Robert S. Daum

Spread of Haemophilus influenzae type b: Recent epidemiologic and therapeutic considerations

DESPlTE isolated case reports of invasive Haemophilus influenzae type b infections in more than one member of a household, 1-~ disease caused by this organism generally has not, until recently, been regarded as contagious. Recent changes in child-care practices, especially increased use of day-care centers, have led to greater exposure of susceptible children to patients with infection. Outbreaks have resulted, with corresponding recognition of the potential for Haemophilus influenzae type b to spread to contacts. ~-~ The purposes of this report are to review the epidemiology ofHaemophilus influenzae type b infections in enclosed populations, and to examine currently available alternatives for management of contacts. INSTITUTIONAL SETTINGS Reports of spread of Haemophilus influenzae type b in day-care centers or in other institutional settings indicate that most of the associated cases are in children less than 2 years of age (Table I). 7-~4 This age distribution corresponds to the pattern observed for endemic infections, ~, with the exception of epiglottitis which occurs in older children. 1~ Disease in contacts has been observed most frequently when the index patient presented with meningitis. However, patients with other invasive type b infections also may be contagious, 1~. ~, ~ and until further data are available the risks probably should be considered the

Haemophilus influenzae type B meningitis: a contagious disease of children.

The families of 126 consecutive patients with Haemophilus influenzae type B meningitis were surveyed for secondary invasive H influenzae disease among household contacts. A total of 120 of the families were contacted. In six cases no contact was possible and the medical record was reviewed. Some 555 household contacts were found; 31% (171) were under 5 years of age. A secondary case was defined as a household contact with H influenzae type B isolated from blood or cerebrospinal fluid more than 24 hours, but less than 30 days, after admission to hospital of the index case. Four secondary cases were identified, all in children aged under 5 years. The secondary attack rate in children under 5 years or less in the month after exposure to an index case was thus 2.3%, 800 times the endemic attack rate for H influenzae meningitis. This is a conservative estimate since five additional contact cases were documented, but not included in the secondary attack rate. Young contacts of a child with H influenzae meningitis are thus at significant risk of life-threatening secondary disease.

Rifampin chemoprophylaxis for household contacts of patients with invasive infections due to Haemophilus influenzae type b

To determine the efficacy of rifampin chemoprophylaxis in eradication of oropharyngeal carriage of Haemophilus influenzae type b, we conducted a multicenter, double-blind, placebo-controlled trial among household contacts of patients hospitalized for invasive HIB infection. Seventy-nine index patients and 400 close contacts were studied; 26.5% of contacts were colonized. The efficacy of rifampin (10 mg/kg/dose, 600 mg/dose maximum, twice daily for two days) in eradicating carriage was 52% and varied with age (75.6% in persons greater than or equal to 5 and 27% in those less than 5 years). Eradication rates in those less than 5 years were not significantly better than for placebo. No resistant isolates were encountered in sensitivity testing. The low efficacy of this rifampin regimen in young children precludes its routine use as a chemoprophylactic agent for family contacts. The occurrence of three cases of invasive HIB infection in individuals outside the defined contact group raises concern regarding the efficacy of any chemoprophylactic regimen.

Unusual presentations of Haemophilus influenzae infections in immunocompromised patients

In two years we encountered five invasive Haemophilus influenzae infections in immunocompromised hosts. The clinical presentations, age distribution, and serotypes of organisms were atypical of H. influenzae infections in the pediatric population. All four infections in patients with cancer occurred during remission. Broad-spectrum antibiotic therapy for immunocompromised patients should include coverage for H. influenzae .

Gentamicin penetration into cerebrospinal fluid in experimental Haemophilus influenzae meningitis.

We studied the effect of meningitis and the method of parenteral gentamicin administration (intramuscular injection, a 30-min intravenous infusion, or intravenous bolus administration) on achievable concentrations of drug in cerebrospinal fluid (CSF). In normal animals, only intravenous bolus administration of 2 to 8 mg/kg produced a gentamicin concentration of greater than 0.1 microgram/ml in CSF in some animals. All CSF samples contained less than the limit of detection (0.1 microgram/ml) after the intramuscular administration of 6 mg/kg. In animals with meningitis, gentamicin penetration into cisternal CSF was increased significantly after a bolus administration of 6 mg/kg (mean, 0.197 +/- 0.063 microgram/ml in normal animals versus 1.68 +/- 0.38 micrograms/ml in animals with meningitis; P less than 0.01). In meningitic animals that received 6 mg/kg as an intravenous bolus, lumbar CSF had the highest maximum concentration (4.25 +/- 1.08 micrograms/ml), in comparison with ventricular CSF (3.10 +/- 0.66 micrograms/ml). The gentamicin concentration in cisternal CSF decreased more slowly than it did in serum (elimination half-life, 238.70 +/- 64.56 min in cisternal CSF versus 82.73 +/- 2.91 min in serum), yielding a relative increase in the percentage of penetration. We conclude that maximum penetration by gentamicin into CSF occurs after intravenous bolus administration and that the maximum concentration occurs in lumbar CSF.

Interrelationships between physical state, phenotypic stability and transferability of beta-lactamase genes in Haemophilus influenzae.

SUMMARY: We have evaluated 66 β-lactamase-producing (β-Lac+) clinical isolates of Haemophilus influenzae for the presence of plasmid DNA, ability to conjugate with an H. influenzae type b recipient, and stability of the β-Lac+ phenotype. Of the 66 strains, 23 (35%) contained a plasmid demonstrable by either of two methods of cleared cell lysate preparation. The molecular weight of 17 of the plasmids was approximately 30 x 106. Four strains carried a plasmid of mol.wt 3 x 106 and single strains carried plasmids of mol. wt 36 x 106 or 50 x 106. Of 23 plasmids that were unstable in vitro, 10 (43%) had molecular weights of 30 x 106 and coded only for β-lactamase. Stability of the β-Lac+ phenotype was observed in all 43 strains in which plasmid DNA was not demonstrated by physical methods. When β-Lac+,‘plasmid-free’ strains were conjugated with strain Eagan (H. influenzae type b), ampicillin-resistant transconjugants were isolated in 21 out of 43 matings (the lower limit of detection was 10-7 per donor cell per 30 min mating). With few exceptions, homologous (type b x type b) crosses with β-Lac+, plasmid-containing donors were more efficient (frequency 10-2—10-3) than heterologous (untypable x type b) matings (frequency 10-6-10-7). β-Lac+,‘plasmid-free’ strains were inefficient donors (frequency 10-6-10-7) irrespective of serotype, but all transconjugants from these crosses were efficient donors (frequency 10-1-10-4) and contained a plasmid of mol. wt 30 x 106 that was unstable in 9 cases out of 26 (35%). The increase in transferability and phenotypic instability of plasmid-bearing β-Lac+ strains reinforces the concept of a chromosomal locus for β-lactamase genes in‘plasmid-free’ β-Lac+ strains. Carriage of β-lactamase genes on 30 x 106 mol. wt R plasmids aids transferability at the expense of genetic stability.

Effect of Ampicillin and Chloramphenicol Alone and in Combination on Ampicillin-Susceptible and -Resistant Haemophilus influenzae Type B

To evaluate ampicillin (Amp) and chloramphenicol (Cm) alone and in combination against Haemophilus influenzae type b, we examined the viability of 5 log10 colony-forming units (CFU) of early-log-phase organisms per ml after 4 and 8 h of incubation with the drug(s). Nine Amp-susceptible (Amps) and five Amp-resistant (Ampr) systemic isolates were examined. Antibiotic concentrations included: the minimum inhibitory concentration (MIC) of Amp, 50% of the MIC of Amp, 25% of the MIC of Amp, the MIC of Cm, 50% of the MIC of Cm, 25% of the MIC of Cm, and nine combinations of these concentrations. Both Amp and Cm at their MIC significantly reduced bacterial titers of AmpsH. influenzae type b after 8 h of incubation (1.36 and 1.47 log10 CFU/ml, respectively; both p < 0.01); only Cm at its MIC significantly reduced the number of viable organisms after 4 h (0.91 log10 CFU/ml; P < 0.001). With Ampr organisms, significant reductions in mean bacterial titers occurred after 4 and 8 h of incubation in the presence of Amp at its MIC (1.66 and 2.06 log10 CFU/ml, respectively; both P < 0.02); smaller but significant reductions were noted after 4 and 8 with Cm at its MIC (0.56 and 0.87 log20 CFU/ml, respectively; both P < 0.025). Antagonism with Amps or Ampr strains was not seen. We conclude that combinations of Amp and Cm have indifferent effects on Amps and AmprH. influenzae type b.

Counterpoint: the red book opts for red urine.

Several important issues regarding chemoprophylaxis for H. influenzae type b infections should be resolved before general adoption of the recommendations of the Committee on Infectious Diseases. Unfortunately the recommendations place the practicing pediatrician in a medical-legal dilemma if he does not comply. The prematurity of the recommendations may preclude further placebo-controlled, prospective evaluation of the efficacy of rifampin or other agents in the prevention of secondary disease. In any event the administration of chemoprophylaxis must not replace close surveillance of all members of the contact group. Band et al. estimated that 150 to 200 of the approximately 20,000 cases of invasive H. influenzae type b disease that occur each year in the United States may be preventable by prophylaxis. Primary prevention with a safe effective vaccine would have a much greater impact.

Trimethoprim and rifampin in combination for chemoprophylaxis of household contacts of patients with invasive infections due to Haemophilus influenzae type b.

We compared the effectiveness of rifampin-trimethoprim in fixed combination (3.75:1) to rifampin alone in the eradication of Haemophilus influenzae type b carriage among contacts of patients with invasive infection caused by this organism. The study population was composed of 127 index patients and 620 contacts. Twenty-six percent of contacts were colonized. Rifampin-trimethoprim eradicated carriage in 77.6% of contacts (71.1% in contacts less than 5 years, 84.2% in contacts greater than or equal to 5 years) whereas rifampin eradicated carriage in 69.9% of contacts (56.4% in contacts less than 5 years, 81.8% in contacts greater than or equal to 5 years). A single isolate resistant to rifampin and rifampin-trimethoprim was encountered. The eradication rate achieved with this regimen of rifampin-trimethoprim was too low to recommend its routine use. However, a higher dose or longer course might merit clinical trial.

COMPARISON OP ANTIGEN DETECTION METHODS IN A PRIMATE MODEL OF HEMOPHILUS MEMINGITIS

An infant primate model of H. influenzae b (HIB) infection was employed to compare the sensitivity of two rapid methods to detect infection. Latex particle agglutination (LPA) and CIE assays developed in this laboratory detect 0.5 and 1.0 ng/ml of HID antigen, respectively. Serial quantitative cultures were performed on 19 monkeys with bacteremia and meningitis and corresponding samples tested for antigen. Both assays detected antigen in the serum of all 19 bacteremic animals, with detection a function of duration and density of infection. LPA became positive earlier:70% of animals were LPA+ after 8 hrs of bacteremia and 100% after 36 hrs, compared with 10% and 60% using CIE. After 80 hrs, 84% of sera were CIE+. LPA detected fewer bacteria in blood:after 30 hrs of constant-density bacteremia, 6/6 animals with 102 - 103 cfu/ml were LPA+ while 1/6 was CIE+. CIE became reliable at bacterial concentrations ≥ 104/ml of blood; at 36 hrs, 8/9 such samples were CIE+. CSF was obtained from 18 monkeys within 24 hrs of meningitis onset (mean bacterial conc. 1.4 × 104 cfu/ml CSF). LPA was positive in 15/17 samples, detecting 4/6 containing 102 cfu/ml and all 11 with > 103/ml. CIE was positive in 9/18 samples, detecting 3/12 containing 102 - 103 cfu/ml but 6/6 with > 104/ml. 24 hrs later, LPA was positive in all CSF samples and CIE in 16/18 (84%). We conclude that LPA is superior to CIE, detecting antigen earlier and with fewer bacteria (102 -103 bacteria/ml of serum or CSF).