Clinical Infection in Practice 1 (2019) 100003

Contents lists available at ScienceDirect

Clinical Infection in Practice

journal homepage: https://www.journals.elsevier.com/clinpr

Case Reports and Series

Clinicopathologic findings in children who died following Klebsiella

pneumoniae bloodstream infection
Heloise Buysa,b,*, Rudzani Muloiwab,c, Gavin L. Williamsa,d, Brian Eleya,b, Komala Pillaya,d,e
a
Red Cross War Memorial Children's Hospital, Klipfontein Road, Rondebosch, Cape Town, South Africa, 7700
b
Department of Paediatrics and Child Health, ICH Building Red Cross War Memorial Children's Hospital, Klipfontein Road, Rondebosch, Cape Town, South Africa, 7700
c
Groote Schuur Hospital, Anzio Road Observatory, Cape Town, South Africa, 7925
d
National Health Laboratory Services, Groote Schuur Hospital, Anzio Road, Observatory, Cape Town, South Africa, 7925
e
University of Cape Town, South Africa

A R T I C L E I N F O A B S T R A C T

Article History: Background: Klebsiella pneumoniae bloodstream infection (KPBSI) is strongly linked to hospitalisation and can
Received 6 March 2019 lead to significant morbidity and mortality. There is little data describing the autopsy findings of KPBSI in
Received in revised form 15 July 2019 children.
Accepted 30 July 2019 Methods: We conducted a retrospective review of clinicopathologic findings of children who died during the
course of KPBSI at a children's hospital in Cape Town, South Africa.
Results: Fifteen hospitalised children who died and had autopsies were included in this analysis. Their
Keywords: median age (interquartile range, IQR) was 4 (1 22) months, 2 (20%) were HIV-infected, and 10 (67%) were
Klebsiella pneumoniae bloodstream infection
moderately or severely underweight. The median time to death from the time of positive blood culture was 1
Children
(IQR 0 5) day. Notable autopsy findings were sepsis-induced multiorgan failure in many cases including thy-
Africa
Autopsy/postmortem mic involution, oesophageal erosions, adrenal haemorrhage, acute tubular necrosis and lung parenchymal
destruction. Other histologic findings included disseminated intravascular coagulopathy, necrotising inflam-
mation in many organs; and in four of five children biofilm formation.
Conclusion: KPBSI may cause widespread necrotic tissue.
© 2019 The Authors. Published by Elsevier Ltd on behalf of British Infection Association. This is an open
access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Introduction including incidence risk, clinical presentation, risk factors associated

with extended-spectrum beta-lactamase producing KPBSI, and fac-
Klebsiella pneumoniae bloodstream infection (KPBSI) is a life- tors associated with 30-day inpatient mortality.1 The current study is
threatening condition that is strongly linked to hospitalisation glob- a sub-study of the original KPBSI study in which we describe the
ally. There has been an increase in the number of publications autopsy findings of children who died as a result of KPBSI.
describing the morbidity and mortality caused by this infection in The study was conducted at RCWMCH, a tertiary referral facility
children.1 6 Although paediatric autopsies have been reported as serving a population of approximately 1.5 million children less than
useful in identifying the cause of death in fatal influenza infection,7 14 years of age. During the original study, 109 children died within
there is little published data if any, focusing exclusively on the 30 days of KP BSI. Autopsies were performed on fifteen of these chil-
autopsy findings in children dying from KPBSI. In this report we dren, comprising gross anatomic and histopathological examinations.
describe the clinicopathological features of 15 fatal cases. The decision to request an autopsy examination was made at the dis-
cretion of the attending clinician.
Clinical information and laboratory results were sourced from the
Methods
patients' clinical notes and National Health Laboratory Services
(NHLS) database respectively.
We previously described aspects of KPBSI in 410 children who
The NHLS histopathology service at RCWMCH performed the autop-
were hospitalised in Red Cross War Memorial Children's Hospital
sies. The findings were reviewed by a senior specialist anatomic pathol-
(RCWMCH), Cape Town, South Africa between 2006 and 2011,
ogist (K. P.), who standardised the reporting terminology. Tissue
samples were sectioned, fixed with formalin, paraffin-embedded and
* Corresponding author at: Room 508, 5th Floor, ICH Building, Department of Paedi-
then examined histologically following routine staining with haematox-
atrics and Child Health, Red Cross War Memorial Children's Hospital, Klipfontein Road,
Cape Town 7700, South Africa. ylin and eosin (H&E). Bacterial presence was confirmed using Gram
E-mail address: [email protected] (H. Buys). staining and Sandiford's counterstain, while Periodic-acid-Schiff (PAS)

https://doi.org/10.1016/j.clinpr.2019.100003
2590-1702/© 2019 The Authors. Published by Elsevier Ltd on behalf of British Infection Association. This is an open access article under the CC BY-NC-ND license.
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
2 H. Buys et al. / Clinical Infection in Practice 1 (2019) 100003

and Grocott histochemical staining were used to detect the presence of study was approved by the Human Research Ethics Committee, Fac-
fungi.8 In cases where exuberant bacterial colonies were observed in ulty of Health Sciences, University of Cape Town (HREC REF: 367/
tissue specimens, an Alcian Blue staining method was used to detect 2009) and the research committee of RCWMCH. The study was com-
the presence of biofilm. This method involved Alcian Blue staining pleted in accordance with the Declaration of Helsinki.
applied at pH 2.5 followed by Sandiford's staining.9 Focal biofilm was
used to describe the presence of biofilm when the Alcian blue staining Definitions
was taken up by < 50% of the bacterial colonies; the term diffuse bio-
film was used when >50% of the bacterial colonies took up the Alcian HIV infection: a positive HIV DNA PCR result confirmed by either a
blue stain. Disseminated intravascular coagulation (DIC) was diagnosed HIV RNA PCR or repeat HIV DNA PCR test, in any child <18 months
when disseminated fibrin thrombi were seen in the small blood vessels old, or 2 positive serological test results (HIV ELISA or HIV Rapitest)
within tissues. or a positive HIV DNA PCR result confirmed by either a HIV RNA PCR
Parental consent was obtained to conduct the autopsies in all or repeat HIV DNA PCR test, in a child > 18 months old were consid-
cases and complete diagnostic autopsy was the standard of care. In ered HIV-infected Unknown HIV status: any infant of child where
eight patients, the brain and spinal cord were not examined because there was no record of HIV testing at the NHLS laboratory database
the parents specifically requested that these organs be left intact. The and whose mother's HIV status was unknown.10 Moderate and severe

Table 1
Characteristics of 15 children who died following Klebsiella pneumoniae bloodstream infection and had an autopsy.

Age (m) Gender WFA HIV status Admission Comorbid ESBL-KPBSI Antibiotic Peripheral Indwelling Time to Time to
at BSI z-score diagnosis condition exposure Intravenous catheter BSI from death
diagnosis pre-BSI* catheter (CVP or admission from
when BSI urinary) (d) BSI (d)
diagnosed when BSI
diagnosed

Case 1 0.2 M -1 unknown Renal IODM, congen- N nil Y Y 4 0

impairment ital hydro-
nephrosis,
PDA, VSD
Case 2 1.3 F 0 infected Sepsis, HIV disease Y ceftriaxone, Y N 4 0
pneumonia cloxacillin
Case 3 1.2 M -1 infected Pneumonia HIV disease Y ceftriaxone, Y N 4 0
ampicillin,
gentamicin,
cotrimoxazole
Case 4 1.2 M -2 uninfected AGE None Y ceftriaxone, Y Y 14 6
amikacin,
metronidazole
Case 5 3.3 M 3 uninfected AGE, shock CMV infection Y ceftriaxone, Y N 3 0
ampicillin,
gentamicin,
cotrimoxazole
Case 6 3.4 F 1 unknown Measles, None Y ampicillin, Y N 3 0
pneumonia gentamicin
Case 7 3.1 M 3 unknown Pneumonia, None Y cefuroxime, Y Y 6 1
PHPT ampicillin,
gentamicin
Case 8 3.7 F < 3 uninfected Chronic DD Cystic fibrosis Y ceftriaxone, N N 0 2
cotrimoxazole
Case 9 5.5 F 2 uninfected Kwashiorkor None Y ceftriaxone, N N 0 1
cefuroxime
Case 10 13.3 F 4 uninfected Pneumonia Short bowel Y ceftriaxone, Y Y 177 5
syndrome meropenem,
piptazobac-
tam,
ciprofloxacin,
amikacin
Case 11 8 F 3 uninfected AGE, SAM None Y ampicillin, Y N 8 5
gentamicin,
cloxacillin,
erythromycin
Case 12 21.8 F 4 unknown vomiting and Caustic soda- N nil N N 0 5
dehydration induced
oesophageal
stricture
Case 13 49.2 F < 3 uninfected JDM JDM Y nil Y Y 18 6
Case 14 72.4 M 3 unknown TEN None Y ceftriaxone, Y Y 20 0
meropenem,
vancomycin
Case 15 80.5 M 1 uninfected Leukaemia Leukaemia in Y cefuroxime, Y Y 13 0
relapse remission azithromicin,
piptazobac-
tam,
meropenem

Legend: m- months; d- days; WFA- weight-for-age; BSI- bloodstream infection; ESBL-KP- extended-spectrum-b-lactamase-producing Klebsiella pneumoniae; neg-negative; pos-
positive; IODM- infant-of-diabetic-mother; PDA- patent ductus arteriosus; VSD- ventricular septal defect; AGE- acute gastroenteritis; PHPT- pulmonary hypertension; DD- diar-
rhoeal disease; SAM- severe acute malnutrition; JDM- juvenile dermatomyositis; TEN- toxic epidermal necrolysis; CMV- cytomegalovirus; CVP-central venous catheter at the
time of the BSI; Y-yes; N-no; *Antibiotic exposure within 30 days preceding the BSI.
 H. Buys et al. / Clinical Infection in Practice 1 (2019) 100003 3

underweight were defined as weight-for-age z score (WAZ) between

Pneumonia
2 and 3 standard deviations (SD) below the median World

Case 15

Legend: BSI- bloodstream infection; UTI: urinary tract infection; GI- upper gastrointestinal; thrombocytopenia-platelet count <150 £ 109/L; anaemia-haemoglobin <11 g/dl; leukopenia- white blood cell count below the reference range for age.
Health Organisation (WHO) growth reference standards, and a

+
WAZ < 3 SD respectively. Healthcare-associated infection: KPBSI
detected within 48 h of hospital admission in children who have had

Case 14
Sepsis
contact with the healthcare service including admission to an inter-

+
+
mediate care facility during the 12-month period preceding
hospitalisation.11

haemorrhage,
Pneumonia,
upper GI

wounds
Case 13

septic
Results

+
+

+
The age at KPBSI diagnosis of the fifteen children ranged from

pneumonia

not done
not done
not done
6 days to 80 months. Seven were male (47%). Ten (67%) were moder-

Case 12
Sepsis,
ately or severely underweight-for-age (UWFA). Two were HIV-
infected and seven were HIV-uninfected, four of whom were known

+
to have been perinatally HIV-exposed. Six children had not been

Pneumonia,
tested for HIV infection including one who was known to be perina-

empyema
Case 11
tally exposed. Characteristics of the fifteen children are shown in
Table 1. Four of the children had no previous hospital admission.

+
+
+

+
+
+
Eight (53%) had been hospitalised in the previous twenty-eight days.
Nine of the fifteen (60%) children had an underlying comorbid condi-

Pneumonia
Case 10
tion.
Bloodstream infection was caused by ESBL-producing KP in 13/15

+
+
+

+
+
+
+
(87%) children and all infections were healthcare-associated. The
two children who had non-ESBL-producing KPBSI had not had any

Pneumonia
prior antibiotic exposure. Seven of the remaining thirteen children

Case 9
had been treated with antibiotics in the previous 12-month period
including five who received antibiotics in the prior 28-day period.

+
+
+
+
+
+
Ten children had been previously exposed to cephalosporin antibiot-
Case 8
Septic
shock
ics Table 1. Fourteen of the fifteen children developed KPBSI during

+
+
+

+
+

+
the hospital's peak diarrhoeal and respiratory disease season months
between February and August, when overcrowding is frequent.
Pneumonia

The median time from admission to the KPBSI was 4 (IQR 3 14)
Case 7

days. All children died within six days of the positive KP blood cul-
ture. Diagnosis at the time of KPBSI and consequent complications

+
are summarised in Table 2. At the time of the KPBSI, septic shock
Pneumonia,

was present in 13/15 cases.

measles
Case 6

Histological findings
+
+

+
+
The autopsy findings are summarised in Table 3.
Case 5

septic
shock
UTI,

Thymic involution was present in all but one case, histological

+
+
+
+
+
+

sections showed that thymic lobular structure was preserved but

Complications during the course of Klebsiella pneumoniae bloodstream infection.

marked lymphocyte depletion had occurred Fig. 1. In the lungs, the

abdomen

most prominent findings were necrotising or haemorrhagic pneu-

Case 4
Acute

monia and diffuse alveolar damage in fourteen children. Bacterial

+

+
+
+
+

colonies were evident in the lung sections of five children. In one

Pneumonia

child, H&E staining demonstrated inflammatory infiltrates in the

lung and pathognomonic multinucleated giant cell clumps of con-
Case 3

comitant measles pneumonia in the acute phase.

+
+
+
+
+
+
+
+

Apart from two children who showed congenital structural car-

diac conditions, the heart was normal in eight children, but in the
Pneumonia,
dysentery

remaining five children endomyocardial haemorrhage or degenera-

Case 2

tion was present. Four children had necrotising oesophagitis with

Candida spp. invasion. The brain, examined in seven children,
+

+
+
+
+
+
+

showed evidence of diffuse hypoxic injury but no evidence of bacte-

Case 1
Septic
shock

rial invasion. Six children had evidence of acute tubular necrosis

+
+
+

+
+

(ATN), one child had multiple renal abscesses and one with dissemi-
nated cytomegalovirus (CMV) co-infection had CMV infection of the
Thrombocytopenia

Respiratory failure
Clinical diagnosis at

renal epithelial and endothelial cells. Ten children had histological

the time of the BSI

Hepatic injury
Coagulopathy

evidence of DIC and nine of these children had frank adrenal hae-
Septic shock

Renal injury
Complications

Leukopenia

morrhages. Of ten children with moderate to severe underweight-

Anaemia

for-age (UWFA), six had evidence of hepatic macrovesicular steato-

Table 2

sis; five children had skin erosions, and splenic congestion was pres-
ent in six children.
4 H. Buys et al. / Clinical Infection in Practice 1 (2019) 100003

Table 3
Histologic and gross anatomic of children who died of Klebsiella pneumoniae bloodstream infection.

Case RS Heart GIT CNS GUS Skin Other

1 Intra-alveolar hae- Congenital PDA, Congested liver and Diffuse hypoxic neu- Bilateral cystic dys- Multiple testicular DIC
morrhages VSD spleen ronal injury with plasia and hydro- abscesses Bilateral adrenal hae-
Focal immaturity gliosis and brain nephrosis; features morrhage
immaturity compatible with Congested spleen
PUV Hypoplastic thymus
2 Extensive necrotising Normal Oral Candidiasis Not examined Numerous thrombi Punched out skin DIC
KP BPN Microvescicular steatosis. and bacterial lesions secondary to Involuted thymus
DAD Necrotising oesophagitis colonies DIC
and enterocolitis with
bacterial colonies
3 KP BPN, alveolar hae- Hydropic Haemorrhagic infarction/ Not examined ATN, mucosal Skin ulcers DIC, adrenal haemor-
morrhage degeneration necrosis of bowel haemorrhage rhage
DAD Depleted splenic
white pulp
Involuted thymus
4 Extensive KP BPN Normal Focal intestinal necrosis Not examined Mild lymphoplas- Macerated neck and DIC, adrenal haemor-
DAD with bacterial colonies macytic infiltrates buttock area with an rhage
Mild hepatitis and ulcerative wound in Lymphocyte depleted
chronic pancreatitis with the right cubital fossa lymph nodes
lymphoplasmacytic infil- Thymic involution
trates
Hepatic macrovescicular
steatosis
5 Confluent necrosis, Myocardial Necrotising oesophagitis Congested; calcified Multiple septic renal Nil Multiple splenic
multiple pyaemic haemorrhages Microvesicular steatosis endothelial cells infarcts / abscesses infarcts
abscesses Involuted thymus
Focal evidence of DAD
6 Extensive necrotising Subendocardial Oesophageal and gastric Not examined ATN Normal Adrenal haemorrhage
BPN:KP, measles and necrosis ulceration Involuted thymus
PJP evident And Candidiasis. Micro- DIC
Candida vesicular hepatic
DAD steatosis
Candida laryngitis
7 BPN, PHPT (hypertro- Hypertrophied right Mild hepatitis, Not examined Normal Nil Congested adrenals
phic arterial walls) ventricle Microvesicular steatosis congested spleen
Small ASD/PDA Incomplete rotation of Normal thymus
Small septal suben- small intestine
docardial infarct
8 BPN with Candida Normal Invasive candida, diffuse Hypoxic injury CMV infiltration Severe excoriation DIC, disseminated
DAD CMV infection CMV infection
Pulmonary Macro- and micro-vesic- PEM
haemorrhage ular hepatic steatosis Features suggestive of
Cystic fibrosis in the
pancreas and gallblad-
der
Congested spleen
Involuted thymus.
Congested adrenals
9 Diffuse alveolar dam- Flabby myocardium Diffuse macrovesicular Hypoxic injury Chronic bladder Excoriated nappy rash Adrenal haemorrhage
age, KP lobar hepatic steatosis, oeso- wall inflammation with Candida PEM
pneumonia phageal congestion Involuted thymus
10 Haemorrhagic pneu- Endomyocardial Haemorrhagic pancreati- Extensive multifocal ATN Nil DIC
monia, septic pulmo- haemorrhage tis cortical Focal congestion Congested spleen
nary emboli Mild microvesicular haemorrhages and haemorrhage Haemorrhagic adre-
Previous pulmonary hepatic steatosis nals
TB Severe thymic
involution
11 Extensive necrotising Normal Diffuse hepatic macrove- Normal ATN Napkin dermatitis DIC, adrenal haemor-
KP BPN+ empyema sicular steatosis (healed) rhage
DAD PEM
Pulmonary infarction Thymic involution
Right pneumothorax
12 Multilobar KP BPN, Normal Oesophageal ulcers, Normal ATN Nil DIC, adrenal haemor-
pulmonary fibrosis, gastritis with rhage
haemorrhage Candida, Thymic involution
Diffuse macrovesicular
hepatic steatosis
13 Haemorrhagic BPN Normal Acute primary peritonitis Not examined ATN Multiple septic skin Skeletal muscle
Haemorrhagic gastritis lesions positive for showed minimal
Moderate macrovesicular Klebsiella on micros- activity of DMS
hepatic steatosis copy and culture Spleen decreased
lymphoid tissue
Thymic involution

(continued)
 H. Buys et al. / Clinical Infection in Practice 1 (2019) 100003 5

Table 3 (Continued)

Case RS Heart GIT CNS GUS Skin Other

14 Focal haemorrhagic Normal Congested liver Not examined Normal Toxic epidermal nec- DIC
pneumonia rolysis 70% skin Congested spleen and
RML embolus +tongue and genitalia adrenals
Focal haemorrhagic Bilateral femoral vein
infarct thrombi
Thymic involution
15 KP BPN Normal Not examined Not examined Not examined Normal Enlarged hilar lymph
Interstitial pneumoni- nodes. Thymic involu-
tis, haemorrhage and tion
haemosiderin Abdominal cavity not
deposition examined
Legend: RCWMCH-Red Cross War Memorial Children's Hospital RS-respiratory system, GIT-gastrointestinal system, CNS-central nervous system, GUS-genitourinary system; KP-
Klebsiella pneumoniae; BPN-bronchopneumonia; DAD-diffuse alveolar damage; PHPT-pulmonary hypertension; DIC-disseminated intravascular coagulation; RML-right middle
lobe; ATN-acute tubular necrosis; PEM-protein energy malnutrition; CMV-cytomegalovirus; PJP-Pneumocystis jirovecii; ASD-atrial septal defect; PDA-patent ductus arteriosus;
VSD-ventricular septal defect; PUV-posterior urethral valves.

Biofilm formation following Klebsiella pneumoniae bloodstream infection. Klebsiella

pneumoniae (KP) frequently causes HAI, which is associated with
Of five children with exuberant bacterial colonies on tissue histol- high case-fatality in children.2 5 Our autopsy study describes find-
ogy sections, four (Cases 2, 3, 5 and 11) exhibited biofilm formation ings of widespread necrotising inflammation and disseminated intra-
in lung and bowel tissue sections (Figs. 2 4). Two were HIV-infected vascular coagulation. These findings are similar to autopsy findings
(Cases 2 & 3), diffuse biofilm formation was seen in the lung and reported in adult patients with Gram-negative sepsis.12
bowel tissue sections. Both experienced pancytopenia during the The bloodstream, lung and kidney are common sites for infection
course of KPBSI and their autopsy findings showed diffuse pneumo- with KP.12 A Bangladeshi study that described autopsy findings in
nia. The other two children were admitted with gastroenteritis and 119 children who were admitted to the International Centre for Diar-
hypovolemic shock, both were HIV-uninfected but were severely rhoeal Disease Research with pneumonia and/or gastroenteritis,
UWFA, and one had concomitant CMV infection; their lung and bowel reported five histologic pneumonia patterns in 93 children with fatal
tissue sections exhibited focal biofilm formation. pneumonia. These included acute alveolar exudate in 38 (41%),
The remaining child (Case 12) had exuberant bacterial colonies on necrotising pneumonia in 24 (26%), interstitial pneumonia in 7 (8%),
tissue section but did not demonstrate biofilm formation. This patient granulomatous pneumonia in 2 (2%) and eosinophilic pneumonia in
was a 22-month-old who had a history of caustic soda-induced oeso- 2 (2%). The pneumonia was most commonly due to Gram-negative
phageal stricture three months previously, and who was admitted bacterial infection, including Klebsiella pneumoniae, in 27% (25/93) of
with vomiting and dehydration to a surgical ward. She was severely cases.13 Most of these children were severely malnourished. In con-
malnourished with evidence of macrovesicular hepatic steatosis and trast, in the present study where ten of fifteen children were mal-
multi-lobar pneumonia. nourished, the three main patterns seen were acute alveolar exudate
in ten cases, necrotising pneumonia in four and alveolar haemor-
Discussion rhage in one case; interstitial pneumonia was seen in only one case.
Additional pathology to the pneumonia included associated pulmo-
We have described the clinicopathologic findings in a case series nary haemorrhage in seven cases and diffuse alveolar damage in
of fifteen African children who underwent autopsy examination eight cases. There was no apparent relationship with age or HIV

Fig. 1. Thymus tissue sections: Legend: A). Relatively normal thymus (from the only child who did not have thymic involution). Outer distinct cortex (darker staining, with densely packed
immature lymphocytes) and medulla- paler staining with less densely packed mature lymphocytes B). Severely involuted thymus with effacement of the corticomedullary junction.
6 H. Buys et al. / Clinical Infection in Practice 1 (2019) 100003

Fig. 2. Patient 2 Lung tissue: (grey hepatisation stage). C). H&E Low power and D). high power showing exuberant bacterial colonies. E). Sandiford's counterstain technique (modi-
fied Gram stain) enhancing the Gram-negative bacterial colonies F). Alcian blue-Sandiford's staining technique showing focal biofilms. (For interpretation of the references to colour
in this figure legend, the reader is referred to the web version of this article.)

infection- the patient numbers were too small to generalise, further- of generalised malnutrition. In the face of protein-energy malnutri-
more the autopsy findings in children with HIV infection were similar tion, zinc deficiency and infection, the thymus is also known to show
to those in children without HIV infection. consistent severe atrophic changes with massive thymocyte apopto-
Malnutrition threatens child survival but can also influence histol- sis.15 Tomashefski et al. commented that although pneumococcus,
ogy findings. Six of the moderately-severely malnourished children Haemophilus influenzae, and common respiratory viral pathogens
who died had typical hepatic macrovesicular steatosis changes of (influenza, adenovirus, respiratory syncytial virus and parainfluenza
kwashiorkor, described by Cicely Williams in 1933.14 Of these six virus) are reported as important causes of mortality from acute respi-
children, three had oedematous malnutrition that would be consis- ratory infection (ARI), the role of Gram-negative bacteria associated
tent with a diagnosis of kwashiorkor; however, the other autopsy ARI may be less well appreciated. Hard evidence from histopatholog-
findings of children with malnutrition were similar to those of chil- ical studies in children with malnutrition complicated by pneumonia
dren without malnutrition. Thymic involution was present in four- and diarrhoea is lacking; this is attributed to the paucity of paediatric
teen children and is likely a consequence of overwhelming systemic autopsy studies from low- and middle-income countries.13 Erosive
infection (experienced by all fifteen children) with the added effect skin changes were more commonly seen in malnourished study

Fig. 3. Patient 3 Lung tissue (grey hepatisation stage). G). Lung tissue: H&E, low power H). Lung tissue: H&E high power showing exuberant bacterial colonies, I). Sandiford's coun-
terstain technique (modified Gram stain) enhancing the Gram-negative bacterial colonies J). Alcian blue-Sandiford's staining technique showing diffuse biofilms. (For interpretation
of the references to colour in this figure legend, the reader is referred to the web version of this article.)
 H. Buys et al. / Clinical Infection in Practice 1 (2019) 100003 7

Fig. 4. Patient 2, Small bowel tissue. K). Small bowel: H&E low power; L). H&E high power showing exuberant bacterial colonies with mucoid appearance. M). Sandiford's counter-
stain showing cerise Gram-negative bacteria. N). Alcian blue-Sandiford's stain showing diffuse biofilms. (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article.)

children, five out of ten malnourished children had erosive skin In the present study, the antemortem clinical diagnoses correlated
lesions versus one out of five children without malnutrition. In our well with the autopsy findings of profound sepsis. Another benefit of
original study on KPBSI, we showed that erosive skin changes were autopsy examination was evidenced in our study where there were
negatively linked to survival.1 two children (cases 1 & 7) who were shown to have previously undi-
The exact pathogenesis of adrenal haemorrhage in septicaemia is agnosed congenital anomalies.
unknown, but it is related to endotoxin production and development
of shock. Adrenal haemorrhage was documented in 60% (39 of 65) of Conclusion
patients with fatal bacterial infection caused by a spectrum of patho-
gens including Neisseria meningitidis, group A streptococci, Strepto- Our study documents that Klebsiella pneumoniae causes destruc-
coccus pneumoniae, Rickettsia rickettsii, and Staphylococcus aureus.16 tive necrotising inflammation and that autopsy studies can provide
In the present study, fatal KPBSI caused frank adrenal haemorrhage good evidence of the extent of the systemic infection allowing corre-
in eight (53.5%) children and adrenal congestion in a further four lations between antemortem clinical and autopsy diagnoses and
(26.7%) children. Of these twelve children, all experienced septic cause of death. Biofilm formation may increase the risk of antibiotic
shock during KPBSI apart from one child (case 7) who had adrenal treatment failure by conferring a survival advantage to the aggre-
congestion but no septic shock. gates of bacteria that produce biofilm.
Biofilms are surface-attached aggregates of microbial cells embed-
ded in a self-produced extracellular polymetric matrix. Biofilms can Declaration of competing interest
develop in many infections including those of the respiratory, gastro-
intestinal and urinary tracts. During colonisation or invasive infec- The authors declare no conflicts of interest.
tion, biofilm may also form on indwelling medical devices. Infections
associated with biofilm formation are difficult to cure because the
Acknowledgements
encased microbes are partially protected from the immune response,
may resist the action of antimicrobial agents, and/or employ toler-
We gratefully acknowledge our medical records department, our
ance mechanisms to evade the action of antimicrobial agents.17 In
doctors and nursing colleagues and mostly, the children we serve.
this study, the use of Alcian blue stain to detect biofilm was novel
and was only used when Haematoxylin and eosin staining showed
the presence of exuberant bacterial proliferation in five of the fifteen References
autopsies; the stain was not used in the remaining 10 autopsies
1. Buys H, Muloiwa R, Bamford C, Eley B. Klebsiella pneumoniae bloodstream infec-
specimens where the bacterial colonies were not seen. This was pre- tions at a South African children's hospital 2006-2011, a cross-sectional study. BMC
viously described in two children from the same institution in the Infect Dis 2016;16.
context of necrotising enterocolitis.18 2. Velaphi S, Wadula J, Nakwa F. Mortality rate in neonates infected with extended-
spectrum beta lactamase-producing Klebsiella species and selective empirical use of
The present study was limited by its retrospective nature and
meropenem. Ann Trop Paediatr 2009;29(2):101–10.
small sample size from a single site, despite this, good histologic data 3. Tanir Basaranoglu S, Ozsurekci Y, Aykac K, Karadag Oncel E, Bicakcigil A, Sancak B,
was available; however, generalizability would be problematic. There et al. A comparison of blood stream infections with extended spectrum beta-lacta-
mase-producing and non-producing Klebsiella pneumoniae in pediatric patients. Ital
have been declining trends in autopsy requests despite the wealth of
J Pediatr 2017;43(1):79.
information, albeit too late for the benefit of the patient, that can be 4. Kuo KC, Shen YH, Hwang KP. Clinical implications and risk factors of extended-spec-
provided by autopsy studies.19 Generally speaking, our parents have trum beta-lactamase-producing Klebsiella pneumoniae infection in children: a case-
always been reluctant to consent to autopsy examination; addition- control retrospective study in a medical center in southern Taiwan. J Microbiol Immu-
nol Infect 2007;40(3):248–54.
ally, clinicians, rightly or wrongly, often regard the case closed after a 5. Kim Y-K, Pai H, Lee H-J, Park S-E, Choi E-H, Kim J, et al. Bloodstream infections by
blood culture report has confirmed the likely explanation for death. extended-spectrum b-lactamase-producing Escherichia coli and Klebsiella
8 H. Buys et al. / Clinical Infection in Practice 1 (2019) 100003

pneumoniae in children: epidemiology and clinical outcome. Antimicrob Agents 12. Dermengiu D, Curca GC, Ceausu M, Hostiuc S. Particularities regarding the etiology
Chemother 2002;46(5):1481–91. of sepsis in forensic services. J Forensic Sci 2013;58(5):1183–8.
6. Marra AR, Wey SB, Castelo A, Gales AC, Cal RG, Filho JR, et al. Nosocomial blood- 13. Tomashefski JF, BT, Islam M. Histopathology and aetiology of childhood pneumonia:
stream infections caused by Klebsiella pneumoniae: impact of extended-spectrum an autopsy study of 93 patients in Bangladesh. Pathology 1989;21:71–8.
beta-lactamase (ESBL) production on clinical outcome in a hospital with high ESBL 14. Williams CD. A nutritional disease of childhood associated with a maize diet. Arch Dis
prevalence. BMC Infect Dis 2006;6:24. Child 1933;8(48):423.
7. World Health Organisation. Updated guideline paediatric emergency triage, assess- 15. Savino W, Dardenne M, Velloso LA, Dayse Silva-Barbosa S. The thymus is a common
ment and treatment. Geneva World Health Organization; 2016. target in malnutrition and infection. Br J Nutr 2007;98(Suppl 1):S11–6.
8. Williams MD, BA. A modified Sandiford's counter stain for smears of blood cultures. J 16. Guarner J, Paddock CD, Bartlett J, Zaki SR. Adrenal gland hemorrhage in patients with
Clin Pathol 1980;33(8):794–5. fatal bacterial infections. Mod Pathol 2008;21(9):1113–20.
9. Bancroft JD, SA. Theory and practice of histological techniques. 2nd ed. New York: 17. Piperaki E_T, SG, Tzouvelekis LS, Daikos GL. Klebsiella pneumoniae: Virulence, bio-
Churchill Livingstone; 1982. p. 1982. film and antimicrobial resistance. Pediatric Infectious Disease Journal 2017;36
10. Schneider E, Whitmore S, Glynn MK, Dominguez K, Mitsch A, McKenna MT. (10):1002–5.
Revised surveillance case definitions for HIV infection among adults, adolescents, and 18. Brisighelli G, Cox S, Theron A, Pillay K, Rode H. Biofilms associated with bowel necro-
children aged < 18 months and for HIV infection and AIDS among children aged 18 sis: a newly recognised phenomenon in infants. S Afr Med J 2016;106(4):39–41.
months to <13 years—United States, 2008. MMWR Recomm Rep 2008;57(10):1–12. 19. Lewis C, Hill M, Arthurs OJ, Hutchinson C, Chitty LS, Sebire NJ. Factors affecting
11. Cardoso T, Almeida M, Friedman ND, Araga ~o I, Costa-Pereira A, Sarmento AE, et al. uptake of postmortem examination in the prenatal, perinatal and paediatric setting.
Classification of healthcare-associated infection: a systematic review 10 years after BJOG 2017;125:172–81.
the first proposal. BMC Med 2014;12(1):40.