The role of cefamandole in the treatment of Haemophilus influenzae infections in infants and children

PEDIATRIC PHARMACOLOGY AND THERAPEUTICS PaulS. Lietman,Editor

The role of cefamandole in the treatment of H a e m o p h i l u s influenzae infections in infants and children Forty patients with suspected non-CNS Hib infections were treated with cefamandole at a dosage of lO0 to 150 mg/kg/day. Hib was isolated from 19patients; three of the isolates were Blac +. All patients responded well without complications except for two children, both infected with Blac + organisms, who 9subsequently developed meningitis, one three weeks following treatment for buccal cellulitis and bacteremia, the other while being treated with cefamandole for empyema. A ll strains of Hib were uniformly susceptible to cefamandole regardless of beta-lactamase production when tested with an inoculum of 10" cfu/ml. With 10~ cfu/ml an inoculum effect was seen which was more pronounced with Blac + strains. Moreover, "heavy" inoculum of Blac + strains inactivated cefamandole in four to eight hours resulting in bacterial overgrowth. Similar results were obtained for ampicillin; chloramphenicol killed all strains regardless of inoeulum size or beta-lactamase production. We conclude that cefamandole may be hydrolyzed by Blac + organisms when present in large numbers, resulting in treatment failure. Extreme care should be taken in the choice of eefamandole for young infants with Hib infections, since this antibiotic neither cures nor prevents meningitis.

Parvin H. Azimi, M.D.,* and Patricia A. Chase, M.D.,

BECAUSE OF the emergence of ampicillin-resistant Haemophilus influenzae type b, chloramphenicol is presently the recommended antibiotic for initial management of systemic Hib infections? However, since the clinical spectrum of childhood infections caused by this organism is diverse, pediatricians are often faced with the dilemma of using chloramphenicol, a drug with rare though potentially serious adverse effects, for the initial treatment of many infections for which H i b may be a cause. Cefamandole, a newly introduced cephalosporin derivative with excellent in vitro activity against Hib, provides a possible alternative. We report our clinical and laboratory experi-

From the Division of Infectious Diseases, Children's Hospital Medical Center. Supported in part by a grant from Eli Lilly & Company, Indianapolis, lndiana. Presented at the 20th lnterscience Conference on Antimicrobial Agents and Chemotherapy in New Orleans, September, 1980. *Reprint address: Division of Infectious Diseases, Children's Hospital Medical Center, 51st & Grove Sts., Oakland, CA 94609.

0022-3476/81/060995+06500.60/0 9 1981 The C. V. Mosby Co.

Oakland, Calif.

ence with this agent in the management of noncentral nervous systemic Hib infections in infants and children.

MATERIALS AND METHODS Patients. Candidates for admission to this study included those patients presenting with symptoms and signs compatible with Hib infections. If there was any clinical suspicion of central nervous system involvement, a lumbar puncture was performed; if the results were indicative of CNS infection, the patient was excluded from the study. Likewise, patients were excluded when pathogens other than Hib were isolated. Blood and the infection site were cultured for bacteria before initiation of cefamandole therapy. Informed conSee related articles, pp. 910 and 1003.

Abbreviations used Hib: Haemophilus influenzae type b CNS: central nervous system cfu: colony-forming units

The Journal ofPED IATR I C S Vol. 98, No. 6, pp. 995-1000

995

996

Azimi and Chase

The Journal of Pediatrics June 1981

Table I, Summary o f presentation and course of culture proven Hib infections treated with cefamandole

Type of infection Periorbital cellulitis

Buccal cellulitis

Otitis media Epiglottitis Pneumonia Osteomyelitis (tibia) Septic arthritis (knee) Empyema

Age 4 mo

Source of Hib

Beta lactamase

Cefamandole MIC (ixg/ml)

Comments

Blood

-

0.35

6 mo 7 mo 9 mo 10 mo 11 mo 36 mo 4 mo 8 mo 11 mo 12 mo

Blood Blood Eye Blood Blood and eye Blood Blood Needle aspirate Blood Blood

---+ +

0.12 0.25 1.0 0.12 0.03 0.50 -0.01 0.25 0.05

16 mo 22 mo 10 mo 4 yr 17 mo 11 mo

Blood Blood MiddIe ear fluid Blood Blood Blood

--

1.0 0.50 0.12 0.50 0.10 0.25

Prompt clinical response, cure of infection

9 mo

Blood and knee aspirate Blood and pleural fluid

-

0.40

Prolonged fever

+

0.10

Development of meningitis

289 mo

sent was obtained for the use of investigational cefaman dole. Forty patients fulfilled the above criteria and were treated with cefamandole from 1975 through 1979. Their ages ranged from 2 months to 7 years; 70% were less than 2 years of age. There were 42 sites of infection: periorbital cellulitis (22), facial cellulitis (9), pneumonia (5), otitis media (2), pyogenic arthritis (1), osteomyelitis (1), epiglottitis (1), and empyema (1). Patients received cefamandole 100 to 150 m g / k g / d a y divided every 6 hours iv or im. The duration of therapy was generally five to ten days but varied from three to 21 days. Patients were examined daily and clinical response was judged by the return of the temperature to normal, by the resolution of signs of inflammation, and by symptomatic improvement. Toxicity was evaluated in most of the patients by means of the following tests performed prior to and at the end of therapy: a complete blood count, urinalysis, blood urea nitrogen, serum creatinine, serum bilirubin, SGOT, SGPT, alkaline phosphatase, and direct Coombs test. Susceptibility of clinical isolates. All Hib isolated from patients were tested for the production of beta-lactamase. ~ Cefamandole susceptibility was determined both by stan-

Prompt clinical response, cure of infection

Development of meningitis

dardized disc diffusion technique ~ and by a tube dilution method, which was performed as follows: serial twofold dilutions of standard cefamandole (Eli Lilly & Co., Indianapolis, Ind.) were made in Mueller-Hinton broth enriched with 5% Fildes reagent (Difco Laboratories, Detroit). Inoculum consisted of an overnight growth o f Hib in the same broth added to each tube to obtain a final concentration of 10' cfu/ml. Minimal inhibitory concentration was determined as the lowest concentration o f antibiotic that inhibited visible growth after overnight incubation. The effect of inoculum size on cefamandole susceptibility. Ten strains of beta-lactamase positive ( B l a c + ) , and ten strains of beta-lactamase negative (Blac--) Hib isolated from blood or CSF of infected children were tested against cefamandole at inoculum sizes of 104, 10~, and 107 c f u / m l by a tube dilution method as outlined above. Minimal bactericidal concentrations were also determined at these inocutum sizes. M B C was defined as the lowest concentration of the antibiotic that killed 99.9% of the original inoculum; this test was performed by quantitatively Subculturing each clear tube onto chocolate agar. For ten o f these strains, five of which were B l a c + , ampicillin (Bristol Laboratories, Syracuse, N. Y.) and

Volume 98 Number 6

Role of cefamandole in treatment of H. influenzae infections

997

Table II. In vitro activities of cefamandole, ampicillin, and chloramphenicol against various inoculum sizes of Hib 104 inoculum Antibiotic Cefamandole Ampicillin Chloramphenicol

No. of strains Blac Blac Blac Blac Blac Blac

+ 10 - 10 + 5 -- 5 + 5 -- 5

10~ inoculum

10~ inoculum

MIC*

MBC*

MIC

MBC

MIC

0.05 - 0.1 0.1 - 0.4 1.6 - 3.1 0.1 - 0.2 0.2 - 0.4 0.2 - 0.8

0.1 - 0.4 0.2 - 0.4 3.1 - 12.5 0.1 - 0.2 0.8 - 1.6 0.8 - 1.6

0.2 - 0.4 0.1 - 0.4 400 0.4 0.4 - 0.8 0.4 - 0,8

0.2 - 0.8 0.1 - 0.4 400 0.8 - 1.6 0.8 - 1.6 1.6

400 400 400 400 1.0 - 2.0 0.5 ~ 2.0

[

MB C 400 400 400 400 1.0 - 2.0 0.5 - 2.0

*/Lg/ml. chloramphenicol (Parke, Davis & Co., Detroit) MICs and MBCs were also determined at similar inoculum sizes. Time-kill studies. Time-kill curves were determined with cefamandole for 5 B l a c + and 5 B l a c - Hib using "light" and "heavy" inocula. The final concentration of bacteria in the "heavy" inoculum was approximately 106 to 107 c f u / m l and in the "light" inoculum l03 to 104 cfu/ml. For four of these strains, two of which were B l a c + , curves were also determined for ampicillin and chloramphenicol using similar inocula. The organisms were grown overnight in Mueller-Hinton broth with 5% Fildes enrichment. A concentration of 10 /~g/ml was chosen for all three antibiotics because it was a readily attainable serum concentration with parenteral therapy and many times greater than the M I C of susceptible organisms. Each test sample (total volume 4 ml) was incubated at 37~ in an enriched CO~ atmosphere and 0.5 ml aliquots were removed at 0, 4, 8, and 24 hours. These were serially diluted and poured into Fildes enriched Mueller-Hinton agar; colonies were counted after 24 hours of incubation. Simultaneous with each series of killing studies, bacterial cell growth was followed in antibiotic-free medium. W h e n cefamandole or ampicillin was used as the test antibiotic, its concentration was determined in each respective aliquot bY an agar well diffusion method using Bacillus subtilis spore suspension (Difco Laboratories) as the test organism." RE-SULTS

Clinical study. All but three patients had an excellent response to cefamondole therapy as determined by the resolution o f fever in 24 to 72 hours, symptomatic Improvement, and the resolution of signs of inflammation by the end of therapy (Table I). A 9-month-old infant with pyogenic arthritis of the knee had intermittent fever for the first six days but thereafter became afebrile and recovered without sequelae, receiving a total of three weeks of parenteral therapy.

Two patients, both infected with Blac + Hib, developed meningitis. One was a 12-month-old infant, with facial cellulitis, bacteremia, and normal CSF, who had a dramatic response to iv c e f a m a n d o l e therapy (100 mg/kg/day), with prompt defervescence and complete resolution of the buccal cellulitis within 48 hours of treatment; he was discharged after four days of cefamandole therapy but received, with excellent compliance at home, seven days of trimethoprim-sulfamethoxazole (10 and 50 m g / k g / d a y , respectively), to which his organism was sensitive. Three weeks after discharge, he suddenly became febrile and lethargic. A lumbar puncture revealed cloudy CSF; both blood and CSF grew B l a c + Hib. He was treated with chloramphenicol for 14 days and recovered without apparent neurologic sequelae. The second patient, a 21A-month-old female infant, was treated with cefamandole (150 m g / k g / d a y ) because of a small right pleural effusion. The analysis of her C S F was normal with a negative culture at the onset of treatment, but both blood and pleural effusion grew B l a c + Hib sensitive to cefamandole (MIC: 0.1 /~g/ml) and resistant to ampicillin (MIC: 8 #g/ml). Because of persistence of the effusion and of high fevers, a chest tube was inserted for drainage of the pleural fluid on the fifth hospitalday. Shortly thereafter a lumbar puncture was repeated because of a suggestive bulging fontanelle. The CSF contained 1,670 W B C / m m :~(85% neutrophils), protein 67 mg/dl, and glucose 55 mg/dl. Cefamandole treatment was discontined and chloramphenicol at a dose of 100 m g / k g / d a y was started. C S F culture grew B l a c + Hib with sensitivities similar to those of the previous blood and empyema fluid isolates. She responded well, became afebrile in 48 hours, and had a normal C S F examination after ten days of chloramphenicol therapy. Safety. Cefamandole was tolerated well by all patients, without evidence of local or systemic reaction. Transient eosinophilia ranging from 4 to 17% was seen in 26% of the patients. Four patients had transient elevation of SGOT (44 to 254 IU). There were no changes in urinalyses,

998

Azimi and Chase

The Journal of Pediatrics June t 98 t

Controls

-~

6-

/

kl..

O

4

C9 O -J

3

.r''""

2

I

CFM IOmcg/ml 0

103-1041lO* IOe-107/lO control llO

,r

8

9.5 I0 I0

9 9 9.5

Time (hours)

24 7t 9.

Fig. 1. Growth curves of five strains of Blac- Hib in 10/~g/ml of cefamandole. Symbols: - o - = "heavy inoculum"; ~l- = "light inoculum"; --- = growth in absence ofcefamandole. Table shows the mean concentrations ofcefamandole at various times of incubation with "'light" and "heavy" inocula of five strains of Blac- Hib. Corresponding concentrations of cefamandole in bacteria free broth are listed (third line) for comparison. */~g/ml.

Coombs tests, serum BUN, creatinine, or alkaline phosphatase measurements. Susceptibility of clinical isolates. Hib was isolated from blood or the site of infection in 47.5% (19/40) of the patients. All isolates were sensitive to cefamandole by disc diffusion. M1Cs ranged from 0.01 to 1.0 b~g/ml; all but two of the 18 strains tested were inhibited by less than or equal to 0.5 /~g/ml. Three of the isolates produced betalactamase. Inoculum effect. Table II shows MICs and MBCs for cefamandole, ampicillin, and chloramphenicol against Blac+ and Blac- Hib strains using different inocnlum sizes. A marked inoculum effect was demonstrated for cefamandole and ampicillin, with both Blac + and Blacorganisms. Chloramphenicol MICs were slightly higher than those for cefamandole at l04 C F U / m l inoculum, but there was no significant change at higher inocula. Time-kill studies. Cefamandole was bactericidal for both "light" and "heavy" inocnla of Blac- Hib within 4 to 24 hours decreasing the CFU below detectable levels in all but one strain (Fig. 1); the concentration of cefamandole' fell slowly to a mean of 7 #g/ml at 24 hours as compared to the bacteria-free control of 9.5 ~g/ml. For Blac+ strains (Fig. 2), a pronounced inoeulum effect was demonstrated. With the "light" inoculum, all strains were killed within eight hours (i.e., CFU below detectable levels). However, with the "heavy" inoculum

there was an initial drop in the number of viable organisms at four and eight hours, but at 24 hours there were 10~ to 10~ cfu/ml, similar to the number of organisms in the antibiotic-free control tubes. The concentration of cefamandole after incubation with the "light" inoculum was similar to that of Blac + strains. However, with the "heavy" inoculum, cefamandole was completely inactivated within four to eight hours. Ampicitlin in a concentration of 10 /~g/ml killed the "light" inoculum of Blac-- Hib within four hours and the "heavy" inoculum within 24 hours. When incubated with the "light" inoculum of Blac+ strains, this concentration killed one strain within eight hours and the other within 24 hours. When a "heavy" inoculum was tested, there was an initial !0 ~ to 103 drop in the number of surviving organisms, but at 24 hours, colony counts exceeded 107 cfu/ml, comparable to antibiotic-free controls. There was a gradual decrease in the concentration of ampicillin through 24 hours for all samples except with the heavy inoculum of the Blac+ Hib. With these two strains, ampicillin was hydrolyzed so rapidly that even at time 0, there was no detectable antibiotic activity. Chloramphenicol was bactericidal for all strains tested, within eight hours for the "light" inoculum and 24 hours for the "heavy" inoculum, regardless of beta-lactamase production, confirming similar studies previously reported?

Volume 98 Number 6

Role o f cefamandole in treatment o f H. influenzae infections

..----"

D

999

Controls

.........

.......................

5-

I

+

co o __1

3 2

I0 103-104[10 * 106-10718

control [1(3

9 CFM IOmcg/ml I

I

4

8

I

Time (hours)

24

9.5

9

I

0

o4

9.5

9.5 I

I0

Fig. 2. Growth curves of five strains of Blac+ Hib in 10/~g/ml of cefamandole. Symbols: - o - = "heavy inoculum"; -o- = "light inoculum"; . . . . growth in absence of cefamandole. Table shows the mean concentrations of cefamandole at various times of incubation with "light" and "heavy" inocula of five strains of Blac+ Hib. Corresponding concentrations of cefamandole in bacteria-free broth are listed (third line) for comparison. *t~g/ml.

DISCUSSION The search for alternative antibiotics to ampicillin for the treatment of systemic Hib infections has been necessitated by the rapidly increasing incidence of ampicillinresistant organisms. Recently a nationwide prevalence of 4.5% has been reported. + At the Children's Hospital Medical Center in Oakland the increase in ampicillinresistant Hib isolates from blood and CSF has been alarming, from 7% (3/44 isolates) in 1977 to 33% (13/39) in 1979. Although chloramphenicol is a very effective antibiotic against Hib and is unaffected by beta-lactamase production, resistance has been reported in this country as well as in EuropeY -8 Moreover, chloramphenicol may produce serious adverse reactions? Our clinical study shows that cefamandole is an effective agent for treatment of Blac- Hib infections. However, our experience with Blac + Hib infections is limited to three cases; the fact that two of the three developed meningitis evoked concern about the efficacy of cefaman~ole in Blac+ Hib infections and prompted our in vitro investigation. In the first patient recurrence of infection may have been caused by inadequate treatment with the initial short course of cefamandole, reinfection by persistence of nasopharyngeal carriage, or reinfection by a different strain.

The second case, however, represents an unquestionable treatment failure. Not only did this infant with empyema fail to respond to cefamandole therapy, but she developed meningitis while receiving the drug. The exact cause of treatment failure in this infant is unknown. The meninges might have been infected even though the initial CSF culture was sterile. Alternatively, the lumbar puncture itself might have caused the seeding of the meninges. A third possible explanation for the development of meningitis is that a large number of Blac+ organisms in the empyema fluid led to inactivation of cefamandole and spread of the infection elsewhere. Our in vitro data demonstrate this inactivation in the presence of large numbers of Blac+ organisms, and support the last explanation as the cause of treatment failure and meningitis. Cefamandole penetration into the CSF is erratic; reported CSF concentrations with high parenteral dosages have ranged from 0 to 17 ~g/ml? ~ In a clinical trial in which cefamandole was used for the treatment of Hib meningitis, patients failed to respond and were treated with other antibiotics?3 In addition, it appears that cefamandole cannot prevent the occurrence of meningeal seeding during the course of therapy for an infection elsewhere. One other case of an infant developing meningitis (with Blac-- Hib [personal communication]) while being treated with cefamandole has been reported? ~

10 0 0

A zimi and Chase

Our in vitro data confirm the previously reported uniform susceptibility of Hib to cefamandole regardless of beta-lactamase production when tested with an inoculum of 104 to 10~ cfu/ml. 1...... 7 Moreover, we have shown that the concentration o f cefamandole is not significantly affected by this size inoculum. With a large inoculum o f Blac+ organisms, on the other hand, cefamandole is inactivated rapidly, resulting in bacterial overgrowth. Like cefamandole, ampicillin was inactivated by a large inoculum of B l a c + organisms; however, it was hydrolyzed much faster than cefamandole. This slower inactivation of cefamandole compared with that of ampicillin was the only difference noted in the behavior of these two antibiotics. These results accord with those o f Yourassowsky et al TM who, by another method, studied the growth characteristics of Hib with ampicillin and cefamandole. The significance Of these in vitro observations, when applied to clinical situations, can only be speculative. Cefamandole could possibly be hydrolyzed and be rendered ineffective in vivo in the presence of large numbers of B l a c + organisms. Chloramphenicol remains the most dependable antibiotic for initial m a n a g e m e n t of severe systemic Hib infections since it is not affected by either inoculum size or beta-lactamase production. Cefamandole is an effective antibiotic for the treatment of n o n - C N S infections caused by beta-lactamase negative Hib, and thus could be an alternative to ampicillin. Experience with beta-lactamase positive organisms is limited. Caution should be exercised when large numbers o f these organisms are present (i.e., pus), since hydrolysis can render cefamandole ineffective. Extreme care should be taken in the choice of cefamandole for young infants with either beta-lactamase positive or negative Hib infections, since this antibiotic neither cures nor prevents meningitis, to which young infants are so prone. The authors thank Mary Gail O'Brien for technical assistance. REFERENCES

1. Committee on Infectious Diseases, American Academy of Pediatrics: Current status of ampicillin-resisant Hemophilus influenzae type b, Pediatrics 57:417, 1976. 2. Thornsberry C, and Kirven LA" Ampicillin resistance in Haemophilus influenzae as determined by a rapid test for betaqactamase production, Antimicrob Agents Chemother 6:653, 1974.

The Journal of Pediatrics June 1981

3. Bauer AW, Kirby W, Sherris, JC, and Turck M: Antibiotic susceptibility testing by a standardized single disk method, Am J Clin Pathol 45:493, 1966. 4. Bennett JV, Brodie JL, and Kirby WMM: Simplified, accurate method for antibiotic assay for clinical specimens, Appl Microbiol 14:170, 1966. 5. Cole FS, Daum RS, Teller L, Goldmann DA, and Smith AL: Effect of ampicillin and chloramphenicol alone and in combination on ampiciltin-susceptible and -resistant Haemophilus influenzae type b, Antimicrob Agents Chemother 15:415, 1979. 6. Ward JI, Tsai TF, Filice GA, and Fraser DW: Prevalence of ampicillin- and chloramphenicol-resistant strains of Haemophilus influenzae causing meningitis and bacteremia: national survey of hospital laboratories, J Infect Dis 138:421, 1978. 7. Manten A, Van Klingersen B, and Dessens-Kroon M: Chloramphenicol resistance in Haemophilus influenzae, Lancet 1:702, 1976. 8. Kenny JF, Isburg CD, and Michaels RH: Meningitis due to Haemophilus influenzae type b resistant to both ampicillin and chloramphenicol, Pediatrics 66:14, 1980. 9. Meissner HC, and Smith AL: The current status of chloramphenicol, Pediatrics 64:348, 1979. 10. Rodriguez WJ, Ross S, Khan WN, and Goldenberg R: Clinical and laboratory evaluation of cefamandole in infants and children, J Infect Dis 137:S150, 1978. 11. Korzeniovgski OM, Carvalho EM, Rocha H, and Sande MA: Evaluation of cefamandole therapy of patients with bacterial meningitis, J Infect Dis 137:S169, 1978. 12. Steinberg EA, Overturf GD, Baraff LJ, and Wilkins J: Penetration of cefamandole into spinal fluid, Antimicrob Agents Chemother 11:933, 1977. 13. Steinberg EA, Overturf GD, Wilkins J, Baraff LJ, Streng JM, and Leedom JM: Failure of cefamandole in treatment of meningitis due to Haemophilus influenzae type b, J Infect Dis 137:S180, 1978. 14. Meyers BR, Leng B, and Hirschman, SZ: Cefamandole: antimicrobial activity in vitro of a new cephatosporin, Antimicrob Agents Chemother 8:737, 1975. 15. Azimi PH: Clinical and laboratory investigation of cefamandole therapy of infections in infants and children, J Infect Dis 137:S155, 1978. 16. Jorgensen JH, and Alexander GA: Comparative activities of selected beta-lactam antibiotics against Haemophilus influenzae, Antimicrob Agents Chemother 13:342, 1978. 17. Syriopoulou VP, Scheifele DW, Sack CM, and Smith AL: Effect ofinoculum size on the susceptibility of Haemophilus influenzae b to beta-lactam antibiotics, Antimicrob Agents Chemother 16:510, 1979. 18. Yourassowsky E, Van Der Linden MP, and Lismont MJ: Growth curves, microscopic morphology, and subcultures of beta-lactamase-positive and-negative Haemophilus influenzae under the influence of ampicillin and cefamandole, Antimicrob Agents Chemother 15:325, 1979.