Circulating 25-Hydroxyvitamin D and 1,25-Dihydroxyvitamin D Concentrations and Postoperative Infections in Cardiac Surgical Patients: The CALCITOP-Study

Background Vitamin D has immunomodulatory properties and seems to reduce the risk of infections. Whether low vitamin D concentrations are independent risk factors for nosocomial postoperative infections in surgical patients remains to be studied in detail. Methods In 3,340 consecutive cardiac surgical patients, we investigated the association of circulating 25-hydroxyvitamin D (25OHD; indicator of nutritional vitamin D status) and 1,25-dihydroxyvitamin D (1,25[OH]2D; active vitamin D hormone) with nosocomicial infections. The primary endpoint was a composite of thoracic wound infection, sepsis, and broncho-pulmonary infection. Vitamin D status was measured on the last preoperative day. Infections were assessed until discharge. Logistic regression analysis was used to examine the association between vitamin D metabolite concentrations and the composite endpoint. Results The primary endpoint was reached by 5.6% (n = 186). In patients who reached and did not reach the endpoint, in-hospital mortality was 13.4% and 1.5%, respectively (P<0.001). Median (IQR) 25OHD and 1,25(OH)2D concentrations were 43. 2 (29.7–61.9) nmol/l and 58.0 (38.5–77.5) pmol/l, respectively. Compared with the highest 1,25(OH)2D quintile (>81.0 pmol/l), the multivariable–adjusted odds ratio of infection was 2.57 (95%CI:1.47–4.49) for the lowest 1,25(OH)2D quintile (<31.5 pmol/l) and 1.85 (95%CI:1.05–3.25) for the second lowest quintile (31.5–49.0 pmol/l). There was no significant association between 25OHD concentrations and the primary endpoint. Conclusions Our data indicate an independent association of low 1,25(OH)2D levels with the risk of postoperative infections in cardiac surgical patients. Future studies should pay more attention on the clinical relevance of circulating 1,25(OH)2D and its regulation.


Introduction
Nosocomial infections occur worldwide and are a significant burden both for the patients and for public health [1]. Infections of surgical wounds, the urinary tract, and the lower respiratory tract are the most frequent infections [1]. Cardiac surgical patients are at an increased risk of developing nosocomial infections. Incidence rates of up to 20% and more have been reported [2]. Although serious postoperative complications are uncommon, they are potentially devastating. For example, in a large study in 331,429 cardiac surgical patients, patients who developed an infection were more likely to have a prolonged in-hospital stay compared with patients who did not develop a postoperative infection [3]. In addition, in-hospital mortality was significantly higher in patients with major postoperative infection (17.3% versus 3.0%).
Besides its pivotal role in musculoskeletal health, vitamin D has a broad range of additional effects, which also cover the immune system [13]. There is likely evidence that vitamin D decreases the risk of airway infections [14,15]. Moreover, vitamin D supplements seem to decrease the need for antibiotics [16], especially in elderly patients [17]. In cardiac surgical patients, the prevalence of vitamin D deficiency (circulating 25-hydroxyvitamin D [25OHD] < 30 nmol/l) is between 15% and 38% [18][19][20].
The present study aimed to investigate whether low vitamin D status is associated with an increased risk of developing postoperative infections.

Methods Patients
This investigation is based on data of the CALCITOP (Calcitriol and clinical outcome in cardiac surgical patients) study, a prospective cohort study of 3,852 cardiac surgical patients [21]. Study participants were recruited between February 2012 and December 2013 at a tertiary heart center in East-Westphalia, Germany (Heart and Diabetes Center North Rhine-Westphalia, Bad Oeynhausen). Eligible for inclusion were cardiac surgical patients aged 18 years and over. Patients with heart transplants and pacemaker/defibrillator implants were excluded. In 3,340 patients, preoperative serum concentrations of 25OHD and 1,25(OH) 2 D as well as postoperative data on infections were available. This dataset was used to perform the statistical analyses. The study was approved by the Ethics Committee of the Ruhr University Bochum at Bad Oeynhausen and was registered at clinicaltrials.gov as NCT02192528. All patients provided their written informed consent.

Data collection
We prospectively collected preoperative, perioperative, and postoperative data using the electronic patient databases of our institution. Besides 25OHD (indicator of nutritional vitamin D status) and 1,25(OH) 2 D (active, hormonal form of vitamin D), we assessed additional preoperative and surgical characteristics such as age, gender, body mass index (BMI), left ventricular ejection fraction (LVEF), current smoking, diabetes mellitus, concentrations of creatinine, glucose, C-reactive protein (CRP) and leukocytes, previous cardiac surgery, operation priority, and type of surgery. Moreover, we assessed four postoperative outcome parameters, namely infection, intensive care unit (ICU) stay, in-hospital stay, and in-hospital mortality.

Antibiotic Prophylaxis and Treatment
According to institutional standards, antibiotic prophylaxis was performed routinely with 2g cephalosporin (Cefazolin-Sandoz, Sandoz Pharmaceuticals, Basel Switzerland), preoperatively, intra-operatively and during postoperative drainage (up to 48 h after surgery). In case of an infection, antibiotic therapy was selected according to the sensitivity of the organism and the clinical response of the patient.

Biochemical Analyses
Circulating levels of 25OHD and 1,25(OH) 2 D, creatinine, CRP, leukocytes, and glucose levels were measured as previously described [21]. Estimated glomerular filtration rate (eGFR) was calculated using the creatinine-based modification of diet in renal disease formula. Routine microbiological tests were used to assess the type of pathogen in case of an infection.

Endpoints
The primary endpoint was a composite of clinically relevant infections such as thoracic wound infection, sepsis, and broncho-pulmonal infection. Infections were diagnosed according to standard procedures, such as the presence of positive results of microbial culture, pyrexia, tachycardia, and tachypnea. Events were assessed until discharge. Additional endpoints were urogenital and other infections, ICU stay, in-hospital stay, and in-hospital mortality.

Statistics
Categorical variables are summarized as percentages. Since several continuous variables were non-normally distributed (CRP, glucose), continuous variables are reported as median and interquartile range (IQR). We used the Kolmogorov-Smirnov test to check normal data distribution of continuous variables. Normal distribution was a consideration when probability values were >0.05. Differences in preoperative categorical variables and continuous variables between the groups with and without infection were assessed using Fisher's exact test and the Mann-Whitney U-test, respectively.
We carried out multiple logistic regression analysis to assess the independent relationship of the preoperative 25OHD or 1,25(OH) 2 D 3 categories with the risk of infection. According to previous classification [21], we used the following cut-off values for classifying 25OHD: risk of deficiency (<30 nmol/L), risk of inadequacy (30 nmol/L to 49.9 nmol/L), borderline status (50-74.9 nmol/L), adequacy (75-100 nmol/L), and potentially harmful (>100 nmol/L, to convert nanomolar to nanogram per milliliter divide by 2.496). The group with adequate vitamin D status was used as the reference group. Regarding 1,25(OH) 2 D, we divided the study cohort into quintiles, since no generally accepted classification exists. We performed age-and genderadjusted analyses and used multivariable-adjusted models to examine the association between vitamin D metabolites and the incidence of the primary endpoint. Since the number of variables that can be included for multivariable testing is equal to the square root of the number of events [22], we restricted the covariates to important demographic parameters (age, gender, and BMI), cardiac surgical-related parameters (redo, operation priority, type of cardiac surgery), and additional parameters that differed significantly between patients who reached or did not reach the primary endpoint (see Table 1). We calculated absolute (incidence) rates and odds ratios (ORs) and corresponding 95% confidence intervals (CI). In subgroup analyses, we restricted the study cohort to non-diabetes patients and patients aged 70 years. In sensitivity analyses, we used 50 nmol/l as cutoff for the lowest 25OHD category and divided the study cohort into tertiles of 1,25(OH) 2 D levels. Moreover, we compared the status of vitamin D metabolites in patients with any infection (broncho-pulmonal infection, thoracic wound infection, sepsis, urogenital infection or other infection) with non-infected patients. To reduce the risk of study bias through already existing preoperative infections, we also restricted the analysis to a comparison between patients with thoracic wound infection and patients without infection. We considered P values <0.05 as statistically significant. All P values are reported two-sided. Analyses were performed using the statistical software package IBM1 SPSS1, version 21.
However, gender distribution, BMI, smoking status, blood glucose levels, and percent onpump surgery did not differ significantly between groups.
During postoperative hospitalization, infected patients had higher median leukocyte and CRP concentrations than patients who did not reach the primary endpoint (Fig 1).
Moreover, ICU-stay and in-hospital stay were longer, and in-hospital mortality was significantly higher in infected patients than in other patients (13.4% vs. 1.5%, Table 2). Of the 25 patients who reached the primary endpoint and died, causes of death were sepsis (n = 8), multiorgan failure (n = 6), heart failure (n = 6), cardiogenic shock (n = 1), and others (n = 4).
In the age-and gender-adjusted model, the OR for patients in the lowest 25OHD and 1,25 (OH) 2 D category was non-significantly and significantly higher, namely 1.77 (95%CI: 0.96-3.28; P = 0.071) and 3.97 (95%CI: 1.74-4.18; P<0.001), compared with the respective reference group. In the fully adjusted model, the OR for patients in the lowest 25OHD category was attenuated and did not differ significantly from the reference category. The ORs remained however significantly higher for the patients in the two lowest 1,25(OH) 2 D categories, compared with the highest 1,25(OH) 2 D category (Table 3). Results did not differ substantially if only non-diabetes patients (n = 2484) or patients 70 years (n = 1887) were included in the data analysis (S1 and S2 Tables). Similar results were also obtained in sensitivity analyses, when 25OHD levels < 50 nmol/l were considered as lowest 25OHD category or 1,25(OH) 2 D levels were divided into tertiles. In detail, the fully adjusted OR for patients in the 25OHD category < 50 nmol/l was 1.44 (95%CI: 0.78-2.67; P = 0.244), compared with the reference category. Regarding 1,25(OH) 2 D, the OR for patients in the lowest tertile was 2.06 (95%CI: 1.36-3.10; P = 0.001), compared with the highest tertile. If all cases of infection (n = 430) were

Discussion
Our data support earlier findings [12] that despite antibiotics prophylaxis, postoperative infection remains a significant complication in cardiac surgical patients. To improve prophylaxis, the use of second-or third-generation cephalosporins as well as prophylaxis prolongation up to 48 h post-operatively has been recommended in cardiac surgical patients [12]. However, the broad spectrum of different pathogens and infections in our study cast doubt that this strategy would completely solve the problem of postoperative infection. In our study, low 1,25(OH) 2 D levels were independently associated with an increased risk of a composite of clinically relevant infections. The risk was lowest in patients with circulating 1,25(OH) 2 D levels above 81.0 pmol/l. No such association was observed for circulating 25OHD levels. Thus, our data do not support a systematic review and meta-analysis of observational studies that found a significantly increased risk of infection and sepsis in critically ill patients with circulating 25OHD levels below 50 nmol/l [23]. However, it is noteworthy that   the meta-analysis reported risk ratios for infection and sepsis of 1.49 (95%CI: 1.12-1.99) and 1.46 (95%CI: 1.27-1.68), respectively, which are on average very similar to the non-significantly higher OR we observed in the sensitivity analysis of our study, compared with adequate 25OHD levels. Thus, our data do not definitively rule out a higher risk of infection at insufficient or deficient 25OHD levels. With respect to 1,25(OH) 2 D, our results support data of a small earlier study in cardiac surgical patients [24]. In that investigation, lower circulating 1,25 (OH) 2 D levels were significantly associated with higher risk of a composite of low cardiac output syndrome, infection, or in-hospital death. Notably, infection was the most prevalent complication (n = 4) among the 7 postoperative events that occurred. Experimental studies provide insights into potential mechanisms of 1,25(OH) 2 D-mediated effects on the immune system: vitamin D receptors are expressed in monocytes and these cells differentiate into macrophages under the influence of 1,25(OH) 2 D [25]. Macrophages express their own 1α-hydroxylase isoenzyme which intracellularly synthesizes 1,25(OH) 2 D from its precursor 25OHD [26]. 1,25(OH) 2 D directly and indirectly regulates the expression of important antimicrobial proteins, such as cathelicidin and defensins [27,28], and of lysosomal enzymes and reactive oxygen species like nitric oxide [29]. The combination of antimicrobial peptides and oxygen species may destroy intracellular viruses, fungi, and bacteria in the autolysosomes. Notably, monocytic cathelicidin production is reduced in individuals with low 25OHD and 1,25(OH) 2 D levels [30].
Normalizing tissue 25OHD may be necessary for providing adequate amounts of substrate for local tissue production of 1,25(OH) 2 D [31]. Interestingly, in patients with low circulating 1,25(OH) 2 D levels administration of 1,25(OH) 2 D can increase 25OHD uptake in monocytes [32], suggesting that adequate circulating 1,25(OH) 2 D levels are necessary for sufficient 25OHD availability in vitamin D target cells.
Usually, 1,25(OH) 2 D synthesis is suppressed by low substrate availability, e.g. deficient 25OHD levels [15]: In children and young female adults, for instance, an increase in circulating 25OHD of 10 nmol/l was associated with an increase in circulating 1,25(OH) 2 D of approximately 5 pmol/l [33,34]. In patients of the CALCITOP study, however, the corresponding increase in circulating 1,25(OH) 2 D levels was only 0.7 pmol/l [21], whereas several clinical parameters such as kidney function, EuroSCORE (a surgical risk score), diabetes, CRP, and diuretic use were inversely correlated with circulating 1,25(OH) 2 D levels in these patients. Similarly, in heart transplant recipients kidney function and inflammatory parameters were much more predictive of circulating 1,25(OH) 2 D levels than circulating 25OHD levels [35]. These differences between healthy individuals and cardiac surgical patients may at least in part explain why vitamin D supplements have beneficial effects on upper respiratory tract infection in otherwise healthy individuals [14,15], whereas deficient 25OHD levels (as an indicator of nutritional vitamin D status) were not significantly associated with infection in the present study. Consequently, we should not be too enthusiastic to believe that in the clinical setting simple vitamin D supplementation would be able to restore all vitamin D-related derangements.
It may well be that in our fully adjusted statistical model the association between low circulating 1,25(OH) 2 D levels and postoperative infections was underestimated by the fact that adjustments were made for those parameters which are related to both clinical outcome and circulating 1,25(OH) 2 D levels. Circulating 1,25(OH) 2 D is related to eGFR, even in individuals without chronic kidney disease [36]. As mentioned before, in the patients of the CALCITOP study 1,25(OH) 2 D was also inversely interrelated with EuroScore, diabetes and CRP values [21]. Therefore, we cannot definitively rule out that our study results are biased by over-adjustment. Of note, our data indicate a stronger association of circulating 1,25(OH) 2 D with the risk of infection in models not adjusted for the aforementioned parameters.
Our study has several strengths, such as the large number of included patients, the availability of data on both 25OHD and 1,25(OH) 2 D levels, and the short follow-up period. However, some limitations also have to be addressed. First, we do not have exact data on postoperative antibiotics use in our study cohort. Second, no data on circulating postoperative 1,25(OH) 2 D levels were available in our study. Earlier investigations could demonstrate a significant transient cardiac surgery-related decline in circulating 1,25(OH) 2 D [24], which may have contributed to the increased risk of infections. Postoperative concentrations were, however, consistently lower in patients with relatively low preoperative 1,25(OH) 2 D levels than in patients with relatively high preoperative levels of this vitamin D metabolite [24], suggesting that the preoperative 1,25(OH) 2 D level is also indicative for the postoperative 1,25(OH) 2 D level of a patient. Third, no subgroup analyses regarding the association of vitamin D with gram positive bacteria, gram negative bacteria, or fungi could be performed, because many patients were simultaneously infected with two or three of these major groups of pathogens.
Finally, there is evidence for novel pathways of vitamin D 3 metabolism initiated by CYP11A1 [37,38], indicating that vitamin D action is not only mediated by the sequence vitamin D 3 ! 25OHD 3 ! 1,25(OH) 2 D 3 , but also by other metabolites which were not analyzed in this study.
In conclusion, our data indicate an independent association of low circulating 1,25(OH) 2 D levels with the risk of postoperative infections in cardiac surgical patients, whereas 25OHD levels were not significantly related to the risk of infection. Future studies should therefore pay more attention on circulating 1,25(OH) 2 D, its regulation, and clinical relevance.
Supporting Information S1 Table. Multivariable-adjusted odds ratio (OR) for the primary endpoint in non-diabetes patients by cutoffs of 25-Hydroxyvitamin D and 1,25-Dihydroxyvitamin D (DOCX) S2 Table. Multivariable-adjusted odds ratio (OR) for the primary endpoint in patients 70 years by cutoffs of 25-Hydroxyvitamin D and 1,25-Dihydroxyvitamin D (DOCX)