Official Statistics

30 day all-cause mortality following MRSA, MSSA and Gram-negative bacteraemia and C. difficile infections: 2021 to 2022 report

Updated 9 November 2023

Main findings

When examining mortality rates over the surveillance periods, there has been:

• a substantial decline in the mortality rate of methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia and Clostridioides difficile infections (CDI), especially for hospital-onset (HO) cases, since the start of surveillance in April 2007

• a particularly evident improvement in the mortality rate for MRSA and CDI before 2015 that was coupled with a decrease in incidence of these infections

• a slowly increasing trend in the mortality rate of methicillin-susceptible Staphylococcus aureus (MSSA) bacteraemia since April 2011, primarily driven by an increase in the mortality rate of community-onset (CO) cases, with a sharper increase in mortality rate and CFR of all cases in the most recent 2 financial years

• a slowly increasing trend in the mortality of CO Escherichia coli (E. coli) bacteraemia between April 2012 and March 2020, which has since reversed; the mortality rate of HO cases has, on the other hand, been stationary since April 2012

• since surveillance started in April 2017 for Klebsiella species (Klebsiella spp.) bacteraemia, the mortality rate for CO cases has been relatively stationary, whilst the mortality rate for HO cases has decreased after the large 2020 increase

• a decrease in the mortality rate for Pseudomonas aeruginosa (P. aeruginosa) bacteraemia between financial years April 2017 to March 2018 and 2018 to 2020, but a peak in financial year 2020 to 2021; a subsequent decline in the most recent financial year was observed

This report presents data on the 30 day all-cause mortality following MRSA, MSSA, E. coli, P. aeruginosa, Klebsiella spp. bacteraemia and Clostridium difficile (C. difficile) infection.

All data tables associated with this report are available in the accompanying dataset for 2021 to 2022. Data in this report are presented as 30 day all-cause mortality rate and case fatality rates (CFRs).

Mortality rate is a widely used outcome for assessing risk of death. Mortality rate is the number of deaths divided by the population at risk. This reflects the incidence of all-cause deaths following these infections in the population.

Case fatality rate (CFR) is the number of deaths as a percentage of all reported cases. This provides a measure for comparing survivability of different infections.

This report presents data on all-cause mortality, and therefore includes deaths that may not be directly attributable to the infections.

Between the financial year April 2020 to March 2021 and April 2021 to March 2022, the CFR slightly decreased for all collections:

• E. coli bacteraemia from 16.0% to 15.0%
• Klebsiella spp. bacteraemia from 21.8% to 19.5%
• MSSA bacteraemia from 23.7% to 22.1%
• MRSA bacteraemia from 28.2% to 26.3%
• P. aeruginosa bacteraemia from 27.7% to 25.0%
• C. difficile infection from 14.9% to 13.7%

These reductions were primarily driven by a more substantial change in the CFR among HO cases (in particular, from 33.9% to 27.0% for P. aeruginosa bacteraemia and from 34.3% to 29.3% for MRSA bacteraemia).

Comparing across financial years April 2020 to March 2021 and April 2021 to March 2022, the mortality rate (expressed in deaths per 100,000 population per financial year) slightly decreased for:

• E. coli bacteraemia from 10.2 to 9.8
• Klebsiella spp. bacteraemia from 4.2 to 3.8
• P. aeruginosa bacteraemia from 2.0 to 1.9
• MSSA bacteraemia from 4.8 to 4.7

These improvements go against the increase seen in the previous 2 financial years. The mortality rate remained stable for MRSA bacteraemia (0.3), but it marginally increased for C. difficile infection (3.2 to 3.4); this is an ongoing trend that is reflective of an increasing incidence of CDI.

Between March 2020 and March 2022, England experienced 3 major waves of coronavirus (COVID-19) infection. During the early part of the pandemic, many elective procedures in hospitals were cancelled, leading to reduced hospital activity. During this decline in hospital activity, reductions were also observed in the number of Gram-negative and Staphylococcus aureus (S. aureus) bacteraemia, and CDI reported during this period.

A small increase in CFR was observed across all organisms in the first year of the COVID-19 pandemic followed by a decrease in the most recent year. It is unclear how much of the increase in CFR during the first pandemic year could be attributed to COVID-19 but 21% of CDI cases had a positive COVID-19 test in the 4 weeks before or after CDI positive sample at this time, as described in the previous edition of this report, covering 2020 to 2021.

Results for all mandatory surveillance infections

Data throughout this report are presented in financial years, with the years starting in April and ending in March. Hereafter, the month names will be omitted for brevity purposes, such that the most recent financial year will be described as 2021 to 2022, rather than April 2021 to March 2022.

During financial year 2021 to 2022, there were a total of 80,913 cases of E. coli, Klebsiella spp., P. aeruginosa, MRSA and MSSA bacteraemia, and CDI in England, 78,956 (97.6%) cases could be linked to NHS Spine data. This was an increase from the previous year’s total cases of 77,150 (financial year 2020 to 2021).

There were 13,501 deaths within 30 days of taking a specimen (blood culture for bacteraemia, faecal sample for CDI). The highest number of these deaths were associated with E. coli at 5,549 deaths. This is 41.1% of deaths covered by this report. The overall mortality rate was 23.9 deaths per 100,000 population, with a CFR of 17.1% (Table S1 of the supplementary dataset and Figure 1).

The number of deaths has decreased from financial year 2020 to 2021 (13,984 deaths) to the current financial year 2021 to 2022 (13,501 deaths). The mortality rate and CFR have decreased by 0.8 deaths per 100,000 population and 1.5%, respectively.

Figure 1. Thirty-day all-cause case fatality rate by infection, England, financial year 2007 to 2008 to financial year 2021 to 2022

Escherichia coli bacteraemia

During financial year 2021 to 2022, 37,965 E. coli bacteraemia cases were reported in England. Information on mortality was available for 97.6% (37,043) of these cases (Table S2 of the supplementary dataset). There were 5,549 deaths within 30 days of an E. coli bacteraemia diagnosis, indicating a mortality rate of 9.8 deaths per 100,000 population and a CFR of 15.0%.

There was a declining trend in CFR starting from 16.8% during financial year 2012 to 2013 to 13.9% during financial year 2018 to 2019, after which the CFR then increased to 16.0% by financial year 2020 to 2021. The mortality rate increased between financial years 2012 to 2013 and 2021 to 2022 from 9.7 to 9.8 deaths per 100,000 population. Mortality rate peaked at 10.7 during financial year 2019 to 2020 but has subsequently declined in the most recent 2 financial years.

Variation by onset of bacteraemia

The surveillance of CO mortality rate peaked during financial year 2019 to 2020 at 7.7 deaths per 100,000 population. This has subsequently decreased to 7.1 (Table S3 of the supplementary dataset and Figure 2). The mortality rate of HO cases increased to 4.8 deaths per 100,000 bed-days during financial year 2019 to 2020 and 5.3 deaths per 100,000 bed-days during financial year 2020 to 2021, coinciding with the COVID-19 pandemic. The mortality rate subsequently decreased in the financial year 2021 to 2022 to 4.7 deaths per 100,000 bed-days (Table S3 of the supplementary dataset and Figure 3).

The CFR of HO cases declined from 23.6% to 22.6% in between financial years 2012 to 2013 and 2021 to 2022. CO cases also declined from 14.8% to 13.3% over the same period (Table S3 of the supplementary dataset and Figure 4). The CFR in HO and CO cases have decreased between financial year 2012 to 2013 and 2021 to 2022, although the decrease was only constant between financial years 2012 to 2013 and 2019 to 2020, which coincides with the start of the COVID-19 pandemic. There was an increase of 2% in CFR of CO cases between financial years 2019 to 2020 and 2020 to 2021 (12.6% to 14.6%). The current financial year (2021 to 2022) then saw a decline in CFR of CO cases to 13.3%.

Figure 2. Thirty-day all-cause mortality rate of community onset E. coli bacteraemia, England, financial year April 2012 to March 2013 to financial year April 2021 to March 2022

Figure 3. Thirty-day all-cause mortality rate of hospital onset E. coli bacteraemia, England, financial year April 2012 to March 2013 to financial year April 2021 to March 2022

Figure 4. Thirty-day all-cause case fatality rate of E. coli bacteraemia by onset, England, financial year April 2012 to March 2013 to financial year April 2021 to March 2022

Variation by NHS commissioning region

During financial year 2021 to 2022, regional mortality rate ranged from 7.7 deaths per 100,000 population in London to 12.4 deaths per 100,000 population in the North East and Yorkshire. Over the same period, CFRs ranged from 13.5% in London to 16.1% in the North East and Yorkshire (Table S4 of the supplementary dataset).

Variation by age and sex

Mortality rate and CFR generally increased with age and was greater in male patients than female patients. The exception to this was patients aged under 1 year where mortality rate and CFR was higher compared with those aged 1 to 14 years (Table S5 of the supplementary dataset, Figure 5 and Figure 6).

During financial year 2021 to 2022, among male patients, the highest mortality rates were in those aged over 85 years (175.6 deaths per 100,000 population) and those aged 75 to 84 years (64.1 per 100,000 population). This equated to CFRs of 22.3% and 17.3% respectively. The CFRs for these groups both decreased from financial year 2020 to 2021 and the mortality rate for those aged 75 to 84 years decreased by 2.4 deaths per 100,000 population. However, the mortality rate for those aged 85 years and over increased by 0.8 deaths per 100,000 population.

Mortality rate in males declined between financial years 2020 to 2021 and 2021 to 2022 in all age groups except for those aged 1 to 14 years. However, numbers of deaths are very small in this age group so differences should be interpreted with caution.

The highest mortality rates among female patients during the financial year 2021 to 2022 were seen in those aged over 85 years at 92.7 per 100,000 population and in those aged 75 to 84 years at 38.2. This equates to a CFR of 18.8% and 14.4% of cases respectively. The mortality rate and CFR in those aged under 1 year increased from financial year 2020 to 2021, with mortality rates for females increasing from 5.5 to 5.8 per 100,000 population and CFR increasing from 7.8% to 8.9% of cases.

Mortality rates and CFRs also increased in females aged 1 to 14 years and 15 to 44 years, although numbers of deaths are small in these groups so differences should be interpreted with caution.

Figure 5. Thirty-day all-cause mortality rate of E. coli bacteraemia by age and sex, financial year April 2012 to March 2013 versus financial year April 2021 to March 2022, England

Figure 6. Thirty-day all-cause case fatality rate of E. coli bacteraemia by age and sex, financial year April 2012 to March 2013 versus financial year April 2021 to March 2022, England

Klebsiella species bacteraemia

During financial year 2021 to 2022, 11,409 Klebsiella spp. bacteraemia cases were reported in England. Information on mortality was available for 97.1% (11,078) of these cases (Table S6 of the supplementary dataset). There were 2,162 deaths within 30 days of a Klebsiella spp. Bacteraemia diagnosis giving a mortality rate of 3.8 deaths per 100,000 population. The CFR was 19.5%.

Mandatory surveillance of Klebsiella spp. bacteraemia started during financial year 2017 to 2018, meaning trends are not as established as those in data collections such as MRSA or E. coli bacteraemia. The mortality rate increased from 3.4 to 3.8 deaths per 100,000 population between financial years 2017 to 2018 and 2021 to 2022. Conversely, the CFR decreased from 20.2% (1,896 deaths) to 19.5% (2,162 deaths) in between financial years 2017 to 2018 and 2021 to 2022.

A large increase in CFR was observed between financial years 2018 to 2019 and 2020 to 2021, coinciding with the COVID-19 pandemic. The CFR increased from 18.7% to 21.8% in between financial years 2018 to 2019 and 2020 to 2021. The mortality rate also increased from 3.5 per 100,000 population between financial year 2018 to 2019 to 4.2 per 100,000 population during financial year 2020 to 2021.

Variation by onset of bacteraemia

The mortality rate in CO cases decreased from 2.4 deaths per 100,000 population (1,367 deaths) during financial year 2020 to 2021 to 2.2 deaths per 100,000 population (1,250 deaths) during financial year 2021 to 2022 (Table S7 of the supplementary dataset and Figure 7).Over the same period, the mortality rate of HO cases decreased from 3.6 deaths per 100,000 bed-days (990 deaths) during financial year 2020 to 2021 to 2.8 deaths per 100,000 bed-days (912 deaths) during financial year 2021 to 2022 (Figure 8).

Between financial years 2017 and 2018 and 2021 to 2022, the CFR of HO cases increased from 24.5% to 25.6%. However, there was a decrease in the latter CFR when compared with financial year 2020 to 2021 (27.1%). Compared with the start of surveillance, CFR in CO cases decreased from 18.3% to 16.6% (Table S7 of the supplementary dataset and Figure 9).

Figure 7. Thirty-day all-cause mortality rate of community-onset Klebsiella species bacteraemia financial year April 2017 to March 2018 to financial year April 2021 March 2022, England

Figure 8. Thirty-day all-cause mortality rate of hospital-onset Klebsiella species bacteraemia financial year April 2017 to March 2018 to financial year April 2021 to March 2022, England

Figure 9. Thirty-day all-cause case fatality rate of Klebsiella species bacteraemia by onset, financial year April 2017 to March 2018 to financial year April 2021 to March 2022, England

Variation by NHS commissioning region

During financial year 2021 to 2022, regional mortality rate ranged from 3.1 deaths per 100,000 population in the East of England to 4.5 deaths per 100,000 population in the North East and Yorkshire. Over the same period, CFRs ranged from 17.9% in the East of England to 21.6% in the Midlands (Table S8 of the supplementary dataset).

There were decreases in mortality rate and CFR in all regions between financial year 2020 to 2021 and 2021 to 2022, except for the South West. The South West saw mortality rates increase from 3.1 deaths per 100,000 population to 3.3 and CFR increase from 17.4% to 18.1% of cases.

Variation by age and sex

During financial year 2021 to 2022, the mortality rate and CFR increased with age, except among children aged under one year. The mortality rate was greater in male patients while the CFR was greater in female patients (Table S9 of the supplementary dataset, Figure 10 and Figure 11).

Among male patients, the highest MRs were in those aged 85 years and over (60.1 deaths per 100,000 population) and those aged 75 to 84 years (24.1 deaths per 100,000 population). This equates to CFRs of 26.0% and 19.7% of cases, respectively. Both mortality rates and CFRs decreased in these age groups compared with financial year 2020 to 2021, when the mortality rate was 63.9 deaths per 100,000 population in those aged over 85 years and 24.5 deaths per 100,000 population in those aged 75 to 84 years. During the same period, CFR was 29.3% of cases in those aged 85 years and over and 21.5% of cases in those aged 75 to 84 years.

In female patients of the same age groups, the mortality rates were 21.1 deaths per 100,000 population (those aged 85 years and over) and 11.1 deaths per 100,000 population (aged 75 to 84 years). CFRs for these groups were 27.5% and 22.3% respectively. Similar to trends seen in the males, the mortality rate and CFR were lower than in the previous financial year, when mortality rate was 21.2 deaths per 100,000 population and CFR was 31.5% for females over 85 years. However, an increase was observed in females aged 75 to 84 years from the previous year, where the mortality rate was 11.1 deaths per 100,000 population and the CFR was 22.3% of cases.

Figure 10. Thirty-day all-cause mortality rate of Klebsiella species bacteraemia by age and sex, financial year April 2017 to March 2018 versus financial year April 2021 to March 2022, England

Figure 11. Thirty-day all-cause case fatality rate of Klebsiella species bacteraemia by age and sex, financial year April 2017 to March 2018 versus financial year April 2021 to March 2022, England

Pseudomonas aeruginosa bacteraemia

During financial year 2021 to 2022, 4,334 P. aeruginosa bacteraemia cases were reported in England. Information on mortality was available for 97.6% (4,230) of these cases (Table S10 of the supplementary dataset). There were 1,056 deaths within 30 days of a P. aeruginosa bacteraemia, giving a mortality rate of 1.9 deaths per 100,000 population and a CFR of 25.0%.

Mandatory surveillance of P. aeruginosa bacteraemia started during financial year 2017 to 2018, meaning trends are not as established as those in data collections such as MRSA or E. coli bacteraemia. The CFR had decreased from 26.9% (1,121 deaths) during financial year 2017 to 2018 to 25.0% (1,056 deaths) during financial year 2021 to 2022. The mortality rate also decreased from 2.0 per 100,000 population to 1.9 between financial years 2017 and 2018 and 2021 to 2022.

The mortality rate and CFR of P. aeruginosa bacteraemia were at the highest during surveillance in financial year 2020 to 2021. These both decreased in the most recent financial year. The mortality rate decreased from 2.0 per 100,000 population during financial year 2020 to 2021 to 1.9 during financial year 2021 to 2022. The CFR decreased from 27.7% during financial year 2020 to 2021 to 25.0% during financial year 2021 to 2022.

Variation by onset of bacteraemia

During financial year 2021 to 2022, the mortality rate of HO cases decreased to 1.3 deaths per 100,000 bed-days (428 deaths) compared with 2.0 (550 deaths) in the previous financial year (2020 to 2021) (Table S11 of the supplementary dataset and Figure 13). Over the same period, the mortality rate in CO cases remained stable at 1.1 deaths per 100,000 population (609 deaths and 628 deaths, respectively) between financial years 2020 to 2021 and 2021 to 2022 (Figure 12).

Between the start of surveillance (financial year 2017 to 2018) and the current financial year (2021 to 2022) the CFR of HO cases decreased from 29.9% to 27.0%, while for CO cases it decreased from 25.1% to 23.8% over the same period (Table S11 of the supplementary dataset and Figure 14). Compared with the previous financial year (2020 to 2021), CFR in HO cases in financial year ending March 2022 dropped from 33.9% to 27.0%, but remained steady in CO cases, with both financial years having a CFR of 23.8%.

Figure 12. Thirty-day all-cause mortality rate of Pseudomonas aeruginosa community-onset bacteraemia financial year April 2017 to March 2018 to financial year April 2021 to March 2022, England

Figure 13. Thirty-day all-cause mortality rate of hospital-onset Pseudomonas aeruginosa bacteraemia financial year April 2017 to March 2018 to financial year April 2021 to March 2022, England

Figure 14. Thirty-day all-cause case fatality rate of Pseudomonas aeruginosa bacteraemia financial year April 2017 to March 2018 to financial year April 2021 to March 2022, England

Variation by NHS commissioning region

During financial year 2021 to 2022, regional mortality rate ranged from 1.3 deaths per 100,000 population in the North West to 2.3 deaths per 100,000 population in the North East and Yorkshire (Table S12 of the supplementary dataset). Over the same period, CFRs ranged from 20.1% in London to 29.8% of cases in the North East and Yorkshire. The mortality rate and CFR in East of England during financial year 2021 to 2022 were the highest for the region since surveillance began, at 2.0 deaths per 100,000 population and 27.8% CFR. In addition, mortality rate in North East and Yorkshire has been increasing year-on-year since financial year 2018 to 2019, from 1.9 deaths per 100,000 population to 2.3 in this financial year.

Variation by age and sex

During financial year 2021 to 2022, mortality rate and CFR increased with age, except among children aged under one year. The mortality rate was greater in male patients while the CFR was greater in female patients (Table S13 of the supplementary dataset, Figure 15 and Figure 16).

During financial year 2021 to 2022, among male patients, the highest mortality rates were in those over 85 years (24.0 deaths per 100,000 population) and those aged 75 to 84 years (12.3 deaths per 100,000 population), which corresponded to CFRs of 24.5% and 25.6% of cases, respectively.

In female patients of the same age groups, the mortality rate were far lower than their male counterparts; 12.1 deaths per 100,000 population (aged 85 years and over) and 5.7 deaths per 100,000 population (aged 75 to 84 years). These equated to CFRs of 40.8% and 34.5% of all cases in these age groups.

Among children aged under one year, the mortality rate in male patients was 3.2 deaths per 100,000 population (37.0% of cases) compared with 2.4 (28.0% of cases) in female patients. Although, caution is required in interpreting these data, as the number of deaths was relatively small in both groups.

Figure 15. Thirty-day all-cause mortality rate of Pseudomonas aeruginosa bacteraemia by age and sex, financial year April 2017 to March 2018 versus financial year April 2021 to March 2022, England

Figure 16. Thirty-day all-cause fatality rate of Pseudomonas aeruginosa bacteraemia by age and sex, financial year April 2017 to March 2018 versus financial year April 2021 to March 2022, England

MRSA bacteraemia

During financial year 2021 to 2022, 673 MRSA bacteraemia cases were reported in England. Information on mortality was available for 97.3% (655 deaths) of these cases (Table S14 of the supplementary dataset). There were 172 deaths within 30 days of an MRSA bacteraemia diagnosis, a mortality rate of 0.3 deaths per 100,000 population. The CFR was 26.3% of cases.

The CFR in the financial year 2021 to 2022 was lower than at the start of surveillance (financial year 2007 to 2008), 26.3% versus 38.9%, respectively. The overall trend of mortality rate decreased from 2.6 to 0.3 deaths per 100,000 population between financial years 2007 to 2008 and 2021 to 2022.

Variation by onset of bacteraemia

The mortality rate of CO cases has remained at 0.2 per 100,000 population since financial year 2018 to 2019 (Table S15 of the supplementary dataset and Figure 17). The mortality rate of HO cases increased from 0.2 deaths per 100,000 bed-days (73 deaths) during financial year 2019 to 2020, to 0.3 deaths per 100,000 bed-days (93 deaths) during financial year 2020 to 2021 (Figure 18). However, this later decreased back to 0.2 deaths per 100,000 bed-days during financial year 2021 to 2022 (67 deaths).

The CFR of HO cases decreased from 42.4% to 29.3% in between financial years 2007 to 2008 and 2021 to 2022, while CO cases decreased from 33.1% to 24.6% of cases over the same period (Table S15 of the supplementary dataset and Figure 19).

However, year-on-year increases in CFR for CO cases have been observed, as seen in the rise from 24.0% in financial year 2018 to 2019 to 24.6% in financial year 2021 to 2022. Between financial years 2019 to 2020 and 2020 to 2021, there was an increase in CFR of HO cases, from 29.2% to 34.3%, followed by a decrease in CFR to 29.3% during financial year 2021 to 2022.

Figure 17. Thirty-day all-cause mortality rate of community-onset MRSA bacteraemia financial year April 2007 to March 2008 to financial year 2021 to 2022 , England

Figure 18. Thirty-day all-cause mortality rate of hospital-onset MRSA bacteraemia, financial year 2007 to 2008 to financial year 2021 to 2022, England

Figure 19. Thirty-day case fatality rate of MRSA bacteraemia by onset, financial year 2007 to 2008 to financial year 2021 to 2022, England

Variation by NHS commissioning region

Like the national trend, the majority of both regional mortality rates and CFRs declined between financial years 2007 to 2008 and 2021 to 2022. The exception was in the South East, where the CFR increased from 35.7% during financial year 2007 to 2008, to 39.5% in financial year 2021 to 2022.

In financial year 2021 to 2022, regional mortality rate ranged from 0.2 deaths per 100,000 population in the Midlands to 0.4 deaths per 100,000 population in the East of England (Table S16 of the supplementary dataset). Over the same period, CFRs ranged from 20.0% in London and the Midlands to 39.5% in the South East. In the East of England region, between financial year 2020 to 2021 and 2021 to 2022, there was an increase in mortality rate from 0.3 deaths per 100,000 population to 0.4 and CFR from 27.6% to 29.3%. In London, the Midlands and North East and Yorkshire regions, the number of deaths, mortality rate and CFR were at their lowest since the inception of surveillance during financial year 2021 to 2022.

Variation by age and sex

During financial year 2021 to 2022, mortality rate and CFR increased with age. Over the same period, the mortality rate was greater in males, while CFR was near equal in both sexes (Table S17 of the supplementary dataset, Figure 20 and Figure 21).

Among male patients, the highest mortality rate was in those aged 85 years and over (6.1 deaths per 100,000 population) and those aged 75 to 84 years (1.5 deaths per 100,000 population), with CFRs being 43.2% and 35.9%, respectively.

In female patients, the mortality rate was also higher in the oldest age groups: 2.0 deaths per 100,000 population among those aged over 85 years and 0.9 deaths per 100,000 population among those in the group aged 75 to 84 years. CFRs of these 2 groups were 46.2% and 50.0%, respectively.

Compared with other infections covered in this report, there were relatively fewer deaths in patients aged under one year compared with other age groups. During financial year 2021 to 2022, there were 2 deaths within 30-days following MRSA bacteraemia in female patients aged under one year, and no deaths in males of this age group.

Figure 20. Thirty-day all-cause mortality rate of MRSA bacteraemia by age and sex, financial year April 2007 to March 2008 versus financial year April 2021 to March 2022, England

Figure 21. Thirty-day all-cause case fatality rate of MRSA bacteraemia by age and sex, financial year April 2007 to March 2008 versus financial year April 2021 to March 2022, England

MSSA bacteraemia

During financial year 2021 to 2022, 12,283 MSSA bacteraemia cases were reported in England. Information on mortality was available for 97.6% (11,983 deaths) of these cases (Table S18 of the supplementary dataset). There were 2,652 deaths within 30 days of an MSSA bacteraemia diagnosis which gave a mortality rate of 4.7 deaths per 100,000 population. The CFR was 22.1%.

There was a declining trend in CFR starting from 21.5% during financial year 2011 to 2012 to 19.1% during financial year 2018 to 2019; this was followed by an increase to 23.7% during financial year 2020 to 2021 and a subsequent decrease to 22.1% during financial year 2021 to 2022. The overall trend of mortality rate increased from 3.3 deaths per 100,000 population to 4.8 in between financial years 2011 to 2012 and 2020 to 2021 (the highest mortality rate for MSSA bacteraemia since the start of mandatory surveillance ), with a slight decrease during financial year 2021 to 2022, to 4.7 per 100,000 population.

Variation by onset of bacteraemia

The CO of mortality rate cases decreased from 3.1 deaths per 100,000 population (1,776 deaths) during financial year 2020 to 2021 to 3.0 deaths (1,699 deaths) during financial year 2021 to 2022 (Table S19 of the supplementary dataset and Figure 22). Similarly, the mortality rate of HO cases decreased from 3.3 deaths per 100,000 bed-days (916 deaths), during financial year 2020 to 2021 to 2.9 deaths per 100,000 bed-days (953 deaths) during financial year 2021 to 2022 (Figure 23).

The CFR of HO cases increased from 26.7% during financial year 2011 to 2012 to 26.3% during financial year 2021 to 2022, while for CO cases it increased from 18.9% to 20.3% in between financial years 2012 to 2013 and 2021 to 2022 (Table S19 of the supplementary dataset and Figure 24). The HO and CO CFR for the previous financial year (2020 to 2021, 28.2% and 21.8%) was the highest recorded for MSSA since its mandatory surveillance began, and this year showed a small decline to 26.3% and 20.3%, respectively. Though the CFR in HO and CO cases has shown a slow decrease between financial year 2012 to 2013 and 2020 to 2021, the decrease was only constant between financial year 2012 to 2013 and 2019 to 2020, which coincides with the start of the COVID-19 pandemic.

Figure 22. Thirty-day all-cause mortality rate of community onset MSSA bacteraemia, financial year 2011 to 2012 to financial year 2021 to 2022, England

Figure 23. Thirty-day all-cause mortality rate of hospital onset MSSA bacteraemia, financial year 2011 to 2012 to financial year 2021 to 2022, England

Figure 24. Thirty-day all-cause case fatality rate of MSSA bacteraemia by onset, financial year 2011 to 2012 to financial year 2021 to 2022, England

Variation by NHS commissioning region

During financial year 2021 to 2022, regional mortality rate ranged from 3.4 deaths per 100,000 population in London to 6.4 deaths per 100,000 population in the North East and Yorkshire. Over the same period, CFRs ranged from 19.1% in London to 23.8% in the North East and Yorkshire (Table S20 of the supplementary dataset).

The number of deaths and mortality rate recorded in the North East and Yorkshire region were the highest since surveillance started in financial year 2021 to 2022 at 557 and 6.4 per 100,000 population. This was also the case for the North West region, where number of deaths and mortality rate at the start of surveillance were 273 per 100,000 population and 4.0, respectively, and 396 and 5.6 per 100,000 population during financial year 2021 to 2022.

In the Midlands, South East and South West regions, the highest mortality rate observed since the start of surveillance was during financial year 2020 to 2021. All of these regions subsequently decreased during financial year 2021 to 2022. The CFR fell in every region between financial years 2020 to 2021 and 2021 to 2022.

Variation by age and sex

During financial year 2021 to 2022, the mortality rate and CFR increased with age, except among children aged under one year. Mortality rate was greater in male patients while CFR was close to equal in both sexes (Table S21 of the supplementary dataset, Figure 25 and Figure 26).

Among male patients, the highest mortality rates were in those aged 85 years and over (85.8 deaths per 100,000 population) and the group aged 75 to 84 years (29.8 deaths per 100,000 population) with CFRs of these age groups at 44.2% and 30.6% respectively. The mortality rate in males aged over 85 years has been increasing over time and was the highest since the start of surveillance during financial year 2021 to 2022. In female patients, the mortality rates were also higher in older age groups, 37.7 deaths per 100,000 population (aged 85 years and over) and 14.3 deaths per 100,000 population (aged 75 to 84 years). These equate to a CFR of 47.4% and 32.0% of all cases in those respective age groups.

Among children aged under one year, the mortality rate in male patients was 1.3 deaths per 100,000 population (2.4% of cases) compared with 4.8 per 100,000 population (13.6% of cases) in female patients. Although the numbers of deaths in this group are small and data should be analysed with caution.

Figure 25. Thirty-day all-cause mortality rate of MSSA bacteraemia by age and sex, financial year April 2011 to March 2012 versus financial year 2021 to 2022, England

Figure 26. Thirty-day all-cause case fatality rate of MSSA bacteraemia by age and sex, financial year April 2011 to March 2012 versus financial year 2021 to 2022, England

Clostridioides difficile infection

During financial year 2021 to 2022, 14,249 CDI cases were reported in England. Information on mortality was available for 98.0% (13,967) of these cases (Table S22 of the supplementary dataset). There were 1,910 deaths within 30 days of a CDI case, giving a mortality rate of 3.4 deaths per 100,000 population and a CFR of 13.7%.

There was a declining trend in CFR from 26.3% during financial year 2007 to 2008 to 13.5% during financial year 2019 to 2020. After this the CFR increased to 14.9% during financial year 2020 to 2021 before decreasing during financial year 2021 to 2022 to 13.7%. The overall trend of mortality rate decreased from 27.1 deaths per 100,000 population during financial year 2007 to 2008 to 2.9 deaths per 100,000 population in financial year 2018 to 2019. It then increased each financial year since, to 3.4 deaths per 100,000 population in the current financial year. The different trend between financial years 2019 to 2020 and 2020 to 2021 is likely a consequence of the COVID-19 pandemic.

Variation by onset of bacteraemia

The mortality rate of CO cases decreased from 8.2 deaths per 100,000 population (4,226 deaths) at the start of surveillance to 1.5 deaths per 100,000 population (862 deaths) during financial year 2021 to 2022 (Figure 27). The same trend is observed with the mortality rate of HO cases, decreasing from 26.0 deaths per 100,000 bed-days (9,747 deaths) at the start of surveillance (financial year 2007 to 2008) to 3.2 deaths per 100,000 bed-days (1,048 deaths) during financial year 2021 to 2022 (Table S23 of the supplementary dataset and Figure 28).

The CFR of HO cases decreased from 30.2% to 20.0% between financial years 2007 to 2008 and 2021 to 2022. CO cases also decreased from 20.2% to 9.9% between financial years 2007 to 2008 and 2021 to 2022 (Table S23 of the supplementary dataset and Figure 29). Though the CFR in HO and CO has shown a slow decrease between financial years 2007 to 2008 and 2021 to 2022, the decrease was only constant between financial years 2007 to 2008 and 2019 to 2020, which coincides with the start of the COVID-19 pandemic.

Variation by prior healthcare exposure

The mortality rate of hospital-onset healthcare associated (HOHA) cases was at 3.1 deaths per 100,000 bed days during financial year 2019 to 2020. It then rose to 3.9 deaths per 100,000 bed days (1,089 deaths) during financial year 2020 to 2021, the highest seen since the start of using these prior healthcare exposure definitions. It then decreased in financial year 2021 to 2022, to 3.5 per 100,000 bed days (1,148 deaths).

The mortality rate of community-onset community associated (COCA) cases has been increasing between financial years 2018 to 2019 and 2021 to 2022, from 0.43 deaths per 100,000 population to 0.58 deaths per 100,000 population. The community-onset healthcare associated (COHA) mortality rate was also the highest observed since the start of using these prior healthcare exposure definitions during financial year 2020 to 2021 at 0.92, before decreasing to 0.74 deaths per 100,000 day admissions and bed days (Table S24 of the supplementary dataset).

Figure 27. Thirty-day all-cause mortality rate of community-onset CDI, financial year April 2007 to March 2008 to financial year April 2021 to March 2022, England

Figure 28. Thirty-day all-cause mortality rate of healthcare-onset CDI financial year April 2007 to March 2008 to financial year April 2021 to March 2022, England

Figure 29. Thirty-day all-cause case fatality rate of CDI by onset financial year April 2007 to March 2008 to financial year April 2021 to March 2022, England

Variation by NHS commissioning region

During financial year 2021 to 2022, regional mortality rate ranged from 2.0 deaths per 100,000 population in London to 4.8 deaths per 100,000 population in the North West. Over the same period, CFRs ranged 11.8% in the London to 14.8% in the Midlands (Table S25 of the supplementary dataset).

Variation by age and sex

CDI surveillance only covers patients aged over 2 years. The CFR following CDI infections increased with age. Mortality rate and CFR increased with age and was greater in male patients compared with female patients. (Table S26 of the supplementary dataset, Figure 30 and Figure 31).

Among male patients, the highest mortality rates were in those aged 85 years and over (62.2 deaths per 100,000 population) and in those aged 75 to 84 years (19.9 deaths per 100,000 population), with a corresponding CFR of 26.1% and 17.6%, respectively.

In female patients, the mortality rates were also higher in older age groups, 43.8 deaths per 100,000 population (aged 85 years and over) and 16.5 deaths per 100,000 population (aged 75 to 84 years). These equated to CFRs of 19.1% and 13.4% of all cases in those respective age groups.

Figure 30. Thirty-day all-cause mortality rate of CDI by age and sex, financial year April 2007 to March 2008 versus financial year April 2021 to March 2022, England

Figure 31. Thirty-day all-cause case fatality rate of CDI by age and sex, financial year April 2007 to March 2008 versus financial year April 2021 to March 2022, England

Discussion

After a decline in incidence of bacteraemia and CDI observed from financial year 2019 to 2020 to financial year 2020 to 2021, the number of infections has since increased in the current financial year, 2021 to 2022. However, the number of deaths within 30 days of being diagnosed with bacteraemia or CDI were higher in financial year 2020 to 2021 compared with financial year 2019 to 2020. Deaths in financial year 2021 to 2022 reduced from 13,984 to 13,501. Mortality rate and CFR also both increased from financial year 2019 to 2020 to financial year 2020 to 2021, and have decreased since, from 24.7 to 23.9 deaths per 100,000 population and from 18.6% to 17.1%, respectively.

Differing trends in the most recent financial years are likely a consequence of the COVID-19 pandemic. Fewer infections were observed than expected across many of the bacteraemia and CDI initially, in financial year 2020 to 2021. However, increases were observed in the mortality rate of hospital onset cases during financial year 2020 to 2021, compared with financial year 2021 to 2022 and 2019 to 2020. This may be due to a decrease in bed-days observed during financial year 2020 to 2021, with a reduction in hospital activity, as well as reduced mixing due to lockdown events (coinciding with the COVID-19 pandemic). The effect of specific factors contributing to the changes in mortality rate has yet to be discerned. The increase in CFR in financial year 2020 to 2021 may reflect poor outcomes in patients with both a bacteraemia or CDI and COVID-19, or a skew in bacteraemia cases with a respiratory source.

During the COVID-19 pandemic, there was an increase in CFR in E. coli bacteraemia cases (14.3% to 16.0%) between financial years 2019 to 2020 and 2020 to 2021. However, there was a decrease in mortality rate (10.7 to 10.2 deaths per 100,000 population) over the same period. The CFR of E. coli bacteraemia (15.0%) was relatively small compared with those for MRSA and P. aeruginosa bacteraemia in financial year 2021 to 2022. However, its higher incidence of infection, and deaths following infection (mortality rate: 9.8 deaths per 100,000 population) compared with MRSA (mortality rate: 0.3 deaths per 100,000 population) and P. aeruginosa bacteraemia (mortality rate: 1.9 deaths per 100,000 population) highlights the public health burden of this infection.

The continued increase in mortality rates following E. coli bacteraemia is of particular concern. During financial year April 2021 to March 2022, there were 5,549 deaths within 30 days of E. coli bacteraemia. This is approximately 4 times the number of deaths following MRSA bacteraemia at the start of mandatory surveillance (financial year April 2007 to March 2008, 1,354) when MRSA bacteraemia cases were at its highest levels. Despite the high mortality rates, the CFR for E. coli bacteraemia remains low.

The overall declining trend of CFRs for CDI and MRSA bacteraemia is indicative of the change in their epidemiology over the years. Over the entire surveillance period, the incidence of these infections declined, and both have shown a shift from predominantly hospital infections to community cases. Since mortality and morbidity are often higher in HO (a proxy for healthcare-acquired) cases compared with CO cases, reductions in the former would most likely be accompanied by reductions in the overall CFR of the infection as seen in MRSA bacteraemia cases (Figure 10, Figure 11; Table S15 of the supplementary dataset).

Background information

UK Health Security Agency (UKHSA) has undertaken mandatory surveillance of key healthcare-associated infections (HCAI) in England since 2001, when NHS acute trusts were mandated to report aggregate counts of S. aureus bacteraemia (bloodstream infection) and the number that were MRSA. Case-level reporting was introduced for MRSA in 2005, and for CDI in patients, aged at least 2 years, since April 2007. The mandatory surveillance programme was expanded to include MSSA and E. coli bacteraemia in January and July 2011, respectively. In April 2017, Klebsiella spp. and P. aeruginosa bacteraemia were also added to the surveillance programme.

Over time, the dominance of HO MRSA, MSSA, Gram-negative bacteraemia and CDI has declined, with a corresponding increase in the proportion of CO cases. Since the composition of the patient population can vary in each setting, this publication reports mortality outcomes by onset of infection, and by prior healthcare exposure for CDI. Due to the impact of HCAI on patients’ morbidity and mortality, monitoring trends in mortality is an important part of HCAI surveillance.

This report presents an analysis of 30 day all-cause mortality among MRSA, MSSA, E. coli, Klebsiella spp. and P. aeruginosa bacteraemia, and CDI patients. A separate report presents an analysis of the incidence of all reported cases of these same infections.

Limitations

The analyses presented here are based on infections reported to UKHSA that could be linked to the NHS Spine to obtain mortality information. While most of infection reports had complete NHS numbers (required for linkage), for some reports the NHS Spine was not able to return patient information. This was due to reasons such as the NHS number and date of birth (DOB) not matching a record on the NHS Spine. As a result, there may be bias in the records with available mortality information, which may over- or under-estimate the number of deaths and associated CFRs if the records without mortality information were for patients with a different likelihood of death. However, this effect on reported outcomes is likely to be low, since the linkage had a high degree of completeness, at around 97.5% of all cases.

Crude CFRs are presented and as such have not been adjusted for potential confounders such as age, sex, or co-morbidities. This may affect comparisons over time, between regions and onset setting.

Finally, while analysis of 30 day all-cause fatality estimates presents the risk of death following an infection within a fixed time frame, it does not provide insight into attributable mortality. However, it is difficult to ascertain attributable mortality in practice, due to clinical and diagnostic uncertainty encountered when trying to determine the exact cause of death in patients, particularly in those with multiple co-morbidities.

The Office for National Statistics (ONS) has historically published statistics on deaths involving MRSA and C. difficile. These statistics are not comparable with those presented here for the reasons highlighted in the methods.

Methods

Infection reports

Data on MRSA, MSSA, E. coli, Klebsiella spp. and P. aeruginosa bacteraemia, and CDI were extracted on 24 April 2022 from the HCAI Data Capture System (DCS). These data cover the period from the start of surveillance for each data collection (April 2007 for MRSA BSI and CDI; April 2011 for MSSA BSI; April 2012 for E. coli BSI; April 2017 for Klebsiella spp. and P. aeruginosa BSI) to March 2022. The disparate time periods reflect the first full financial year available for each of the infections. Reports of CDI from patients aged under 2 years at the time of specimen collection were excluded from all analyses because these data are not subject to mandatory surveillance, as carriage rates of C. difficile are high with little evidence for disease in this age group.

Mortality data

Mortality data were obtained by linking the HCAI DCS extract with the NHS Spine, a central repository of patient demographic and medical information managed by the Health and Social Care Information Centre. Records were traced using the NHS number and DOB. Only records that match on both the NHS number and the DOB can be successfully traced and have the potential for fatality information to be returned. These are referred to as ‘linked or traced reports’ in this document and the accompanying dataset.

Within the HCAI DCS, NHS number and DOB are mandatory fields for entering and saving a case onto the surveillance system. Users can enter nines in place of a valid NHS number if the NHS number is unknown, while ‘01/01/1900’ is used for DOB if it is unknown. Only traced reports are considered when calculating CFR.

For infection reports with a death reported in the NHS Spine, the time in days between specimen date and date of death was calculated to identify whether it was within a 30-day window. Bacteraemia reports with a date of death 2 or more days prior to the specimen date were excluded from the analysis.

In publications prior to September 2018, CDI cases with dates of death 3 or more days before the sample data were excluded. However, since then such cases have been included and are considered a 30-day all-cause death. On the HCAI DCS, MRSA, MSSA, E. coli, Klebsiella spp. and P. aeruginosa bacteraemia episode lengths are 14 days, and CDI is 28 days, therefore it is possible to have multiple cases within 30 days of a death.

Where multiple records from the data collection (bacteraemia or CDI) had the same NHS number and DOB within the 30-day fatality window, only the record with the specimen date closest to the date of death was used to calculate 30 day all cause CFR. This was done to prevent estimate bias by overestimating of the numbers of deaths. This deduplication algorithm was applied to both the 30-day fatality, traced and total number of reports to prevent an inflated count of deaths and reports.

Records between financial years 2007 to 2008 and 2014 to 2015 were originally traced on 4 July 2015. A secondary trace was conducted on all records between financial years 2013 to 2014 and 2016 to 2017 on 3 July 2017. Records after financial year 2016 to 2017 were traced in the same financial year they were published: for the latest data, this occurred on 8 August 2022.

Caveats

The number of infections and deaths presented here may differ from those in earlier publications due to late reports, inclusion of new reports since previous publications, or mortality re-tracing.

Data in this report are presented by financial year based on the date of sampling of the first positive specimen rather than the date of death. It is therefore possible that a death occurred in a different financial year from what is presented in this report. When the date of sampling was not known, the date when the sample was received in the laboratory was used as a proxy.

De-duplication algorithm

This report uses the same base data as UKHSA’s 2021 to 2022 Annual HCAI Epidemiological Commentary. Unlike the annual report, in this publication counts of infections and deaths within 30 days of specimen collection have been deduplicated by infection to a patient level. If a patient was positive for multiple types of infection within 30 days of their death, they will be counted for each infection. This ensures that a patient’s mortality outcome is reflected for each infection.

Blood (faecal for CDI) sample with the earliest specimen date before the patient’s date of death was considered the index sample. However, if there was a bacteraemia positive post-mortem sample within one day after a patient’s date of death, and no other case was reported within 30 days prior to the patient’s date of death, that post-mortem sample is taken as the index sample and considered a 30-day all-cause death. This short period of allowance is included to account for the possibility of late reports and data entry errors. Post-mortem bacteraemia samples after this period of allowance are excluded and not included in either the denominator (linked cases) or numerator (30-day all-cause deaths) of CFR analyses.

In contrast, CDI positive post-mortem samples are included in CFR where applicable since they are subject to mandatory surveillance. For the purpose of this report, patients with a post-mortem CDI faecal sample are assumed to have died on the same date as their most recent sample. If no other CDI positive sample was reported within 30 days prior to the patient’s date of death, these CDI positive post-mortem samples are taken as index cases and considered 30-day all-cause deaths

For bacteraemia and CDI samples which could be linked to fatality records, the most recent sample was taken as an index sample. Bacteraemia and CDI positive samples with specimen dates before index samples but within a 30-day window period prior to the patient’s date of death are excluded as duplicates.

Deduplication is done at a data collection-level. This means that each data collection will have its own index sample if a patient tests positive for multiple infections covered in this report. Additionally, only duplicates of the same type of infection are excluded. For example, if a patient tests positive for MRSA, E. coli and CDI samples within 30 days of each other, they will be included in CFR calculation 3 times, once for each data collection.

Calculation of case fatality rates

Case fatality rates (CFR) indicate the risk of death per case of a particular infection. They were calculated as:

CFR equals 100 multiplied by (number of deaths from any cause occurring within 30 days of sampling date) divided by (number of cases)

Thus, if the ratio of deaths to cases remains constant over time, so will the CFR, even if there has been an overall increase or decrease in both the number of deaths and cases. By contrast, if the number of deaths increases but the number of cases remains constant, or if the number of deaths remains constant but the number of cases decreases, the CFR will increase. Thus, the CFR facilitates comparison between clinical outcomes of diseases with differing incidence.

Estimation of number of 30-day all-cause deaths

An estimate of the number of deaths that might be observed in each period if all infection reports could have been linked to mortality records was calculated.

Number of 30-day all-cause deaths equals (number of de-duplicated infection reports submitted to the HCAI DCS for a given financial year) divided by (30-day CFR expressed as a proportion)

This is rounded to the nearest whole number. Care should be taken with interpretation of this metric, as it assumes the risk of death for those cases that could and could not be linked are the same.

Calculation of mortality rates

Mortality rate is a measure of deaths in the population at risk. This contrasts with the CFR, which shows the percentage of people with an infection who die within 30 days. The mortality rate for each infection is calculated as:

Mortality rate equals (number of deaths from any cause occurring within 30 days of sampling date) divided by (population estimate)

The population estimate is based on England’s mid-year estimates by ONS when calculating the mortality rate among CO infection. For HO cases, the population estimate used were NHS overnight admissions. However, for COHA cases, NHS overnight and day admissions were used.

Comparability with previous ONS publications on mortality

The Office for National Statistics (ONS) previously published data on deaths in England involving MRSA and C. difficile. The ONS data on MRSA bacteraemia and CDI are not comparable to the data published here for several methodological reasons.

Pathogens

UKHSA outputs cover deaths relating to MRSA, MSSA, E. coli bacteraemia and C. difficile infection whereas ONS outputs cover deaths relating to MRSA bacteraemia and C. difficile only.

Mortality data source and definition

UKHSA uses NHS Spine reports of death from all causes, within 30 days of the date of a positive specimen. ONS uses death certificates, only including deaths for which the pathogen was explicitly listed as a cause of death. All-cause mortality is a common epidemiological convention; while it is not known if the deaths were attributable to the HCAI, the use of all-cause mortality provides a consistent methodology to determine the temporal trends and reduces the subjectivity of cause of death or changes in priorities for death certification.

Geography

UKHSA fatality outputs cover England only whereas ONS outputs cover England and Wales.

Time period covered

UKHSA outputs are timed by financial year whereas ONS outputs use the calendar year.

Denominator

UKHSA outputs use all traced reports of MRSA, MSSA, E. coli, P. aeruginosa and Klebsiella spp. bacteraemia or CDI in the given time period as the denominator whereas ONS use all deaths in the given time period and population in the given time period (2 different denominators used).

Revisions to data included are covered by a data-specific revisions and correction policy.

Citations

Citation to UK Health Security Agency (UKHSA), healthcare associated infections (HCAI) and antimicrobial resistance (AMR) division is required, if these data are used for publication elsewhere, using the content above.

Citation: UK Health Security Agency. MRSA, MSSA, Gram-negative bacteraemia and CDI: 30-day all-cause fatality, London: UK Health Security Agency, March 2023.