Association between prehospital vitamin D status and hospitalacquired bloodstream infections1–3 Sadeq A Quraishi, Augusto A Litonjua, Takuhiro Moromizato, Fiona K Gibbons, Carlos A Camargo Jr, Edward Giovannucci, and Kenneth B Christopher

INTRODUCTION

Hospital-acquired infections (HAIs)4 are common in the United States with an incidence of 22 cases/1000 admissions (1). Each year, w2 million new cases of nosocomial infections account for nearly 100,000 potentially avoidable deaths (2). Patients with HAIs incur substantial excess annual health care expenditures ($28 billion–$45 billion), driven primarily by an

952

increase in the hospital length of stay (3). The most common HAIs in the United States in order of frequency are catheterrelated urinary tract infections, surgical site infections, pneumonias, and hospital-acquired bloodstream infections (HABSIs) (2). Although HABSIs account for only 11–14% of all nosocomial infections, they account for a disproportionate share of the hospital length of stay, cost, and mortality attributable to HAIs (2–4). Strategies to decrease the incidence of HABSI focus on the control of environmental factors including provider hand hygiene, skin decontamination before the placement of invasive lines, antiseptic-impregnated catheters, and the use of prophylactic antibiotics (5). Significant declines in HABSI rates, but not eradication, have been noted after adoption of prevention strategies (6). Although it is known that derangements in immune function and disruption of natural barrier sites can predispose patients to infectious complications (7), few studies have explored potentially modifiable host factors that may decrease risk of HABSI. Evidence has suggested that vitamin D is a key regulator of the innate and adaptive immune system (8). Indeed, low serum 25-hydroxyvitamin D [25(OH)D] concentrations have been hypothesized to be an important risk factor for infection susceptibility and the development of HABSIs (9). Because suboptimal 1

From the Departments of Anesthesia, Critical Care and Pain Medicine (SAQ) and Emergency Medicine (CAC), Massachusetts General Hospital, Boston, MA; the Channing Division of Network Medicine and Pulmonary and Critical Care Division (AAL and FKG) and The Nathan E Hellman Memorial Laboratory, Renal Division (TM and KBC), Department of Medicine, Brigham and Women’s Hospital, Boston, MA; and the Departments of Nutrition and Epidemiology, Harvard School of Public Health, Boston, MA (EG). 2 Supported by the NIH [5T32GM007592-33 and UL1 RR025758 (to SAQ); R01 AI093723 and U01 AI087881 (to CAC); and K08AI060881 (to KBC)]. 3 Address correspondence to KB Christopher, Renal Division, Department of Medicine, Brigham and Women’s Hospital, 75 Francis Street, MRB 418, Boston, MA 02115. E-mail: [email protected] 4 Abbreviations used: BWH, Brigham and Women’s Hospital; CoNS, coagulase-negative staphylococcus; DRG, Diagnostic Related Group; HABSI, hospital-acquired bloodstream infection; HAI, hospital-acquired infection; LC-MS, liquid chromatography–mass spectroscopy; MGH, Massachusetts General Hospital; 25(OH)D, 25-hydroxyvitamin D. Received January 16, 2013. Accepted for publication July 10, 2013. First published online August 14, 2013; doi: 10.3945/ajcn.113.058909.

Am J Clin Nutr 2013;98:952–9. Printed in USA. Ó 2013 American Society for Nutrition

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

ABSTRACT Background: Alterations in immune function can predispose patients to nosocomial infections. Few studies have explored potentially modifiable host factors that may improve immune function and decrease risk of hospital-acquired bloodstream infection (HABSI). Vitamin D is a key regulator of innate and adaptive immune systems that may influence host susceptibility to infections. Objective: We investigated the association between prehospital serum 25-hydroxyvitamin D [25(OH)D] concentrations and risk of HABSI. Design: We performed a retrospective cohort study of 2135 adult patients from 2 Boston teaching hospitals. All patients had 25(OH)D concentrations measured before hospitalization between 1993 and 2010. The main outcome measure was HABSI, which was defined as positive blood cultures from samples drawn 48 h after hospital admission. Coagulase-negative Staphylococcus isolates were not considered to be bloodstream infections. Associations between 25(OH)D groups and HABSI were estimated by using bivariable and multivariable logistic regression models. Adjusted ORs were estimated with the inclusion of covariate terms thought to plausibly interact with both 25(OH)D concentration and HABSI. Results: Compared with patients with 25(OH)D concentrations $30 ng/mL, patients with concentrations ,30 ng/mL had higher odds of HABSI. For 25(OH)D concentrations ,10 ng/mL, the OR was 2.33 (95% CI: 1.45, 3.74); for 25(OH)D concentrations from 10 to 19.9 ng/mL, the OR was 1.60 (95% CI: 1.04, 2.46); and for 25(OH)D concentrations from 20 to 29.9 ng/mL, the OR was 1.13 (95% CI: 0.69, 1.84). After adjustment for age, sex, race (nonwhite compared with white), patient type (medical compared with surgical), and DeyoCharlson index, the ORs of HABSI were 1.95 (95% CI: 1.22, 3.12), 1.36 (95% CI: 0.89, 2.07), and 0.98 (95% CI: 0.60, 1.62), respectively. Conclusions: The analysis of 2135 adult patients showed that 25(OH)D concentrations ,10 ng/mL before hospitalization were associated with significantly increased odds of developing HABSI. These data support the initiation of randomized trials to test the role of vitamin D supplementation in HABSI prevention. Am J Clin Nutr 2013;98:952–9.

VITAMIN D DEFICIENCY AND HABSI

25(OH)D concentrations are increasingly prevalent in the general population of the United States (10–13), we performed a 2-center observational study of adult patients in whom 25(OH)D concentrations had been measured for nonspecific reasons #1 y before hospitalization. The objective of this study was to test our hypothesis that vitamin D status before hospital admission is associated with risk of developing HABSI. SUBJECTS AND METHODS

Source population

Data sources Data on all patients admitted to the BWH or MGH between 3 August 1993 and 5 January 2011 were obtained through the Research Patient Data Registry, which is a computerized registry that serves as a central data warehouse for all inpatient and outpatient records at Partners HealthCare sites, which include the BWH and MGH. The Research Patient Data Registry has been used for other clinical research studies (14–18). Institutional review board approval for the study was granted by the Partners Human Research Committee.

analysis, we categorized a prehospital serum 25(OH)D concentration ,20 ng/mL as a cutoff for vitamin D inadequacy, as suggested in a recent Institute of Medicine report (20). We used the Deyo-Charlson index to assess the burden of chronic illness (21) by using International Classification of Diseases (Ninth Edition) coding algorithms ,which are well studied and validated (22, 23). Patient type’ was defined as medical or surgical and incorporated the DRG methodology (24). Intensive care unit admission was determined by Current Procedural Terminology code 99291 (critical care, first 30–74 min) assignment during hospital admission and has been previously validated in the Research Patient Data Registry database (18). 25(OH)D assays Between 1993 and 2011, the following different assays were used at the 2 hospitals: a chemiluminescence assay, radioimmunoassay, or liquid chromatography–mass spectroscopy (LCMS). Dates, times, and types of 25(OH)D assay were recorded. The clinical laboratories where the assays were performed were Clinical Laboratory Improvement Amendments certified. The 25(OH)D assays were tested for imprecision by clinical laboratories at the 2 hospitals. Imprecision testing with human serum specimens showed within-run CVs of #4.5% for the chemiluminescence assay, #10.8% for the radioimmunoassay, and #8.6% for the LC-MS. Method corrections were not used when institutions changed assays. Data on the comparability of 25(OH)D assays has been performed by other groups (25), but because this was a retrospective, observational study, the determination of between-method assay differences was not possible.

Study population During the study period, there were 24,787 individual patients admitted, aged $18 y, with serum 25(OH)D measured between 7 and 365 d before hospitalization and who were assigned a Diagnostic Related Group (DRG). Exclusions included the following: 23 foreign patients without Social Security Numbers because vital status in this study was determined by the Social Security Administration Death Master File, 618 patients with missing laboratory data, 216 patients who received high-dose vitamin D supplementation (50,000 IU ergocalciferol) between the 25(OH)D draw and hospital admission, 3093 patients who received critical care because the data had been previously analyzed (14), and 18,702 patients who did not have blood cultures drawn .48 h after hospital admission. Thus, 2135 patients constituted the total study population. Compared with the study cohort, the parent cohort of 24,787 patients had the same average age but more women (62%), more surgery patients (40%), a higher mean 25(OH)D value (28 ng/mL), and a lower 30-d mortality rate (4%). Exposure of interest and comorbidities The exposure of interest was prehospital serum 25(OH)D concentration obtained 7–365 d before the date of hospital admission and categorized a priori as ,10 ng/mL (,25 nmol/L), 10–19.9 ng/mL (25–49.9 nmol/L), 20–29.9 ng/mL (50–74.9 nmol/L), and $30 ng/mL (75 nmol/L). All cutoffs were adapted from existing national clinical guidelines (19). In a secondary

Endpoints The primary endpoint was the presence of HABSI. Information on blood cultures for the study cohort was obtained from microbiology reports from the computerized registry at the hospitals under study. All cohort patients had blood cultures drawn. Blood cultures were defined as positive if aerobic, anaerobic, or fungal blood cultures grew identifiable organisms. Patients with positive blood cultures were considered to have bloodstream infections (26–28). Bloodstream infections were characterized as hospital acquired if they were drawn .48 h after hospital admission. Patients with coagulase-negative staphylococcus (CoNS) isolates were considered not to have a bloodstream infection. Power calculations and statistical analysis We assumed that the bloodstream infection incidence would increase 5% in patients with prehospital 25(OH)D concentrations ,20 ng/mL compared with in patients with concentrations $20 ng/mL (29). With an a error level of 5% and a power of 80%, the minimum sample size required for our primary endpoint (HABSI) was 1414 total patients. Categorical variables were described by frequency distribution and compared across 25(OH)D groups by using contingency tables and chi-square testing. Continuous variables were examined graphically (eg, by using a histogram or box plot) and in terms of summary statistics (means 6 SDs, medians, or IQRs)

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

We abstracted laboratory and administrative data from the electronic medical records of the of the following 2 teaching hospitals in Boston, MA: Brigham and Women’s Hospital (BWH), with 777 beds and Massachusetts General Hospital (MGH) with 902 beds. The 2 hospitals provide primary as well as tertiary care to an ethnically and socioeconomically diverse population within eastern Massachusetts and the surrounding region.

953

954

QURAISHI ET AL

RESULTS

Characteristics of the study population are shown in Table 1. Most patients were women (55%), were white (80%), and had medically related DRGs (69%). The mean (6SD) age at hospital admission was 61 6 18 y. Eight percent of patients had positive blood cultures. The mean 25(OH)D concentration was 25 6 17 ng/mL; 58% of 25(OH)D measurements had occurred in the 3 mo before hospital admission. There was a nonsignificant effect of season on 25(OH)D concentrations (F[3] = 2.37, P = 0.068). The percentage of patients with HABSI did not differ by the season of hospital admission [chi-square(df = 3; n = 2135) = 5.54, P = 0.136]. The variability in 25(OH)D measurements was present because of the different assays used over the study period. In the cohort, 55% of 25(OH)D concentrations $30 ng/mL were measured via LC-MS, whereas 50% of 25(OH)D concentrations ,10 ng/mL were measured via a chemiluminescence assay. Patient characteristics of the study cohort were stratified according to prehospital 25(OH)D concentrations (Table 2). Factors that significantly differed between stratified groups included age, sex, and race. As indicated in Table 3, the 25(OH)D concentration, age, sex, and patient type were significant predictors of HABSI. The in-hospital mortality rate was 6%. Thirty- and 90-d mortality rates were 7% and 14%, respectively. Primary outcome A prehospital 25(OH)D concentration ,10 ng/mL was a strong predictor of HABSI (Table 4). The odds of HABSI in patients with 25(OH)D concentrations ,10 ng/mL were 2.3-fold those in patients with concentrations $30 ng/mL. A 25(OH)D concentration ,10 ng/mL remained a significant predictor of the odds of HABSI after adjustment for age, sex, race, DeyoCharlson index, and patient type. Adjusted odds of HABSI in the group of patients with 25(OH)D concentrations ,10 ng/mL were 2.0-fold those in patients with concentrations $30 ng/mL (Table 4). Additional adjustment for assay type (chemiluminescence, radioimmunoassay, or LC-MS) did not materially alter point estimates; the group with 25(OH)D concentrations ,10 ng/mL had an OR of HABSI of 1.91 (95% CI: 1.17, 3.13) relative to

that in the groups with 25(OH)D concentrations $30 ng/mL after adjustment for age, sex, race (nonwhite compared with white), patient type (medical compared with surgical), DeyoCharlson index, and assay type. Although limited by the statistical power, results did not materially differ by hospital site. Subanalyses Although the statistical power was compromised, when patients who had vitamin D status measured .90 d before hospital admission were excluded (n = 1238), the odds of HABSI in patients with 25(OH)D concentrations ,10 ng/mL were 2.18 (95% CI: 1.20, 3.93) relative to those in patients with concentrations $30 ng/mL. The multivariable adjusted odds of HABSI in patients with 25(OH)D concentrations ,10 ng/mL was 1.87 (95% CI: 1.03, 3.42) relative to those in patients with concentrations $30 ng/mL. As per the recommendation of the Institute of Medicine (20) that a 25(OH)D concentration ,20 ng/mL should be the cutoff for vitamin D inadequacy, we also analyzed the full cohort with the exposure of interest as prehospital serum 25(OH)D concentration categorized as ,20 or $20 ng/mL. The odds of HABSI in patients with prehospital 25(OH)D concentrations ,20 ng/mL was 1.8-fold those in patients with concentrations $20 ng/mL (Table 4). After multivariable adjustment per our primary analysis, the OR of HABSI in patients with prehospital 25(OH)D concentrations ,20 ng/mL was 1.7-fold that of those with concentrations $20 ng/mL (Table 4). Another sensitivity analysis was performed to evaluate the data relative to the 25(OH)D assay used. Although the statistical power was compromised, we showed that patients with 25(OH)D TABLE 1 Baseline demographic characteristics of the study population1 Characteristic Cases (total n) Age (y) Sex [n (%)] F M Race [n (%)] Nonwhite White Patient type [n (%)] Medical Surgical White blood cells [n (%)] 0–3.9 3 103/mL 4.0–9.9 3 103/mL $10 3 103/mL 25(OH)D (ng/mL) 25(OH)D by season (ng/mL) Fall (n = 529) Spring (n = 563) Summer (n = 525) Winter (n = 518) Bloodstream infection [n (%)] Absent Present 1 2

25(OH)D, 25-hydroxyvitamin D. Mean 6 SD (all such values).

Value 2135 61 6 182 1171 (55) 964 (45) 435 (20) 1700 (80) 1475 (69) 660 (31) 168 (8) 858 (40) 1109 (52) 25 6 17 25 25 26 24

6 6 6 6

15 21 16 15

1967 (92) 168 (8)

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

and compared across exposure groups by using ANOVA. The outcome considered was HABSI. Unadjusted associations between 25(OH)D groups and HABSI were estimated by using bivariable logistic regression models. Adjusted ORs were estimated by using multivariable logistic regression models with the inclusion of covariate terms thought to plausibly be associated with both the 25(OH)D concentration and HABSI. For the primary model (HABSI), the specification of each continuous covariate (as linear compared with categorical terms) was adjudicated by the empiric association with the primary outcome by using Akaike’s information criterion; the overall model fit was assessed by using the Hosmer-Lemeshow test. Models for secondary analyses were specified identically to the primary model. Locally weighted scatter plot smoothing was used to graphically represent (30, 31) the relation between the prehospital 25(OH)D concentration and risk of HABSI. All P values were 2-tailed; values ,0.05 were considered statistically significant. All analyses were performed with STATA 12.0MP statistical software (StataCorp LP).

955

VITAMIN D DEFICIENCY AND HABSI TABLE 2 Patient characteristics by prehospital vitamin D status1 Prehospital 25(OH)D2 10–19.9 ng/mL

20–29.9 ng/mL

$30 ng/mL

P

323 59 6 183

667 60 6 17

462 60 6 18

683 64 6 17

— ,0.0001* ,0.0001

177 (55) 146 (45)

325 (49) 342 (51)

238 (52) 224 (48)

431 (63) 252 (37)

100 (31) 223 (69)

147 (22) 520 (78)

85 (18) 377 (82)

103 (15) 580 (85)

238 (74) 85 (26)

471 (71) 196 (29)

309 (67) 153 (33)

457 (67) 226 (33)

61 (19) 71 (22) 191 (59)

155 (23) 158 (24) 354 (53)

120 (26) 114 (25) 228 (49)

166 (24) 176 (26) 361 (53)

27 (8) 129 (40) 167 (52)

63 (9) 262 (40) 342 (51)

30 (6) 193 (42) 239 (52)

48 (7) 274 (40) 361 (53)

283 (88) 40 (12)

608 (91) 59 (9)

433 (94) 29 (6)

643 (94) 40 (6)

,0.0001

0.09

0.11

0.61

0.002

1

P values were determined by using the chi-square test unless otherwise designated. *Determined by using the Kruskal-Wallis test. 2 25(OH)D, 25-hydroxyvitamin D. 3 Mean 6 SD (all such values).

concentrations ,10 ng/mL had consistently higher (but nonsignificant) odds of HABSI than those in patients with concentrations $30 ng/mL, regardless of the assay used [chemiluminescence assay OR: 1.84 (95% CI: 0.73, 4.66); radioimmunoassay OR: 1.53 (95% CI: 0.63, 3.70); and LC-MS OR: 2.22 (95% CI: 0.91, 5.47); all fully adjusted]. To assess the discrimination of 25(OH)D concentrations for HABSI, we used a receiver operating characteristic curve analysis and determined the AUC. The estimation of the AUC showed that the 25(OH)D concentration had a similar discriminative power for HABSI regardless of the assay used [chemiluminescence assay AUC: 0.57 (95% CI: 0.50, 0.64); radioimmunoassay AUC: 0.63 (95% CI: 0.55, 0.71); and LC-MS AUC: 0.55(95% CI: 0.46, 0.64)]. When the 25(OH)D concentration was expressed as a continuous variable and adjusted for age, sex, race, Deyo-Charlson index, and patient type, the odds for HABSI were 0.98 (95% CI: 0.97, 0.99). This result was multiplicative, and thus, a 10-ng/mL increase would have an OR of 0.9810 effect (= 0.82 or a HABSI odds reduction of 18%). The locally weighted scatter plot smoothing plot (Figure 1) showed a near-inverse linear association between vitamin D status and risk of HABSI to 25(OH)D concentrations near 20 ng/mL. Between 25(OH)D concentrations of 20 and 40 ng/mL, there was a progressive flattening of the curve. For 25(OH)D concentrations .40 ng/mL, the curve appeared flat. Finally, in the parent cohort of 20,837 inpatients with serum 25(OH)D measured between 7 and 365 d before hospital admission who did not receive critical care, the proportion of patients with blood cultures drawn was highest in the group with prehospital 25(OH)D concentrations ,10 ng/mL [16%, 12%, 10%, and 8% in patients with 25(OH)D concentrations ,10,

10–19.9, 20–29.9, and $30 ng/mL, respectively [chi-square(df = 3, n = 20,837) = 154; P , 0.001]. With the assumption that vitamin D status is not a factor in the decision to order a blood culture, TABLE 3 Multivariable-adjusted associations between covariates and hospitalacquired bloodstream infections1

Age (/1 y) Sex F M Race Nonwhite White Patient type Medical Surgical Deyo-Charlson index 0–3 4–6 $7 Prehospital 25(OH)D2 ,10 ng/mL 10–19.9 ng/mL 20–29.9 ng/mL $30 ng/mL 1

OR (95% CI)

P

0.99 (0.98, 1.00)

0.021

1.00 (reference) 1.51 (1.09, 2.09)

0.013

1.00 (reference) 0.89 (0.60, 1.30)

0.53

1.00 (reference) 0.55 (0.37, 0.81)

0.003

1.00 (reference) 1.53 (0.93, 2.52) 1.45 (0.92, 2.28) 1.95 1.36 0.98 1.00

(1.22, 3.12) (0.89, 2.07) (0.60, 1.61) (reference)

0.097 0.11 0.005 0.16 0.95

Adjusted ORs were estimated by using a multivariable logistic regression model with the inclusion of covariate terms thought to plausibly be associated with vitamin D status and hospital-acquired bloodstream infections. Estimates for each variable were adjusted for all other variables in the table. 2 25(OH)D, 25-hydroxyvitamin D.

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

Cases (n) Age (y) Sex [n (%)] F M Race [n (%)] Nonwhite White Patient type [n (%)] Medical Surgical Deyo-Charlson index [n (%)] 0–3 4–6 $7 White blood cells [n (%)] 0–3.9 3 103/mL 4.0–9.9 3 103/mL .10 3 103/mL Bloodstream infection [n (%)] Absent Present

,10 ng/mL

956

QURAISHI ET AL

TABLE 4 Unadjusted and adjusted associations between prehospital vitamin D status and hospital-acquired bloodstream infections in patients with confirmed blood culture results (n = 2135)1 Prehospital 25(OH)D2

Unadjusted OR (95%CI) P Adjusted OR (95%CI) P

,10 ng/mL

10–19.9 ng/mL

20–29.9 ng/mL

$30 ng/mL

,20 ng/mL

$20 ng/mL

2.33 (1.45, 3.74) ,0.0001

1.60 (1.04, 2.46) 0.032

1.13 (0.69, 1.84) 0.62

1.00 (referent) —

1.84 (1.34, 2.53) ,0.0001

1.00 (referent) —

1.95 (1.22, 3.12) 0.005

1.36 (0.89, 2.07) 0.16

0.98 (0.60, 1.62) 0.95

1.00 (referent) —

1.68 (1.22, 2.33) 0.002

1.00 (referent) —

1 Unadjusted associations between 25(OH)D groups and hospital-acquired bloodstream infections were estimated by using bivariable logistic regression models. Adjusted ORs were estimated by using multivariable logistic regression models with the inclusion of covariate terms thought to plausibly be associated with both 25(OH)D concentrations and hospital-acquired bloodstream infections. Estimates were adjusted for age, sex, race (white compared with nonwhite), patient type (surgical compared with medical), and the Deyo-Charlson index. 2 25(OH)D, 25-hydroxyvitamin D.

prehospital 25(OH)D concentrations ,20 ng/mL were 2.3-fold those in patients with concentrations $20 ng/mL (Table 5). Effect modification Analyses that were based on fully adjusted models were performed to evaluate the 25(OH)D-HABSI association, and the P-interaction was determined to explore any evidence of an effect modification. We individually tested for the effect modification by creatinine, blood urea nitrogen, white blood count, assay type, and season of 25(OH)D draw by adding an interaction term to the multivariate models. None of these variables emerged as an effect modifier of the association between 25(OH)D and HABSI (P-interaction: HABSI . 0.20 for all variables tested). An additional effect-modification analysis showed that the association between 25(OH)D and HABSI was not modified by serum calcium .10.5 mg/dL (P-interaction: HABSI = 0.50). Additional effect modification analysis showed that the association between 25(OH)D and HABSI was not modified by the hospital that provided care (P-interaction: HABSI = 0.14) or the time

FIGURE 1. Vitamin D status compared with risk of hospital-acquired bloodstream infections. Locally weighted scatter plot smoothing was used to represent the near inverse linear association between prehospital 25(OH)D concentration and risk of HABSI. With bandwidth variable = 0.99, 2114 cohort patients were used to construct the curve. HABSI, hospital-acquired bloodstream infection; 25(OH)D, 25-hydroxyvitamin D.

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

and it is unlikely that patients who did not have any blood cultures drawn were likely to be true negatives and not false negatives, we performed a subanalysis on 20,837 inpatients (2135 patients from the study cohort and 18,702 patients who did not have blood cultures drawn .48 h after hospital admission) to determine the association between vitamin D status and bloodstream infection. In this subanalysis, a prehospital 25(OH)D concentration ,20 ng/mL was associated with an increased odds of HABSI (Table 5). The odds of HABSI in groups with 25(OH)D concentrations ,10 and 10–19.9 ng/mL were 4.4- and 2.4-fold, respectively, those in the group with 25(OH)D concentrations $30 ng/mL. The multivariable-adjusted odds of HABSI in groups with 25(OH)D concentrations ,10 and 10–19.9 ng/mL were 3.3- and 1.9-fold, respectively, those in the group with 25(OH)D concentrations $30 ng/mL. In this subanalysis, we also categorized the exposure of interest as prehospital serum 25(OH)D concentrations ,20 or $20 ng/mL. The odds of HABSI in patients with prehospital 25(OH)D concentrations ,20 ng/mL were 2.8-fold those in patients with concentrations $20 ng/mL. The multivariable adjusted odds of HABSI in patients with

957

VITAMIN D DEFICIENCY AND HABSI

DISCUSSION

In this study, we investigated whether prehospital vitamin D status was associated with risk of HABSI. Our data showed that prehospital 25(OH)D concentrations ,10 ng/mL were associated with a significant increase in odds of HABSI in hospitalized adults. The finding remained on multivariable analyses and suggested that vitamin D supplementation may provide a novel approach to lowering HABSI risk. However, because our study was observational and not interventional, a causal relation between vitamin D status and outcomes could not be inferred from these data alone. Our observation of increased odds of HABSI with 25(OH)D concentrations ,10 ng/mL may have been related to a depressed innate immunity. Vitamin D receptors are expressed in cells of the immune system, such as T cells, activated B cells, and dendritic cells (32–36). Low 25(OH)D concentrations are associated with depressed macrophage phagocytosis, chemotaxis, and proinflammatory cytokine production (37). Vitamin D links Tolllike receptor activation and innate immunity to upregulate the

expression of antimicrobial peptides cathelicidin and b-defensin (38–40). In humans, cathelicidin has been shown to have potent activity against bacteria, viruses, fungi, and mycobacteria. In addition, vitamin D is important for the interferon-g–dependent T cell response to infection (41). In murine sepsis, vitamin D supplementation improves coagulation variables and inhibits endotoxemia (42–44). The convergence of mechanistic evidence of vitamin D–mediated immunodulation and our clinical data on risk of HABSI in patients with 25(OH)D concentrations ,10 ng/mL raised a number of questions that merit additional investigation. The current study had potential limitations. Although cohort studies are not the highest level of clinical evidence, they can illustrate the existence or absence of a true effect and direct future research (45). Observational studies may be limited by potential confounding, reverse causation, and the lack of a randomly distributed exposure. An ascertainment bias may exist because the patient cohort under study had vitamin D status measured for unknown reasons that may have been absent in other hospitalized patients. These differences may have decreased the generalizability of our results to all hospitalized patients. Despite adjustment for multiple potential confounders, there may still have been residual confounding that contributed to observed differences in outcomes. Specifically, low 25(OH)D concentrations may have been a marker for the general condition of the patient, for which we were unable to fully adjust. In addition, we were not able to adjust for use of glucocorticoids, which may have adversely affected vitamin D status (46) and may have increased patient susceptibility to infection (47). We were also unable to adjust for sun exposure, immobilization, excessive alcohol intake, smoking status, or genetic factors (48) that could have altered 25(OH)D concentrations. An additional potential limitation was that 25(OH)D concentrations were not drawn at the time of hospital admission. It has been previously shown that the intraperson Pearson’s correlation coefficient for 25(OH)D in outpatients after adjustments for age, race, and season is 0.70 at 3 y between blood draws (49). The 25(OH)D-HABSI association appears to be preserved when 25(OH)D is obtained within 90 d of admission. Despite this observation, 25(OH)D concentrations at the time of hospitalization may be different from when prehospital values were

TABLE 5 Unadjusted and adjusted associations between prehospital vitamin D status and hospital-acquired bloodstream infections in patients with confirmed blood culture results relative to all hospitalized patients (n = 20,837)1 Prehospital 25(OH)D2

Unadjusted OR (95%CI) P Adjusted OR (95%CI) P

,10 ng/mL

10–19.9 ng/mL

20–29.9 ng/mL

$30 ng/mL

,20 ng/mL

$20 ng/mL

4.40 (2.83, 6.84) ,0.0001

2.35 (1.57, 3.51) ,0.0001

1.29 (0.80, 2.08) 0.30

1.00 (referent) —

2.81 (2.06, 3.82) ,0.0001

1.00 (referent) —

3.25 (2.07, 5.11) ,0.0001

1.91 (1.27, 2.88) 0.002

1.17 (0.72, 1.90) 0.52

1.00 (referent) —

2.29 (1.67, 3.14) ,0.0001

1.00 (referent) —

1 Unadjusted associations between 25(OH)D groups and hospital-acquired bloodstream infections were estimated by using bivariable logistic regression models. Adjusted ORs were estimated by using multivariable logistic regression models with the inclusion of covariate terms thought to plausibly be associated with both 25(OH)D concentrations and hospital-acquired bloodstream infections. Estimates were adjusted for age, sex, race (white compared with nonwhite), patient type (surgical compared with medical), and the Deyo-Charlson index. Patients with positive bloodstream infection were analyzed relative to all patients with either negative blood cultures or patients without cultures drawn. 2 25(OH)D, 25-hydroxyvitamin D.

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

between 25(OH)D draw and hospital admission (P-interaction: HABSI = 0.16). In addition, individually running the adjusted model with and without terms for assay type, creatinine, blood urea nitrogen, white blood count, calcium, hematocrit, and the time between 25(OH)D draw and hospital admission, the HABSI estimates in each case were similar. The odds of HABSI in the group with 25(OH)D concentrations ,10 ng/mL were 1.91 (95% CI: 1.17, 3.13) with additional adjustment for assay type; there was an OR of 1.90 (95% CI: 1.18, 3.04) with the time between 25(OH)D draw and admission, an OR of 1.95 (95% CI: 1.22, 3.12) with calcium, and OR of 1.99 (95% CI: 1.24,v3.21) with creatinine, an OR of 1.95 (95% CI: 1.22, 3 .13) with the white blood cell count, and ad OR of 1.98 (95% CI: 1.24, 3.18) with season, all relative to the group with 25(OH)D concentrations $30 ng/mL. These results indicated that the 25(OH)D-HABSI relation was not materially confounded by assay type, year of hospital admission, creatinine, blood urea nitrogen, white blood count, hematocrit, or time between 25(OH)D draw and hospital admission.

958

QURAISHI ET AL We dedicate this article to the memory of our dear friend and colleague Nathan Edward Hellman. The authors’ responsibilities were as follows—SAQ and KBC: jointly conceived the study and designed and implemented the analysis: CAC, AAL, and EG: assisted with the design and implemented the analysis; KBC: assembled input data, wrote code, ran the model, and analyzed output data; SAQ, CAC, and KBC: wrote the manuscript; and all authors: edited the manuscript and provided conceptual advice. None of the authors had a conflict of interest.

REFERENCES 1. Al-Rawajfah OM, Stetzer F, Hewitt JB. Incidence of and risk factors for nosocomial bloodstream infections in adults in the United States, 2003. Infect Control Hosp Epidemiol 2009;30:1036–44. 2. Klevens RM, Edwards JR, Richards CL Jr, Horan TC, Gaynes RP, Pollock DA, Cardo DM. Estimating health care-associated infections and deaths in U.S. hospitals, 2002. Public Health Rep 2007;122:160–6. 3. Scott R. The direct medical costs of healthcare-associated infections in US hospitals and the benefits of prevention. Atlanta, GA: Centers for Disease Control and Prevention, 2009. 4. Roberts RR, Scott RD 2nd, Hota B, Kampe LM, Abbasi F, Schabowski S, Ahmad I, Ciavarella GG, Cordell R, Solomon SL, et al. Costs attributable to healthcare-acquired infection in hospitalized adults and a comparison of economic methods. Med Care 2010;48:1026–35. 5. Flanagan ME, Welsh CA, Kiess C, Hoke S, Doebbeling BN. A national collaborative for reducing health careassociated infections: current initiatives, challenges, and opportunities. Am J Infect Control 2011;39: 685–9. 6. Centers for Disease Control and Prevention (CDC). Vital signs: central line-associated blood stream infections–United States, 2001, 2008, and 2009. MMWR Morb Mortal Wkly Rep 2011;60:243–8. 7. Al Mohajer M, Darouiche RO. Sepsis syndrome, bloodstream infections, and device-related infections. Med Clin North Am 2012;96: 1203–23. 8. Adams JS, Hewison M. Unexpected actions of vitamin D: new perspectives on the regulation of innate and adaptive immunity. Nat Clin Pract Endocrinol Metab 2008;4:80–90. 9. Youssef DA, Ranasinghe T, Grant WB, Peiris AN. Vitamin D’s potential to reduce the risk of hospital-acquired infections. Dermatoendocrinol 2012;4:167–75. 10. Kuriacose R, Olive KE. Prevalence of vitamin D deficiency and insufficiency in northeast Tennessee. South Med J 2008;101:906–9. 11. Pellicane AJ, Wysocki NM, Schnitzer TJ. Prevalence of 25-hydroxyvitamin D deficiency in the outpatient rehabilitation population. Am J Phys Med Rehabil 2010;89:899–904. 12. Fish E, Beverstein G, Olson D, Reinhardt S, Garren M, Gould J. Vitamin D status of morbidly obese bariatric surgery patients. J Surg Res 2010;164:198–202. 13. Bogunovic L, Kim AD, Beamer BS, Nguyen J, Lane JM. Hypovitaminosis D in patients scheduled to undergo orthopaedic surgery: a single-center analysis. J Bone Joint Surg Am 2010;92:2300–4. 14. Braun A, Chang D, Mahadevappa K, Gibbons FK, Liu Y, Giovannucci E, Christopher KB. Association of low serum 25-hydroxyvitamin D levels and mortality in the critically ill. Crit Care Med 2011;39:671–7. 15. Hivert MF, Grant RW, Shrader P, Meigs JB. Identifying primary care patients at risk for future diabetes and cardiovascular disease using electronic health records. BMC Health Serv Res 2009;9:170. 16. Hug BL, Lipsitz SR, Seger DL, Karson AS, Wright SC, Bates DW. Mortality and drug exposure in a 5-year cohort of patients with chronic liver disease. Swiss Med Wkly 2009;139:737–46. 17. Linder JA, Bates DW, Williams DH, Connolly MA, Middleton B. Acute infections in primary care: accuracy of electronic diagnoses and electronic antibiotic prescribing. J Am Med Inform Assoc 2006;13:61–6. 18. Zager S, Mendu ML, Chang D, Bazick HS, Braun AB, Gibbons FK, Christopher KB. Neighborhood poverty rate and mortality in patients receiving critical care in the academic medical center setting. Chest 2011;139:1368–79. 19. Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney RP, Murad MH, Weaver CM. Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2011;96:1911–30.

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

drawn. It is not clear if illness severity influences 25(OH)D concentrations at the time of hospital admission (50). Measurements of vitamin D status before hospital admission may allow inferences to be made regarding the potential importance of 25(OH)D concentrations $30 ng/mL before the onset of illness. Despite the exclusion of patients with known high-dose vitamin D supplementation, we do not know if patients received vitamin D or multivitamins including vitamin A, which may have antagonized the vitamin D action (51). In addition, we do not know if patients who had 25(OH)D concentrations determined by different assays had differences in treatment outside our health system. These issues will need to be addressed by other groups as they try to replicate and extend our findings. Last, we note that CoNS isolates were not considered to be bloodstream infections because CoNS are common skin commensals and are frequent contaminants of blood cultures (52). True CoNS bacteremia has an increasing incidence, is associated with mortality and morbidity, and has gained greater importance as a pathogen (53). Although the exclusion of CoNS isolates may have underestimated the incidence of true bloodstream infections, the analysis of our study cohort with the inclusion of all CoNS isolates did not materially alter the 25(OH)D-HABSI association (data not shown). In addition, indwelling central venous catheters are important risk factors for HABSI (54). However, because of limitations of the data, we were unable to determine which patients had central venous catheters or the site from which the blood cultures may have been obtained. The current study has several strengths. For example, our study had sufficient numbers of patients to ensure the adequate reliability of our estimates (n = 2135; HABSI rate = 8%). We had sufficient statistical power to detect a clinically relevant difference in HABSI if one existed. And although the Deyo-Charlson index accounts for chronic conditions such as diabetes and chronic renal failure that may alter immune function (55), it is not clear if illness severity influences vitamin D status at the time of hospital admission (20, 56). However, the preservation of the 25(OH)D-HABSI association after the exclusion of patients with 25(OH)D drawn .90 d before hospitalization and the absence of effect modification or confounding relative to the time between the 25(OH)D draw and hospital admission suggested that the likelihood of reverse causation was low. In conclusion, these data show that prehospital 25(OH)D concentrations ,10 ng/mL are associated with increased odds of HABSI in a large cohort of hospitalized patients. Our work supports the hypothesis that better vitamin D status may play an important protective role against nosocomial infections (9). Although Vitamin D supplementation may decrease the incidence of acute respiratory tract infections in ambulatory settings (57), we believe that our results provide clinical evidence of a potential link between vitamin D and outcomes of hospitalized patients. We illustrate the possibility that even small changes in vitamin D status may affect HABSI. Despite our observations, the supplementation of vitamin D in the hospitalized adult cannot currently be advocated solely for HABSI. Longitudinal studies are required to confirm our findings and to further investigate the mechanisms underlying these observations. If confirmed, randomized, placebo-controlled trials will be needed to determine whether vitamin D–supplementation therapy before or at the time of hospital admission might have a benefit in improving outcomes in hospitalized patients.

VITAMIN D DEFICIENCY AND HABSI

40. Wang TT, Nestel FP, Bourdeau V, Nagai Y, Wang Q, Liao J, TaveraMendoza L, Lin R, Hanrahan JW, Mader S, et al. Cutting edge: 1,25dihydroxyvitamin D3 is a direct inducer of antimicrobial peptide gene expression. J Immunol 2004;173:2909–12. (Published erratum appears in J Immunol 2004;173:following 6489.) 41. Fabri M, Stenger S, Shin DM, Yuk JM, Liu PT, Realegeno S, Lee HM, Krutzik SR, Schenk M, Sieling PA, et al. Vitamin D is required for IFN-gamma-mediated antimicrobial activity of human macrophages. Sci Transl Med 2011;3:104ra2. 42. Asakura H, Aoshima K, Suga Y, Yamazaki M, Morishita E, Saito M, Miyamoto K, Nakao S. Beneficial effect of the active form of vitamin D3 against LPS-induced DIC but not against tissue-factor-induced DIC in rat models. Thromb Haemost 2001;85:287–90. 43. Møller S, Laigaard F, Olgaard K, Hemmingsen C. Effect of 1,25dihydroxy-vitamin D3 in experimental sepsis. Int J Med Sci 2007;4: 190–5. 44. Horiuchi H, Nagata I, Komoriya K. Protective effect of vitamin D3 analogues on endotoxin shock in mice. Agents Actions 1991;33:343–8. 45. Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med 2000;342:1887–92. 46. Skversky AL, Kumar J, Abramowitz MK, Kaskel FJ, Melamed ML. Association of glucocorticoid use and low 25-hydroxyvitamin D levels: results from the National Health and Nutrition Examination Survey (NHANES): 2001-2006. J Clin Endocrinol Metab 2011;96:3838–45. 47. Cutolo M, Seriolo B, Pizzorni C, Secchi ME, Soldano S, Paolino S, Montagna P, Sulli A. Use of glucocorticoids and risk of infections. Autoimmun Rev 2008;8:153–5. 48. Wang TJ, Zhang F, Richards JB, Kestenbaum B, van Meurs JB, Berry D, Kiel DP, Streeten EA, Ohlsson C, Koller DL, et al. Common genetic determinants of vitamin D insufficiency: a genome-wide association study. Lancet 2010;376:180–8. 49. Platz EA, Leitzmann MF, Hollis BW, Willett WC, Giovannucci E. Plasma 1,25-dihydroxy- and 25-hydroxyvitamin D and subsequent risk of prostate cancer. Cancer Causes Control 2004;15:255–65. 50. Krishnan A, Ochola J, Mundy J, Jones M, Kruger P, Duncan E, Venkatesh B. Acute fluid shifts influence the assessment of serum vitamin D status in critically ill patients. Crit Care 2010;14:R216. 51. Jenab M, Bueno-de-Mesquita HB, Ferrari P, van Duijnhoven FJ, Norat T, Pischon T, Jansen EH, Slimani N, Byrnes G, Rinaldi S, et al. Association between pre-diagnostic circulating vitamin D concentration and risk of colorectal cancer in European populations:a nested casecontrol study. BMJ 2010;340:b5500. 52. Tokars JI. Predictive value of blood cultures positive for coagulasenegative staphylococci: implications for patient care and health care quality assurance. Clin Infect Dis 2004;39:333–41. 53. Diekema DJ, Pfaller MA, Schmitz FJ, Smayevsky J, Bell J, Jones RN, Beach M. Survey of infections due to Staphylococcus species: frequency of occurrence and antimicrobial susceptibility of isolates collected in the United States, Canada, Latin America, Europe, and the Western Pacific region for the SENTRY Antimicrobial Surveillance Program, 1997-1999. Clin Infect Dis 2001;32(suppl 2):S114–32. 54. Wenzel RP, Edmond MB. The impact of hospital-acquired bloodstream infections. Emerg Infect Dis 2001;7:174–7. 55. Esper AM, Moss M, Lewis CA, Nisbet R, Mannino DM, Martin GS. The role of infection and comorbidity: Factors that influence disparities in sepsis. Crit Care Med 2006;34:2576–82. 56. 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;121:452–5. 57. Quraishi SA, Camargo CA Jr. Vitamin D in acute stress and critical illness. Curr Opin Clin Nutr Metab Care 2012;15:625–34.

Downloaded from ajcn.nutrition.org at KAINAN UNIVERSITY on May 3, 2015

20. Ross AC, Manson JE, Abrams SA, Aloia JF, Brannon PM, Clinton SK, Durazo-Arvizu RA, Gallagher JC, Gallo RL, Jones G, 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;96:53–8. 21. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. 22. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, Saunders LD, Beck CA, Feasby TE, Ghali WA. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130–9. 23. Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, Januel JM, Sundararajan V. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol 2011;173:676–82. 24. Rapoport J, Gehlbach S, Lemeshow S, Teres D. Resource utilization among intensive care patients. Managed care vs traditional insurance. Arch Intern Med 1992;152:2207–12. 25. van den Ouweland JM, Beijers AM, Demacker PN, van Daal H. Measurement of 25-OH-vitamin D in human serum using liquid chromatography tandem-mass spectrometry with comparison to radioimmunoassay and automated immunoassay. J Chromatogr B Analyt Technol Biomed Life Sci 2010;878:1163–8. 26. Shorr AF, Jackson WL, Kelly KM, Fu M, Kollef MH. Transfusion practice and blood stream infections in critically ill patients. Chest 2005;127:1722–8. 27. Blaine KP, Tuohy MJ, Wilson D, Procop GW, Tisch DJ, Shrestha NK, Hall GS. Progression to bacteremia in critical care patients colonized with methicillin-resistant Staphylococcus aureus expressing PantonValentine leukocidin. Diagn Microbiol Infect Dis 2010;68:28–33. 28. Miller RS, Norris PR, Jenkins JM, Talbot TR 3rd, Starmer JM, Hutchison SA, Carr DS, Kleymeer CJ, Morris JA Jr. Systems initiatives reduce healthcare-associated infections: a study of 22,928 device days in a single trauma unit. J Trauma 2010;68:23–31. 29. Lange N, Litonjua AA, Gibbons FK, Giovannucci E, Christopher KB. Pre-hospital vitamin D concentration, mortality and bloodstream infection in a hospitalized patient population. Am J Med 2013;26:640. e19–27. 30. Cleveland W. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 1979;74:829–36. 31. Cleveland W, Devlin S. Locally-weighted regression: an approach to regression analysis by local fitting. J Am Stat Assoc 1988;83:596–610. 32. Dickson I. New approaches to vitamin D. Nature 1987;325:18. 33. Minghetti PP, Norman AW. 1,25(OH)2-vitamin D3 receptors: gene regulation and genetic circuitry. FASEB J 1988;2:3043–53. 34. Mahon BD, Wittke A, Weaver V, Cantorna MT. The targets of vitamin D depend on the differentiation and activation status of CD4 positive T cells. J Cell Biochem 2003;89:922–32. 35. Heine G, Anton K, Henz BM, Worm M. 1alpha,25-dihydroxyvitamin D3 inhibits anti-CD40 plus IL-4-mediated IgE production in vitro. Eur J Immunol 2002;32:3395–404. 36. Adorini L, Penna G, Giarratana N, Roncari A, Amuchastegui S, Daniel KC, Uskokovic M. Dendritic cells as key targets for immunomodulation by vitamin D receptor ligands. J Steroid Biochem Mol Biol 2004;89-90:437–41. 37. Kankova M, Luini W, Pedrazzoni M, Riganti F, Sironi M, Bottazzi B, Mantovani A, Vecchi A. Impairment of cytokine production in mice fed a vitamin D3-deficient diet. Immunology 1991;73:466–71. 38. Liu PT, Stenger S, Li H, Wenzel L, Tan BH, Krutzik SR, Ochoa MT, Schauber J, Wu K, Meinken C, et al. Toll-like receptor triggering of a vitamin D-mediated human antimicrobial response. Science 2006; 311:1770–3. 39. Holick MF. Vitamin D deficiency. N Engl J Med 2007;357:266–81.

959

Association between prehospital vitamin D status and hospital-acquired bloodstream infections.

Alterations in immune function can predispose patients to nosocomial infections. Few studies have explored potentially modifiable host factors that ma...
180KB Sizes 0 Downloads 0 Views