Vitamin D level is associated with mortality predictors in ventilator-associated pneumonia caused by Acinetobacter baumannii

Introduction: Vitamin D plays a role in host defense and is known to be associated with mortality in patients in the intensive care unit (ICU). We aimed to evaluate the relationships between vitamin D levels and predictors of mortality in patients with ventilator-associated pneumonia (VAP) caused by extensively drug-resistant Acinetobacter baumanii (XDR A. baumanii). Methodology: A retrospective single-center study was conducted in an 18-bed adult ICU of a teaching hospital, including all patients with VAP due to XDR A. baumanii. Levels of 25(OH)D, procalcitonin (PCT), C-reactive protein (CRP), n-terminal pro-BNP (NT-proBNP), as well as clinical scores (Sequential Organ Failure Assessment [SOFA], Acute Physiology And Chronic Health Evaluation [APACHE II], Clinical Pulmonary Infection Score [CPIS) were recorded. Results: Forty-for patients were studied over six months. All patients had vitamin D deficiency. The 28-day mortality in patients with 25(OH)D levels ≤ 10 ng/mL was higher than in patients with 25(OH)D > 10ng/mL (p = 0.001). The fourthand seventh-day SOFA scores (p= 0.04 and p= 0.001) and firstand fourth-day procalcitonin levels (p = 0.03 and p = 0.004) were higher in patients with 25(OH)D levels ≤ 10 ng/mL. The clinical scores (SOFA, CPIS, and CEPPIS) and biomarkers (NT-proBNP, PCT) were negatively correlated with 25(OH)D levels in all study groups. Conclusions: Severe vitamin D deficiency was associated with adverse outcome in VAP due to XDR A. baumanii. Vitamin D levels may be a prognostic predictor of VAP. It is also important to evaluate the effect of rapid vitamin D replacement on mortality.


Introduction
Ventilator-associated pneumonia (VAP) is a nosocomial infection that develops 48 hours after intubation and mechanical ventilation [1].It is a common and serious problem in the intensive care unit (ICU), with an estimated incidence of 10% to 25%, and the risk of VAP increases 1%-3% for each day of invasive mechanical ventilation [2].VAP accounts for prolonged mechanical ventilation and length of stay, as well as poor outcomes [1].Despite technological advances, the mortality rate for VAP ranges between 24% and 50% [1,3].Studies aiming to find prognostic markers to quickly distinguish patients who will have unfavorable outcomes are ongoing [4,5].Early identification of patients with poor prognosis may improve treatment strategies.
Recent studies reported that Acinetobacter baumannii has become an increasingly significant cause of VAP in ICUs, and the mortality rate can reach up to 84.3% [6,7].This may be attributed to pathogens that are multidrug resistant (MDR) and extensively drug resistant (XDR) [8].In our center, A. baumanii is the most common pathogen that causes VAP, and new XDR species have recently emerged.For this reason, it is crucial to investigate the prognostic factors related to VAP by XDR A. baumanii.
Vitamin D has well-known effects on bone metabolism as a hormone, but there are increasing data about its additional effects.One of these extramusculoskeletal effects is on both the innate and adaptive immune systems.Vitamin D increases the production of cathelicidin and defensin and hence plays a role in the innate immune system [9].In addition, it shows its effects on the adaptive immune system via vitamin D receptors on monocytes and T and B cells [9].
Vitamin D deficiency is very common worldwide and there are increasing data about hypovitaminosis D and increased susceptibility to infections [10][11][12].Additionally, it was also documented that a relationship between vitamin D deficiency and increased mortality in community-acquired pneumonia (CAP) and sepsis was found [13][14][15][16].Despite the increasing number of studies about CAP, there are scarce data about VAP, and to our knowledge, there are no studies related to the effects of vitamin D on specific pathogens such as XDR A. baumanii in the literature.
Our single-center retrospective study was designed to investigate the relationships between vitamin D levels and mortality and known predictors of mortality in patients with VAP caused by XDR A. baumannii.

Subjects and study design
This retrospective study recruited patients admitted to an 18-bed adult ICU of Marmara University Pendik Training and Research Hospital in Istanbul (41°N), from November 2014 to May 2015; a total of 50 adult patients with microbiologically-confirmed VAP caused by XDR A. baumanii were selected.Six patients whose 25-hydroxyvitamin D [25(OH)D] levels were not measured were excluded from the final analysis.After obtaining approval from the Marmara University Ethics Committee, the study was conducted in accordance with the Declaration of Helsinki including current revisions and Good Clinical Practice guidelines.Informed consent was not obtained from all patients due to the retrospective study design.
Data were obtained from medical charts and electronic records.For each patient, the following data were collected: age, sex, medical history, duration of mechanical ventilation, clinical scores (sequential organ failure assessment [SOFA], acute physiology and chronic health evaluation [APACHE II], clinical pulmonary infection score [CPIS]), laboratory and microbiologic data, and 28-day mortality.
The diagnosis of microbiologically confirmed VAP was made when a patient who had been intubated and mechanically ventilated for ≥ 48 hours developed a new or persistent lung infiltrate and the tracheal aspirate culture results were positive (count > 10 4 colonyforming units/mL) [17].The day of VAP diagnosis was defined as day 1.XDR A. baumanii was defined as resistance to all standard antimicrobial agents except colistin or tigecycline [18].Patients' progress was followed until the 28th day after the diagnosis of VAP, when they were categorized as survivors.Patients who died before the 28th day were non-survivors.Patients discharged from the ICU before the 28th day were also considered survivors.

Clinical scores
The APACHE scoring system appraises disease severity by quantifying 34 physiologic variables [19].The SOFA score was designed as a tool for describing the severity of organ dysfunction, and serial assessments of the scores were associated with outcomes [20].CPIS is a diagnostic scoring system that predicts VAP; patients were assumed to have VAP when the CPIS score was > 6 [21].APACHE II was assessed on the first day of ICU admission and SOFA at day 1, day 4, and day 7 of diagnosis of VAP.CPIS was determined on day 1.
Vitamin D deficiency was considered as 25(OH)D < 20 ng/mL, insufficiency as 25(OH)D = 20-29 ng/mL, and sufficiency as 25(OH)D ≥ 30 ng/mL in accordance with the guidelines of the Endocrine Society [22].Additionally, 25(OH)D ≤ 10 ng/mL was defined as very severe vitamin D deficiency.Serum PCT, CRP, and NT-proBNP levels were evaluated every day in thr ICU but only day 1, day 4, and day 7 measurements were evaluated in this study.Serum 25(OH)D, calcium, and PTH values were evaluated routinely as part of the ICU care on the day the patients were admitted to the ICU.

Statistical analysis
R version 2.15.3 programme was used for statistical analysis.When evaluating study data, in addition to descriptive statistical methods (mean, median, standard deviation, Q1, Q3, frequency, ratio, minimum, maximum), student's t-test was performed to compare two groups of normally distributed variables and the Mann-Whitney U test was performed to compare two groups of non-normally distibuted variables in the evaluation of quantitative data.The Kruskal-Wallis test was used to test the difference of non-normally distributed variables between more than two groups, with the Mann-Whitney U test as post-hoc analysis.Pearson's Chi-square test, Fisher-Freeman-Halton exact test, Fisher's exact test, Yates's continuity corrected Chi-square test, McNemar test, and sensitivity values were used to test the association between nominal variables.Wilcoxon's signed-rank test was used to assess the significance of the changes at different time points.Pearson's correlation analysis and Spearman's correlation analysis were used to evaluate the relationships of the variables.Significance level was evaluated at p < 0.01 and p < 0.05.

Results
There were 27 (61.3%)men and 17 (38.6%)women with a mean age of 54.6 ± 16.1 years.The most frequent cause of ICU admission was postoperative respiratory failure.Other causes were polytrauma, intracranial hemorrhage, and thrombotic stroke.The mean duration of mechanical ventilation before VAP onset was 14 days.
The baseline characteristics of the 44 patients and stratifications as survivors or non-survivors are given in Table 1.Patients with cardiovascular diseases (p = 0.049) or diabetes mellitus (p = 0.039) had higher 28day mortality than did the others.Clinical scores and laboratory parameters are shown in Table 2.
All patients involved in the study had vitamin D levels < 20 ng/mL.The median 25(OH)D levels of the survivors were significantly higher than that of the nonsurvivors (10 [8.2-11.9]and 3 [2-4.8],respectively; p < 0.001).The relationship between 25(OH)D levels and demographic parameters are shown in Table 3.There was no statistically significant difference between age and duration of mechanical ventilation and 25(OH)D levels.Diabetic patients had lower vitamin D levels than did non-diabetics (p = 0.001).Similarly, vitamin D levels were significantly lower in patients with cerebrovascular diseases (p = 0.035).

Discussion
In this study, our data showed that hypovitaminosis D was very prevalent in patients with VAP caused by XDR A. baumanii, and very severe vitamin D deficiency was related to poor prognosis.We also demonstrated that vitamin D concentrations at admission to the ICU were an important predictor of 28day mortality, and as such, it could function as an adjunctive parameter to clinical scores and inflammatory biomarkers.
There are very few studies in the literature about the relationship of vitamin D status with A. baumanii infections.Turkoglu el al. reviewed D deficiency and the development of A. baumanii infections in critically ill patients and showed that vitamin D deficiency was an independent risk factor of A. baumanii infections [23].Consistent with their study, we also found that the vitamin D levels of patients in both survivors and non-survivors were within deficient status (< 20 ng/mL).Additionally, the vitamin D status of non-survivors was lower than that in the survivors.The high prevalence of vitamin D deficiency found in patients with VAP was comparable to that reported in previous studies [23][24][25].
It has been demonstrated that vitamin D had immunomodulatory activities [26,27].In cases of hypovitaminosis D, these immunomodulatory effects are disruptive, and a tendency to infections occurs.One of the most frequently studied infections with respect to vitamin D deficiency is CAP.Increased mortality in CAP was found to be related to vitamin D deficiency [13,14].There are also growing data about vitamin D status on the prediction of outcomes in CAP, sepsis, and critically ill patients [24,25,28].Nonetheless, to our knowledge, no studies have specifically related vitamin D status to patients with VAP.In this study, we evaluated a more specific subgroup of patients with VAP who had microbiologically documented XDR A. baumanii infections.The reason for choosing this subgroup was that XDR A. baumanii was the most prevalent and mortal infectious agent in our institute.Consistent with the literature, we also found that vitamin D deficiency could be a predictor of mortality in patients with VAP caused by XDR A. baumanii.
The prognostic performance of PCT, CRP, SOFA, and APACHE II scores were previously evaluated, which produced variable results [5,29].The prognostic value of serum 25(OH)D was not evaluated previously in VAP associated with XDR A. baumannii.The comparision of the prognostic value of 25(OH)D with the predictive ability of clinical scores and biomarkers was a novel aspect of our study.Vincent et al. showed that multiple organ dysfunction and high SOFA scores for any individual organ were associated with increased 572 mortality [30].SOFA scores in our study were significantly lower in survivors on days 1, 4, and 7 compared with the scores of non-survivors.There was also a negative correlation found between vitamin D levels and day 4 and day 7 SOFA scores.A body of evidence showed that PCT, the prehormone of calcitonin, was a predictor of severity in sepsis, antimicrobial efficiency, and hospital mortality [31,32].The level of PCT correlates with the severity of pneumonia [33,34].Also, one small study found that procalcitonin levels increased over time in nonsurvivors but decreased in survivors [34].Luyt et al. studied the prognostic value of PCT kinetics in patients with VAP and concluded that elevated serum levels of PCT on days 1, 3, and 7 of VAP were strong predictors of unfavorable outcomes [35].CRP, an acute-phase protein produced by the liver, appears to be less sensitive than procalcitonin for the detection of bacterial pneumonia [34].In our study, we showed a significant relationship between PCT and vitamin D levels but no correlation between 25(OH)D concentrations and serum CRP levels, consistent with the literature.
In terms of CPIS and cause of admission, the two groups were comparable.Similary to what we found in our study, Seligman et al. found that CPIS and the causes of admission were not significantly pronounced in both groups [36].APACHE II scores on admission were also similar between survivors and non-survivors.In addition, we found no significant correlations between serum 25(OH)D and APACHE II and CPIS scores.
NT-proBNP has recently been used to predict sepsis outcomes.A recent meta-analysis suggested that elevated NT-proBNP levels could be a predictor of mortality in patients with sepsis [37].Another study demonstrated that elevated serum NT-proBNP was an independent predictor for poor ICU outcomes in the presence of clinical severity scores [38].In our study, the NT-proBNP levels in the non-survival group were higher than those in the group on days 4 and 7. Additionally, there was a negative correlation found between vitamin D levels and day 1, day 4, and day 7 NT-proBNP levels.
Underlying disease of patients has also been related with death as much as the infection sites and antibiotic susceptibility of pathogens [39].Diabetes mellitus and cardiovascular diseases were negatively associated with mortality in our study.These susceptible populations may be considered for special follow-up of response to treatment.
There were some limitations to our study.First, the patient population was small and no control group was included.Second, this was not a prospectively designed study, so causal relationship could not be established.These limitations reduce the strength of the conclusions; thus, the results should be verified by larger scale, prospectively designed studies.Nevertheless, to our knowledge, this is the first study to investigate the clinical and microbiologic outcomes of patients with VAP caused by XDR A. baumanii, one of the most frequent and mortal VAP pathogens, and extremely low 25 (OH)D levels.Further studies are needed to address the benefits of rapid vitamin D replacement in this population.

Conclusions
Vitamin D deficiency is highly prevalent in patients with VAP caused by XDR A. baumanii.We confirmed that severe 25(OH)D deficiency was associated with adverse outcomes in patients with VAP associated with XDR A. baumanii.Vitamin D levels may be a significant predictor of 28-day mortality.It is also important to evaluate the effects of rapid vitamin D replacement on mortality.

Table 1 .
Baseline characteristics of 44 patients who developed VAP from XDR A. Baumanii.

Table 2 .
Evaluation of clinical scores and laboratory parameters.
* Results are given as mean ± standard deviation; β Results are given as median and [interquartile range]; SOFA: sequential organ failure assessment; CRP: Creactive protein; NT-ProBNP: n-terminal pro-brain natriuretic peptide.

Table 3 .
The relationship between 25(OH)D levels and demographic parameters.
*Data are expressed as mean ± standard deviation; β Data are expressed as median [interquartile range]; DM: diabetes mellitus; COPD: chronic obstructive pulmonary disease.

Table 4 .
The correlations between 25(OH)D levels and biochemical parameters and clinical scores.
APACHE II: acute physiology and chronic health evaluation; CPIS: clinical pulmonary infection score; CRP: C-reactive protein; proBNP: pro-brain natriuretic peptide; SOFA: sequential organ failure assessment.

Table 5 .
Clinical and biochemical prognostic parameters according to stratification of vitamin D status.