In order to investigate any correlation between the different serotypes/serogroups and age, two-tailed Likelihood Ratio (LR) and a multivariable logistic regression was performed. Serotype/serogroup was significantly associated with age (≥65 years; P < 0.001). Subsequent single serotype analysis showed that cases with serotypes/serogroups 6A, 23F, 6B, 11, 14 and 15 infection were most significantly (OR > 2) associated with the age ≥65 years compared to those

infected Rigosertib in vivo with the serotypes of the reference group (1 and 7F) ( Fig. 1A). Serotype was also associated with case fatality (P = 0.001) and scrutinizing the individual serotypes revealed that serotypes 3, 19A and 19F were saliently associated with increased case-fatality, compared to the reference

group ( Fig. 1B). As for the manifestations, suffering from pneumonia (P < 0.001), meningitis (P < 0.01) and bacteremia without focus (P < 0.01) was associated by serotype, too. IPD due to serotypes/serogroups 35, 23, 19F and 15 infection were clearly (OR > 2) associated with a bacteremia without focus compared to infection with a reference group serotype Cyclopamine 1 and 7F ( Fig. 2A). In addition, meningitis was associated with serotypes 35, 15, 11, 18C and 23F (OR > 6) compared to the reference group. These findings were independent of age, sex and number of comorbidities. As for pneumonia, none of the serotypes was more likely than the chosen reference group. In more detail, serotypes 15, 35, 18C, 19F, Resminostat 23, 23F, 6B and 11 were the rarest and resulted in OR < 0.5. As for morbidity, serotype was associated with different numbers of comorbidities (i.e. having at least one versus no comorbidity; P < 0.001). Results displayed that cases infected

with serotypes other than serogroup 8 suffered from one or more comorbidities significantly more often than those infected with the serotypes of the reference group (1 and 7F) ( Fig. 2B). Among these serotypes/serogroups OR were highest for 23, 35, 6B, 19F and 20. Of those, serogroups 20 and 35 are neither covered by PCV7 nor PCV13. Regarding type of comorbidity, immunosuppression (P < 0.001) but not chronic diseases (P = 0.2) and pre-existing underlying respiratory disease (P = 0.4) were significantly associated with serotype using the two-tailed Likelihood Ratio (LR) test. As for the first, cases infected with serotypes/serogroups other than 4 and 8 were more often immunocompromised than those infected with the serotypes of the reference group (1 and 7F) ( Fig. 2C). This population-based study evaluates the serotype epidemiology of invasive S. pneumoniae isolates, from 2003 to 2012 including association of causing serotype with IPD characteristics and case-fatality in adult Swiss residents aged ≥16 years reported from 2007 to 2010. The study period for the latter covered the years after recommendation of the complementary vaccination with PCV7, but before recommendation of PCV13 for infants [13] and [22].