Maternal sepsis

Marian KnightProfessor of Maternal and Child Population Health

National Perinatal Epidemiology Unit

Nuffield Department of Population Health

Objectives

• To describe the main causes of maternal

sepsis in the UK

• To identify women at greater risk of

maternal sepsis

• To highlight actions to prevent and treat

maternal sepsis

Anna’s story

Hinton L

Anna was 21 years old. She already had

one son at home. When she came home

with her new baby she started to feel

unwell.

Anna’s story

Hinton L

Maternal deaths from genital

tract sepsis

Test for trend p=0.01

Maternal deaths from genital

tract sepsis

0.0

0.5

1.0

1.5

1985-87 1988-90 1991-93 1994-96 1997-99 2000-02 2003-05 2006-08

Rate

per

100,0

00 m

ate

rnit

ies

Source: Lewis G et al. Saving Mothers Lives 2006-8. BJOG 2011

Test for trend p=0.01

UK national study of severe sepsis

associated with pregnancy

• Affected women were identified between

2011 and 2012

• Information collected on the source of

infection, management, and the

subsequent consequences to both mother

and baby

0.0

0.5

1.0

1.5

1985-87 1988-90 1991-93 1994-96 1997-99 2000-02 2003-05 2006-08

Rate

pe

r 1

00

,00

0 m

ate

rnit

ies

Sepsis morbidity

• 365 women diagnosed with severe sepsis

associated with pregnancy over one year

(47 per 100,000 maternities)

• 71 women developed septic shock

(9 per 100,000 maternities)

• 5 women died

• For every woman who dies, 14 have septic

shock; a further 60 have severe sepsisAcosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Sources of severe sepsis

Genital tract 31%

Urinary tract20%

Wound9%

Respiratory5%

Other9%

Unknown26%

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

37% (N=134) before delivery

UTI (34%)

63% (N=231) after delivery

Genital-tract (37%)

(P<0.0001)

Sources of sepsis – intensive care

Pregnant or recently pregnant women admitted to

UK intensive care units 2008-2010 (ICNARC data) Source of infection

Number of women

(n=646) Pneumonia/ respiratory infection 257 (39.8)

Genital tract 157 (24.3)

Urinary tract/kidney 59 (9.1)

Surgical trauma 24 (3.7)

Septicaemia 20 (3.1)

Appendicitis 19 (2.9)

Other infection 62 (9.6)

Unknown 48 (7.4)

Acosta C et al. 2016 (BMJ Open; 6(8):e012323)

Sources of infection

• Genital tract infection (classic “puerperal

fever”) forms only a small proportion of

maternal morbidity from infectious disease

• Maternal sepsis should be considered as a

wider definition

• Consideration of all types of infection is

important in training, when planning

management and services

Causative organisms

Escherichia coli 21%

Group A streptococcus

9%

Group B streptococcus

8%

Other streptococcus

6%Staphylococcus6%

Mixed organisms5%

Other7%

Unknown2%

No laboratory confirmed infection

36%

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Causative organism by source

of infection

0

20

40

60

80

100

120

Nu

mb

er

of

ca

se

s

Source of infection

No Lab confirmed infection

Unknown

Other

Mixed

Staph

Other strep

Group B strep

Group A strep

E.coli

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Causative organism by mode of

delivery

0

20

40

60

80

100

120

Spontaneousvaginal

Operativevaginal

Pre-labourcaesarean

Caesarean afterlabour onset

Nu

mb

er

of

ca

se

s

No Lab confirmed infection

Unknown

Other

Mixed

Staph

Other strep

Group B strep

Group A strep

E.coli

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Rapid progression to severe

sepsis

• <24 hours between the first signs of an inflammatory

response to infection and sepsis:

– 83% of cases and 85% of septic shock cases

• <48 hours between the first signs of an inflammatory

response to infection and sepsis:

– 89% of cases and 95% of septic shock cases

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Causative organism according to

whether septic shock diagnosed

0

5

10

15

20

25

30

35

40

45

50

E.coli Group Astrep

Group Bstrep

Other strep Staph Mixed Other Unknown No Labconfirmedinfection

Pro

po

rtio

n o

f w

om

en

(%

)

Septic shock No shock

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Rapid progression to severe

sepsis for women with Group A

streptococcal infection

• 50% <2 hours between the first signs of an inflammatory

response to infection and sepsis diagnosis

• 75% <9 hours between the first signs of an inflammatory

response to infection and sepsis diagnosis

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Significant medical risk factors

Cases n (%)

Controls n (%) aOR* 95% CI

n=365 n=757

Parity

0 197 (54) 330 (44) 1.6 (1.2-2.2)

≥1 167 (46) 427 (56) 1

Pre-existing medical problems

Yes 120 (33) 171 (23) 1.4 (1.0-1.9)

No 245 (67) 583 (77) 1

Febrile illness or antibiotics in 2 wks before delivery

Yes 153 (42) 42 (6) 12.1 (8.1-18.0)

No 212 (58) 715 (94) 1

*Adjusted for all other factors examined

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Significant delivery risk factors

Postpartum cases n (%)

Controls n (%) aOR* 95% CI

n=302 n=757

Mode of delivery

Spontaneous vaginal 57 (21) 443 (59) 1

Operative vaginal 39 (14) 100 (13) 3.4 (1.7-7.0)

Pre-labour caesarean 67 (25) 119 (16) 3.5 (2.0-6.1)

Caesarean after labour onset 108 (40) 92 (12) 6.7 (3.8-12.0)

Complications of delivery

Yes 103 (34) 279 (37) 1.7 (1.1-2.5)

No 199 (66) 478 (63) 1

*Adjusted for all other factors examined

Acosta, Kurinczuk, Lucas at al 2014 (PLoS Med 11(7):e1001672)

Significant factors associated

with mortality in intensive careSurvivors

n (%)Deaths n (%) aOR* 95% CI

n=610 n=29

Maternal age

<25 234 (38) 5 (17) 1

25-34 254 (42) 15 (52) 2.2 (0.7-7.0)

≥ 35 121 (20) 9 (31) 3.3 (0.9-11.0)

BMI

Unknown 317 (52) 13 (45) 1.2 (0.2-9.1)

<25 126 (21) 3 (10) 1

25-29.9 90 (15) 7 (24) 5.2 (1.4-18.9)

≥ 30 76 (13) 6 (21) 6.3 (1.5-27.0)

IMD Quintiles 4&5 354 (58) 17 (58) 2.6 (1.0-6.7)

*Adjusted for all other factors examined

Acosta C et al. 2016 (BMJ Open; 6(8):e012323)

Causes of maternal death

UK 2009-12

Causes of maternal death

UK 2009-12

Influenza deaths 2009-12

Causes of maternal death

2009-12• Major contribution from influenza and

other non-genital tract sepsis deaths

• Overall rate of maternal mortality from

infectious causes in 2009-12 was 2.0 per

100,000 maternities (95%CI 1.6-2.6)

Maternal sepsis deaths 2000-

2012

0.00

0.50

1.00

1.50

2.00

2.50

3.00

2000-02 2003-05 2006-08 2009-12

Rate

per

100,0

00 m

ate

rnit

ies

Sources: CMACE and MBRRACE-UK

Key actions in suspected sepsis

• The key actions for diagnosis

and management of sepsis

are:

1. Timely recognition

2. Fast administration of intravenous antibiotics

3. Quick involvement of experts - senior review is

essential

NHS England Patient Safety Alert NHS/PSA/R/2014/015 (NHS England 2014)

Vignette: sepsis

Two hours after delivery a woman became unwell on the

postnatal ward feeling faint. Her oxygen saturation was

65%. She was reviewed by junior staff and found to be

shocked, with moderate PV bleeding. Her temperature

was never measured. A diagnosis of haemorrhage was

made and she was treated with fluids. She failed to

improve and was taken to theatre where she had a cardiac

arrest. A laparotomy and hysterectomy were carried out but

resuscitation failed. The postmortem found an extensive

blistering skin rash, swollen labia and disseminated

intravascular coagulation all as a result of overwhelming

Group A Streptococcal sepsis.

Antibiotic administration among

women who died 2009-12

Symptom onset Diagnosis Antibiotics Death

43 40 37

3

<1 hour

3

1 – 2 hours

2 – 3 hours

3 – 24 hours

< 24 hours

(missing time)

> 24 hours

Death at home or on arrival

Never commenced on antibiotics

Progression of disease

Tim

e

14

7

4

4

3

5

Mohamed-Ahmed O et al. 2015 BJOG;122(11):1506-15.

43 women died

from bacterial

sepsis

14 (33%)

received

antibiotics within

one hour

Vignette; sepsis

A woman who was seven days post spontaneous vaginal

delivery became unwell at home with a fever. She was advised

to attend the maternity unit immediately. On admission she

was noted to be tachycardic, tachypnoeic and febrile. She

was prioritised for urgent medical review. A diagnosis of

acute sepsis from retained products was made and fluid

resuscitation started immediately. Intravenous antibiotics were

started within one hour of the diagnosis and she was transferred

to the high dependency unit. The retained products of

conception were removed promptly and she made a full

recovery. Blood culture subsequently grew Klebsiella. Early

recognition, clear advice and prompt treatment led to a good

outcome without any further complications.

Screening and action tools for women who

are pregnant or who have given birth within

the last six weeks

https://sepsistrust.org/education/clinical-tools/

MATERNAL DEATHS FROM

SEPSIS – WHERE ARE WE

NOW?

Maternal mortality rate due to sepsis per 100,000 maternities, UK 2013-15

But..A woman with a large for dates baby was admitted at term in early

labour. CTG abnormalities were not recognised or responded to and an intrauterine death occurred. She then developed signs of sepsis and

acute kidney injury during labour, which were not promptly recognised. A sepsis bundle was not initiated. After a prolonged second stage she

had a massive haemorrhage, developed coagulopathy and died. Although her death was probably multifactorial, sepsis was the

underlying cause of her deterioration. Reviewers noted several aspects of her resuscitation that could have been improved. Senior staff were

not involved until after her collapse.

After her death it was thought amniotic fluid embolus was implicated, although this was reported as a secondary phenomenon by the

pathologist at postmortem. This led to a superficial review of her care, significantly limiting the lessons that should have been learnt from her death. A pathologist, who would have provided a better understanding

of the post-mortem findings, was not involved in the review.

WHO Global maternal sepsis

study• Over one week in the West Midlands 93

women were identified across 8 hospitals

who were treated for infection associated

with pregnancy

• Nationally, 82 maternity units reported that

they had no women with suspected or

confirmed pregnancy-associated infection

in the same week

Key points

• Genital tract infection forms only a small

proportion of maternal mortality and morbidity

from infectious disease

• Women with pre-existing medical problems and

women undergoing operative delivery are at

higher risk of sepsis morbidity

• Obesity, older maternal age and deprivation are

associated with mortality

• Symptoms of sepsis, particularly postnatally,

may be interpreted as part of the norm

Anna’s story

Hinton L

Anna had a life-saving hysterectomy and

after several weeks in hospital, she came

home to her sons. But being the Mum she

wanted to be was hard after being so

gravely ill.

Anna’s story

Hinton L

Acknowledgements and funding

MBRRACE-UK•Department of Health, England

•NHSSPS Northern Ireland

•Scottish Government Health Department

•NHS Wales

•Channel Islands and the Isle of Man Government Offices

The contract is managed on behalf of the funding bodies by the

Healthcare Quality Improvement Partnership (HQIP)

ICNARC/UKOSS studies•NIHR PGfAR And thanks to numerous

co-authors and

collaborators!