nutrition xxx (2014) 1e6E-mail address: emma.owers@health.qld.gov.au (E.L. Owers).symptoms, thereby reducing complication rates and hospital2tween dietitians requires determination. 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism

1. Introduction

Refeeding Syndrome (RFS) can occur when re-commencingnutrition provision to patients who have been starved or severelymalnourished.1 Fluid and electrolyte imbalances can result

including hypophosphataemia, hypokalaemmagnesaemia, with potential for neurological, pneuromuscular, and haematological complicidentication and management of RFS and malnand medical staff is vital for prevention and tcation of at-risk patients was in accordance with RFS guidelines. Electrolyte supplementation waspositively associated with electrolyte levels post-assessment. Consistency of RFS-risk identication be-Refeeding syndromeAdultHospitalMalnutritionIdenticationElectrolyteshttp://dx.doi.org/10.1016/j.clnu.2014.02.0030261-5614/ 2014 Elsevier Ltd and European Society

Please cite this article in press as: Owers ELClinical Nutrition (2014), http://dx.doi.org/1ics, Meadowbrook, QLD 4131, Australiaietetics, Beaudesert, QLD 4285, Australia

s u m m a r y

Background & aims: Identication of Refeeding Syndrome (RFS) is vital for prevention and treatment ofmetabolic disturbances, yet no information exists that describes identication rates by dietitians in acutecare. We aimed to describe rates and demographics of inpatients identied by dietitians as at-risk of RFSand factors associated with electrolyte levels post-dietetic assessment.Methods: Eligible participants were adult (18 yrs) acute care inpatients reviewed by dietitians betweenMarch 2012eFebruary 2013 and not admitted to intensive care prior to rst dietetic assessment. Patientinformation was sourced from medical charts. Chi-squared, t-tests and linear regression analyses wereconducted.Results: Of 1661 eligible inpatients (55%F, 65 18yrs), 9% (n 151) were documented as at-risk of RFS inthe rst dietetic medical chart entry. On average, patients identied with RFS-risk had four days greaterhospital stay, were 13 kg lighter, more likely classied SGA C (36% vs. 7%), and on a modied diet (52% vs.35%) than non-RFS patients (p < 0.05). Very low and low electrolyte values occurred within seven dayspost-dietetic assessment in 7% and 52%, respectively, of inpatients with RFS-risk. Regression analysisshowed that electrolyte supplementation was positively associated (b 0.145e0.594), and number ofRFS-related risk factors negatively associated (b 0.044e0.122), with potassium, magnesium andphosphate levels within seven days post-dietetic assessment (p < 0.05).Conclusion: Nine percent of adult inpatients were documented as at-risk of RFS by dietitians. Identi-Sally A. Noble , Helen E. PoRachel A. Roth a, Clare E. By , Michelle A. PalmerEmma L. Owers , Anneli I. Reeves , Susan Y. Ko , Aleshia K. Ellis , Shannon L. Huxtable ,a rteous a, Eli J. Newman a, Christine A. Josephson a,Rates of adult acute inpatients documensyndrome by dietitiansq

a,* a afor Clinical Nutrition and Metabol

, et al., Rates of adult acute in0.1016/j.clnu.2014.02.003d as at risk of refeeding

a a

le at ScienceDirect

utrition

w.elsevier .com/locate/c lnuism. All rights reserved.

patients documented as at risk of refeeding syndrome by dietitians,

al Nuassociated with oral intake. The lack of consensus on identifyingRFS in RFS guidelines4e13 likely contributes to the lack of evidenceon rates of RFS risk in inpatients. Internationally, hospital malnu-trition rates vary between 19 and 48%,14 indicating a high likelihoodfor inpatients being at risk of RFS. As such, we aimed to describerates and demographics of inpatients identied by dietitians as at-risk of RFS and factors associated with electrolyte levels post-dietetic assessment.

2. Materials and methods

This retrospective study investigated rates and outcomes ofmedical and surgical patients reviewed and identied by dietitiansas at risk of RFS in a 350-bed acute care hospital. This study wasexempted from ethical review (HREC/13/QPAH/415).

2.1. Patient eligibility criteria

Eligible subjects were acute adult (18 years) patients admittedto medical and surgical wards who were rst reviewed by a dieti-tian between March 2012 and February 2013. Patients wereexcluded if, prior to the rst dietetic assessment, they were: forpalliative care; admitted to the intensive care unit; or aged

al NuE.L. Owers et al. / Clinichypophosphataemia were more likely to be on an at-risk diet codeprior to dietetic assessment (p < 0.001, Table 2). Patients who hadlow potassium levels were more likely to be receiving thiaminesupplementation (p < 0.05). Patients with low magnesium levelsand normal serum phosphate levels were more likely to havesubstance abuse issues documented (p < 0.001). Patients with lowmagnesium levels had 4 days greater length of stay and a greaternumber of RFS risk factors (p < 0.05). Patients who developed lowlevels of potassium, magnesium or phosphate were more likely tohave received the relevant nutrient supplementation (p < 0.001).

Regression analysis showed, however, that the lowestbiochemical level within seven days post-rst dietetic assessmentwas positively associated with the relevant nutrient being supple-mented (p < 0.01), and potassium and magnesium levels werenegatively associated with number of RFS risk factors (p < 0.05,Table 3). Lower phosphate levels were associated with a patientrst being assessed by a dietitian on a Friday compared to any otherweekday and the presence of substance abuse (p < 0.01). Othervariables, including caloric restriction, were not signicantly asso-ciated (p > 0.05).

Fig. 1. Flowchart of eligible an

Please cite this article in press as: Owers EL, et al., Rates of adult acute inClinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.02.003trition xxx (2014) 1e6 34. Discussion

Our retrospective study is the rst internationally to describeidentication of RFS risk amongst medical and surgical inpatientsby dietitians.

Nine percent of eligible patients admitted to acute medical orsurgical wards were documented with RFS risk by dietitians on rstdietetic assessment. Yet Rio et al.3 found that 55% (n 133/243) ofadults on articial nutrition support in wards and intensive carehad one or more risk factors associated with RFS. Wagstaff18 alsosurveyed dietitians in NHS Trusts in London who reported thatpatients most likely at risk of RFS were those seen for enteralnutrition support (141 responses), oral nutrition support (92 re-sponses) and then parenteral nutrition support (67 responses). Ourstudy may have a lower prevalence of at-risk patients as weexcluded those admitted to ICU rst, and only 12% were receivingenteral or parenteral nutrition. Our ratemay also be under-stated aspatients may also develop RFS risk later in their hospital stay. Themajority of patients were seen in the rst 1e5 days of admission.Yet patient lengths of stay were up to 143 days and hospital-

d ineligible participants.

patients documented as at risk of refeeding syndrome by dietitians,

al NuTable 1Demographics of eligible inpatients reviewed during the rst dietetic consultationon medical and surgical wards.

Descriptive variable At riskof RFS

NOT atrisk of RFS

(n 151) (n 1510)n(%) n(%)

DemographicsGender, female 82(54) 823(55)Age, years, mean SD 63 19 65 18Place of residence, nursing home 12(8) 104(7)

AnthropometryWeight, kg, mean SD 63 20*** 76 25***BMI, kg/m2, mean SD 22 6*** 27 8***Percent ideal body weight,mean SD

89 26*** 109 33***

Nutritional statusSGA A 54*** 461(46)***

E.L. Owers et al. / Clinic4acquiredmalnutrition can occur in 19e48% of patients.14Wagstaff18

showed that variation in classifying the level of RFS risk can occurwhen different dietitians examine the same case study. RFSguidelines also vary in identifying at-risk patients.4e13 Whileseveral dietitians worked on the acute care wards, our workforce isexperienced (5 years of clinical dietetic experience). Our currentstudy was also conducted in a real clinical setting, so our results arelikely generalisable to many acute care facilities. While the retro-spective nature of our study is a limitation, it is also a strength asthere was no bias placed on dietitians when identifying risk of RFS.Future studies could examine how consistently dietitians workingin acute care wards identify patients as at risk of RFS.

Dietitians may be identifying RFS risk appropriately in acutecare settings. At-risk patients identied in our study had signi-cantly lower weight, BMI, percentage of ideal body weight, weremore likely severely malnourished (SGA C), on a diet code oftenassociated with poor intake,16 and more likely admitted for meta-bolic, mental health issues, pancreatic or hepatic conditions, which

SGA B 78(61) 479(48)SGA C 46(36)*** 68(7)***

Admission detailsDiagnosis-related group, n(%)Hepatobiliary system & pancreas 18(12)*** 72(5)Skin, subcutaneous tissue & breast 0(0)*** 60(4)Endocrine, nutritional & metabolism 21(14)*** 104(7)Nervous system 13(9) 223(15)Respiratory system 18(12) 287(19)Circulatory system 8(5) 128(9)Digestive system 25(17) 238(16)Musculoskeletal & connective tissue 14(9) 109(7)Kidney and urinary tract 10(7) 75(5)Mental health 5(3)*** 11(1)Medically complex, other

factors inuencing health5(3) 72(5)

Other 14(9) 131(9)Initial at-risk diet code, (nil bymouth, uids, texture modied,high protein high energy)

78(52)*** 369(35)***

Number of days from admissionuntil rst seen bydietitian, median(range)

3(1e20) 3(1e38)

Length of stay, days, median(range) 11(2e143)*** 7(1e82)***Number of times reviewedby dietitian, median(range)

3(1e30)*** 1(1e21)***

Number of times reviewed bydietitian as a proportion oflength of stay, median(range)

0.35(0.06e1)*** 0.23(0.01e1)***

Patient rst seen by thedietitian on a Friday

44(29)** 298(20)**

Days between admissiondate & date rst seen bydietitians, median (range)

3(1e20) 3(1e38)

*p< 0.05 between groups, **p< 0.01 between groups, ***p< 0.001 between groups.

Please cite this article in press as: Owers EL, et al., Rates of adult acute inClinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.02.003could include substance abuse or malabsorption, than those notidentied as at risk. These descriptors are well known for beingassociated with the development of RFS.4e13 The increased likeli-hood of patients being assessed as at risk of RFS on Fridays may bedue to dietitians undertaking conservative identication andmanagement of RFS given the absence of a dietetic workforce onweekends. Yet our regression analysis showed that at-risk patientsrst assessed by a dietitian on a Friday had associated lowerphosphate levels, suggesting that our identication process may beappropriate. Furthermore, at least half (52%) of the patients iden-tied as at risk of RFS developed low levels of potassium, magne-sium or phosphate and up to 43% of patients who had normal levelshad received nutrient supplementation which may have preventedRFS from developing. Dietitians may therefore be appropriatelyidentifying patients at risk of RFS in acute care settings andcontributing towards the effective management of these patients.

Nutrient supplementation was positively associated with po-tassium, magnesium and phosphate levels in the seven days afterthe rst dietetic assessment. Caloric intake restriction was notsignicantly associated with these levels. OConnor et al.19 arguesthat increased calorie provision does not cause lower biochemicalvalues. Caloric restriction recommendations for at-risk patients areoften linked to the provision of enteral or parenteral feeding20

rather than oral feeding. Intake is often easier to measure withenteral or parenteral nutrition, rather than via oral intake, partic-ularly during retrospective chart audits. The lack of associationobserved between calorie restriction and biochemical levels mighttherefore be due to most (88%) of our patients receiving oral regi-mens and our inability to determine whether they were consumingall of their prescribed hypo- or normo-caloric regimens. Our studywas also retrospective and therefore not able to establish causationbetween variables and outcomes. Several risk factors of RFS,including substance abuse, were signicantly associatedwith lowerbiochemical levels post-dietetic assessment. This is supported byRFS guidelines.4e13 Up to 18% of biochemical data was missing, andbiochemical tests were often not done daily in the seven days post-dietetic assessment. While this could also have inuenced theregression results, and better patient care can occur when resultsare available, limited data availability is consistent with usualpractise. Similarly, a systematic literature review reported that 52%(n 14/27) of published RFS cases located between 2000 and 2011had not checked all three levels of potassium, magnesium andphosphate.21 Other factors unrelated to RFS might also explainlowered potassium, magnesium and phosphate levels. Hypo-phosphataemia may occur in neoplastic disease and diabetesmellitus, and is attributable to certain medications such as insulin,epinephrine and dopamine.21 We did not attempt to collect allpossible biochemical confounders. Some relevant RFS-related riskfactors for each patient might have been missed. This may alsoexplain the limited number of variables that were signicantlyassociated with biochemical levels. Our department conductedchart audits annually to ensure high standards for chart docu-mentation but, despite best efforts, variation in data collectionbetween dietitians could still occur. The inability to obtain all in-formation from patients is also challenging, but consistent with ausual clinical environment.

This study had several additional strengths and weaknesses. Wecollected data for one year and reviewed 2222 admissions to obtainan adequate representation of acute care patients seen by di-etitians. While we had a large number of data collectors, everyeffort was made to ensure that data was entered consistently,including regular training, and data checks by the project lead. Thelist of all patients seen by dietitians in 2012e2013 may have beenunderstated due to incomplete recording in PI5. However, regular

trition xxx (2014) 1e6accuracy checks on PI5 are done by management to mitigate this

patients documented as at risk of refeeding syndrome by dietitians,

f pat

/L)a Serum magnesium (mmol/L) Serum phosphate (mmol/L)

15% loss of body weight 27(29) 8(16)

Number of reasons for poorintake, median(range)

1(1e3) 1(1e3)

GIT, gastro-intestinal tract; RFS, Refeeding syndrome, *p < 0.05 between groups, **p < 0a n 8 potassium values not tested.b n 26 magnesium values not tested.c n 24 phosphate values not tested.

Table 3Multinomial stepwise regression analyses assessing factors associated with variation ininitial dietetic assessment.

Variable Potassium (mmol/L)

(n 141)b coefcient

Total number of risk factors for RFS 0.122Relevant nutrient was supplemented, yes 0.594Patient rst seen by dietitian on a Friday, yes NSRFS risk factor e substance abuse (drugs or alcohol), yes NS

NS, not signicant (p > 0.05); RFS, refeeding syndrome; Additional variables included idietitian during admission, number of days between date of admission and date rst seengender, patient lived in a nursing home, additional risk factors for RFS (including cognitweight, poor intake, gastrointestinal and malabsorption, social/nancial issues), risk factodentition, substance abuse, mental health or cognitive issues, increased requirements omentation with thiamine, supplementation with multivitamin, and whether on intra-ve

Please cite this article in press as: Owers EL, et al., Rates of adult acute inClinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.02.00315(19)*** 34(76)*** 12(16)*** 27(54)***54(68) 35(78) 57(74) 33(66)57(71) 38(84) 61(79) 35(70)10(13) 4(9) 6(8) 8(16)

2(1e3)** 2(1e4)** 2(1e4) 2(1e3)

1(1) 3(7) 2(3) 2(4)) Normal (0.7) Low (15% loss of body weight,

References

1. Mehanna HM, Moledina J, Travis J. Refeeding syndrome: what it is, and how toprevent and treat it. BMJ 2008;336(7659):1495e8.

2. Correia MI, Waitzberg DL. The impact of malnutrition on morbidity, mortality,length of hospital stay and costs evaluated through a multivariate modelanalysis. Clin Nutr 2003;22(3):235e9.

3. Rio A, Whelan K, Goff L, Reidlinger DP, Smeeton N. Occurrence of refeedingsyndrome in adults started on articial nutrition support: prospective cohortstudy. BMJ Open 2013;3. http://dx.doi.org/10.1136/bmjopen-2012e002173.

4. Wakeeld A, Williams H. Practice recommendations for the nutritional man-agement of anorexia nervosa in adults. Dietitians Association of Australia; 2009.http://dmsweb.daa.asn.au/les/DINER/Anorexia%20Nervosa_Final.pdf[accessed November 2013].

5. National Collaborating Centre for Mental Health. Eating disorders: core in-terventions in the treatment and management of anorexia nervosa, Bulimianervosa and related eating disorders. Available at: The British PsychologicalSociety; 2004. Available from: http://www.nice.org.uk http://www.nice.org.uk/nicemedia/pdf/cg9fullguideline.pdf.

6. MARSIPAN Working Group e Royal College of Psychiatrists and Royal Collegeof Physicians London. MARSIPAN: Management of Really Sick Patients withAnorexia nervosa. 2010 [Updated October 20102013], http:www.rcpsych.ac.uk/publications/collegereports/cr/cr162.aspx.

7. Beumont P, Hay P, Beumont D, Birmingham L, Derham H, Jordan A, et al.Australian and New Zealand clinical practice guidelines for the treatment ofanorexia nervosa. Aust N Z J Psychiatry 2004;38(9):659e70.

8. Stanga Z, Brunner A, Leuenberger M, Grimble RF, Shenkin A, Allison SP, et al.Nutrition in clinical practice e the refeeding syndrome: illustrative cases andguidelines for prevention and treatment. Eur J Clin Nutr 2008;62(6):687e94.

9. National Institute for Health and Clinical Excellence. Nutrition support in adults:oral nutrition support, enteral tube feeding and parenteral nutrition. London:National Collaborating Centre for Acute Care at The Royal College of Surgeonsof England; 2006. Available from: http://www.nice.org.uk/nicemedia/live/10978/29981/29981.pdf.

10. Boland K, Solanki D, OHanlon C. IrSPEN guidelines document 1: prevention andtreatment of refeeding syndrome in the acute care setting: Irish society for clinicalnutritrion & metabolism [updated November 2013]; [accessed November 2013],http://www.irspen.ie/wp-content/uploads/2013/11/Refeeding-Guideline1.pdf;2013.

11. Dewar H, Horvath R. Refeeding syndrome e a pocket guide to clinical nutrition.2nd ed. British Dietetic Assocation; 2004. editor.

12. Kraft MD, Btaiche IF, Sacks GS. Review of the refeeding syndrome. Nutr ClinPract 2005;20(6):625e33.

13. McCray S, Walker S, Parrish CR. Much ado about refeeding. Pract Gastroenterol2004;28(12):26e44.

14. Agarwal E, Ferguson M, Banks M, Bauer J, Capra S, Isenring E. Nutritional statusand dietary intake of acute care patients: results from the Nutrition Care DaySurvey 2010. Clin Nutr 2012;31(1):41e7.

15. Detsky A, McLaughlin JR, Baker JP, Johnston N, Whittaker S, Mendelson RA,et al. What is subjective global assessment of nutritional status? JPEN1987;11(1):8e13.

16. Huxtable S, Palmer M. The efcacy of protected mealtimes in reducing meal-time interruptions and improving mealtime assistance in adult inpatients in anAustralian hospital. Eur J Clin Nutr 2013;67:904e10.

17. Australian Health Services Research Institute. Australian rened diagnosisrelated groups version 7.0 denitions manual. Wollongong: National Casemixand Classication Centre; 2013. Available from: http://nccc.uow.edu.au/ardrg/denitionsmanuals/index.html.

18. Wagstaff G. Dietetic practice in refeeding syndrome. J Hum Nutr Diet2011;24(5):505e15.

19. OConnor G, Goldin J. The refeeding syndrome and glucose load. Int J Eat Disord2011;44(2):182e5.

20. OConnor G, Nicholls D. Refeeding hypophosphatemia in adolescents withanorexia nervosa: a systematic review. Nutr Clin Pract 2013;28(3):358e64.

21. Skipper A. Refeeding syndrome or refeeding hypophosphatemia: a systematicreview of cases. Nutr Clin Pract 2012;27(1):34e40.

E.L. Owers et al. / Clinical Nutrition xxx (2014) 1e66Please cite this article in press as: Owers EL, et al., Rates of adult acute inClinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.02.003patients documented as at risk of refeeding syndrome by dietitians,

Rates of adult acute inpatients documented as at risk of refeeding syndrome by dietitians1 Introduction2 Materials and methods2.1 Patient eligibility criteria2.2 Data collection2.3 Data analysis

3 Results4 Discussion5 ConclusionStatement of authorshipConflict of interest statement and funding sourcesAcknowledgementsReferences