- Top of page
- Materials and Methods
- Transparency Declaration
Haemophilus influenzae (Hi) are classified according to their capsular polysaccharides into six serotypes (a–f) and non-capsular strains . Encapsulated strains, especially Hi type b (Hib), are aggressive pathogens that primarily cause infections in children <5 years of age because of their lack of T-cell independent response to polysaccharides . This capsular type caused more than 95% of invasive H. influenzae infections in Scandinavian children before routine vaccination was initiated against Hib . On the other hand, Hi carrying other capsular types and non-capsular strains (Haemophilus sensu lato not serotype b, hereafter referred to as non-type b Hi) commonly cause invasive infections in persons with comorbidities [4,5]. In May 1989, routine vaccination against Hib was started in Iceland with the vaccine PRP-D (ProHIBiT®), followed by a dramatic decrease in Hib disease .
Despite the early success of vaccination against Hib, the concern remains that during longer follow-up, non-type b Hi with other capsular types or non-typeable Hi (NTHi) strains may replace Hib. Recent studies have been inconclusive regarding this matter [5,7,8]. Furthermore, with reduced pharyngeal carriage a natural booster to antibody production may be lost, resulting in waning immunity . The purpose of this study was to analyse trends in invasive disease caused by H. influenzae in Iceland for 26 years, from 1983 to 2008, at the 20-year anniversary of nationwide childhood vaccination against Hib in Iceland.
- Top of page
- Materials and Methods
- Transparency Declaration
During the 26 years of study, 202 cases of invasive Haemophilus infections were diagnosed in Iceland. During 1983–1989, before routine childhood vaccination against Hib, 123 infections were identified, 104 (85%) of them in children <16 years of age. In the post-vaccination era 1990–2008, 79 cases were diagnosed, 25 (32%) of them children. The annual numbers of Hib and non-type b Hi infections in Iceland, 1983–2008, are shown in Fig. 1. Age-specific incidence rates for invasive Hib and non-type b Hi disease, before and after Hib vaccination, are presented in Table 1. The yearly incidence of Hib dropped by 96% (95% CI, 92.8–97.6), from 6.4/100 000/year before 1990, to 0.3/100 000/year following adoption of routine childhood immunisation (p <0.0001). The incidence of non-type b Hi during 1983–1989 had a yearly incidence of 0.9/100 000/year and increased to 1.2/100 000/year post-Hib vaccination, but this was not a significant increase (p 0.16). The reduction in incidence of invasive Hib infections was most pronounced among children aged <1 and 1–5 years (p <0.05).
Table 1. Age-specific incidence for infections caused by Haemophilus influenzae type b (Hib) and other (non b Hi)
|Hib||Non-b Hi||Hib||Non-b Hi|
| n ||Incidence|| n ||Incidence|| n ||Incidence|| n ||Incidence|
About one in five of all Hib isolates tested for beta-lactamase production were positive (24 of 115 isolates tested, 21%), whereas 16% of non-type b Hi isolates tested were positive. The association between clinical diagnosis and species or serotypes, as well as timing of the infection, is summarised in Table 2. As shown, NTHi were most commonly isolated from patients with pneumonia or bacteraemia in the post-vaccine era, followed by H. influenzae type f (Hif), which primarily caused bacteraemic pneumonia. Only seven cases (3.5%) of all invasive disease were caused by H. parainfluenzae and the epidemiology of this species remained stable. Meningitis was the most frequent diagnosis (63/202, 31%) in the cohort, most commonly caused by Hib in children in the pre-vaccination period. Almost one-quarter (13/57, 23%) of the children with meningitis suffered from long-term sequelae, most commonly hearing loss. In the post-vaccination era bacteraemic pneumonia caused by non-type b Hi became the most frequent clinical presentation of invasive disease.
Table 2. Clinical diagnosis according to species, serotype and time interval
|a||b||c||d||e||f||NTHi|| Parainfluenzae ||Non b||Uncertain||Hib||Non-Hib||Hib||Non-Hib|
|Pneumonia||1||12||1|| || ||6||16|| || ||4||9||4||3||24|
|Meningitis|| ||58|| || || ||3|| || || ||1||55||2||3||2|
|Bacteremia||2||16|| || ||1||2||16||3|| || ||14||5||2||19|
|Epiglottitis|| ||4|| || || ||1|| || || || ||4|| || ||1|
|Cellulitis|| ||10|| || || ||2||1||1|| ||1||10||1|| ||4|
|OM/SA|| ||10|| || || || ||1||1|| || ||9|| ||1||2|
|Other|| ||5|| ||1|| || ||6||2||3|| ||1||2||4||10|
|NA|| ||8||1|| || || || || || ||2||7|| ||1||3|
Predisposing conditions of children and adults with Hib and non-type b Hi are summarized in Table 3. As shown, severe underlying disease was more frequent among both children and adults with non-type b Hi infections compared with those with Hib. Antecedent viral infection diagnosed clinically was commonly reported in children with Hib, whereas children with non-type b Hi were more likely to suffer from chronic conditions.
Table 3. Predisposing conditions in children and adults with Hib infections compared with non b Hi (n = 191)
|Hib, n = 101 (%)||Non b Hi, n = 24 (%)||OR (95% CI)||p value||Hib, n = 14 (%)||Non b Hi, n = 52 (%)||OR (95% CI)||p value|
|Smoking|| || || || ||6 (43)||19 (37)||1.3 (0.4–4.4)||0.7|
|Alcoholism|| || || || ||1 (7)||4 (8)||1.1 (0.1–29)||0.9|
|Viral symptoms||43 (43)||6 (25)||2.2 (0.8–6.5)||0.12|| ||4 (8)|| || |
|Pregnancy|| || || || ||3 (21)||6 (12)||2.1 (0.4–9.7)||0.4|
|Otitis media||36 (36)||2 (8)||6.0 (1.5–39.9)||0.01||1 (7)||1 (2)||3.8 (0.1–155)||0.4|
|CV disease||2 (2)||2 (8)||4.4 (0.4–44.4)||0.2||2 (14)||22 (42)||4.3 (1.0–31)||0.06|
|Lung disease||3 (3)||3 (13)||4.5 (0.7–28.4)||0.09||3 (21)||12 (23)||1.1 (0.3–5.6)||0.9|
|Cancer||1 (1)||2 (13)||8.9 (0.6–270)||0.1||2 (14)||14 (27)||2.2 (0.5–15)||0.4|
Among 18–40-year-old women, 13 were diagnosed with Haemophilus disease, nine of whom were pregnant, and five of them lost their foetus (median gestational age 12 weeks compared with 40 weeks for those who gave birth). According to nationwide data on fertility, late foetal deaths and induced abortions, combined with previously published data on frequency of spontaneous abortions , the incidence of invasive Haemophilus disease in pregnant women was 9.2 infections/100 000/year (nine infections during 98 014 person-years), whereas non-pregnant women in the same age group had an incidence of 0.4/100 000/year (four infections during 1 118 167 person-years). The rate ratio between the two groups is 25.7 (95% CI, 8.0–95.9; p <0.0001). Non-type b Hi caused 6/9 of the infections in pregnancy and 8/11 of the neonatal infections, where the majority was caused by NTHi (5/6 and 4/8, respectively). The infection was fatal in two neonates (NTHi and non-type b Hi) while all the pregnant women survived.
Of the 14 fatal cases, three were caused by Hib; one severely disabled child had not received vaccination, in the post-vaccine era, and two adults in the pre-vaccine era died from septicaemia and pneumonia. Thus, the case fatality rate for Hib was 2.4%. The case fatality rate for non-type b Hi was 14%. The highest fatality rates were observed at the extremes of age, <1 year (2/10, 20%), 41–60 years (2/9, 22%) and >60 years (7/33, 21%). Of the 11 non-type b Hi strains associated with fatal infections, two were encapsulated, serogroups c and f, six were NTHi and three were only classified as non-type b Hi. The most common causes of death were pneumonia (7/11) and bactaeremia (3/11).
- Top of page
- Materials and Methods
- Transparency Declaration
This study confirms a remarkable and sustained success of Hib conjugate vaccination in Iceland. During the two decades of follow-up where all species of invasive Haemophilus spp. were included, an increase in Hib infections was not noted. This contrasts with the early experience from the United Kingdom [11,12]. In Iceland a booster dose was used from the beginning, which may explain this difference . During the last 5 years of our study, Hib was all but eliminated, occurring in only one of 24 (4%) invasive infections. Similar to the results of others, the reduction in incidence of Hib was most pronounced among young children, and from 1990 onwards non-type b Hi infections became more than twice as common as Hib in young children [9,12]. Interestingly, the incidence of invasive Hib infections also decreased markedly in the elderly after childhood vaccination was initiated. This reduction is most likely caused by the herd effect, where the unvaccinated elderly benefit from the reduction in nasopharyngeal carriage of Hib in vaccinated children, leading to reduced exposure and subsequent infections [13,14]. Most studies have concentrated on the age-group 0–4 years although Perdue et al. and Peltola et al. [3,15] also demonstrated a reduction in incidence of invasive disease among adults in Alaska and Scandinavia.
The age group 21–40 is of particular interest, childbearing women in particular. It has been proposed that the immunosuppressive effects of pregnancy increase the susceptibility of the host to H. influenzae infections . We found a significant increase in incidence among pregnant women compared with other women of the same age. A prospective study found that child-bearing women in the age group 18–39 years had a six-fold increased risk of H. influenzae bacteraemia compared with adults of the same age . Our results show that the relative risk associated with pregnancy may be even higher. Carriage studies have shown that only 1.8/1000 parturient women carry non-encapsulated Hi in their genital tract , but they were found in 8/110 women with preterm rupture of membranes . It has been suggested that unusually virulent NTHi biotype II or IV strains may preferentially cause invasive infections among neonates and pregnant women [19–21]. Further characterization of the strains in our study may therefore be of interest. The incidence of non-type b Hi infections among infants younger than 1 year from 1990 to 2008 was 8.9/100 000/year, which is similar to the results of Ladhani et al. .
Currently, NTHi are the most common types of Hi causing invasive disease in Iceland, similar to results of large-scale studies from Australia, Canada and Denmark . Hif are currently the most common encapsulated invasive H. influenzae isolates in Iceland. Hif infections are similar to NTHi in that 10/14 (72%) cases had severe predisposing conditions and the most common clinical diagnosis was pneumonia. Other studies have reported a high prevalence (60–80%) of predisposing conditions among patients with Hif infections [12,23] and they are most frequent in young individuals and the elderly [4,23]. Resman et al.  recently reported a significant increase in serious invasive infections ascribed to Hif, mainly in individuals >60 years of age in Sweden. This resonates well with our study, as 9/14 (64%) Hif patients were >60 years of age. A recent study from Utah shows the emergence of H. influenzae type a as the most common in children, leading to a disease similar to Hib . A similar increase was not noted in our study.
Viral infections of the upper airway make patients more prone to colonization and infection with both Hib and non-type b Hi [26,27]. Interestingly, a higher proportion of patients with Hib reported a history of preceding viral infection, which may be explained in part by the different age composition of the two groups. However, a similar difference was also noted with respect to otitis media, which was the second most frequent predisposing condition among children with Hib infections, 36/101 (36%), while children with NTHi infections had a history of otitis media in only 2/24 (8%) cases (p 0.01). This is somewhat surprising, because NTHi frequently cause otitis media in children , but this may be indirect evidence of the relative lack of invasive potential of these strains. Similarly, children who were infected with non-type b Hi were more likely to have severe predisposing conditions. This is in agreement with other studies and reinforces that non-type b Hi are opportunistic pathogens with less virulence compared with Hib [4,7,12]. The most common predisposing condition among adults with Hib infections was smoking, which has been linked to increased susceptibility to infections . As other authors have concluded, adults with non-type b Hi infections frequently had serious predisposing conditions, most commonly coronary artery disease, lung disease and malignant disease [8,12,29].
The age-related fatality rate for patients with Hib was 2.4% and for non-type b Hi infections it was 14%, being highest at the extremes of age, which is in accordance with data from Scandinavia, the United Kingdom and the USA [5,11,30]. With the availability of a conjugated vaccine for NTHi, it may therefore be prudent to study the effectiveness of this vaccine in select groups of adult patients with predisposing conditions, including young women. The strength of our study is its population-based design, giving a clear picture of the epidemiology of invasive Haemophilus disease before and after vaccination for Hib. Despite the long time-frame, clinical information was still available for 95% of the cases. The main limitation of our study is the limited power in statistical analyses of small subgroups.
In conclusion, we present a population-based study of invasive Haemophilus disease in Iceland, spanning 26 years. Following vaccination, which started in 1989, invasive infections caused by Hib were reduced by 96%, in part by the herd effect. The concern that other non-type b Hi species might replace Hib has not materialized after 20 years of follow-up. Pregnancy is an important risk factor, which may be amenable to prevention, either by selective screening and antibiotic therapy or by vaccination. The high case fatality rate of non-type b Hi infections in the elderly with underlying conditions also raises questions regarding the potential efficacy of vaccination for NTHi in this risk group.