Whenever one prescribes antibiotics for a newborn infant with a potentially serious infection, it is important to give thought to three factors: the baby or host, the probable organisms, and the drug under consideration for use. In this way the most rational therapy can be prescribed for the infant, therapy that should lead to prompt cure of the patient without producing serious side effects. The following discussion will attempt to highlight certain features of the host, the infectious organism, and antibiotic drugs that are particularly pertinent to the newborn infant with suspected sepsis.
The signs and symptoms of neonatal sepsis (and often meningitis) are nonspecific but well known. They include poor temperature control (often hypothermia), poor feeding, respiratory distress, jaundice, and irritability. An infant is more likely to have sepsis when there have been complications of pregnancy, such as maternal urinary tract infection. Once sepsis is suspected, bacterial cultures of blood, urine, throat, umbilicus, and cerebrospinal fluid (CSF) should be obtained and therapy promptly begun. The most common causes of neonatal sepsis are Escherichia coli and group B betahemolytic streptococci. Less commonly encountered organisms are Staphylococcus aureus, Pseudomonas aeruginosa, Listeria monocytogenes, and Klebsiella species.
Pediatricians are fond of pointing out that children are different from adults, and this difference is especially profound when the child is a newborn infant. We have lately become more and more aware of the differences between adults and newborns, but there was a time when the medical profession was less cognizant of these differences. For example, in 1959 many physicians routinely administered prophylactic antibiotics to premature infants whose mothers' membranes had ruptured many hours before delivery, since these infants were at high risk of developing severe infections. One of the antibiotics used was chloramphenicol. To the surprise of all concerned, it was found that the mortality for babies weighing 2,000-2,500 gm. was actually higher in those who were receiving chloramphenicol than in those who were receiving no antibiotic. The infants had received chloramphenicol in doses of 100-165 mg. /kg. of body weight per day. Approximately one-half of the infants so treated developed a syndrome characterized by vomiting, anorexia, respiratory distress, cyanosis, abdominal distention, and loose green stools. It was subsequently realized that the host's immature renal and hepatic function led to decreased metabolism of chloramphenicol, causing the drug to accumulate to toxic levels and leading to increased morbidity and mortality in the infants.1
From the example given, it is easy to see the importance of knowing the clinical pharmacology of various antibiotics in the host - in this particular instance, the newborn infant. The clinical pharmacology of penicillin in newborn infants was recently studied by McCracken and coworkers in Dallas, Texas.2 These investigators studied blood levels of penicillin at various times after administration of a standard dose to newborns. It was found, as had been suspected, that serum levels of penicillin varied immensely with the age of the infant and the infant's creatinine . clearance. It was found that 50,000 units of penicillin G per kilogram per day (in two or three doses) in newborn infants provided peak serum levels 100 to 1,000 times the minimal inhibitory concentrations for group B beta-hemolytic streptococci. Ampicillin in the neonate has similarly been studied by the Dallas group.3 Again, serum levels of ampicillin varied immensely with the age of the infant. These investigators were able to make recommendations for dosage of ampicillin for infants of different ages on the basis of their studies. These doses could be predicted to result in high enough levels of drug to kill an infecting organism (streptococci, most E. coli) without harming the child.
The host in neonatal sepsis is also different from an adult because the neonate seems to be almost "immunocompromised"; this may also explain the infant's lack of obvious symptoms when he is infected.
The neonatal host is thus at high risk from the following bacterial and viral pathogens, which ordinarily cause mild or no disease in older hosts: group B streptococci, E. coli, herpes simplex virus, cytomegalovirus, and varicella-zoster virus. Finally, the neonate is often the inadvertent victim of such modern-day advances in neonatal care as respirators and umbilical catheters, which, while necessary for some of his problems, predispose him to infections. All these factors must be taken into account when antibiotic therapy is planned for the newborn.
One frequently encountered problem in management of neonatal sepsis has been the emergence of organisms resistant to certain antibiotics. Drug resistance is an important and practical problem with the use of drugs in newborn infants. This is particularly so with regard to E. colt infection. Drug resistance also has long been a problem with staphylococci. This problem persists, and it was recently noted that even most non-hospital-acquired staphylococci in Washington, D.C., were resistant to penicillin.4
Organisms can acquire resistance to drugs genetically. Part of the DNA of a bacterium present in an episome (R factor or plasmid), which is separate from the other chromosomal material, codes for production of enzymes that lead to resistance of the organism to antibiotics.5 The episome can be passed from one type of bacterium to another - e.g., from a Salmonella to an E. coli - by conjugation. Resistant strains are obviously "selected" by the use of antibiotics: as the susceptible organisms are killed, the resistant ones live and multiply. It is important for the clinician to be aware of this phenomenon, since overuse of antibiotics will lead to the emergence of resistant strains. It was predicted, for example, that with the increased use of kanamycin, resistance to it by such organisms as E. coli would emerge, and it did.6 E. coli resistant to gentamicin has also been reported recently, as was anticipated from services where gentamicin was used freely.7
Resistance of E. coli to kanamycin seems to vary from one region to another. For example, in Dallas, Texas, in 1970, 30 per cent of the strains of E. coli isolated from blood cultures of neonates were resistant to kanamycin. The experience of the group in Houston, Texas, was different, in that 95 per cent of E. coli isolates tested were sensitive to kanamycin.8 This group has recommended that gentamicin generally be reserved for infants with suspected hospital-derived infection and gramnegative bacillary meningitis. Of course, a knowledge of the pattern of sensitivity of E. coli to kanamycin and gentamicin (and other common infecting organisms to antibiotics) in the particular nursery is essential in planning the most rational therapy for the infected neonate.
ANTIBIOTIC THERAPY FOR NEONATAL SEPSIS AND MENINGITIS
Fortunately, drug resistance has not been a problem with treatment of streptococcal infections. The main problem with this organism has been that in babies infected with it during the first few days of life, initiation of treatment is almost invariably too late.9,10 The organism is often carried by the mother and father of the baby as part of the "normal flora" of their genital tracts. The baby becomes infected just before or at birth and, by the time symptoms are apparent, the disease is often too far gone to be effectively treated. It is thought, therefore, that one possible approach to this infection in babies might be prevention. Studies of prophylactic antibiotic treatment of pregnant women and their sexual partners who harbor group B streptococci and the effects in their offspring are currently in progress in several medical centers. Whether this approach will be successful is not yet known.
One further point concerning E. coli that cause neonatal meningitis deserves mention. It has recently been observed that most of the E. coli causing this infection possess a certain envelope or capsular antigen, termed the K-1 antigen.11 This antigen has previously been related to virulence of this organism in adults. The identification of this antigen and its relationship to virulence opens up several possible approaches to treatment of E. coli infections in neonates - for example, passive immunization of the mother and/or infant with antibody to the K-1 antigen and active immunization of women with the K-1 antigen itself.
THE DRUGS (TABLE 1)
A detailed pharmacologic analysis of drugs commonly utilized in the neonate is beyond the scope of this article. However, one basic concept regarding successful antibiotic therapy should be mentioned. This is that in order to be effective, a certain concentration of antibiotic (depending on the sensitivity of the organism) must reach the part of the body where the infection exists. For example, it is well known that nontoxic doses of the aminoglycoside drugs (streptomycin, kanamycin, and gentamicin) will provide serum concentrations adequate to rid the blood of E. coli. Penetration of these drugs into the CSF, however, is poor, so that while serum levels of the drug may be therapeutic, CSF levels may not.
It has been found that in newborns with meningitis due to gram-positive organisms who are treated by penicillin or ampicillin, the CSF is usually promptly sterilized. In contrast, viable organisms may persist for up to 10-12 days after institution of therapy for neonatal meningitis due to gramnegative organisms most often treated with aminoglycosides. McCracken has attributed this to the poor penetration of the aminoglycoside antibiotics into the CSF.12
Two possible regimens to combat this problem with neonatal meningitis due to gram-negative organisms have been devised. It has been suggested that ampicillin be used along with an aminoglycoside to treat this condition. Many E. coli are sensitive to ampicillin, and this drug penetrates more readily into the CSF. In addition, an in vitro synergy between ampicillin and aminoglycosides has been noted; it is not known, however, if this synergy occurs in vivo.
It has also been suggested that gentamicin be given intrathecally to infants with meningitis due to gram-negative organisms.13 This maneuver will serve to increase levels of antibiotic in the CSF. Whether intrathecally administered gentamicin will result in an improved survival rate in these newborns with meningitis is currently being studied at a number of medical centers in the United States. It appears, on first analysis, that the advantages gained by intrathecal gentamicin therapy (more rapid sterilization of CSF and decreased mortality) may be offset by the disadvantage of increased serious neurologic sequelae in those who survive meningitis.14 Thus there may be no advantage to this form of treatment.
1. Burns, L E.. Hodgman, J. E., and Cass. A. B. Fatal circulatory collapse in premature infants receiving chloramphenicol. N. Engl. J. Med. 261 (1959). 1318-1321.
2. McCracken, G. H., Jr., et al. Clinical pharmacology of penicillin in newborn infants. J. Pediatr. 82 (1973), 692-698.
3. Kaplan, J. M., et al. Pharmacologic studies in neonates given large dosages of ampicillin. J. Pediatr. 84 (1974). 571-577.
4. Ross, S.. et al. Staphylococcal susceptibility to penicillin G: The changing pattern among community strains. J.A.M.A. 229 (1974). 1075-1077.
5. Davies, J. Bacterial resistance to aminoglycoside antibiotics. J. Infect. Dis. 124 (1971). S7S10.
6. McCracken, G. H., Jr. Changing pattern of the antimicrobial susceptibilities of Escherichia coli in neonatal infections. J. Pediatr. 78 (1971), 942-947.
7. Franco. J. A.. Eitzman. D. V.. and Baer. H. Antibiotic usage and microbial resistance in an intensive care nursery. Am. J. Dis. Child. 126 (1973). 318-321.
8. Baker, C. J.. Barrett. F. F.. and Clark. D. J. Incidence of kanamycin resistance among Escherichia coli isolates from neonates. J. Pediatr. 84 (1974).
9. Franciosi, R. A.. Knostman, J. D., and Zimmerman, R. A. Group B streptococcal neonatal and infant infections. J. Pediatr. 82 (1973), 707-718.
10. Baker, C. J., et al. Suppurative meningitis due to streptococci of Lancefield group B: A study of 33 infants. J. Pediatr. 82 (1973). 724-729.
11. Robbins, J. B. , et al. Escherichia coli K- 1 capsular polysaccharide associated with neonatal meningitis. N. Engl. J. Med. 290 (1974). 1216-1220.
12. McCracken, G. H., Jr. The rate of bactériologie response to antimicrobial therapy in neonatal meningitis. Am. J. Dis. Child. 123 (1972). 547-553.
13. McCracken, G. H., Jr. Pharmacological basis for antimicrobial therapy in newborn infants. Am. J. Dis. Child. 128 (1974). 407-419.
14. McCracken. G. H.. Jr. Coordinator, the Cooperative Neonatal Meningitis Study Group: Evaluation of intrathecal therapy for meningitis due to gram-negative enteric bacterium. Pediatr. Res. 9 (1975), 342.
15. McCracken. G. H.. Jr., and Eichenwald, H. F. Antimicrobial therapy: Therapeutic recommendations and a review of newer drugs. J. Pediatr. 85 (1974). 297-312.
ANTIBIOTIC THERAPY FOR NEONATAL SEPSIS AND MENINGITIS