Flow chart figures

This initial search was a very broad search which used infective endocarditis as the main search term. The consideration of the titles and abstracts from this search yielded a large number of papers which included those which would be relevant to the consideration of the risk/outcomes of developing IE with cardiac conditions but also included papers which it was identified would be appropriate for other sections under consideration in this guideline.

Initial search – infective endocarditis search and including risk and outcomes

Bacteraemia and interventional procedures

The third search was devised in consideration of antibiotic prophylaxis against IE.

Antibiotic prophylaxis

The final search was devised to consider patient perspectives on prophylaxis against IE.

Patient perspectives

Evidence Tables

Evidence Tables

Footnotes

1

According to the guidelines of the endocarditis working party of the British Society of Antimicrobial Chemotherapy (1993) and the American Heart Association (1997)

2

According to the guidelines of the endocarditis working party of the British Society of Antimicrobial Chemotherapy (1993) and the American Heart Association (1997)

3

Derived from several sources

4

Derived solely from this study

5

Derived from several sources

6

As some individual’s records covered less than a year or several years the number of times a dento-gingival manipulative procedure was carried out was standardised on a year

7

Allocated by coded sealed letter from Bayer Co, opened by the anaesthetist when the patient came to the theatre

8

Databases originated in locations in the USA (2 locations) and Europe (Spain, France, Sweden, UK)

9

Randomised by the opening of sealed opaque envelopes

10

Study had 80% power to detect a difference of 15%, from 1% to 16% at the 5% significance level

11

the doctor performing suture removal was unaware of whether or nor a patient had used a rinse; power calculation completed

12

Complicated PVE was defined as infection associated with any of the following; a new or increasing murmur of prosthetic valve dysfunction; new or worsening CHF related to dysfunction of the prosthesis; fever doe 10days or more days during appropriate antibiotic therapy; new or progressive abnormalities of cardiac condition

Relapse – if infection occurred in the 12mths after hospital discharge and either was caused by the same organism or else no pathogen was identified

13

Medical-surgical therapy was considered to be where patient underwent repair or replacement of the infected prosthesis during the initial hospitalisation for treatment of PVE

14

The one exception was the inclusion of those with antecedent findings of isolated mitral regurgitation, since mitral valve prolapse is commonly accompanied by auscultatory findings of mitral regurgitation

15

The eligibility of patients was determined by a ‘blinded’ researcher, without knowledge of the echocardiograph findings

16

Mitral valve prolapse diagnosed by echocardiography

17

The authors note that the frequency of mitral valve prolapse in the general population varies according to the diagnostic criteria used

18

randomisation not specified; power calculated

19

PCC were defined according to the French recommendations for IE prophylaxis

20

Data used was taken from a 1-yr French epidemiological study on IE in 1999

21

Sample drawn from 2 studies ongoing in 1998, a structured and previously validated questionnaire was administered by phone interview to classify subjects as having a PCC or not

22

2.7 administered antibiotic courses, corresponding to 2,228,545 for those with native valve PCC and 512,829 for those with prosthetic valve PCC

23

An additional blood sample was obtained 5mins after ET intubation but before endoscopy in those who had a GA to assess if the endotracheal intubation may have resulted in bacteraemia

24

The First Natural History Study of Congenital Heart Defects between 1958–1965

25

The Second Natural History Study between 1983–1989

26

VSD and VSD plus aortic regurgitation

27

Early PVE; PVE occurring during the first 12mths after surgery Late PVE; PVE occurring after 12mths

28

Randomisation not specified

29

Randomisation not specified

30

Titles or abstracts identified by the search were reviewed independently by two investigators regarding suitability for inclusion in the meta-analysis, if there was disagreement an assessment was made by a third investigator

31

There was little heterogeneity between the results

32

7 of the cases were on chronic haemodialysis and 6 were parenteral drug users

33

In no set was MVP present in more than one of the 3 controls

34

Taken from the New South Wales State hospital morbidity and mortality statistics for 1980

35

Any positive blood culture in which the isolated microorganism is one of the following: coliform bacteria (including Escherichi coli and Proteus) Bacteroides, Hemophilus, group A Streptococcus, and Streptococcus pneumoniae, or more than one blood culture, drawn at different times, positive for the same organism. Patients were not necessarily symptomatic

36

A single positive blood culture in which the isolated microorganism is one of the following: Staphylococcus coag negative (including S. epidermidis anf S. warneri), Corynebacterium, Propionibacterium and Bacillus species, unless a patient has a prosthetic valve, graft or shunt, or a single blood culture for Clostridia (including C. perfringens and C. sordelli) without clinical correlation of active infection

37

none of the blood culture results drawn before endoscopy were positive

38

The prevention used was similar to that recommended earlier by the AHA

39

PVE was considered related to the diagnostic or therapeutic procedure only if symptoms of endocarditis occurred within 2weeks

40

The bacteriologic determinations were made in the laboratory without the investigator having any knowledge of the nature of the individual samples

41

To adjust for factors which could potentially influence the risk of IE associated with procedures, the questionnaire requested items concerning general co-morbid conditions such as alcohol and tobacco consumption, and diabetes mellitus

42

Abdominal surgery N=3, soft tissue surgery N=6, gynaecological surgery N=2. Two of the 7 clean surgical procedures were done with antibiotic prophylaxis and five without antibiotic prophylaxis

43

usual procedure is to administer an enema with antibiotics before the procedure and a repeat dose of antibiotics 12hrs after the procedure

44

organisms identified; Staphylococcus, Streptococcus, Diptheroids, Bacteeroides fraglis, E coli, Proprionibacterium, Gemella morbillum (all N=1 patient), Enterobacter, Gram-pos rods (all N=2 patients)

45

study patients were selected consecutively from a large pool of outpatients who underwent dental extractions; randomised by a random number generator in the hospital pharmacy, unmarked identical bottles; power analysis

46

A computer-generated random number system was used by our pharmacy to assign identically appearing syringes containing placebo or study drug. All investigators were blinded as to the assigned treatment

47

All bacteria were considered in the analysis of the intubation blood draw, number 1, because the skin and nasopharynx are more likely to harbour other bacteria

48

Resumption of clinical picture of endocarditis in the first 6mths after treatment, an infecting micro-organism of the same genus and species, no change in underlying cardiac condition

49

Clinical picture and isolation of a micro-organism different from previous episode of endocarditis, change in underling cardiac condition, clinical picture and micro-organism consistent with previous episode of endocarditis greater than 6mths since the previous episode

50

If prophylactic antibiotics were deemed necessary they were not administered until after the collection of the second set of cultures

51

To inform the registry, medical records departments in all Oregon hospitals that performed cardiac or thoracic surgery were asked to identify cases

52

To obtain long term follow-up information, subjects were traced through next of kin, physicians, employment records, motor vehicle registrations, city and telephone directories, and the National Death Index

53

allocation decided at random

54

N=2 vomited the oral amoxicillin and were given IV 250mg upon attainment of anaesthesia

55

ampicillin was used as first choice, the use of teicoplanain and amikacin combined was required as part of the hospital infection control policy where there were concerns about antibiotic resistant Staphylococcus aureus

56

randomisation was using random number tables, there were three exceptions, extractions which could only be performed if clinically needed, mucoperiosteal flap because of its relative infrequency was studied each time it was needed for treatment of the patient, the third was the cardiac group all of whom had antibiotic prophylaxis and therefore formed a separate group of patients

57

for the study groups, one of the three injection techniques was selected using random number tables

58

only one sample of blood was taken from each child

59

randomisation by random number table

60

some of the staphylococci may be contaminants, it is not possible to identify the skin as a source of contamination without carrying out DNA typing of the isolates and matching them to skin swabs taken at the time of the blood sample

61

Patients were sequentially assigned using computer-generated randomisation tables

62

in addition in group II cotton swabs were used to take smear samples from the surface of the endoscope after the procedure

63

the smear samples from the surface of the endoscope after the procedure were positive in N=29/38 (79%), the organisms were essentially haemolytic Streptococcus or Neisseria

64

for those who were lost to follow-up the authors assumed that patients would have been referred back to them in the event of an episode of endocarditis

65

Randomised using a closed envelope method

66

Allocation by selection of a sealed envelope containing a random number

67

N=1 with staphylococcus epidermis not considered in further analysis

68

Endocarditis diagnosed within 60days of operation was classed as early endocarditis, cases that occurred after 60days were classed as late endocarditis

69

Random code and coded envelopes containing either erythromycin or placebo of identical appearance, the number of the envelope issues to each patient was recorded and numbers decoded only after the trial on side-effects was completed

70

the following organisms were excluded as contaminants unless they were found in both flasks, or in one flask and in significant numbers on the corresponding pour plate; Staphylococcus epidermis, Bacillus species and aerobic diptheroids

71

A cotton swab took smear samples from the throat 30 to 60mins before the procedure

72

The threshold of the growth value indicating a positive result was set at 25 to 30, a change in growth value of >10 to 15 between two consecutive readings was also indicative of a positive result

73

minimum of 3 colonies of the same germ species in ≥ 2/6 media, provided the germ is known to be local flora at the site of intervention; strictly anaerobic germs were isolated, irrespective of the number of their colonies on the media

74

germ count below 3/10ml of blood, germs were not identified in the patient’s secretion and not known to be at the site of intervention, 3 or more colonies of the same germ species grew on one culture medium only, one and the same germ species was found at time 0, 3to5, and 15min

75

The elevated OR for skin infection disappeared after the analysis was restricted to subjects with cardiac valvular abnormalities

76

Adjusted for cardiac valvular abnormality and diabetes

77

Interviewers and medical records abstractors were not blinded but were extensively trained in good interviewing and abstracting techniques

78

Due to the study focusing on indications for an antibiotic prophylaxis

79

Of the 15 case-participants with negative cultures, 12 received antibiotics before admission, 4 had histopathologic evidence of endocarditis, 10 had echocardiographic evidence of valve vegetation

80

Adjusted for socioeconomic status variables (ethnic group, education, occupation, health insurance status, and dental insurance status)

81

Adjusted for socioeconomic status variables (ethnic group, education, occupation, health insurance status, and dental insurance status), diabetes mellitus and severe kidney disease

82

N=7 were given prophylactic antibiotics, their regimen did not follow the recommendations of the AHA

83

Randomisation was based on a single sequence of random assignments created by applying a computer-generated randomisation list

84

The final blood sample could only be obtained from N=50 patients

85

the final blood sample could only be obtained from 50 patients because of technical reasons

86

Population figures of January 1988 adjusted for births, deaths, and migration were used to estimate incidence rates

87

Recommendations for the use of antibiotic prophylaxis to prevent endocarditis were established by a working group from the Netherlands Heart foundation and corresponded with the recommended prophylaxis in the US and GB

88

in accordance with Netherlands Heart Foundation in N=6/8 cases, in the other N=2 the antibiotics could be considered equivalent

89

in accordance with Netherlands Heart Foundation in N=15 controls, in the other N=11 the antibiotics could be considered equivalent

90

stratified for age and certainty of indication

91

stratified for age and certainty of indication

92

Death within 30days after admission or operation was defined as early mortality; late mortality was defined as death after 3months; relapsing endocarditis was defined as persistent infection after termination of antibiotic treatment after a complete course

93

details of randomisation not given

94

Group A β-haemolytic streptococci

95

those with valvular disease had antibiotics according to AHA guidelines