Rates of, and risk factors for, septic arthritis in patients with invasive pneumococcal disease: prospective cohort study

Streptococcus pneumoniae is thought to account for about 3-5% of all cases of septic arthritis [1]. There are many case reports of this entity [2] and there have been several reviews, the most comprehensive of which examined the features of 190 cases (108 adults) reported between 1965 and 2003 [3]. While the clinical features of this entity have been adequately described there has been no study that has compared patients with invasive pneumococcal disease (IPD) with septic arthritis to those who didn’t develop septic arthritis. Such a comparison would allow a delineation of the risk factors for septic arthritis. A 15-year study of IPD in Northern Alberta gave us the opportunity to do this. In addition to examining rates of, and risk factors for, septic arthritis we hypothesized that patients with septic arthritis would have more severe infection as manifested by intensive care admission, higher complication rate and higher mortality rate.

Over the course of the study 3251 patients had IPD of whom 51 (1.6%) had septic arthritis. The patients with septic arthritis were less likley to walk independently and less likely to have chronic obstructive pulmonary disease (Table 1). 27.4% of the septic arthritis patients had underlying joint disease compared with 8.5% in the non-septic arthritis group (Table 1). In 9 (18%) patient’s joint involvement was polyarticular. The knee was the most frequently infected joint – 26 (51%), followed by shoulder 7 (13.7%); elbow 6 (11.7%); wrist 6 (11.7%), hip 4 (7.8%); ankle 4 (7.8%) and acromioclavicular joint 1 (1.9%). In multivariate analyses, septic arthritis patients were more likely to have osteoarthritis (adjusted odds ratio (aOR) 3.4, 95% confidence interval (95%CI) 1.55-7.34) and rheumatoid arthritis (aOR 5.05, 95%CI 1.85-13.78), more likely to be males (aOR 2.07, 95%CI 1.12-3.83), lesslikely to walk independently (aOR 0.48, 95%CI 0.24-0.96) and less likely to have chronic obstructive pulmonary disease (aOR 0.14, 95%CI 0.03-0.57). No other characteristics were associated with septic arthritis (p > 0.1).

Table 2 shows that the septic arthritis group were less likely to be admitted to ICU which remained after adjustment for potential characteristics in Table 1 (aOR 0.41, 95%CI 0.18-0.92), P = 0.03) There was no difference in other measures of severity of illness such as mechanical ventilation, acute kidney injury, empyema, cellulitis, bacteremia in either univariate or adjusted analyses between those with and without septic arthritis. With respect to laboratory data, no differences were observed between those with and without septic arthritis in either univariate or multivariate analyses with the exception that septic arthritis patients tended to have higher platelet counts even after adjustment for potential characteristics in Table 1.

Table 3 shows the five most common serotypes of S. pneumoniae causing infection in the septic arthritis group. Two of these serotypes, 22F and 12F were more frequent in the septic arthritis group than in those without septic arthritis.

This study found that septic arthritis was not that common (1.6% incidence in this cohort of IPD patients). Although this is the first study we are aware of that focused exclusively on rates of septic arthritis among patients with IPD, previous studies of septic arthritis patients have reported that pneumococcal septic arthritis accounts for 3-8% [1‐3, 7] of all septic arthritis cases. Factors associated with septic arthritis in patients with IPD included underlying joint disease, male sex, and mobility.

Pre-existing joint disease has been documented in other studies as predisposing factors for septic arthritis [1‐3]. Furthermore, we suspect that many of the patients who reported limitations in walking did so as a result of (osteo)arthritis which may not have been captured in the medical record. In keeping with other reports, we found that the knee is most the commonly involved joint and that polyarticular infection was not uncommon [1‐3]. A damaged joint is more likely to become infected in patients with bacteremia than an undamaged joint. In a prospective study of 75 patients with septic arthritis 46 (61%) had underlying joint disease [8]. In our study (27.4%) of the patients with pneumococcal septic arthritis had underlying joint disease compared with 8.5% of those without septic arthritis. The finding that male sex is at increased risk is not consistent with previous studies where males and females appear to have similar risk or potentially a slight female predominance [1‐3, 9].

Although previous studies have suggested meningitis may be a risk factor for septic arthritis [1, 7], among our IPD patients we observed the opposite with no patients having concurrent meningitis and septic arthritis. The reason why meningitis could be a risk factor for septic arthritis is not well understood. One possibility is that high bacterial load may be more likely to result in seeding of the joints and the meninges.

With respect to microorganism specific factors, while 4 of the 5 most common serotypes in septic arthritis patients were more frequently represented compared with non-septic arthritis patients none achieved statistical significance. Interestingly the two serotypes that were significantly higher in septic arthritis patients, 12 F and 22 F, are high case fatality rate serotypes [10]. Thus, whether or not there are organism specific factors remains to be determined.

A strength of this study is the large number of patients with IPD and extensive clinical data. A limitation of our study is that we did not measure bacterial load in patients with IPD. It is possible that a higher bacterial load results in seeding of the joints. In addition, we did not undertake detailed examinations of fluid and aspirates for alternate etiologies of inflammatory arthritis in this population of IPD patients, such as gout or pseudogout. Last, although patients reported limitations in ambulation prior to hospitalization, we do not know to what degree this is simply a proxy measure for undiagnosed joint disease or perhaps the presence of a prosthetic joint.

In conclusion, in patients with IPD, septic arthritis is uncommon. Certain risk factors such as inability to walk independently without assistance and underlying joint disease make biological sense as damaged joints are more likely to be infected in the presence of bacteremia.