 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 105 No. 5 May 2000, pp. 1041-1045
,, and
From the * Section of Infectious Diseases, Department of
Pediatrics, Baylor College of Medicine; and Department of Infection
Control, Woman's Hospital of Texas, Houston, Texas.
 |
ABSTRACT |
|---|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Background and Objective. An increase in the incidence of systemic candidiasis (SC) followed a change in skin care for extremely low birth weight (ELBW) infants in our neonatal intensive care unit (NICU). We sought to determine whether the use of topical petrolatum ointment (TPO) for skin care of ELBW infants was associated with risk for SC.
Study Design. Case-control study.
Setting. A 48-bed NICU in a private hospital in Houston, Texas.
Patients. Ten ELBW infants with and 30 without SC admitted to the NICU from December 1, 1997 through July 31, 1998.
Methods. ELBW infants with SC were identified using
hospital microbiology and infectious disease consultation databases. A
case was defined as an infant weighing 
1000 g at birth with
Candida spp isolated from a normally
sterile body site. Three infants without SC were matched to each case
by birth weight, admission date, and survival to the age of SC onset
for the case. Data were collected by retrospective medical record
review. Molecular analysis of Candida isolates was
performed by karyotyping and restriction fragment length polymorphism
using pulsed-field gel electrophoresis.
Results. Case infants had a mean (± standard deviation) age of onset of 21.5 ± 24 days. Infants with SC and controls did not differ in birth weight, gestational age, or duration of therapy with steroids, antibiotics, insulin, or total parenteral nutrition. Although cases were more likely to be born vaginally and had a longer duration endotracheal intubation than controls, these differences were not significant. The odds ratio for skin care with TPO in case infants versus control infants was 11 (95% confidence interval: 1.9-63). Skin care with TPO was discontinued and the incidence of SC decreased to baseline. Several Candida spp and genetic profiles were identified, suggesting that there was not a common source outbreak.
Conclusions. We conclude that the use of TPO promoted an increase in the incidence of SC in ELBW infants. Additional investigation of potential infectious risks for ELBW infants receiving TPO skin care is warranted. Key words: Candida, neonate, premature, low birth weight, petrolatum, skin care.
Systemic candidiasis (SC) is a frequent cause of
infection in low birth weight (BW) premature infants.1,2 In a recent multicenter study of late-onset sepsis in neonates, Candida spp accounted for 7% of bloodstream
infections.3 In the report of the National Nosocomial
Infections Surveillance System, Candida spp were
second in frequency among blood culture isolates from neonates in
high-risk nurseries.4 Risk factors associated with SC
include prematurity, broad-spectrum antibiotics, steroids, prolonged
endotracheal intubation, central venous catheters, and parenteral
alimentation.1,2,5 The frequency of Candida spp infections has increased in some institutions and may be
associated with new therapies and improved survival of extremely low
birth weight (ELBW) premature infants.6,7
Infants born before 34 weeks of gestation have immature epidermal
barrier function that results in increased risk for excessive transepidermal water loss, dermatitis, and infection.8 Since 1990, several investigations of methods to reduce complications associated with this immature skin barrier have been
reported.9-12 In a study of infants <33 weeks'
gestation, Nopper et al12 concluded that use of
preservative-free petrolatum-based emollient (Aquaphor, Beiersdorf Inc,
Norwalk, CT) reduced transepidermal water loss, improved the skin
condition, and reduced bacterial colonization as well as nosocomial
infections in premature neonates. Although a small number of ELBW
infants were enrolled in that study, the use of topical petrolatum
ointment (TPO) skin care has been implemented for ELBW infants in some
neonatal intensive care units (NICUs). However, the potential risk for
nosocomial infection associated with this therapy has not been
investigated in a large, prospective, clinical trial.
In the NICU of Woman's Hospital of Texas, in Houston, Texas, the use
of TPO for skin care of ELBW infants was initiated in December 1997. The incidence of SC per 1000 NICU patient-days in this nursery
increased from 5.9% (.5 per 1000 patient-days) in 1997 to 17.4% (1.7 per 1000 patient-days) during the first quarter of 1998. The purpose of
this study was to determine whether the use of TPO for skin care was
associated with this increase in cases of SC in ELBW infants.
Research Design
A case-control study was performed to determine whether there
was an association between SC and TPO skin care.
Study Infants
The study population was selected from all inborn ELBW (BW:
Study Methods
Medical records were reviewed to ascertain BW, gestational age
(GA), gender, mode of delivery, duration of endotracheal intubation, and final outcome (death or discharge). Medical and pharmacy records were reviewed to determine duration of therapy with antibiotics, dexamethasone, total parenteral nutrition, and TPO for skin care. Treatment with insulin was used as a marker for hyperglycemia. All data
were recorded on a standardized form and entered into a computer
database.
The investigation of the increase of SC incidence consisted of
observation of NICU practices and review of pharmacy protocols for
dispensing medications. Swabs of the NICU environment, equipment, and
new and used vials of petrolatum were cultured on sheep blood agar
plates (Baltimore Biological Laboratory, Cockeysville, MD) and
Sabouraud dextrose media (Remel, Lenexa, KS) for isolation of
Candida spp. All bloodstream isolates of
Candida spp were subjected to genetic analysis by
karyotyping and restriction fragment length polymorphism using
pulsed-field gel electrophoresis (provided by J. E. Patterson,
University of Texas Health Science Center, San Antonio, TX).
Statistical Methods
An unpaired t test was used when appropriate for
comparisons between case and control groups. Fisher's exact test and
odds ratios were used to compare the relative risk of TPO skin care between case and control groups. P < .05 was
considered significant. Statistical calculations were performed using
GraphPad statistical software (Prism, San Diego, CA).
Case-Control Study
During the 7-month study, 10 cases were identified. Clinical
characteristics of these infants are summarized in Table
1. The mean BW was 716 g (standard
deviation ± 37) and mean GA was 24.8 weeks (standard
deviation ± .2). The mean age of onset was 21.5 days (median: 10;
range: 6-82). SC was caused by Candida albicans in 6 infants, C parapsilosis in 3, and C lusitaniae in 1. Each case infant had candidemia; other foci of infection also were
present in 8 infants. Three infants had urinary tract infection, 3 had
ileal perforations with peritonitis, and 2 had meningitis (both C
albicans). None of the infants had invasive fungal dermatitis. Each case was treated with intravenous amphotericin B deoxycholate (10-30 mg/kg total cumulative dose).
TABLE 1
METHODS
Top
Abstract
Methods
Results
Discussion
References
1000 g) infants between December 1, 1997 and July 31, 1998. Our NICU
does not have a transport service; thus, the study did not include
outborn infants. The hospital microbiology laboratory and infectious
disease consultation service databases were reviewed to identify
possible SC cases diagnosed during the study. A case was defined as an
infant with signs of systemic infection and isolation of
Candida spp from blood, cerebrospinal fluid,
peritoneal fluid, joint aspirate, or urine (collected by
catheterization or suprapubic aspiration). Three consecutive controls
were selected for each case. Controls were matched by date of birth and
BW category (<750 g or 751-1000 g). Controls had to survive to at
least the age of diagnosis for the matched case. Two potential controls were replaced because they had received empiric antifungal therapy for
respiratory tract colonization with Candida.
RESULTS
Top
Abstract
Methods
Results
Discussion
References
Clinical Characteristics of 10 ELBW Infants With SC
A summary of the analysis for potential risk factors related to SC is found in Table 2. Because of the small sample size, multivariate analysis was not possible. Case and control infants did not differ by BW or GA. Although infants with SC were more likely to be male and delivered vaginally than controls, this difference was not statistically significant. Dexamethasone was frequently administered to case and control infants to treat refractory hypotension or an exacerbation of brochopulmonary dysplasia. The mean duration of dexamethasone therapy before SC infection did not differ in case and control infants. The duration of insulin therapy and total parenteral nutrition were similar in both groups of infants. Although the duration of antibiotic therapy and endotracheal intubation were longer in cases than controls, neither was statistically significant. All case and control infants survived to discharge.
|
White petrolatum United States Pharmacopeia (Sherwood Medical, St Louis, MO) was used for skin care in ELBW infants. Infants did not routinely receive TPO skin care, but this was ordered within the first 48 hours of life at the discretion of the admitting neonatologist. Therefore, the dosage, frequency, and duration of therapy varied. Recorded indications for TPO included fragile or cracked skin, excessive insensible water loss, or for skin care prophylaxis because of extreme prematurity. TPO was used for skin care significantly more frequently in infants with SC than in control infants (P < .01). Eight (80%) infants with SC, but only 8 (27%) control infants were treated with TPO. The odds ratio for skin care with TPO in SC case infants versus control infants was 11 (95% confidence intervals: 1.9-63).
Outbreak Investigation and Control
Fig 1 summarizes the incidence of Candida spp infection in the NICU before, during, and after the study. The introduction of TPO skin care was the only identified change in patient care practice that preceded the increased number of SC cases. Shared, multidose containers of medications were not used in the NICU before or during the outbreak. Individual vials of petrolatum were dispensed for each infant. Although some work areas were contaminated with ointment, environmental cultures of these work areas as well as of new and used vials of petrolatum did not grow Candida spp. Adherence to infection control policies for hand-washing and cohorting of infants was emphasized through inservice education. TPO therapy was discontinued on April 15, 1998. The incidence of SC decreased to <2% (.2 per 1000 NICU patient-days) during the third quarter of the year and was <5% the subsequent year (June 1998 through June 1999)
|
Genetic analysis of the 10 blood isolates of Candida spp confirmed that several strains were responsible for the cluster of SC cases (Fig 2). Among the C albicans isolates from 6 cases, 3 different patterns were identified. The 3 C parapsilosis isolates had a similar pattern and were likely related strains. The C lusitaniae isolate had a distinctive pattern.
|
| |
DISCUSSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
The overall incidence of SC in our NICU of 5.9% during 1997 was comparable to that reported by other institutions.6,7 When our incidence increased threefold, we sought to determine what factors might account for this rise. TPO skin care for ELBW infants was initiated in December 1997 in an attempt to prevent complications associated with skin immaturity. A case-control study documented that each of the infants included in the study had several factors known to increase risk for SC, including prematurity, previous treatment with broad-spectrum antibiotics, steroids, parenteral alimentation, endotracheal intubation, and hyperglycemia. Vaginal delivery and prolonged endotracheal intubation were more common in infants with SC, but these variables did not reach statistical significance. It is possible that with a larger sample size and multivariate analysis these factors might be covariables for SC. However, infants with SC were significantly more likely to have skin care with TPO, compared with other infants matched for BW. We speculate that TPO provided a milieu in which the colonizing yeast could proliferate. Experimental models of mucocutaneous candidal infection are established by epicutaneous inoculation of Candida spp followed by an occlusive dressing.13 Occlusion of the skin also alters pH, CO2 emission rate, and microbial flora of the skin and this enhances hyphal formation of Candida spp.14,15 Invasion of hyphae into the epidermis is enhanced by treating the experimental animals with prednisolone.16 These conditions mimic those of ELBW infants exposed to Candida spp at birth and treated with topical occlusive ointment for skin care and dexamethasone for refractory hypotension.
Some infants treated with TPO had fragile or cracked skin. Skin condition may have been a confounding variable because the interrupted integument could be an additional risk factor for SC. However, given the retrospective study design and the lack of a standardized scoring for skin condition, this could not be determined.
The petrolatum-based emollient used in the study by Nopper et al12 was Aquaphor (Beiersdorf Inc), and this product contains lanolin alcohol, panthenol, bisobalol, and glycerin in addition to petrolatum. In our study, the product used was white petrolatum United States Pharmacopeia (Sherwood Medical). Neither preparation contains preservatives that would inhibit the growth of microorganisms, and their effect on the growth of fungi has not been studied. Other yeast, such as Malasezzia spp, grow in lipid-containing environments. Another possibility is that TPO enhanced the adherence of C albicans to mucocutaneous surfaces. Established factors that influence hyphal formation and adherence to host cells include glucose concentration, endogenous epithelial lipids, and glycoproteins.17-19 It is possible that alteration of host-parasite interactions could contribute to the progression from colonization to disease.
We considered the possibility that the increased incidence was caused by a common source outbreak. Although reported clusters of Candida spp infection in premature infants have been attributed to common source outbreaks, most temporal clusters are caused by genetically distinct strains.20,21 This was the case in our outbreak. SC cases were caused by 3 distinct Candida spp and 3 distinct DNA patterns among C albicans isolates were identified. Further, an environmental or medication source was not identified. Most cases were delivered vaginally and probably became colonized after exposure to Candida at birth. In addition, 5 cases occurred in infants <10 days of age suggesting that acquisition of the organism in the first week of life may have been a contributing factor for SC, as has been described by Rowen et al.22 An investigation of the mode of transmission of Candida spp to premature infants demonstrated that most infants with C albicans infection acquire the organism vertically, whereas C parapsilosis is more often associated with nosocomial acquisition.23 The genetic similarity between C parapsilosis isolates from our 3 cases suggests that horizontal or nosocomial transmission may have been a contributing factor.24 Weems et al25 reported a cluster of cases of candidemia in a NICU after the use of a topical emollient for skin care. The incidence of candidemia declined after procedures to assure adequate skin cleansing had been applied. In our investigation, oily and inadequately cleaned surfaces in the NICU were noted and possibly contributed to horizontal transmission via the hands of health care workers. Although environmental culture results were negative, these were performed after the investigation and other control measures were in progress.
The study by Nopper et al12 suggested that a petrolatum-based emollient reduced the risk for nosocomial infections in premature infants. However, only 9 of the patients in the treatment group were ELBW infants, and other potential factors related to risk for SC were not provided by those authors. Several infants in their treatment group were colonized with Candida and Malassezzia spp, but patients were followed only for the first 14 days of life. It is conceivable that SC was not detected in study patients because the median age of onset of SC occurs after 14 days of life.7
In the NICU setting, changes in clinical practice may alter the incidence of infections that occur in this high-risk population, and especially ELBW infants. TPO may improve skin condition and reduce insensible water loss in some ELBW infants. However, in our NICU, it was associated with an increased risk for SC. The risk for nosocomial infections associated with this therapy has not been adequately investigated. A prospective, controlled trial of TPO skin care in a sufficiently large number of ELBW infants will be needed to detect potential infectious complications. NICUs that use TPO for skin care of ELBW infants are advised to carefully monitor the incidence of SC after TPO skin care is initiated.
| |
ACKNOWLEDGMENTS |
|---|
We thank the neonatology, laboratory, and administrative staff for their assistance in the investigation, and Robin Schroeder for her secretarial expertise.
| |
FOOTNOTES |
|---|
Received for publication Mar 23, 1999; accepted Jul 16, 1999.
Reprint requests to (J.R.C.) Baylor College of Medicine, Department of Pediatrics, 1 Baylor Plaza, Room 302A, Houston, TX 77030. E-mail: judithc{at}bcm.tmc.edu
| |
ABBREVIATIONS |
|---|
SC, systemic candidiasis; BW, birth weight; ELBW, extremely low birth weight; TPO, topical petrolatum ointment; NICU, neonatal intensive care unit; GA, gestational age.
| |
REFERENCES |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. K. Fridkin, D. Kaufman, J. R. Edwards, S. Shetty, T. Horan, and the National Nosocomial Infections Surveillance Sy Changing Incidence of Candida Bloodstream Infections Among NICU Patients in the United States: 1995-2004 Pediatrics, May 1, 2006; 117(5): 1680 - 1687. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. A. Bhutta, G. L. Darmstadt, B. S. Hasan, and R. A. Haws Community-Based Interventions for Improving Perinatal and Neonatal Health Outcomes in Developing Countries: A Review of the Evidence Pediatrics, February 1, 2005; 115(2/S1): 519 - 617. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Yan Preventing Sepsis in Premies: Do Emollients Help or Hurt? AAP Grand Rounds, November 1, 2004; 12(5): 55 - 55. [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kaufman and K. D. Fairchild Clinical Microbiology of Bacterial and Fungal Sepsis in Very-Low-Birth-Weight Infants Clin. Microbiol. Rev., July 1, 2004; 17(3): 638 - 680. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. H. Edwards, J. M. Conner, R. F. Soll, and for the Vermont Oxford Network Neonatal Skin Care The Effect of Prophylactic Ointment Therapy on Nosocomial Sepsis Rates and Skin Integrity in Infants With Birth Weights of 501 to 1000 g Pediatrics, May 1, 2004; 113(5): 1195 - 1203. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Topical Petrolatum Ointment May Increase the Risk of Neonatal Candidiasis Journal Watch Infectious Diseases, June 1, 2000; 2000(601): 16 - 16. [Full Text]
|
|
|
|
 |
|
 Topical Petrolatum for Premature Infants Journal Watch Dermatology, June 1, 2000; 2000(601): 5 - 5. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2000 American Academy of Pediatrics. All rights reserved. |
||
|
||
-ladder
DNA size standards, respectively. Lanes 2 to 8 contain C
albicans and lanes 9 to 13 contain C
parapsilosis isolates from 9 infant cases. The single C
lusitaniae case isolate is shown in lane 14.