Blood cultures - do they make a difference to patient management?


Blood cultures are an expensive test, and optimising the pathway requires considerable effort and commitment. Is it worth the trouble? We looked at 102 consecutive episodes of bacteraemia with clinical details in microbiology system, excluding contaminants and deduplicated to see how it affected patient management.

To spoil the ending, these are the conclusions :


Conclusions


A positive blood culture significantly altered management in 72/102 cases of significant bacteraemia

This was on the basis of

            Gram stain                  27 cases
            Culture                       25 cases
            Sensitivities                20 cases



There may be debate as to the proportion of septic patients who are bacteraemic (anywhere from 10% to 95% depending on what you believe), but even if we take the lower end of this as being true, bacteraemia would seem an important way of influencing management, even if it is just prompting the microbiologist to get out of the lab and go and review the patient.

Effect on diagnosis

Blood cultures had a significant impact on diagnosis in 25 cases (25%)
            In 18 cases this was predominantly on the basis of the gram stain
            In 7 cases this was predominantly on the basis of culture results

Review of incorrect diagnosis

In 15 cases the initial diagnosis was wrong and was modified on the basis of the blood culture result

            In 10 cases the diagnosis was switched on the basis of the gram stain
            In 5 cases the diagnosis was switched on the basis of the culture

Initial diagnosis when wrong

           
CAP
6
UTI
4
HAP
2
Neutropenic sepsis
1
Cellulitis
1
Endocarditis
1

Corrected diagnosis when initially wrong

UTI
5
Cryptic
2
Endocarditis
1
Infected ulcer
1
Infected prosthetic joint
1
Septic arthritis
1
Cholecystitis
1
CAP
1
Abdominal infection
1
Discitis
1


Review of uncertain diagnosis

In 10 cases, the blood culture suggested a cause of infection that was not otherwise apparent.

            In 8 cases this was suggested from the gram stain
            In 2 cases this was suggested from the culture results

Final diagnosis when initially uncertain

Line infection
3
Septic arthritis
2
UTI
1
Surgical site infection
1
Discitis
1
Cellulitis
1
Abdominal infection
1

Effect on antibiotic usage

In 47 cases when the diagnosis was supported by the blood culture, antibiotics were changed on the basis of the result and microbiologist review

Narrowed
20
Broadened
11
Switched
11
Started
5

The switch was instigated on the basis of :

gram
9
organism
18
sensitivity
20

Narrowed :

taz to amox
3
Amox/gent to amox
2
gent to amox
2
flu to amox
1
Co-amox/clari to amox
1
Mero to trim.
1
Not recorded
1
Taz to fluclox
1
mero to cipro
1
mero to mtz
1

Broadened

Metronidazole added to fluclox
2
gent r to mero
2
amox to cef
1
taz to mero - res gent taz
1
fluclox to taz
1

Switched

taz R switched to cipro
1
cef r to cip
1
taz to teic line lock
1
fluclox to amox
1


These were the organisms : 
1

Coliform
48
Other
18
Haemolytic streptococcus
11
S. aureus
9
anaerobe
6
Pneumococcus
5
Faecal streptococcus
5

These were the final diagnoses:

UTI
41
Cholecystitis
11
Cryptic
11
Cellulitis
7
CAP
6
Line infection
5
Septic arthritis
5
Abdominal infection
3
Infected prosthetic joint
2
Infected ulcer
2
Discitis
2
Infected pressure ulcer
2
Surgical site infection
1
Mycotic aneurysm
1
Endocarditis
1
SBP
1
Diverticulitis
1


Comments

Post a Comment

Popular posts from this blog

Artefactual hyperkalemia - dangerous, expensive. And preventable.

Image
Potassium is part of a renal profile, the most commonly ordered test in pathology.  In N Devon we do just over 100000 primary care tests per year for a population of 160000 Assuming we are similar to others, that would be about 40 million primary care tests per year in the UK. Most potassium in a blood sample sits inside the blood cells, and there is a large concentration gradient across the cell membranes. This is maintained through the the Na/K ATPase, which uses energy to drive sodium out of the cell, in return for potassium entering the cell. Because this is an enzymatic process it is extremely sensitive to environmental conditions. For instance, as temperature rises then the enzyme works faster, pumping more potassium out of the serum into the cell. Conversely, when it is cold the pump doesn't work well and potassium leaks out of the cells. All this means that the level of potassium in the serum can be affected by the conditions in which a non-stabilised sample is kep
Read more

Use of Peezy device to try to improve quality of urine culture

Image
Background Urine cultures are often contaminated, making interpretation difficult. Pure growths may often not reflect infection and may be due to contamination - in our lab only one third of pure growths from pregnant women are reproducible.  In pre-op urines mixed growths make optimising prophylactic antibiotics difficult. We often end up giving broad spectrum antibiotics to cover all the organisms in a mixture as there is insufficient time to get a repeat. Alternatively, multi-resistant organisms may not be covered adequately. For patients with bacteraemia from a urinary source, many have a mixture in their urine culture. This compromises early rationalisation of antibiotic therapy. The Peezy has been designed to try to improve the quality of urine specimens by automatically discarding the beginning and end of the urine sample. In 2018 we trialled its use in pre-operative patients. These are the results. Number of patients The Peezy was used for both orthopaedic and
Read more

Updated on SoS - Introducing CARIS

Introducing CARMIS - Codes of Antibiotic Responsive Medical Infections associated with Sepsis. In a previous blog I developed the work on SoS codes of Matt Inada-Kim to focus on those codes that are are associated with high mortality infections. Having just completed the first GIRFT deep dive, this concept worked, but was clear it needs development. The first is with the "branding", and we need a new name that more clearly distinguishes this from sepsis, and makes it clear that it is derived from Hospital Episode Statistics. I also felt we needed more categories, to make the reasoning clearer. It is hard to know how to deal with things like aspiration pneumonia and COPD. These have amongst the highest mortalities of any common disease, but it is debatable how much antibiotics contribute to management. Similarly, surgical conditions that are principally managed through intervention (certainly perforation, but also probably cholangitis) muddy the waters. My latest feeling is
Read more