Queen Elizabeth University Hospital: case note review - overview report

4. Description of Cases and Episodes included in the Review

4.1 Overview

The criteria for the inclusion of patients in our Review were defined in our ToR (Chapter 2), and the associated methodology for identification of these cases is further described in Chapter 3, section 3.2. The work undertaken using these criteria before we began our work suggested that 85 patients, who had experienced 120 infection episodes, were eligible for review. In the course of our work, however, we identified some adjustments.

1. We identified one patient who had had two episodes of eligible infection, the earliest of which had occurred shortly before the move of the Children’s Hospital from Yorkhill to the new QEUH campus. As this fell outside the timeline of the Review and did not relate to the QEUH/RHC site, we considered that this first episode was ineligible for inclusion. However, the patient remained in the review by virtue of a second qualifying episode.

2. We subsequently identified a patient who had been identified for the Review with a single episode of bacteraemia caused by Moraxella catarrhalis. This is a Gram-negative bacterium, but is not considered to be environmental and spreads predominantly from person-to-person by droplet contamination. We considered this ineligible for inclusion and both the patient and the episode have been excluded from our analysis. However, as the family had been notified of, and subsequently agreed for the Case Note Review, we reviewed this child’s records and will provide the family with an individual report.

3. One further patient, who otherwise fulfilled the criteria for the Review, was not included at the family’s request. This patient had had four episodes of infection and although no other records were extracted or reviewed by the Panel, the timings and types of these infections were included in the microbiology data provided to the Panel because this could have contributed to our understanding of any clustering with other cases with similar infections.

In summary, in this report we provide findings for 84 patients who, between them, had 118 episodes of infection and were eligible for the Review.

4.2 Demographic and Clinical Profiles of Patients included in the Review

The characteristics of the patients included in the Review are summarised in Table 4.1.

Table 4.1 Demographic and Clinical characteristics of cases included in the Review

Total no. of cases	84	100%
Gender	Male 32 Female 52	38% 62%
Diagnosis	Leukaemia 36 Lymphoma 7 CNS tumour 11 Solid tumour 23 Non-malignant Disease 7	43% 8% 13% 27% 8%
Age at diagnosis	Median (Range): 3y 9m (Birth–18y 4m)
No. of infection episodes in the Review	One episode 65 Two episodes 10 Three or more episodes 9 (n=3 in 6; n=4 in 2; n=8 in 1)	77% 12% 11%
Age at first episode of infection	Median (Range): 5y 11m (3m–18y 10m)
Alive* at the time of the publication of this report	62	74%

*Further discussion of patients who have died is provided in section 6.2.

4.2.1 Gender

The observation that 62% of the cases in this series were female is of interest. Age Standardised Rates for cancer in children to the age of 15 in northern European countries (and in most developed countries) indicate a slight excess of boys with a M:F ratio in the range of 1.1-1.2. The ratio in older teenagers and young adults is closer to 1.0. This is confirmed in the most recent publication of data for cancer in children and young people in Scotland^[48], which states “In the ten year period 2009-2018, 1,298 children (aged 0-14, 53% male) were diagnosed with cancer and 1,996 young people (aged 15-24, 51% female) were diagnosed with cancer”.

The great majority of patients in our Review were aged under 15 years at diagnosis and we are not able to offer any obvious explanation for the reversal of the expected gender balance. There is no reason to believe that gender should influence the risk of infection at this age, and the finding of a female excess is unexpected. As the number of cases in this series is relatively small, the likelihood of this being a real effect is also small. It would nevertheless be appropriate for the staff in the Paediatric Haematology Oncology service at NHS GGC to audit gender patterns of all bacteraemias in children under their care to assess this further.

4.2.2 Age

The patients included in the Case Note Review were young, both at the diagnosis of their cancer or other condition (median age 3 years 9 months) and at the time of their first Gram Negative Environmental (GNE) infection (median age 5 year 11 months). The young median age at diagnosis is not unexpected and reflects the peak of diagnosis of the commonest form of childhood leukaemia, and some solid tumours, seen in the pre-school age range.

We consider the distribution of diagnoses in patients included in the Case Note Review to be representative of the age range expected to be under treatment in the Paediatric Haematology Oncology service at NHS GGC.

4.2.3 Diagnosis

The classification of cancer in children and young people uses a different system to that applied in adults. Individual diagnoses may be very rare and analyses typically group patients into four main groups – leukaemias, lymphomas, central nervous system (CNS) tumours, and solid tumours. The data shown in Table 4.1 for the distribution of diagnoses amongst patients included in the Review are broadly in line with that expected, although there is a small excess of leukaemia (43% of the cases in the Review group vs 31% in the Scottish data for 2009-2018) and a corresponding deficit of both CNS (13% vs 27%) and solid tumours (27% vs 34%). The proportion of children with lymphoma is as expected (8% vs 8%).

The leukaemia excess is consistent with two observations: i) almost all children with leukaemia require periods of intensive treatment with chemotherapy, and are therefore more susceptible to infection; and ii) NHS GGC is the national center for paediatric bone marrow stem cell transplantation (SCT) in Scotland and receives referrals of patients with high risk leukaemia and other blood disease from other centres. SCT patients are especially at risk of serious infection.

The small number of children in the Review with non-malignant diagnoses included those with serious blood diseases such as aplastic anaemia and other bone marrow failure syndromes (n = 5), haemophilia (1), and two patients who had initially been diagnosed with a malignant condition but were subsequently shown to have alternative but nevertheless serious non-malignant conditions.

An additional factor to consider is that the Paediatric Haematology Oncology service at NHS GGC is the designated national bone marrow stem cell transplant service for children in Scotland. Some of the children in the series had been referred for stem cell transplantation after initial treatment elsewhere. The requirement for such treatment is typically seen amongst children with high risk, including relapsed, leukaemia and those with severe bone marrow failure syndromes. Overall, however, we consider the population of patients seen in the Case Note Review to be representative of the case mix expected to be under treatment at NHS GGC.

4.2.4 Frequency of infection episodes

It is noteworthy that although the large majority (77%) of patients included in the Review had only one episode of GNE infection, almost one quarter had more than one, and several patients had >2 episodes. We believe this indicates the persistence of risk in this population, with the continuing presence of a central venous line and, in most, ongoing exposure to chemotherapy. It may also imply the persistence of environmentally associated risk.

Further detail about the frequency and type of organisms causing the bacteraemias in the whole case series is discussed in section 4.3.

4.3 Microbiology profile of the isolates identified in the Review

We have described (section 3.2) how cases were selected for the Case Note Review and have identified the adjustments we made to arrive at the final figures of 84 cases and 118 infection episodes eligible for our Review (section 4.1).

Table 4.2 provides a summary of all bacteraemias at genus level. Table 4.3 provides a summary of the same data but at the species level. Note that data from 2015 represent only a partial year (from May 15^th 2015^[49]) and that the isolates from the patient who was eligible but whose family did not wish them to be part of the Review (patient 3 discussed in section 4.1) are included within these data. Note also that, as some episodes were polymicrobial (i.e. more than one bacterium was identified in the same blood culture), the totals given in these tables exceed the total number of episodes considered in the Review.

Table 4.2 Frequency of infection by organism (defined at genus level) and year

Organism by genus	2015	2016	2017	2018	2019	Total
Achromobacter					1 (3.6%)	1 (0.6%)
Acinetobacter		2 (7.7%)	6 (11.8%)	2 (4.2%)	1 (3.6%)	11 (7.1%)
Aeromonas		1 (3.8%)			1 (3.6%)	2 (1.3%)
Brevundimonas			1 (2.0%)			1 (0.6%)
Burkholderia			1 (2.0%)	1 (2.1%)		2 (1.3%)
Chryseobacterium		1 (3.8%)	1 (2.0%)	2 (4.2%)	1 (3.6%)	5 (3.2%)
Citrobacter			3 (5.9%)	2 (4.2%)		5 (3.2%)
Cupriavidus			1 (2.0%)	1 (2.1%)		2 (1.3%)
Delftia			2 (3.9%)		1 (3.6%)	3 (1.9%)
Elizabethkingia		2 (7.7%)	3 (5.9%)		1 (3.6%)	6 (3.9%)
Enterobacter		1 (3.8%)	8 (15.7%)	10 (20.8%)	8 (28.6%)	27 (17.4%)
Herbaspirillum			1 (2.0%)			1 (0.6%)
Klebsiella	1 (50.0%)	10 (38.5%)	10 (19.6%)	7 (14.6%)	2 (7.1%)	30 (19.4%)
Mycobacterium		1 (3.8%)		2 (4.2%)	1 (3.6%)	4 (2.6%)
Pantoea			1 (2.0%)	1 (2.1%)	1 (3.6%)	3 (1.9%)
Pseudomonas	1 (50.0%)	3 (11.5%)	3 (5.9%)	6 (12.5%)	4 (14.3%)	17 (11.0%)
Raoultella		1 (3.8%)	1 (2.0%)			2 (1.3%)
Rhizobium		1 (3.8%)				1 (0.6%)
Roseomonas			1 (2.0%)			1 (0.6%)
Serratia		2 (7.7%)	1 (2.0%)	2 (4.2%)	2 (7.1%)	7 (4.5%)
Sphingomonas			1 (2.0%)			1 (0.6%)
Stenotrophomonas		1 (3.8%)	6 (11.8%)	12 (25.0%)	4 (14.3%)	23 (14.8%)
Totals	2	26	51	48	28	155

Table 4.3 Frequency of infection by organism (defined at species level) and year

Organism by species	2015	2016	2017	2018	2019	Total
Achromobacter spp.					1 (3.6%)	1 (0.6%)
Acinetobacter baumannii		1 (3.8%)	3 (5.9%)			4 (2.6%)
Acinetobacter baumannii complex			1 (2.0%)			1 (0.6%)
Acinetobacter ursingii		1 (3.8%)	2 (3.9%)	2 (4.2%)	1 (3.6%)	6 (3.9%)
Aeromonas hydrophila		1 (3.8%)				1 (0.6%)
Aeromonas spp.					1 (3.6%)	1 (0.6%)
Brevundimonas spp.			1 (2.0%)			1 (0.6%)
Burkholderia cepacia			1 (2.0%)	1 (2.1%)		2 (1.3%)
Chryseobacterium indologenes		1 (3.8%)	1 (2.0%)	1 (2.1%)	1 (3.6%)	4 (2.6%)
Chryseobacterium spp.				1 (2.1%)		1 (0.6%)
Citrobacter braakii			1 (2.0%)			1 (0.6%)
Citrobacter freundii			1 (2.0%)	1 (2.1%)		2 (1.3%)
Citrobacter koseri				1 (2.1%)		1 (0.6%)
Citrobacter youngae			1 (2.0%)			1 (0.6%)
Cupriavidus pauculus			1 (2.0%)	1 (2.1%)		2 (1.3%)
Delftia acidovorans			2 (3.9%)		1 (3.6%)	3 (1.9%)
Elizabethkingia meningoseptica		2 (7.7%)	1 (2.0%)			3 (1.9%)
Elizabethkingia miricola					1 (3.6%)	1 (0.6%)
Elizabethkingia spp.			2 (3.9%)			2 (1.3%)
Enterobacter cloacae		1 (3.8%)	7 (13.7%)	7 (14.6%)	6 (21.4%)	21 (13.5%)
Enterobacter cloacae complex				1 (2.1%)	2 (7.1%)	3 (1.9%)
Enterobacter cloacae ESBL				1 (2.1%)		1 (0.6%)
Enterobacter hormaechie			1 (2.0%)	1 (2.1%)		2 (1.3%)
Herbaspirillum spp.			1 (2.0%)			1 (0.6%)
Klebsiella oxytoca	1 (50.0%)	4 (15.4%)	2 (3.9%)	1 (2.1%)	1 (3.6%)	9 (5.8%)
Klebsiella pneumoniae		6 (23.1%)	8 (15.7%)	6 (12.5%)	1 (3.6%)	21 (13.5%)
Mycobacterium chelonae		1 (3.8%)		2 (4.2%)	1 (3.6%)	4 (2.6%)
Pantoea septica					1 (3.6%)	1 (0.6%)
Pantoea species			1 (2.0%)	1 (2.1%)		2 (1.3%)
Pseudomonas aeruginosa		1 (3.8%)	1 (2.0%)	5 (10.4%)	2 (7.1%)	9 (5.8%)
Pseudomonas putida	1 (50.0%)	2 (7.7%)	1 (2.0%)	1 (2.1%)	2 (7.1%)	7 (4.5%)
Pseudomonas stutzeri			1 (2.0%)			1 (0.6%)
Raoultella planticola		1 (3.8%)	1 (2.0%)			2 (1.3%)
Rhizobium radiobacter		1 (3.8%)				1 (0.6%)
Roseomonas mucosa			1 (2.0%)			1 (0.6%)
Serratia liquefaciens				1 (2.1%)		1 (0.6%)
Serratia marcescens		2 (7.7%)	1 (2.0%)	1 (2.1%)	2 (7.1%)	6 (3.9%)
Sphingomonas paucimobilis			1 (2.0%)			1 (0.6%)
Stenotrophomonas maltophilia		1 (3.8%)	6 (11.8%)	12 (25.0%)	4 (14.3%)	23 (14.8%)
Totals	2	26	51	48	28	155

In the following sections, we briefly consider the frequencies and distributions of bacteraemias caused by 4 particularly common GNE species groups.

4.3.1 Enterobacter spp.

In total, there were 38 bacteraemias in 31 children. In 2017, between mid-July and mid-December, there were 7 episodes in 6 children. In 2018, of the 10 affected children, all occurred in a 6-month period (February–August 2018). Similarly, in 2019, 11 children had bacteraemias, but none after May 2019.

4.3.2 Stenotrophomonas spp.

21 bacteraemias occurred in 19 children. There were 12 episodes of S. maltophilia bacteraemia in 11 children during 2018, but none after September 2018, until the first of 5 episodes in 5 children between April - September 2019.

4.3.3 Klebsiella spp.

22 children had a Klebsiella spp. bacteraemia. In 2016, there were 9 episodes affecting 8 children; all except one of these bacteraemias occurred in between June-November 2016. In 2017, 9 bacteraemias occurred in 7 children, with all except one occurring in a 5-month period (July-December). In 2018, 6 children had a Klebsiella spp. bacteraemia, 5 of which occurred between late January and mid-May.

4.3.4 Pseudomonas spp.

16 bacteraemias occurred in 14 children; in 2018, all 5 episodes (in 4 children) occurred between February-June. Similarly, in 2019, there were 4 bacteraemias in 4 children; with respect to time, there were two pairs, one five days apart in March and the others 16 days apart in June.

4.3.5 Conclusions

The above observations demonstrate two notable points. Firstly, while it is not possible to state this with certainty, the frequency of these bacteraemias caused by GNE appears to be higher than would be expected, particularly for the infections caused by Enterobacter spp. and Stenotrophomonas spp.. As Klebsiella spp., and Pseudomonas spp are the second and third most common Gram-negative bacteria (after Escherichia coli) causing blood stream infections, it is less clear that the frequencies of these two bacteria are higher than would normally be expected.

The second notable point is the clustering of bacteraemias in time; by virtue of this Review they are all broadly clustered in place. We consider the chances of the cluster patterns identified above occurring by chance is small.

Thus, we conclude from this simple analysis of the epidemiology of a large proportion of the bacteraemias in this Review that there is evidence for both increased frequency of specific GNE bacteraemia and episode clustering in time (and place). Neither phenomena prove that some of the bacteraemias had hospital environment sources, but the observations are consistent with this hypothesis.