Colonization with antibiotic-resistant Gram-negative bacilli in the neonatal intensive care unit.

Infections with Gram-negative bacilli have affected critically ill newborns throughout the era of modern neonatal intensive care. Selected topics regarding the acquisition of Gram-negative bacteria, particularly those expressing antibiotic resistance, in the neonatal intensive care unit (NICU) are reviewed. Some data suggest that the recent introduction of intrapartum antibiotic prophylaxis programs, through which selected women are administered penicillin or ampicillin during labor to prevent vertically transmitted Group B streptococcal disease, has increased the incidence and resistance of early onset Gram-negative bacillary infection in the newborn over the past 5 years. Additional data suggest that late onset, nosocomially-acquired Gram-negative bacillary infections also are increasing among critically-ill newborns. The seminal event in Gram-negative bacterial infection in the hospitalized infant is colonization of the gastrointestinal tract and nasopharynx. Newborns admitted to the intensive care unit acquire hospital-associated Gram-negative bacterial colonization throughout their NICU stay, presumably transmitted from the environmental flora via the hands of caregivers. Colonization specifically by antibiotic-resistant bacilli follows a similar pattern, suggesting that such bacteria make up part of the modern NICU ecology and are acquired by infants similar to susceptible microorganisms. Althou-gh some clinically-undetected cross-transmission of resistant bacilli occurs during non-outbreak periods, most colonizing antibiotic-resistant bacilli are unique to each infant. The role of antibiotic exposure on the acquisition of antibiotic resistant bacilli in the intensive care nursery is difficult to calculate given the covariance of such exposure to other markers of severe disease. Experience has demonstrated, however, that use of the aminoglycosides is infrequently associated with emergence of resistance in the newborn, whereas the use of higher-generation cephalosporins may be associated with the rapid appearance of bacilli resistant to betha-lactams.

Molecular epidemiology of antibiotic-resistant gram-negative bacilli in a neonatal intensive care unit during a nonoutbreak period.

OBJECTIVE: Gram-negative organisms that are resistant to parenteral antibiotics are a growing threat to hospitalized patients. This study was conducted to define the epidemiologic characteristics of these organisms during a nonoutbreak period in a neonatal intensive care unit (NICU). METHODS: Nasopharyngeal and rectal swab specimens were obtained 3 times a week from every infant in a tertiary care NICU during a 12-month period. Specimens were processed to identify aerobic Gram-negative species resistant to gentamicin, piperacillin-tazobactam, or ceftazidime. Selected clinical parameters were tested for their association with colonization with a resistant organism. Restriction endonuclease digests of genomic DNA were derived from isolates of the most frequently occurring species. The fragments were analyzed by pulsed-field gel electrophoresis (PFGE) to determine the genetic relatedness of the various isolates and thereby determine the length of colonization, the frequency of horizontal transmission, and the size and duration of clusters. RESULTS: A total of 101 infants (8.6%) of 1180 admissions were colonized with at least 1 antibiotic-resistant bacillus before NICU discharge. Multiple parameters indicating a prolonged, complicated NICU course were associated with resistant colonization, including gestational age, length of stay, and exposure to several classes of antibiotics. Colonization with resistant bacilli occurred as early as the first NICU day, but acquisition continued throughout the infants' stay. A total of 436 isolates were analyzed by PFGE. On the basis of this molecular analysis, it was determined that duration of colonization was usually very short; the median for all species tested was <1 week. In addition, cross-colonization occurred in only 12% of all PFGE-analyzed isolates. Most clusters of cross-colonized infants were small, with the majority involving only 2 patients. CONCLUSIONS: During endemic periods, acquisition of antibiotic-resistant Gram-negative bacilli in the NICU may occur very soon after admission, but colonization continues over many weeks of NICU stay. The duration of colonization with resistant bacilli is short, and horizontal transmission is unusual. These characteristics suggest a gradual but temporary incorporation of these organisms from the NICU environment into the nascent newborn microflora over time with little cross-colonization. These observations may aid the rational development of infection-control strategies to contain the reservoir of resistant Gram-negative organisms in the NICU.antibiotic resistance, Gram-negative bacilli, neonatal intensive care, antibiotic utilization, colonization, pulsed-field gel electrophoresis.

Etiology of fever and opportunities for reduction of antibiotic use in a pediatric intensive care unit.

OBJECTIVE: To determine the cause of fever in critically ill children and to identify opportunities for reducing antibiotic use in this population. DESIGN: Prospective case series. SETTING: A tertiary-care medical-surgical pediatric intensive care unit (PICU). PATIENTS: Children admitted to the PICU who experienced fever (axillary temperature >38.3 degrees C). MEASUREMENTS: Consecutive children who were febrile at any point in their PICU stay were investigated over two winter seasons. Etiology of the fever was determined by physical examination and routine microbiology and radiographic tests. Three subgroups were reviewed to approximate the number of antibiotic-days that could have been reduced; namely, those with an indeterminate source, those with a documented viral infection, and those receiving a prolonged course of antibiotics. A set of standards reflecting common antibiotic use then was applied to these three patient groups. RESULTS: Of 211 subjects, the majority (83.3%) had either a definitive or suspected focus for their fever, and nearly all of these patients were judged to have an infectious etiology. The study population received a total of 2,036 antibiotic-days. Despite the high incidence of infectious causes of fever in our subjects, however, approximately 15% of total antibiotic-days could have been reduced by applying common-use standards. CONCLUSIONS: Fever in the PICU was usually of defined focus and infectious in origin. However, among febrile patients in the PICU, substantial opportunity exists for reduction of antibiotic use. Trials determining the safety of antibiotic reduction in this population should be pursued vigorously.

Infection control in pediatric hospitals.

Important characteristics of hospital infection control are specific to pediatric facilities. For example, colonization and infection with vancomycin-resistant enterococci, which are widely spread in many units housing adult patients, are uncommon in children, especially in the neonatal intensive care unit where vancomycin use is heavy. Characteristics of the neonatal intensive care unit, such as the insulated environment and infrequent treatment with antibiotics with broad anaerobic activity, likely account for this finding. Artificial fingernails have been discovered to promote colonization with potential pathogens; their implication in recent nursery epidemics emphasizes the need to prohibit their use in this environment in particular. Finally, nosocomial viral infections occur with regularity in pediatric hospitals. Programs that successfully and cost-effectively control hospital spread of respiratory syncytial virus, however, demonstrate that rational, multifaceted interventions can nearly eliminate transmission of certain viral pathogens on the pediatric wards.

Use of real time pulsed field gel electrophoresis to guide interventions during a nursery outbreak of Serratia marcescens infection.

BACKGROUND: Pulsed field gel electrophoresis (PFGE) is a commercially available technique that can establish clonal relationships among many common hospital-derived organisms with a high degree of accuracy and can yield results in a sufficiently short time to guide interventions during an outbreak investigation. METHODS: The CHEF Genomic Bacterial DNA Plug Kit (Bio-Rad) was applied to an unfolding nursery outbreak of Serratia marcescens infections according to the manufacturer's guidelines. Bacterial genomic DNA was digested with XbaI or SpeI and separated on 1% agarose gels, and the isolates were grouped by restriction endonuclease patterns according to established standards. RESULTS: S. marcescens was isolated from nine patients in an intensive care nursery during an 8-week period. Initial PFGE analysis performed after identification of the first eight patients, when closure of the nursery was imminent, revealed that the epidemic was caused by two groups of four isolates each. In both instances the group was geographically contained, and the nursery remained open. A second PFGE analysis indicated that a ninth S. marcescens isolate, recovered in Week 8, was genetically unrelated to the other two. Surveillance during an additional 6 weeks revealed no new cases, and the epidemic was declared over. No cases of invasive S. marcescens infection were identified during the subsequent 10 months. CONCLUSION: Real-time PFGE determined that an apparent nursery outbreak of S. marcescens infection was, in fact, caused by three genetically distinct strains. This information allowed the nursery to remain open after other appropriate infection control measures had been imposed.

Factors that predict preexisting colonization with antibiotic-resistant gram-negative bacilli in patients admitted to a pediatric intensive care unit.

OBJECTIVE: To predict which patients hospitalized in a pediatric intensive care unit (ICU) are colonized with antibiotic-resistant gram-negative rods on admission. METHODS: Consecutive children admitted to a pediatric ICU over a 6-month period were entered into the study. A questionnaire soliciting information regarding the child's medical history and home environment was completed by the parent or guardian. Nasopharyngeal and rectal cultures were obtained on each of the first 3 days of ICU admission, and organisms resistant to ceftazidime or tobramycin were identified. Only clonally distinct organisms, as confirmed by pulsed field gel electrophoresis, were analyzed. The association between identification of colonization with an antibiotic-resistant gram-negative rod within 3 days of ICU admission and factors included in the questionnaire was tested by chi2 or t test. RESULTS. In 64 (8.8%) of 727 admissions, an antibiotic-resistant gram-negative bacillus was isolated within the first 3 ICU days. More than half were identified on the day of admission. Colonization was associated with two factors related to the patient's medical history, namely, number of past ICU admissions (1.98 vs.87) and administration of intravenous antibiotics within the past 12 months (67.9% vs 28.2%). No association was found between colonization and exposure to oral antibiotics. In addition, factors related to the child's environment were also associated with presumed importation of an antibiotic-resistant gram-negative rod into the ICU. Specifically, residence in a chronic care facility was strongly associated with colonization (28.3% vs 2.6%); exposure to a household contact who had been hospitalized in the past 12 months also predicted colonization (41.7% vs 18.5%). CONCLUSIONS: These data suggest that a profile can be established characterizing children colonized with resistant gram-negative bacilli before admission to a pediatric ICU. Infection control measures may help to contain these potentially dangerous bacteria once they have been introduced into the unit.

Human immunodeficiency virus (HIV)-related risk-taking behaviors in women attending inner-city prenatal clinics in the mid-west.

OBJECTIVE: Federal guidelines recommend the provision of human immunodeficiency virus education to all attendees of prenatal clinics. The current study was conducted to assess risk-taking behaviors among urban women voluntarily pursuing prenatal care. DESIGN AND SETTING: African American women attending urban prenatal clinics in Cleveland, Ohio were subjected to an extensive interview before receiving an experimental AIDS education curriculum. The interview sought detailed information regarding demographics, lifetime and recent sexual activity, condom use, and lifetime and recent illicit drug use. RESULTS: A total of 1017 women were interviewed; of those women, approximately 73% were single. The majority had a monthly income of less than $500. A total of 66% had only one partner in the past year, and almost 90% had < or = 1 partner in the past 6 months. A total of 98% identified a main partner. Nearly all subjects were at least fairly certain that this partner did not use intravenous drugs, and 71% were at least fairly certain that he was monogamous. Only 19% used condoms most or all the time. Intravenous drug use among study subjects was very infrequent. CONCLUSION: These data indicate that inner-city Cleveland women seeking prenatal care are largely monogamous around the time of their gestation, and that a history of intravenous drug use is infrequent. They suggest that prenatal counseling in urban clinics will need to address women who largely are engaged in single-partner relationships at the time of the intervention.

Antibiotic restriction does not alter endemic colonization with resistant gram-negative rods in a pediatric intensive care unit.

OBJECTIVE: To test whether a ceftazidime restriction policy in a pediatric intensive care unit (ICU) could decrease the endemic rate of colonization with ceftazidime-resistant gram-negative bacilli. DESIGN: Prospective, pre- vs. postintervention study. SETTING: University hospital pediatric ICU. PATIENTS: Consecutive children admitted to the pediatric ICU over a 19-mo period. INTERVENTIONS: After an observation period in which antibiotic use was not controlled, ceftazidime was prohibited unless the patient's microbiological results indicated that the drug was necessary for cure. Aminoglycoside use was not regulated. The size of the endemic reservoir of ceftazidime- and tobramycin-resistant organisms was determined by daily nasopharyngeal and rectal swab specimens obtained on all admissions to the ICU. MEASUREMENTS AND MAIN RESULTS: Despite a 96% reduction in ceftazidime use, the incidence density (number of isolates/100 patient-days) of ceftazidime-resistant organisms increased through the course of the study, from 1.57 to 2.16. The incidence density of tobramycin-resistant organisms was unchanged. Ceftazidime restriction resulted in a small but nonsignificant decrease in the proportion of ceftazidime-resistant organisms acquired late (beyond 72 hrs) in the patients' ICU course (56.5% vs. 45.9%). However, there was a more substantial decrease in the proportion of ceftazidime-resistant organisms derived from species known to harbor derepressible amp C beta-lactamases (68.2% vs. 45.9%, p < .05). CONCLUSIONS: These data indicate that antibiotic restriction policies in an ICU fail to diminish the size of the endemic reservoir of antibiotic-resistant gram-negative rods, and suggest that such policies in the absence of broader efforts to limit antibiotic use will have little impact.

Compliance with universal precautions among pediatric residents.

BACKGROUND: There are few data on the rate of compliance with universal precautions among pediatricians. We hypothesized that compliance in pediatrics would be poor because of the intrinsic difficulties in performing invasive procedures in small subjects. DESIGN: Prospective, observational study. SETTING: Tertiary care children's hospital. STUDY PARTICIPANTS: A convenience sample of pediatric house staff. MAIN OUTCOME MEASURES: Pediatric house staff members were observed while performing invasive procedures. Procedure type, number of attempts required, and patient's age and diagnosis were recorded. Degree of compliance with universal precautions was judged by means of Centers for Disease Control and Prevention guidelines. Comparisons between the compliant and noncompliant groups were analyzed by chi2 and 2-tailed t test. RESULTS: A total of 128 procedures performed by 43 house officers, 4 advanced medical students, and 3 chief residents or fellows were observed. Sixty-nine (53.9%) of the 128 procedures were performed correctly according to universal precaution guidelines. Rate of compliance did not appear to be influenced by small patient size, as judged by the lack of association with the age of the patient (mean+/-SD, 4.8+/-5.7 years among those in whom universal precautions were properly used vs 4.9+/-5.4 years among patients in whom precaution guidelines were breached; P=.96). Moreover, the number of attempts required in compliant procedures (1.31+/-0.53) was almost identical to that in noncompliant procedures (1.28+/-0.49; P=.73). Additionally, compliance did not improve with advanced level of training. CONCLUSIONS: Failure of compliance among pediatricians has no apparent association with procedure difficulty, and compliance rates continue to be poor through the course of pediatric training. These findings underline the need for effective education concerning universal precautions throughout pediatric residency, and they suggest that such efforts will not be precluded by obstacles intrinsic to performing invasive procedures on young subjects.

Colonization with antibiotic-resistant gram-negative organisms in a pediatric intensive care unit.

OBJECTIVE: To measure the prevalence of colonization with antibiotic-resistant Gram-negative organisms and its association with potential risk factors, including antibiotic exposure, in a pediatric intensive care unit (ICU). DESIGN: Prospective, observational study. SETTING: A 16-bed tertiary care pediatric ICU. PATIENTS: All children admitted to the pediatric ICU for > 24 hrs over a 5-month period. MEASUREMENTS AND MAIN RESULTS: Two hundred ninety-six patients, approximately half of all patients admitted to the ICU, were enrolled in the study; 236 patients had sufficient data collected for analysis and were prospectively examined for nasopharyngeal and gastrointestinal colonization by antibiotic-resistant Gram-negative organisms (ceftazidime minimal inhibitory concentration of > 16 micrograms/mL, or tobramycin minimal inhibitory concentration > 8 micrograms/mL). Association between colonization and potential predisposing factors including demographics, diagnosis, Pediatric Risk of Mortality (PRISM) score, invasive instrumentation, and prior ICU antibiotic exposure was assessed. More than 20% of patients were found to be colonized with an antibiotic-resistant Gram-negative organism. Examination of the timing of colonization indicated that more than half were identified within 72 hrs of admision. Colonization was associated by unadjusted analysis to prior ICU antibiotic exposure, as well as by factors associated with the severity of illness (PRISM score and invasive instrumentation) and young age. However, when the independence of these factors was tested by logistic regression, prior antibiotic exposure was no longer associated with resistant organism colonization. CONCLUSIONS: These data suggest that antibiotic-resistant Gram-negative organisms are a significant risk to intensively III children and that in many instances, they are imported into the unit or rapidly acquired from environmental reservoirs. Since risk factors for colonization are multiple, policies confined to antibiotic utilization within the ICU may have fixed, and possibly limited, benefit in their control.

Problems with resistance in pediatric intensive care.

Antibiotic resistance in the neonatal and pediatric intensive care environments has not been rigorously investigated. There is reason to believe that the epidemiology of antibiotic resistance in these settings may be different from that in other hospital settings because the patients' preadmission health status, the maturity of their immune systems, and their outpatient exposure to antibiotics are different from those seen in adults. At the present time, the areas of greatest concern are the outbreaks of infection caused by methicillin-resistant Staphylococcus aureus in the neonatal ICUs and the emergence of antibiotic-resistant Gram-negative bacteria in pediatric intensive care units (PICU). In the former, colonization and transmission by nursery personnel remains one of the great challenges in infection control. In the latter, new information is emerging which challenges the notion that antibiotic-restriction policies might be an effective means for modulating the emergence of antibiotic-resistant Gram-negative pathogens in the pediatric intensive care environment. It appears that these organisms are largely imported into the ICU from the community and are not a result of antibiotic practices within the unit itself. This observation requires that strategies to control these organisms in the PICU be reassessed.

Antibiotic-resistant gram-negative bacteria in the critical care setting.

Gram-negative bacilli that are resistant to commonly used antibiotics are a growing problem in seriously ill, hospitalized patients. Numerous outbreaks involving these organisms have been reported in intensive care nurseries and among critically ill adults. In endemic situations, the major reservoir for these pathogens is the patient; occasionally, transmission from patient to patient occurs through the hands of caregivers. Although the degree of antibiotic use probably plays some role in the emergence of antibiotic-resistant gram-negative bacilli, this relationship has not been uniformly demonstrated, and other factors intrinsic to the organisms themselves and to the critically ill patient may play an important role.

Predictive factors of short-term neurologic outcome in children with encephalitis.

When a child is admitted to the hospital with presumed encephalitis, the physician must use clinical criteria to gauge the appropriate level of hospital care and to give a preliminary assessment of outcome to the family because the etiology is unknown. This study attempted to determine which clinical factors gathered on hospital admission would be most helpful to the physician. The records of 106 children (ages 1 month to 20 years), admitted to Rainbow Babies and Childrens Hospital between 1978-1989 who had discharge diagnoses of encephalitis, were reviewed. Seventy-five met the case definition of presumed viral encephalitis, with viral etiology established in 23% of patients. Poor short-term outcome was defined as the presence of an abnormal neurologic examination at hospital discharge, and was present in 32% of patients. Focal signs on neurologic examination (odds risk: 16.30, P < .05) and abnormal neuroimaging studies (odds risk: 5.66, P < .05) were the only 2 factors present at admission that predicted a poor short-term outcome. Glasgow coma scale at admission was predictive of an abnormal neurologic examination at discharge only when profoundly depressed (6 or less); otherwise, this scale was not useful as a prognostic tool. Factors that were not correlated with adverse outcomes included age younger than 1 year, any type of seizure occurrence, status epilepticus, diffuse or focal electroencephalographic abnormalities, or abnormal cerebrospinal fluid findings.(ABSTRACT TRUNCATED AT 250 WORDS)

Use of semiquantitative human immunodeficiency virus type 1 culture to guide therapy in a premature infant.

Semiquantitative cultures for human immunodeficiency virus type 1 were used to initiate and guide therapy in an asymptomatic, heavily infected premature infant. The cultures were performed on fivefold serial dilutions of peripheral blood mononuclear cells, and viral growth was detected by the appearance of p24 antigen in the supernatant. While the patient was receiving low-dose zidovudine, her titer diminished from 206 infectious units per 10(6) cells to undetectable levels over 20 weeks, and she remained asymptomatic throughout her 1st year of life. This case suggests that semiquantitative HIV-1 culture should be further evaluated for its ability to guide practical therapeutic decisions in patients who do not fulfill currently established criteria for treatment.

Comparative embryonic cytotoxicity of antiretroviral nucleosides.

Previous experiments have indicated that zidovudine is cytotoxic to early murine embryos both in vivo and in vitro. Newer nucleoside analogs (ddI, ddC, and d4T) with antiretroviral activity were tested to determine whether they had similar toxicity. Exposure of two-cell embryos to each of these three drugs inhibited blastocyst formation only at concentrations > or = to 100 microM. Sublethal preblastocyst exposure to d4T resulted in failure to develop beyond the blastocyst stage at 10 microM; no effect was seen with ddC or ddI at concentrations up to 100 microM. In each instance, however, cytotoxicity of all three drugs was significantly less than with zidovudine at equivalent concentration. These experiments suggest that newer antiretroviral nucleosides may be safer to use in early pregnancy than zidovudine.

Effect of zidovudine on preimplantation murine embryos.

It previously has been demonstrated that zidovudine (AZT) is lethal to early murine embryos. The effect of the drug on pre- and postimplantation embryos was examined to delineate the timing of this toxicity and to investigate its possible mechanisms. Embryos exposed in the whole mouse during preblastocyst development were unable to proceed beyond the blastocyst stage. Similarly, when two-cell embryos harvested from unexposed females were exposed to low-concentration (1 microM) AZT in vitro over 24 h, development beyond the blastocyst stage was inhibited. In contrast, drug exposure during in vitro blastocyst and postblastocyst development resulted in little or no morphologic toxicity. Further investigation revealed that preblastocyst AZT exposure resulted in the development of blastocysts with significantly lower cell numbers than control embryos. While embryonic exposure to AZT at the blastocyst and postblastocyst stages also resulted in retarded cell division, the effects were milder than those recorded after preblastocyst exposure. These data demonstrate that the critical period of AZT toxicity toward murine embryos is between ovulation and implantation and indicate that AZT directly suppresses cell division in the preimplantation embryo.