Incidence Trends and Epidemiology of Staphylococcus aureus Bacteremia: A Systematic Review of Population-Based Studies

Objectives: To determine incidence trends of Staphylococcus aureus bacteremia (SAB) from population-based studies from multiple countries. Methods: A contemporary systematic review was conducted using Ovid Cochrane Central Register of Controlled Trials (1991+), Ovid Embase (1974+), Ovid Medical Literature Analysis and Retrieval System Online (MEDLINE) (1946+ including epub ahead of print, in-process & other non-indexed citations), and Web of Science Core Collection (Science Citation Index Expanded 1975+ and Emerging Sources Citation Index 2015+). Two authors (J.R.H. and J.A.Q.M.) independently reviewed all studies and included those that reported population-based incidence of SAB in patients aged 18 years and older. Results: Twenty-six studies met inclusion criteria with the highest number (n=6) of studies conducted in Canada. The incidence of SAB ranged from 9.3 to 65 cases/100,000/year. The median age of patients with SAB ranged from 62 to 72 years and SAB cases were more commonly observed in men than in women. The most common infection sources were intravascular catheters and skin and soft tissue infections. SAB incidence trends demonstrated high variability for geographic regions and calendar years. Overall, there was no change in the incidence trend across all studies during the past two decades. Conclusion: Multiple factors, both pros, and cons are likely responsible for the overall stable SAB incidence in countries included in this systematic review. Some of these factors vary in geographic location and prompt additional investigations from countries not included in the current review so that a more global characterization is defined.


Introduction And Background
Staphylococcus aureus is a predominate pathogen in cases of bacteremia in both the community and healthcare settings [1,2], with a mortality rate as high as 30% [3]. The incidence of S. aureus bacteremia (SAB) varies depending on risk factors of the studied population and infection control practices in healthcare facilities. Surveillance of SAB trend incidence is crucial to evaluate its impact on public health agencies and to enhance the development of infection prevention and control strategies [4]. Unlike hospital-based studies, population-based studies prevent selection biases and allow standardization of incidence rates for a reference population, making them one of the best tools to assess infectious diseases epidemiology [5]. The most recent systematic review of population-based studies assessing the incidence of bacteremias due to a variety of organisms was conducted in 2013, with incidence rates reported through 2008 [6]. Therefore, we conducted a systematic review of contemporary population-based studies investigating the incidence of SAB from numerous geographic regions during the past two decades as presented herein.

Data Sources and Searches
A literature search was done by a medical librarian (D.J.G.) for the concepts of SAB and incidence rates. Search strategies were created using a combination of keywords and standardized index terms. Searches were run on September 2, 2021

Inclusion and Exclusion Criteria
All population-based investigations published in the English language from 2001 through 2020 on SAB incidence trends in the adult population were included in this review. Studies that did not report population-based estimates were not included, such as case reports, single-center, and multi-center studies, clinical trials, conference abstracts, systematic reviews, and studies on animals [7]. Studies conducted that focused on specific sub-groups (e.g., patients with HIV infection or end-stage renal disease [ESRD]) or published in a language other than English were also excluded.

Data Extraction
The following information was extracted from the included studies: first author's last name, year of publication, country of origin (+/-specific region studied), population size, and incidence rate (per 100,000 person-year). Two reviewers (J.R.H. and J.A.Q.M.) extracted data from the studies independently; study investigators were contacted if additional information was needed. The incidence rate of SAB was the primary outcome. Secondary outcomes were age, sex, comorbidities, methicillin resistance of the blood culture isolate, site of onset, and mortality rates.

Statistical Analysis
When available, incidence rate estimates from separate periods within each study were extracted. If estimates from separate periods were not available, overall incidence rate estimates were used. When necessary, these incidence rate estimates were transformed to represent incidence rates per 100,000 population. To visualize the data, incidence rate estimates were plotted by time point and study. For periods where reported incidence rates spanned multiple years, the median of the period was chosen for plotting. To look at trends across the entire period of interest, a simple linear regression model was fit to the individual estimates to help visually interpret the overall trend over time. The SAS, version 9.4 (SAS Institute, Cary, NC, USA) was used for plotting and regression.

Results
After removing all duplicates, a total of 16,499 citations were identified from the search engines (supplemental data) and their abstracts were screened. Thirty-eight of them were chosen for full-text review and 26 studies met inclusion criteria ( Table 1).   The highest number (n=6) of studies were from Canada; other countries that were included are shown in Figure 2. Of note, 20 (76.9%) of the 26 investigations were regional studies.  Trends across all included studies demonstrated high variability for geographic regions and calendar years with a SAB incidence ranging from 9.3 to 65 cases/100,000/year. While some showed an increasing incidence of SAB, others revealed decreasing or even unchanging rates over time. Overall, there was no change in the incidence trend across all studies during the past two decades (Figure 4).

FIGURE 4: Overall incidence trend of SAB in included studies
SAB: Staphylococcus aureus bacteremia *The multinational study included five countries: Australia, Canada, Denmark, Finland, and Sweden.
In the Danish study by Mejer et al., the proportion of SAB patients with a Charlson comorbidity index (CCI) >0 increased from 1995 to 2008 and the majority of this population had a CCI of 1-2 (39.1%) [8]. In addition, another Danish study restricted to Funen County showed that most of the SAB patients had no comorbidities [9]. However, the majority of the Danish population according to Thorlacius-Ussing et al. had a CCI >3 [10], similar to the SAB populations in Quebec, Canada, and Oslo, Norway [18,20]. Three-fourths of the Canadian population in Calgary from 2000 to 2006 had significant chronic comorbid conditions and/or alcohol use disorder [23]. Compared to SAB patients with no comorbidities, there was increased mortality with increased CCI scores [10].  [27], which was comparable to that reported in Australian and Icelandic studies (0.74% to 5.7%) [16,29,30]. In addition, Lam et al. noted that 18.7% of the population was abusing substances and was not limited to IDU [24].
Based on the New Zealand deprivation index, the least deprived patients had a significantly lower incidence rate of SAB compared to the most deprived [32].
The most common infection sources of SAB were intravascular catheters and skin and soft tissue infections [20,23,29,30], although four studies from Norway, Iceland, USA, and Canada reported that the majority of their observed SAB did not have an identified source of infection [16,18,24,27]. Other sources were also identified, including infective endocarditis, bone or joint, and respiratory tract infections [16,20,23,24,27,30].

Discussion
The current systematic review represents a contemporary review of SAB incidence in 26 geographical population-based investigations over two decades and demonstrates the striking burden of disease due to SAB. The incidence of SAB ranged from 9.3 to 65 cases/100,000/year. The highest SAB incidence appeared to be an outlier; it was seen in an aboriginal population of the Northern Territory of Australia and was likely due to unique risk factors associated with SAB [29]. This considerable variability in SAB incidence rates between geographic regions and calendar year is likely due to changes in population demographics, socioeconomic status, cultural influences, clinical practices, including those related to infection prevention and control, and surveillance processes. While some studies showed an increased incidence of SAB, others revealed decreasing rates over time. Overall, SAB incidence trends observed during the study period were stable.
The most recent systematic review of population-based studies was conducted in 2013 and assessed both overall bacteremia incidences and SAB incidence more specifically [6]. It showed high variability in rates according to the geographic location with an overall SAB incidence rate of 25 per 100,000 annually. Moreover, this review did not provide overall incidence trends of SAB or demographic data regarding the SAB population.
North America had higher percentages of MRSA causing SAB [20,24,26,27] as compared to Iceland and Denmark [10,15,32]. In the early 21st century, patients in Northern European countries were undergoing aggressive screening for MRSA with eradication in colonized patients [35]. The implementation of this new policy to search and destroy MRSA could account for the low percentages of MRSA in the region [10,15].
During the past two decades, the highest rates of SAB were observed in the Northern Territory of Australia's population [29]. When stratified by ethnicity, Aboriginal patients had a six-times higher rate of SAB than non-Aboriginal patients. This significantly correlates with socioeconomic disadvantages related to access to healthcare and education, as well as housing and employment in this population as decreasing socioeconomic status was linked to an increased risk of SAB [32,36].
A Danish study reported high incidence rates of SAB in an HIV-infected population that had IDU from 1995 to 2007 [37]. The percentage of IDU in a SAB population was the highest (28%) in Connecticut, USA in 1998 [26]. In contrast, a much lower (1.6%) percentage was reported more recently in MN, USA [27]; of course, this could be an underestimation since IDU can be underreported, or related to geographic heterogeneity related to opioid use in more rural areas of the USA [38]. The highest percentages (9% to 32%) of MRSA causing SAB were also observed in the USA [26,27], which could be explained by the rising IDU prevalence since onefourth of SAB caused by MRSA were IDU-related in the USA [39].
One-third of SAB cases were device-related [30]; the increasing utilization of indwelling foreign devices including vascular catheters, and orthopedic prostheses in clinical practice is a well-established risk factor associated with SAB [40]. Also, patients undergoing hemodialysis have an increased SAB incidence rate, particularly when chronic indwelling vascular catheters are used for hemodialysis access [23,41]. Of note, patients are most at risk during the first three months of hemodialysis [41].
There were several limitations to this systematic review despite its rigorousness. First, while multinational, studies included in our review do not represent a global aspect of SAB incidence since it does not include coverage of many geographic regions, namely Africa, Asia, and South and Central America, highlighting the need for population-based investigations from regions worldwide. Second, while our systematic review included all studies, regardless of geographic location, the inclusion of non-English language articles could have enhanced the likelihood of finding data from other countries, particularly developing countries. Third, not all included studies provided consistent data regarding their population demographics, comorbidities, and mortality rates. Fourth, some studies reported incidence trends of SAB, whereas others reported only one time point; this may have contributed to the absence of clear change in SAB incidence trends overall.

Conclusions
The current systematic review represents the most contemporary review of SAB incidence in 26 geographical population-based investigations during the past two decades. It shows an overall stable SAB incidence in the 11 countries included ranging from 9.3 to 65 cases/100,000/year. Multiple factors, both pros, and cons are likely responsible for the overall stable SAB incidence. Some of these factors vary in geographic location and prompt additional investigations from countries not included in the current review are required so that a more global characterization of SAB is defined.