The Southern Alberta Clinic (SAC) and Calgary Laboratory Services (CLS) are the exclusive providers of HIV care and all routine laboratory services respectively in Calgary, Canada. Patients accessing care at SAC sign a voluntary informed consent form allowing administrative data to be used in research projects as approved by the University of Calgary Ethics Committee. Hospitalization data including discharge diagnoses from all city hospitals on SAC patients are collected and available within the SAC database . Hospitalizations were included if a primary cause for admission was pneumococcal disease. Length of stay was evaluated using previously described methodology .
All HIV patients with cultures positive for S. pneumoniae growth between January 1st, 2000 and January 1st, 2010 were identified by reviewing the clinical database, which contains all of the microbiology results, provided to us by CLS. The results were reconciled for both completeness and for verification with hospitalization data. IPD was diagnosed when the culture was isolated from a normally sterile site such as blood, peritoneum, pleura, or cerebrospinal fluid (CSF); all such isolates were serotyped. Pneumococcal pneumonia was defined as isolation from a sputum sample or via bronchoalveolar lavage (BAL); most BAL isolates were also serotyped. Sputum and BAL samples were obtained because of strong clinical suspicion for lower respiratory tract infection, generally associated with radiologic evidence and severe illness.
The follow-up period was measured from the initial HIV-care visit until the patient moved, died, was disconnected from care, or until January 1st, 2010; this was used to calculate incidence rates . If a patient presented with recurrent pneumococcal disease multiple times in one year, only the first instance was counted in an effort to reduce data distortion by a few patients with multiple recurrent cases. Clinical information obtained through routine care was evaluated with multivariate binary logistic regression. These risk factors included gender, age group (≤30, 31-45, 46-60, or ≥61), race (white, Aboriginal, black, or other), education (< high school or ≥ high school), a smoking history (if >1 month), alcohol abuse (>9 drinks/week for females, >14 drinks/week for males), IDU (at any time), nadir CD4 cell count (≤200, 201-500, or >500/μL), co-trimoxazole use, frequency of pneumococcal immunization (never, < every 5 years, or ≥ every 5 years), history of active tuberculosis (any clinical event), history of a common HIV-associated pulmonary disease (mycobacterium avium complex or Pneumocystis Jiroveci pneumonia), asthma, chronic obstructive pulmonary disease (COPD), or hepatitis C. All risk factors were adjusted for gender, age, a smoking history, IDU, and nadir CD4 cell count. These confounders were selected a priori based on expert opinion and previous studies [2, 8, 9]. Pearson's χ2 and adjusted odds ratios with 95% confidence intervals (or P < 0.05) were used to evaluate significance. All demographic information was self-defined.
The effect of immunization with a 23-valent pneumococcal polysaccharide vaccine (PPV-23, Merck Frosst, Kirkland, Quebec, Canada) on the incidence of IPD was analyzed by comparing number of episodes during the total patient-time without immunization with the time post-immunization. The post immunization time was then sub-divided into categories: within three years of immunization, three to five years post-immunization, and greater than five years post-immunization. Any episodes of IPD within 30 days of immunization were to be excluded. Patients were immunized with two doses of PPV-23 five years apart and with a repeat dose every five years based on clinical discretion.
A nested case-control study was performed in patients who received PPV-23 comparing patients who developed IPD versus those who did not in a 1:3 ratio to evaluate the interaction of HIV-specific markers, while controlling for other known risks. Each patient diagnosed with IPD was randomly matched with three patients with the same age group, gender, smoking history, and history of IDU who also received PPV-23 but had not been diagnosed with IPD. Differences in HIV viral load and CD4 cell count at the time of immunization were evaluated with Levene's test for non-significance, followed by a 2-tailed t-test of means. An additional multivariate analysis included both CD4 count and viral load. Differences in nadir CD4 count, a history of COPD, asthma, and hepatitis C were also evaluated.
Statistical analyses were conducted with SPSS, version 15.0 for Windows (SPSS Inc., Chicago, Illinois).