nejm protocol sample 1

86
Protocol This trial protocol has been provided by the authors to give readers additional information about their work. Protocol for: Albert RK, Connett J, Bailey WC, et al. Azithromycin for prevention of exacerbations of COPD. N Engl J Med 2011;365:689-98. (PDF updated November 21, 2011.)

Upload: razoraz

Post on 30-Dec-2015

59 views

Category:

Documents


3 download

DESCRIPTION

Sample Research Protocol

TRANSCRIPT

Page 1: NEJM Protocol Sample 1

Protocol

This trial protocol has been provided by the authors to give readers additional information about their work.

Protocol for: Albert RK, Connett J, Bailey WC, et al. Azithromycin for prevention of exacerbations of COPD. N Engl J Med 2011;365:689-98.

(PDF updated November 21, 2011.)

Page 2: NEJM Protocol Sample 1

1

Azithromycin for Prevention of COPD Exacerbations

Complete Protocol

Section A. Complete protocol………………………………………………. 1-71

Section B. Protocol changes……………………………………………… 72-72

Section C. References…………………………………………………….. 73-85

Page 3: NEJM Protocol Sample 1

2

Section A. Complete protocol

Azithromycin for Prevention of COPD Exacerbations

Summary:

The objective of this proposal is to determine if chronic administration of a

macrolide antibiotic will reduce the morbidity of COPD.

The rationale supporting the importance of this study is that (1) the

prevalence, morbidity, mortality and treatment cost of COPD are high and

increasing; (2) a large fraction of the morbidity and cost is attributable to acute

exacerbations; (3) macrolide antibiotics have a variety of antibacterial and

antiinflammatory properties that, in theory, could reduce acute exacerbations;

and (4) long-term administration of macrolide antibiotics has resulted in clinically

important improvements in patients with a number of other pulmonary disorders.

The specific aim is to determine if administration of azithromycin for one year

will decrease the frequency and/or the severity of COPD exacerbations.

The research design is a prospective, randomized, double-blind, placebo-

controlled clinical trial.

The methods involve selecting 1130 patients with at least moderately severe

COPD who, based on clinical indicators, have an increased likelihood of

experiencing an acute exacerbation during the study period. Exclusion criteria

include a variety of conditions or medications known to adversely interact with

macrolides. Monitoring will occur monthly and include a careful assessment of

possible macrolide-related side effects. The primary endpoint is time to first

Page 4: NEJM Protocol Sample 1

3

COPD exacerbation. Secondary endpoints include the incidence of macrolide-

resistant bacterial colonization, quality of life and cost-effectiveness.

Research Design

This is a prospective, randomized, double-blind, placebo-controlled trial that

will enroll 1130 patients with at least moderately severe COPD over a two year

period and follow them at monthly intervals for one year (Figure 1). The primary

endpoint is time to first acute exacerbation. Secondary endpoints are described

below.

Hypothesis

Administration of azithromycin for one year will decrease the frequency and

the severity of COPD exacerbations when added to the usual care of these

patients.

Figure 1. Patient Contact Flow Diagram (Time in Months)

0 1 2 3 4 5 7 8 9 10 11 12 13 6

Clinic Visit

Enroll

Phone Contact

Clinic Visit

Clinic Visit

Clinic Visit

Clinic Visit

Phone Contact

Washout Visit

Phone Contact

Phone Contact

Page 5: NEJM Protocol Sample 1

4

Inclusion Criteria

1. Male and female subjects, ≥ 40 years of age

2. Clinical diagnosis of at least moderate COPD as defined by the GOLD

criteria (Pauwels, 2001):

a. Postbronchodilator FEV1/FVC < 70%,

b. Postbronchodilator FEV1 < 80% predicted, with or without chronic

symptoms (i.e., cough, sputum production).

3. Cigarette consumption of 10 pack-years or more. Patients may or may

not be active smokers.

4. To enrich the population for patients who are more likely to have acute

exacerbations (Niewoehner, 2004), each subject must meet one or more

of the following 4 conditions

a. Be using supplemental O2. or will have a history of

b. Receiving a course of systemic corticosteroids for respiratory

problems in the past year,

c. Visiting an Emergency Department for a COPD exacerbation within

the past year, or

d. Being hospitalized for a COPD exacerbation within the past year

5. Willingness to make return visits and availability by telephone for duration

of study.

Page 6: NEJM Protocol Sample 1

5

Exclusion Criteria

1. A diagnosis of asthma established by each study investigator on the basis

of the recent American Thoracic Society/European Respiratory Society

guidelines (Table 3).

If, after applying the above criteria, the clinicians are still unsure about the

distinction in a specific patient bronchodilator testing with inhaled albuterol will be

performed and patients with changes in FEV1 > 400 mL will be excluded.

Table 3: Clinical Features Differentiating COPD & Asthma (Modified from

NICE Guidelines, 2004)

History COPD Asthma

Smoker or ex-smoker Nearly all Possibly

Symptom onset < 35 yrs Rare Common

Chronic productive cough Common Uncommon

Breathlessness Persistent and

progressive

Variable

Nighttime waking with

breathlessness and wheeze

Uncommon Common

Significant diurnal or day-to-

day variation of symptoms

Uncommon Common

Page 7: NEJM Protocol Sample 1

6

2. The presence of a diagnosis other than COPD that results in the patient

being either medically unstable, or having a predicted life expectancy < 3

years.

3. Special patient groups: prisoners, pregnant women, institutionalized

patients

4 Women who are at risk of becoming pregnant during the study (pre-

menopausal) and who refuse to use acceptable birth control (hormone-

based oral or barrier contraceptive) for the duration of the study.

5. Patients with a history of hypersensitivity to any macrolide antibiotic.

6. Patients taking any of the following medications

a. Cisapride (which has been removed from the U.S. market)

b. Ergot derivatives: Cafergot, Ergomar, Wigraine, Migrainal,

D.H.E.45

c. Pimozide (Orap®)

d. Disopramide (Norpace®)

e. Cyclosporin (Gengraf™; Neoral®; Restasis™; Sandimmune®)

f. Tacrolimus (Prograf®; Protopic®)

g. Nelfinavir (Viracept®)

h. Bromocriptine (Parlodel®)

i. Hexobarbital (Evipan®; Hexenal®; Hexobarbitone®)

7. A manually determined QTc interval measured at least one hour after use

of any short-acting inhaled β2 agonist that exceeds 450 ms on two

Page 8: NEJM Protocol Sample 1

7

occasions separated by at least one week. Manual determinations are

only to be done if the automated QTc estimate exceeds the 450 ms limit.

The method that should be used to determine the QTc interval manually is

to measure the mean QT interval from the beginning of the QRS complex

to the end of the T wave, in a minimum of 3 cardiac cycles, in leads II and

V5 or V6 (using the lead in which the QT is the longest). This QT interval

is then corrected for heart rate using Federicia’s formula (QTc=

QT/RR(sec)½).

Bundle branch blocks prolong the QT interval. Accordingly, patients with

either left or right bundle branch blocks should have the JT interval

measured. Subjects with a JT interval > 420 msec (corrected for heart

rate) should be excluded.

It is recommended that a cardiologist specializing in arrhythmias and/or

ECG interpretation perform the above manual analysis and evaluate all

ECGs of patients with bundle branch blocks.

Patients should also be excluded if they have a history suggesting

increased risk for developing torsade de pointes (e.g., heart failure,

hypokalemia, family history of Long QT syndrome), or if they are taking

other medications that prolong the QT interval (table).

Page 9: NEJM Protocol Sample 1

8

Amiodarone Cordarone® Anti-arrhythmic / abnormal heart

rhythm

Females>Males,TdP risk

regarded as low

Amiodarone Pacerone® Anti-arrhythmic / abnormal heart

rhythm

Females>Males,TdP risk

regarded as low

Arsenic trioxide Trisenox® Anti-cancer / Leukemia

Bepridil Vascor® Anti-anginal / heart pain Females>Males

Chloroquine Arelan® Anti-malarial / malaria infection

Chlorpromazine Thorazine® Anti-psychotic/ Anti-emetic /

schizophrenia/ nausea

Cisapride Propulsid® GI stimulant / heartburn Restricted availability;

Females>Males.

Clarithromycin Biaxin® Antibiotic / bacterial infection

Disopyramide Norpace® Anti-arrhythmic / abnormal heart

rhythm Females>Males

Dofetilide Tikosyn® Anti-arrhythmic / abnormal heart

rhythm

Domperidone* Motilium® Anti-nausea / nausea not available in the United

States

Droperidol Inapsine® Sedative;Anti-nausea / anesthesia

adjunct, nausea

Erythromycin Erythrocin® Antibiotic;GI stimulant / bacterial

infection; increase GI motility Females>Males

Erythromycin E.E.S.® Antibiotic;GI stimulant / bacterial

infection; increase GI motility Females>Males

Halofantrine Halfan® Anti-malarial / malaria infection Females>Males

Haloperidol Haldol® Anti-psychotic / schizophrenia,

agitation

Page 10: NEJM Protocol Sample 1

9

Ibutilide Corvert® Anti-arrhythmic / abnormal heart

rhythm Females>Males

Levomethadyl Orlaam® Opiate agonist / pain control,

narcotic dependence

Mesoridazine Serentil® Anti-psychotic / schizophrenia

Methadone Methadose® Opiate agonist / pain control,

narcotic dependence Females>Males

Methadone Dolophine® Opiate agonist / pain control,

narcotic dependence Females>Males

Pentamidine NebuPent® Anti-infective / pneumocystis

pneumonia Females>Males

Pentamidine Pentam® Anti-infective / pneumocystis

pneumonia Females>Males

Pimozide Orap® Anti-psychotic / Tourette's tics Females>Males

Procainamide Pronestyl® Anti-arrhythmic / abnormal heart

rhythm

Procainamide Procan® Anti-arrhythmic / abnormal heart

rhythm

Quinidine Quinaglute® Anti-arrhythmic / abnormal heart

rhythm Females>Males

Quinidine Cardioquin® Anti-arrhythmic / abnormal heart

rhythm Females>Males

Sotalol Betapace® Anti-arrhythmic / abnormal heart

rhythm Females>Males

Sparfloxacin Zagam® Antibiotic / bacterial infection

Thioridazine Mellaril® Anti-psychotic / schizophrenia

Page 11: NEJM Protocol Sample 1

10

8. Patients taking rifabutin or rifampin (as these medications reduce

macrolide plasma levels).

9. Patients with chronic hepatic insufficiency (defined as an INR > 2.0 for

subjects not taking warfarin, a serum albumin < 3.0 g/dL, or levels of

aspartate and alanine aminotransferases ≥ three times the normal value).

10. Patients with chronic renal insufficiency (defined as a serum creatinine >

1.5 mg/dL or an estimated creatinine clearance < 20 mL/min using the

simplified equation derived from the Modified of Diet in Renal Disease

study [i.e., GFR, in mL/min per 1.73 m2 = 186.3 x ((serum creatinine) -1.154)

x (Age-0.203) x (0.742 if female) x (1.21 if African American) (Levey, 1999;

2000).

11. Patients otherwise meeting the inclusion criteria will not be enrolled until

they are a minimum of four weeks from their most recent acute

exacerbation (i.e., they will not have received a course of systemic

corticosteroids, an increased dose of chronically administered systemic

corticosteroids, and/or antibiotics for an acute exacerbation for a minimum

of four weeks).

12. A clinical diagnosis of bronchiectasis defined as production of > one-half

cup of purulent sputum/day.

13. Patients with age-adjusted hearing thresholds ≥ 95th percentile at any one

of 500, 1000, 2000, and 4000 Hz in either ear on initial screening

audiometry, or on formal audiometry testing if the latter is obtained

Page 12: NEJM Protocol Sample 1

11

Outcome Measures

The primary endpoint of this study is the time to first occurrence of an acute

COPD exacerbation during the one-year treatment period.

Secondary outcomes are:

1. The number of acute exacerbations occurring within 12 months of

randomization

2. The number of ED visits resulting from acute exacerbations

3. The number of hospital admissions resulting from acute exacerbations

4. The number of hospital days resulting from acute exacerbations

5. The incidence of presumed macrolide-related side-effects

6. The incidence of presumed macrolide-related side effects that require

cessation of treatment

7. The incidence of macrolide-resistant bacterial colonization of the

nasopharynx or sputum

8. The incidence of pneumonia or acute bronchitis in patients who do and do

not become colonized with macrolide-resistant bacteria

9. Quality of life

10. Cost-effectiveness

Acute exacerbations are defined as a "complex of respiratory symptoms

(increase or new onset) of more than one of the following: cough, sputum,

wheezing, dyspnea, or chest tightness with a duration of at least three days

Page 13: NEJM Protocol Sample 1

12

requiring treatment with antibiotics and/or systemic steroids " (Niewoehner,

2004).

Acute exacerbations will be graded according to the following scale:

1. Mild (home management, with or without contacting a health care

provider)

2. Moderate (requiring a visit to an Emergency Department)

3. Severe (requiring hospitalization)

4. Very severe (requiring intubation and mechanical ventilation)

When patients report that they have had a moderate, severe or very severe

exacerbation the hospital records of that encounter will be obtained and

reviewed.

Statistical Design

Sample Size

The primary endpoint is time to first acute COPD exacerbation within 12

months of randomization. Because Niewoehner and colleagues (2004) found

that identically selected patients had an incidence of acute acute exacerbations

of 38% over 6 months, we anticipate that approximately 50% of the subjects in

our Control group will have an exacerbation within 12 months. Previous data

also indicate a one-year death rate of approximately 3% in similarly selected

patients. The treatment effect and the ‘dropout rate’ (i.e., non-compliance with

Page 14: NEJM Protocol Sample 1

13

study medication) are not known. Table 1 shows the required total projected

sample size for the study, based on plausible assumptions for these parameters.

The projected reduction in event rates from the Control group to the

Azithromycin-treated group is the most important determinant of sample size. A

25% reduction (50% vs. 37.5%) in perfect compliers yields a required sample

size of 724, assuming there is a 20% noncompliance rate with the assigned

medication at the end of a year, and that 6% of participants either die or are lost

to follow-up. But a 20% reduction, with the other assumptions not changed,

requires that 1130 subjects be randomized. Based on the findings of Sin and

colleagues (2003) with respect to the effect of other treatments on acute COPD

exacerbations, we believe that a 20% treatment effect in perfect compliers is

more likely, so our total sample size goal is N = 1130.

Randomization Method

Table 1: Total Required Sample Size Under Various Assumptions (Two-sided

significance level 0.05, power = 85%)

Event Rates

(Control vs. Azithromycin, Death + Loss Total Sample

Perfect Compliers) Non-Compliance Rate To Followup Rate Size

50% vs 40% 0% 6% 870

50% vs 40% 5% 6% 957

Page 15: NEJM Protocol Sample 1

14

Randomization was stratified by each site with separate schedules created for

each. A matched schedule was retained in each site's pharmacy, where

separate supplies of capsules containing study drug or placebo were also kept.

Only the pharmacist and the staff of the Data Coordinating Center (DCC) knew

this schedule and the pharmacists could not know the actual assignment until

after the DCC specified an accession number. Treatment assignment was only

disclosed in cases of emergency.

Randomization will be carried out by linking to the Data Coordinating Center

through a website, (http://www.copdcrn.org) using a required COPD Clinical

Research Network (CCRN) User ID and password. After securing entry a menu

listing the clinical sites appears. The user must choose one of these – for

example, the Health Partners site is labeled as ‘Center 3b: Health Partners,

Minnesota CCRN’. After choosing a clinical site and validating the identity of the

clinic, a second menu appears that lists a series of activities (e.g., ID assignment,

Data Entry, Error Correction, Randomization, Adverse Event Reporting, etc.). To

randomize a patient into this protocol the coordinator clicks on the hyperlink

labeled “Randomization” and from there, to the hyperlink labeled “Macrolide

Randomization”.

The program will verify that eligibility criteria are met. Verification

necessitates that data from all forms bearing on eligibility have been completed

and transmitted to the Data Coordinating Center including:

1. Demographics

2. Physical exam

Page 16: NEJM Protocol Sample 1

15

3. Medical history

4. Medications

5. Lung function

6. Informed consent

Randomization will be halted if required data are missing or if eligibility criteria

are not met. If any of these forms are not in the DCC database, or if any

eligibility criteria contained in them are not met, a message will appear explaining

what the problem is and randomization will be halted. The associated treatment

assignment will not be issued until all criteria are met.

Each participant may receive only one randomization assignment.

Randomization is stratified by each designated site in each clinical center. As an

example, the Minneapolis clinical center has three sites: the VA Hospital in

Minneapolis, Health Partners, and the Mayo Clinic. Separate schedules are

created for each.

If all eligibility criteria are met the randomization program will issue a

treatment assignment number such as ‘113’. This number matches a schedule

that is retained in each site's clinical pharmacy, where separate supplies of

capsules containing the active drug and the placebo are also kept. The only

person knowing this schedule will be the pharmacist and the DCC. When the

clinic coordinator requests pills for treatment assignment #113, the pharmacist

will check the schedule and distribute the assigned drug from the appropriate

pharmacy supply. The actual assignment will only be revealed in cases of

emergencies where caregivers need to know what drugs the person was taking

Page 17: NEJM Protocol Sample 1

16

to provide treatment, or to avoid prescribing other medications that might

adversely interact with the study drug.

Active-drug and placebo capsules will be identical in appearance. Clinic staff

will make no attempt to determine the content of any capsules except in cases of

emergency.

It will be clearly explained to each study participant that they will be assigned

to use either active drug (azithromycin) or placebo, that the treatment assignment

is random (i.e., cannot be predicted in advance, like the outcome of a coin-flip),

and that they should not attempt to discover the content of the capsules.

Each time the participant returns for a renewed supply of capsules, the clinic

coordinator must request that the pharmacist distribute pills for that specific

participant’s treatment assignment number, referring to the schedule in the

pharmacy.

If the participant’s treatment assignment number is lost or forgotten, the

information can be obtained by contacting the pharmacy or the DCC.

Data Analysis

An appropriate analysis for the primary endpoint is an intent-to-treat analysis

based on life-table (i.e., survival analysis) methods. To adjust for differences in

pre-randomization factors such as age, gender, prior history of exacerbations,

lung function, and other important predictors of exacerbation, a Cox proportional-

hazards model analysis will also be carried out, with time-to-first-exacerbation as

Page 18: NEJM Protocol Sample 1

17

the outcome variable and treatment group (i.e., azithromycin vs placebo) as the

primary variable of interest.

Additional analyses will be conducted in which the variable of interest is the

repeated occurrence of COPD exacerbations. The methodology in this case will

be either Poisson regression, or repeated measures regression for categorical

data (implemented in PROC GENMOD in SAS). This analysis also permits

entering other variables in the analysis.

Secondary analyses will be carried out in which the effects of compliance with

the study medication are taken into account. This can be done using Cox

regression with time-dependent covariates (for time-to-event outcomes), or a

generalized estimating equation approach. Both permit the inclusion of time-

varying risk factors.

Additional secondary analyses will be conducted to detect differences in

severity of COPD exacerbations. These will include (1) time-to-event analyses

for the more severe outcomes (i.e., exacerbations requiring hospitalization or

resulting in death), and (2) analyses based on scoring of the severity of

exacerbation as described above. These analyses will be carried out using

repeated measures analyses based on generalized estimating equation (GEE)

methods, which can take the effects of baseline risk factors into account.

Analyses of other secondary outcomes (side-effects, bacterial colonization,

secondary bacterial infections, etc) will be done similarly using either survival-

analytic approaches to time-to-event data, or repeated measures analyses.

Page 19: NEJM Protocol Sample 1

18

Sequential Analysis and Monitoring Boundaries

Because approximately 1130 participants will be randomized over an

estimated period of two years and will receive either a macrolide antibiotic

(AzithromycinTM) or a matched placebo for a period of one year, the study will

take a minimum of three years to complete.

The Data and Safety Monitoring Board (DSMB) will meet approximately every

6 months to review recruitment, follow-up rates, compliance, safety and efficacy.

Repeated reviews of endpoint data involve the problem of multiple statistical

testing performed on a set of accumulating data. As a solution to this problem

we will adopt a group sequential method of analysis related to that proposed in

the 1970s (O'Brien, 1979). Such procedures typically require large critical values

(or boundaries) early in the trial that decrease as the trial progresses. Because

of the conservatism early in the trial, the critical value of the final analysis is close

to the “nominal” critical value. The specific method we propose is a general

approach to group sequential testing developed by Kim and DeMets (1987) for

which neither the number of looks nor the increments between looks need to be

pre-specified. Rather, the Kim-DeMets approach requires only specification of

the rate at which the Type I error (which here will be chosen to be α = 0.05) will

be “spent”. This procedure allows “spending” a little of the α at each interim

analysis in such a way that, at the end of the study, the total Type I error does

not exceed 0.05.

Page 20: NEJM Protocol Sample 1

19

The exact sequential monitoring plan ultimately adopted for the trial must be

approved by the DSMB before any looks at interim data occur. The following is

an example of the kind of plan we will propose:

We expect to present efficacy-related analyses to the DSMB beginning 12

months after the first patients are randomized, and continuing at 6-month

intervals thereafter until all patients have completed one year of follow-up. Thus,

there will be interim looks at efficacy data at the time-points and corresponding

information times indicated in Table 2. A final analysis will be conducted at 36

months. Two-sided

tests of significance will be assumed. For this trial, if randomization occurs at a

uniform rate over a two-year period, and the hazard of events is constant with the

probability that an exacerbation occurs within 12 months of randomization, the

following table illustrates the cumulative information time at each of the DSMB

looks at the data [we have chosen an alpha-spending function of the form

f(t) = α * tρ,

Table 2. Sequential Analysis

Information Time 0.00 0.25 0.50 0.75 0.88 1.00

Real Time 0 M 12 M 18 M 24 M 30 M 36 M

Upper Boundary +3.36 +2.76 +2.36 +2.23 +2.07

Lower Boundary - 3.36 -2.76 -2.36 -2.23 -2.07

Nominal 2-sided p-values -0.0008 0.0058 0.0182 0.0258 0.0384

Page 21: NEJM Protocol Sample 1

20

where α is the two-sided significance level (α = 0.05), and t is information time

(0 < t ≤ 1)]. We have specified the exponent ρ = 3.0, which produces

conservative boundaries for Z-statistics (Table 5).

Use of the alpha-spending function approach to sequential monitoring has the

advantage that, if necessary, additional looks at the data can be accommodated

without affecting the overall probability of Type I error (α = 0.05). Thus, if

recruitment takes longer than expected, analyses of the data for DSMB meetings

may occur at information times which are different from those shown above.

A disadvantage of the alpha-spending function approach (versus doing a

single analysis at the end) is that, although it preserves the overall Type I error

level, it results in a small decrease in power (Jennison, 2000). We have

specified a total sample size of N = 1130 for this trial to achieve 85% power, as

shown in Table 3, on the assumption that one significance test will be performed

at the end, and this is probably near the limit of what the participating clinical

centers will be able to recruit in a 2-year period. Taking into account the

decrease in power associated with this sequential monitoring plan, the power of

the trial will be approximately 83% rather than 85%.

Judgment concerning the continuation or termination of the study will involve

not only the degree of statistical significance observed at the interim analysis, but

also the likelihood of achieving significance should enrollment continue to the

original projected sample size. As an aid in this assessment, the DCC will

supplement the group sequential analysis outlined above with calculations of

conditional power based on the method of stochastic curtailment (also known as

Page 22: NEJM Protocol Sample 1

21

futility analysis, Halperin, 1982; Lan, 1982; Ware, 1985). This procedure

evaluates the conditional probability that a particular statistical comparison will be

significant (or not significant) at the end of the trial at the α level used in the

design, given the hypothesized treatment difference and the endpoint data

accumulated to date. Conditional power for the primary endpoint will be

computed and provided to the DSMB as part of the interim study reports, and will

include calculations based on the originally hypothesized treatment difference as

well as the observed treatment difference up to that point in the trial.

The DSMB will have a particularly valuable role in its recommendations to the

Steering Committee and the NHLBI if there emerges any statistically extreme

benefit or harm. The DSMB will need to put any such interim in proper

perspective. If protocol modifications are warranted, close consultation among

the Steering Committee, the DSMB, and the NHLBI staff will be needed.

Page 23: NEJM Protocol Sample 1

22

Schedule of Study Interventions (Figure 2)

Enrollment (two visits)

1. Demographics (including age, gender, ethnic group, smoking history, 02

use were obtained by history), information regarding where the patient

might be seen in an emergency department (ED) and/or hospitalized,

should this be required during the course of the study, primary care

physician name and contact information.

2. A medical history and physical examination that includes recording a list of

all medications, allergies, menopausal status (women), whether the

Page 24: NEJM Protocol Sample 1

23

patient has received influenza and/or pneumococcal pneumonia

vaccination and if so, when, whether they have subjective hearing loss, a

list of all coexisting diseases or conditions, whether they have chronic

bronchitis (defined in American Thoracic Society, 1987), previous COPD-

related ED visits and hospitalizations (including whether they have

previously required invasive or non-invasive mechanical ventilation), a

BMI determination, and a resting SpO2 on the prescribed amount of

supplemental O2 being given, if any, or on air.

3. FEV1 and FEV6 measured 15 min after inhalation of two puffs of albuterol

(Ferguson, 2000; Hankinson, 2003).

4. An ECG

5. The Hearing Handicap Inventory for the Elderly (screening version)

questionnaire. Those with scores ≥ 10 will be referred for formal

audiometry testing. All others will undergo a screening audiogram (with

testing at 500, 1000, 2000, and 4000 Hz). Patients with ≥ 95th percentile

age-adjusted hearing thresholds at 500, 1000, 2000, or 4000 Hz in either

ear on either screening audiograms or formal audiometry testing will be

excluded unless there is an obvious extenuating explanation (e.g.,

complete cerumen impaction). Those with ≥ 75th and < 95th percentile

age-adjusted hearing thresholds at any two frequencies in either ear on

screening audiometry will undergo formal audiometry testing to establish

their basal level of hearing.

Page 25: NEJM Protocol Sample 1

24

6. Quality of life assessment. General quality of life will be assessed by the

Medical Outcomes Study 36-Item Short Form (the SF-36) to allow the

results to be compared with those in other studies. General quality of life

will also be assessed by the Quality of Well-Being Scale to allow

estimation of quality-adjusted life years (QALYs), an important element for

cost-utility analysis (see below), and to assess the net result of possible

gains from the treatment being offset by side-effects that adversely affect

health-related quality of life. Sleep quality will be assessed using the

Pittsburgh Sleep Quality Index (PSQI). Disease-specific quality of life will

be assessed with the St. George's Respiratory Questionnaire, an

instrument that was developed and validated in patients with COPD and

may be more responsive to the intervention planned than the two general

instruments being administered. A change of ≥ 4 units between the time

point at which the SGRQ is assessed will be considered the minimal

clinically important difference and the percent of patients achieving ≥ 4

unit change will also be assessed.

7. Cost effectiveness will be assessed by calculating the ratio of incremental

costs to the ratio of incremental QALYs. The same methods used to

estimate the cost effectiveness of lung volume reduction surgery (National

Emphysema Treatment Trial Research Group, 2003) will be employed.

QALYs will be determined from the Quality of Well-Being questionnaire as

noted above.

Page 26: NEJM Protocol Sample 1

25

8. Depression screen. All patients will be given the Hospital Anxiety and

Depression Scale (HADS) survey. Patients with depression subscale

scores of 11 or greater who are receiving their care from one of the

enrolling centers will be referred for depression evaluation and/or

treatment. Those receiving their care from providers outside the enrolling

centers will, with the patient's permission, have these providers informed

of the result of the survey, with the suggestion that they might consider

referral for evaluation and/or treatment

9. Laboratory studies to include:

a. Complete blood count and platelet count

b. Blood urea nitrogen and creatinine

c. Liver function tests (i.e., aspartate and alanine aminotransferases,

alkaline phosphatase, albumin, total bilirubin, and INR)

d. Nasopharyngeal swabs and, when possible, expectorated sputum

for culture. Sensitivity testing will be done on all pathogens

identified.

e. Blood will be drawn for measurement of several biomarkers of

inflammation. (See biomarker protocol addendum.)

f. Blood will also be drawn and stored for future analyses after the

white blood cells and the DNA are extracted and the serum is

aliquoted. (See genetic protocol addendum.)

g. An α-1-antitrypsin level should either be available in the chart or

obtained at the time of admission.

Page 27: NEJM Protocol Sample 1

26

h. HCG (urine or blood) for pre-menopausal women.

10. Basic education about COPD, its treatment and how to recognize acute

exacerbations. Patients will also be given a diary to assist with recalling

acute exacerbations that are treated at home, without contact with a

health care provider.

11. Azithromycin, 250 mg tablets, or an identically appearing placebo, with

instructions that the medication should be taken daily.

12. A wallet card indicating that they are participating in the study, may be

receiving azithromycin, and that they should not receive any of the

medications listed in the exclusion criteria noted above. The card will also

contain instructions about stopping the study medication when they are

treated for an acute exacerbation with any antibiotic, and about restarting

the study medication after the exacerbation has resolved (defined as

being one week after completing the course of antibiotics). A letter

containing similar information, along with basic treatment guidelines for

COPD will also be sent to each patient’s primary care physician whenever

one is identified.

13. A daily diary, on which subjects may keep track of medication changes

and illnesses between visits.

Clinic visits

Patients will be seen at months 1, 3, 6, 9, and 12 of treatment at which time

they will again queried about

Page 28: NEJM Protocol Sample 1

27

1. Whether any acute exacerbations occurred within the previous month, and

2. Possible macrolide-related side effects including:

(a) Neural: headache, hallucinations, paresthesias, syncope,

seizures, aggressive behavior, somnolence, tinnitus, hearing loss,

or vertigo.

(b) Hypersensitivity: rash, pruritis, tongue or facial swelling

(c) Gastrointestinal: diarrhea, vomiting, nausea, abnormal taste,

heartburn, or abdominal pain.

3. Patients will also have the following laboratory tests targeting macrolide-

related side effects and microbiologic changes:

(a) Liver function tests to include: alkaline phosphatase, albumin,

aspartate and alanine aminotransferase, INR and bilirubin

(b) Nasopharyngeal swabs and, when possible, expectorated sputum

for culture. Sensitivity testing will be done on all pathogens

cultured.

(c) Blood will be drawn for measurement of several biomarkers of

inflammation. (See biomarker protocol addendum.) This will

occur at visit months 3 and 12.

4. The patient's self-assessment of compliance with the medical regimen will

be recorded, as will a pill count and a recording of the output of the Med-

icTM ECMTM electronic compliance monitor.

5. At each clinic visit patients will be asked to show their study identification

wallet card.

Page 29: NEJM Protocol Sample 1

28

6. A repeat ECG will be obtained at one month seeking evidence of QT

prolongation.

7. Patients receiving phenytoin or digoxin will have drug levels checked

every week after beginning the study, with appropriate dose adjustments,

until therapeutic levels have been attained. Patients receiving warfarin will

have prothrombin times and INRs measured weekly, and those taking

theophylline will have theophylline levels measured weekly after starting

treatment with appropriate adjustments in dosages until the INR or the

theophylline levels are stable and therapeutic.

8. Pre-menopausal women will be queried about their contraceptive use and

pregnancy status. Those not using acceptable birth control will be

removed from study medication.

9. Patients complaining of worsening tinnitus or the development of, or

worsening hearing loss, will have repeat screening audiograms. Those

with ≤ 95th percentile age-adjusted hearing thresholds at 500, 1000, 2000,

or 4000 Hz in either ear will have the study drug discontinued and will

undergo formal audiology testing. All others will continue the medication

until they undergo formal audiology testing. If, on formal testing ≥ 95th

percentile age-adjusted hearing loss at 500, 1000, 2000, or 4000 Hz in

either ear is found, or if ≥ 10 db hearing loss at two frequencies in either

ear is confirmed in comparison with previous formal audiometry testing,

the patient will no longer take the study medication and will undergo

formal testing every months x 3, or until the patient returns to < 95th

Page 30: NEJM Protocol Sample 1

29

percentile loss or to their pre-study dB thresholds. All others will remain

on the study drug until their subsequent contact.

10. At months six and 12 months the SF-36, the St. George Respiratory

Questionnaire, the Quality of Well-Being Scale and post-bronchodilator

spirometry will be repeated as will the screening audiogram. Patients with

≥ 95th percentile age-adjusted hearing thresholds at 500, 1000, 2000, or

4000 Hz in either ear will have the study drug discontinued and will

undergo formal audiology testing. If ≥ 95th percentile age-adjusted hearing

thresholds in either ear is found, or if ≥ 10 db hearing loss at two

frequencies is confirmed in either ear in comparison with previous formal

audiometry testing, they will no longer take the study medication and

formal testing will be repeated every months x 3, or until the patient

returns to < 75th percentile loss. All others will remain on the study drug

until the subsequent contact.

11. At months three and 12 the Hospital Anxiety and Depression

Questionnaire (HADS) and the Pittsburgh Sleep Quality Index (PSQI) will

be repeated.

Monthly phone contact and intermittent studies.

Patients will be contacted by phone monthly, between clinic visits and queried

about

1. Whether any acute exacerbations occurred within the previous month, and

Page 31: NEJM Protocol Sample 1

30

2. Whether the patient has any possible macrolide-related side effects (see

above)

3. Patients with any of these side effects will be assessed for severity,

specifically with regard to whether the medication should be discontinued

or reduced in dose.

4. Pre-menopausal women will be queried about their contraceptive use and

pregnancy status. Those not using acceptable birth control will be

removed from study medication.

5. Patients complaining of worsening tinnitus or the development of, or

worsening hearing loss, will have repeat screening audiograms. Those

with ≥ 95th percentile age-adjusted hearing thresholds at 500, 1000, 2000,

or 4000 Hz in either ear will have the study drug discontinued and will

undergo formal audiology testing. All others will continue the medication

until they undergo formal audiology testing. If, on formal testing ≥ 95th

percentile age-adjusted hearing loss at 500, 1000, 2000, or 4000 Hz in

either ear is found, or if ≥ 10 db hearing loss at two frequencies in either

ear is confirmed in comparison with previous formal audiometry testing,

the patient will no longer take the study medication and will undergo

formal testing every months x 3, or until the patient returns to < 95th

percentile loss or to their pre-study dB thresholds. All others will remain

on the study drug until their subsequent contact.

Wash-out visit

Page 32: NEJM Protocol Sample 1

31

Patients will be seen one month after completing the one-year period of

treatment and have nasopharyngeal swabs and, when possible, expectorated

sputum sent for culture. Sensitivity testing will be done on all pathogens

identified. Spirometry will be repeated. Blood will be drawn for measurement of

several biomarkers of inflammation. (See biomarker protocol addendum.)

Patients will subsequently be contacted by phone at six-month intervals for

the duration of the study to inquire about hospitalizations in the preceding month

and whether they are or are not continuing to use azithromycin.

Nasopharyngeal swab processing.

Nasopharyngeal samples were obtained on enrollment and at months 1, 3, 6,

9 and 12 for culture. Local laboratories identified isolates of S. aureus, Moraxella

spp., Haemophilus spp. and S. pneumoniae in any quantity. Isolates were

shipped to a central laboratory for storage and batch processing. Organism

identification was verified by the central laboratory according to standard

guidelines (P. R. Murray, E. J. Baron, J. H. Jorgensen, M.L. Landry, and M. A.

Pfaller, (ed.), Manual of clinical microbiology, 9th ed. ASM Press, Washington,

DC., 2007).

Susceptibility testing was performed by broth microdilution using commercially

available panels (TREK Diagnostics Inc., Cleveland, OH). Clarithromycin

susceptibility was performed using Etest (AB bioMérieux, Solna, Sweden). All

susceptibility testing was performed according to the methods (Clinical and

Laboratory Standards Institute. Methods for dilution antimicrobial susceptibility

Page 33: NEJM Protocol Sample 1

32

tests for bacteria that grow aerobically; approved standard. 8th ed. Document

M07-A8. Wayne, PA: CLSI; 2009) and interpretative criteria recommended by the

Clinical Laboratory Standards Institute (Clinical and Laboratory Standards

Institute. 2010. Methods for antimicrobial dilution and disk susceptibility testing of

infrequently isolated or fastidious bacteria. Approved guideline M45-A2. CLSI,

Wayne, PA.) with the exception of Moraxella spp, for which EUCAST breakpoints

were used (European Committee on Antimicrobial Susceptibility Testing

(EUCAST, Home page at: http://www.eucast.org).

When patients had more than one isolate for which macrolide susceptibility

was performed at a given visit, only the most resistant isolate was included.

Differences between groups were determined using chi square analysis.

Method for Analyzing Hearing Changes

The mean difference in hearing thresholds in patients receiving

azithromycin or placebo was determined by comparing the mean thresholds

measured at 1000, 2000, 3000 and 4000 Hz in both ears. Differences in

thresholds were compared between measurements made on enrollment and at

three months, and between enrollment and 12 months.

Method for determining compliance taking study drug as prescribed.

The stop date was taken as the date of death if the patient died during the

study or the date of the last clinic visit attended prior to the wash-out visit (with a

30 day window for the last scheduled visit).

Page 34: NEJM Protocol Sample 1

33

Compliance was determined by dividing the actual number of pills taken by

number of days pills should have taken, and was recorded as being 100% if

patients took more than the prescribed number of pills during a given time period.

Compliances over all time periods were averaged to determine an overall percent

compliance/patient. Calculations excluded 25 patients who attended no visits.

Since adherence was assessed post hoc the correlations sought between

compliance and AECOPDs should not be considered as intention to treat

analyses.

Reasons for study drug discontinuation There are five instances in which the study drug might be discontinued: 1. At the start, and for one week following, an acute exacerbation

Patients are instructed as to what constitutes an acute exacerbation as

described above. In brief, qualifying exacerbations require that the patient be

treated with antibiotics and/or systemic corticosteroids. Cards summarizing this

information are also given to each patient and sent to their primary care

physicians. Patients are instructed to stop the study medication if and when they

begin antibiotics, and restart the study medication one week following completion

of the course.

2. Development of symptoms that might represent medication-related side

effects.

Prior to beginning the study patients are instructed about what might

constitute a macrolide-related side effect and whom they should notify should

such symptoms develop. They are also queried at each clinic visit and phone

Page 35: NEJM Protocol Sample 1

34

contact about a specific list of possible macrolide-related side effects (see

above). If potential side effects develop, study personnel will assess the severity

of the problem and decide whether the study medication should be discontinued

or reduced in dose.

Diarrhea is expected to be the most common complaint. Should this be

severe enough to warrant intervention we will reduce the study medication to one

pill three times a week. If the diarrhea persists on this lower dose the study drug

will be discontinued.

Although the literature predicts that hearing loss should be rare, if patients

complain of a change in their hearing we will repeat their audiograms. Those

with documented decrements will have the study drug discontinued.

Hypersensitivity and toxic hepatic reactions rarely or never occur as a result

of metabolism of azithromycin. Nonetheless, we will measure the standard

battery tests used to evaluate the liver at each clinic visit. The study drug will be

discontinued in all who have values > three times normal.

3. Intercurrent illness.

We anticipate that during the one year treatment period some patients will

develop medical and/or surgical problems that are unrelated to COPD or to a

possible azithromycin-related side effect but warrant treatment. In these

instances the patient's treating physicians will decide whether the specific

problem encountered warrants discontinuation of the study medication. Each

patient will carry a wallet card for the duration of the study that provides

Page 36: NEJM Protocol Sample 1

35

information regarding the study and how unmasking of treatment can be

accomplished should the indication merit (see below).

4. If the QTc on ECGs recorded one month after starting the study medication

increases to > 500 ms or > 60 ms from the patient’s baseline value.

5. For patients with a bundle branch block, if the JTc interval increases 60 ms

from the patient’s baseline value.

Recruitment and Retention

Enrollment

Each of the ten Clinical Centers in the COPD Clinical Research Network

plans to enroll patients from up to four separate medical centers and/or from a

number of associated outpatient practices (Table 6). We have budgeted 3

hours/week, 48 hours/year for two years for Clinical Coordinator time to be spent

on recruiting. This time will be used to visit primary and appropriate specialty

care clinics, describe the study to the treating physicians and be available for

screening. Each Clinical Center includes outpatient clinics that treat large

numbers of patients with COPD and each has experience with advertising to the

public and/or to selected groups (e.g., COPD patient support groups, local lung

associations). Many of the Centers have dedicated personnel and standardized

approaches to advertising that will facilitate the process.

The Clinical Coordinator will screen records of patients giving HIPPA-

compliant consent, and will seek consent for study participation from all those

meeting entry criteria.

Page 37: NEJM Protocol Sample 1

36

Retention

A variety of factors contribute to patients failing to complete trials or comply

with prescribed regimens. We will address issues in each area in an effort to

achieve maximal retention and compliance as Rotor and colleagues (1998) found

that programs employing educational, behavioral and motivational components

are more effective than any single approach. We have developed a

comprehensive plan to address each of five domains identified as contributing to

patient's adhering to prescribed medical regimens.

Socioeconomic factors. In addition to the cost of buying medications patients

frequently need sufficient financial resources to cover the cost of traveling to the

health care provider, arranging childcare, taking time off from work. Unstable

living conditions, low levels of education and poor social support systems

contribute to these problems. Patients agreeing to participate in this study will

bear no additional expense for medications or testing, and will be reimbursed for

their travel-related time and effort. Information regarding the study will be printed

in easily understood language aimed at patients with no more than a grade

school education.

Healthcare team and healthcare system-related factors. Organizational

factors such as the time spent with the health care provider, the continuity of

care, and the communication style of the provider may improve retention and

adherence to a prescribed regimen (Albaz, 1997; Abbott, 1999). We will

structure the initial evaluation and all follow-up phone contacts and clinic visits

Page 38: NEJM Protocol Sample 1

37

such that the patients interact with the same study coordinator on each occasion.

Lack of awareness and knowledge about adherence problems on the part of

health care providers may contribute to the problem. Accordingly, all the

investigators and coordinators will receive and review the recent WHO report that

summarizes the pertinent issues pertaining to adherence (Sebate, 2003). The

importance of continuous attention to retention and adherence, and the fact that

facilitating compliance is a dynamic process will be emphasized.

Condition-related factors. Condition-related factors include those related to the

severity of symptoms, the level of physical, psychological, social and vocational

disabilities, the rate of progression, the availability of effective treatments, and

the co-existence of depression. To reduce the possibility that these factors will

limit compliance with the protocol we will attempt to assure that patients are

receiving the optimal treatment for their disease. The co-investigators in this trial

concur with the COPD treatment guidelines that are summarized in the GOLD

initiative and will utilize these when caring for all of the patients for whom they

have responsibility. Unfortunately, this standardized treatment approach cannot

be used for all patients as some will be treated by physicians who are not part of

the COPD Clinical Research Network. To increase the likelihood that as many

patients as possible will receive what is currently perceived to be optimal therapy

for their disease all patients will receive an informational pamphlet that reviews

appropriate treatment of COPD and recommendations concerning this treatment

will also be communicated to their primary care physicians. All patients will be

Page 39: NEJM Protocol Sample 1

38

screened for depression. When found treatment will either be administered or

recommended.

Therapy-related factors. The most important therapy-related factors affecting

compliance are the complexity of the medical regimen, the duration of treatment,

frequent changes in treatment, the immediacy of beneficial effects, the frequency

of side effects, and the availability of medical support to deal with them. In an

attempt to facilitate compliance with the prescribed regimen we have chosen a

daily dosing regimen (250 mg once daily) despite the fact that it will provide a

slightly higher total weekly dose, and run the risk of causing a slightly higher rate

of side effects than the 500 mg three times weekly dose that was alternately

considered. Patients will be educated as to the side effects that might be

expected, and what changes in dosing might occur should they develop.

Patients will have ready access to study personnel who will assist them with the

appropriate response to any side effect that develops.

Patient-related factors. Patient-related factors include their knowledge and

perceptions about their illness and their motivations and expectations regarding

self-management of their symptoms. To address these factors all patients will

receive education about their disease as described above and information

pertaining to COPD Support Groups in their respective areas, and they will be

queried on a monthly basis about their disease management, their understanding

of their treatment instructions and the success or lack thereof they are seeing in

response to the intervention. We will also attempt to increase the patient’s

motivation by discussing the perceived importance of adherence, and by

Page 40: NEJM Protocol Sample 1

39

teaching self-management skills (e.g., correct bronchodilator inhalation

techniques, the importance of exercise, purse-lipped breathing).

Choice of macrolide

Azithromycin was chosen over the other macrolides for the following reasons:

1. The incidence of medication-related side effects and drug-drug

interactions is considerably lower with azithromycin than with

clarithromycin or erythromycin.

Clarithromycin and erythromycin are nonspecific P450 enzyme inhibitors,

specifically CYP3A4, the enzyme involved with the metabolism of the

largest number of medications (i.e., calcium channel blockers, statins,

dihydropyridines, disopyramide, benzodiazepines, quinidine, protease

inhibitors, non-sedating antihistamines, and many psychiatric

medications).

Azithromycin does not interact with the cytochrome P-450 system. There

is a single case report of azithromycin interacting with disopyramide, and a

number of studies have shown that azithromycin does not interact with

statins, non-sedating antihistamines, psychiatric medications, and

benzodiazepines.

Contraindications.

Page 41: NEJM Protocol Sample 1

40

The only contraindication listed in the package insert for azithromycin is

giving it to patients with known hypersensitivity to macrolide antibiotics.

Warnings.

Warnings included in the package insert are for serious allergic reactions

and antibiotic-associated pseudomembranous colitis from C. difficile toxin.

Precautions.

The package insert cautions against giving azithromycin to patients with

impaired hepatic function and those with glomerular filtration rates < 10

mL/min. We are excluding patients with INRs > 2.0 (unless they use

warfarin), a serum albumin < 3.0 g/dL, a serum AST or ALT > 3 times

normal, or patients with a serum creatinine > 1.5 mg/dL or an estimated

creatinine clearance < 20 mL/min.

Prolonged QTc intervals, with the attendant risk of cardiac arrhythmias

(including torsades de pointes), have been seen in patients treated with

macrolides other than azithromycin. This repolarization delay is results

from blockade of specific potassium channels, and is most commonly

seen with erythromycin (generally with intravenous administration),

clarithromycin, and patients with preexisting heart disease (Iannini, 2002).

QTc prolongation is also seen when macrolides are given with other

medications that are metabolized by cytochrome CYP3A4 (e.g., the non-

Page 42: NEJM Protocol Sample 1

41

sedating antihistamines and cisapride). Harris and colleagues (1995)

found that azithromycin (500 mg for 1 day and 250 mg on 4 subsequent

days) did not prolong the QTc interval and had no additive effect on

terfenadine-induced QTc prolongation. Strle and Maraspin (2002)

prospectively studied 47 patients given azithromycin (3 g divided over 5

days). All were previously healthy. The median QTc increased from 406

to 419 ms (P = NS) and the same proportion of patients had QTc > 440

ms prior to, and after treatment. In vitro studies indicate that, despite have

similar effects on QT prolongation, erythromycin and clarithromycin have

greater pro-arrhythmic potentials compared with azithromycin (Milberg,

2002).

Nonetheless, the package insert states cautions that an effect of

azithromycin on the QTc interval "cannot be completely ruled out in

patients at increased risk for prolonged cardiac repolarization". We are

excluding patients with a QTc interval > 440 msec on either their

screening ECG or on an ECG repeated one month after beginning the

study.

The risk of development of drug-resistant bacteria is also mentioned as a

precaution (see below).

Drug Interactions.

The package insert lists the following drug interactions:

Page 43: NEJM Protocol Sample 1

42

(a) Co-administration of nelfinavir increases azithromycin concentrations.

Accordingly, use of nelfinavir is one of our exclusion criteria.

(b) Azithromycin does not affect the prothrombin time or INR in response

to a single dose of warfarin but concurrent use of macrolides and warfarin

has been associated with increased anticoagulant effects. We will follow

the package insert recommendation and monitor the INR in all patients

taking warfarin as noted above.

(c) Therapeutic doses of azithromycin cause a "modest" effect on the

pharmacokinetics of atorvastatin, carbamazepine, cetirizine, didanosine,

efavirenz, fluconazole, idinavir, midazolam, rifabutin, sildenafil,

theophylline, triazolam, trimethoiprim/sulfamtehoxazole and zidovudine.

Co-administration with efavirenz or fluconazole has a "modest" effect on

the pharmacokinetics of azithromycin. "[N]o dosage adjustment of either

drug is recommended".

(d) Although interactions with digoxin, ergotamines, terfenadine,

cyclosporine, hexobarbital and phenytoin have not been reported in

clinical trials, no specific drug-interaction studies have been performed to

evaluate potential interactions with these medications, and other

macrolides have been found to increase digoxin levels or result in severe

peripheral vasospasm and dysesthesias in patients taking ergotamines,.

Monitoring terfenadine, cyclosporine, hexobarbital and phenytoin

concentrations is advised. Accordingly, we will monitor drug levels on all

patients enrolling in the study who are taking phenytoin or digoxin.

Page 44: NEJM Protocol Sample 1

43

Patients taking ergot alkaloids, terfenadine, cyclosporine, or hexobarbital

are excluded.

Medication-related side effects

Gastrointestinal. With long-term administration of azithromycin diarrhea

occurs in from 0% [with 250 mg given either daily (Wolter, 2002) or three

times per week for 3 months (Gerhardt, 2003)] to 25% of subjects [in

those receiving 500 mg twice-weekly for six months (Cymbala, accepted

pending review, personal communication)]. In the largest study, diarrhea

occurred in 20% of treated patients compared with 8% of controls

(Saiman, 2003). The need to discontinue therapy because of diarrhea is

rare, however, (maximum reported occurrence 8% (Cymbala, accepted

pending revision, personal communication). Nausea occurs in 29% of

patients (compared with 16% of controls), but does not seem to result in

study drop-out (Saiman, 2003).

If patients develop diarrhea that is severe enough to warrant intervention

we will first attempt to reduce the frequency of the study medication to

three times a week. If the diarrhea persists on the lower dose the study

drug will be discontinued.

Ototoxicity. The incidence of ototoxicity is 21% in patients taking 4 g/day

of erythromycin (Swanson, 1992). Symptoms can begin within the first

Page 45: NEJM Protocol Sample 1

44

week of treatment but are usually reversible within 30 days of

discontinuing the drug (Brummett, 1993). Irreversible changes have been

reported with intravenous administration, however. The mechanism of this

adverse reaction is unknown but is thought to be dose-dependent (Taylor,

1981; Swanson, 1992), and predisposing factors include renal

abnormalities, advance age, high doses, and concurrent use of other

ototoxic medications (Haydon, 1984; Umstead, 1986; Vasquez, 1993).

Ototoxicity has only rarely been reported with azithromycin but the

incidence is unknown (Wallace, 1994). Saiman and colleagues (2003)

reported that two of 185 patients reported hearing impairment and that two

others noted tinnitus. Interestingly, these complaints were equally divided

between patients receiving azithromycin and placebo.

We will inquire about hearing and balance problems in all patients prior to

instituting therapy, obtain a screening audiogram, and obtain additional

audiograms on all patients reporting tinnitis or the development of hearing

impairment during the course of the study.

Hepatotoxicity and hypersensitivity reactions. Hypersensitivity and toxic

hepatic reactions occur in response to nitrosoalkanes formed from the

metabolism of erythromycin. Nitrosoalkanes rarely or never form as a

result of metabolism of azithromycin. We will measure the standard

Page 46: NEJM Protocol Sample 1

45

battery tests used to evaluate the liver at each clinic visit and stop

treatment if they are elevated to the extent described above.

2. Although clarithromycin produces higher mean and maximum plasma

concentrations, a greater area under the 24-hour plasma concentration-

time curve, and a higher concentration in alveolar lining fluid and alveolar

macrophages than azithromycin when given in comparable doses (Patel,

1996; Rodveld, 1997; Zhanel, 2001), the studies documenting these

differences were all done in normal subjects. All of the macrolides

concentrate in areas of inflammation due to their basic charge.

Azithromycin is dibasic so this occurs to an even greater extent.

Accordingly, the data reported in normals may not be relevant to patients

with pulmonary inflammation.

3. Although many studies document the clinical benefit of erythromycin or

clarithromycin when given to patients with a variety of pulmonary

conditions (Table 2), most of these studies utilized observational designs.

Two randomized, controlled trials have documented the clinical benefit of

azithromycin in patients with cystic fibrosis (Wolter, 2002; Saiman, 2003),

and a third showed utility in patients with panbronchiolitis [although a third

small, randomized, crossover trial only found that treatment had no

substantive effect on airway inflammation and only was associated with

fewer uses of additional antibiotics (Equi, 2002)].

Page 47: NEJM Protocol Sample 1

46

Choice of dose.

The dose of any medication should be the lowest possible that produces

clinical effectiveness, taking into account issues that pertain to maximizing

compliance with the prescribed regimen. There are no published studies

comparing the efficacy or side-effects of different doses of long-term

azithromycin. Accordingly, the choice of dose for this study must be empiric. A

review of the pertinent literature is as follows:

Efficacy.

The two double-blind trials showing efficacy in patients with cystic fibrosis

used azithromycin, either 250 mg once daily for three months, or 250 or 500 mg

(depending on whether body weight was above or below 40 kg) three days a

week for 168 days (Wolter, 2002; Saiman, 2003). One unpublished study in 12

patients with bronchiectasis found reduced sputum volume and fewer

exacerbations with azithromycin, 500 mg twice weekly (Cymbala, in press

pending revision).

Side-effects and compliance.

No subjective sides effects were reported in 61 cystic fibrosis patients who

received 250 or 500 mg of azithromycin daily for 6 months (Equi, 2002). In a

second study of cystic fibrosis patients, 6 (25%) of those receiving azithromycin

(250 mg daily) did not complete three months of treatment, as did 9 (30%)

receiving placebo. Failure to complete therapy was thought to be “likely” or

Page 48: NEJM Protocol Sample 1

47

“possibly” related to medication-related side effects only 3 of these 15.

Compliance, assessed by retrospective questioning, indicated that 29% of the

patients receiving azithromycin reported missing doses compared with 40% of

those receiving placebo (Wolter, 2002). Side effects in the largest study in cystic

fibrosis patients (N = 185) who were treated with 250 or 500 mg of azithromycin

three times a week or a placebo were considered to be “mild or moderate” and

included only nausea in 33% vs 16% (azithromycin vs placebo, respectively),

diarrhea (23 vs 8%) and wheezing (17 vs 4%) (Saiman, 2003). The dose of the

study drug was reduced, or stopped and restarted in 4 (5%) of the treated

patients and in 4 (4%) of the controls, and was discontinued in 3 (3%) treated

patients and 2 (2%) controls. Pill counts indicated that the patients took 93% of

the azithromycin doses and 89% of the placebo. Cymbala and colleagues

(unpublished data provided courtesy of Guy Amsden, PharmD) treated 12

patients with bronchiectasis (mean age 71 years, mean weight 76 kg) with

azithromycin, 500 mg twice a week, for six months. Three (25%) complained of

diarrhea. In one, the dose was decreased to 250 mg three times a week with

improvement. Compliance assessed by pill counts ranged from 85 to 108%.

We have opted to use a dose of 250 mg daily because it is high enough to

exclude the possibility that a negative result is not due to an insufficient dose,

and daily (as opposed to three times a week) administration is likely to facilitate

compliance (see below).

Page 49: NEJM Protocol Sample 1

48

Other or Concomitant Therapy

All the investigators in the COPD Clinical Research Network agree with the

treatment guidelines summarized in the GOLD initiative (Pauwels, 2001) and,

accordingly, will be treating those patients for whom they are the primary care

giver in a similar fashion. Patients whose primary care giver is not one of

investigators in the Network may be treated differently. To minimize this, each

patient will receive a tutorial and printed material summarizing the recommended

treatment of his or her disease at the beginning of the study. Letters will also be

sent to their primary care physician summarizing these same recommendations.

Non-COPD-related

Given the expected age of the patients we will be enrolling many are likely to

have one or more medical conditions in addition to COPD. Therapy for these

problems will be continued at the direction of each patient's treating physician

with the exceptions of the medications listed in the exclusion criteria.

Unmasking

Emergency situations may arise in which it is necessary to discontinue the

study drug and unmask treating physicians, Emergency Department personnel,

clinic personnel, or the patient to the assigned treatment. Participants will be

given a wallet-size card explaining that they are in a research study and that they

are assigned to take either azithromycin capsules (250 mg/day) or a matched

placebo. The card will include telephone numbers of their site's pharmacy, the

Page 50: NEJM Protocol Sample 1

49

responsible investigator and the DCC. Patients will be instructed to carry the

card at all times; compliance with this instruction will be checked at clinic visits

(see above).

Situations that require unmasking are expected to be rare. In most

emergency situations it will be sufficient to have the participant discontinue taking

either the drug or the placebo until the event is resolved. It is possible, however,

that emergency personnel or treating physicians might feel it is necessary to

know the medication the participant was taking to decide upon a rational course

of treatment, or to ensure that other medications are not given that might

adversely interact with azithromycin. In such cases the treating medical

personnel should call the clinic pharmacy (first choice) or the DCC to obtain the

treatment assignment.

Adverse events or hypersensitivity reactions may occur which are believed to

be definitely or probably associated with the use of azithromycin, and which

result in discontinuation of the use of the assigned medication by the participant.

In such cases, the participant should be informed of the study drug prescription.

The reason for this is to provide information to the participant regarding their use

of azithromycin in the future. If the study drug is azithromycin, then the

participant should be told that there have been indications that he/she is

hypersensitive or reactive to this drug and use of it in the future should be

avoided. The participant should also be informed if the study drug is placebo, so

that his/her use of azithromycin in the future will not be restricted.

Page 51: NEJM Protocol Sample 1

50

Such events must be thoroughly and carefully documented, and an Unmasking

Report for the event must be completed and transmitted promptly to the DCC.

Such events must be reported to the DSMB on at least a quarterly basis.

12. Data Management

Training

A training session for the study personnel who are conducting the protocol

(clinic coordinators PIs, pharmacists, lung function techs, others) will be held in a

central location (Denver or Minneapolis) prior to startup. Material will be

presented covering:

1. Background and rationale for the study (Dr. Albert)

2. Design of the study (Dr. Connett)

3. Recruiting goals and strategies (Dr. Connett and others)

4. Determining eligibility (Dr. Albert and others)

5. Informed consent

6. Randomization procedures and drug dispensation; role of the pharmacy

(DCC)

7. Follow-up visit schedule

8. Forms and forms completion

9. Data entry, error correction (DCC)

10. Exacerbations: determination and documentation (Dr. Albert)

11. Adverse events: notification and documentation (DCC)

Page 52: NEJM Protocol Sample 1

51

12. Dose adjustment or termination; breaking the blind (Dr. Albert)

13. Monitoring of recruitment, adherence to protocol, compliance, visit

completion, other; role of the DSMC (DCC)

14. Post-treatment follow-up (Dr. Niewoehner)

15. Technical procedures (Arranged by DCC):

a. Lung function testing

b. Blood draws, lab work, and shipping

c. Audiometry

d. Other

16. Ancillary study: biomarkers (Drs. Albert and Woodruff)

a. Background, rationale, and goals

b. Procedures to obtain specimens, specimen handling; schedule

17. Closeout and notification of participants (DCC, Dr. Albert)

The session will be conducted as an interactive process with questions and

answers and practice in procedures, interviews, forms completion, data entry and

transmission, and other activities and testing built in.

Training of new or replacement personnel will be conducted by sending them

to the DCC. Personnel who have unsatisfactory records with respect to

performance or error rates will be required to undergo central re-training.

Page 53: NEJM Protocol Sample 1

52

Site Visits

Each clinical center will be visited at least twice during follow-up by a team

assembled by the DCC. Site visitors will review the following aspects of

operation:

1. Organizational structure of the clinical center

2. Recruiting methods and strategies

3. Adherence to protocol

4. Visit scheduling, visit completion rates, adherence rates

5. Error rates and timeliness of corrections in data entry

6. Handling of adverse events; unblinding events

7. Compliance

8. Agreement of entered data with raw source documents (review of records)

9. Satellite clinics; communications

10. Exit interview with clinic PI, clinic coordinators, and other staff

A brief written report on each site visit will be prepared by the DCC after each

site visit and sent to the clinical center PI, the clinic coordinator, and the NHLBI

Project Office.

Data flow

Data from interviews, lab reports, and most other sources of information will

be collected on printed paper forms. Spirometry will be obtained using study-

standard systems.

Page 54: NEJM Protocol Sample 1

53

After forms have been completed, personnel in each Clinical Center will enter

the data by accessing the website (http://www.copdcrn.umn.edu). A CCRN User

ID and password will be required to link to this site. After securing entry, a menu

of clinical sites will appear. The user must choose one of these. After choosing

a site and validating the identity of the clinic, a second menu will appear that lists

the available activities (see above). For data entry, clinic personnel will need to

click on 'Data Entry'. A list of currently open studies will appear. The data

enterer will select ‘Macrolide Study’ from a hyperlinked list of all forms. The data

enterer will then select the specific form to be used.

Data will be edited at the time of entry. The data entry program inserts data

directly into the database at the DCC. An additional and more comprehensive

edit will take place each day at the DCC. This may result in requests for error

correction that will be e-mailed to the clinical center.

Data management at the DCC makes use of Oracle databases running on a

Unix network. All files are automatically backed up, with backup files stored off-

site, on a daily basis. To ensure confidentiality and security, all data transmitted

to or from the DCC, whether through the CCRN website or via e-mail, is fully

encrypted.

The DCC will make available current reports on recruitment, randomization,

visit completion rates, forms completion and accuracy, adherence to protocol,

compliance, serious adverse events, and other information, both by clinical

center and for all centers combined. These will be regularly posted on the CCRN

Page 55: NEJM Protocol Sample 1

54

website. They will be accessible only to personnel with registered User IDs and

passwords.

Quality Assurance and Quality Control

Assurance of data quality will be facilitated by:

1. Online availability of a study-specific Manual of Procedures

2. Clear, well-designed forms and form-modules with most terminology

defined on the forms themselves

3. Training of clinic personnel in the following aspects of the study:

a. Design and rationale

b. Target population and recruitment

c. Eligibility

d. Informed consent procedures

e. Randomization

f. Allocation of assigned medication

g. Follow-up visit scheduling and procedures

h. Reporting COPD exacerbations

i. Reporting adverse events

j. Data entry, transmission, and error correction

k. Spirometry and associated data transmission

l. Audiometry

m. Unmasking treatment assignment

n. Reporting protocol violations

Page 56: NEJM Protocol Sample 1

55

o. Closeout of the study for the participant

4. Interactive editing at the time of data entry

5. Comprehensive editing of data at the DCC

6. Site visits by DCC and other study personnel

7. Regular, frequent preparation of monitoring reports, and review of these

reports with clinic coordinators

8. Monitoring of quality of lung function testing; over reading of spirometry

data

9. Frequent communication by phone and e-mail between the DCC and clinic

personnel

10. Periodic review of data on safety, toxicity, efficacy, and study operation by

the Data and Safety Monitoring Committee

A special consideration regarding data quality in this study is the review and

definitive classification of COPD exacerbations by (1) an expert coder in the

DCC, and (2) a 3-member panel of investigators who will be masked to

participant ID and treatment assignment. The panel will review admission and

discharge summaries for all hospitalizations, emergency department visits, and

physician summaries for all outpatient events, assess whether each event was in

fact a COPD exacerbation and, if so, whether it was mild, moderate, severe or

very severe. The same panel will review any patient deaths that occur during the

study attempt to determine the primary and secondary causes.

Reports to Data and Safety Monitoring Board

Page 57: NEJM Protocol Sample 1

56

Interim reports to the Data and Safety Monitoring Board (DSMB) will be

carried out at monthly and 6-monthly intervals. Monthly reports will include a

graphical comparison of actual and target enrollments. Six-monthly reports will

include follow-up rates, adverse events, mortality, levels of compliance, losses to

follow-up, and data on the primary and secondary outcomes.

We are proposing a group-sequential monitoring plan to provide guidelines for

termination of the trial in the event that the results are strongly positive, as

discussed in the Sequential Analysis and Monitoring Boundaries section on page

9.

The DCC will establish a secure website and a process for e-mail notification

that will expedite reviews of serious adverse events by the DSMB while

maintaining subject privacy. Reports of adverse events and study results will be

presented in a manner that maintains blinding of the DSBM with regard to

treatment assignment but the DCC will be prepared to identify the treatment

groups (independently for adverse events and study results) if requested by the

DSMB.

Limitations

Exclusion of patients with asthma.

The degree of reversibility after bronchodilator inhalation has previously been

used to distinguish patients with asthma from those with COPD. There are,

however, a number of problems with this approach. First, there is no consensus

regarding what variable to use to define reversibility. Second, regardless of how

Page 58: NEJM Protocol Sample 1

57

one defines reversibility, there is no consensus regarding the degree of

reversibility needed to conclude that patients have asthma. A recent study of

660 COPD patients found that the maximum change in FEV1 in response to

salbutamol (expressed as a percent of the normal predicted FEV1) was 12%.

When both salbutamol and ipratropium were given together, however, the

percent increase was as much as 26% (Calverley, 2003). The maximum

absolute change in FEV1 was 0.7 L, and the change in FEV1 (expressed as a

percent of the pre-bronchodilator FEV1) was as much as 70% (Calverley, 2003).

In patients with very low FEV1s some suggest using the absolute increase in the

FEV1 to define reversibility, but the degree of change distinguishing asthma has

been variable (e.g., 200 or 400 mL) and arbitrary. Others have found that the

absolute change in FEV1 is unrelated to the FEV1 whereas the percent change

was correlated (Tweeddale, 1987; Calverley, 2003). Third, there is considerable

day-to-day variability in the change in FEV1 in response bronchodilators, both

between and within individuals (Anthonisen, 1986; Tweeddale, 1987; Nisar,

1992). Fourth, bronchodilator testing can yield different results if the test is

performed with different inspiratory maneuvers (Santus, 2003).

Bronchodilator testing was included in previous British Thoracic Society

guidelines and in the GOLD initiative, but is not recommended in the latest

ATS/ERS guidelines that goes on to suggest that asthma can be excluded by a

combination of the clinical history, signs and baseline spirometry, and that

reversibility testing does not add any additional information. This position is

supported by a recent study of patients diagnosed as having either COPD or

Page 59: NEJM Protocol Sample 1

58

asthma on the basis of the clinical history who subsequently underwent bronchial

biopsy and had induced sputum evaluated for differential cell counts. The

authors found that the clinical diagnosis was correct when the pattern of

inflammation observed was used as the gold standard (Fabbri, 2003).

13. Microbial Resistance.

Long-term antibiotic treatment raises the concern of colonization or infection

with macrolide-resistant organisms.

Prevalence of S. pneumoniae resistance. Resistance of S. pneumoniae to

macrolide antibiotics is increasing. From 1995-1998, resistance to erythromycin

increased from 11% to 15% (Whitney, 2000). From 1995 to 1999 data from the

Centers for Disease Control found resistance rates increased from 10.6% to

20.4% (Hyde, 2001). Data from the SENTRY Antimicrobial Surveillance Program

collected from 1997-1999 found resistance rates to clarithromycin of 16% (and

2.1% and 0.4% for H. influenza and M. catarrhalis, respectively) (Hoban, 2001).

Doern and colleagues (2001) found a 26% rate of resistance (and 78%

resistance in penicillin-resistant S. pneumoniae). Guchev and colleagues (2004)

reported that resistance rates of S. pneumoniae to azithromycin increased from 0

to as much as 40% when patients were cultured 14 weeks after 8 weeks of

therapy had been completed.

Mechanisms of resistance. The most common mechanisms of resistance to

macrolides are ribosomal methylation and upregulation of efflux pumps.

Page 60: NEJM Protocol Sample 1

59

A. Methylation of rRNA.

Ribosomal methylation is the most common mechanism of resistance to

macrolides, occurring in S. aureus, S. pneumoniae, and H. influenzae. It is

mediated by the erm gene that codes for a methyltransferase that methylates

adenine 2058 in 23S rRNA, probably resulting in steric hindrance of the

macrolide attaching to its binding site (Weisblum, 1995; Poehlsgaard, 2003).

Erm(A) is common in staphylocci and S. pyogenes but rare in S. pneumoniae.

Erm(B) is most common in S. pneumoniae (depending on geography, see

below), other streptococci and enterococci.

The mechanism by which macrolides induce erm RNA is not yet known but

may be dependent on concentration (i.e., at low concentration too few ribosomes

might be occupied to allow for sufficient erm synthesis).

B. Efflux pumps

Constitutively expressed efflux pumps having a broad spectra of substrates

are thought to account for the poor susceptibility of many Gram-negative

organisms to macrolides (Leclercq, 2002; Van Bambeke, 2003). Macrolide-

inducible pumps in Gram-positive bacteria have narrow spectra resulting in

isolates of S. pneumoniae that are susceptible to clindamycin but resistant to

macrolides as a result of the effects of mefA (Jones, 1996; Nishijima, 1999).

Waites and colleagues found that 97.6% of mefA+ S. pneumoniae isolates had

MICs ≤ 32 µg/ml for clarithromycin, while 97.1% of ermB+ isolates had MICs >

Page 61: NEJM Protocol Sample 1

60

256 µg/ml. Waites and colleagues (2000) have suggested that the

predominance of low-level resistance conferred by mefA may account for the

observation that there are only rare reports of treatment failure with macrolides

despite increasing in-vitro macrolide resistance, and widespread use of these

antibiotics to treat respiratory ailments. A similar conclusion was reached by

Nuermberger and Bishai (2004).

The etiology of macrolide resistance varies markedly by geographic region

(Lynch, 2002): two-thirds of resistant isolates in North America are the result of

mefA whereas ermB accounts for the majority of the resistance found in Europe

and South Africa, and the marked increase in resistance seen since 1993 is the

result of mefA (Gay, 2000; Hyde, 2001). This geographic disparity is such that in

many European countries macrolides are no longer recommended as first-line

therapy for conditions in which S. pneumoniae or S. aureus are likely organisms

(Mulazimoglu, in press).

Animal studies. Clarithromycin improves survival in neutropenic mice infected

with 8 of 9 macrolide-resistant mefA-producing strains of S. pneumoniae with

MICs to clarithromycin ranging from 0.5 to 16 ug/ml, but azithromycin improved

survival in only 2 of the 9 strains (with azithromycin MICs of 1-32 ug/mL). Neither

medication improved survival in ermB-producing isolates that had MICs > 63

ug/mL (Hoffman, 2003). Tessier and colleagues (2002) found that colony counts

decreased in the same model when animals with macrolide-resistant S.

pneumoniae were treated with clarithromycin when MICs were as high as 4

µg/mL but not when the strain had MICs of 8 µg/mL. Unfortunately, these data

Page 62: NEJM Protocol Sample 1

61

are difficult to extrapolate to what might be the case in human subjects. First,

use of a neutropenic model is a problem in that macrolides (particularly

azithromycin) are concentrated within, and carried to sites of infection by,

neutrophils. Second, administering S. pneumoniae via the trachea generally

results in bacteremia such that outcomes may more likely to depend on serum

rather than intraparenchymal drug levels and the latter is thought to be more

important in human S. pneumoniae infections as most are not bacteremic.

Clinical data. A number of case reports and small series report break-through

bacteremia in patients with macrolide-resistant S. pneumoniae who were treated

with macrolides [Fogarty, 2000 (N=3); Kelley 2000 (N = 4); Waterer, 2000 (N=1);

Musher, 2002 (N=1); Lonks, 2002 (N=19); van Kerkhoven, 2003 (N=4)]. Lisby

and colleagues (2001) note, however, that break-through bacteremia also occurs

with macrolide-sensitive strains. In addition, most (but not all) of the instances of

break-through bacteremia occurred in patients taking oral therapy for species

with MICs > 8 µg/mL. The study by Lonks and colleagues (2002) in which 19

cases were reported, was conducted over a 13-year period in four hospitals.

Only one of these isolates had MICs < 16 µg/mL. Nuermberger and Bishai

(2004) suggest that (1) while high-level ermB-mediated resistance is associated

with occasional treatment failure, the risk of treatment failure in this group of

patients is unknown, (2) the risk of treatment failure in patients whose organisms

have MICs < 16 mediated by efflux pumps has not been substantiated, and (3)

because treatment failures can also occur in patients with drug-susceptible

Page 63: NEJM Protocol Sample 1

62

organisms, observing failure in patients with resistant organisms does not

establish cause and effect.

Several studies have found a survival benefit when macrolides are used to

treat either community acquired pneumonia, or bacteremic S. pneumoniae

pneumonia (Gleason, 1999; Stahl, 1999; Waterer, 2001; Rello, 2002; Martinez,

2003), and the morbidity and mortality associated with treating community

acquired pneumonia with macrolides seems to be the same regardless of

whether patients have macrolide-sensitive or macrolide-resistant infections

(Ewig, 1999). In summary, many patients are treated with macrolide antibiotics

for pneumonia and reports of treatment failure are extremely rare, despite

evidence of increasing in vitro macrolide resistance.

Clinical consequences of long-term macrolide treatment. Colonization with

macrolide-resistant bacteria occurred in a group of HIV-positive patients who

were given 12 weeks of either clarithromycin (500 mg twice daily) or azithromycin

(1200 mg weekly) as prophylactic treatment for M. avium (Aberg, 2001), but was

not observed in a study of cystic fibrosis patients who received azithromycin, 250

mg daily, for three months, (Wolter, 2002). In 10 patients with panbronchiolitis

treated with long-term clarithromycin sputum cultures turned negative at 0.5

years in two, one year in two more, and at two years in another, and these

patients continued to have negative sputum cultures up to four years of

treatment. The remaining 5 patients had continually positive sputum cultures, but

with fewer colony forming units (Kadota, 2003). S. pneumoniae that was

resistant to erythromycin was isolated from one of these five at one and two

Page 64: NEJM Protocol Sample 1

63

years of treatment, but the organism was replaced by S. aureus during the next

follow-up period. No patient was hospitalized with acute exacerbation of infection

during the four years of treatment (Kadota, 2003). In 73 children receiving

clarithromycin for eight weeks or longer for chronic otitis media with effusion

and/or sinusitis no change was found in the carriage of drug-resistant pathogens,

and the authors concluded the medication worked as a result of its anti-

inflammatory rather than its antibacterial effects (Iino, 2003).

We will obtain nasopharyngeal swabs and, when possible, expectorated sputum

quarterly, and will do sensitivity testing on any pathologic organisms cultured.

This will allow us to determine the emergence rate of macrolide-resistance

organisms prospectively in both macrolide-treated patients and controls.

Macrolide-related side-effects

Discussed above.

14. Medication compliance

A recent review of medication compliance indicates that the rate of

nonadherence to a prescribed medical regimen ranges from 0% to 100% (20% to

80% in patients with asthma), averages between 50% and 80%, is not related to

sociodemographic factors, intelligence or education but decreases as the

frequency of dosing, cost and duration of treatment increase (Claxton, 2001;

MacDonald, 2002).

Page 65: NEJM Protocol Sample 1

64

Although an extensive literature addresses issues relating to compliance and

ways to improve it in patients with hypertension, diabetes, epilepsy, psychiatric

disorders and, to a lesser extent, asthma and transplantation, information on

patients with COPD is limited. On the basis of responses to a questionnaire

James and colleagues (1985) found that only 33% of 185 patients with COPD

were fully compliant with their prescribed treatment regimen and compliance was

not related to the number of medications prescribed. Gallefoss and Bakke

(1999) performed a randomized controlled trial of a standardized education

program and self-management plan in patients with mild to moderate asthma (N

= 78) and COPD (N = 62) assessing compliance from monthly printouts of

pharmacy registers. The education program had no effect on compliance with

inhaled steroids in the patients with COPD (58% and 50% compliance at one

year in the control and intervention group, respectively) but did reduce the

amount of short-acting inhaled β2-agonists dispensed.

The ability of a macrolide antibiotic to reduce the frequency or severity of

acute COPD exacerbations can be determined by assessing its efficacy or its

effectiveness. Efficacy studies determine the benefit the drug under optimal

circumstances with considerable effort directed at obtaining maximum adherence

to the regimen and careful monitoring of compliance so that the effects of poor

compliance can be distinguished from the regimen being ineffective.

Effectiveness studies are designed to test the effect of the intervention under

conditions of standard practice in which compliance is not formally monitored.

Effectiveness studies have a greater likelihood of demonstrating that the

Page 66: NEJM Protocol Sample 1

65

intervention being tested is beneficial but the results may not extrapolate to what

occurs under normal daily practice.

In response to the Protocol Review Committee's suggestions to increase the

extent of compliance monitoring, to change the dosing regimen and educate the

patients to facilitate compliance, we propose the following:

Compliance monitoring.

Compliance with a medical regimen can be quantified by patient self-

assessment, pill counting, inspection of pharmacy databases, checking blood or

urine samples for the presence of a medication or a marker substance, and/or by

electronic monitoring of dispenser use. Unfortunately, is no “gold standard” for

measuring compliance because (1) all of the current approaches have important

limitations, and (2) there are varying recommendations regarding what defines

“good” and “bad” compliance and little experimental data to support any of the

recommendations (Timmreck, 1993; Farmer, 1999).

Patient self-assessment of compliance is easy to record, but the accuracy of

the information is questionable. Patients who admit to not following prescribed

regimens tend to describe their compliance accurately (Cramer, 1987) while

those who claim good compliance may be consciously falsifying their response or

have inaccurate recall (Spector, 1986).

Pill counting is also easy to accomplish, but also generally overestimates

compliance. Finding excessive numbers of pills clearly indicates poor

Page 67: NEJM Protocol Sample 1

66

compliance, but finding the appropriate number of pills does not confirm that the

medication was actually taken, or that it was taken regularly (Matsui, 1994).

Similar problems apply to inspection of pharmacy databases. While finding

that patients do not refill prescriptions at the appropriate rates indicates poor

compliance, the opposite may not true. In addition, the data may be inaccurate if

patients use more than one pharmacy.

Electronic monitoring systems utilize microprocessors to record the time and

date that a medication container was opened or when a blister pack was

accessed. Although these are considered to be the best single method of

compliance assessment they are expensive and still have limitations. As with the

other methods, data indicating poor compliance are accurate, but patients can

still remove their medications regularly but fail to ingest them. Systems that

assess pill bottle or blister pack opening have limited utility if patients use daily

pill containers that are filled once weekly.

Measuring blood or urine levels of a medication, or of a biological marker that

can be added to a medication, is expensive, and will identify whether patients

have recently taken recently taken a dose. These methods cannot confirm

regular use and, depending on the medication being monitored, the tests can

partially depend on factors that are unrelated to compliance (e.g., diet, absorption

and rate of excretion) (Vitolins, 2000).

Finally, all monitoring methods are susceptible to the Hawthorne effect, i.e.,

compliance being improved because the patients know that their drug use is

being monitored.

Page 68: NEJM Protocol Sample 1

67

No single measurement strategy has won uniform approval as a

representative 'gold standard' and, accordingly, a multi-method approach that

combines self-reporting and other objective measures is generally recommended

(Sabate, 2003). We have elected to monitor compliance by self-reporting, pill

counting, and blister packs containing microprocessors (MedicTM ECMTM).

Human Subject Issues

Risks and measures designed to minimize the risks. Discussed above.

Procedure for obtaining consent. Described above.

Enrollment of minorities and women

Minority patients

The prevalence of at least moderate COPD (defined by spirometry), the

number of COPD-related office visits and the COPD-related death rate are lower

in minority patients than in Whites, but Blacks have a higher prevalence of ED

visits and hospitalization (Table 6).

Table 6. Prevalence, Morbidity and Mortality of COPD by Race (Centers of

Disease Control and Prevention, 2002)

Black White Other

Prevalence of Moderate COPD by

Spirometry (per 1,000)

55.8 67.1 52.8

Office Visits for COPD (per 1,000) 30.4 46.9

Page 69: NEJM Protocol Sample 1

68

ED Visits (per 10,000) 130.8 84.8

Hospitalizations (per 10,000) 36 31.5 21.3

Death rates (per 100,000) 42.9 70.1 30.6

Data describing the prevalence, morbidity and mortality in Hispanics are both

very limited and over 20 years old.

The ability to compare morbidity and mortality between races is limited,

however, because there are numerous potential confounders of these data.

The observed minority visits in seven clinical centers is summarized in Table

7 (prior to completing the protocol three additional centers were added to the

original nine to facilitate patient enrollment).

Table 7. The percentages of women and minority patients seen in seven Clinical

Centers

Center Hospital Women

(%)

Black Hispanic American

Indian/

Alaska

Native

Asian/

Pacific

Islander

White

Baltimore University inpt* 58 51 0 0 0 45

University outpt* 53 75 0 0 0 22

VA* 6 35 0 0 0 51

Birmingham University 50 30 0 0 0 70

VA

Page 70: NEJM Protocol Sample 1

69

Boston Brigham/Womens* 56 12 5 0 81

MGH* 49 3 3 0 0 92

BDMC 52 8 2 0 0 85

Denver Denver Health*1 55 14 55 1 0 5

National Jewish*1 50 4 3 0 1 80

University*1 50 8 7 4 51

VA*1 10 4 6 0 30 58

Los Angles Harbor/UCLA* 30 15 4 0 0 74

Minneapolis Health Partners4 50 7

Mayo Clinic 50

VA 10

San

Francisco

Moffit-Long* 55 10 8 13 60

Mean 43 21 7 0 4 60

* Data pertain to COPD patients specifically

1 Combined in-patient and out-patient

Previous studies indicate that fewer Black and Hispanic patients agree to

participate in clinical studies. Accordingly, we estimate that despite 21% of the

patients seen in the seven Clinical Centers being Black, Blacks may represent

only 12% of the study population after enrollment is complete. Similarly, while

Page 71: NEJM Protocol Sample 1

70

7% of the patients in the Clinical Centers are Hispanic, perhaps only 4 or 5% of

the final study population will be Hispanic.

Our target minority enrollment estimate is that the final study population will

include 15% minority patients. This would seem to be a realistic objective given

the percentage of minority patients seen in seven Clinical Centers after adjusting

for lower rates of enrollment in minority patients.

Women

Although women have a higher rate of self-reported COPD than men, and the

prevalence of COPD is increasing in women but not in men, there is a higher

prevalence of moderate COPD (defined by spirometry) in men, and men have a

higher death rate (Table 8, Centers of Disease Control and Prevention, 2002).

Table 8. Prevalence, Morbidity and Mortality of COPD by Gender (MMWR,

2002)

Men Women

Prevalence of Moderate COPD by

Spirometry (per 1,000)

74.3 58.2

Office Visits for COPD (per 1,000) 46.8 43.4

ED Visits (per 10,000) 80.7 94.4

Hospitalizations (per 10,000) 42.4 40.2

Death rates (per 100,000) 82.6 56.7

Page 72: NEJM Protocol Sample 1

71

The well documented spirometric progression of COPD, together with the fact

that the annual COPD-related death rate is increasing in women but not in men,

suggests that, on the average, the severity of COPD in women is less than in

men. This difference could affect enrollment in the proposed study. The higher

prevalence of ED visits in women argues against this suggestion, however.

Accordingly, our target enrollment estimate is that the final study population will

include 45% women.

The large male predominance in the armed services results in the fact that

patients in any study that has Veteran's hospitals as enrollment sites will be

disproportionately male unless special enrollment strategies are employed to

assure a more equal gender distribution. Seven of the ten Clinical Centers in the

COPD CRN include VA hospitals as enrollment sites. Accordingly, strategies

specifically targeting enrollment of women from the other sites will be will be

needed for this study. Plans for accomplishing this at each Center were

developed prior to starting the study.

Page 73: NEJM Protocol Sample 1

72

Section B. Protocol Changes Pertaining to Possible QT Prolongation During the trial two protocol changes were made:

(1) Because only 15.4%, 11.2% and 9.4% of participants were able to

produce sputum at enrollment and at the one- and three-month clinic visits,

respectively, we elected to stop collecting expectorated sputum after month

three.

(2) Fifteen months after enrollment began the heart rate exclusion criteria

and the methods used to determine the QTc interval were modified in response

to a new publication suggesting that clarithromycin was associated with an

increased risk of cardiovascular mortality (Jespersen, 2006). New exclusion

criteria included:

a. A resting heart rate > 100 beats/min

b. A manually determined QTc interval measured at least one hour

after use of any short-acting inhaled β2 agonist that exceeds 450 ms on two

occasions separated by at least one week. Manual determinations are only to be

done if the automated QTc estimate exceeds the 450 ms limit.

The method for determining the QTc interval manually is to measure the

mean QT interval from the beginning of the QRS complex to the end of the T

wave, in a minimum of 3 cardiac cycles, in leads II and V5 or V6 (using the lead

in which the QT is the longest). This QT interval is then corrected for heart rate

using Federicia’s formula (QTc= QT/RR(sec)1/3).

Bundle branch blocks prolong the QT interval. Accordingly, patients with

either left or right bundle branch blocks should have the JT interval measured.

Page 74: NEJM Protocol Sample 1

73

Subjects with a JT interval > 420 msec (corrected for heart rate) should be

excluded.

It is recommended that a cardiologist specializing in arrhythmias and/or

ECG interpretation perform the above manual analysis and evaluate all ECGs of

patients with bundle branch blocks.

In addition, two additional reasons for drug discontinuation were added:

a. If the QTc on ECGs recorded one month after starting the study

medication increases to > 500 ms or > 60 ms from the patient’s baseline value.

b. For patients with a bundle branch block, if the JTc interval

increases 60 ms from the patient’s baseline value.

Page 75: NEJM Protocol Sample 1

74

References

Abbott PJ et al.Retrospective analyses of additional services for

methadonemaintenance patients. Journal of Substance Abuse Treatment,

1999;17:129–137.

Aberg JA, Wong MK, Flamm R, Notario GF, Jacobson MA. 2001. Presence of

macrolide resistance in respiratory flora of HIV infected patients receiving

either clarithromycin or azithromycin for Mycobacterium avium complex

prophylaxis. HIV Clin Trials 2:453-9.

Albaz RS. Factors affecting patient compliance in Saudi Arabia. Journal of Social

Sciences, 1997, 25:5-8.

American Thoracic Society. 1987. Standards for the diagnosis and care of

patients with chronic obstructive pulmonary disease (COPD) and asthma.

Am Rev Respir Dis 136:225-44.

Anthonisen NR, Wright EC, IPPB Trial Group. 1986. Bronchodilator response in

chronic obstructive pulmonary disease. Am Rev Respir Dis 133:814-9.

Brummett RE. 1993. Ototoxic liability of erythromycin and analogues.

Otolaryngol Clin North Am 26:811-819.

Calverley PMA, Burge PS, Spencer S, Anderson JA, Jones PW, for the ISOLDE

Study Investigators. 2003. Bronchodilator reversibility testing in chronic

obstructive pulmonary disease. Thorax 58:659-64.

Centers of Disease Control and Prevention. 2002. Surveillance Summaries,

August 2, MMWR 51:(No. SS-6).

Page 76: NEJM Protocol Sample 1

75

Claxton AJ, Cramer J, Pierce C. 2001. A systematic review of the associations

between dose regimens and medication compliance. Clin Ther 23:1296-

1310.

Cramer JA, Mattson RH, Prevey ML, Scheyer RD, Ouellette VL. 1989. How

often is medication taken as prescribed. JAMA 261:3273-7.

Cymbala AA, Edmonds LC, Bauer MA, et al. An open label, randomised,

crossover design pilot study of the disease modifying effects of twice

weekly oral azithromycin in patients with bronchiectasis. Treat Resp Med,

in press pending revision.

Doern GV, Heilmann KP, Huynh HK, et al. Antimicrobial resistance among

clinical isolates of Streptococcus pneumoniae in the United States during

1999-2000, including a comparison of resistance rates since 19994-1995.

Antimicrob Agents Chemother 45:1721-9.

Equi A, Balfour-Lynn IM, Bush A, Rosenthal M. 2002. Long-term azithromycin in

children with cystic fibrosis: a randomized, placebo-controlled crossover

trial. Lancet 360:978-84.

Ewig S, Ruiz M, Torres A, et al. 1999. Pneumonia acquired in the community

through drug resistant Streptococcus pneumoniae. Am J Respir Crit Care

Med 159:1835-42.

Fabbri LM, Romagnoli M, Corbetta L, Casoni G, Busljetic K, Turato G, et al.

2003. Differences in airway inflammation in patients with fixed airflow

obstruction due to asthma or chronic obstructive pulmonary disease. Am J

Respir Crit Care Med. 167:418–424

Page 77: NEJM Protocol Sample 1

76

Farmer KC. 1999. Methods for measuring and monitoring medication regimen

adherence in clinical trials and clinical practice. Clinical Therapeutics

21:1074-1090.

Ferguson, GT, Enright, PL, Buist, AS, et al Office spirometry for lung health

assessment in adults: a consensus statement from the National Lung

Health Education Program. Chest 2000;117,1146-1161

Fogarty C, Goldschmidt R, Bush K. 2002. Bacteremic pneumonia due to

multidrug-resistant pneumococci in 3 patients treated unsuccessfully with

azithromycin and successfully with levofloxacin. Clin Infect Dis 31:613-5.

Gallefoss F, Bakke PS. 1999. How does patient education and self-

management among asthmatics and patients with chronic obstructive

pulmonary disease affect medication? Am J Respir Crit Care Med 160:

2000–5.

Gay K, Baughman W, Miller Y, et al. 2000. The emergence of Streptococcus

pneumonia resistant to macrolide antimicrobial agents: a 6-year

population-based assessment. J Infect Dis 182;1417-24.

Gearhardt SG, McDyer JF, Girgis RE, Conte JV, Yange SC, Orens JB. 2003.

Maintenance azithromycin therapy for bronchiolitis obliterans syndrome.

Results of a pilot study. Am J Respir Crit Care Med 168:121-5.

Gleason PP, Meehan TP, Fine JM, Galusha DH, Fine MJ. 1999. Associations

between initial antimicrobial therapy and medical outcomes for

hospitalized elderly patients with pneumonia. Arch Intern Med 159:2562-

72.

Page 78: NEJM Protocol Sample 1

77

Guchev IA, Gray GC, Klochkov OI. 2004. Two regimens of azithromycin

prophylaxis against community-acquired respiratory and skin/soft tissue

infections among military trainees. CID 38:1095-1101.

Hankinson, JL, Crapo, RO, Jensen, RL Spirometric reference values for the 6-s

FVC maneuver. Chest 2003;124,1805-1811

Halperin M, Lan KKG, Ware JH et al. (1982). And aid to data monitoring in long-

term clinical trials. Controlled Clin Trials 3:311-323.

Harris S, Hilligoss DM, Colangeol PM, Eller M, Okerhaom R. 1995.

Azithromycin and terfenadine: lack of drug interaction. Clin Pharmacol

Ther 58:310-5.

Haydon RC, Thelin JW, Davis WE. 1984. Erythromycin ototoxicity: analysis and

conclusions based on 22 case reports. Otolaryngol Head Neck Surg

92:678-684.

Hoban DJ, Doern GV, Fluit AC, Roussel-Delvallez M, Jones RN. 2001.

Worldwide prevalence of antimicrobial resistance in Streptococcus

pneumonia, Haemophilus influenzae, and Moraxella catarrhalis in the

SENTRY Antimicrobial Surveillance Program, 1997-1999. Clin Inf Dis

32:S81-S93.

Hoffman HL, Klepser ME, Ernst EJ, Petzold CR, Sa'adah LM, Doern DV. 2003.

Influence of macrolide susceptibility on efficacies of clarithromycin and

azithromycin against Streptococcus pneumoniae in a murine lung infection

model. Antimicrob Agents Chemother 47:739-46.

Page 79: NEJM Protocol Sample 1

78

Hyde TB, Gay K, Stephens DS, et al. 2001. Macrolide resistance among

invasive Streptococcus pneumoniae isolates. JAMA 286:1857-62.

Iannini PB. 2002. Cardiotoxicity of macrolides, ketolides and fluoroquinolones

that prolong the QTc interval. Expert Opin Drug Saf 1:121-128.

Iino Y, Sasaki Y, Miyazawa T, Kodera K. 2003. Nosopharyngeal flora and drug

susceptibility in children with macrolide therapy. Laryngoscope 113:1780-

5.

James PN, Anderson JB, Prior JG, White JP, Henry JA, Cochrane GM. 1985.

Patterns of drug taking in patients with chronic airflow limitation. Postgrad

Med J 61:7-10.

Jennison C, Turnbull BW (2000) Group Sequential Methods with Applications to

Clinical Trials. Chapman & Hall/CRC, Chap. 7, Table 7.1.

Jespersen CM, Als-Nielsen B, Damgaard M, Hansen JF, Hansen S, Helo OH,

Hildebrandt P, Hilden J, Hensen GB, Kolmos HJ, Keller E, Lind I, Nikelsen

J, Petersen L, Gluud C, the CLARICOR Trial Group. Randomized placebo

controlled multicenter trial to assess short term clarithromycin for patients

with stable coronary heart disease: CLARICOR trial. BMJ 2006;332:22-27

Jones RN, Cormican MC, Wanger A. 1996. Clindamycin resistance among

erythromycin-resistant Streptococcus pneumoniae. Diagn Microbiol Infect

Dis 25:201-4.

Kadota J, Mukae H, Kshii H, Nagata T, Kaida H, Tomono K, Kohno S. 2003.

Long-term efficacy and safety of clarithromycin treatment inpatients with

diffuse panbronchiolitis. Resp Med 97:844-50.

Page 80: NEJM Protocol Sample 1

79

Kelley MA, Weber DJ, Gilligan P, Cohen MS. 2000. Breakthrough

pneumococcal bacteremia in patients treated with azithromycin and

clarithromycin. Clin Infect Dis 31:1008-11.

Kim K, DeMets DL (1987) Design and analysis of group sequential tests based

on the type I error spending rate function. Biometrika 74, 149-154.

Lan KKG, Simon R, Halperin M.(1982) Stochastically curtailed tests in long-term

clinical trials. Commun. Stat., Sequential Analysis. 1:207-219.

Leclercq R. 2002. Mechanisms of resistance to macrolides and lincosamides:

nature of the resistance elements and their clinical implications. Clin Infect

Dis 34:482-492.

Levey AS; Bosch JP; Lewis JB; Greene T; Rogers N; Roth D. 1999. A more

accurate method to estimate glomerular filtration rate from serum

creatinine: a new prediction equation. Modification of Diet in Renal

Disease Study Group. Ann Intern Med 130:461-70.

Levey, AS, Greene, T, Kusek, JW, Beck, GJ, and MDRD study group. 2000. A

simplified equation to predict glomerular filtration rate from serum

creatinine (abstract). J Am Soc Nephrol 11:155A.

Lisby G, Brasholt MS, Teglbjerg L. 2001. Bacteremia and meningitis caused by

macrolide-sensitive strain of Streptococcus pneumoniae during treatment

with azithromycin. Clin Infect Dis 33:415-6.

Lonks JR, Garau J, Gomez L. 2002. Failure of macrolide antibiotic treatment in

patients with bacteremia due to erythromycin-resistant Streptococcus

pneumoniae. Clin Infect Dis 35:556-64.

Page 81: NEJM Protocol Sample 1

80

Lynch JP, Martinez FJ. 2002. Clinical relevance of macrolide-resistant

Streptococcus pneumoniae for community-acquired pneumonia. Clin

Infect Dis 34 (Suppl 1):S27-46).

MacDonald HP, Garg AX, Haynes RB. 2002. Interventions to enhance patient

adherence to medication prescriptions. JAMA 288:2868-79.

Malazimoglu L, Tulkens PM, Van Bambek R. Macrolides. In: VL Yu, R Weber, D

Raoult (eds), Antimicrobial Therapy and Vaccines (Volume II), in press.

Martinez JA, Horcajada JP, Almela M, et al. Addition of a macrolide to a β-

lactam-based empirical antibiotic regimen is associated with lower in-

hospital mortality for patients with bacteremic pneumococcal pneumonia.

Clin Inf Dis 36:389-95.

Matsui D et al. 1994. Critical comparison of novel and existing methods of

compliance assessment during a clinical trial of an oral iron chelator.

Journal of Clinical Pharmacology 34:944-949.

Milberg P, Eckardt L, Bruns W, et al. 2002. Divergent proarrhythmic potential of

macrolide antibiotics despite similar QT prolongation: fast phase 3

repolarization prevents early after depolarizations and torsade de pointes.

J Pharmacol Exp Ther 303;218-25.

Musher DM, Dowel ME, Shortridge VD. 2002. Emergence of macrolide

resistance during treatment of pneumococcal pneumonia. N Engl J Med

346:630-1.

Page 82: NEJM Protocol Sample 1

81

National Emphysema Treatment Trial Research Group. Cost effectiveness of

lung-volume-reduction surgery for patients with severe emphysema. N

Engl J Med 2003;348:2092-102.

NICE Guidelines on COPD. 2004. National clinical guideline on management of

COPD in adults in primary and secondary care. Diagnosing COPD.

Thorax 59 (Suppl 1):i27-38.

Niewoehner D, Rice K, Cote C, Paulson D, Cooper JA, Korducki L, Cassimo C,

Kesten S. 2004. Reduced COPD exacerbations and associated health

care utilization with once-daily tiotropium (TIO) in the VA Medical System.

Am J Respir Crit Care Med 169: A207 (abstract).

Nisar M, Earis JE, Pearson MG, Calverley PM. 1992. Acute bronchodilator trials

in chronic obstructive pulmonary disease. Am Rev Respir Dis 146:555-9.

Nishijima T, Saito Y, Aoki A, Toriya M, Toyonaga Y, Fujii R. 1999. Distribution of

mefE and ermB genes in macrolide-resistant strains of Streptococcus

pneumoniae and their variable susceptibility to various antibiotics. J

Antimicrob Chemother 43:637-43.

Nuermberger E, Bishai WR. 2004. The clinical significance of macrolide-

resistnat Streptococcus pneumoniae: It’s all relative. CID 38:99-103.

O’Brien PC, Fleming TR (1979). A multiple testing procedure for clinical trials.

Biometrics 35:549-556.

O'Connor CM, Dunne MW, Pfeffer MA, Muhlestein JB, Yao L, Gputa S, Benner

RJ, Fisher MR, Coor TD for the investigators in the WIZARD study. 2003.

Azithromycin for the secondary prevention of coronary heart disease

Page 83: NEJM Protocol Sample 1

82

events. The WIZARD study: A randomized controlled trial. JAMA

290:1459-66.

Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS; The GOLD

Scientific Committee. 2001. Global strategy for the diagnosis,

management, and prevention of chronic obstructive pulmonary disease.

NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease

(GOLD) Workshop summary. Am J Respir Crit Care Med 163:1256-76.

Poehlsgaard J, Douthwaite S. 2003. Macrolide antibiotic interaction and

resistance on the bacterial ribosome. Curr Opin Investig Drugs 4:140-148.

Rello J, Catalan M, Diaz E, Bodi M, Alvarez B. 2002. Associations between

empirical antimicrobial therapy at the hospital and mortality in patients with

severe community-acquired pneumonia. Inten Care Med 28:1003-5.

Roter DL et al. Effectiveness of interventions to improve patient compliance: a

meta-analysis. Medical Care, 1998, 36:1138-1161.

Sabate E. 2003. Adherence to Long-Term Therapies. Evidence for Action.

World Health Organization, Geneva.

Saiman L, Marshall BC, Mayer-Hamblett N, Burns JL, Quittner AL, Cibene DA,

Coquillette S, Fieberg AY, Accurso FJ, Campbell PW, III. 2003.

Azithromycin in patients with cystic fibrosis chronically infected with

Pseudomonas aeruginosa: a randomized controlled trial. JAMA 290:1749-

1756.

Page 84: NEJM Protocol Sample 1

83

Santus P, Pecchiari M, Carlucci P, Boveri B, Di Marco F, Castagna F, Centanni

S. 2003. Bronchodilation test in COPD: effect of inspiratory manoeuvre

preceding forced expiration. Eur Respir J 21:82-5.

Sin DD, McAlister FA, Man SF, Anthonisen NR. 2003. Contemporary

management of chronic obstructive pulmonary disease: scientific review.

JAMA 290(17):2301-12.

Spector SL et al. 1986. Compliance of patients with asthma with an

experimental aerosolized medication: implications for controlled clinical

trials. Journal of Allergy & Clinical Immunology 77:65-70.

Stahl JE, Barza M, DesJardin J, Martin R, Eckman MH. 1999. Effect of

macrolides as part of initial empiric therapy on length of stay in patients

hospitalized with community-acquired pneumonia. Arch Intern Med

159:2576-80.

Strle R, Maraspin V. 2002. Wien Kiln Wochenschr 112:396-9.

Swanson DJ, Sung RJ, Fine MJ, Orloff JJ, Chu SY, Yu VL. 1992. Erythromycin

ototoxicity: prospective assessment with serum concentrations and

audiograms in a study of patients with pneumonia. Am J Med 92:61-68.

Taylor R, Schofield IS, Ramos JM, Bint AJ, Ward MK. 1981. Ototoxicity of

erythromycin in peritoneal dialysis patients. Lancet 2:935-936.

Tessier PR, Kim MK, Zhou W, Xuan D, Li C, Ye M, Nightingale CH, Nicolau DP.

2002. Pharmacodynamic assessment of clarithromycin in a murine model

of pneumococcal pneumonia. Antibmcrob Agents Chemother 46:1425-34.

Page 85: NEJM Protocol Sample 1

84

Timmreck TC, Randolph JF. 1993. Smoking cessation: clinical steps to improve

compliance. Geriatrics 48:63-66.

Tweeddale PM, Alexander F, McHardy GJ. 1987. Short term variability in FEV1

and bronchodilator responsiveness in patients with obstructive ventilatory

defects. Thorax 42:487-90.

Umstead GS, Neumann KH. 1986. Erythromycin ototoxicity and acute psychotic

reaction in cancer patients with hepatic dysfunction. Arch Intern Med

146:897-899.

Van Bambeke F, Glupczynski Y, Plesiat P, Pechere JC, Tulkens PM. 2003.

Antibiotic efflux pumps in prokaryotic cells: occurrence, impact on

resistance and strategies for the future of antimicrobial therapy. J

Antimicrob Chemother 51:1055-1065.

Van Kerkhoven D, Peetermans WE, Verbist L, Verhaegen J. 2003.

Breathhrough pneumococcal bacteraemia in patients treated with

clarithromycin or oral β-lactams. J Antimicrob Chemother 51:691-6.

Vasquez EM, Maddux MS, Sanchez J, Pollak R. 1993. Clinically significant

hearing loss in renal allograft recipients treated with intravenous

erythromycin. Arch Intern Med 153:879-882.

Vitolins MZ et al. 2000. Measuring adherence to behavioral and medical

interventions. Controlled Clinical Trials, 2000, 21:188S-194S.

Waites K, Johnson C, Gray B, Edwards K, Crain M, Benjamin W Jr. (2000) Use

of clinicamycin disks to detect macrolide resistance mediated by ermB and

Page 86: NEJM Protocol Sample 1

85

mefE in Streprococcus pneumoniae isolates from adults and children. J

Clin Microbiol 38:1731-4.

Wallace MR, Miller LK, Nguyen MT, Shields AR. 1994. Ototoxicity with

azithromycin. Lancet 343:241.

Ware JH, Muller JE, Braunwald E. (1985) The futility index, a approach to the

cost-effective termination of randomized clinical trials. Am J Cardiol

78:636-664.

Waterer GW, Wunderink RG, Jones CB. 2000. Fatal pneumococcal pneumonia

attributed to macrolide resistance and azithromycin monotherapy. Chest

118:1839-40.

Waterer GW, Somes GW, Wunderink RG. 2001. Monotherapy may be

suboptimal for severe bacteremic pneumococcal pneumonia. Arch Intern

Med 161:1837-42.

Whitney CG, Farley MM, Hadler J, Harrison LH, Lexau C, Reingold A, Lefkowitz

L, Cieslak PR, Cetron M, Zell ER, Jorgensen JH, Schuchat A. Active

Bacterial Core Surveillance Program of the Emerging Infections Program.

2000. Increasing prevalence of multidrug-resistant Streptococcus

pneumoniae in the United States. N Engl J Med 343:1917-24.

Weisblum B. 1995. Erythromycin resistance by ribosome modification.

Antimicrob Agents Chemother 39:577-585.

Wolter J, Seeney S, Bell S, Bowler S, Masel P, McCormack J. 2002. Effect of

long term treatment with azithromycin on disease parameters in cystic

fibrosis: a randomized trial. Thorax 57:212-6.