Effect of an expanded physical on nosocomial infections in a neonatal intensive care unit
Elaine Larson, R.N., Ph.D., F.A.A.N. Clarice 0. Hargiss, R.N., M.S. Lorna Dyk, R.N., B.S.N. Philadelphia,
of a threefold
space per infant in a neonatal intensive care unit on rates of nosocomial infections (NIs) and colonization with Staphylococcus ~UEU.S (39 months in a crowded l&bed unit and 13 months in a spacious 32-bed unit). Mean length of stay, survival rates, mean birth weights, and other parameters indicated that infant populations in the old and new units were similar. NI rates were not significantly different in the old and new units (11.7% and 9.6%, respectively; p = 0.17) nor were rates of colonization of anterior nares with S. uwzus (11.7% and 10.7%; p = 0.5). NI rates, but not S. atirez4.s colonization rates, were significantly higher during months of high patient turnover (p < 0.01). Sites of infection were similar in the old and new units. There was, however, a significant change in bacterial species causing NI. Klebsiella pneumoniae and Pseudomonas aeruginosa caused 20.4% of NIs in the old unit, but only 2.1% in the new unit (p < 0.001) and NIs caused by S. epidemidis increased from 4.7% to 14.9% (p = 0.02) in the new unit. There was also a marked decrease in the numbers of clusters of NI occurring in the new unit, indicating that cross-infections between infants were probably minimized. (AM J INFECT CONTROL 13:16-20,
In March 1980 the neonatal intensive care unit (NICU) at University Hospital, Seattle, moved from a crowded 18-bed unit to a spacious 32-bed unit. Conditions in the old unit were extremely cramped. Equipment and even babies in Isolettes were sometimes of necessity located in the hallways. There were six small cubicles that housed as many as four babies and the equipment such as ventilators, cardiac monitors, and phototherapy lamps that was required for their care. Obviously in such situFrom School
University Hospital of Medicine.
Presented at the ence, Washington, Reprint NEEY52, delphia,
Eleventh Annual APIC D.C., June 1984.
requests: Elaine Larson, University of Pennsylvania PA 19104.
ations there is concern about the high risk of cross-infection and nosocomial outbreaks. Such outbreaks are common in nurseries, especially among infants at high risk such as premature babies and those undergoing invasive monitoring.‘, ’ When the move to a new unit transpired, it was assumed that rates of nosocomial infections (NIs), given a stable patient population, would decrease. The purpose of this study was to determine the influence of the spacious new physical environment on NI rates in the NICU.
R.N., Ph.D., F.A.A.N., School of Nursing,
The NICU at this 360-bed university-affiliated tertiary care facility serves as a referral center for critically ill infants throughout the Pacific Northwest. Space in the old unit aver-
Volume 13 Number February, 1985
aged about 30 square feet per infant and in the new unit, 100 square feet per infant, a threefold increase. In addition, two isolation rooms were available in the new unit. In both units, average occupancy rates were 95% to 100%. The average length of stay in the old unit during 1977 was 21 days for survivors and 9 days for babies who died, and in the new unit, 21 and 6 days, respectively. When the unit moved, the nursing and medical staff remained the same, although additional positions were filled. Nurse : patient ratio remained the same, 1: 1 or 1: 2. Except for the architectural change there were no major changes in clinical practice or patient management between the old and the new units during the time periods studied. The nursing staff, however, has taken increasing responsibility for many procedures (e.g., drawing arterial blood for determination of blood gas) that were previously physician tasks, and nursing practice has become more standardized by means of written protocols and procedures. Rates and types of NIs in the NICU were determined by the hospital’s nurse epidemiologist for approximately 10 years, ending in April 1981, as part of the total surveillance conducted in collaboration with the National Nosocomial Infections Study (NNIS) sponsored by the Centers for Disease Control (CDC). Five days each week, all microbiology reports were screened and patient rounds were conducted by this individual. For this study, 52 months’ data were used: 39 months in the old NICU (January 1977 to March 1980) and 13 months (April 1980 to April 1981) in the new unit. All of these data were collected by the same individual using CDC criteria for determining the presence of a NI.” In addition to total surveillance of NI, the colonization rates of the anterior nares and umbilicus of neonates with Staphylococcus am-em were monitored for 23 months in the old unit and 23 months in the new unit. Cultures were obtained on approximately 75% of babies once, immediately prior to hospital discharge. The sites were swabbbed; both swabs were placed in a single tube of Stuart’s transport medium,4 streaked on blood agar plates, and incubated at 37” _t 2” C in a CO, incubator. Sampling and
Table 1. Comparison
an expanded physical facility of NICU patients
old and new units Characteristics patients
Mean length of stay Survivors Nonsurvivors Range of birth weight for most patients Birth weight 51500 gm With respiratory assistance With hyaline membrane disease
21 days 9 days 1751-2000 gm
21 days 6 days 1751-2000 gm
microbiologic techniques have been previously described.5> 6 Chi-square tests for differences in proportions were used for statistical analyses.’ RESULTS In the 39 months during which the old unit was studied, there were 1443 patient discharges and 169 NIs in 142 patients (overall NI rate of 11.7%). During 13 months in the new unit there were 502 patient discharges and 48 NIs in 44 babies (9.6% NI rate) (p = 0.17). Characteristics of babies admitted to the old and new units were similar (Table 1). Rates of colonization of the anterior nares or umbilicus with S. aureus in the old unit and the new unit were similar (11.7% and 10.7%, respectively; p = 0.5). For the entire 52 months of the study there were no significant seasonal variations in NI rates (p = 0.6). Antimicrobial susceptibility patterns of strains of S. aureus isolated remained the same. Rates of NI were not affected by the number of babies admitted to the unit with infections (transfers from other hospitals), but there were significant differences in NI rates between times of high and low patient discharges. Although the number of discharges does not necessarily reflect average daily census, it does indicate busy times for the staff, since patient admissions and discharges require a considerable amount of work. NI rates were significantly higher when there were more than 45 discharges per month (12.1%) than when there
Journal of CONTROL
Old unit (n= 165)
New unit (n= 47)
t! $ 12
n” 12 8
Fig. 1. Sites of Nls in neonates significant changes are seen units.
New unit (n= 47)
GNB= Gram-Nepolive Bacteria GPB =Gmm-Positive Bacteria
in ND-J. No statistically between old and new
were less than 45 discharges per month (6.6%) (p < 0.01). There were no significant differences in colonization rates with S. aureus between low and high discharge months (9.7% and 10.8%, respectively; p = 0.75). There was a decrease in peritoneal infections, primarily necrotizing enterocolitis (NEC), from 28.5% of the total NIs in the old unit to 19.1% in the new unit (p = 0.15). There was an increase in conjunctivitis from 12.1% of total NIs in the old unit to 23.4% in the new unit (p = 0.06). Other sites of NIs were similar in both units (Fig. 1). Overall, the most common organisms causing NIs were S. aureus (22.5%) and Escherichia coli (17.9%). Klebsiella pneumoniae and Pseudomonas aeruginosa caused 20.4% of NIs in the old unit, but only 2.1% in the new unit (p < 0.001). There was also an increase in NI caused by S. epidermidis in the new unit, from 4.7% to 14.9% (p = 0.02) (Fig. 2). In the old unit there were three clusters of two or more premature infants who developed NEC associated with K. pneumoniae resistant to ampicillin and carbenicillin while their stays were overlapping in the NICU. The similar antimicrobial susceptibility pattern and proximity in time indicated that these clusters could have represented cross-contamination. There were also four clusters of infections due to S. aweus involving 10 babies and one cluster of five babies who developed nosocomial enteritis, presumed to be viral in origin because no potential bacterial pathogens were isolated. There were no such clusters of NIs in the new unit (Table 2).
(8 q pe
Fig. 2. Organisms significant changes units.
causing Nls in NICU. Statistically are seen between old and new
Although Goldmann et al.’ found a significant decrease in NI rates associated with improved facilities, we found only a slight decrease. They noted that most infants weighing less than 1000 gm became infected and that lower birth weight was significantly associated with increased risk of NI.’ In our old and new units, birthweights were similar, with an average range of 1751 to 2000 gm. Survival rates of infants weighing less than 1500 gm were also similar: 66.7% in the old unit and 69.1% in the new unit. We concluded that patients in the old and new units were comparable (Table 2); hence a change in patient characteristics could not explain why NI rates did not significantly decrease. Although we evaluated changes in medical practice and patients to the best of our ability, time is always a confounding variable in studies such as this. The slight decrease in NI rates in the new unit was accounted for by a decrease in the number of clusters of NEC associated with K. pneumoniae, clusters of infections due to S. aureus, and a decrease in infections caused by P. aeruginosa. It appeared that the change to a more spacious physical environment did decrease the risk of certain cross-infections between babies. That S. aureus and E. coli were the two most common causes of NI in the NICU is consistent with the findings of others.’ Likewise, Klebsiella has also been found to be one of the most pre-
Volume 13 Number February, 1985
Table 2. Cluslers
Effect of an expanded of Nls (possible
in the old and new NlCUs
No. of oeses
NEC NEC Cofqunctivitls Pneumonia Scalp (from fetal electrode) NEC Skin Blood Surgical wound NEC Skin
pathogen isolated) 6177 7177
K. Pneumonlae K. Pneumorxae
2178 6178 7178
K. pneumoniae S. aureus S. aureus
4180 t0 4’81
S. aureus None
dominant causes of outbreaks in NICU.8-” We did not, however, find NI caused by either Serrutia marcescens or methicillin-resistant staphylococci, two other organisms causing outbreaks among neonates.“-‘” Whereas NIs due to Klebsiella and P. aevtcginosa decreased, we found an increase in those caused by S. epidermidis in the new unit. Munson et al.‘” nol:ed such infections among premature infants, especially in association with central venous pressure (85%) and total parenteral nutrition (78%) lines, as might be expected because this organism is so common on the skin. S. epidermidis may become increasingly of concern as a cause of significant clinical infections, especially among compromised hosts such as the premature infant. As evidenced in other studies, an increasing prevalence of this organism probably would have occurred whether the move to a new unit had taken place or not. We do not attribute this change to a reduction in overcrowding. On the other hand, NEC has been shown to occur in time and place clusters, possibly as a result of increased virulence of colonizing organisms. 17-‘9 Therefore, a reduction in cross-infection of infants with endemic strains of K. pneumoniat’, the organism most frequently isolated in the old unit, could explain in part why NEC was seen with less frequency in the new unit. An increase in conjunctivitis in the new unit was partially due to increased recognition of Chlamydia and one treatment failure for
2 2 2 2 4 22
prevention of gonococcal conjunctivitis. There were five cases of scalp infections in 1977 to 1978 where fetal monitoring electrodes were placed. Two were due to S. aureus and three to Peptococcus. None have occurred since 1978. Nurses in labor and delivery stations stated that at approximately that period of time there was increased emphasis placed on careful technique for various procedures because of an increase in postpartum infections. SUMMARY We did not find a significant decrease in NI rates after expanding the physical facilities of an NICU, despite a comparable patient population. We did, however, find that clusters of infections indicative of possible cross-infection occurred with less frequency. Thus, the incidence of preventable NIs probably decreased in the new unit. We attribute this to decreased crowding rather than to changes in clinical practice over a period of time. In a study of a move from an old to a new hospital, Maki et als20 concluded that the inanimate environment had little effect on NI rates. We agree with his conclusion that most organisms present in the environment originate from patients or from the flora of hospital personnel. A less-crowded environment appeared to reduce the transmission of infectious agents between patients and on the hands of hospital personnel.
Larson, Hargiss, and Dyk
We also found
a shift away from NIs caused by gram-negative bacteria, with the emergence of S. epidermidis as a major cause. It will be important to follow this trend in this and other NICUs. We gratefully acknowledge the helpful critique of David Woodrum, M.D., professor of neonatal biology and pediatrics, University of Washington School of Medicine, Seattle.
References 1. Goldmann DA, et al: Nosocomial infections in a neonatal intensive care unit. J Infect Dis 144:449, 1981. 2. Townsend TR, Wenzel RP: Nosocomial bloodstream infections in a newborn intensive care unit. Am J Epidemiol 114:73, 1981. 3. Guidelines for determining presence and classification of infection. Atlanta, 1970, Centers for Disease Control. 4. Stuart RD: Transport problems in public health bacteriology. Can J Public Health 47:114, 1956. 5. Williams CPS, Oliver TK: Nursery routines and staphylococcal colonization of the newborn. Pediatrics 44: 640, 1969. 6. Hargiss C, Larson E: The epidemiology of Staphybcoccus az4rem in a newborn nursery from 1970 through 1976. Pediatrics 61:348, 1978. 7. Armitage P: Statistical methods in medical research. Oxford, 1971, Blackwell Scientific Publications. 8. Mayhall CG, et al: Nosocomial KZebsieh infections in a neonatal unit: Identification of risk factors for gastrointestinal colonization. Infect Control 1:239, 1980. 9. Chicon MJ, et al: Nosocomial Klebsiellu infections in an intensive care nursery. South Med J 70~33, 1977.
10. Eidelman AL, Reynolds J: Gentamicin-resistant Klebsiella infections in a neonatal intensive care unit. Am J Dis Child 132:421, 1978. 11. White RD, et al: Are surveillance of resistant enteric bacilli and antimicrobial usage among neonates in a newborn intensive care unit useful? Pediatrics 68:1, 1981. 12. Scheidt A, et al: Nosocomial outbreaks of resistant Serrutia in a neonatal intensive care unit. NY State J Med 82:1188, 1982. 13. Gilbert GL, et al: Methicillin-resistant Staphylococcus aureus in neonatal nurseries. Med J Aust 1:455, 1982. 14. Primavesi R, et al: Serrutia marcescens in a special baby unit (letter). Lancet 2:1164, 1982. 15. Christensen GD, et al: Epidemic Serratia marcescens in a neonatal intensive care unit: Importance of the gastrointestinal tract as a reservoir. Infect Control 3:127, 1982. 16. Munson DP, et al: Coagulase-negative staphylococcal septicemia: Experience in a newborn intensive care unit. J Pediatr 101:602, 1982. 17. Lawrence G, et al: Pathogenesis of neonatal necrotising enterocolitis. Lancet 1:137, 1982. 18. Stevenson OK, Stevenson JK: Neonatal necrotizing enterocolitis. In Kelly VC, editor: Practice of pediatrics, Scranton, Pa, 1980, Harper & Row Publishers, Inc. 19. Kliegman RM: Neonatal necrotizing enterocolitis implications for an infectious disease. Pediatr Clin North Am 26:327, 1979. 20. Maki DG, et al: Relation of the inanimate hospital environment to endemic nosocomial infection. N Engl J Med 307:1562, 1982.
The Centers for Disease Control (CDC) of the U.S. Public Health Service, Department of Health and Human Services, is initiating a national neonatal herpes simplex virus (HSV) surveillance system. We want to identify every case of neonatal HSV infection occurring in infants born after September 30, 1983. We ask that health professionals report all cases of suspected or confirmed HSV infection that occur in infants within 30 days of life. Neonatal HSV is a devastating infection, and the CDC is attempting to describe the magnitude of the problem and to identify possible means of prevention. To report a case or for further information, call Louise Ritz or Gail Bassin Stempler collect at (301)589-6760.