Roger G. Faix

Systemic Candida infections in infants in intensive care nurseries: High incidence of central nervous system involvement

The clinical courses in 27 infants with culture or autopsy evidence of systemic candidiasis were reviewed. Twenty-two infants (group 1) had persistent signs of sepsis and clinical deterioration or died before institution of antifungal therapy. Five infants (group 2) improved markedly before culture results were reported, and recovered without systemic antifungal therapy. Fourteen infants in group 1 (64%) had central nervous system infection. Of four patients in whom CNS involvement was diagnosed only postmortem, antemortem cerebrospinal fluid from three was abnormal despite sterile cultures; no antemortem CSF was obtained in the other. In meningitis caused by susceptible organisms addition of flucytosine sterilized CSF within 5 days, although prior amphotericin monotherapy had been unsuccessful. Of 14 patients in group 1 who received systemic antifungal therapy, only one died with Candida infection. Toxicity from antifungal agents occurred in 11 of 13 successfully treated infants, but was reversible in every case except one by modifying the dosage. Our data indicate that (1) CNS infection is very common in infants with systemic candidiasis, (2) combined flucytosine-amphotericin therapy may facilitate treatment of CNS infection and should be the initial therapy for systemic candidiasis in infants, (3) Gram stains of CSF and urine enhance early diagnosis, (4) isolation of Candida from normally sterile body fluids in high-risk infants should be considered pathogenic and therapy initiated unless the clinical course strongly suggests otherwise, and (5) toxicity from antifungal agents is common but usually reversible.

Frequent resistance of clinical group B streptococci isolates to clindamycin and erythromycin

Objective To determine both the frequency of reported penicillin allergy in parturients and the frequency of resistance in vitro of clinical isolates of group B streptococci to clindamycin and erythromycin. Methods One hundred clinical isolates of group B streptococci were tested to determine the frequency of resistance to clindamycin, erythromycin, penicillin G, vancomycin, and cefazolin. The frequency of beta-lactam allergy and reported allergic reaction also were recorded for all consecutive laboring women during the 4-month study. Results The frequency of group B streptococcal resistance to clindamycin was 15% and to erythromycin was 16%. No isolates were resistant to penicillin G, vancomycin, or cefazolin. Twelve percent of the 963 women who delivered during the study reported a penicillin allergy, but only 30% of those could describe their allergic reaction. Conclusion In vitro resistance of group B streptococci to clindamycin and erythromycin occurred frequently in this population. Whereas the importance of this finding in vivo is uncertain, it raises concern about the possibility of inadequate prophylaxis using currently recommended alternatives in penicillin-allergic patients. Artful questioning of women reporting penicillin allergy may lessen the likelihood of using these less desirable agents in the setting of intrapartum antimicrobial prophylaxis.

Invasive neonatal candidiasis: Comparison of albicans and parapsilosis infection

Severe infections caused by non-albicans Candida species are being increasingly reported among infants in neonatal intensive care units. To assess relative severity, mortality rates for C. albicans (CA) and C. parapsilosis (CP) infections in one neonatal intensive care unit from 1980 to 1990 were compared. Invasive candidiasis was defined as Candida recovery from a normally sterile body fluid or site with clinical signs of infection. Invasive candidiasis was diagnosed and systemic antifungal therapy initiated in 45 infants, 29 with CA and 16 with CP. No differences were found between CA and CP for birth weight, gestational age, age or weight at onset, presence of necrotizing enterocolitis, gastrointestinal or genitourinary anomalies, death (all causes), prior incidence or duration of antibiotics, parenteral nutrition, steroids or endotracheal intubation. Candida infection as the cause of death was more frequent with CA than CP (7 of 29 vs. 0 of 16; P = 0.034). Infants with CA were more likely to have antecedent thrush (P = 0.007) and perineal Candida dermatitis (P less than 0.02); those with CP were more likely to have vascular catheters at the time of positive culture (P less than 0.02). Though both pathogens occur in similar neonatal intensive care unit infants and can cause severe disease, CA appears more likely to result in death than CP.

Persistently positive cultures and outcome in invasive neonatal candidiasis.

Background. A persistently positive culture >24 h after starting antibiotic therapy has been correlated with adverse outcome in several invasive bacterial infections, but few reports address persistent positivity and outcome in infections caused by fungi and other pathogens that replicate more slowly and therefore may succumb less quickly to therapy. Methods. To assess whether positive culture >24 h after achieving target doses (amphotericin ≥0.5 mg/kg/day or fluconazole ≥6 mg/kg/day) of systemic antifungal therapy predicts focal infectious complication(s) or death from infection, we compared neonatal intensive care unit infants who had persistent (P+) or nonpersistent (P−) positive cultures with invasive candidiasis (clinical signs of infection and recovery of Candida from a normally sterile site) at this center from January 1, 1981, through June 30, 1999. Infants who died ≤ 24 h after attaining target dosing, recovered without therapy, had a focal infectious complication already present at the time target dosing was achieved or were diagnosed with invasive candidiasis only postmortem were excluded. Results. We identified 58 P+ (29, 12 and 7 had positive cultures for >7, >14 and ≥21 days, respectively) and 38 P− infants. No differences were found between P+ and P− for birth weight; gestational age; gender; onset age; central vascular catheters; necrotizing enterocolitis, surgery or bacterial sepsis; or duration of parenteral nutrition, antibiotics, tracheal intubation or postnatal steroids. P+ were more likely to have blood or cerebrospinal fluid involvement (68 vs. 45%, P = 0.03). Distribution of Candida species was similar (albicans in 53 vs. 63% for P+vs. P−). P+ were significantly more likely to develop later “fungus ball” uropathy (16 of 56 vs. 2 of 32, P = 0.01), to develop renal infiltration (11 of 56 vs. 1 of 32, P = 0.03) and to die from invasive candidiasis (11 of 58 vs. 0 of 38, P = 0.003) than P−. P+ were also more likely to develop endocarditis, abscess, ventriculitis and invasive dermatitis, although P > 0.05. Focal complication increased as duration of P+ increased (48, 55, 67 and 71% at >1, >7, >14 and ≥21 days, P = 0.06). When comparing only those with positive blood and/or cerebrospinal fluid culture, similar patterns were observed, although only death and focal complication or death from invasive candidiasis attained significance. Conclusions. These observations suggest that in neonatal invasive candidiasis: (1) cultures usually remain positive >24 h after attaining target antifungal doses; (2) aggressive imaging for focal complications may be reserved for infants with persistently positive cultures after several days of antifungal therapy at target doses or have signs strongly suggestive of focal complication; (3) focal complications and/or death from candidiasis increase with persistence; (4) focal complications increase with duration of persistence; (5) serial culture of infected site(s) helps predict outcome and the need for aggressive surveillance and intervention for focal complications.

Randomized, blind trial of dopamine versus dobutamine for treatment of hypotension in preterm infants with respiratory distress syndrome.

To compare the efficacy of dopamine and dobutamine for the treatment of hypotension (mean arterial blood pressure, < or = 30 mm Hg) in preterm (< or = 34 weeks of gestation) infants with respiratory distress syndrome in the first 24 hours of life, we enrolled 63 hypotensive preterm infants in a randomized, blind trial. Inclusion criteria required an arterial catheter for measurement of mean arterial blood pressure, treatment with exogenous surfactant, and persistent hypotension after volume expansion with 20 ml/kg (packed erythrocytes if hematocrit < 0.40, 5% albumin if > or = 0.40). Intravenous study drug infusions were initiated at 5 micrograms/kg per minute and then increased in increments of 5 micrograms/kg per minute at 20-minute intervals until a mean arterial blood pressure > 30 mm Hg was attained and sustained for > or = 30 minutes (success) or a maximum rate of 20 micrograms/kg per minute was reached without resolution of hypotension (failure). The study groups at entry were comparable for birth weight, gestational age, postnatal age, gender, birth depression, hematocrit < 0.40, heart rate, oxygenation index, delivery route, maternal chorioamnionitis, and maternal magnesium or ritodrine therapy. No infants in the dopamine group had a treatment failure (0/31; 0%); (16%) of 32 infants failed to respond to dobutamine (p = 0.028). Success was attained at < or = 10 micrograms/kg per minute in 30 (97%) of 31 infants given dopamine and in 22 (69%) of 32 infants given dobutamine (p < 0.01). Among those treated successfully, the increase in mean arterial blood pressure was significantly higher in those given dopamine (mean, 11.3 vs 6.8 mm Hg; p = 0.003). We conclude that dopamine is more effective than dobutamine for the early treatment of hypotension in preterm infants with respiratory distress syndrome.

Maternal and transplacental pharmacokinetics of cefazolin

OBJECTIVE To evaluate the intrapartum pharmacokinetics of cefazolin, including delivery to amniotic fluid (AF) and fetal compartments, and to ascertain that adequate cefazolin concentrations are attained to exceed the mean concentration inhibiting 90% (MIC90) of group B streptococcus strains. METHODS Cefazolin (1 g) was administered intravenously at five separate time intervals (0.5, 1, 2, 4, and 6 hours) before elective cesarean at term to 26 women with intact membranes and with no significant infections or cardiovascular, liver, or renal disease. Samples of maternal blood, cord blood, and AF were obtained at the time of delivery. Exact collection times relative to cefazolin infusion were noted. Amniotic fluid contaminated with blood or meconium was excluded. Cefazolin concentration was measured by high‐pressure liquid chromatography. RESULTS All maternal and cord plasma cefazolin levels, except one, were above the MIC90 for Streptococcus agalactiae (group B streptococcus). For AF, all cefazolin levels, except two, were above the MIC90. CONCLUSIONS Cefazolin concentrations greater than or equal to the MIC90 for group B streptococcus were attained in nearly all maternal, fetal, and AF samples. This information, together with the knowledge that there is rare resistance of group B streptococcus to cefazolin, supports the use of cefazolin as a better alternative than clindamycin or erythromycin for group B streptococcus prophylaxis in patients with a nonanaphylactic penicillin allergy.

Successful treatment of neonatal arterial thromboses with recombinant tissue plasminogen activator

Seven newborns were treated with recombinant tissue plasminogen activator for arterial thromboses. Complete lysis occurred in four of seven and partial in two of seven patients. Serious bleeding complications were observed in two of seven patients. This and published experience suggest that successful lysis with recombinant tissue plasminogen activator occurs in most patients and that hemorrhagic complications are unusual but are not.

The population pharmacokinetics of theophylline in neonates and young infants

The population pharmacokinetics of theophylline were evaluated using 391 theophylline serum concentration measurements from 108 neonates and young infants (postnatal age 0–26 weeks), who received theophylline for the treatment of neonatal apnea. A one-compartment pharmacokinetic model with first-order elimination was used, with intravenous aminophylline and oral theophylline administration modeled as zero-order infusions. The effect of a variety of developmental and demographic factors on clearance (CL) and volume (V) were investigated. Hypothesis testing to evaluate potentially significant factors produced a final model in which clearance was based on weight (kg) raised to an exponential power and postnatal age (weeks), with CL (ml/hr)=17.5 (weight)1.28 + 1.17 (postnatal age). Clearance was reduced by 12% for patients receiving parenteral nutrition. Volume of distribution in this population was adequately described using only weight, with V (L)=0.858 L/kg. Bioavailability of orally administered drug was not significantly less than unity. Interindividual variability in clearance was modest, with a coefficient of variation for clearance of 16%. An estimate of interindividual variability in volume could not be obtained. As a measure of residual variability in theophylline serum concentrations, the coefficients of variation for theophylline serum concentrations of 5.0, 10.0, and 13.0 mg/L were found to be approximately, 25, 12, and 9%, respectively. The identification of influential patient factors and the quantification of their influence on theophylline disposition allow for a priori estimates of theophylline pharmacokinetic parameters in these patients.

Persistent bacteremia and outcome in late onset infection among infants in a neonatal intensive care unit.

Background. Persistent bacteremia despite antibiotic therapy has been correlated with adverse outcomes, including focal suppurative complications and death. Coagulase-negative staphylococci (CONS) are the most common cause of nosocomial infection in infants requiring neonatal intensive care and might yield more substantial pathology if infection were persistent. Methods. To compare the severity and features of persistent infection by CONS with those of other bacteria, we reviewed infants admitted to our neonatal intensive care unit from 1990 through 2001 who developed bacteremia at >5 days of age with recovery of the same bacterial species from blood for >24 h after initiation of antibiotic therapy to which the organism was susceptible. Cases were excluded if a focal complication was already present with the initial positive culture or if the medical record was unavailable. Outcomes of interest included focal suppurative complications, death attributable to infection and duration of hospitalization among survivors. Results. We identified 62 infants with sustained infection, caused by CONS in 30 and by other organisms in 32 [10 Gram-negative, 22 Gram-positive (16 Staphylococcus aureus)]. Infants with persistent CONS had significantly lower birth weight and gestational age, but no difference was found for multiple other clinical and demographic risk factors. Indwelling vascular catheters were present at diagnosis in 85% of the infants (CONS 26 of 30, non-CONS 27 of 32). Responses of bacteremia to catheter removal vs. in situ treatment did not differ between the groups. No differences were observed for death from all causes (27 vs. 34%), death attributable to infection (6 vs. 12%) or duration of hospitalization among survivors [median (interquartile range): 102 (73 to 167) vs. 107.5 (89 to 130) days]. Focal suppurative complications were significantly more frequent in infants persistently infected with non-CONS (28 vs. 3%;P = 0.01). Duration of persistence correlated with focal complication in non-CONS infants (r = 0.988;P < 0.001). Conclusions. Although persistent infection with CONS occurs in significantly smaller and less mature infants than with non-CONS, death is no more frequent and focal complications are significantly less frequent. Infants with persistent infection should undergo aggressive evaluation for focal complications, with the yield expected to be higher in those with non-CONS.

Ureaplasma urealyticum and chronic lung disease in very low birth weight infants during the exogenous surfactant era.

BACKGROUND An association between recovery of Ureaplasma urealyticum from the respiratory tract of very low birth weight (VLBW) infants (< or =1500 g) and later chronic lung disease (CLD) was reported by several authors before the routine use of exogenous surfactant (SURF). We sought to assess whether this relation persists in the era of routine SURF. METHODS We prospectively studied a cohort of 105 VLBW infants who required mechanical ventilation at < 12 h of age. Tracheal aspirates for U. urealyticum culture were obtained before administration of SURF or antibiotics. Clinicians were unaware of U. urealyticum status. Chest radiographs at 28 days were reviewed by a single pediatric radiologist, blinded to U. urealyticum status. Sample size was predetermined to detect a 30% increase in CLD among those with U. urealyticum recovery from tracheal culture (U. urealyticum-positive) with alpha <0.05 and beta <0.20. RESULTS Of the study infants 22 were U. urealyticum-positive and 83 were U. urealyticum-negative. No differences were found between the groups for birth weight, gestational age, gender, inborn, antenatal or postnatal steroid use, SURF therapy, non-U. urealyticum infection, necrotizing enterocolitis, patent ductus arteriosus, intraventricular hemorrhage or cystic periventricular leukomalacia. At 28 days U. urealyticum-positive patients were significantly more likely to have CLD than U. urealyticum-negative [15 of 22 (68%) vs. 30 of 83 (36%); P < 0.02]. The U. urealyticum-positive patients also required significantly longer courses of supplemental oxygen and mechanical ventilation. No significant differences were found for CLD at 36 weeks postconception or duration of hospitalization, although type II error could not be excluded for these secondary endpoints. CONCLUSIONS Respiratory U. urealyticum at or shortly after birth remains associated with CLD at 28 days despite routine use of SURF. Controlled trials of anti-Ureaplasma therapy in U. urealyticum-positive VLBWs as soon after birth as possible may determine whether CLD, duration of respiratory support and attendant costs can be decreased.