Effectiveness of Short-course Combination Therapy of Cloxacillin with Gentamicin for Right-sided Endocarditis in Intravenous Drug Abusers: A Systematic.

Effectiveness of Short-course Combination Therapy of Cloxacillin with Gentamicin for Right-sided Endocarditis in Intravenous Drug Abusers: A

Systematic Review

Binod Neupane McMaster University

05/10/06 2

A typical schenerio Population: P (e.g., smokers, or drug users, or elderly male) Disease: D (Cancer, Diabetes, MI) Outcome: O (e.g., death, cure)

Available threapies for the treatment of D in P in terms of O: Therapy of Interest (T) vs. Therapy of comparisons (T1, T2, T3)

e.g., T can be any types of monotherapies (only one drug), and T1, T2, T3 can be any, same or different, combination (two or more drugs) therapies for the disease D

Suppose there are 6 studies altogether, none of them was large enough to conclude the effectiveness of any therapy from a single study.

Study 1 (T vs T1): T was more effective

Study 2 (T vs T2): T

2 was more effective

Study 3 (T vs T2): T was more effective

Study 4 (T vs T2): T was more effective

Study 5 (T vs T3): T appeared to be equally effective to T

3

Study 6 (T vs T3): T was more effective

So coflicting evidences......

Which therapy should a physician give to his patients?

A systematic review might help him to make a decision in such situation

05/10/06 3

What is Systematic Review?

Systematic review a scientific investigation in which original studies are its “subjects” it synthesizes the results of multiple primary investigations by limiting bias

and random error

(Qualitative) systematic review: When the results of primary studies are summarized but not statistically

combined (when results of studies can not be combined? How do we know it?)

Meta-analysis (quantitative systematic review): If the statistical methods are used to combine the results of two or more studies

05/10/06 4

Liming bias and random error? How?

Clearly define the population, disease, outcome of interest, and therapy of interest and broadly the therapy of comparison (focussed research problem!)

Identify all the relevant studies done so far, published or unpublished, in any language (thus limiting selection bias, publication bias and language bias)

How to identify them? Develop effective search strategy and search articles in databases

Hand-search individual journals and conference proceedings; scan references of each relevant articles, contact experts and industries

Note: Randomized controlled trials are considered to be best evidences Unknown variables are expected to be controlled

05/10/06 5

Liming bias and random error.....: Example Suppose there are only 5 studies considered in a systematic review:

Study 1 (T vs T1): T is more effective ... ... ... Study 3 (T vs T2): T is more effective Study 4 (T vs T2): T is more effective: poor evidence Study 5 (T vs T3): T appeared to be equally effective to T3

... ... ... Study 7 (T vs T3): T is more effective: poor evidence ... ... ... Total patients in 5 trials: 200, say

Information from three studies are missing or not included in the overview. Does the missing trials suggest the same thing about the effectiveness of therapy T? Does this overview limits bias or random error?

Good systematic review should identify all are 8 studies. Suppose the evidences from them are: Study 1 (T vs T1): T was more effective Study 2 (T vs T2): T2 was more effective (unpublished) Study 3 (T vs T2): T was more effective Study 4 (T vs T2): T was more effective (poor quality: e.g., improper randomization, no stratification) Study 5 (T vs T2): T appeared to be equally effective to T2

Study 6 (T vs T3): T was more effective Study 7 (T vs T3): T was more effective (poor quality: e.g., poor diagnosis: Influenza, or flu-like illness?) Study 8 (T vs T4): T appeared to be equally effective to T4 (published in language other than English) Total patients in 8 trials: 300, say

Random error decreases as sample size (n) increases!!!!!

05/10/06 6

The method of synthesis of information from studies

Suppose, k studies are consider in a systematic review

Suppose, relative risk is the effect measure of ith study (RRi, i = 1,2, ..., k)

If studies are heterogeneous, then generally study samples might have different characteristics or drugs under comparisons are of very different types– Just present the charateristics of patients in all reviewed studies and corresponding effect

measure (Qualitative systematic review)

If studies are similar (more or less homogeneous):– Also pool RR1, RR2, ..., RRk into a single estimate statistically (Meta-analysis)

» Using fixed effect model, or

» Using random effect model

05/10/06 7

Test of Homogeneity

H0: RR1 = RR2 = ...= RRk

H1: At one of them is different from others

Ti = log(estimate of RRi from ith study) = log(observed RRi) linear

Vi = variance

Wi= 1/Vi = weight

T = ∑WiTi/ ∑ Wi, weighted average of Ti

Q = ∑Wi (Ti – T)2 ~ χ2(k-1)

Reject H0 if p < 0.10

– Power of the test is often low

05/10/06 8

Dealing with heterogeneity

Fixed effect moded: Ignore heterogeneity RR

1 = RR

2 = ...= RR

k = RR (= theta), common underlying treatment effect

Ti = log(RRi) = θ + ε

i, i = 1, 2, ..., k

Random effect moded: Incorporate heterogeneity Underlying effect vary from trial to trial, that is, T

i = log(RRi) = θi + ε

i, i = 1, 2, ..., k

where, Ti ~ N(θ, τθ

2), i = 1, 2, ..., k;

θ = log(random effect); τθ2 = between study variance

05/10/06 9

Background

Right-sided endocarditis Dominant IE in intravenous drug abusers (1, 2, 3) Mostly tricuspid valve is involved (1, 4) S. aureus is the dominant infective organism (1, 4, 6, 5)

Diagnosis: Clinical evidence, radiographic finding, and (1) Positive blood culture (three or more) (7, 8, 9) Positive echocardiogram (Duke criteria) (1, 4, 8, 9)

Treatment: In intraveneous drug abusers (IVDAs) with uncomplicated right-sided IE

compliance with lengthy therapy is often a major problem for such population, Short-course antibiotic therapies may be adequate (11, 13, 14, 15)

combination antibiotic therapy may have an enhanced potential synergism (interaction!) when compared to the additive effect of each of the antibiotics assessed separately (18)

Cloxacillin is a semisynthetic penicillin widely used in nonmethicillin resistant Staphylococcus aureus infections.(19) Gentamicin in native valve endocarditis is beneficial in earlier defervescence of fever and the sterilisation of blood cultures.

05/10/06 10

Objective

Objectives: To determine whether a combination therapy of short-course (<= 2 weeks)

cloxacillin and gentamicin compared to any other drug or placebo (combination or monotherapy, short-course or longer-course) is effective in treating right-sided bacterial endocarditis due to S. Aureus in intravenous drug abusers.

05/10/06 11

Study Selection Criteria Patients:

Intravenous drug abusers with bacterial right-sided endocarditis

Diagnosis criteria:

Clinical symptoms and laboratory test (blood culture) and/or echocardiography

Exclusion: Patients with extra-pulmonary metastatic infection

Intervention in one treatment arm Short-course antibiotic therapy of cloxacillin and gentamicin (<= 2 weeks, combination)

Intervention in Comparison arm Any other antibiotic therapy (short- or long-course, single or combination)

05/10/06 12

Selection criteria…

Outcome: At least Clinical Cure at the end of therapy, assessed by clinical symptoms and at least one blood culture.

Follow up: at least 2 weeks of follow-up after the completion of therapy but within 6 months

Design type: (Parallel) Randomized controlled trials (RCTs)

Study year: 1966-present

Setting: Inpatients

Publicaition type: Published or unpublished

Language No restriction

05/10/06 13

Search strategy for the identification of studies

Major Databases: Ovid search of MEDLINE (1996-present) and EMBASE (1980-present) with the following text words, corresponding MeSh and index terms and exploding them

“Endocarditis”

cloxacillin$, gentamicin$,

right$ or tricuspid$ or mitral$ or intravenous$ or injection$ or parenteral$ or substance or drug abuse/,

Human(s)/,

Clinical Trials/, clinical trial(pt), randomized controlled trials/

Combine all “related articles” of relevant articles and limit by “random* or random$”

Hand searches in individual journals and conferences proceedings “Endocarditis” limited to “random*”

Scanning reference lists of relevant atricles, reviews, guidelines

Contact experts, industries

Google search for conference procedings

05/10/06 14

Validation assessment criteria

Reviewers independently examine the articles for quality assessment cross-reference to screen for additional information Make decision of whether to "include" or "not include" Any discrepancy resolved by consensus

Major Criteria for assessment were: Diagnosis: two or more blood culture, and/or two or more echocardiography Appropriate randomization Similarity in baseline characteristics extent of bias in data collection (allocation concealment, observer blinding) follow-up > 80% At least two weeks of follow-up after completion of therapies

05/10/06 15

Validated study from initial searches

3 articles were obtained through MEDLINE search: Long-course (> 2 weeks) monotherapy therapy: 1 study (Fortun 1995)

Short-course antibiotic therapy: 2 (monotherapy: Ribera 1998, and combination: Fortun 2001)

No more randomized controlled trials were found through other sources

05/10/06 16

Characteristics of included studies

Fortun 1995: Short-course (2 weeks, Combination) vs. longer-course (4 weeks, Single) Study carried in a hospital in Spane, published in 1995 Source of funding: Not reported

Methods: Interventions: intravenous

Treatment arm (2 weeks): Cloxacillin (2 g/4h) Gentamicin (1.5 mg/kg body weight/8h)

Comparison arm (4 weeks): Teicoplanin (7 mg/kg/24h)

Randomization issues Method of randomization: 1 of group A: 1 of group B Allocation sequence generation: not reported Allocation concealment: No Observer blinding: No

Follow-up: 2--4 weeks after the end of treatment, and when clinically indicated Outcome assessment: Blood culture

First in 2-5th days, and then in 2-4 weeks after therapy

05/10/06 17

Characteristics of included studies…

Ribera 1998: Short-course (2 weeks, Single) vs. short-course (2 weeks, Combination) Publication Year: 1998 Period of Study: March 1988 to Feb 1993 Study country, setting: Spain, single center (academic hospital), inpatients Source of funding: Not reported

Methods: Interventions (intravenously):

Treatment arm (2 weeks): Cloxacillin (2 g/4h for 2 weeks)

Comparison arm (2 weeks): Cloxacillin (2 g/4h for 2 weeks) Gentamicin (1 mg/kg body weight/8h for first week)

Randomization issues Method of randomization: 5 of group A: 5 of group B (in a set of 10) Allocation sequence generation: Yes (using random number table) Allocation concealment: Not clear (sealed envelope opened at the start of treatment) Observer blinding: No Note: Some patients randomly assigned to treatments before complete screening

Outcome assessment and follow-up: transthoracic echocardiography (at the end); blood culture (2 days after therapy); standard

lab tests, chest X-ray and two blood cultures (2 weeks after therapy); blood cultures if evidence indicated in subsequent visits at 1, 3 and 6 months after the end of therapy.

05/10/06 18

Characteristics of included studies…

Fortun 2001 Publication Year: 2001 Period of Study: 30 months Study country: Spain, single center Source of funding: Not reported

Methods: Interventions (all for 2 weeks except indicated, intravenously):

Treatment arm: Cloxacillin (2 g/4h) Gentamicin (1.5 mg/kg body weight/8h)

Comparison arm I: Vancomycin (500 mg/6h) Gentamicin (1.5 mg/kg body weight/8h)

Comparison arm II: Teicoplanin (12 mg/kg body weight/24 h (first day=24 mg/kg)) Gentamicin (1.5 mg/kg body weight/8h)

Randomization issues Method of randomization: randomly allocated to 1 of 3 therapeutic groups Allocation sequence generation: not reported Allocation concealment: No Observer blinding: No

Follow-up: 12 weeks after the end of treatment Outcome assessment: Blood culture, transthoracic echocardiogram (not performed to all)

05/10/06 19

Meta Analysis

A total of 140 patients were randomized in three studies: Treatment failure: T= Cloxacillin + Gentamicin vs. comparisons

1. Fortun 1995: 2/9 vs. 5/7

2. Ribera 1996: 14/45 vs. 11/45

3. Fortun 2001: .5/11 vs. 10/23

Meta-Analysis (Total: 16/65 vs. 26/75):

RR (lower 95% upper)

Fortun1995 0.31 0.08 1.15

Ribera1996 1.27 0.65 2.49

Fortun2001 0.10 0.01 1.63

RR (Fixed effect)= 0.71, 95% CI = (0.45, 1.12)

Test for heterogeneity: X2(2) = 6.28 (p-value 0.0434)

RR (Random effect)= 0.51, 95% CI = (0.13,1.96)

Test for heterogeneity: X2(2) = 5.88 (p-value 0.0529)

Estimated random effects variance: 0.89

05/10/06 20

Forest Plot

05/10/06 21

Forest Plot...

05/10/06 22

Conclusion of the meta-analysis

Based on the pooled estimate from the meta-analyses using both the fixed and random effect models, the short-course combination therapy of Cloxacillin and Gentamicin did not appear to be effective in the treatment of right-sided endocarditis in intravenous drug abusers

However, there are substantial heterogeneity between study results (Chi-square test of homogeneity, p < .10) This might be due to different treatment duration or due to single and combination

therapy in comparison arms or due to both.

Funnel plot to assess the publication bias could not be produced due to small number of studies.

05/10/06 23

Subgroup Analysis

To address the heterogeneity in the study results, we planned to do separate analysis for each group of long-course and single, long course and combination, short-course and single, short-course and combination, therapies used in the comparison arms

05/10/06 24

Subgroup Analysis: Long-course monotherapy

1 study (Fortun 1995)

Cloxacilliin + Gentamicin (2 weeks) vs. Teicoplanin (4 weeks) Treatment failure: 2/9 vs. 5/7

Observed RR = RR = 0.31 95% CI ( 0.15, 0.66 )

Conclusion: large effect size: short-course treatment of right-sided endocarditis with

combination of Cloxacillin and Gentamicin may be better regimen than longer course therapy of teicoplanin

However, very small study size, In sufficient data and follow-up period was relatively short

Trial had to be stopped prematurely due to adverse events in comparison arms

05/10/06 25

Subgroup Analysis: Short-course monotherapy

1 study (Ribera 1996) Cloxacilliin + Gentamicin (2 weeks) vs. Cloxacillin (2 weeks) Treatment failure: 14/45 vs. 11/45

Observed RR = 1.27 and 95% CI = ( 0.63, 2.57 ) Conclusion:

Drugs under comparison are had similar efficacy

Trial was of moderate size (90 randomized), follow-up period was adequate

05/10/06 26

Subgroup Analysis: Short-course Combination therapy

Separate A: Cloxacilliin + Gentamicin vs. Vancomycin + Gentamicin

0.5/11 vs. 5/11 Observed RR = RR =0.1 95% CI ( 0.01, 0.79 )

B: Cloxacilliin + Gentamicin vs. Teicoplanin + Gentamicin 0.5/11 vs. 5/12 Observed RR = 0.11 95% CI ( 0.01, 0.93 )

Combined Cloxacilliin + Gentamicin vs. (Vancomycin or Teicoplanin) + Gentamicin

0.5/11 vs. 10/23 Observed RR = 0.1 95% CI ( 0.01, 0.82 )

Conclusion : Cloxacillin + Gentamicin may be effective, however CI of observed effect size

RR is relatively wide Insufficient data, a size of 34 is small

Trial had to be stopped prematurely due to adverse events in comparison arms

05/10/06 27

References1. Moss R and Munt B. Injection drug use and right sided endocarditis. Heart 2003; 89:577-581

2. Crane LR, Levine DP, Zervos MJ, et al. Bacteremia in narcotic addicts at the Detroit Medical Center. I. Microbiology, epidemiology, risk factors, and empiric therapy. Rev Infect Dis 1986; 8:364-73

3. Robbins MJ, Soeiro R, Frishman WH, Strom JA. Right-sided valvular endocarditis: etiology, diagnosis, and an approach to therapy.Am Heart J 1986, 111:128-135.

4. Prendergast B. The changing face of infective endocarditis. Heart online. Oct 10, 2005

5. Hecht SR, Berger M. Right-sided endocarditis in intravenous drug users. Prognostic features in 102 episodes. Ann Intern Med. 1992; 117:560-6.

6. Chambers HF, Korzeniowski OM, Sande MA. Staphylococcus aureus endocarditis: clinical manifestations in addicts and nonaddicts. Medicine 1983;62:170-7.

7. DeWitt DE, Paauw DS. Endocarditis in injection drug users. Am Fam Physician 1996; 53(6): 2045-9

8. Horstkotte D, Follath F, Gutschik E, et al. Guidelines on prevention, diagnosis and treatment of infective endocarditis: Executive summary—The Task Force on infective endocarditis of the Europian Society of Cardiology

9. Elliott TSJ, Foweraker J, Gluld FK, et. Al. Guidelines for the antibacterial treatment of endocarditis in adults: report of the working party of British Society for Antimicrobial Chemotherapy. J Antimicrobial Chemotherapy 2004; 54:971-81

05/10/06 28

References10. Heldman AW, Hartert TV, Ray SC, Daoud EG, Kowalski TE, Pompili VJ, et al. Oral antibiotic treatment of right-sided

staphylococcal endocarditis in injection drug users: prospective randomized comparison with parenteral therapy. Am J Med 1996;101:68-76.

11. DiNubile MJ. Short-course antibiotic therapy for right-sided endocarditis caused by Staphylococcus aureus in injection drug users. Ann Intern Med 1994;121:873-6.

12. Bayer AS, Norman DC. Valve site-specific pathogenic differences between right-sided and left-sided bacterial endocarditis. Chest. 1990; 98:200-5.

13. Chambers HF, Miller RT, Newman MD. Right-sided Staphylococcus aureus endocarditis in intravenous drug abusers: two-week combination therapy. Ann Intern Med. 1988; 109:619-24.

14. Torres-Tortosa M, deCueto M, Vergara A, Sanchez-Porto, Perez-Guzman, Gonzalez-Serrano M, et al. Indications and therapeutic results of an antibiotic regime lasting two weeks in intravenous drug users with right-sided S. aureus infective endocarditis: a multicentre study of 139 consecutive cases. Eur J Clin Micro Infec Dis. 1994; 13:533-4.

15. Baddour LM, Wilson WR, Bayer AS, et al. Infective Endocarditis: Diagnosis, Antimicrobial Therapy, and Management of Complications: A Statement for Healthcare Professionals From the Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease, Council on Cardiovascular Disease in the Young, and the Councils on Clinical Cardiology, Stroke, and Cardiovascular Surgery and Anesthesia, American Heart Association. Circulation 2005; 111(23):e394-e434

16. Devlin RK, Andrews MM and Reyn CF. Recent trends in infective endocarditis: influence of case definitions. Current Opinion on Cardiology 2004, 19: 134-9

17. Palepu A. Cheung SS, Montessori V, et al. Factors other than the Duke criteria associated with infective endocarditis among injection drug users.Clin Invest Med. 2002; 25(4):118-25.

1. Le T and Bayer AS. Combination antibiotic therapy for infective endocarditis. Clinical Infectious Diseases 2003; 36:615-21

2. Dominguez-Ortega J, Martinez-Alonso JC, Marcos-Perez MC, Kindelan C, Frades A. Allergy to cloxacillin with normal tolerance to amoxicillin and cefuroxime. Allergol Immunopathol (Madr). 2006 Jan-Feb;34(1):37-8

05/10/06 29

References of included studies1. Short vs. longer therapy:

1. Combination vs. Single 1. Fortun 1995

Fortun J, Perez-Molina JA, Anon MT, et al. Right-sided endocarditis caused by staphylococcus aureus in drug abusers. Antimicrobial Agents and Chemotherapy 1995; 39(2): 525-8

2. Short vs. short therapy:1. Single vs. combination

1. Ribera 1998Ribera E, Gomez-Jimenez J, Cortes E, et al. Effectiveness of cloxacillin with and without gentamicin in short-term therapy for right-sided staphylococcus aereus endocarditis. Ann Intern Med 1996; 125(12): 969-74

1. Combination vs. combination 1. Fortun 2001

Fortun J, Navas E, Martinez-Beltran J, et al. Short-course therapy for right-sided endocarditis due to staphylococcus aureus in drug abusers: Cloxacillin versus glycopeptides in combination with gentamicin. Clinical Infectious Diseases 2001; 33:120-5

Thank You!