OBJECTIVES: To assess trends in incidence of Candida spp. central line-associated bloodstream infections (CLABSIs) in US NICUs, 1999–2009.
METHODS: Data from NICUs participating in the National Nosocomial Infections Surveillance (1999–2004) and National Healthcare Safety Network (2006–2009) were analyzed. Overall and birth weight-specific incidence rates of Candida spp. CLABSIs per 1000 central line-days were calculated. Trends in incidence were assessed by using Poisson regression, and trends in proportion of CLABSIs identified as Candida albicans were assessed by using weighted-linear regression.
RESULTS: Overall, 398 NICUs reported 1407 Candida spp. CLABSIs (706 due to C albicans) among 1400 neonates. Of the 1400 neonates, 963 (69%) were ≤1000 g at the time of birth, and 182 (13%) died. From 1999 to 2009, the overall incidence decreased significantly for CLABSIs due to Candida spp. (0.92 vs 0.2), C albicans (0.53 vs 0.09), and non-albicans Candida spp. (0.39 vs 0.1). Birth weight-specific incidence significantly decreased across all birth weight categories for C albicans. For CLABSIs due to non-albicans Candida spp., significant decreases were detected among all birth weight categories, except among neonates 1501 to 2500 g. The proportion of Candida spp. CLABSIs due to C albicans did not significantly change over time, remaining at ∼50%.
CONCLUSIONS: Incidence of Candida spp. CLABSIs decreased substantially among NICU patients, regardless of birth weight. Decreases in incidence across all birth weight categories, and not only among neonates ≤1000 g in whom antifungal prophylaxis may be more common, suggest that multiple factors contributed to the declining incidence.
- central line-associated bloodstream infections
- health care-associated infection
- BSI —
- bloodstream infection
- CDC —
- Centers for Disease Control and Prevention
- CI —
- confidence interval
- CLABSI —
- central line-associated bloodstream infection
- HAI —
- health care-associated infection
- ICI —
- invasive Candida spp. infection
- IP —
- infection preventionist
- NHSN —
- National Healthcare Safety Network
- NNIS —
- National Nosocomial Infections Surveillance
What’s Known on This Subject:
Emphasis on preventing central line-associated bloodstream infections (CLABSIs) in US health care facilities and prophylactic antifungal medication use in neonates may impact incidence of Candida spp. CLABSIs. However, data on trends in incidence of neonatal Candida spp. CLABSIs are lacking.
What This Study Adds:
Data from a large sample of US NICUs was analyzed to assess trends in incidence over time. This analysis provides a description of the epidemiology of Candida spp. CLABSIs in a national health care-associated infections surveillance system.
Central line-associated bloodstream infections (CLABSIs) are the most common health care-associated infections (HAIs) in NICUs.1,2Candida spp. are the third most common pathogens associated with CLABSIs in NICUs.3 States4,5 and hospitals6,7 that adopted central line insertion and maintenance bundles have noted decreases in the overall CLABSI rate. Also, use of prophylactic fluconazole among neonates ≤1000 g has been shown to decrease occurrence of invasive Candida spp. infections (ICIs) in randomized trials.8–12 In the past decade, increased attention to prevention of CLABSIs and use of prophylactic antifungal medications among neonates ≤1000 g in NICUs may have led to decreases in Candida spp. CLABSIs. However, few data exist regarding trends in incidence of Candida spp. CLABSIs in NICUs by using surveillance data from health care facilities throughout the United States.
We analyzed data reported to the Centers for Disease Control and Prevention (CDC) to examine trends in incidence of Candida spp. CLABSIs among US NICU patients during 1999–2009.
From 1970 to 2004, CDC’s National Nosocomial Infections Surveillance (NNIS) was the most established and largest hospital-based surveillance system for monitoring HAIs. NNIS was succeeded by CDC’s National Healthcare Safety Network (NHSN) during 2005. Hospitals started reporting to NHSN in 2006. Since 2007 there has been a major influx of hospitals reporting to NHSN because of state-based mandates for public reporting of hospitals’ CLABSI rates. We analyzed data reported to the high-risk nursery component of NNIS (1999–2004) and to the Device-Associated Module of NHSN (2006–2009). Surveillance data are not available for 2005, because 2005 was the transition year from NNIS to NHSN.
Both NNIS and NHSN surveillance processes have been described elsewhere.13,14 In brief, infection preventionists (IPs) in participating hospitals select the patient care locations they want or are required to monitor based on state mandates, and they perform surveillance for ≥1 calendar month in each location. Both NNIS and NHSN use the American Academy of Pediatrics’ definitions for level of neonatal care,15 and both surveillance systems have standard HAI definitions.16,17 The major difference between NNIS and NHSN is related to the type of HAIs hospitals are required to monitor. In NNIS, hospitals reporting NICU data had to prospectively monitor all neonates in level II/III or III NICUs for all HAIs, whereas NHSN hospitals that are not required under state mandates to report to NHSN can select the type of HAI they want to monitor. For each month of surveillance, IPs collect data regarding number and type of HAIs, admissions, patient-days, and device-days for all neonates in 4 birth weight categories (≤1000 g, 1001–1500 g, 1501–2500 g, >2500 g) in NNIS and 5 birth weight categories (≤750 g, 751–1000 g, 1001–1500 g, 1501–2500 g, >2500 g) in NHSN.
Standard NNIS and NHSN CLABSI definitions were used.16,17 In NNIS, a central line was defined as a vascular device access that terminates at or close to the heart in 1 of the great vessels; an umbilical artery or vein catheter was also considered a central line. However, in NHSN, umbilical artery or vein catheters are separated from other types of central lines, and CLABSI rates are calculated separately for patients with umbilical catheters and for patients with other types of central lines. Both NNIS and NHSN define CLABSI as a laboratory-confirmed bloodstream infection (BSI) that is not secondary to an infection at another body site in a patient who had a central line or umbilical catheter at time of onset or within 48 hours before BSI onset. There is no minimum period of time that the central line must be in place for the BSI to be considered central line associated.
Patients with a laboratory-confirmed BSI who had a central line and an umbilical catheter in place at BSI onset would be reported as having a CLABSI in NNIS, whereas in NHSN that same patient would be reported as having an umbilical catheter-associated BSI. In NNIS, a neonate with a central line or umbilical catheter would be reported as contributing 1 central line-day, and in NHSN this same patient would be counted as contributing 1 umbilical catheter-day. All laboratory-confirmed Candida spp. BSIs that met the CLABSI definition, including an umbilical catheter-associated BSI, were included in analyses. Candida spp. CLABSIs were classified as Candida albicans or non-albicans Candida spp. The latter category included all CLABSIs reported as being due to any non-albicans Candida spp., and all CLABSIs reported as being due to Candida spp. not otherwise specified. Polymicrobial Candida spp. CLABSIs (ie, C albicans and non-albicans Candida spp.) were included in the trend analysis of both C albicans and non-albicans Candida spp., but only included once when trend analysis was performed for overall Candida spp. CLABSIs. Because of the variability in ability and accuracy of hospital-based laboratories and IP staff to correctly report species identification of non-albicans Candida spp., we limited our comparisons of species-specific incidence to C albicans and non-albicans Candida spp.
For the purposes of this analysis, and to ensure uniform definitions were used across NNIS and NHSN, we combined key variable and variable categories in NHSN. To ensure comparability of data between NNIS and NHSN, we included all CLABSIs that were associated with central lines and an umbilical artery or vein catheter in NHSN; the 5 NHSN birth weight categories were classified into the 4 NNIS birth weight categories; and level II/III and level III NICUs were combined as in NNIS data for these levels of NICUs were reported in aggregate.
Data regarding number of Candida spp., C albicans, and non-albicans Candida spp. CLABSIs; total number of patients; central line-days; and deaths were pooled by birth weight category and calendar year for each NICU. Overall annual NICU-specific Candida spp., C albicans, and non-albicans Candida spp. CLABSI rates were determined by pooling all data by year within each NICU across the analysis period. Annual birth weight-specific C albicans and non-albicans Candida spp. CLABSI rates were determined by pooling all data by birth weight category within each NICU across the analysis period. Central line-days, including umbilical catheter-days, were used as the denominator for all incidence measures. The proportion of Candida spp. CLABSIs due to C albicans during each year of the analysis was calculated by using the yearly total for Candida spp. CLABSIs as a denominator. Overall annual NICU specific central line utilization (defined as the ratio of central line-days to patient-days) was determined by pooling all data by year within each NICU across the analysis period.
Poisson regression was used to assess trends in the annual overall and birth weight-specific CLABSI incidence per 1000 central line-days. The Mood median test was used to compare the overall median CLABSI incidence per 1000 central line-days in 1999 versus 2009, and the Kuiper 2-sample test was used to compare the distribution of CLABSI rates. Sensitivity analysis was conducted to confirm observed trends in incidence of CLABSIs detected in the Poisson regression model and to control for sample migration (ie, NICUs coming in and out of the system during the analysis period). The sensitivity analysis was conducted by restricting incidence trend analysis to NICUs reporting ≥1 month of data in each of the 10 years and ≥50 central line-days in each year of the analysis period. Weighted linear regression was used to model changes in annual proportion of C albicans CLABSIs and in central line utilization ratio. Proportion of C albicans was adjusted for Candida spp. CLABSIs reported, and central line utilization was adjusted for patient-days. All reported P values are 2-sided, and a P value of <.05 was considered statistically significant. All analysis was conducted by using SAS version 9.2 (SAS Institute, Inc, Cary, NC).
Description of NNIS and NHSN Population
From January 1, 1999 to December 31, 2009, 398 NICUs reported 1407 Candida spp. CLABSIs among 9 494 886 patient-days and 2 941 522 central line-days. The median age of disease onset was 22 days (range, 0–330 days), and deaths due to all causes were reported among 182 (13%) of 1400 patients with Candida spp. CLABSIs. Among the 4 birth weight categories, neonates ≤1000 g had the lowest median age of disease onset, the highest number of patient-days and central line-days, and the highest proportion of deaths (Table 1).
Over the analysis period, NICUs submitted a median of 24 months of data (range, 1–120 months). Of the 398 NICUs, 358 (90%) were located in general hospitals, 29 (7%) in children’s hospitals, and 11 (3%) in other types of hospitals.
Pathogen Distribution of Candida spp. CLABSIs
Of the 1407 Candida spp. CLABSIs, 1400 (99.5%) CLABSIs were monomicrobial and 7 (0.5%) CLABSIs were polymicrobial. Of the Candida spp. CLABSIs, 706 (50%) and 701 (50%) were identified as Candida albicans and non-albicans Candida spp., respectively. Of the 701 non-albicans Candida spp. CLABSIs, 486 (69%), 90 (13%), 40 (6%), 35 (5%), and 30 (4%) were identified as Candida parapsilosis, Candida spp. not otherwise specified, Candida tropicalis, Candida lusitaniae, and Candida glabrata, respectively.
Trends in Overall and Birth Weight-Specific CLABSI Incidence
The overall incidence of Candida spp. CLABSIs among all birth weight categories combined decreased 75% (95% confidence interval [CI]: −78% to −71%) from 1999 through 2009; the pooled mean rate of Candida spp. CLABSIs per 1000 central line-days decreased from 0.9 in 1999 to 0.2 in 2009 (P < .001), and the median rate decreased from 4.5 in 1999 to 0.0 in 2009 (P < .001).
Overall incidence for C albicans and non-albicans Candida spp. CLABSIs among all birth weight categories combined decreased 79% (95% CI: −83% to −74%) and 70% (95% CI: −76% to −63%), respectively, during the analysis period. From 1999 to 2009, the pooled mean rate of CLABSIs decreased for C albicans (0.5/1000 vs 0.1/1000; P < .001) and for non-albicans Candida spp. (0.4/1000 vs 0.1/1000; P < .001). Median rate of C albicans CLABSIs per 1000 central line-days also decreased from 0.51 in 1999 to 0.0 in 2009 (P < .001). Although median incidence rate of non-albicans Candida spp. CLABSIs was 0.0/1000 in 1999 (interquartile range, 0.0–4.0) and 2009 (interquartile range, 0.0–0.0), the distribution of these rates differed (P = .04). To assess if decreases in incidence were similar across birth weight categories, birth weight-specific incidence trend analyses were conducted for both C albicans and non-albicans Candida spp. CLABSIs.
Birth weight-specific rates of C albicans CLABSIs significantly decreased among each of the 4 birth weight categories (Fig 1). The greatest decreases in pooled mean rates of C albicans CLABSIs were detected among neonates ≤1000 g (−83%; 95% CI: −87% to −77%), 1001 to 1500 g (−81%; 95% CI: −90% to −65%), and 1501 to 2500 g (−73%; 95% CI: −85% to −51%) compared with neonates >2500 g (−49%; 95% CI: −73% to −5%). For non-albicans Candida spp. CLABSIs, significant decreases were detected for all birth weight categories except among neonates 1501 to 2500 g (Fig 2). The greatest decreases in pooled mean rates of non-albicans Candida spp. CLABSIs were detected among neonates ≤1000 g (−72%; 95% CI: −78% to −64%), 1001 to 1500 g (−75%; 95% CI: −87% to −55%), and >2500 g (−77%; 95% CI: −89% to −51%) compared with neonates 1501 to 2500 g (−5%; 95% CI: −56% to 113%).
Sensitivity analysis restricted to 44 NICUs confirmed observed trends in both C albicans and non-albicans Candida spp. CLABSIs among all birth weight categories with the exception of birth weight category >2500 g, where sensitivity analysis did not show a significant decrease in C albicans (7%; 95% CI: −62% to 200%) or non-albicans Candida spp. (−62%; 95% CI: −89% to 26%) CLABSIs.
Trends in Yearly Proportion of C albicans CLABSIs
The overall proportion of Candida spp. CLABSIs identified as C albicans was 0.58 in 1999 and 0.48 in 2009 (Fig 3); however, in weighted linear regression, there were no significant changes (P = .17) in this yearly proportion. Similar findings were obtained when each birth weight category was analyzed separately with the exception of birth weight category 1501 to 2500 g where a significant decrease in the proportion of C albicans CLABSI was observed (P = .03).
Trends in Central Line Utilization
During 1999–2009, statistically significant decreases in usage of central lines in NICU patients were detected only among neonates 1501 to 2500 g and >2500 g at birth. However, these decreases were modest. Central line utilization ratio for birth weight categories 1501 to 2500 g and >2500 g went from 0.23 and 0.34 in 1999 to 0.19 and 0.24 in 2009, respectively, representing an annual decrease in central line utilization of 0.4% for neonates 1501 to 2500 g and of 1.0% for neonates >2500 g (Fig 4).
From 1999 to 2009, Candida spp. CLABSI incidence decreased 75% among neonates in US NICUs as a result of decreases in both C albicans and non-albicans Candida spp. CLABSIs. These decreases in C albicans and non-albicans Candida spp. CLABSIs were observed among neonates in most birth weight categories, but especially among neonates weighing ≤1000 g and 1001 to 1500 g at birth, for whom central line utilization is high. The observed decreases across most birth weight categories, and not only among neonates ≤1000 g in whom anti-fungal prophylaxis may be more common, suggest that multiple factors may have contributed to the declining incidence. In addition, our data suggest that, if NICUs included in our analyses used antifungal prophylaxis during 1999–2009, such use did not result in a significant shift in the proportion of Candida spp. CLABSIs because of less susceptible non-albicans Candida spp.
The decreases in Candida spp. CLABSI incidence are unlikely to be influenced by the large influx of NICUs in NHSN after 2007 because of state-based mandates of CLABSI reporting or by changes in central line utilization for several reasons. First, reductions in incidence have been detected since 2001, before implementation of state-based mandates. Second, sensitivity analysis including only NICUs reporting data in every year of the analysis period confirmed the observed declines in Candida spp. CLABSI incidence. Finally, only small changes (≤1% per year) in central line utilization ratios were detected across all birth weight categories; therefore, reductions in central line use among neonates are unlikely to explain the significant decreases in Candida spp. CLABSIs detected. In addition, the decreases also are unlikely to reflect a reduction in CLABSIs solely among NICUs with initially high rates, because the median rate of Candida spp. CLABSIs significantly decreased during the analysis period.
Improvements in central line insertion and maintenance practices in NICUs during the analysis period likely contributed to rate reductions of Candida spp. CLABSIs. NICUs in New York,4 California,5 and in single academic institutions6,7 that participated in infection prevention initiatives targeted toward improving adherence to central line insertion and maintenance bundles noted a 25% to 85% decrease in overall CLABSI incidence. Although findings from these states and academic institutions were not limited to assessing changes in rates of specific pathogens (eg, Candida spp. or Staphylococcus aureus), they confirm findings from similar efforts in pediatric and adult ICUs.18 In these US non-NICU settings there was an estimated 46% reduction in Candida spp. CLABSI incidence in 2009 in comparison with 2001.18 These reductions have, in part, been attributed to regional and state-based CLABSI prevention programs.18 Therefore, it is possible that similar decreases in Candida CLABSIs may have occurred in NICUs as result of prevention collaboratives targeting reductions in CLABSIs.
Increased use of prophylactic fluconazole in NICUs during the analysis period also may have contributed to rate reductions in Candida spp. CLABSIs. Prophylactic fluconazole has been shown to be effective for preventing ICIs among neonates ≤1000 g in multiple randomized trials8–12 and retrospective studies19–28 conducted in NICUs that generally had high rates of ICI. A survey of neonatologists conducted in 2004, before publication of more robust multicenter studies of fluconazole prophylaxis in NICUs, revealed that only 34% of respondents used prophylactic antifungal medications and the antifungal agent routinely used varied.29 However, antifungal prophylaxis practices in neonates may have changed in more recent years. The updated guidelines on clinical management of candidiasis published in 2009 recommends considering routine fluconazole prophylaxis in premature neonates whose birth weight is ≤1000 g in NICUs with high rates of ICI.30 In addition, some experts in neonatology have recommended extending the use of fluconazole prophylaxis to any neonate ≤27 weeks’ gestational age who requires intravenous access with a central or peripheral line.31 In light of these recent recommendations, use of antifungal prophylaxis in US NICU patients needs to be further evaluated.
Although reductions in rates of Candida spp. CLABSIs were detected in US NICUs during 1999–2009, the same has not been observed for S aureus CLABSIs where rates of methicillin-resistant S aureus CLABSIs increased 49% and methicillin-susceptible S aureus CLABSIs remained stable over this time period.32 This suggests that not all CLABSIs are going down in US NICUs and that some interventions specifically for Candida spp. infections may have played a more important role in the decreases we observed. Nevertheless, additional studies are needed to identify specific interventions that contributed to these rate reductions. The results of these studies can help define overall and pathogen-specific prevention strategies for CLABSIs.
As a result of the significant decreases detected in rates of C albicans and non-albicans Candida spp. CLABSIs across all 4 birth weight categories, the 2009 rates of Candida spp. CLABSIs are currently low, varying from 0.03 to 0.16 per 1000 catheter-days, compared with 1999 when rates varied from 0.15 to 0.94 per 1000 catheter-days across birth weight categories. Whether these current low rates of Candida spp. CLABSIs among neonates in different birth weight categories can be further lowered by either central line prevention practices or current antifungal prophylaxis practices is unclear. Neonates ≤1000 g who have longer lengths of stay in NICUs and received antifungal prophylaxis earlier during their hospitalization, and those neonates >1000 g who are admitted to NICUs, but never received antifungal prophylaxis because of current antifungal prophylaxis guidelines, may be at risk for translocation of Candida spp. from the gastrointestinal tract to the blood, which would not be preventable by central line insertion and maintenance practices. Nevertheless, the proportion of Candida spp. CLABSI that are not preventable by current practices has yet to be determined.
There are some limitations to this analysis. First, NICUs participating in NNIS and NHSN may not be representative of all US NICUs; therefore, our findings may not be generalized to all NICUs. Second, NNIS and NHSN surveillance data do not represent a consistent cohort of facilities followed over the analysis period. There was a large influx of NICUs in NHSN after 2007 because of state-based mandatory reporting of CLABSI. However, the decreases detected in Candida spp. CLABSIs when we limited the analysis to NICUs that participated in surveillance for all 10 years of the analysis period (ie, continuous reporters) confirmed downward trends in both C albicans and non-albicans Candida spp. Third, there may have been potential misclassification when categorizing non-albicans Candida spp., because Candida spp. not otherwise specified were included in this category. Because 90 of 701 C albicans spp. CLABSIs were reported as Candida spp. not otherwise specified, it is unlikely that this misclassification substantially affected the incidence trend analysis. Fourth, we were unable to assess trends in C glabrata CLABSI incidence during the analysis period because of issues related to pathogen selection in the NHSN application. During the transition from NNIS to NHSN, users would have needed to select “Torulopsis glabrata” from a large dropdown pathogen menu to report a CLABSI due to C glabrata. Because NHSN users may not have been familiar with this older terminology for C glabrata, it is likely that many C glabrata CLABSIs were reported as being due to “Candida species not otherwise specified.” In 2010, “C glabrata” was added to the pathogen menu in the application. Additional analyses using NHSN data starting in 2010 will be helpful to evaluate trends in C glabrata CLABSIs in US NICUs. Nevertheless, the fact that overall non-albicans Candida spp. incidence decreased from 1999 through 2009 during a period of increased use of antifungal prophylaxis in US NICUs is encouraging. Despite these limitations, the data analyzed represent the largest number of NICU patients evaluated systematically for Candida spp. CLABSIs in the United States, providing a better national picture of disease trends and disease burden in comparison with previous studies.
We observed a significant decrease in Candida spp. CLABSIs among all birth weight categories combined in US NICUs from 1999 through 2009. The decrease did not result in a shift in proportion of CLABSIs due to C albicans versus non-albicans Candida spp. Understanding the reasons for the decreases in rates of Candida spp. CLABSIs may provide important information to guide future prevention efforts in NICUs.
- Accepted March 8, 2012.
- Address correspondence to Amit S. Chitnis, MD, MPH, 1600 Clifton Rd, MS A-24, Atlanta, GA 30333. E-mail:
All authors made substantial contributions to the following areas of this study: (1) conception and design, acquisition of data, or analysis and interpretation of data; (2) drafting the article or revising it critically for important intellectual content; and (3) final approval of the version of the manuscript to be published.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
- ↵Dudeck MA, Horan TC, Peterson KD, et al. National Healthcare Safety Network (NHSN) report, data summary for 2009, device-associated module. Available at: www.cdc.gov/nhsn/PDFs/dataStat/2010NHSNReport.pdf. Accessed December 5, 2011
- ↵Hocevar S, Horan T, Morrell G, Iwamato M, Edwards J, Lessa F. Rates and pathogen profile of device-associated infections among neonatal intensive care unit patients: data from the National Healthcare Safety Network, 2006–2008. In: Proceedings from The Society for Healthcare Epidemiology of America Fifth Decennial Meeting; March 18–22, 2010; Atlanta, GA
- Schulman J,
- Stricof R,
- Stevens TP,
- et al.,
- New York State Regional Perinatal Care Centers
- Bizzarro MJ,
- Sabo B,
- Noonan M,
- Bonfiglio MP,
- Northrup V,
- Diefenbach K,
- Central Venous Catheter Initiative Committee
- Aly H,
- Herson V,
- Duncan A,
- et al
- Aydemir C,
- Oguz SS,
- Dizdar EA,
- et al
- ↵Centers for Disease Control and Prevention. NHSN patient safety manual. Available at: www.cdc.gov/nhsn/TOC_PSCManual.html. Accessed December 5, 2011
- Stark AR,
- American Academy of Pediatrics Committee on Fetus and Newborn
- Manzoni P,
- Arisio R,
- Mostert M,
- et al
- Uko S,
- Soghier LM,
- Vega M,
- et al
- McCrossan BA,
- McHenry E,
- O’Neill F,
- Ong G,
- Sweet DG
- Healy CM,
- Campbell JR,
- Zaccaria E,
- Baker CJ
- Burwell LA,
- Kaufman D,
- Blakely J,
- Stoll BJ,
- Fridkin SK
- Pappas PG,
- Kauffman CA,
- Andes D,
- et al.,
- Infectious Diseases Society of America
- ↵Lessa F, Edwards JR, Horan TC, Kupronis B, Fridkin SK. Trends in incidence of methicillin-resistance Staphylococcus aureus central line-associated bloodstream infections in U.S. neonatal intensive care units, 1998–2008. In: Proceedings of The Society for Healthcare Epidemiology of America Fifth Decennial Meeting; March 18–22, 2010; Atlanta, GA
- Copyright © 2012 by the American Academy of Pediatrics