Streptococcus pneumoniae is the most frequent isolate from clinical samples of respiratory tract infection, including acute exacerbations of chronic bronchitis and community-acquired pneumonia 123. The emergence of multiple drug-resistance has complicated the empirical treatment of pneumococcal infections. Several surveillance programs that span numerous countries indicate that the proportion of drug-resistant Streptococcus pneumoniae isolates continues to increase worldwide4. However, in Spain, where there was a great increase of penicillin-resistance in 1980s, stabilization in 1990s and a decrease in penicillin-resistant isolates in the last few years has been reported 567.

In elderly people, the population group with highest rates of pneumococcal infection, the penicillin intermediate-susceptibility strains has not been associated with increased mortality or unsuccessful treatments, so these strains can be treated with β-lactams. If we consider penicillin-resistant pneumococci, data reported on mortality has not been uniform489.

Updated data about drug-resistance in Streptococcus pneumoniae isolates from different populations may be important in recognizing changes in disease patterns. The present study describes current levels of drug resistance among Streptococcus pneumoniae causing community-acquired pneumonia in Spanish middle aged and older patients, and evaluates clinical factors associated with increasing drug resistance or mortality.

Hospital case-series that included 104 consecutively recruited isolates of Streptococcus pneumoniae recovered from patients 50 years or older with a radiographycally confirmed community-acquired pneumonia in the region of Tarragona (a mixed residential industrial urban area in the Northeast Mediterranean coast of Spain) with Antimicrobial susceptibility to penicillin tested between January 2002 and April 2007.

The isolates of S. pneumoniae were identified from blood and sputum samples in the microbiology laboratory of two reference hospitals in the study area (Joan XXIII and Sant Pau i Santa Tecla) by standard methods. Pneumococcal serotyping was carried out by the Quellung reaction using the Statens Serum Institute (Copenhagen, Denmark) typing sera. Penicillin susceptibility was determined by the Kirby-Bauer disk diffusion method with a disk containing 1 μg oxacillin (BD BBL, Sparks, MD USA). Isolates were defined as resistant when the growth inhibition diameter was 19 mm or less. Minimal inhibitory concentration (MIC) for penicillin and cefotaxime was determined by E-test method (AB Biodisk, Solna, Sweden) following the manufacturer's instructions. Penicillin-susceptible was considered when MIC ≤ 0.06 μg/mL, penicillin intermediate-resistant when MIC 0.12 – 1 μg/mL and high resistance when MIC ≥ 2 μg/mL. Cefotaxime susceptibility was considered when MIC ≤ 1 μg/mL, cefotaxime intermediate when MIC 2 μg/mL and cefotaxime resistant when MIC ≥ 4 μg/mL. Erythromycin and levofloxacin susceptibility was determined by the Kirby-Bauer disk diffusion method with a disk containing 15 μg erythromycin and 5 μg levofloxacin respectively (BD BBL, Sparks, MD USA). Erythromycin susceptibility was considered when the growth inhibition diameter was ≥ 21 mm, erythromycin intermediate between 16 and 20 mm, and erythromycin resistant ≤ 15 mm. Levofloxacin susceptibility was considered when the growth inhibition diameter was ≥ 17 mm, levofloxacin intermediate between 14 and 16 mm, and levofloxacin resistant ≤ 13 mm, according to the guidelines established by the National Committee for Clinical and Laboratory Standards (formerly NCCLS)10. Multidrug resistance was defined as intermediate resistance or resistance to penicillin plus intermediate resistance or resistance to ≥ 2 antimicrobial agents11.

Cases were validated by checking clinical records and, according to the presence of underlying medical conditions, all patients were grouped into 3 risk strata on the basis of the degree of immunocompromise and risk for pneumococcal disease. High-risk level included persons with conditions associated with possible immunocompromise: immunodeficiency (including AIDS), asplenia, cancer (solid organ or haematological neoplasia), chronic nephropathy (nephrotic syndrome, renal failure, dialysis or transplantation), and long-term corticosteroid therapy (20 mg/day of prednisone or equivalent). Moderate-risk level included patients without a level 1 condition but who had a history of chronic lung disease (chronic bronchitis, emphysema or asthma), liver disease (cirrhosis or alcoholic hepatitis), heart disease (congestive heart failure or chronic angina) and diabetes mellitus. Low-risk level included patients without the above mentioned risk conditions.

Pneumonia severity index (PSI) was calculated according to criteria described in classical meta-analysis, and case-fatality was considered when the patient died (in-hospital or not) within the first 30 days after the diagnosis of pneumonia1213.

Penicillin-resistance was determined for early cases (2002–2004) and contemporary cases (2005–2007). For the statistical analysis, continuous variables were compared by Student's t test, whereas categorical values were compared using chi-squared or Fisher's test as appropriate.

The study was approved by the ethical committee of the Catalan Health Institute and was conducted in accordance with the general principles for observational studies set out by this institution (Exp FIS 050231).

In the present study, the overall rate of penicillin-resistance among isolates of Streptococcus pneumoniae was considerable (25.9%). The proportion of penicillin-resistance was higher among early isolates than in contemporary isolates, which concurs with other Spanish studies5714. Among Spanish individuals, Fenoll et al7 found that a high-level penicillin-resistance rate (MIC > 2 mg/l) decreased from 13.3% in 1997 to 12.9% in 2001. In Catalonia, the high-level penicillin-resistance fell from 15% (1989–1993)3 to 8.9% (1999–2002)14, which fits with the 9% found in our study. In contrast, the results found in other countries showed an increase in penicillin-resistance rate. In United States, the high-level penicillin-resistance increased from 14.7% (1998–1999) to 18.4% (2001–2002)15. Although several factors explain the emergence of drug resistant Streptococcus pneumoniae, the single most important consideration, however, is the selective pressure of use of antimicrobial agents. Reductions in the rate of resistance found in our study could be explain by a reduction in consumption of antibiotics in Catalonia, going from 17.77 defined daily doses per 1000 inhabitants per day (DHD) in 1999 to 12.58 DHD in 2006, and could be partially attributed to the vaccination that covers serotypes typically associated with drug resistance14.

We found a statistically higher age among patients infected with penicillin-resistant strains than in those infected with penicillin-susceptible strains. However, because of low statistical power, we could not reject that prevalence of underlying conditions was the predisposing cause for this age-related differences in the aetiology of pneumococcal pneumonia.

In accordance with previous studies514 we found a high proportion of underlying diseases among patients with pneumococcal pneumonia. Although not statistically significant, the proportion of penicillin resistance in our study was higher among patients with low or moderate level risk in early isolates. Conversely, in contemporary isolates this proportion was higher among patients in high level of risk. Some studies616 have reported that patients with chronic pulmonary disease, HIV infected, suspected aspiration or who were admitted to the hospital within the previous 3 months were more likely than other patients to have penicillin resistant pneumococcal pneumonia. In our study, only two patients had been hospitalized in prior 3 months, and none of them were HIV infected. Chronic pulmonary disease was highly prevalent in our study but differences were not observed in prevalence by penicillin resistance status.

Almost three quarter (74.1%) of penicillin-resistant pneumococcal pneumonia were PSI classes IV-V. There was a similar percentage of penicillin-resistant pneumonia between PSI classes IV-V and PSI classes I-III. This results agree with those found in other studies616 that only reported a trend for higher PSI classes to be associated with penicillin-resistant pneumococcal pneumonia.

In the present study, the rate of penicillin susceptibility was higher in bacteremic than non-bacteremic cases (65.8% vs. 42.9%). These fits with Aspa's report6, where penicillin susceptibility was 71% among blood culture isolates and 50% among sputum isolates. In the USA, a higher prevalence of resistance to penicillin among lower respiratory tract than among isolates recovered from blood samples (16.2% vs. 13.7%) has also been reported15.

In our study, most penicillin resistant S. pneumoniae isolates were resistant to at least one additional antimicrobial agent, and 3.8% (four cases) of all isolates were identified as MRSP. As described in previous studies penicillin resistance, particularly high-level resistance, predicts resistance to other antimicrobials1114171819. We found that all MRSP were highly resistant to penicillin and cefotaxime, and three of them were highly resistant to erythromycin too.

We found that an inadequate initial antimicrobial treatment was started in 7.7% of patients. Although in our study there was no excess mortality due to initial inappropriate treatment (12.5%), likely due to a small sample size, it is important to take into account underlying condition or regional pattern of drug resistance for S. pneumoniae in determining adequate antimicrobial therapy111720.

In accordance with other studies5142122, penicillin resistance was not related with higher mortality. It has been suggested that the primary determinant of virulence might be the serotype rather than its sensitivity to penicillin, and it is possible that the mechanisms conferring penicillin resistance are related to those leading to a loss of virulence. Previous studies of pneumococcal bacteremia did not show differences in mortality between those with susceptible and those with non-susceptible pneumococci after adjustment for age, underlying disease, severity of illness on presentation, and appropriate concordant treatment5616.

We found that three serotypes (14, 6B and 23B) were associated with increasing risk of penicillin resistance, accounting for a 53% of serotyped pneumococci causing penicillin-resistant infections and the serotype 6B was responsible for 50% of cases of MRSP. In accordance with our results, Fenoll et al7 reported that 56% of penicillin-resistant pneumococcal strains belonged to serotypes 6 (18%), 14 (19%) and 19 (19%). In a long term follow-up between 1983 and 2003 Mufson et al23 found that all blood isolates of penicillin-resistant were capsular serotypes 6, 9, 14, 19 and 23.

Main limitation in this study includes the relatively low number of cases and the difficulty to generalize results. According to classical breakpoint criteria, it could be concluded that, in the study area, resistance to penicillin among Streptococcus pneumoniae remains high, although such resistance does not result in increased severity or mortality. However, it must be emphasized that, according to the new breakpoint criteria for intravenous penicillin24, almost all strains in our study would be susceptible.

To date, criteria for penicillin susceptibility of Streptococcus pneumoniae remains unchanged for patients without meningitis who can be treated with oral penicillin (e.g., for outpatient pneumonia). However, since 2008, for patients without meningitis who are treated with intravenous penicillin, the new breakpoints are ≤ 2, 4, and ≥ 8 μg/mL for susceptibility, intermediate-resistance and high-resistance, respectively24. Thus, considering the new pneumonia breakpoints, 100 (96.2%) of the 104 cases in our study would be penicillin-susceptible (MIC ≤ 2 μg/mL), 4 (3.8%) would be penicillin intermediate-resistant (MIC = 4 μg/mL) and none would be high resistance (MIC ≥ 8 μg/mL).

According to the change in pneumonia breakpoint for susceptibility24, percentages of strains resistant to penicillin have dramatically decreased worlwide. In Spain, a large surveillance study that included 2,721 isolates of S. Pneumoniae obtained from November 2001 to October 2002 in diferent Spanish Hospitals25 reported that (taking into account the old pneumonia breakpoint) 20% of overall strains were penicillin resistant (MIC >2 mg/mL) and 24% were intermediate resistant strains (MIC = 0.12–1 mg/mL). However, if we consider the new pneumonia breakpoint, 94.3% of the total strains in the forementioned study would be currently considered penicillin susceptible, which is also in agreement with data found in the present study (96.2% penicillin susceptible strains in our study considering the new breakpoint).

According to classical pneumonia breakpoints for susceptibility to penicillin, the resistance among Streptococcus pneumoniae was considerable, but with the use of new breakpoints, the percentage of Streptococcus pneumoniae penicillin-resistant decreases dramatically.

In vitro penicillin resistance is not associated with increasing mortality, even considering new breakpoints for susceptibility to intravenous penicillin.

The low mortality rates reported among patients with pneumonia due to penicillin resistant strains could likely be due to the fact that patients were treated with active agents, such as penicillins or third generation cephalosporins.

The authors declare that they have no competing interests.

AV-C, FB, ES, OO-G and CdD designed the study, assessed outcomes, and wrote and edited the paper. FG-B, XR and XC obtained the data. VA did the statistical analysis. AV-C coordinated the study.