SciELO - Scientific Electronic Library Online

vol.8 issue4A new broom?The cost of nephroblastoma treatment in South Africa: A very cost-effective investment with guidelines for the rest of Africa author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand



Related links

  • On index processCited by Google
  • On index processSimilars in Google


South African Journal of Child Health

On-line version ISSN 1999-7671
Print version ISSN 1994-3032

S. Afr. j. child health vol.8 n.4 Pretoria Nov. 2014 



Hospital-acquired Klebsiella pneumoniae infections in a paediatric intensive care unit



S T HlopheI; N H McKerrowII

IMB ChB, DCH (SA), FC Paed (SA); Department of Paediatrics and Child Health, Nelson R Mandela School of Medicine, Faculty of Health Sciences, University of KwaZulu-Natal, Durban, South Africa
IIMB ChB, BA, DCH (SA), FC Paed (SA), MMed, PG Dip Int Res Ethics; Department of Paediatrics and Child Health, Nelson R Mandela School of Medicine, Faculty of Health Sciences, University of KwaZulu-Natal, Durban, South Africa; Department of Health , Pietermaritzburg, South Africa





BACKGROUND: Hospital-acquired infections (HAI) are a significant problem in the delivery of intensive care services. Each nosocomial infection prolongs an affected patient's stay in hospital by 5 - 10 days.
METHODS: A retrospective case control chart review of children admitted to the paediatric intensive care unit (PICU) in Grey's Hospital between July 2003 and December 2010, who developed a hospital-acquired Klebsiella pneumoniae infection, was undertaken to describe the trend in HAI in a newly commissioned PICU and to identify any association with the patient demographics and modalities of care. Patients with a K. pneumoniae infection were identified through the PICU infection control surveillance system. Each case was matched to a control of the same age admitted during the same period, with a similar clinical diagnosis.
RESULTS: During the 7.5-year period, 2 266 children <12 years of age were admitted to the PICU. Of these, 113 had K. pneumoniae cultured from a body fluid >48 h after admission, including 23 cultured from the blood. Clinical records were obtained for 14 of these patients and matched to control cases of similar age and gender who were admitted at the same time. The length of stay in both the PICU and hospital was longer in children with an HAI compared with the control group (3.7 v. 2.9 and 18.5 v. 9.14, respectively; p=0.04). There was no significant difference in the treatment modalities provided to the two groups, although most patients in the sample group required invasive treatment.
CONCLUSION: K. pneumoniae nosocomial infection was a significant problem encountered in Grey's Hospital paediatric intensive care. It has major cost implications, as it prolongs the length of stay in intensive care and hospital.



Hospital-acquired infections (HAI) are a significant problem in the delivery of intensive care services. Each nosocomial infection prolongs an affected patient's stay in the hospital by 5 - 10 days. This is an important preventable cause of increased cost, morbidity and mortality among hospitalised patients. The incidence of HAI increases with the use of invasive devices and with increased duration of hospitalisation.[1] Hospital-acquired Klebsiella pneumoniae in neonatal intensive care units (ICUs) has increased recently with numerous reports of outbreaks in neonatal ICUs throughout South Africa (SA).[2] Investigations into these outbreaks have identified an underlying context characterised by overcrowding, understaffing and a breakdown in infection control measures.

The incidence of HAI varies across ICUs and with different patient profiles.[3] Patients in facilities offering a higher level of care are at greater risk of developing HAI compared with those facililties offering a lower level of care.[4] Patient characteristics and treatment modalities play an important role in susceptibility to, and outcomes from, HAI. In multivariate analyses, modalities of care - specifically central venous catheters, arterial catheters and other procedures in the paediatric ICU (PICU) - as well as brief periods out of the PICU, were all associated with bloodstream infection.[5]

The PICU in Grey's Hospital, Pietermaritzburg, was established in 2003. It was commissioned as a 7-bed unit in July 2003 and increased to an 8-bed unit in 2006. It was closed for 9 months from November 2008 to July 2009 owing to lack of staff, and for 1 month in 2010 during a public sector strike. Since it was commissioned, surveillance systems have noted a low but rising incidence ofKlebsiella nosocomial infections in the unit. This poses the question whether this is an independent variable or whether it is related to the spectrum of patients admitted to, or the modalities of care provided in, the unit. This study attempted to answer this question by:

  • determining the rate of HAI
  • identifying the contribution of K. pneumoniae to HAI
  • describing the profile of patients, treatment modalities and the relationship between these and the incidence of K. pneumoniae HAI.



This was a retrospective case-control chart review of children admitted to the PICU in Grey's Hospital, Pietermaritzburg, over a 7.5-year period from July 2003 to December 2010.

Children of all ages admitted to the PICU and who developed a K. pneumoniae infection were eligible for enrolment. Only children with a positive blood culture obtained >48 h after admission to the PICU were considered for enrolment, while tracheal aspirates and urinary cultures were not considered in order to exclude colonisation of these sites. Patients with hospital-acquired K. pneumoniae were identified via an infection control surveillance system and were matched with children of the same age and gender who were admitted during the same period with a similar diagnosis.

Exclusion criteria included readmission to the PICU during a single hospital stay, positive blood cultures taken within 48 h of admission, positive cultures from sites other than blood and children admitted from another ICU.

Clinical records were retrieved from the medical registry and data extracted from the chart for analysis. The χ2 test was applied for studying the significance of these risk factors in K. pneumoniae HAI. The probability value was obtained and considered significant if below 0.05.



During the 7.5-year study period, 2 266 patients were admitted to the PICU at Grey's Hospital (Table 1). Of these, 113 had a positive culture for K. pneumoniae, although only 23 (20%) of these were positive blood cultures. Clinical records were retrieved for 14 of these children and matched with control cases identified as described above.

The average number of patients admitted to the PICU during the study period was 28.7 patients per month (median 27, range 13 - 35.3). This figure fluctuated between 27 and 34 patients per month from 2004, once the PICU was fully commissioned. This equates to an incidence of K. pneumoniae HAI of 10/1 000 patients. The average bed occupancy was 75.0%, ranging from 57.8% in 2003 to a peak of 79.9% in 2007. The lower rate seen from 2008 to 2010 was associated with the closure of the PICU and the public sector strike. The average length of stay ranged from 4.7 to 9.5 days, with an average of 7.6 days.

During the study period, there were 113 positive cultures for K. pneumoniae, including 23 blood cultures (20.4%). K. pneumoniae was grown in 24.2% of all positive blood cultures during this period.

The background data of the sample and control patients are presented in Table 2. The mean age of patients in the sample was younger than that of the control group (18.3 v. 26.3 months, respectively), although the range was similar in the two groups (1 - 145 v. 2 - 144 months, respectively). There were equal numbers of girls and boys in the study, but more girls in the control and more boys in the sample groups. The HIV status was unknown in 35.7% (5/14) of the sample and 21.4% (3/14) of the control group. Although more children in the control group were HIV-positive (28.6% v. 21.4% in the sample group), a higher proportion of children in the sample group had evidence of disease progression manifest by a World Health Organization stage of 2 or more (14.3% v. 7.1% in the control group). Almost twice as many children in the sample group were wasted compared with those in the control group. However, none of these differences was statistically significant.



Table 3 shows the treatment modalities received by children in this study. Although more children with K. pneumoniae nosocomial sepsis required nasotracheal intubation, second- or third-line antibiotics and steroids, these differences were not statistically significant.



Table 4 shows the outcome of children in this study. The duration of hospitalisation before a positive blood culture ranged from 2 to 14 days. Children in the sample group spent more time in hospital prior to admission to the PICU than those in the control group (3.7 v. 2.9 days, range 0 - 21 v. 0 - 26 days, respectively). The total period of hospitalisation of children in the sample group was double that of those in the control group (18.5 v. 9.1 days, range 5 - 60 v. 2 - 29 days, respectively), which was statistically significant (p=0.04).




Incidence and trends of K. pneumoniae HAI

Yogaraj et al.[5] found that in their experience, in Washington DC, bloodstream infections accounted for 28% of HAI in the PICU, followed by ventilator-associated pneumonias, which accounted for a further 21%. Lakshmi et al.[6] reported on 116 episodes of bloodstream infection in 86 patients in India in 2006, with an incidence of 31.2 episodes/1 000 patient days. Marra et al.[7] found that K. pneumoniae ranked among the top ten pathogens causing bloodstream infection in the USA and Canada, and that it was the third most prevalent pathogen isolated in Latin America.

K. pneumoniae HAI is a significant problem in the delivery of intensive care services, and prolongs hospital stay. In this study, there was an incidence for K. pneumoniae HAI of 10/1 000 admissions, with a range of 0 - 28.8/1 000 patients per annum once the PICU was fully commissioned. Unfortunately, as this was a retrospective study, we were unable to access data on patient days, therefore a figure for comparison with international experience cannot be provided. There are no similar reports of K. pneumoniae HAI from SA, although outbreaks have been reported in RK Khan Hospital, Durban,[8] in 1992, Mahatma Ghandi Memorial Hospital, Durban,[2] in 2005 and Tygerberg Hospital, Cape Town, in 2001.[9]

Patient profiles

During the study period, 2 266 children aged 0 - 12 (average 2) years were admitted to the PICU in Grey's Hospital. Even though the number of patients with HIV infection was higher in the control compared with the sample group (28.6% v. 21.4%), more patients in the sample group were severely immunocompromised (14.3% v. 7.1%). Donowitz et al. [10] revealed that HIV-infected children and those with prolonged steroid treatment have a significantly reduced immunity which, in turn, is a major risk factor for the acquisition of nosocomial infections. Although not statistically significant, a similar trend was seen in this study, with a higher proportion of patients in the sample group receiving steroids.

Cell-mediated immunity is greatly suppressed in malnourished children, which predisposes these children to nosocomial infection. Isaack et al. [11] have reported a higher incidence of nosocomial infection among these children. The current study reveals a similar trend, with 50% of children in the sample group being wasted, compared with only 28.6% in the control group - although, once again, this was not statistically significant.

Impact of management on the incidence of K. pneumoniae HAI

As modalities of care increase, susceptibility of patients to nosocomial infection increases. The association between common invasive procedures, including intubation, urinary catheterisation and central venous lines, and K. pneumoniae HAI was studied in the current study. A higher rate of intubation and ventilation in the sample group (71.3%) compared with the control group (50.0%) was documented.

Tullu et al.[3] found an incidence of 27.54% for hospital-acquired pneumonia in patients with endotracheal intubation. This translated to 7.96/100 days of endotracheal intubation and 8.92/100 days of mechanical ventilation. The incidence of ventilator-associated pneumonia increased with the duration of ventilation; this has been reported previously.[12] In the current study, more patients were ventilated in the sample group, increasing the risk of nosocomial ventilator-associated pneumonia.

Although Darbyshire et al.[13] and Narendaran et al.[14] reported indwelling urinary catheters as a risk factor for HAI, this was not so in the current study, where there was little difference between the two groups; 92% of patients in sample group were catheterised as compared with 86% in control group.

Central venous catheters are often used in PICUs for vascular access, and Tacconelli et al.[15] and Kunin and McCormack.[16] have reported on the association between the duration of catheterisation and HAI. In the current study, there was no difference between both groups, since 42.8% in each group had central venous catheters. Yogaraj et al.[5] reported that PICUs have one of the highest central venous catheter-associated bloodstream infection rates, namely 7.7 infections per 1 000 central venous catheter days, surpassed only in burn units and neonatal ICUs. This study did not find a significant difference in the proportion of children in each group who had a central line.

In total, a higher proportion of children with HAI had one or other recognised risk factor, underlying immune status or modality of care, compared with those without HAI. However, unlike other published reports, this was not statistically significant for any individual risk factor, probably owing to the small sample size in this study.

The only statistically significant finding in this study was an increased duration of hospitalisation in children with an HAI (18.5 v. 9.14 days), which is similar to the findings of Deep et al.,[1] who reported that the duration of stay was longer in patients with nosocomial infections (9.8 v. 1.8 days), as was the mortality (30.5%) of the studied population. They reported common causes of HAI being Klebsiella and Pseudomonas. Mortality rate was not calculated in the current study.



K. pneumoniae HAI was a significant problem encountered in the PICU at Grey's Hospital. There was an association between HAI and recognised risk factors, as well as a doubling in the length of stay in hospital. This prolonged stay would not only be associated with an increased cost of care, but also with an ongoing increased risk of further HAI.

Further study is required to establish the broader consequences of these infections and, given the frequency of reported HAI outbreaks in neonatal nurseries, into the prevalence of such infections in other vulnerable groups.



1. Deep A, Ghildiyal R, Kandian S, Shinkre N. Clinical and microbiolgical profile of nosocomial infections in the paedriatric intensive care unit (PICU). Indian Pediatr 2004;41(12):1238-1246.         [ Links ]

2. Moodley P, Coovadia YM, Sturm WA. Intravenous glucose preparation as a source of an outbreak of extended-spectrum β-lactamase-producing Klebsiella pneumoniae infections in the neonatal unit of a regional hospital in KwaZulu-Natal. S Afr Med J 2005;95(11):861-864.         [ Links ]

3. Tullu MS, Deshmukh CT, Baveja SM. Bacterial nosocomial pneumonia in paediatric intensive care unit. J Postgrad Med 2000;46(1):18-22.         [ Links ]

4. Sacho H, Schoub BD. Current perspective on nosocomial infections. Halfway House, South Africa: GlaxoWellcome, 1997.         [ Links ]

5. Yogaraj JS, Elward AM, Fraser VJ. Rate, risk factors, and outcomes of nosocomial primary bloodstream infection in pediatric intensive care unit patients. Pediatrics 2002;110(3):481-485. []        [ Links ]

6. Lakshmi KS, Jayashree M, Singhi S, Ray P. Study of nosocomial primary bloodstream infections in pediatric intensive care unit. J Trop Pediatr 2006;53(2):87-92. []        [ Links ]

7. Marra AR, Wey SB, Castelo A, et al. Nosocomial bloodstream infections caused by Klebsiella pneumoniae: Impact of extended-spectrum β-lactamase (ESBL) production on clinical outcome in a hospital with high ESBL prevalence. BMC Infect Dis 2006;6:24 []        [ Links ]

8. Coovadia YM, Johnson AP, Bhana RH, Hutchinson GR, George RC, Hafferjee IE. Multiresistant Klebsiella pneumoniae in a neonatal nursery: The importance of maintenance of infection control policies and procedures in the prevention of outbreaks. J Hosp Infect 1992;22(3):197-205.         [ Links ]

9. Cotton MF, Pieper CH, Kirsten GF, Orth H, Theron DC. Necrotising enterocolitis as an infectious disease: Evidence from an outbreak of invasive disease due to extended-spectrum beta-lactamase-producing Klebsiella pneumonia. S Afr Med J 2001;91(2):133-135.         [ Links ]

10. Donowitz LG, Wenzel RP, Hoyt JW. High risk of hospital-acquired infection in the ICU patient. Crit Care Med 1982;10(6):355-357.         [ Links ]

11. Isaack M, Mbise RL, Hirji KF. Nosocomial bacterial infections among children with severe protein energy malnutrition. East Afr Med J 1992;69(8):433-436.         [ Links ]

12. Langer M, Mosconi P, Cigada M, Mandelli M. Long-term respiratory support and risk of pneumonia in critically ill patients. Intensive Care Group of Infection Control. Am Rev Respir Dis 1989;140(2):302-305. []        [ Links ]

13. Darbyshire PJ, Weightman NC, Speller DC. Problems associated with indwelling central venous catheters. Arch Dis Child 1985;60(2):129-134.         [ Links ]

14. Narendran V, Gupta G, Todd DA, John E. Bacterial colonization of indwelling vascular catheters in newborn infants. J Pediatr Child Health 1996;32(5):391-396.         [ Links ]

15. Tacconelli E, Tumbarello M, Cauda R. Central venous catheters and bloodstream infection. JAMA 2000;283(4):478-479. []        [ Links ]

16. Kunin CM, McCormack RC. Prevention of catheter-induced urinary-tract infections by sterile closed drainage. N Engl J Med 1966;274(21):1155-1161. []        [ Links ]



S T Hlophe

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License