Journal of Parenteral and Enteral Nutrition

http://pen.sagepub.com/ Relationship of Vitamin D Deficiency to Clinical Outcomes in Critically Ill Patients

David M. Higgins, Paul E. Wischmeyer, Kelly M. Queensland, Stefan H. Sillau, Alexandra J. Sufit and Daren K. Heyland JPEN J Parenter Enteral Nutr 2012 36: 713 originally published online 20 April 2012 DOI: 10.1177/0148607112444449 The online version of this article can be found at: http://pen.sagepub.com/content/36/6/713

Published by:
http://www.sagepublications.com

On behalf of:

The American Society for Parenteral & Enteral Nutrition

Additional services and information for Journal of Parenteral and Enteral Nutrition can be found at: Email Alerts: http://pen.sagepub.com/cgi/alerts Subscriptions: http://pen.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav

>> Version of Record - Oct 17, 2012 OnlineFirst Version of Record - Apr 20, 2012 What is This?

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

444449
ns et alJournal of Parenteral and Enteral Nutrition 2012

PENXXX10.1177/0148607112444449Higgi

Original Communication

Relationship of Vitamin D Deficiency to Clinical Outcomes in Critically Ill Patients

David M. Higgins, MS1; Paul E. Wischmeyer, MD1; Kelly M. Queensland, BA1; Stefan H. Sillau, MS2; Alexandra J. Sufit, BA1; and Daren K. Heyland, MD3, 4

Journal of Parenteral and Enteral Nutrition Volume 36 Number 6 November 2012 713-720 2012 American Society for Parenteral and Enteral Nutrition DOI: 10.1177/0148607112444449 http://jpen.sagepub.com hosted at http://online.sagepub.com

Abstract
Background: Despite the numerous disease conditions associated with vitamin D deficiency in the general population, the relationship of this deficiency to outcome in critically ill patients remains unclear. The objective of this study is to determine the burden of vitamin D deficiency in intensive care unit (ICU) patients and determine if it is associated with poor patient outcomes. Methods: The authors conducted an analysis of samples collected from a prospective study of 196 patients admitted to a medical/surgical ICU in a tertiary care hospital. They measured serum 25-hydroxyvitamin D at admission and up to 10 days following admission and followed patients prospectively for 28-day outcomes. Results: Of analyzable patients, 50 (26%) were deficient (30 nmol/L) and 109 (56%) were insufficient (>30 and 60 nmol/L). Baseline 25(OH)D levels decreased significantly in all patients after 3 days in the ICU and remained significantly lower through 10 days (P < .001). 25(OH)D status was not significantly associated with 28-day all-cause mortality (hazard ratio [HR], 0.89; 95% confidence interval, [CI] 0.372.24). Higher levels of 25(OH)D were associated with a shorter time-to-alive ICU discharge (HR, 2.11; 95% CI, 1.273.51). 25(OH)D-deficient patients showed a nonstatistically significant trend toward a higher infection rate (odds ratio [OR], 3.20; 95% CI, 0.78413.07; P = .11) compared with patients with sufficient levels of 25(OH)D. Conclusions: This study demonstrates significant decreases in vitamin D status over the duration of the patients ICU stay. Low levels of vitamin D are associated with longer time to ICU discharge alive and a trend toward increased risk of ICU-acquired infection. (JPEN J Parenter Enteral Nutr. 2012;36:713-720)

Keywords
immunonutrition; research and diseases; vitamins; nutrition; critical care; research and diseases

Clinical Relevancy Statement

Vitamin D deficiency has been associated with a myriad of disease conditions in the general population. However, the impact of vitamin D deficiency on outcome in critically ill patients remains unclear. Although studies have described a high burden of vitamin D deficiency in this population and suggested a correlation with adverse outcomes, few studies have addressed specific outcomes. Furthermore, the usefulness of obtaining vitamin D levels in critically ill patients is not clear because most studies have evaluated preintensive care unit (ICU) vitamin D levels, and no large studies to date have trended vitamin D levels during the ICU stay. We confirm in this study that vitamin D deficiency is common in this population and is associated with a longer ICU length of stay, may increase the risk of ICU-acquired infections and pneumonia, and may be associated with elevated organ failure scores. In addition, we describe the trend of vitamin D status during the stay of patients in the ICU and demonstrate that there is a statistically significant decrease in the levels of vitamin D. These results add to the growing fund of knowledge concerning the association of vitamin D deficiency with adverse outcomes in the critical care population and indicate that further research is warranted in defining whether vitamin D supplementation is beneficial and what is the optimal dose and timing of vitamin D in this population.

From the 1Department of Anesthesiology, University of Colorado School of Medicine, Aurora, Colorado; 2Department of Biostatistics and Informatics, Colorado School of Public Health, University of Colorado, Aurora, Colorado; and 3Clinical Evaluation Research Unit, Kingston General Hospital, Kingston, Ontario, Canada; 4Department of Medicine, Queens University, Kingston, Ontario, Canada. This work was presented at the following meetings: Society for Critical Care Medicine, San Diego, California, January 16, 2011; American Society for Parenteral and Enteral Nutrition, Vancouver, Canada, January 30, 2011; and SHOCK society, Norfolk, Virginia, January 12, 2011. Financial disclosure: This study was a secondary analysis of an existing database. Received for publication December 7, 2011; accepted for publication January 6, 2012. Corresponding Author: Paul Wischmeyer, MD, University of Colorado School of Medicine, Department of Anesthesiology, RC-2, Mail Stop 8602, 12700 E 19th Ave, Aurora, CO 80045; e-mail: Paul.Wischmeyer@ ucdenver.edu

Download a QR code reader on your smartphone, scan this image, and listen to the podcast for this article instantly. Or listen to this and other JPEN podcasts at http://pen. sagepub.com/site/misc/Index/Podcasts. xhtml.

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

714

Journal of Parenteral and Enteral Nutrition 36(6)

Introduction
Vitamin D deficiency is a common disorder that is associated with excess morbidity and mortality in general population studies.1-4 The role of vitamin D in the regulation of calcium, phosphorus, and skeletal growth has been extensively described.5,6 Evidence is mounting that vitamin D deficiency also has an association with the immune system, affecting cancer, multiple sclerosis, diabetes, and autoimmune disorders.7-13 The recent discovery that many cells of the immune system express both the enzyme 25-hydroxyvitamin D-1hydroxylase to convert circulating 25(OH)D to hormonally active 1,25(OH)2D and vitamin D receptors may reveal a new and significant function of vitamin D on immune system regulation.10,14 Specifically, vitamin D may play a key physiologic role in attenuating potentially pathogenic cell-mediated immune responses and thus has been linked to an increased incidence of autoimmune disorders.7-13 Vitamin D may also be important for optimizing host defense against infection as children with rickets are more susceptible to infection,15 and vitamin D has been implicated in the pathogenesis of infections such as tuberculosis.16 Vitamin D deficiency in the critically ill patient was observed more than 20 years ago, primarily in hypocalcemic patients.17 Recent studies in normocalcemic critically ill patients have suggested a very high burden of vitamin D deficiency in this population and suggested a previously unrecognized contribution of deficient states to non-calcium-related consequences.18-24 However, to our knowledge, no large studies have specifically addressed vitamin D status trends through the duration of stay in the intensive care unit (ICU). Critically ill patients are at an especially high risk for many of the disorders that vitamin D deficiency may be associated with. Yet, evaluation of 25(OH)D status is rarely performed in the critical care setting.25 The hypothesis of this study is that vitamin D status in ICU patients is significantly decreased at admission and continues to decrease throughout their hospital stay. Furthermore, we hypothesize that vitamin D deficiency is associated with poor patient outcomes, particularly longer ICU stay and greater risk of infection.

were enrolled over the period of 1 year from October 2002 to October 2003. Research ethics boards from Queens University approved the protocol for the original study; informed consent was obtained from next of kin before enrollment. The inclusion criteria of the patients included all consecutive patients 18 years and older who were expected to stay more than 24 hours in the ICU and were enrolled within the first 24 hours of ICU admission. Only patients admitted for overdoses were excluded from the study. The study protocol at Kingston General Hospital included blood sample collection daily for a maximum of 10 days. Serum samples were also collected and stored at 80C prior to analysis.

Data Collection
Demographics, past medical history, and medications were obtained from the patients charts and included age, race, sex, body mass index (BMI; determined prior to ICU admission or estimated as pre-ICU dry weight [in kg] divided by height [in cm] squared in the ICU on admission), season of admission (summer admission was defined as admission between June 21 and September 22 of the year of the study), admission category (cardiovascular, respiratory, neurological, metabolic, gastrointestinal, hematologic, sepsis, postoperative), and comorbidities (coronary artery disease, hypertension, chronic obstructive pulmonary disease, diabetes, renal disease, liver disease, cancer). Necessary variables were recorded to calculate Acute Physiology and Chronic Health Evaluation II (APACHE II)27 and Sequential Organ Failure Assessment (SOFA) scores28 at admission. Clinical outcomes were assessed at 14 and 28 days. No patients were lost to follow up for the 14- or 28-day clinical end points. Outcomes recorded included mortality, time-to-alive ICU discharge, and infection status. The diagnosis of ICU-acquired infection was defined as infection present after 48 hours of ICU admission. Suspected infection was defined by the presence of a new positive culture result or initiation of new antibiotics after 48 hours of ICU admission. The charts of all patients with suspected infections were reviewed by 2 independent blinded physicians using standardized definitions.29 Adjudicators then met to determine the presence or absence of infection. In case of discrepancy, the physicians would meet and resolve the issue by reanalyzing the chart. Given the uncertainty around the absolute diagnosis of a new infection, we used standard definitions of probable and possible for each type of infection, where probable reflected a higher degree of probability, in the opinion of the adjudicator, that infection was present compared with possible infection. This adjudication technique has been validated and published in previous large clinical ICU trials.30,31 The max SOFA score was calculated as the worse score for each organ component over the patients ICU stay. The delta SOFA score was determined by subtracting the maximum SOFA score from the admission total SOFA score. As vitamin D insufficiency has been associated with an increased susceptibility to pulmonary infections,32 the subgroup of patients with ICU-acquired pneumonia >48 hours after admission was analyzed separately.

Materials and Methods

The original purpose of the prospective observational study was to evaluate a diagnostic marker for sepsis in critical illness. This trial was conducted at the Kingston General Hospital, a tertiary care, mixed medical-surgical ICU,26 and 2 other hospitals. In the original study, 597 patients were enrolled at the 3 sites. However, additional blood was collected and available for analysis only from patients at the Kingston General Hospital site, which included 203 total patients, and this is the cohort studied in this report. Here we report a secondary analysis of the serum samples collected in this study evaluating the relationship between 25(OH)D levels and clinical outcomes. The patients

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

Vitamin D Deficiency in Critically Ill Patients / Higgins et al

715

Serum was collected on the day of ICU admission (within 24 hours of ICU admittance) and daily for 10 days while in the ICU. All samples were stored at 80C for future analysis and did not undergo more than 1 freeze-thaw cycle before running assays. Serum 25(OH)D levels were assessed using a radioimmunoassay (RIA) method (DiaSorin, Antony, France) at The Childrens Hospital Clinical and Translational Research Center Core Lab in Denver, Colorado, at baseline and with samples from days 3, 5, 7, and 10 (when available). This laboratory participates biannually in the Vitamin D External Quality Assessment Scheme to ensure the analytical reliability of the 25(OH)D assay. The coefficient of variation for the 25(OH)D RIA assay is 0.046. Samples were run in 4 batches according to the day after admission. The following 25(OH)D values were used for deficient, insufficient, and sufficient patients according to previously reported subgroups: sufficient, >60 nmol/L; insufficient, >30 to 60 nmol/L; and deficient, 30 nmol/L,19 which are similar to the recent Institute of Medicine report cutoffs.33

Results
Patients and 25(OH)D Status
In the original study, 203 patients were enrolled at Kingston General Hospital. However, baseline serum samples were not available for analysis in 7 patients. Thus, a total of 196 patients on admission day were included in this analysis. Of the 196 patients analyzed, 66 (33.7%) were admitted as postoperative patients and 130 (66.3%) were admitted for medical reasons. The primary admission categories, including cardiovascular, respiratory, neurological, metabolic, gastrointestinal, and hematologic disorders, were not statistically different between the groups (see Table 1). Comorbid conditions existing prior to ICU admission, including coronary artery disease, hypertension, diabetes, renal disease, liver disease, or cancer, were not significantly different between the groups. In addition, APACHE II scores, presence of sepsis on admission, and baseline SOFA scores were similar between the 25(OH)D groups (see Table 1). Furthermore, there were no significant differences in mean parenteral nutrition (PN) and enteral nutrition (EN) over ICU length of stay between the groups. Upon admission to the ICU, a low baseline serum 25(OH) D (60 nmol/L) level was observed in 159 patients (82% of total cohort), as seen in Figure 1. Of analyzable patients, 50 (26%) were deficient (30 nmol/L), 109 (56%) were insufficient (>30 and 60 nmol/L), and 37 (18%) were sufficient (>60 nmol/L) at baseline. At baseline, the mean 25(OH)D level was 47.0 nmol/L. The mean dropped significantly by post admission day 3 to 44.4 nmol/L (n = 138, P < .001) and remained significantly decreased compared with baseline in patients who stayed 10 days or longer (45 nmol/L at baseline to 39 nmol/L on day 10, n = 44, P < .001). The mean 25(OH)D for patients who stayed at least 10 days (n = 44) in the insufficient or sufficient groups significantly decreased compared with the baseline on days 3, 7, and 10, as seen in Figure 2. Of the 23 patients who were sufficient at ICU admission and stayed longer than 3 days (14 patients excluded), 9 of 23 sufficient patients (39.1%) developed insufficiency at some point during their stay.

Statistics
For baseline SOFA scores, multiple regression was used to calculate adjusted least squares means using PROC GLM in SAS Version 9.2 (SAS Institute, Inc, Cary, NC) between 25(OH)D groups. In all multivariate models, we controlled for age, gender, BMI, and APACHE II score. Ethnicity was not included in the analysis because 201 of the 203 patients were white. Except for gender, all covariates were continuous. Logistic and proportional hazards models used Wald tests and confidence intervals. To assess the change in 25(OH)D status over the patients stay in the ICU, the PROC MIXED function was used in SAS to run a mixed model regression. 25(OH)D status was expressed as a categorical variable, and a multivariate Cox proportional hazard ratio model was used to identify variable association with time to ICU discharge or mortality. The PROC PHREG function in SAS was used to calculate the hazard ratios for time-to-alive ICU discharge and mortality variables. Patients who died were censored because their time to recovery was unknown. Because a 1-nmol/L increase in 25(OH)D is not clinically relevant, we represented a 1-unit increase as a 30-nmol/L increase in 25(OH)D levels for hazard ratio estimates. This unit selection scales the hazard ratio estimates but does not have any effect on the statistical significance. The P value for this analysis was determined using a type 3 Wald test. Logistic regression was used to model probable and possible ICU-acquired infections in the 25(OH)D groups using the PROC LOGISTIC function in SAS Version 9.2. The Student t test (unpooled [Satterthwaite]) was used to compare the means of baseline 25(OH)D between patients with and without pneumonia. The SAS Version 9.2 statistical package was used for all analyses. Means (SD) are reported. All tests were 2-sided, and P < .05 was considered significant. A statistical trend was considered P < .20.

Effect of 25(OH)D Level on Mortality and Time-to-Alive ICU Discharge

All-cause 28-day mortality in the study population was 26.2%. The mortality rates and average time to death for 25(OH)D sufficient, insufficient, and deficient groups were 27.0% and 8.0 7.0 days, 27.5% and 8.8 7.2 days, and 22.0% and 9.0 8.8 days, respectively. Using a multivariate Cox proportional hazard ratio (HR) model, 25(OH)D category was not associated with 28-day all-cause mortality (HR, sufficient vs deficient groups: 0.89; 95% confidence interval [CI], 0.372.24), as shown in Table 2. The mean SD time-to-alive ICU discharge for all patients in the study population was 7.5 8.5 days. The mean time-to-

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

716 Table 1. Comparison of Demographics Between 25(OH)D Groups

Sufficient (>60 nmol/L) 25(OH)D status, No. (%) Age, y, mean SD Male sex, No. (%) Body mass index, mean SD Summer admission, No. (%) APACHE II, mean SD Baseline SOFA, mean SD Sepsis, No. (%)c Parenteral + enteral nutrition, kcal/d, mean SDd Comorbidity, No. (%) CAD Hypertension COPD Diabetes Renal disease Liver disease Cancer Admission category, No. (%) Cardiovascular Respiratory Neurological Metabolic Gastrointestinal Sepsis Postoperative 37 (18.5) 65.0 13.3 17 (46) 26.7 5.0 7 (26) 20.8 8.4 7.1 2.9 15 (40) 485.5 82.5

Journal of Parenteral and Enteral Nutrition 36(6)

Insufficient (>30 to 60 nmol/L) 109 (55.9) 64.3 14.4 72 (66) 30.1 6.9 28 (36) 19.7 7.8 6.8 3.6 52 (48) 613.3 57.7

Deficient (30 nmol/L) 50 (25.6) 62.9 14.1 32 (64) 29.8 8.8 18 (36) 20.8 7.4 7.6 3.7 22 (44) 573.5 73.4

P Value .41 .09b .76a .19a .47a .45a .29b .49a .74b .35b .44b .93b .15b .15b .99b .77b .66b .99b .99b .26b .18b .81b

6 (17) 12 (33) 7 (19) 8 (22) 4 (11) 3 (8) 3 (8) 5 (14) 13 (36) 3 (8) 2 (6) 0 (0) 2 (6) 8 (22)

23 (21) 50 (46) 18 (17) 29 (27) 9 (8) 1 (1) 9 (8) 14 (13) 24 (22) 8 (7) 9 (8) 2 (2) 4 (4) 45 (41)

8 (16) 21 (42) 12 (24) 14 (28) 10 (20) 3 (6) 4 (8) 8 (16) 20 (40) 3 (6) 3 (6) 3 (6)6 0 (0) 13 (30)

APACHE II, Acute Physiology and Chronic Health Evaluation II; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; SOFA, Sequential Organ Failure Assessment. A total of 196 patient blood specimens were available for analysis. P values are from univariate analysis. a Analysis of variance test P value. b 2 test P value. c Patients were classified as having sepsis if at any time during their stay they met systemic inflammatory response syndrome criteria and had a source of infection. d Mean parenteral and enteral nutrition per day.

alive ICU discharge for 25(OH)D sufficient, insufficient, and deficient groups was 5.9 5.4, 6.8 6.0, and 10.6 8.4, respectively. Sufficient levels of 25(OH)D were significantly associated with a shorter time-to-alive ICU discharge (HR, 1.74 per 30 nmol/L; 95% CI, 1.192.53 at 14 days and HR, 3.49; 95% CI, 1.368.95 at 28 days). Using the multivariate Cox proportional hazards model with 25(OH)D represented as a categorical variable, we found that 25(OH)D sufficient or insufficient patients left the ICU sooner than patients with deficient levels of 25(OH)D, as shown in Table 2. Max SOFA scores were calculated and showed nonsignificantly higher SOFA scores in deficient patients compared with the other groups (deficient max SOFA 9.7 4.2, insufficient max SOFA 8.4 4.2, and sufficient max SOFA 8.4 3.3; P = .12). Delta SOFA scores were 1.3 1.7, 1.6 2.1, and 2.1 2.3 in sufficient, insufficient, and deficient groups, respectively (P = .29).

In addition, the mean duration of ventilation was determined and found to be 141.5 18.1 hours in vitamin Ddeficient patients compared with 138 16.9 hours in insufficient and 120 20.9 hours in sufficient groups (P = .49), as shown in Table 3.

Infection Status
Patients were followed for infection status, and the relationship of baseline 25(OH)D status to new infections diagnosed after 48 hours in the ICU is summarized in Table 3. 25(OH)D-deficient and insufficient patients had a higher rate of at least 1 probable infection compared with patients with sufficient levels of 25(OH)D, but this was not statistically significant (odds ratio [OR], 3.20; 95% CI, 0.78413.07, P = .11). There was no relationship between rate of culture-confirmed infections and possible infections and 25(OH)D levels (Table 3).

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

Vitamin D Deficiency in Critically Ill Patients / Higgins et al

717

Admission 25(OH)D (nmol/L) levels

200 180 160 140 120 100 80 60 40 20 0

1-002 1-008 1-014 1-020 1-026 1-032 1-038 1-045 1-051 1-057 1-063 1-069 1-075 1-081 1-087 1-093 1-099 1-105 1-111 1-117 1-123 1-129 1-135 1-141 1-147 1-153 1-159 1-165 1-171 1-177 1-183 1-189 1-195 1-201

Table 2. Cox Proportional Hazard Model of Mortality and Time-to-Alive ICU Discharge
Adjusted HR (95% CI) Mortality Sufficient vs deficient Insufficient vs deficient Sufficient vs insufficient Time-to-alive ICU discharge Sufficient vs deficient Insufficient vs deficient Sufficient vs insufficient 0.89 (0.37, 2.24) 1.23 (0.61, 2.49) 0.724 (0.33, 1.63) 2.11 (1.27, 3.51) 1.59 (1.05, 2.40) 1.33 (0.86, 2.06) P Value .46 .01

Paents

Figure 1. Admission levels of 25(OH)D (nmol/L) in 196 intensive care unit (ICU) patients. The following 25(OH) D values were used for deficient, insufficient, and sufficient patients according to previously reported subgroups: sufficient, >60 nmol/L; insufficient, >30 to 60 nmol/L; and deficient, 30 nmol/L.

Age, sex, body mass index, and Acute Physiology and Chronic Health Evaluation II were controlled for in all analyses. Patients who died were censored because their time to recovery was unknown. CI, confidence interval; HR, hazard ratio; ICU, intensive care unit.

infection or pneumonia from documented infection or cultureconfirmed pneumonia diagnosed >48 hours after admission had significantly lower levels of baseline 25(OH)D compared with those who did not have pneumonia, as seen in Figure 3.

95

**

** **

Discussion
All paents (n=44)

25(OH)D (nmol/L)

75

55

** **

** **

** **

Decient (n=15) Insucient (n=21) Sucient (n=8)

35

15 1 3 7 10

Days aer admission

Figure 2. Levels of 25(OH)D decreased over intensive care unit (ICU) patient stay. Levels of 25(OH)D on days 1, 3, 7, and 10 following admission were assayed from patient serum samples. A total of 44 patients (only patients who stayed 10 days or longer) are displayed in this figure. The data represent the mean SE bars. The All patients line represents all patients who stayed 10 days or longer. The Sufficient, Insufficient, and Deficient lines represent patients in the respective group who stayed at least 10 days or longer. **P values <.001 based on a mixed model regression analysis and are relative to the day 1 sample within the group.

Pneumonia rates were 5%, 10%, and 16% in sufficient, insufficient, and deficient patients, respectively. Although the rate of culture-positive and probable pneumonia between the groups was higher in the 25(OH)D-deficient group, these results were not significant (culture confirmed: OR, 4.24; 95% CI, 0.5038.8, P = .21; probable pneumonia: OR, 3.07; 95% CI, 0.76312.36, P = .11). Patients with a probable pulmonary

Here we report an observational study evaluating the relationship between vitamin D deficiency and clinical outcomes in an unselected, heterogeneous critically ill patient population. This study showed that a significant number of ICU patients had inadequate vitamin D status (82% of patients were insufficient or deficient) on admission. In addition, this is the first study to describe a significant decrease in 25(OH)D levels from admission 25(OH)D levels as early as 3 days after admission and as late as 10 days. Our data also demonstrate a high incidence (39.1%) of sufficient patients becoming insufficient during their stay. We did not observe an association between baseline 25(OH)D status and 28-day all-cause mortality, but 25(OH)D-deficient patients stayed significantly longer in the ICU, tended to develop more organ failure, and tended to experience a higher number of infections, particularly pneumonia, compared with sufficient patients. The outcomes associated with inadequate 25(OH)D status seen in this study corroborate the findings of a recent smaller study (n = 41) by Lee et al19 that showed that 93% of patients had inadequate baseline 25(OH)D levels. However, little is known about the measurement of 25(OH)D status during a patients duration of stay. Only 1 smaller study21 (12 patients) indirectly evaluated 25(OH)D status beyond admission levels in the ICU, and to our knowledge, no studies have demonstrated a significant decrease in 25(OH)D status during the patients stay. In our study, we show a significant decrease in 25(OH)D levels over the first 310 days in the ICU in the sufficient and insufficient groups but no changes in patients who are already deficient. The explanation for this acute decrease in 25(OH)D

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

718 Table 3. Infection Outcomes in 25(OH)D Groups

Sufficient, No. (%) Possible ICU infection Probable ICU infection Probable pneumonia Culture-positive pneumonia Duration of ventilation, h, mean SD 5 (13%) 4 (11%) 2 (5%) 2 (5%) 120 20.9 Insufficient, No. (%) 19 (17%) 15 (14%) 10 (10%) 8 (7%) 138 16.9

Journal of Parenteral and Enteral Nutrition 36(6)

Deficient, No. (%) 12 (24%) 9 (18%) 8 (16%) 6 (12%) 141 18.1

Deficient vs Sufficient OR 1.7 (0.48, 2.7) 3.2 (0.78, 13.0) 4.2 (0.46, 38.8) 3.0 (0.76, 12.3)

P Value .35 .11 .11 .20 .49

Intensive care unit (ICU) infections were defined as infections diagnosed >48 hours after ICU admission. Possible/probable infection or probable pneumonia status was adjudicated by 2 independent clinicians. OR, odds ratio.

A)
25(OH)D (nmol/L)

60 50 40 30 20 10 0 No infecon No infecon (n=160) Infecon Infecon (n=36)

* p=0.01

B)
25(OH)D (nmol/L)

50 40 30 20 10 0 No pneumonia (n=176) Pneumonia (n=20) Culture posive (n=16)

* p=0.01

* p=0.05

Figure 3. (A) Baseline levels of 25(OH)D in patients with a probable new infection vs no new infection diagnosed >48 hours after admittance to the intensive care unit (ICU). The data represent the mean SD and P values compared with no infection. (B) Baseline levels of 25(OH)D in patients with new culture-positive pneumonia, probable pneumonia, and no pneumonia diagnosed >48 hours after admission to the ICU. The probable pneumonia group included culture-positive pneumonia patients in addition to 4 patients without culture positivity but clinical presentation of pneumonia. The data represent the mean SD and P values compared with patients with no pneumonia. *P < 0.05.

in the sufficient and insufficient groups is not clear but could be due to changes in vitamin D binding protein (VDBP), which approximately 85% of 25(OH)D is bound to.34 VDBP is decreased in ICU patients and has been shown to be a predictor of ICU mortality.35 Another recent study demonstrated that ICU patients with sepsis and worse outcomes had significantly lower levels of VDBP compared with those without sepsis.20 A

decrease in VDBP may be due to decreased protein synthesis or increased clearance of VDBP in ICU patients with liver, renal, or intestinal diseases. However, because VDBP concentrations are much higher than 25(OH)D and VDBP has a high affinity for vitamin D metabolites, reduced levels of VDBP often result in lower total 25(OH)D but do not affect free concentrations.36 The VDBP relationship should be considered in any future studies examining the acute changes in 25(OH)D observed in this study. In our study, we did not observe an association with 28-day all-cause mortality and 25(OH)D status. This is in contrast to a recent study by Braun et al24 that evaluated preadmission vitamin D levels and found vitamin D deficiency to be significantly associated with 30-, 90-, and 365-day mortality. However, the prior study was limited by selection bias and measurement of vitamin D status up to 365 days prior to ICU admission. The lack of association of vitamin D status with mortality in our study may be due to our study being underpowered. We were not able to run a power analysis a priori because the relationship of vitamin D to mortality was not the initial end point of the trial at conception, but post hoc power analysis suggests that this study did not have the power to detect small differences in mortality or infection rates between the groups (data not shown). Therefore, the failure to demonstrate statistically significant differences in mortality as well as infection rates in this study should not be interpreted as evidence for lack of a true difference. In fact, the magnitude of the association with infection and vitamin D levels is quite large (OR, 1.73.2), but we lacked statistical precision because of our limited sample size. The data presented here are the largest thus far to report an increased time-to-alive ICU discharge associated with admission vitamin D deficiency. Our study also suggests a hypothesis for the observed increased time-to-alive ICU discharge. We used a very rigorous adjudication process to determine the presence or absence of infection and a rigorous definition of pneumonia. Our observation of increased infection rates and decreased organ function may be factors associated with vitamin D deficiency that could contribute to prolonged time-toalive ICU discharge and poorer outcomes in ICU patients. This potentially could be due to the role vitamin D has on the immune system. In line with this, a recent report found

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

Vitamin D Deficiency in Critically Ill Patients / Higgins et al

719

significantly lower levels of the endogenous antimicrobial peptide cathelicidin (LL-37) in critically ill patients who were vitamin D deficient, suggesting a role for vitamin D in maintaining innate immunity to infection in the ICU.20 Furthermore, 1,25(OH)D has been shown to have a stimulatory effect on the innate immune system by increasing interleukin-1 (IL-1) production and stimulating monocyte proliferation.37 Without these important innate immune system stimulating effects, patients may be more susceptible to ICU infections. On the other hand, sufficient levels of vitamin D have been associated with lower rates of autoimmune disorders, and it may be that adequate levels of vitamin D are necessary to regulate overactive, pathologic inflammatory immune responses.10,14 Therefore patients may be more susceptible to pathologic, overactive immune responses to common ICU infections. Limitations to this study include a relatively small sample size and confounding variables that could not be accounted for in our multivariate analysis. For instance, the observed vitamin D deficiency in this study could be due to a number of confounding causes, including the fact that many patients have limited exposure to UV-B light during chronic disease and often have significant dietary alterations, and the calcium parathyroidvitamin D axis may be disrupted in critical illness, so a consideration of PTH and Ca levels should be considered in future studies.25,38 Although our study suggests an association between inadequate vitamin D levels and poor patient outcomes, this does not mean vitamin D is the direct cause of the poor outcomes. It is possible that the conditions leading to poor outcomes cause a decrease in vitamin D levels, which are reactionary and not causal. More studies are needed to determine if there is direct causality between admission vitamin D status and patient outcomes. Furthermore, although there was not a statistically significant difference in the rates of infection and pneumonia, the significant differences in baseline levels of 25(OH)D between infected and noninfected patients suggest an association that this study may not have had the power to observe in actual infection rates. A recent report from the Institute of Medicine (IOM) points out that cutoffs for vitamin D deficiency have not been systematically developed.33 The cutoffs used for this study to define sufficient, insufficient, and deficient groups were selected according to recently published articles on vitamin D in ICU patients to allow for across-study comparison in ICU patients. Our cutoffs, although not exactly matching, are similar to the IOM report that suggests 25(OH)D levels <50 nmol/L (<20 ng/ mL) are insufficient for bone health.33 However, the cutoff values set for other concerns, such as bone health in the general population, may not be applicable to the critically ill population. Therefore, without sufficient evidence for the appropriate sufficient/insufficient cutoffs in this select ICU population, we decided on the cutoffs used in this article for cross-study comparison to other ICU vitamin D studies. Another consideration that adds to the challenge of setting hard cutoff values for vitamin D is the various assays available to measure 25(OH)D. The major assays used include the

RIA method used in this study, electrochemiluminescence (ECLIA) methods, and high-performance liquid chromatography (HPLC), which is often considered the gold standard.34 Some studies have suggested that RIA and ECLIA methods underestimate or overestimate levels of 25(OH)D because of how they measure the major forms of 25(OH)D found in the blood.34 This is a potential limitation of this study. However, reports are inconsistent, and the RIA method used in our study has been demonstrated in many studies to correlate with other assay methods, including HPLC,34,39,40 even after long-term storage at 80C.41,42 Future studies of vitamin D in the ICU also need to consider the assay method, and proper laboratory standardization should be ensured to avoid a bias in the levels. The findings of this study highlight the need for further studies to evaluate the optimal dose of 25(OH)D levels in critically ill ICU patients. Unfortunately, a limitation of our study was not being able to examine differences in vitamin D intake between groups during the ICU stay, although general nutrition intake was not significantly different between the groups. It cannot be assumed that the dose used to correct a healthy individual will be sufficient to correct a critically ill patient. The current American Society for Parenteral and Enteral Nutrition recommendation for hospitalized patients of 200 IU of vitamin D daily, which is intended for maintenance and not repletion, has been shown to be insufficient to normalize vitamin D levels.43,44 Recently, a small study demonstrated that serum 25(OH)D levels in critically ill patients may be normalized by the oral administration of larger doses (two 60,000 IU doses) of vitamin D.21 Furthermore, a recent small study using a single oral dose of 540,000 IU of vitamin D3 in critically ill patients demonstrated a significant increase in serum 25(OH)D levels without adverse effects.45 However, the optimal dosing of vitamin D required to both normalize the deficiency observed in the ICU setting and potentially have a favorable effect on infection and other outcomes is unknown and requires further study. It will be vital, prior to any large clinical outcome trial, to first conduct a welldesigned dosing trial to ensure that sufficient doses of vitamin D are administered to correct the deficiency observed in many ICU patients. In addition, these dosing studies will have to consider the form of vitamin D repletion because it is available in various forms (ergocalciferol vs cholecalciferol) and various routes depending on location (no intravenous vitamin D is available in the United States). In conclusion, in this study, we demonstrate that 25(OH)D insufficiency is very common in the ICU (82% of patients) and is significantly associated with longer time-to-alive ICU discharge, a trending toward increased organ failure, and a nonsignificant trend toward a higher number of infections. In addition, we report for the first time that 25(OH)D levels continue to significantly decrease through the duration of a patients ICU stay. Further research defining whether vitamin D supplementation is beneficial and what is the optimal dose of vitamin D in this population is warranted. In addition, future studies with larger sample sizes are required so that potential differences regarding mortality, infection, and immune status can be further evaluated.

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013

720

Journal of Parenteral and Enteral Nutrition 36(6)

Acknowledgments
The authors thank the Childrens Hospital Clinical and Translational Research Center Core Lab in Denver, Colorado, for technical support in analyzing vitamin D levels from serum.

References
1. Holick MF. High prevalence of vitamin D inadequacy and implications for health. Mayo Clin Proc. 2006;81(3):353-373. 2. Looker AC, Dawson-Hughes B, Calvo MS, Gunter EW, Sahyoun NR. Serum 25-hydroxyvitamin D status of adolescents and adults in two seasonal subpopulations from NHANES III. Bone. May 2002;30(5):771-777. 3. Rosen CJ. Clinical practice: vitamin D insufficiency. N Engl J Med. 2011;364(3):248-254. 4. Ross AC, Manson JE, Abrams SA, et al. The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know. J Clin Endocrinol Metab. 2011;96(1):53-58. 5. DeLuca HF. Overview of general physiologic features and functions of vitamin D. Am J Clin Nutr. 2004;80(6)(suppl):1689S-1696S. 6. Holick MF. Vitamin D deficiency. N Engl J Med. 2007;357(3):266-281. 7. Munger KL, Levin LI, Hollis BW, Howard NS, Ascherio A. Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis. JAMA. 2006;296(23):2832-2838. 8. Holick MF. Sunlight and vitamin D for bone health and prevention of autoimmune diseases, cancers, and cardiovascular disease. Am J Clin Nutr. 2004;80(6)(suppl):1678S-1688S. 9. Holick MF. Vitamin D: importance in the prevention of cancers, type 1 diabetes, heart disease, and osteoporosis. Am J Clin Nutr. 2004;79(3): 362-371. 10. Moro JR, Iwata M, von Andriano UH. Vitamin effects on the immune system: vitamins A and D take centre stage. Nat Rev Immunol. 2008;8(9):685-698. 11. Feskanich D, Ma J, Fuchs CS, et al. Plasma vitamin D metabolites and risk of colorectal cancer in women. Cancer Epidemiol Biomarkers Prev. 2004;13(9):1502-1508. 12. Holick MF. Calcium plus vitamin D and the risk of colorectal cancer. N Engl J Med. 2006;354(21):2287-2288; author reply 2287-2288. 13. Merlino LA, Curtis J, Mikuls TR, Cerhan JR, Criswell LA, Saag KG. Vitamin D intake is inversely associated with rheumatoid arthritis: results from the Iowa Womens Health Study. Arthritis Rheum. 2004;50(1):72-77. 14. DeLuca HF. Evolution of our understanding of vitamin D. Nutr Rev. Oct 2008;66(10)(suppl 2):S73-S87. 15. van Etten E, Mathieu C. Immunoregulation by 1,25-dihydroxyvitamin D3: basic concepts. J Steroid Biochem Mol Biol. 2005;97(1-2):93-101. 16. Shapira Y, Agmon-Levin N, Shoenfeld Y. Mycobacterium tuberculosis, autoimmunity, and vitamin D. Clin Rev Allergy Immunol. 2010;38(23):169-177. 17. Zaloga GP, Chernow B. The multifactorial basis for hypocalcemia during sepsis: studies of the parathyroid hormonevitamin D axis. Ann Intern Med. 1987;107(1):36-41. 18. Lucidarme O, Messai E, Mazzoni T, Arcade M, du Cheyron D. Incidence and risk factors of vitamin D deficiency in critically ill patients: results from a prospective observational study. Intensive Care Med. 2010;36(9):1609-1611. 19. Lee P, Eisman JA, Center JR. Vitamin D deficiency in critically ill patients. N Engl J Med. 2009;360(18):1912-1914. 20. Jeng L, Yamshchikov AV, Judd SE, et al. Alterations in vitamin D status and anti-microbial peptide levels in patients in the intensive care unit with sepsis. J Transl Med. 2009;7:28. 21. Mata-Granados JM, Vargas-Vasserot J, Ferreiro-Vera C, Luque de Castro MD, Pavon RG, Quesada Gomez JM. Evaluation of vitamin D endocrine system (VDES) status and response to treatment of patients in intensive care units (ICUs) using an on-line SPE-LC-MS/MS method. J Steroid Biochem Mol Biol. 2010;121(1-2):452-455. 22. McKinney JD, Bailey BA, Garrett LH, Peiris P, Manning T, Peiris AN. Relationship between vitamin D status and ICU outcomes in veterans. J Am Med Directors Assoc. 2011;12(3):208-211.

23. Cecchi A, Bonizzoli M, Douar S, et al. Vitamin D deficiency in septic patients at ICU admission is not a mortality predictor. Minerva Anestesiol. 2011;77(12):1184-1189. 24. Braun A, Chang D, Mahadevappa K, et al. Association of low serum 25-hydroxyvitamin D levels and mortality in the critically ill. Crit Care Med. 2011;39(4):671-677. 25. Lee P, Nair P, Eisman JA, Center JR. Vitamin D deficiency in the intensive care unit: an invisible accomplice to morbidity and mortality? Intensive Care Med. 2009;35(12):2028-2032. 26. Heyland DK, Govert, J, Garrard C. Waveform abnormalities of APPT in critically ill hospitalized patients: the WATTCH study. Crit Care Med. 2006;34(12):A100. 27. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13(10):818-829. 28. Moreno R, Vincent JL, Matos R, et al. The use of maximum SOFA score to quantify organ dysfunction/failure in intensive care: results of a prospective, multicentre study. Working Group on Sepsis related Problems of the ESICM. Intensive Care Med. 1999;25(7):686-696. 29. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health careassociated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control. 2008;36(5):309-332. 30. Heyland DK, Dodek P, Muscedere J, Day A, Cook D. Randomized trial of combination versus monotherapy for the empiric treatment of suspected ventilator-associated pneumonia. Crit Care Med. 2008;36(3):737-744. 31. A randomized trial of diagnostic techniques for ventilator-associated pneumonia. N Engl J Med. 2006;355(25):2619-2630. 32. Ginde AA, Mansbach JM, Camargo CA Jr. Vitamin D, respiratory infections, and asthma. Curr Allergy Asthma Rep. 2009;9(1):81-87. 33. Ross AC, Institute of Medicine Committee to Review Dietary Reference Intakes for Vitamin D and Calcium. Dietary Reference Intakes for Calcium and Vitamin D. Washington, DC: National Academies Press; 2011. 34. Jafri L, Khan AH, Siddiqui AA, et al. Comparison of high performance liquid chromatography, radio immunoassay and electrochemiluminescence immunoassay for quantification of serum 25 hydroxy vitamin D. Clin Biochem. 2011;44(10-11):864-868. 35. Dahl B, Schiodt FV, Nielsen M, Kiaer T, Williams JG, Ott P. Admission level of Gc-globulin predicts outcome after multiple trauma. Injury. 1999;30(4):275-281. 36. Bikle DD, Halloran BP, Gee E, Ryzen E, Haddad JG. Free 25hydroxyvitamin D levels are normal in subjects with liver disease and reduced total 25-hydroxyvitamin D levels. J Clin Invest. 1986;78(3):748-752. 37. Mora JR, Iwata M, von Andrian UH. Vitamin effects on the immune system: vitamins A and D take centre stage. Nat Rev Immunol. 2008;8(9):685-698. 38. Kelly A, Levine MA. Hypocalcemia in the critically ill patient. J Intensive Care Med. 2011 Aug 12. [Epub ahead of print] 39. Binkley N, Krueger D, Gemar D, Drezner MK. Correlation among 25-hydroxy-vitamin D assays. J Clin Endocrinol Metabolism. 2008;93 (5):1804-1808. 40. Leino A, Turpeinen U, Koskinen P. Automated measurement of 25-OH vitamin D3 on the Roche Modular E170 analyzer. Clin Chem. 2008;54(12):2059-2062. 41. Ersfeld DL, Rao DS, Body JJ, et al. Analytical and clinical validation of the 25 OH vitamin D assay for the LIAISON automated analyzer. Clin Biochem. 2004;37(10):867-874. 42. Antoniucci DM, Black DM, Sellmeyer DE. Serum 25-hydroxyvitamin D is unaffected by multiple freeze-thaw cycles. Clin Chem. 2005;51(1): 258-261. 43. A.S.P.E.N. Board of Directors and the Clinical Guidelines Task Force. Guidelines for the use of parenteral and enteral nutrition in adult and pediatric patients. JPEN J Parenter Enteral Nutr. 2002;26(1)(suppl):1SA-138SA. 44. Van den Berghe G, Van Roosbroeck D, Vanhove P, Wouters PJ, De Pourcq L, Bouillon R. Bone turnover in prolonged critical illness: effect of vitamin D. J Clin Endocrinol Metab. 2003;88(10):4623-4632. 45. Amrein K, Sourij H, Wagner G, et al. Short-term effects of high-dose oral vitamin D3 in critically ill vitamin D deficient patients: a randomized, double-blind, placebo-controlled pilot study. Crit Care. 2011;15(2):R104.

Downloaded from pen.sagepub.com at A.S.P.E.N on August 28, 2013