Surveillance Cultures for Antimicrobial Resistance in the NICU
Surveillance Cultures for Antimicrobial Resistance in the NICU
During the study period, 84% (1751/2081) of transferred infants were eligible for inclusion in this study; 79% (1182/1497) of infants transferred to NICU 1 and 97% (569/584) of infants transferred to NICU 2 were included. Eligible infants who did not have surveillance cultures performed had a higher risk of crude mortality, a higher mean birth weight and gestational age and a lower mean age at admission than infants who had surveillance cultures performed, when adjusted for year of study (data not shown).
The site-specific characteristics of eligible infants are shown in Table 1. Infants transferred to NICU 2 had a lower mean gestational age, were younger at admission and more likely to have the admitting diagnosis of a neurologic condition than infants transferred to NICU 1. In contrast, infants transferred to NICU 1 were more likely to have GI disease, congenital heart disease or ROP. NICU 1 had a significantly higher crude mortality rate of 7.2% (85/1169) compared with NICU 2's rate of 3.5% (20/569), P = 0.002.
Of the 1751 eligible infants, 83% (n = 1725), 75% (n = 1566) and 71% (n = 1475) were cultured for MRSA, VRE and AR-GNR, respectively. The rate of obtaining surveillance cultures increased during the study period. In 2004, approximately 60% of infants were cultured for MRSA and VRE and approximately 30% were cultured for AR-GNR. From 2008 to 2010, greater than 80% of eligible infants were cultured for all 3 types of AROs.
Among those with surveillance cultures performed, 5.2% (91/1751) were colonized with 1 or more AROs; 3% (52/1725) of infants were colonized with MRSA, 1.7% (27/1566) with VRE and 1% (15/1475) with AR-GNR. Only 3 infants were colonized with 2 AROs; 1 was colonized with MRSA and AR-GNR and 2 were colonized with MRSA and VRE. The yield of surveillance cultures for all AROs combined as well as for individual AROs were similar in the 2 study NICUs (Table 1).
The proportion of infants with positive surveillance cultures increased with age at the time of transfer (Fig. 1). The overall yield of surveillance cultures for infants transferred on day of life 1–3, 1–6 and ≥7 was 1.4% (n = 11/804), 1.2% (n = 12/975) and 10.2% (79/776), respectively. AR-GNRs were detected from 1 infant transferred on day of life 1 and from none transferred on day of life 2–6. MRSA was the most common ARO cultured at each day of life strata generally followed by VRE (Fig. 1). The rate of positive surveillance cultures for AROs did not change over time as shown in Figure 2. The highest yield for MRSA, VRE and AR-GNR was 2006, 2004 and 2010, respectively.
(Enlarge Image)
Figure 2.
Yield of surveillance cultures by year. The percent of positive surveillance cultures for MRSA, VRE and AR-GNRs resistant to third generation cephalosporin agents is shown from 2004 to 2010. Data from both NICUs are combined. Numeric values on each bar represent the number of infants who underwent surveillance for the respective organism.
In bivariate analysis stratified by site, infants colonized with any ARO were significantly older when transferred (odds ratio [OR] = 1.13 per increasing week of life, P < 0.0001), of younger gestational age (OR = 1.1 per decreasing week of gestation, P < 0.0001), of lower birth weight (OR = 1.35 for every 100 g decrease in birth weight, P < 0.0001) and more likely to have been previously hospitalized in the study NICUs (OR = 4.3, P < 0.0001) as shown in Table 2. The diagnosis of ROP was associated with an increased risk of ARO colonization, whereas GI disease was associated with a decreased risk of ARO colonization.
Risk factors for individual AROs were also assessed by bivariate analysis (data not shown). The same risk factors noted for colonization with any ARO remained significant for colonization with MRSA. Similar risk factors were noted for colonization with VRE, that is, older age at transfer, lower gestational age and lower birth weight, but VRE colonization was not associated with previous hospitalization in the study NICUs. Risk factors for AR-GNR included older age at transfer and previous hospitalization in the study NICUs.
Multivariable logistic regression (stratified by site) included birth weight, gestational age, age at transfer, admitting diagnosis and previous hospitalization in study NICUs. In the final model, age upon transfer remained the only significant predictor of ARO colonization (OR = 1.1 per increased week of life, P < 0.0001). At all dichotomized days of life strata, older age at transfer remained a significant predictor of colonization status (Table 3).
An unadjusted analysis of individual outcomes of infants colonized with AROs are shown in Table 4. Crude mortality was not significantly different in infants with and without ARO colonization. Infants colonized with any ARO had a decreased length of hospitalization when compared with uncolonized infants. There was no difference in the risk of ARO infection between ARO colonized and uncolonized infants.
Overall, 1.7% (27/1751) of infants subsequently had a positive clinical culture for an ARO (Table 4). Thirteen had a positive culture for MRSA, 6 for VRE and 8 for AR-GNR. Only 3 (11.1%) of these 27 infants had concordant surveillance and subsequent positive clinical cultures and all had MRSA. Three infants without surveillance cultures obtained at the time of transfer (day of life 1, 3 and 69) also had positive clinical cultures for an ARO (2 with MRSA, including an infant with epidermolysis bullosa, and 1 infant with an AR-GNR urinary tract infection).
The overall cost of surveillance cultures at both study NICUs was $58,425 (Table 5). The cost of supplies for negative versus positive cultures was $21,237 and $1258, respectively, and the cost of labor was $34,203 and $1727, respectively. If there had been 100% compliance with surveillance cultures, laboratory costs would have increased to $66,512. The estimated cost of supplies for negative versus positive cultures would increase to $23,899 and $1554, respectively, and the estimated cost of labor for negative versus positive cultures would increase to $38,890 and $2169, respectively.
Results
Study Subjects
During the study period, 84% (1751/2081) of transferred infants were eligible for inclusion in this study; 79% (1182/1497) of infants transferred to NICU 1 and 97% (569/584) of infants transferred to NICU 2 were included. Eligible infants who did not have surveillance cultures performed had a higher risk of crude mortality, a higher mean birth weight and gestational age and a lower mean age at admission than infants who had surveillance cultures performed, when adjusted for year of study (data not shown).
The site-specific characteristics of eligible infants are shown in Table 1. Infants transferred to NICU 2 had a lower mean gestational age, were younger at admission and more likely to have the admitting diagnosis of a neurologic condition than infants transferred to NICU 1. In contrast, infants transferred to NICU 1 were more likely to have GI disease, congenital heart disease or ROP. NICU 1 had a significantly higher crude mortality rate of 7.2% (85/1169) compared with NICU 2's rate of 3.5% (20/569), P = 0.002.
Yield of Surveillance Cultures
Of the 1751 eligible infants, 83% (n = 1725), 75% (n = 1566) and 71% (n = 1475) were cultured for MRSA, VRE and AR-GNR, respectively. The rate of obtaining surveillance cultures increased during the study period. In 2004, approximately 60% of infants were cultured for MRSA and VRE and approximately 30% were cultured for AR-GNR. From 2008 to 2010, greater than 80% of eligible infants were cultured for all 3 types of AROs.
Among those with surveillance cultures performed, 5.2% (91/1751) were colonized with 1 or more AROs; 3% (52/1725) of infants were colonized with MRSA, 1.7% (27/1566) with VRE and 1% (15/1475) with AR-GNR. Only 3 infants were colonized with 2 AROs; 1 was colonized with MRSA and AR-GNR and 2 were colonized with MRSA and VRE. The yield of surveillance cultures for all AROs combined as well as for individual AROs were similar in the 2 study NICUs (Table 1).
The proportion of infants with positive surveillance cultures increased with age at the time of transfer (Fig. 1). The overall yield of surveillance cultures for infants transferred on day of life 1–3, 1–6 and ≥7 was 1.4% (n = 11/804), 1.2% (n = 12/975) and 10.2% (79/776), respectively. AR-GNRs were detected from 1 infant transferred on day of life 1 and from none transferred on day of life 2–6. MRSA was the most common ARO cultured at each day of life strata generally followed by VRE (Fig. 1). The rate of positive surveillance cultures for AROs did not change over time as shown in Figure 2. The highest yield for MRSA, VRE and AR-GNR was 2006, 2004 and 2010, respectively.
(Enlarge Image)
Figure 2.
Yield of surveillance cultures by year. The percent of positive surveillance cultures for MRSA, VRE and AR-GNRs resistant to third generation cephalosporin agents is shown from 2004 to 2010. Data from both NICUs are combined. Numeric values on each bar represent the number of infants who underwent surveillance for the respective organism.
Risk Factors for Colonization With AROs
In bivariate analysis stratified by site, infants colonized with any ARO were significantly older when transferred (odds ratio [OR] = 1.13 per increasing week of life, P < 0.0001), of younger gestational age (OR = 1.1 per decreasing week of gestation, P < 0.0001), of lower birth weight (OR = 1.35 for every 100 g decrease in birth weight, P < 0.0001) and more likely to have been previously hospitalized in the study NICUs (OR = 4.3, P < 0.0001) as shown in Table 2. The diagnosis of ROP was associated with an increased risk of ARO colonization, whereas GI disease was associated with a decreased risk of ARO colonization.
Risk factors for individual AROs were also assessed by bivariate analysis (data not shown). The same risk factors noted for colonization with any ARO remained significant for colonization with MRSA. Similar risk factors were noted for colonization with VRE, that is, older age at transfer, lower gestational age and lower birth weight, but VRE colonization was not associated with previous hospitalization in the study NICUs. Risk factors for AR-GNR included older age at transfer and previous hospitalization in the study NICUs.
Multivariable logistic regression (stratified by site) included birth weight, gestational age, age at transfer, admitting diagnosis and previous hospitalization in study NICUs. In the final model, age upon transfer remained the only significant predictor of ARO colonization (OR = 1.1 per increased week of life, P < 0.0001). At all dichotomized days of life strata, older age at transfer remained a significant predictor of colonization status (Table 3).
Outcomes for Infants With ARO Colonization
An unadjusted analysis of individual outcomes of infants colonized with AROs are shown in Table 4. Crude mortality was not significantly different in infants with and without ARO colonization. Infants colonized with any ARO had a decreased length of hospitalization when compared with uncolonized infants. There was no difference in the risk of ARO infection between ARO colonized and uncolonized infants.
Overall, 1.7% (27/1751) of infants subsequently had a positive clinical culture for an ARO (Table 4). Thirteen had a positive culture for MRSA, 6 for VRE and 8 for AR-GNR. Only 3 (11.1%) of these 27 infants had concordant surveillance and subsequent positive clinical cultures and all had MRSA. Three infants without surveillance cultures obtained at the time of transfer (day of life 1, 3 and 69) also had positive clinical cultures for an ARO (2 with MRSA, including an infant with epidermolysis bullosa, and 1 infant with an AR-GNR urinary tract infection).
Laboratory Costs of Surveillance Cultures
The overall cost of surveillance cultures at both study NICUs was $58,425 (Table 5). The cost of supplies for negative versus positive cultures was $21,237 and $1258, respectively, and the cost of labor was $34,203 and $1727, respectively. If there had been 100% compliance with surveillance cultures, laboratory costs would have increased to $66,512. The estimated cost of supplies for negative versus positive cultures would increase to $23,899 and $1554, respectively, and the estimated cost of labor for negative versus positive cultures would increase to $38,890 and $2169, respectively.
Source...