Incidence of IPD in young children vary widely worldwide [2–4]. Decline has been observed in countries which introduced immunization programs based on the PCV7 [5–8]. Continuous surveillance of S. pneumoniae isolates causing invasive disease is necessary to evaluate the actual effect of vaccine  and to identify the circulating serotypes. This is the first active surveillance study conducted in Lombardy, Italy, to estimate the incidence of IPD in children aged less than 5 years and to determine the serotype distribution of S. pneumoniae isolates. In the monitored area the estimated annual incidence of IPD in children aged less than 5 years was 50.0/100,000. It was around 31% higher in infants than children aged 2–4 years.. In Italy, few prospective regional studies have been conducted [18–21]. In children aged less than 5 years, incidence of IPD varied from 2.8/100,000 in Puglia (South Italy) to 5.7/100,000 in Piemonte . In children aged ≤36 months the incidence was 58.9/100,000 in North-East . The incidence of IPD in young children reported from different European countries before the introduction of pneumococcal vaccination ranged from 3.1/100,000 in Portugal to 110.2/100,000 in Spain (pooled 31.1/100,000) , with higher values in infants. Rates of IPD were lower in Europe than USA, where blood culture is performed for every febrile young child. In 1998–1999, before of introduction of the PCV7, the IPD incidence in children aged less than 5 years was 96.7/100,000. After the introduction of PCV7, incidence declined to 23.9/100,000 in 2003, with a percentage reduction of 75% . Although in Europe there is lack of extensive epidemiological data on the effect of PCV7, a recent study estimated that an infant vaccination program would prevent 39,000 IPD cases in the 20 years after PCV7 introduction in the UK . In Denmark, one year after introduction (2008) of PCV7, incidence of IPD declined of about 57% in infants and of 10% in children aged 2 years or older . In the Netherlands a reduction of 44% was found two years after introduction of PCV7 in infants . The PCV7 was introduced in Italy in 2001 and firstly recommended in at-risk children aged less than 5 years. Currently, either risk-based or universal vaccination programs are used, depending on the region, and PCV7 is offered either free of charge or with cost sharing . Unfortunately, there is still paucity of information on the epidemiological effect of PCV7 .
Heterogeneity of IPD serotypes was observed in this study. Less than a third of the isolates belonged to serotypes covered by PCV7 while around 85% of them was covered by PCV13. In particular, the serotype 19A accounted 23.1% of typified IPD isolates. This result is comparable with recent national pooled data that estimated 19.2% of pneumococcal meningitis or sepsis occurred among children aged less than 5 years during in 2009 in Italy being caused by the serotype 19A . A recent study conducted in Massachusetts showed throughout 2001–2007 period an increase in frequency of serotype 19A, that was the most cause of identified IPD cases in children (28%) . The relevance of this emerging serotype, included in PCV13 but not in PCV7, has been as well indicated in South-West Asian countries [12, 13], and emphasized recently [14, 15, 31, 32].
The coverage of PCV7 was lower in this study (30.8%) compared to values reported from European countries. In Europe, PCV7 coverage in children aged less than 5 years ranges from 50% in Spain and Sweden to 100% in Greece, with a mean of 71%. In infants, the coverage of PCV7 ranges from 37% in Norway to 100% in Greece, with a mean of 72% . The introduction of the PCV13 showed average increments of additional coverage of about 16% in children aged less than 2 years over PCV7 [2, 27]. A study conducted in Spain in children aged less than 5 years showed that compared to PCV7 the PVC13 would increase the coverage by 17% (due to serotypes 1, 3, 5 and 7F) . It may be handy to point out that in the population examined in this study, the 23-valent pneumococcal polysaccharide vaccine would show the same coverage of PCV13. However, it should be noted that current international guidelines recommend in children aged less than 5 years initial administration of the conjugate vaccine, above all in infants [25, 34]. Indeed, infants may show poor antibody response to the polysaccharide vaccine  and level of antibody may decline already one year after polysaccharide vaccination .
The antimicrobial susceptibility of different serotypes of S. pneumoniae isolates may vary across geographic areas. In this study about 31% (4/13) of typified isolates was resistant to erythromycin (serotypes 14, 19A and 19F), 23% (3/13) was resistant to tetracycline (serotypes 19F and 15) and one isolate (serotype 15C) was resistant to penicillin G. In Europe the rate of penicillin G resistant serotypes in children aged less than 5 years ranges from 5% in Denmark to 55% in the Czech Republic (average 29%) . Penicillin G resistance was often associated with the serotypes 19A (especially in Germany) and 14 in Poland . For children aged less than 5 years, the mean proportion of erythromycin-resistant isolates was 35% (range 7% in Denmark to 53% in Spain), and erythromycin resistance was often associated with pneumococcal serotype 14 . A warning may emerge from this study about the serotype 15C, as also reported in USA . This serotype, currently not included in any available pneumococcal vaccine, caused meningitis and was resistant to penicillin and tetracycline.
Lastly, it is important highlight that all four IPD children, who had received pneumococcal vaccine previously, exhibited serotypes (19A and 1) covered by PCV13 but not by PCV7. On the whole, this information and the above results might be helpful in determining which conjugate vaccine is preferable for this population. Caution should be exercised anyway in inferring any definite conclusion due to the small number of isolates examined (n = 13). The small sample size of the current data base of IPD cases is a limitation, indeed. Continuous active surveillance, possibly based on a more branched network would be desirable to evaluate accurately the burden of disease caused by S. pneumoniae in this population, and to assess the effect of the vaccine on the incidence of IPD.