BRIEF REPORTS

Vitamin D Insufficiency and Sepsis Severityin Emergency Department Patients WithSuspected InfectionAdit A. Ginde, MD, MPH, Carlos A. Camargo, Jr., MD, DrPH, and Nathan I. Shapiro, MD, MPH

AbstractObjectives: Vitamin D is increasingly recognized as an important mediator of immune function and mayhave a preventive role in the pathogenesis of sepsis. We sought to evaluate the association between vita-min D status and sepsis severity and hypothesized that vitamin D insufficiency would be associated withincreased sepsis severity.

Methods: This was a pilot study of emergency department (ED) patients age ‡18 years evaluated for sus-pected infection at an urban, teaching hospital. The authors measured illness severity using the followingassessments at baseline and 24 hours: 1) severe sepsis, defined as suspected infection plus two or moreelements of systemic inflammatory response syndrome criteria and acute dysfunction of one or moreorgan systems; 2) Acute Physiology Age Chronic Health Evaluation (APACHE) II scores; and 3) Sepsis-related Organ Failure Assessment (SOFA) scores. Vitamin D insufficiency was defined as baseline serum25-hydroxyvitamin D (25OHD) levels <75 nmol ⁄ L.

Results: Eighty-one patients were enrolled, with a median age of 62 years (interquartile range[IQR] = 48–76 years), 47% were female, and 77% were white. At baseline, 64 (79%) had 25OHD levels of<75 nmol ⁄ L, and 43 (53%) had severe sepsis. At 24 hours, 48 (59%) had severe sepsis. Patients with base-line 25OHD levels of <75 nmol ⁄ L, compared to patients with 25OHD levels of ‡75 nmol ⁄ L, were morelikely to have severe sepsis (61% vs. 24%; p = 0.006) and SOFA scores ‡2 (44% vs. 18%; p = 0.049). Addi-tionally, at 24 hours, those with 25OHD levels of <75 nmol ⁄ L were more likely to have severe sepsis(67% vs. 29%; p = 0.005), dysfunction of two or more organ systems (50% vs. 18%; p = 0.02), APACHE IIscore of ‡25 (19% vs. 0%; p = 0.06), and SOFA scores of ‡2 (63% vs. 29%; p = 0.02). Additionally, all fourpatients who died during the index hospitalization had 25OHD levels of <75 nmol ⁄ L.

Conclusions: Vitamin D insufficiency was associated with higher sepsis severity in ED patients hospital-ized for suspected infection. Larger observational studies, mechanistic studies, and ultimately random-ized controlled trials are needed to determine causation and to evaluate if vitamin D supplementationcan reduce the risk of sepsis as a preventive or therapeutic strategy.

ACADEMIC EMERGENCY MEDICINE 2011; 18:551–554 ª 2011 by the Society for Academic EmergencyMedicine

W hile vitamin D is best known for its regula-tion of calcium-phosphate homeostasis andeffects on bone health, recent studies sug-

gest that it also regulates innate and adaptive immunefunction, including differentiation and activation of mac-rophages, dendritic cells, and lymphocytes.1 Welldescribed is the ability of 25-hydroxyvitamin D (25OHD),the primary circulating form, to induce its own conver-sion to the active form 1,25-dihydroxyvitamin D (calcitri-ol) and produce cathelicidin antimicrobial peptides in thepresence of an antigen challenge.2 Accordingly, lowserum 25OHD levels have been linked to increased riskof infection,3 and a recent clinical trial demonstrated thatvitamin D supplementation reduced the risk of incidentinfluenza A infection.4

In addition to infection prevention, vitamin D insuffi-ciency may also be important in the pathogenesis ofsepsis, the most fatal consequence of severe infection.In vitro and animal models of sepsis suggest that vita-min D treatment modulates proinflammatory cytokine

ª 2011 by the Society for Academic Emergency Medicine ISSN 1069-6563doi: 10.1111/j.1553-2712.2011.01047.x PII ISSN 1069-6563583 551

From the Department of Emergency Medicine, University ofColorado School of Medicine (AAG), Aurora, CO; the Depart-ment of Emergency Medicine, Massachusetts General Hospital(CAC), and the Department of Emergency Medicine, BethIsrael Deaconess Medical Center (NIS), Harvard MedicalSchool, Boston, MA.Received July 7, 2010; revision received October 5, 2010;accepted October 18, 2010.The authors have no relevant financial information or potentialconflicts of interest to disclose.Supervising Editor: Nina T. Gentile, MD.Address for correspondence: Adit A. Ginde, MD, MPH; e-mail:adit.ginde@ucdenver.edu.

production, deranged coagulation, and activation of thevascular endothelium associated with the sepsis syn-dromes.5,6 Although 77% of the U.S. population hasinsufficient vitamin D (serum 25OHD levels of<75 nmol ⁄ L) to support normal immune function,7 criti-cally ill individuals with sepsis may have even lower25OHD levels, compared to healthy controls.8

Vitamin D supplementation is a simple interventionthat holds the potential to reduce illness severity inpatients with infection. However, the associationbetween vitamin D insufficiency and infection severity isnot known. We hypothesized that among emergencydepartment (ED) patients with suspected infection, thosewith vitamin D insufficiency, defined as serum 25OHDlevels of <75 nmol ⁄ L, would have higher infection sever-ity than those with 25OHD levels of ‡75 nmol ⁄ L.

METHODS

Study DesignThis was a prospective observational pilot study. Thisstudy was approved by our institutional human subjectscommittee and informed consent was obtained.

Study Setting and PopulationThe population was a convenience sample of patientsage ‡18 years presenting to the ED with hospital admis-sion planned for suspected infection. Suspected infec-tion was defined as a clinical suspicion of an infectiousetiology as determined by the treating clinician. Thesetting was Beth Israel Deaconess Medical Center, anurban, teaching hospital with an emergency medicineresidency training program.

Study ProtocolStudy procedures are as previously described.9 Briefly,treating clinicians notified study staff of patients withsuspected infection; patients were then recruited forthe study during their ED stay with clinical and labora-tory measures assessed at two time points: baseline (inthe ED) and 24 (±2) hours later. Patients were classifiedas having ‘‘sepsis’’ if they had suspected infection plustwo or more elements of systemic inflammatoryresponse syndrome criteria and ‘‘severe sepsis’’ if theyhad sepsis plus acute dysfunction of one or more organsystems. Additional markers of illness severity includedthe Acute Physiology Age Chronic Health Evaluation(APACHE) II, and Sepsis-related Organ Failure Assess-ment (SOFA) scores, which are validated and frequentlyused scales to measure severity of medical illness andinfection, respectively.10,11 All assessments were madebased on worst parameters in the ED (for baseline timepoint) or during the first 24 hours (for the 24-hour timepoint). If a patient was discharged before 24 hours or asample was unavailable at 24 hours, the closest assess-ment before that time was carried forward for analysis.Outcome assessors were blinded to vitamin D statusand to the objective of this study.

Serum samples were promptly refrigerated at )4�Cand transferred to )80�C freezers for long-term storagewithin 24 hours. The 25OHD levels were measuredby liquid chromatography–tandem mass spectrometryin the Massachusetts General Hospital laboratory.

The between-run coefficient of variation for a qualitycontrol serum with 25OHD concentration of 57 nmol ⁄ Lis 7.5%. Serum lactate levels were assessed either bythe hospital laboratory or using a point-of-care i-STATsystem (Abbot Lifesciences Point of Care, Princeton,NJ). Activation of the inflammatory cascade wasassessed using established markers of the body’s inflam-matory response (interleukin [IL]-1b and IL-6). These lev-els were measured at baseline using Quantikine enzyme-linked immunosorbent assay kits (R&D Systems, Minne-apolis, MN).

Data AnalysisWe performed statistical analyses using Stata 10.1(StataCorp, College Station, TX) and summarized datausing basic descriptive statistics. Vitamin D status wasdefined by 25OHD levels as insufficient (<75 nmol ⁄ L) orsufficient (‡75 nmol ⁄ L), based on prior analysis of vita-min D and immune responses2 and previous epidemio-logic data.3,7 APACHE II and SOFA scores were alsodichotomized based on established thresholds to facili-tate clinically relevant comparisons. We comparedillness severity by vitamin D status using chi-squaretests, or Fisher’s exact test as appropriate, and continu-ous laboratory measurements by Student’s t-test. Allp-values were two-tailed, with p < 0.05 consideredstatistically significant. For this hypothesis-generatingpilot study, we did not correct for multiple compari-sons. Because age was a large potential confounder ofthe vitamin D sepsis severity association, we adjustedthe primary results for age using standard multivariablelogistic regression, with 25OHD level and age as pre-dictors. We assessed model goodness of fit using theHosmer-Lemeshow test.

RESULTS

We enrolled 81 patients, who were median age 62 years(interquartile range [IQR] = 48–76 years), 47% female,and 77% white. At baseline, 16 (20%) had infectionwithout sepsis, 22 (27%) had sepsis, and 43 (53%) hadsevere sepsis. At 24 hours, 20 (25%) had sepsis and 48(59%) had severe sepsis. The median serum 25OHDlevel was 52 nmol ⁄ L (IQR = 32–71 nmol ⁄ L), with 64(79%) having vitamin D insufficiency defined by 25OHDlevel of <75 nmol ⁄ L.

Demographic and clinical characteristics, stratified byvitamin D status, are presented in Table 1. Vitamin Dinsufficiency was associated with all markers of illnessseverity at the baseline and 24-hour time points, andwith an increase in SOFA score from baseline to24 hours. While vitamin D insufficiency was associatedwith older age, adjusting for age did not materiallychange the direction of any effect sizes (Table 2). Addi-tionally, all four patients who died during the indexhospitalization had 25OHD levels of <75 nmol ⁄ L.

Although the study was underpowered to detectsmaller differences in laboratory parameters statisti-cally, patients with baseline 25OHD levels of<75 nmol ⁄ L, compared to patients with 25OHD levels of‡75 nmol ⁄ L, had higher mean initial lactate levels (2.6vs. 2.2 mmol ⁄ L; p = 0.20), IL-1 levels (3.1 vs. 1.4 pg ⁄ mL;p = 0.19), and IL-6 levels (276 vs. 148 pg ⁄ mL; p = 0.11).

552 Ginde et al. • VITAMIN D AND SEPSIS

DISCUSSION

To our knowledge, this is the first study to evaluatethe association between vitamin D status and sepsisseverity. In a cohort of ED patients with suspectedinfection, vitamin D insufficiency was consistentlyassociated with severe sepsis, APACHE II scores, andSOFA scores. Additionally, vitamin D insufficiency was

associated with worsening clinical course, as measuredby increase in SOFA score from baseline to 24 hours.

Laboratory and animal studies provide support forour hypothesis that vitamin D may have an importantprotective effect in the pathogenesis of sepsis. Forexample, vitamin D pretreatment of human microvesselendothelial cells reduced lipopolysaccharide-inducedproduction of proinflammatory cytokines (IL-6 and IL-8)and inhibited NF-jB activation.5 In the present clinicalstudy, our results also suggest that higher 25OHDlevels may be associated with lower systemic levels ofproinflammatory cytokines, such as IL-1 and IL-6,although these nonsignificant results require furtherstudy. Additionally, Møller and colleagues6 found that1,25-dihydroxyvitamin D3 supplementation improvedsepsis-induced coagulation disturbances in a rat cecalligation and puncture model. Although we did notassess coagulation in this study, these data suggestanother possible mechanism of vitamin D-modulatedprotection against the pathogenesis of sepsis.

Severe sepsis affects more than 750,000 Americansannually, of whom 51% require intensive care unitassistance and 29% die during the index hospitaliza-tion.12 Despite years of research, the treatment ofpatients with severe sepsis remains an important chal-lenge. To date, therapies that inhibit individual compo-nents of the inflammatory or coagulation pathwayshave had little impact on survival. The present resultsadd to the existing literature suggesting that vitamin Dis associated with sepsis severity. Further studies areneeded to determine the role of vitamin D in preventionof infection and mitigation of the inflammatory andcoagulation associated with sepsis.

Vitamin D supplementation, particularly of higherrisk populations such as older adults, holds the poten-tial to lower risk of incident infection and associatedmorbidity such as sepsis. In addition to prevention,vitamin D supplementation in the acute care of patientswith infection has potential as an adjuvant treatment tomodulate inflammatory and coagulation-induced sepsissyndromes. Vitamin D supplementation is a simple andsafe intervention. In the present study, and consistentwith prior epidemiologic data from the U.S. popula-tion,7 79% of patients had 25OHD levels of <75 nmol ⁄ L,below what is likely needed for optimal immune func-tion. These hypotheses require further study throughlarger observational studies to confirm the presentfindings and ultimately randomized controlled trials ofvitamin D supplementation.

LIMITATIONS

As a pilot study, this analysis was exploratory in nature,as our modest sample size was underpowered to per-form robust multivariable analyses or other inferentialanalyses. Although adjusting for age did not materiallychange results, other factors may have confounded theassociations between vitamin D status and illness sever-ity. Larger studies are needed to assess the effects ofother potential confounders, such as season, obesity,and medical comorbidities. This was an observationalstudy, and the reported associations do not provecausal relationships. Specifically, although we are not

Table 1Association of Vitamin D Insufficiency and Infection SeverityOutcomes Among 81 ED Patients Hospitalized for SuspectedInfection

Characteristic

Serum 25OHD Level(nmol ⁄ L)

p-value<75, n (%) ‡75, n (%)

DemographicsAge ‡ 65 years 34 (53) 4 (24) 0.03Female sex 27 (42) 11 (65) 0.10Nonwhite race 14 (22) 5 (29) 0.53

BaselineSevere sepsis 39 (61) 4 (24) 0.006SOFA score ‡ 2 28 (44) 3 (18) 0.049

24 hoursSevere sepsis 43 (67) 5 (29) 0.005Two or more organsystem dysfunction

32 (50) 3 (18) 0.02

APACHE II score ‡ 25 12 (19) 0 0.06SOFA score ‡ 2 40 (63) 5 (29) 0.02Increase in SOFA score(baseline to 24h)

29 (45) 3 (18) 0.04

25OHD = 25-hydroxyvitamin D; APACHE = Acute PhysiologyAge Chronic Health Evaluation; SOFA = Sepsis-related OrganFailure Assessment.

Table 2Age-adjusted Association of Vitamin D Insufficiency andInfection Severity Outcomes Among 81 ED Patients Hospitalizedfor Suspected Infection

Characteristic

UnadjustedOdds Ratio*

(95% CI)

Age-adjustedOdds Ratio*

(95% CI)

BaselineSevere sepsis 0.20 (0.06–0.67) 0.26 (0.07–0.96)SOFA score ‡ 2 0.28 (0.07–1.05) 0.34 (0.09–1.34)

24 hoursSevere sepsis 0.20 (0.06–0.65) 0.28 (0.08–1.01)Two or moreorgan systemdysfunction

0.22 (0.06–0.82) 0.30 (0.07–1.25)

APACHE II score ‡ 25 NC NCSOFA score ‡ 2 0.25 (0.08–0.80) 0.35 (0.10–1.24)Increase in SOFAscore (baselineto 24h)

0.26 (0.07–0.99) 0.34 (0.85–1.36)

Hosmer-Lemeshow test results were p > 0.05 for each model,indicating appropriate model goodness of fit.25OHD = 25-hydroxyvitamin D; APACHE = Acute PhysiologyAge Chronic Health Evaluation; NC = not calculable due toperfectly associated outcome; SOFA = Sepsis-related OrganFailure Assessment.*Odds of outcome for 25OHD ‡75 nmol ⁄ L, compared<75 nmol ⁄ L group.

ACAD EMERG MED • May 2011, Vol. 18, No. 5 • www.aemj.org 553

aware of any evidence that serum 25OHD levels changeacutely in response to severe illness, it is difficult toaddress this issue (potential reverse causation) in thisobservational study. Clinical trials are needed beforevitamin D supplementation can be recommended toprevent infection and resulting sepsis.

CONCLUSIONS

Vitamin D insufficiency was associated with higher sep-sis severity in ED patients hospitalized for suspectedinfection.

References

1. Hewison M. Vitamin D and the immune system:new perspectives on an old theme. EndocrinolMetab Clin North Am. 2010; 39:365–79.

2. Liu PT, Stenger S, Li H, et al. Toll-like receptor trig-gering of a vitamin D-mediated human antimicro-bial response. Science. 2006; 311:1770–3.

3. Ginde AA, Mansbach JM, Camargo CA Jr. Associa-tion between serum 25-hydroxyvitamin D level andupper respiratory tract infections in the ThirdNational Health and Nutrition Examination Survey.Arch Intern Med. 2009; 169:384–90.

4. Urashima M, Segawa T, Okazaki M, Kurihara M,Wada Y, Ida H. Randomized trial of vitamin D sup-plementation to prevent seasonal influenza A inschool children. Am J Clin Nutr. 2010; 91:1255–60.

5. Equils O, Naiki Y, Shapiro AM, et al. 1,25-Di-hydroxyvitamin D3 inhibits lipopolysaccharide-

induced immune activation of human endothelialcells. Clin Exp Immunol. 2006; 143:58–64.

6. Møller S, Laigaard F, Olgaard K, Hemmingsen C.Effect of 1,25-dihydroxy-vitamin D3 in experimentalsepsis. Int J Med Sci. 2007; 4:190–5.

7. Ginde AA, Liu MC, Camargo CA Jr. Demographicdifferences and trends of vitamin D insufficiency inthe US population, 1988–2004. Arch Intern Med.2009; 169:626–32.

8. Jeng L, Yamschikov AV, Judd SE, et al. Alterationsin vitamin D status and anti-microbial peptide levelsin patients in the intensive care unit with sepsis.J Transl Med. 2009; 7:28.

9. Shapiro NI, Yano K, Okada H, et al. A prospective,observational study of soluble FLT-1 and vascularendothelial growth factor in sepsis. Shock. 2008;29:452–7.

10. Knaus WA, Draper EA, Wagner DP, ZimmermanKE. APACHE II: a severity of disease classificationsystem. Crit Care Med. 1985; 13:818–29.

11. Vincent J, Moreno R, Takala J, et al. The SOFA(sepsis-related organ failure assessment) score todescribe organ dysfunction ⁄ failure. On behalf of theWorking Group on Sepsis-Related Problems of theEuropean Society of Intensive Care Medicine.Intensive Care Med. 1996; 22:707–10.

12. Angus DC, Linde-Zwirble WT, Lidicker J, ClermontG, Carcillo J, Pinsky MR. Epidemiology of severesepsis in the United States: analysis of incidence,outcome, and associated costs of care. Crit CareMed. 2001; 29:1303–10.

554 Ginde et al. • VITAMIN D AND SEPSIS