Filetype PDF | Posted on 09 Jan 2023 | 4 years ago
Authentication
digital resources
148x Filetype PDF File size 0.60 MB Source: 61.8.75.226
Clinical Nutrition 38 (2019) 883e890
Contents lists available at ScienceDirect
Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu
Original article
Timing of PROTein INtake and clinical outcomes of adult critically ill
patients on prolonged mechanical VENTilation: The PROTINVENT
retrospective study
a,1 a,1 a
W.A.C. (Kristine) Koekkoek , C.H. (Coralien) van Setten , Laura E. Olthof ,
b a, *
J.C.N. (Hans) Kars , Arthur R.H. van Zanten
a Department of Intensive Care Medicine, Gelderse Vallei Hospital, Willy Brandtlaan 10, 6716 RP, Ede, The Netherlands
b Department of Information Technology and Datawarehouse, Gelderse Vallei Hospital, Willy Brandtlaan 10, 6716 RP, Ede, The Netherlands
articleinfo summary
Article history: Background & aims: Optimal protein intake during critical illness is unknown. Conflicting results on
Received 16 March 2017 nutritional support during the first week of ICU stay have been published. We addressed timing of
Accepted 11 February 2018 protein intake and outcomes in ICU patients requiring prolonged mechanical ventilation.
Methods: Weretrospectivelycollectednutritional and clinical data on the first 7 days of ICU admission of
Keywords: adult critically ill patients, who were mechanically ventilated in our ICU for at least 7 days and admitted
Critical care nutrition betweenJanuary1st2011andDecember31st2015.Basedonrecentliterature,patientsweredividedinto
Protein 3 protein intake categories, <0.8 g/kg/day, 0.8e1.2 g/kg/day and >1.2 g/kg/day. Our primary aim was to
Energy identify the optimum protein dose and timing related to the lowest 6 month mortality. Secondary
Non-nutritional calories endpoints were ventilation duration, need for renal replacement therapy (RRT), ICU length of stay (LOS)
Autophagy
Mortality and mortality and hospital LOS and mortality.
Results: In total 455 patients met the inclusion criteria. We found a time-dependent association of
protein intake and mortality; low protein intake (<0.8 g/kg/day) before day 3 and high protein intake
(>0.8 g/kg/day) after day 3 was associated with lower 6-month mortality, adjusted HR 0.609; 95% CI
0.480e0.772, p < 0.001) compared to patients with overall high protein intake. Lowest 6-month mor-
tality was found when increasing protein intake from <0.8 g/kg/day on day 1e2 to 0.8e1.2 g/kg/day on
day 3e5 and >1.2 g/kg/day after day 5. Moreover, overall low protein intake was associated with the
highest ICU, in-hospital and 6-month mortality. No differences in ICU LOS, need for RRT or ventilation
duration were found.
Conclusions: Our data suggest that although overall low protein intake is associated with the highest
mortality risk, high protein intake during the first 3e5 days of ICU stay is also associated with increased
long-term mortality. Therefore, timing of high protein intake may be relevant for optimizing ICU, in-
hospital and long-term mortality outcomes.
©2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction provision [2e4], however there is growing evidence that protein
intake may be more important than caloric intake [5e7].Asfixed
Nutritional support during critical illness is heavily debated [1]. protein to energy ratios in most feeding regimens are used, it is
Many studies have evaluated effects of nutritional support on complex to separate effects of protein intake from those of energy
clinical outcomes in ICU. Most studies have focused on energy intake. Furthermore, in several studies both energy and protein
intake were similarly associated with clinical outcomes in univar-
iateanalyses[8].Otherstudiesshowedthathighproteinintakewas
* Corresponding author. Fax: þ31 318 43 41 16. associated with reduced mortality risk [9], whereas energy over-
E-mail addresses: koekkoekk@zgv.nl (W.A.C. Koekkoek), ch.vinkvansetten@ feeding was associated with increased mortality risk [10].Lower
online.nl (C.H. van Setten), ao.laura@gmail.com (L.E. Olthof), karsh@zgv.nl mortality and more ventilator free days were reported in patients
(J.C.N. Kars), zantena@zgv.nl (A.R.H. van Zanten). with sepsis or severe pneumonia reaching higher protein and
1 Shared first authorship.
https://doi.org/10.1016/j.clnu.2018.02.012
0261-5614/©2018TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).
884 W.A.C. Koekkoek et al. / Clinical Nutrition 38 (2019) 883e890
caloric intake in the early phase of ICU stay [11]. This might even be administered non-nutritional calories (dextrose infusion, propofol
more relevant for patients with Body Mass Index (BMI)<25 or and trisodium citrate) [14].
2 DatatocalculatetheCharlsonComorbidityIndex(CCI)[19]were
>35 kg*m [12]. Recently, a retrospective analysis of energy pro-
vision during the first week of ICU stay in 475 patients with pro- obtained from the quality management system for hospital mor-
longedmechanicalventilationshowedbeneficialeffectsfromearly tality registration. All deaths in the Netherlands are registered in
full energy feeding on mortality and quality of life 3 months post the municipal personal records database of the Dutch government.
ICU discharge [13]. In this study protein intake was not studied Asourelectronicpatientmanagementsystemisdirectlyconnected
separately. From the non-nutritional calories (e.g. dextrose, citrate to this database date of death could be extracted. When date of
and propofol infusions) only propofol infusions were taken into death was not registered the patient was presumed alive. Days
account,althoughnon-nutritionalcaloriesmaycontributeforupto weredefinedascalendar days.
20% of total caloric intake in individual patients [14]. Moreover, in
thatstudycumulativecaloricintakeoveroneweekwasstudiedand 2.3. Nutritional parameters
daily effects of intake were not assessed.
Casaer and co-workers, based on a post hoc analysis of the Wecollecteddataonnutritionalintakeforthefirst7daysofICU
EPANIC randomized trial, suggested a time-dependent association admission,includingproteinandenergytargets,actualgivendoses
of protein intake and clinical outcome, with possible harmful ef- of proteins (g) and calories (kcal) from enteral (EN) and parenteral
fects of protein intake during the first 3 days of ICU admission [15]. nutrition(PN).Additionallynon-nutritionalcaloriesfromtrisodium
In order to achieve a personalized nutritional approach several citrate, glucose and propofol infusions were calculated and added
questions need to be answered [16,17]. Therefore, we addressed to calculate total caloric intake [14]. We divided total caloric intake
how protein intake during the first week of ICU admission in- into adequacy categories based on recent literature [6,10] (3
fluences clinical outcomes among prolonged mechanically venti- groups:hypocaloric:<80%ofenergytarget,normocaloric:80e110%
lated critically ill patients. of energy target and hypercaloric: more than 110% of energy
Our primary aim was to determine the best timing and dose of target).
protein intake to support the lowest 6-month mortality. Secondary
outcome measures were the effect of timing and dose of protein 2.4. Calculation of nutritional goals
intake on ICU and hospital length of stay (LOS), ICU and hospital
mortality, ventilation duration and need for renal replacement In all patients body weight and height were measured on ICU
therapy (RRT). admission. The World Health Organization/Food and Agricultural
2. Materials and methods OrganizationoftheUnitedNations(WHO/FAO)formulaswereused
tocalculatecaloricandproteintargetsbyourcomputerizedfeeding
For this single center cohort study we retrospectively collected protocol [14]. According to BMI, the actual, corrected (weight on
data from patients fulfilling inclusion criteria, who were admitted BMI 27) or ideal body weight (women weight on BMI 21, men
to our ICU between January 1st 2011 and December 31st 2015. In- weight on BMI 22.5) was used. An addition to resting energy
clusion criteria were: adult critically ill patients (18 years), expenditure (REE) of 20% was used to correct for disease activity.
requiring invasive mechanical ventilation for a minimum duration Ourtargetproteinintakewas1.5gperkilogrambodyweightper
day(g*kg1 1 2
of7days.Patientswereexcludedifthetimefromadmissiontostart *day )forpatientswithBMIsupto27kg*m .Incase
2 2
of mechanical ventilation exceeded 48 h, if data on nutritional of BMI 27e30 kg*m , weight was corrected to BMI 27 kg*m .In
2
needs were incomplete, in case of contraindications to full nutri- case of BMI >30 kg*m we used ideal body weight and protein
1 1
tion, if their condition influenced their nutritional needs in a way administrationwassetto2.0g*kg *day ,whereaspatientswitha
that we were unable to estimate or compare results with other BMI>40kg*m2prescriptionwas2.5gperkgidealweightperday
patients, such as pregnancy, preexistent neuromuscular diseases, according to international guidelines [20].
known protein malabsorption or metabolic abnormalities. In pa-
tients with multiple ICU admissions during the study period, to 2.5. Protein categories
avert bias we excluded data from ICU readmissions. An ICU
admission was considered a readmission when the patient was We used protein targets in grams per kilogram uncorrected
admitted within 6 months of the primary ICU admission. body weight on ICU admission to divide patients into categories
according to their mean protein intake during the first week of ICU
2.1. Ethical approval admission. The chosen cut-off values are based on recent literature
1 1 1 1
[10]; protein intake less than 0.8 g*kg *day ,0.8g*kg *day to
1 1 1 1
The institutional review board of Gelderse Vallei Hospital 1.2 g*kg *day and more than 1.2 g*kg *day .
approvedthestudyandwaivedinformedconsentforreasonsofthe
retrospective design and anonymization of patient identifiers 2.6. Study end points
before analysis.
Our primary endpoint was the association of 7-days protein
2.2. Data collection intake and 6-months survival. We considered this to be the most
appropriate time window, because the effects of protein provision
Data extraction was performed using SAS Enterprise Guide maynotbeexpectedwithinashorttimeframeandpreviousstudies
queries (version 7.12HF1), from our MetaVision (Patient Data on critical care nutrition feeding interventions show effects on
Management System, iMDsoft, Tel Aviv, Israel) database and other long-term but not early mortality endpoints [21]. Moreover, long-
hospital electronic patient records. Baseline characteristics were term outcomes are clinically very important for patient prognosis
listed; age, gender, primaryadmission diagnosis, baseline APACHE- and recovery. Secondary endpoints included ICU and in hospital
II and SOFA-scores, several baseline blood tests, admission type mortality, ICU and hospital LOS, ventilation duration, need for and
(medical, elective and non-elective surgery), comorbidities, modi- durationofRRTandallcausehospitalreadmissionwithin6months
fied Nutrition Risk in Critically ill (mNUTRIC) score [18] and from ICU admission.
W.A.C. Koekkoek et al. / Clinical Nutrition 38 (2019) 883e890 885
2.7. Data analysis 3. Results
Descriptive data are reported as means and standard deviation 3.1. Patients
(SD) or median and interquartile range (IQR) in case of skewed
distributions, or as frequencies and percentages or ranges Duringthestudyperiod2237patientswereadmittedtoourICU,
(minimumemaximum). of which 546 were considered eligible for inclusion. We excluded
91 patients; reasons were delayed intubation (N ¼ 59), ICU
admission within the six months previous to the selected admis-
2.8. Statistical analysis sion (N ¼ 25) and insufficient data on nutritional intake due to
Baseline characteristic differences and secondary endpoints participation in a blinded tube feeds study (N ¼ 7, Fig. 1). In total,
455 individual patients were enrolled in our study, of which four
were assessed with Chi square tests or Fisher's exact tests and were enrolled twice.
ANOVA or KruskaleWallis tests where appropriate. Six-month Baseline characteristics and feeding parameters are shown in
survival was assessed by Kaplan Meier survival estimate curves Tables1and2.Significantdifferenceswereobservedbetweenthe3
and Cox Proportional Hazards Models. A P-value <0.05 was protein intake subgroups for BMI, SOFA-score, admission type,
considered statistically significant. For univariate analysis all vari- hours to start feeding, route of feeding, daily protein target, total
ables considered to be relevant based on literature were included. protein and caloric intake, adequacy of protein and caloric intake
For the primary outcome measure, when univariate analysis and percentage of non-nutritional calories.
revealed p < 0.10 multivariate analysis was performed. Multi-
collinearity of variables included into multivariate analyses was
assessed by calculation of the variance inflation factor (VIF), we 3.2. Primary outcome
consideredaVIFabove2asanindicatorofrelevantcollinearity.IBM
SPSSStatisticsforWindows,version24.0(IBMCorporation,released The 6-months survival was 65.6%, 68.9% and 55.6% in the low
1 1 1 1
2014, Armonk, NewYork, USA) was used toperformanalyses. (0.8 g*kg *day ), intermediate (0.8e1.2 g*kg *day ) and high
Fig. 1. Flow chart of the study population.
886 W.A.C. Koekkoek et al. / Clinical Nutrition 38 (2019) 883e890
Table 1
Baseline characteristics.
Total population Protein intake categories p-valuea
Protein intake in g*kg1*day1 LOW INTERMEDIATE HIGH
<0.8 0.8e1.2 >1.2
N(%) 455 (100) 128 (28.1) 264 (58.0) 63 (13.8)
Females N (%) 170 (37.4) 47 (36.7) 98 (37.1) 25 (39.7) 0.933
Age, median [IQR] 70 [61e77] 68 [60e77] 70 [61e76] 70 [61e79] 0.633
2
BMI,kg*m , median [IQR] 26.4 [23.5e30.0] 28.4 [24.7e32.9] 26.2 [23.6e29.4] 24.6 [21.4e26.6] <0.001
BMIcategories, N (%) <0.001
<18.5 16 (3.5) 4 (3.1) 10 (3.8) 2 (3.2)
18.5e25 157 (34.5) 31 (24.2) 93 (35.2) 33 (52.4)
25e35 234 (51.4) 68 (53.1) 140 (53.0) 26 (41.3)
>35 48 (10.5) 25 (19.5) 21 (8.0) 2 (3.2)
ICU admission year, N (%) 0.818
2011 105 (23.1) 35 (27.3) 59 (22.3) 11 (17.5)
2012 84 (18.5) 20 (15.6) 51 (19.3) 13 (20.6)
2013 81 (17.8) 21 (16.4) 48 (18.2) 12 (19.0)
2014 91 (20.0) 25 (19.5) 50 (18.9) 16 (25.4)
2015 94 (20.7) 27 (21.1) 56 (21.2) 11 (17.5)
APACHEII score, median [IQR] n ¼ 433 22 [18e28] 24 [19e29] 22 [18e27.5] 23 [18e28.5] 0.167
SOFA score, median [IQR] N ¼ 435 8.0 [6e10] 8.0 [6e11] 8.0 [6e9] 7.0 [5e9.75] 0.050
CCI, median [IQR] 4.0 [2e5] 4.0 [2e6] 4.0 [3e5] 4.0 [2e6] 0.985
mNUTRICscore, median [IQR] 5 [4e6] 5 [4e6] 5 [4e6] 5 [3e6] 0.648
mNUTRICrisk group 0.524
Low(<5), N (%) 183 (40.2) 46 (35.9) 111 (42.0) 26 (41.3)
High (5e9), N (%) 272 (59.8) 82 (64.1) 153 (58) 37 (58.7)
Admission categories, N (%)
Surgical emergency 90 (19.8) 35 (27.3) 47 (17.8) 8 (12.7) 0.032
Surgical 63 (13.8) 17 (13.3) 41 (15.5) 5 (7.9)
Medical 302 (66.4) 76 (59.4) 176 (66.7) 50 (79.4)
1 1
N¼numberofpatients,g*kg *day ¼gramperkilogramuncorrectedbodyweightperday,BMI¼bodymassindex,APACHEIIscore¼AcutePhysiologicandChronicHealth
EvaluationIIscore,SOFAscore¼sequentialorganfailureassessmentscore,CCI¼CharlsonComorbidityIndex,mNUTRICscore¼modifiedNutritionRiskinCriticallyIllscore,
IQR ¼ interquartile range (1ste3th quartile), percentiles by Tukey's Hinges distributions.
a Calculated by Pearson's Chi square or Fishers exact test, Anova or Kruskal Wallis test as appropriate.
1 1 1 1
(>1.2 g*kg *day )protein intake groups, respectively. Univariate day1e3butadvancedtomorethan0.8g*kg *day (group2(g2))
analysis showed a significant survival benefit of the intermediate on day 4 and later and patients who had protein intake of more
1 1
protein intake category compared with the high protein intake than 0.8 g*kg *day during the whole week (group 3 (g3)). A
category (p ¼ 0.043). However, this significance was lost in Cox significant difference in 6-month survival was observed between
regression multivariate analysis (p ¼ 0.209). g1 and g2 (p ¼ 0.005) and g2 and g3 (p ¼ 0.004) in univariate
analysis (Fig. 2). In multivariate analysis the significance between
3.3. Time dependent effect of protein intake g1 and g2 was lost. However, the survival benefit was confirmed
between g2 and g3, HR 0.609 (95% 0.480e0.772; p < 0.001).
Wesubsequently analysed the early (days 1e3) and late phase Moreover, a significant difference was observed between and g1
(days4e7)ofICUadmissionseparately(Table3).Proteinintakewas and g3 in multivariate analysis, HR 1.495 (95% CI 1.020e2.190; p
classified for mean daily protein intake during earlyand late phase. 0.039).
Lowproteinintake during days 1e3 was associated with a statisti-
cally significant reduction in 6-month mortality, whereas higher 3.5. Time-dependent optimal protein intake
proteinintakeduringdays4e7wasassociatedwithbetteroutcome
byunadjustedCoxproportionalhazardregression(Table3).Fordays Furthermore, we analysed the 6-month mortality risk of low,
1e3aHazardRatio(HR)of1.231(95%CI:1.040e1.457;p¼0.016)in intermediate and high protein intake of each ICU admission day
1 1 1 1
the >0.8 g*kg *day group compared to the <0.8 g*kg *day separately for the first week of admission in order to find daily
1 1 optimum protein intake. On day 1e2 the lowest mortality was
groupwasfound.Lowproteinintake<0.8g*kg *day duringdays
4e7hasaHRof1.605(95%CI1.118e2.186;p¼0.003)comparedto foundwithlowproteinintake,day3and5forintermediateprotein
thehighproteinintakegroup.ThelowestHRwasfoundinthegroup intake and day 6 and 7 for high protein intake. When comparing
withintermediateproteinintakeduringdays4e7(HR0.71695%CI this modeltothepreviousmentionedproteinintakecategoriesand
0.558e0.917; p ¼ 0.008). Further validation of these results was groups a survival benefit was shown with a 6-month survival of
done by assessing days 1e2, showing similar association of low 76.6% for the group advancing from low, to intermediate to high
protein intake and 6-month survival. When considering days 1e4, intake.
nodifferencebetweenthelowandhighintakegroupwasobserved
(data not shown). 3.6. Secondary outcomes
3.4. Time-dependent protein intake subgroups Secondary outcome measures were assessed based on time-
dependent subgroups. Statistical significant differences between
We subsequently compared patients with protein intakes less groups were found in 6-month mortality (g1 48.6%, g2 28.7%, g3
1 1
than 0.8 g*kg *day during the whole week (group 1 (g1)), with 42.7%, p ¼ 0.004) ICU mortality (g1 40.0%, g2 13.5%, g3 22.2%,
1 1 p ¼ 0.001) and hospital mortality (g1 48.6%, g2 20.8%, g3 33.3%,
patients who initially received less than 0.8 g*kg *day during
no reviews yet
Please Login to review.
