Analysis of bacterial infections in a neonatal intensive care unit

Analysis of bacterial infections in a neonatal intensive care unit

Journal of Hospital Infection (1982) 3, 275-284 Analysis of bacterial infections in a neonatal intensive care unit Jacomina A. A. HoogkampKorstanje”...

709KB Sizes 0 Downloads 9 Views

Journal of Hospital Infection (1982) 3, 275-284

Analysis of bacterial infections in a neonatal intensive care unit Jacomina

A. A. HoogkampKorstanje”, B. Cats?, R. Ch. Senderst Ingrid van Ertbruggen-f


*Laboratory for Public Health, Jelsumerstraat 6,8917 EN Leeuwarden, The Netherlands twilhelmina

Children’s Hospital, Nieuwe Gracht 137, 3512 IK Utrecht, The Netherlands

Summary: An analysis was made of all cases of infection among 181 neonates admitted to the neonatal intensive care unit (NICU) during one year. Twentyfour per cent had an infection on admission; their infections correlated with prolonged ruptured membranes and the degree and site of colonization. The predominant organisms found in perinatal infections were Staphylococcus aureus, Group B streptococci and Escherichia cola’. Thirty per cent acquired a hospital infection. This correlated with the length of the period of instrumentation. The majority of the hospitalacquired infections was caused by Gram-positive cocci (micrococci, Staph. saprophyticus, Staph. aureus, forming 65 per cent of the total), E. coli and Pseudomonas aeruginosa. Most hospital infections were nosocomial and not auto-infection. The outcome of the neonates with hospital infection was good, except for those with pseudomonas infection. Acquisition of hospital infection prolonged the period of hospitalization up to twice that required for neonates without infection.


Neonates with severe infection acquired shortly after birth and very premature or immature babies are aniimportant part of those admitted to neonatal intensive care units (NICUs). Because of their compromised immunological response the latter are uniformly regarded as a high risk group for infection in the neonatal period. As intensive care for all of these babies frequently includes extensive instrumentation (e.g. prolonged intubation, indwelling catheterization or peripheral venous fluid administration), continuous surveillance for bacterial colonization and/or infection has been accepted in many centres as a part of the NICU routine. We recorded all bacteriological data and clinical data relating to intensive care treatment procedures in all neonates admitted to the NICU of the Wilhelmina Children’s Hospital during one year. The results reported in this study may serve to demonstrate the significance of bacterial infections with respect to mortality 0195-6701/82/030275

+ IO $oz.oo/o

0 1982 The Hospital Infection




and morbidity bacteriological

J. A. A. Hoogkamp-Korstanje

ef al.

of neonates admitted to our NICU and the importance of including surveillance in routine NICU procedures. Materials

and methods

All neonates admitted to the NICU of the Wilhelmina Children’s Hospital from 1 January 1977 to 1 January 1978 have been included in the study. This year was chosen out of a period of eight years surveillance, as it yielded the most complete data for evaluation. Because there is no obstetric ward in the hospital, all neonates were referred from other hospitals and were transferred mostly by a special neonatal intensive care transport system. Swabs from the nose, throat, external auditory meatus and skin were taken and cultured immediately after admission and repeated twice a week. All catheters and tubes were cultured when removed. Other specimens were cultured when indicated. All samples were cultured by routine methods and micro-organisms were identified by routine biochemical, serological and phage-typing methods. Nose, throat, ear and skin swabs were considered bacteriologically negative when the cultures revealed no growth, or growth of diphtheroids and viridans or non-groupable streptococci only. The umbilicus was regarded as a wound and was considered to be negative only when cultures revealed no growth. This last criterion was also used for the other specimens. Blood samples for haematological and biochemical analyses were taken on admission and repeated when indicated. Chest X-rays were taken on the day of admission and subsequently if indicated. Diagnostic criteria were as follows: meningitis was diagnosed on clinical symptoms, with a positive CSF culture; septicaemia was diagnosed on clinical symptoms and/or haematological abnormalities (leucocytosis, leucopenia, bands, toxic granulation of white cells, etc.) with a positive blood culture; pneumonia was diagnosed on X-ray evidence and if possible confirmed with cultures from trachea/bronchial aspirates ; urinary tract infection (UTI) was diagnosed by pyuria [more than 10 leucocytes/ microscopic field (400 x ) in a direct preparation] and on cultures of urine showing not more than one type of organism and >105 organisms/ml, or from a culture of urine obtained by bladder puncture containing >lOs organisms/ml; criteria for soft tissue infection included erythema and/or suppuration with positive cultures from the site involved. Neonates with negative cultures on admission and without symptoms of infection were considered ‘negative’, and those with positive cultures without symptoms of infection were considered to be ‘colonized’. Antibiotics were administered on admission to every neonate with one or more clinical symptoms of infection, to those born to mothers with prolonged rupture of membranes, to those with an umbilical catheter present for more than 24 h and those treated previously with antibiotics in the referring hospitals. The standard antibiotic therapy consisted of penicillin (20~000-50~000 E/kg/24 h i.v.) plus kanamycin (7.5-15 mg/kg/24 h i.v. or i.m.).


in a neonatal



care unit


pe&atal infection Incidence of infection at time of admission. One hundred and eighty-one neonates were admitted to the NICU in the study period and were investigated during the neonatal period (O-28 days post partum). One hundred and two (56 per cent) were males and 79 (44 per cent) were females and birthweights ranged from 740 to 3650 g (mean 2010 g). The most frequent reasons for admission are shown in Table I. Forty-three (24 per cent) required assisted respiration before admission Table I. Indications for admission to the neonatal intensive care unit of neonateswith negative and positive cultures

Culture Indication for admission Respiratory distress Birth asphyxia Low birth weight Congenital malformation Infection




(n 3 0 62)

@ 3 0 76)

@ 4 0 43)

0 (n =Y81)





:: 1.5



66 26

1: 40

:i 10

n = Total in group. and during transport. Sixty-two (34.3 per cent) were bacteriologically negative, 119 (65.7 per cent) were positive. Among the positives 76 (64 per cent of all positive) were colonized and 43 (36 per cent) appeared to have an infection with clinical, haematological and radiographic signs. None in the_ culture-nega&e ...-___..--_-. group&ad an infection. It was noted that about 4$ller cent of the neonates had ‘m&e than one indE&on for admission, especially in the infected group. Most culture-negative children were admitted within 12 h of birth in contrast to only half of the infected babies. The majority of the negatives were admitted because of respiratory distress from a variety of causes (e.g. hyaline membrane disease, meconium aspiration or pneumothorax). Fifty per cent of the infected neonates were referred for the same reason, 40 per cent because of symptoms of infection and 25 per cent because of congenital malformations. Gestational age and circumstances of birth in relation to colonization and infection at time of admission. No differences could be detected between colonization and perinatal infection in infants of different states of maturity (Table II). Birth after prolonged rupture of the membranes yielded a higher percentage of neonates with infection when compared to birth following membrane rupture for less than 24 h. Birth by Caesarian section resulted in a significantly lower colonization rate than after vaginal delivery, either with or without prolonged rupture of the membranes. The infection rate in the Caesarian section group was significantly lower than in the group with prolonged ruptured membranes, but did not differ significantly from the infection rate in neonates born vaginally without prolonged ruptured membranes.

J. A. A. Hoogkamp-Korstanje

278 Table II. Association



of conditions at birth with culture-negative andpositive neonates Culture

Conditions birth


Prolonged rupture of membranes ( >24 h) Caesarian section Immaturity Prematurity Full-term infants

Negative (n = 62) No. (%I

Colonized (n = 76) No. (%)

7 (ll)j+yr,

Infected (n = 43) No. (%) 10 (23)+3 .yt

6 (8)q~l

23 (37) ~];ili8)


23 (13)

10 (13) VL

98 (71) 29 (21) clsqtl,

n = Total in group.

Bacteriological data at time of admission. Forty-three neonates had infections on admission, including meningitis (one), septicaemia (nine), pneumonia (13), urinary tract infection (UTI) (one), peritonitis (one), omphalitis (eight) and severe systemic infection without localization (16). Twelve babies required mechanical ventilation and/or exchange transfusion merely because of their infections. Three died within a few hours after admission (septicaemia was present in all three cases, one with a Group B streptococcus, one with Pseudomonas aeruginosa and one with Candida albicans). The organisms isolated from all positive neonates are shown in Table III. Table III.

Organisms from colonized neonates and neonates with perinatal or hospital-acquired infections Perinatal

Organisms Coagulase negative cocci Staphylococcus aureus Group B streptococcus Pneumococcus Group D streptococcus Escherichia coli Klebsiella pneumonia Pseudomonas aeruginosa Other bacteria Candiak albicans

Colonized (n = 76) No. (%)

Infected (n = 43) No. (%I

47 (62)i.w 22 (2% I./ 11 (14)o.i 0 mdbq 20 (26);. 6

5 11 11 2 7

2: c;;;+ -

(16) 127 (28)

: $1 0 (0)

(12) (26) (26) (5) (16)

z $1 1 (2)

Hospitalacquired (n = 55) No. (%) 18 (33) 1i ‘“(2 1 (2) 3 (5) ‘: % E ‘{E,’ 3 (5)

n = Total in group.

Colonized babies mainly yielded cultures positive for Gram-positive coagulasenegative cocci, Staphylococcus aura, Group D streptococci and Escherichia coli. Staphylococcusaureus, Group B streptococci, Group D streptococci, E. coli and Klebsiklla pneumoniaewere the most important organisms responsible for infection.


in a neonatal


care unit


The number of positive culture sites was significantly higher in the infected than in the colonized neonates (Table IV). The umbilicus was most frequently positive in both groups (Table V), but this site was significantly more frequently Table IV. Number of positive-culture

sites frd~r~wized

or infected neonates (per cent) on

Number of positive sites






Colonized (n = 76) Infected (n = 43) n = Number

in group.

Table V. Sites of positive cultures from colonized or infected neonates on admission Culture sites (% positive) Neonates Colonized Infected (n(n= =43) 76) n = Number







Tracheal tube


23 32

30 37





in group.

positive in the colonized group than in the infected group. The other culture site found to be mainly positive in colonized neonates was the ear (30 per cent). The infected neonates had significantly higher percentages of positive cultures from the nose, skin, blood and tracheal tubes. Antibiotic treatment. One hundred and twenty-five neonates (69 per cent) were receiving antibiotics on admission. Retrospectively, 79 per cent of all neonates, later considered negative, 59 per cent of all colonized and 75 per cent of all infected neonates were treated. The antibiotics were stopped in the culture negative and colonized group when the results of the cultures were known. Neonates with infection were treated with the appropriate antibiotics. Intubation and arterial or venous cannulation alone were not considered indications for antibiotic prophylaxis. Hospital-acquired infection Incidence of hospital-acquired

infection. Fifty-five neonates (23 females and 32 males) acquired an infection during hospitalization in the NICU (30.4 per cent of all patients). These infections became manifest on the 5th to the 12th day after admission and were found in neonates previously culture-negative (46) or positive (9). These infections included septicaemia (15), pneumonia (25), UT1 (12), enteritis (1) and soft tissue infections (omphalitis, pyoderma, conjunctivitis, phlebitis) (19). F our children died due to these infections alone (all Ps. aeruginosa), another four died from other complications together with infections (three caused


J. A. A. Hoogkamp-Korstanje

et al.

by Ps. aeruginosa). The organisms isolated from these acquired infections are also presented in Table III. Coagulase-negative cocci, Staph. auwus, E. coli and Ps. wuginosa were the most important organisms. The origin of infections by the Gram-positive cocci and Ps. aeruginosa were studied in more detail. Staphylococcus aureus. Hospital infections caused by Staph. nureus were both auto- and nosocomial infections. Thirty-eight babies, negative for Staph. aureus on admission yielded cultures of Staph. aureus during hospitalization; 18 of them developed an infection caused by this organism. Eleven different phage-types were isolated. Thirteen neonates had a phage-type only lysed by phage 94, six a phagetype lysed by phages 3a, 3c, 55, 71 (group III) and five had a phage-type lysed by phages 29,52,52a, 80,81,42e, 83a (groups I and III). The first and the third types mentioned were also found in cultures from eight members of the medical staff. The other 14 neonates had eight different phage-types, not found on staff members. Mothers and other visitors of the ward were not screened for staphylococci. Coagulase-negative cocci. Hospital infections caused by Gram-positive coagulasenegative cocci were considered to be mainly nosocomial infections. Seventy-one babies showed these cocci on one or more culture sites during hospitalization, whereas these cocci were not found on admission. Eighteen had an infection, including septicaemia (four), pneumonia (nine) and soft tissue infections (five). Seventy-one per cent of all strains were micrococci belonging to subgroup 2 (21 per cent), 5 (17 per cent), 6 (30 per cent) and 7 (3 per cent). Twenty-nine per cent were staphylococci belonging to subgroup III (25 per cent), II, IV and V (4 per cent). The strains were not phage-typable. All strains produced betalactamase and were resistant to penicillin, methicillin, cloxacillin, cephalothin, tetracycline, kanamycin and gentamicin. The same strains were isolated from hands and nails of 10 members of the staff and not from other specimens of the environment or from hands of laboratory staff outside the ward. Pseudomonas aeruginosa. All neonates with pseudomonas infection were negative for this organism on admission and were considered to have a nosocomial infection. Ps. aeruginosa could be isolated from many sources in the environment: floor, furniture, washing stands (taps and sinks), suction equipment, respiratory apparatus, hands and clothes of nurses and ointment (Vaseline) used on baby lips! These sources have been eliminated by rigid disinfection procedures during the next year. Factors injluencing the acquisition of a hospital infection. Table VI presents the data on birth conditions and the frequency of instrumentation of the neonates with and without hospital-acquired infection. The mean gestational age was the same in both groups. The birth weight was slightly lower in the group with acquired infections, the percentages of neonates born by Caesarian section was similar in both groups. The use of mechanical ventilation and continuous positive airway pressure was similar in both groups, but the period of mechanical ventilation was twice as long in the group with infection. The need for an arterial catheter (umbilical or radial) was similar in both groups. Forty-nine per cent of the neonates with acquired infections had long-term venous catheters in contrast to 23 per cent of the neonates


in a neonatal


care unit

Table VI. Association of conditions at birth and instrumentation hospital-acquired infection


in neonates with or without


&A gestation period (weeks) Mean birth weight (g) Caesarian section (%) Positive cultures on admission (%) Infection on admission (%) Mechanical ventilation (%) Mean duration (days) Arterial catheter (%) Mean duration (days) Venous catheter (%) Mean duration (days) n = Number


Positive (n = 55)

Negative (n = 126)

33.0 1810 ;i

33.9 2091 22 55



14.3 50.9 8.7 49 20.3

5f.Z 4.4 23 10.2

in group.

in the negative group. The period of venous cannulation was twice as long in the group with hospital-acquired infection. The mean period of hospitalization was 12-9 days for the neonates without, and 25.8 days for the neonates with hospital infections. Discussion

The acquisition of certain bacteria during or after birth by the upper respiratory tract, gastrointestinal tract and the skin is part of the normal development. Coagulase-negative staphylococci and diphtheroids are commonly found on the skin, whereas cultures of the nose, throat, umbilicus and rectum taken immediately after birth are reported to be sterile in the majority of newborn babies (Sarkany and Gaylarde, 1967). Colonization with organisms other than those mentioned and heavy colonization of various sites increase the risk of development of bacterial infection. Therefore, knowledge of the degree of colonization of various sites may be helpful in the early detection of neonatal infection, especially in those neonates most at risk, such as newborns requiring intensive care for a variety of reasons. The value of bacteriological screening on admission has been demonstrated by the finding that 24 per cent of all our newborns appeared to have a perinatal infection, whereas only 10 per cent had been admitted with suspicion of infection. All newborns with infection (a total of 43) had positive cultures with organisms not belonging to the normal colonizing flora from more than one site. None of the newborns considered bacteriologically negative had an infection. Infection occurred significantly more often among babies born after prolonged rupture of the membranes or with congenital anomalies. While this is a well known phenomenon, another statement, the higher risk of bacterial infection in infants with low birthweight (i.e. premature and immature), could not be demonstrated in our population, although 138 premature and immature infants were involved in the study.


J. A. A. Hoogkamp-Korstanje

ef al.

The high nasal colonization rate of the neonates with infection when compared to those only colonized was striking. Respiratory tract colonization rates in healthy full-term and premature infants vary from 8.4 to 10 per cent after birth (MacAllister, Givan and Turner, 1974) which is in agreement with the percentages we found in colonized babies. The incidence of initial respiratory tract colonization in critically ill newborns has been reported at 42 to 43 per cent (Harris, Wirtschafter and Cassidy, 1976; Sprunt, Leidy and Redman, 1978). Our rate was 44 per cent in neonates with infection. The predominant pathogens isolated from neonates with infections were E. coli, Staph. aureus and Group B streptococci. It was noted that the same organisms were found in a number of babies who were only colonized. Isolation of these pathogens does not, therefore, indicate the existence of infection. Other factors as mentioned above, particularly the degree of colonization and localization of the pathogens, may determine the occurrence of infection. It is clear that antibacterial drugs have been greatly overprescribed in our NICU. This was shown by the fact that 69 per cent of all neonates received antibiotics on admission whereas only 24 per cent had an infection. This again demonstrates the difficulties in diagnosing neonatal infections on clinical and early laboratory evidence at a time when treatment cannot be delayed and the results of cultures are not yet known. Fifty-five neonates (30 per cent) acquired a hospital infection. Nine of these had a previous perinatal infection, different from the infection acquired during hospitalization. Pneumonia, soft tissue infections and septicaemia were most frequently observed. There was an obvious correlation between the period of instrumentation and the occurrence of an infection. As might be expected, tracheal intubation will decrease the local defence mechanism and increases the risk of colonization of the lower respiratory tract with the subsequent risk of subsequent systemic infection. The percentage of neonates requiring mechanical ventilation was similar in infected and non-infected groups. Therefore, intubation alone did not increase the incidence of infection. As about 70 per cent of all neonates had been treated with antibacterial drugs briefly on admission, it was difficult to estimate the role of antibiotics. We have the impression that antibiotics do not reduce the incidence of infection in contrast to the opinion of Harris et al. (1976). The length of period of intubation appeared to be the most important factor influencing the incidence of infection. The longer the period of intubation the higher the risk of infection. In our opinion intubation for a short time is not an indication for antibiotic treatment or prophylaxis and the same can be said of the role of arterial and venous cannulation: the incidence of infection is not related to insertion of i.v. needles and catheters, but the longer the period of cannulation the greater is the risk of infection. The NICU mortality rate in 1977 was 19.8 per cent (36 neonates). Four neonates died from a hospital infection. The overall mortality rate was not therefore markedly influenced by hospital infections. The morbidity, however, was influenced significantly by hospital infection: hospitalization of neonates with a hospital infection was twice as long as that of neonates without a hospital infection. Gram-positive cocci were involved most frequently in hospital infections.


in a neonatal


care unit


Staph. aureus and coagulase-negative cocci were found in equal numbers. The most important Gram-negative organisms were E. coli and Ps. aeruginosa. A number of neonates had mixed infections, in others Gram-positive and Grdmnegative organisms succeeded each other. We concluded that the infections caused by staphylococci and Ps. aeruginosa were usually nosocomial infections. The outcome of the Gram-positive infections was satisfactory. All neonates recovered after adequate antimicrobial therapy. The outcome of pseudomonas infection was poor: four neonates died from this infection alone, another three died from other complications together with pseudomonas infection. The isolates of Gram-positive coagulase-negative cocci were of special interest. Classification of our strains demonstrated that a number were so-called nonpathogenic micrococci, frequently found on the human skin. Others were staphylococci mostly from group III or Staph. saprophyticus (Baird Parker, 1963), usually regarded as non-pathogenic, although some strains are associated with urinary tract infections (Mitchell, 1964; Namavar et al. ; 1978). Staphylococcus epidermidis, isolated from a variety of infections in man (Holt, 1969), was only found once. We concluded that these strains were circulating inside the NICU ward as they were not isolated from the neonates on admission and were frequently found on hands of the staff members and not outside; the staff from the laboratory yielded other types. Additionally, these strains had a particular resistance pattern: they were multiple-resistant, including methicillin and cloxacillin. We demonstrated that resistance to these drugs was based on tolerance. Colonization of 71 neonates (40 per cent) a few days after admission indicates a high colonization rate with extensive cross-contamination. As these micrococci and Staph. saprophyticus caused infection in 18 babies, they must be considered as pathogenic for this special group of patients. These strains have also been reported from another NICU where the use of antibiotics was widespread (Bernick, Baublis and Borer, 1979). We consider this is the reason for the high colonization rate of our NICU with these resistant strains. At the same time a decrease of Gramnegative infections was observed compared to the preceding iyears. This is shown by the decrease in the incidence of hospital infections caused by Ps. aeruginosa in the’ preceding four and the following three years. The incidence of pseudomonas infection (as a percentage of the whole NICU population) was 14.1 (1973), 13.2 (1974), 8.6 (1975), 8.9 (1976), 4.9 (1977), 5*4( 1978), 4.6 (1979), 2.6 (1980). A similar observation of changing a pattern from Gram-negative to Gram-positive infections was made recently by Battisti, Mitchison and Davies (1981). Although this changing pattern might have some advantages because of the better outcome of Grampositive infections, we recognize that a number of the Gram-positive infections were due to cross-infection and probably could have been avoided or diminished with even more stringent and meticulous aseptic procedures. However, it is likely that these same procedures associated with handling the smallest babies, often immuno-compromised, will limit the effects of our effort.



wish to thank


ten Brug for preparing

the manuscript.


J. A. A. HoogkampKorstanje

et al.

References Baird Parker, A. C. (1963). A classification of micrococci and staphylococci based on physiological and biochemical tests. Journal of General Microbiology 30, 409-427. Battisti, O., Mitchison, R. & Davies, P. A. (1981). Changing blood culture isolates in a referral intensive care unit. Archives of Diseases in Childhood 56, 775-778. Bemick, J. J., Baublis, J. V. & Borer, R. C. (1979). Nosocomial bacteremia in neonates. Pediatric Research 13, 388. Harris, H., Wirtschafter, D. & Cassidy, G. (1976). Endotracheal intubation and its relationship to bacterial colonization and systemic infection of newborn infants. Pediatrics 56, 816-823. Holt, R. J. (1969). The classification of staphylococci from colonized ventriculoatrial shunts. Journal of Clinical Pathology 22, 475-482. McAllister, T. A., Givan, J. & Turner, M. J. (1974). The natural history of bacterial colonization of the newborn in a maternity hospital. Scottish Medical Journal 19, 11912s. Mitchell, R. G. (1968). Classification of Staphylococcus albus strains isolated from the urinary tract. Journal of Clinical Pathology 21, 93-96. Namavar, F., GraafI, J. de, With, C. de & MacLaren, D. M. (1978). Novobiocin resistance and virulence of strains of Staphylococcus saprophyticus isolated from urine and skin. Journal of Medical Microbiology 11, 243-248. Sarkany, I. & Gaylarde, C. C. (1967). Skin flora of the newborn. Lmcet i, 589-590. Sprunt, K., Leidy, G. & Redman, W. (1978). Abnormal colonization of neonates in an intensive care unit. Pediatric Research 12, 998-1002.