Previous Article | Next Article ![]()
Journal of Clinical Microbiology, February 2004, p. 524-529, Vol. 42, No. 2
0095-1137/04/$08.00+0 DOI: 10.1128/JCM.42.2.524-529.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Centro Nacional de Microbiología, Instituto de Salud Carlos III,1 División de Fármaco-Vigilancia, Agencia Española del Medicamento, Ministry of Health, Majadahonda,3 Instituto de Salud Pública, Consejería de Sanidad, Comunidad de Madrid, Madrid, Spain2
Received 6 October 2003/ Returned for modification 14 October 2003/ Accepted 21 October 2003
|
|
|---|
|
|
|---|
Vaccination campaigns generate a novel epidemiological environment. In theory, the decline in the rate of Hib infections could encourage the emergence of diseases caused by other H. influenzae serotypes (7, 35). An increase in invasive infections caused by H. influenzae type f has been described in United States (35). In our experience, infections due to other H. influenzae serotypes occur mainly in adult patients with underlying diseases (9).
In the Madrid area, a widespread program of vaccination against H. influenzae type b has been in place since 1998, although conjugate vaccines had been available for private use almost 2 years before. Conjugate Hib vaccine was given at 2, 4, 6, and 18 months of age. Population coverage was >95%. The incidence of invasive infections due to H. influenzae in the Madrid area was 20 cases/100,000 inhabitants <5 years of age in 1994, although detailed serotype distribution information was not available (15).
Some time ago, Spain recorded high rates of antimicrobial resistance in H. influenzae (11, 12). Spain is also a country with high antibiotic consumption in comparison with other European countries (13).
Little information is available about the epidemiology, microbiology, and molecular epidemiology of invasive H. influenzae after the widespread vaccination campaigns with Hib conjugate vaccines. Accordingly, we sought to evaluate here the incidence of invasive H. influenzae in the population, to study the demographic features of the patients infected, to learn about their clinical patterns and possible predisposing underlying conditions, to determine the molecular epidemiology of the isolates, and to study their antimicrobial susceptibility patterns in connection with antimicrobial consumption in the same population.
|
|
|---|
Population and clinical study. From January 1999 to December 2000, all patients with H. influenzae invasive infection diagnosed in clinical laboratories in the Madrid area were prospectively studied. Invasive infection was defined as "infection with an H. influenzae isolate from a normally sterile site." All available microbiological laboratories and hospitals of the Madrid area (Autonomous Community of Madrid [ACM]), which serve a population of 5,022,289 (1996 census data), were contacted at the beginning of the study and regularly afterward; all of them were asked to send all of their invasive H. influenzae isolates to the reference laboratory. According to official census data, the structure of the Madrid population was very similar to the rest of Spain and other countries of the European Union. A medical epidemiologist examined the clinical records of the patients. An individual clinical protocol that included patient's identity data, clinical pictures, diagnosis, and a description of underlying conditions and outcome was filled out for each patient. General incidence, expressed as cases per 100,000 people per year, and specific incidence by age group were calculated. Incidence in children <5 years of age was compared to data obtained in 1994, before conjugate vaccines were available in the ACM (15).
Molecular studies. All clinical strains were examined by pulsed-field gel electrophoresis after digestion of bacterial DNA with SmaI (MBI Fermentas, Vilnius, Lithuania) as described previously (9). DNA fragments were analyzed by the UPGMA (unweighted pair-group method with arithmetic averages) clustering method and Dice coefficient, which are available as part of the Molecular Analyst program (Bio-Rad, Madrid, Spain).
Screening of large conjugative plasmid carriage was carried out by PCR, as described by Leaves et al. (22). Screening of capsule-deficient type b strains (b- phenotype) was determined by PCR as described earlier (23).
Antimicrobial susceptibility testing. Testing for antimicrobial susceptibility to 14 antimicrobial agents was performed according to National Committee for Clinical Laboratory Standards guidelines (29) by using a semiautomated microdilution method (Wider; Fco. Soria Melguizo S.A., Madrid, Spain). H. influenzae ATCC 49247 and H. influenzae ATCC 49766 were used as quality control strains as recommended elsewhere (15).
ß-Lactamase production was determined by the chromogenic cephalosporin test with nitrocephin as substrate. Chloramphenicol acetyltransferase was measured as described previously (4).
Antibiotic consumption. Due to the published high rates of antibiotic resistance in H. influenzae in Spain (11, 12), we studied the evolution of antibiotic consumption in the ACM. The Spanish Ministry of Health and Consumer Affairs maintains a drug database with a packet-by-packet record of all retail pharmacy sales of all medicines acquired with National Health System prescriptions (21, 31). This database was used to gather information on sales in the ACM over the period from 1996 to 2000. The information was tabulated, and the number of units sold was converted into defined daily doses (DDDs) of active drug ingredients according to World Health Organization (WHO) methodology (37). The number of DDDs per 1,000 inhabitants per day (DDD/1,000 inhabitants/day) for each of the active drug ingredients was then calculated.
Statistical analyses. Data were managed and statistics calculated by using the SPSS program (SPSS, Inc., Chicago, Ill.). Categorical variables were compared by two-tailed Fisher exact test. Association was determined by calculation of the odds ratio and its 95% confidence interval. The null hypothesis was rejected for P values of <0.05.
|
|
|---|
(i) Population study. A total of 113 patients had H. influenzae invasive infections in the ACM. Males accounted for 75 (66.4%) cases and females for 38 (33.6%) (P < 0.001). The median incidence in the total population was 1.0 cases/100,000 persons per year, without differences between the 2 years studied. The median incidence of noncapsulated and capsulated H. influenzae strains was 0.7 and 0.3 cases/100,000 persons per year, respectively. No differences were observed between the 2 years studied.
In four (3.3%) cases, the patient's age was missing; of the remaining 109, 81 (74.3%) were adults (>14 years) and 28 (25.7%) were children (
14 years), 22 (19.5%) patients were <5 years old, and 14 (12.4%) were <1 year old.
The incidence of H. influenzae invasive infections was higher in patients
14 years old than in those >14 years (P < 0.001) (Fig. 1). The highest incidence was detected in children <1 year of age, with 16.3 cases/100,000 inhabitants (Fig. 1). In adults, the highest incidence (2.6/100,000 inhabitants) was in patients >70 years old (Fig. 1).
![]() View larger version (27K): [in a new window] |
FIG. 1. Distribution of incidence (cases per 100,000 inhabitants) of H. influenzae invasive infections according to patient age.
|
14 years old (P = 0.005). No significant differences were detected between the two age groups in the number of infections caused by capsulated strains. In children <5 years old, the prevalence of invasive infections due to typeable and nontypeable H. influenzae strains did not increase over the 2-year study period; in this age group, eight capsulated strains (four serotype b in 1999 and three serotype b and one serotype e in 2000) were isolated from CSF or blood; all serotype b cases were found in unvaccinated or partially vaccinated children. In comparison with data obtained before the implementation of conjugate Hib vaccines, the incidence of invasive H. influenzae decreased by 86% in children <5 years old (15).
(ii) Clinical study. H. influenzae caused bacteremia in 59 (52.2%) cases, pneumonia in 23 (20.4%) cases, and meningitis in 12 (10.6%) cases. In capsulated H. influenzae infections, meningitis and pneumonia were the second and third most frequent clinical diagnosis, after bacteremia. Of the pneumonia cases, 79% were caused by nontypeable H. influenzae strains and 21% were caused by capsulated strains (P < 0.001).
In all age groups, bacteremia was the most important clinical presentation. Pneumonia was more prevalent in adult patients (>14 years) than in children (
14 years): 25.9% versus 7.1% (P = 0.03). In contrast, meningitis was more frequent in children than in adults, at 14.3% versus 9.8%, although this difference was not statistically significant; however, meningitis due to capsulated H. influenzae was more prevalent in children (P = 0.03).
Eighty-four patients (74.3%) had previous underlying predisposing conditions. The most prevalent underlying diseases were those that impaired immunity: tumor pathologies, organ transplantation, and primary impaired immunity in 36 (31.8%) cases, respiratory diseases in 26 (23%) cases (17 of whom [15%] had chronic obstructive pulmonary disease), intravenous drug use in 15 (13.3%) cases, human immunodeficiency virus infection in 10 (8.8%) cases, and premature birth and/or obstetric complications in 10 (8.8%) cases.
Of all of the 113 patients studied, 17 (15%) died. All of the patients who died were adults, and 55% of them were >65 years old.
Molecular studies.
Cluster analyses of results of DNA fingerprinting of capsulated H. influenzae is shown in Fig. 2. SmaI did not digest the DNA of seven isolates, all of them nontypeable H. influenzae. Nontypeable H. influenzae isolates showed little genetic homology (data not shown); in contrast, capsulated strains were more homologous, since most serotype b strains had a genetic distance of
18% (Fig. 2).
![]() View larger version (26K): [in a new window] |
FIG. 2. Dendrogram tree illustrating the genetic relationship of 24 strains of capsulated H. influenzae, as determined by pulsed-field gel electrophoresis. The numbers at the right are strain numbers. The letters indicate biotypes (before hyphen) and serotypes (after hyphen).
|
|
View this table: [in a new window] |
TABLE 1. Frequency of antimicrobial resistance of invasive H. influenzae in relation to large conjugative plasmid carriage
|
Susceptibility testing. MICs for the two control strains were always within recommended limits (29). A total of 24.2% of the isolates were resistant to ampicillin by ß-lactamase production. Decreased susceptibility to cotrimoxazole, clarithromycin, tetracycline, chloramphenicol, and rifampin was detected in 48.4, 19.8, 7.7, 6.6 and 2.2% of strains, respectively. Resistance to amoxicillin-clavulanic acid, cefuroxime, cefotaxime, cefixime, cefepime, meropenem, ciprofloxacin, and levofloxacin was not observed.
The prevalence of susceptibility and the 50 and 90% MICs for H. influenzae isolates are presented in Table 2.
|
View this table: [in a new window] |
TABLE 2. Susceptibility of 91 H. influenzae isolates from sterile sites in relation to noncapsulated and capsulated strainsa
|
Pediatric isolates were more resistant to ampicillin and cotrimoxazole than were adult isolates (44% versus 17.1% and 48% versus 28.6%, respectively), although only the difference with ampicillin was statistically significant (P = 0.007).
Ampicillin resistance was associated with decreased susceptibility to tetracycline, cotrimoxazole, and chloramphenicol (P < 0.001). However, decreased susceptibility to clarithromycin was more frequent in ampicillin-susceptible isolates (24.6%) than in ampicillin-resistant isolates (4.5%) (P = 0.03).
Isolates from blood and CSF had higher rates of resistance to ampicillin (27.5%) than those from other sterile sites (0%) (P = 0.05). In contrast, resistance to clarithromycin was less prevalent in isolates from blood and CSF than in those from other samples (17.7% versus 36.4%).
Multidrug resistance (nonsusceptibility to three or more antibiotics) was recorded for 10 (11%) of the 91 H. influenzae isolates. The most prevalent resistance phenotype was ampicillin-tetracycline-chloramphenicol, which was detected in three isolates, representing 33.3% of the multiresistant strains and 3.3% of strains overall.
Antibiotic consumption. Antibiotic use in the ACM decreased from 18.40 DDD/1,000 inhabitants/day (in 1996) to 15.91 DDD/1,000 inhabitants/day (in 2000). The same trend was observed in almost all single antibiotic groups. ß-Lactams (J01C group), principally penicillins, were the most widely used antibiotics. The consumption of tetracyclines (J01A group) and trimethoprim-sulfamethoxazole (J01EE group) was very low and decreased over time. Very little use of amphenicols (J01B group) was observed (Table 3).
|
View this table: [in a new window] |
TABLE 3. Antimicrobial consumption in the ACM, 1996 to 2000
|
|
|
|---|
Worldwide reports of invasive H. influenzae isolates during the prevaccination era noted that >80% of the cases were caused by Hib (16, 20, 38). In a study carried out in England and Wales, Hib strains represented ca. 84% of all invasive isolates, whereas only 1% were of another capsular serotype (34). In a recent report from Brazil, Hib was by far the most common serotype (97.8%), followed in frequency by nontypeable strains (1.5%) and by serotype a (0.5%) (38). In comparison, epidemiology studies after vaccine implementations reveal a very different pattern (17, 34). In the present study, Hib invasive infections accounted for only 16.5% of invasive H. influenzae infections, results similar to the rates obtained in another Spanish study (17); both datasets were obtained after effective conjugate vaccination. The data from England showed that, since the introduction of routine immunization of infants with conjugate Hib vaccine, there has been a 16-fold reduction in the annual attack rate of invasive Hib disease recorded in children <5 years of age (34).
We noticed that 9 of 24 capsulate isolates were not serotype b; 6 of them were serotype f. In the United States the decrease in Hib infections has led to the report of increased incidence of H. influenzae type e and f (35, 36). The proportion of all invasive H. influenzae disease caused by serotype f rose from 1% in 1989 to 17% in 1994 in the United States (35). Previous Spanish results showed a 3% invasive H. influenzae type f infections (17). In comparison, in the present study we detected 6.6% of invasive H. influenzae infections due to this serotype. In our experience (9), infections caused by H. influenzae types e and f in Spain have not increased after vaccination campaigns; these pathogens produced mostly opportunistic infections in adults with underlying diseases.
The H. influenzae nontypeable isolation rate in England was 60% in the postvaccination era (34). In another study, 72.7% of invasive strains isolated after the start of vaccination were noncapsulated (17); these strains were also the most common H. influenzae invasive strains isolated in the present study (73.6%).
Nearly one in every four H. influenzae isolates was ampicillin resistant; all ampicillin-resistant isolates were ß-lactamase producers. These rates are very similar to those recorded by The Spanish Surveillance Group for Respiratory Pathogens in 1,730 strains from respiratory tract isolates in Spain in 1998 and 1999 (27). In a collaborative European study (The Alexander Project), the overall prevalence of ß-lactamase production was less than 12% in 1997 and 1998, although with marked geographical variation (32). In the present study, we have shown that capsulated strains were significantly more resistant to ampicillin, tetracycline, and chloramphenicol than were noncapsulated strains. This may explain why ampicillin resistance was more prevalent among children and isolates from CSF and blood. In a previous Spanish study, ß-lactamase production was detected in 50% of Hib invasive isolates (12), a figure very similar to the 45.8% found in capsulated strains in the present study, even though only 62.5% of them were serotype b.
In the present study, a substantial prevalence of resistance to ampicillin, tetracycline, chloramphenicol, and trimethoprim-sulfamethoxazole was found in invasive H. influenzae against a background of decreased antibiotic consumption in the community. From 1996 to 2000, tetracycline and chloramphenicol consumption decreased from 0.70 to 0.58 and from 0.71 to 0.43 DDD/inhabitants day, respectively; amphenicol consumption was undetectable (Table 3). Decreases of 75 and 88% in tetracycline and cotrimoxazol consumption, respectively, have been reported in Spain from 1985 to 2000 (21). Antibiotic resistance can rapidly become more prevalent in a population, whereas its decline when antibiotic consumption is reduced is much slower (18). Furthermore, in areas with very high resistance rates, reduction in antibiotic pressure may have an even slower effect, especially where there is multidrug resistance (3, 24). There are various explanations for this phenomenon, including the possibility of resistance to different classes of antibiotics and coselection when only one of them is used (2, 18) and the reservoir of molecular mechanisms in species of the commensal flora and the exchange between these and pathogen species (1, 33).
Carriage of high-molecular-weight conjugative plasmids in Hib constitutes the genetic basis of resistance to ampicillin, chloramphenicol, and tetracycline (11, 25). Our results show a strong association between single and multiple resistances to these antibiotics and carriage of large conjugative plasmids (Table 1). However, the trimethoprim and clarithromycin resistance determinants are usually chromosomally mediated and are not associated with plasmid carriage (11). In the present study, plasmids were not found in any of the 18 strains nonsusceptible to clarithromycin or the 21 strains with nonsusceptibility to cotrimoxazole alone. It is noteworthy that in our study, carriage of large conjugative plasmids, as detected by PCR, was common in resistant capsulated and noncapsulated H. influenzae strains.
In summary, we carried out a community study into the significance of invasive H. influenzae after widespread vaccination against H. influenzae serotype b. We collected data about the clinical diagnoses and the microbiological characteristics of the strains, including their molecular epidemiology, antibiotic resistance patterns, and antibiotic consumption in the community.
This study was supported by a research grant from the Fondo de Investigaciones Sanitarias, Ministerio de Sanidad y Consumo, Madrid, Spain (98/0020-01).
Members of the Group of Invasive Haemophilus Infections of the Autonomous Community of Madrid, Spain, include E. Cercenado, Hospital Gregorio Marañón, Madrid; J. Merino, Unilab, Madrid, Spain; S. Concheiro, Hospital de Getafe, Madrid, Spain; R. Velasco, Cínica Moncloa, Madrid, Spain; G. Herruz, Hospital Principe de Asturia, Alcalá de Henares, Spain; C. Rodríguez-Avial, Hospital Clínico San Carlos, Madrid, Spain; F. Sanz, Hospital 12 de Octubre, Madrid, Spain; A. Carvajal, Hospital La Paz, Madrid, Spain; García Picazo, Hospital El Escorial; I. Wilhelmi, Hospital Severeo Ochoa, Madrid, Spain; E. Loza, Hospital Ramón y Cajal, Madrid, Spain; R. Fernández Roblas, Fundación Jiménez Díaz, Madrid, Spain; and J. L. Gómez Garcés, Hospital de Móstoles, Madrid, Spain.
Contributing members of the Group of Invasive Haemophilus Infections of the Autonomous Community of Madrid, Spain, are listed in the Acknowledgments. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»