Secondary Statistical Analyses. Additionally, all cases were analyzed by site of infection, location (inpatient or outpatient) of positive culture, and season of infection. The same season definitions were used for infection as were used for vitamin D measurement (winter: December- February, spring: March-May, summer: June-August, and fall: September-October). Infections that occurred in 2005 were excluded from this analysis, since infections were only recorded for part of this year. Because the first serum 25(OH)D level measured for each patient was used in the analysis, and each case did not simultaneously present with an MRSA infection, the number of days between positive MRSA culture (day 0) and serum 25(OH)D recording was determined. In addition, all vitamin D determinations in the database were used to calculate the median serum 25(OH)D for all subjects by year. Because the patient population at the VAMC is not necessarily representative of the general population, additional analyses were carried out in order to further characterize the patients included in the study. Both age and race were analyzed by the other patient characteristics.
Secondary Statistical Analyses. Most cases (70.1%) experienced one infection while relatively small percentages of cases experienced 2 to 5 infections over the duration of the study period (Figure 1). The most common site of MRSA infection was skin and soft tissue (n=232; 57.9%), followed by genitourinary tract (n=53; 13.2%), bloodstream and cardiovascular system (n=37; 9.2%), respiratory (n=32; 8.0%), bone and joint (n=22; 5.5%), surgical site (n=13; 3.2%), eye/ear/nose/throat/mouth (n=10; 2.5%), reproductive tract (n=1; 0.3%), and gastrointestinal tract (n=1; 0.3%) (Figure 2). Almost all of the MRSA infections were diagnosed in the outpatient setting (n=366; 91.3%) as opposed to the inpatient setting (n=35, 9%). There was no significant difference in frequency of infection by season (p=0.1343) (Table VIII). The number of days between vitamin D determination and positive MRSA culture ranged from 1669 days (approximately 4.57 years) before infection to 1360 days (3.72 years) after the infection with a median of 77 days prior to infection (Figure 3). Median vitamin D levels were approximately the same from 2007 to 2009, and then dropped significantly to 19 ng/mL in 2010 (p<0.0001) (Table IX). In this group of veterans, males were significantly older than females (p<0.0001), white subjects were significantly older than black subjects and subjects in the “other race” category (p<0.0001), and non-Hispanics were significantly older than Hispanics (p=0.0308) (Table X). BMI related inversely to age (p<0.0001). HIV positive individuals were significantly younger than HIV negative subjects (p<0.0001). Race was significantly associated with gender (p<0.0001), ethnicity (p<0.0001), and HIV status (p<0.0001) (Table XI). The other race category was comprised of the highest percentage of females (29.8%) and Hispanic subjects (7.0%). Of the black subjects, 23.2% were HIV positive, which was significantly more than in the white and other race categories (6.9% and 0.0%, respectively).
Secondary Statistical Analyses
