Haemophilus influenzae (Hi) was first identified as a pathogen in 1883 by Koch who described small Gram-negative rods in pus from patients with conjunctivitis. 1 In 1931 Pittman 2 identified nonencapsulated strains of Hi as well as six antigenically distinct encapsulated strains that could be identified serologically by characteristics of the polysaccharide capsule, now defined as serotypes a to f. She also recognized that one capsular type, serotype b, was responsible for almost all cases of Hi meningitis. Subsequently population-based studies have confirmed that H. influenzae type b (Hib) is responsible for ∼95% of all invasive Hi disease in unvaccinated children. 3, 4 The recent introduction of Hib conjugate vaccines has led to a substantial fall in the incidence of Hib disease in childhood. 5
With the decline in Hib disease, non-type b Hi have become a more important cause of Hi disease. Noncapsulated, i.e. nontypable (nt), strains are well-recognized causes of sinusitis and otitis media. 6 They are also capable of causing invasive diseases such as bacteremia, 7 pneumonia 8 and meningitis. 9 They are particularly recognized as a cause of neonatal sepsis. 10, 11 Serotype f (Hif) is reported as the most common cause of invasive encapsulated non-b Hi disease in children. 12 Serotype a is a rare cause of pneumonia and meningitis 13 and there are even less frequent reports of invasive disease in children caused by Hi serotype c (Hic), 14 Hi serotype d 15 and Hi serotype e (Hie). 16, 17
Since October 1, 1992, Hib conjugate vaccines have been used routinely in the UK and the Republic of Ireland (ROI). After the introduction of routine Hib vaccination, we initiated a prospective national study to detect cases of invasive Hi disease in vaccinated and unvaccinated children. This allowed us to assess the clinical and epidemiologic factors associated with non-type b H. influenzae disease in all British and Irish children and in vaccinated children to compare the features of non-type b disease cases with those of type b disease cases.
This study was performed under the auspices of the British Paediatric Surveillance Unit (BPSU) of the Royal College of Paediatrics and Child Health. The BPSU has a program of active surveillance for selected rare pediatric conditions. Every month >90% of pediatricians routinely report to the BPSU using the report card sent to them on a monthly basis. 18
Microbiologists and consultants in communicable disease control were also informed about the study and contributed to surveillance. In the ROI there was also active laboratory surveillance in which the investigator telephoned all pediatric laboratories every 2 weeks.
Pediatricians were asked to provide early notification by telephone. They were sent a questionnaire requesting clinical, demographic and laboratory information.
For the first phase of the study (October 1, 1992 to October 31, 1995), reports were requested of any child with invasive Hi disease who had received Hib conjugate vaccine. From November 1, 1995, the case definition was broadened to include all children with invasive Hi disease regardless of vaccination status. Invasive disease was defined as isolation of Hi from a normally sterile site (such as blood, cerebrospinal fluid, joint aspirate) or a positive Hib antigen test (in cerebrospinal fluid) combined with a clinical picture compatible with meningitis. The diagnosis of pneumonia required the presence of radiologic changes together with positive blood cultures for Hi. In the case of non-type b isolates the case was not included if other organisms were isolated at the same time.
The dates of all primary immunizations were obtained from the child’s general practitioner or from the district child health immunization computer records.
The local microbiologist was asked to send the isolate to the Public Health Laboratory Service Haemophilus Reference Laboratory in Oxford where the identity was verified by standard serotyping and PCR techniques. 19 In the ROI isolates were sent to Oxford via the microbiology department at the Waterford General Hospital.
To encourage maximum reporting of cases, the study was widely covered in the medical press before its commencement, and pediatricians were then updated through quarterly BPSU bulletins and an annual BPSU report. 18 Updates were also posted in the Communicable Disease Weekly Report for England and Wales and the Weekly Report of the Scottish Centre for Infection and Environmental Health. Reports of Hi in children made to other agencies were checked against those reported to the study including the two pharmaceutical companies supplying the vaccines, the UK Department of Health Medicines Control Agency, the Public Health Laboratory Service Communicable Disease Surveillance Centre for England and Wales and the Scottish Centre for Infection and Environmental Health.
We compared the clinical features of type b and non-type b disease in fully vaccinated children. A case of Hi disease in a fully vaccinated child was defined as invasive disease occurring >1 week after two or more doses of Hib conjugate vaccine had been given to a child in the first year of life.
Statistical analysis was performed using SPSS (SPSS, Inc., Chicago, IL) and EPI-Info Version 6 (CDC, Atlanta, GA). Ages are given as medians (range) and compared with the Mann-Whitney test. Proportions are compared using the chi square test or Fisher’s exact test where an expected number was <5. No adjustments were made for multiple comparisons. Population and birth figures were obtained from the Office for National Statistics. During the period of the study the annual UK birth cohort varied between 717 000 and 781 000.
This study was approved by the Central Oxford Research Ethics Committee.
During the 6 years and 3 months between October 1, 1992, and December 31, 1998, there were 147 cases of invasive non-type b Hi disease reported in vaccinated children. During the 3 years and 2 months between November 1, 1995, and December 31, 1998, there were 84 cases of invasive non-type b Hi disease reported in unvaccinated children.
Hi disease in vaccinated children
Of the 147 non-b isolates in vaccinated children, nt strains accounted for 78% (115), Hif for 16% (23), Hie for 5% (7) and there was 1 each of Hi serotype a and Hi serotype c. The median age of presentation for all non-b cases was 16.8 months (2.4 to 95.2 months), and the most common modes of presentation were bacteremia (34%), meningitis (33%) and pneumonia (20%). There were no cases of epiglottitis.
During the 6 years between October 1, 1992, and October 1, 1998, there were 106 cases of invasive type b disease in fully vaccinated children (i.e. vaccine failures). Table 1 compares age, mode of presentation, presence of associated medical conditions and mortality in these children with children who were also fully vaccinated but had invasive non-b Hi disease during this same period (102 cases). Non-b cases are further divided into those with nt and those with non-b capsulate strains. The major differences include age at disease presentation (younger in those with non-b disease, especially non-b capsulate disease) and clinical presentation (epiglottitis only found and meningitis more common and bacteremia and pneumonia less common among Hib cases). Associated medical conditions were also more common among cases of nt Hi disease as compared with cases of type b disease (32%vs. 20%, P = 0.08).
The numbers of cases of type b and non-type b Hi in fully vaccinated children varied during the 6-year period. During the first 24 months the numbers of cases were similar (22 non-b vs. 18 Hib), whereas for the last 24-month period there were more cases of non-b Hi than Hib (58 non- b Hi vs. 38 Hib).
Hi disease in all children (vaccinated and unvaccinated)
For the second period of the study (November 1, 1995, to December 31, 1998; 38 months) notifications of Hi disease were made for all children, regardless of vaccination status. Of a total of 188 non-b cases, 104 were vaccinated (included in analysis above) and 84 were not vaccinated. The majority were nt strains (165, 88%) followed by Hif (19, 10%), Hie (3) and Hic (1). The major modes of presentation were bacteremia (52%), meningitis (22%) and pneumonia (18%). The median age of presentation for non-b disease was 10.8 months (0 to 167.1). Associated medical conditions were recorded in 76 (40%) of all non-b cases and are detailed in Table 2.
There were 16 deaths (8.5%) all of which were in children with nt Hi (16 of 165; 9.7%) and associated medical conditions were present in 9 deaths (prematurity 7, Down’s syndrome, 1; Tay-Sachs disease, 1).
Nontypable H. influenzae disease. The median age of nt Hi disease was 9.2 months (0 to 167.1) The major mode of presentation was bacteremia (89, 54%), followed by pneumonia (32, 19%), meningitis (30, 18%) and cellulitis (7, 4%). Forty-one percent had an associated medical condition.
Twenty-nine percent (48) of nt cases occurred in the neonatal period (≤30 days of age). Of the 48 cases 85% (41) occurred at <7 days of age (early onset disease), 77% (37) in the first 2 days of life. Of those with early onset disease 73% (30) were born prematurely (<37 weeks), 95% (39) presented with septicemia (2 with meningitis) and 17% (7) died (all were premature). The annual incidence of early onset neonatal nt disease in the UK is shown in Table 3.
H. influenzae type f disease. Hif presented at a median age of 16.8 months (5.4 to 154.4), and the majority of cases had meningitis (11, 58%) followed by bacteremia (5, 26%), pneumonia (2) and cellulitis (1). Five had associated medical conditions.
H. influenzae type e disease. The three cases of Hie disease presented with bacteremia at 2.9 and 54 months of age and peritonitis at 47 months of age, respectively. All had associated medical conditions.
H. influenzae type c disease. The one case of Hic developed bacteremia at 12 months of age. This child also had diabetes insipidus.
There were two strains that phenotypically appeared to be nontypable but genotypically were type b (b− strains). One was obtained from a fully vaccinated child and the other from a partially vaccinated child.
We report the results of 6 years of surveillance of non-type b H. influenzae disease in children who have received the Hib conjugate vaccine and 3 years of surveillance of non-type b H. influenzae disease in all children (regardless of vaccination status) in the UK and ROI. This surveillance has been national, active and prospective, and a number of steps were taken to ensure that ascertainment was as complete as possible.
The Hib vaccine program has been well-accepted by children and parents in the UK and ROI. In the UK the impact on invasive Hib disease has been impressive, with a 98% fall in cases in children <5 years of age. 20 Accurate Hib vaccine coverage figures are known, and >90% of UK children have received a full course of Hib vaccination by their fifth birthday (3 doses ≤13 months of age or 1 dose 13 to 48 months of age). For example 94% of children who were born at around the midpoint of this surveillance period (born April through June, 1995) have been fully vaccinated. 21
We estimate an annual incidence of non-b disease in children <5 years of age (shown in Table 3) which is ∼30-fold lower than that of Hib disease in the prevaccine era 3 and in accord with that calculated in previous UK and US studies. 4, 11, 22 Non-b Hi is therefore a rare cause of disease. However, in 1998 the incidence of non-b Hi disease was twice that of Hib disease in children <5 years of age and three times the incidence of Hib disease in children <1 year of age (Table 3). Furthermore in fully vaccinated children Hi disease is now more likely to be caused by non-b Hi than by Hib. In the absence of full identification of the strain causing disease, this might conceivably result in a significant underestimation of vaccine effectiveness; in this study at least twice as many vaccine failures would have been reported if the strain causing disease had not been fully verified.
Comparison of type b and non type-b cases of Hi reveals a striking difference in clinical presentation; Hib presents predominantly with meningitis and epiglottitis and infrequently with pneumonia and bacteremia, whereas non-b Hi presents less commonly with meningitis and much more commonly with bacteremia and pneumonia. When non-b disease is further divided into encapsulated and nonencapsulated strains, it becomes apparent that capsulated strains (predominantly serotype f) occur in younger children with a predominance of meningitis and bacteremia whereas noncapsulated (nontypable) strains account for the majority of pneumonia among non-b cases. Epiglottitis appears to be overwhelmingly a syndrome associated with serotype b.
Consistent with other population-based studies in the UK, 11, 22 US, 4 the Netherlands 23 and Switzerland, 24 nontypable Hi accounts for most non-b Hi disease. One exception to this is a US report of Hi disease in 1994 and 1995 which appeared to show that non-b capsulated strains were a more frequent cause of invasive disease than nontypable strains in children <5 years of age. 15 However, only 56% of isolates in this age group were serotyped; therefore the reported distribution of serotypes may not be accurate.
Underlying or associated medical conditions appear to be frequent in children with non-b Hi disease. In contrast such conditions were uncommon among cases of Hib disease in the prevaccine era 25 and reinforce the difference in virulence between non-b and type b Hi.
It must be emphasized, however, that non-b cases are capable of causing significant disease. Of a total of 208 cases of non-b Hi in this study, 58 (25%) presented with meningitis and the mortality of non-b cases was ∼9%, twice the mortality of type b disease in the prevaccine era. 3
Urwin et al. 12 reported 19 cases of Hif in US children in a 6-year period. We describe a total of 28 children, 9 from the period in which only vaccinees were eligible for inclusion and 19 from the 38-month period in which all childhood cases were sought. In keeping with Urwin et al., 12 of those children <5 years of age, the majority were <12 months of age. Meningitis and pneumonia accounted for 40% of disease respectively in the US children, whereas in the UK children there was a slightly greater proportion of meningitis (58%) and more bacteremia (26%). We had no deaths as compared with 4 of 19 deaths in the US study and there were similar proportions with underlying conditions (26%). These data add to and are compatible with that of Nitta et al. 26 who described 3 cases of Hif and reviewed another 12 cases from the literature and also found a predominance of meningitis, underlying abnormalities (in 1 of 3) and no deaths.
It may be concluded from this and other reports that invasive disease in our population caused by serotypes a (one case), c (one case) and d (no cases) is extremely rare. Hia as a cause of meningitis and pneumonia has been reported more frequently in certain native populations. 27, 28
Noncapsulate (nontypable) Hi is recognized as an uncommon pathogen in the neonatal period. 11, 29–31 Quentin et al. reported an incidence of disease of 2.8/1000 live births, but others have reported rates of 4.6/100 000 live births, 11 2.8/100 000 31 and 1.6 to 1.9/100 000 in this report. The predominance of prematurity and the onset of disease soon after birth are features common to all reports. However, we, like Falla et al. and Takala et al., did not find a preponderance of biotype 4 as has been described by others (Refs. 29 and data not shown). To put this in perspective the rate of early onset neonatal nontypable disease determined in this study is one-fortieth that of early onset group B streptococcal disease in the UK. 32
Two strains were genotypically b−. If identification of isolates relied solely on conventional capsular serotyping, the presence of such strains could result in an overestimation of Hib vaccine efficacy. They may also represent a theoretical means by which Hib could evade the effect of Hib vaccine. There is no way of knowing, however, whether these particular isolates were expressing capsule at the time of invasion.
We conclude that non-type b Hi is a rare cause of disease in children. With the decline of Hib, however, it is a relatively more important pathogen and is now the most common strain of Hi isolated from children who have received Hib conjugate vaccine. Care must therefore be taken in verifying the isolate in cases of possible vaccine failure. This group of organisms is capable of causing severe diseases such as meningitis; their most common modes of presentation are with bacteremia and pneumonia. They also may signal an underlying host abnormality and we would recommend that underlying conditions should be sought.
We thank Richard Lynn and the British Paediatric Surveillance Unit of the Royal College of Paediatrics and Child Health as well as pediatricians, microbiologists and consultants in communicable disease control.
1. Koch R. Bericht uber die Thatigkeit der deutschen Cholera-kommission in Aegpytten and Ostindien. Wien Med Wochenschr 1883; 33: 1548–51.
2. Pittman M. Variation and type specificity in bacterial species of Haemophilus influenzae
. J Exp Med 1931; 53: 471–92.
3. Booy R, Hodgson SA, Slack MP, Anderson EC, Mayon-White RT, Moxon ER. Invasive Haemophilus influenzae
type b disease in the Oxford region(1985–91). Arch Dis Child 1993; 69: 225–8.
4. Wenger JD, Pierce R, Deaver K, et al. Invasive Haemophilus influenzae
disease: a population-based evaluation of the role of capsular polysaccharide serotype:Haemophilus influenzae
Study Group. J Infect Dis 1992; 165 (Suppl 1): S34–5.
5. Heath PT. Haemophilus influenzae
type b conjugate vaccines
: a review of efficacy data. Pediatr Infect Dis J 1998; 17: S117–22.
6. Klein JO. Role of nontypeable Haemophilus influenzae
in pediatric respiratory tract infections. Pediatr Infect Dis J 1997; 16: S5–8.
7. Williams M, Thompson AB. Fatal Haemophilus influenzae
type f infection. Lancet 1982; 1: 630.
8. Weinberg GA, Ghafoor A, Ishaq Z, et al. Clonal analysis of Haemophilus influenzae
isolated from children from Pakistan with lower respiratory tract infections. J Infect Dis 1989; 160: 634–43.
9. Faden H. Meningitis caused by nontypable Haemophilus influenzae
in a four-month-old infant. Pediatr Infect Dis J 1991; 10: 254–5.
10. Van Deynse E, Vermijlen P, Van Noyen G, Van Noyen R. Neonatal sepsis due to nonencapsulated Haemophilus influenzae
biotype IV. Acta Clin Belg 1997; 52: 204–6.
11. Falla TJ, Dobson SR, Crook DW, et al. Population-based study of non-typable Haemophilus influenzae
invasive disease in children and neonates. Lancet 1993; 341: 851–4.
12. Urwin G, Krohn JA, Deaver-Robinson K, Wenger JD, Farley MM. Invasive disease due to Haemophilus influenzae
serotype f: clinical and epidemiologic characteristics in the H. influenzae serotype b vaccine era. The Haemophilus influenzae
Study Group. Clin Infect Dis 1996; 22: 1069–76.
13. Rutherford GW, Wilfert CM. Invasive Haemophilus influenzae
type a infections: a report of two cases and a review of the literature. Pediatr Infect Dis 1984; 3: 575–7.
14. Turk DC, Cruickshank JC. Significance of Haemophilus influenzae
type-c. Lancet 1981; 2: 536–7.
15. Bisgard KM, Kao A, Leake J, Strebel PM, Perkins BA, Wharton M. Haemophilus influenzae
invasive disease in the United States, 1994–1995: near disappearance of a vaccine-preventable childhood disease. Emerg Infect Dis 1998; 4: 229–37.
16. Shishido H, Matsumoto K. Meningitis due to Haemophilus influenzae
type e biotype 4. J Clin Microbiol 1979; 10: 926–7.
17. Waggoner-Fountain LA, Hendley JO, Cody EJ, Perriello VA, Donowitz LG. The emergence of Haemophilus influenzae
types e and f as significant pathogens. Clin Infect Dis 1995; 21: 1322–4.
18. British Paediatric Surveillance Unit. 12th Annual Report 1997–8. London: Royal College of Paediatrics and Child Health, 1998:20–1.
19. Falla TJ, Crook DW, Brophy LN, Maskell D, Kroll JS, Moxon ER. PCR for capsular typing of Haemophilus influenzae
. J Clin Microbiol 1994; 32: 2382–6.
20. Heath PT, Booy R, Azzopardi HJ, et al. Antibody concentration and clinical protection after Hib conjugate vaccination in the United Kingdom. JAMA 2000; 284: 2334–40.
21. CDSC. Cover program: April to June, 2000. Commun Dis Rep CDR Wkly 2000; 10: 343–4.
22. Slack MP, Azzopardi HJ, Hargreaves RM, Ramsay ME. Enhanced surveillance of invasive Haemophilus influenzae
disease in England, 1990 to 1996: impact of conjugate vaccines
. Pediatr Infect Dis J 1998; 17: S204–7.
23. van Alphen L, Spanjaard L, Dankert J. Non-typable Haemophilus influenzae
invasive disease. Lancet 1993; 341: 1536.
24. Muhlemann K, Balz M, Aebi S, Schopfer K. Molecular characteristics of Haemophilus influenzae
causing invasive disease during the period of vaccination in Switzerland: analysis of strains isolated between 1986 and 1993. J Clin Microbiol 1996; 34: 560–3.
25. Tudor-Williams G, Frankland J, Isaacs D, et al. Haemophilus influenzae
type b conjugate vaccine trial in Oxford: implications for the United Kingdom. Arch Dis Child 1989; 64: 520–4.
26. Nitta DM, Jackson MA, Burry VF, Olson LC. Invasive Haemophilus influenzae
type f disease. Pediatr Infect Dis J 1995; 14: 157–60.
27. Gratten M, Morey F, Hanna J, et al. Type, frequency and distribution of Haemophilus influenzae
in central Australian aboriginal children with invasive disease. Med J Aust 1994; 160: 728–9.
28. Wall RA, Mabey DC, Corrah PT. Haemophilus influenzae
non type b. Lancet 1985; 2: 845.
29. Wallace RJ, Baker CJ, Quinones FJ, Hollis DG, Weaver RE, Wiss K. Nontypable Haemophilus influenzae
(biotype 4) as a neonatal, maternal, and genital pathogen. Rev Infect Dis 1983; 5: 123–36.
30. Quentin R, Musser JM, Mellouet M, Sizaret PY, Selander RK, Goudeau A. Typing of urogenital, maternal, and neonatal isolates of Haemophilus influenzae
and Haemophilus parainfluenzae
in correlation with clinical source of isolation and evidence for a genital specificity of H. influenzae
biotype IV. J Clin Microbiol 1989; 27: 2286–94.
31. Takala AK, Pekkanen E, Eskola J. Neonatal Haemophilus influenzae
infections. Arch Dis Child 1991; 66: 437–40.
32. Nicoll A, Heath P. Group B streptococcal infection (commentary). Arch Dis Child Fetal Neonatal Ed 2000; 82: F207.